Tabulating SPost results

The SPost package by Scott Long and Jeremy Freese is a suite of post-estimation commands used to compute additional tests and effects representations for a variety of regression models (see http://www.indiana.edu/~jslsoc/spost.htm). To facilitate and automate the task of processing result from SPost for inclusion in reports and publications, estadd provides tools to integrate SPost results with estout or esttab.

Supported commands are brant, fitstat, listcoef, mlogtest, prchange, prvalue, and asprvalue from SPost for Stata 9 or newer (spost9_ado). SPost for Stata 8 (spostado) is not supported. See the SPost section in estadd's documentation for further details. Below is a range of examples covering various models and applications.

Introductory examples

From SPost to esttab/estout

The general procedure to tabulate results from an SPost command in esttab or estout is to

  1. fit one or more models,
  2. use estadd to apply the SPost command and add the results to the models' e()-returns, and
  3. include the added returns in the the main(), aux(), and scalars() options of esttab or the cells() and stats() options of estout.

For example, to tabulate a number of fitstat information measures for a linear regression model, type:

. spex regjob2
(Academic Biochemists / S Long)

. regress job fem phd ment fel art cit

      Source |       SS       df       MS              Number of obs =     408
-------------+------------------------------           F(  6,   401) =   17.78
       Model |  81.0584763     6  13.5097461           Prob > F      =  0.0000
    Residual |  304.737915   401  .759944926           R-squared     =  0.2101
-------------+------------------------------           Adj R-squared =  0.1983
       Total |  385.796392   407  .947902683           Root MSE      =  .87175

------------------------------------------------------------------------------
         job |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         fem |  -.1391939   .0902344    -1.54   0.124    -.3165856    .0381977
         phd |   .2726826   .0493183     5.53   0.000     .1757278    .3696375
        ment |   .0011867   .0007012     1.69   0.091    -.0001917    .0025651
         fel |   .2341384   .0948206     2.47   0.014     .0477308    .4205461
         art |   .0228011   .0288843     0.79   0.430    -.0339824    .0795846
         cit |   .0044788   .0019687     2.28   0.023     .0006087     .008349
       _cons |   1.067184   .1661357     6.42   0.000     .7405785     1.39379
------------------------------------------------------------------------------

. estadd fitstat, bic
AIC:                             2.580   AIC*n:                        1052.793
BIC:                         -1371.725   BIC':                          -60.162
BIC used by Stata:            1080.872   AIC used by Stata:            1052.793

added scalars:
               e(aic0) =  2.5803757
              e(aic_n) =  1052.7933
               e(bic0) =  -1371.7248
              e(bic_p) =  -60.162312
           e(statabic) =  1080.8722
           e(stataaic) =  1052.7933

. esttab, cells(none) scalars(aic0 aic_n bic0 bic_p)

-------------------------
                      (1)
                      job
-------------------------
N                     408
aic0                2.580
aic_n              1052.8
bic0              -1371.7
bic_p              -60.16
-------------------------

[do-file]

If you are working with multiple models you can either add results to each model individually after estimation (as above), or you can first estimate and store a set of models and then apply estadd to all of them in one call using the colon syntax. Here is an example of the latter, using eststo to store the models:

. spex regjob2
(Academic Biochemists / S Long)

. eststo: regress job fem phd ment

      Source |       SS       df       MS              Number of obs =     408
-------------+------------------------------           F(  3,   404) =   23.77
       Model |  57.8903644     3  19.2967881           Prob > F      =  0.0000
    Residual |  327.906027   404  .811648583           R-squared     =  0.1501
-------------+------------------------------           Adj R-squared =  0.1437
       Total |  385.796392   407  .947902683           Root MSE      =  .90092

------------------------------------------------------------------------------
         job |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         fem |  -.1769641   .0915984    -1.93   0.054    -.3570331     .003105
         phd |   .3307536   .0495896     6.67   0.000     .2332678    .4282395
        ment |   .0015841   .0007207     2.20   0.029     .0001673    .0030009
       _cons |   1.171768   .1635376     7.17   0.000     .8502769    1.493259
------------------------------------------------------------------------------
(est1 stored)

. eststo: regress job fem phd ment fel art cit

      Source |       SS       df       MS              Number of obs =     408
-------------+------------------------------           F(  6,   401) =   17.78
       Model |  81.0584763     6  13.5097461           Prob > F      =  0.0000
    Residual |  304.737915   401  .759944926           R-squared     =  0.2101
-------------+------------------------------           Adj R-squared =  0.1983
       Total |  385.796392   407  .947902683           Root MSE      =  .87175

------------------------------------------------------------------------------
         job |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         fem |  -.1391939   .0902344    -1.54   0.124    -.3165856    .0381977
         phd |   .2726826   .0493183     5.53   0.000     .1757278    .3696375
        ment |   .0011867   .0007012     1.69   0.091    -.0001917    .0025651
         fel |   .2341384   .0948206     2.47   0.014     .0477308    .4205461
         art |   .0228011   .0288843     0.79   0.430    -.0339824    .0795846
         cit |   .0044788   .0019687     2.28   0.023     .0006087     .008349
       _cons |   1.067184   .1661357     6.42   0.000     .7405785     1.39379
------------------------------------------------------------------------------
(est2 stored)

. estadd fitstat, bic: *

. esttab, cells(none) scalars(aic0 aic_n bic0 bic_p)

--------------------------------------
                      (1)          (2)
                      job          job
--------------------------------------
N                     408          408
aic0                2.639        2.580
aic_n              1076.7       1052.8
bic0              -1359.9      -1371.7
bic_p              -48.30       -60.16
--------------------------------------

. eststo clear

[do-file]

A key difference between the two approaches is that with the first method output from estadd fitstat is displayed, whereas execution with the second syntax is silent.

Tabulating results from prchange

The default for estadd prchange is to return a matrix called e(dc) containing the 0 to 1 change effects for binary variables and the standard deviation change effects for continuous variables in the first row, followed by additional rows containing separate results for the different effect types computed by prchange. To tabulate the contents of the first row simply refer to dc in esttab or estout. Example:

. spex binlfp2
(Data from 1976 PSID-T Mroz)

. quietly logit lfp k5 k618 age wc hc lwg inc, nolog

. estadd prchange

logit: Changes in Probabilities for lfp

      min->max      0->1     -+1/2    -+sd/2  MargEfct
  k5   -0.6361   -0.3499   -0.3428   -0.1849   -0.3569
k618   -0.1278   -0.0156   -0.0158   -0.0208   -0.0158
 age   -0.4372   -0.0030   -0.0153   -0.1232   -0.0153
  wc    0.1881    0.1881    0.1945    0.0884    0.1969
  hc    0.0272    0.0272    0.0273    0.0133    0.0273
 lwg    0.6624    0.1499    0.1465    0.0865    0.1475
 inc   -0.6415   -0.0068   -0.0084   -0.0975   -0.0084

         NotInLF     inLF
Pr(y|x)   0.4222   0.5778

            k5     k618      age       wc       hc      lwg      inc
   x=  .237716  1.35325  42.5378  .281541  .391766  1.09711   20.129
sd_x=  .523959  1.31987  8.07257  .450049  .488469  .587556  11.6348

added scalars:
            e(predval) =  .57779419
              e(delta) =  1
           e(centered) =  1

added matrices:
                 e(dc) :  6 x 7      (main, min->max, 0->1, -+1/2, -+sd/2, Marg
> Efct)
            e(pattern) :  1 x 7
                  e(X) :  4 x 7      (X, SD, Min, Max)

first row in e(dc) contains:

  01 change for binary variables
  sd change for continuous variables

. esttab, aux(dc) nopar wide

-----------------------------------------
                      (1)                
                      lfp                
-----------------------------------------
lfp                                      
k5                 -1.463***       -0.185
k618              -0.0646         -0.0208
age               -0.0629***       -0.123
wc                  0.807***        0.188
hc                  0.112          0.0272
lwg                 0.605***       0.0865
inc               -0.0344***      -0.0975
_cons               3.182***             
-----------------------------------------
N                     753                
-----------------------------------------
dc in second column
* p<0.05, ** p<0.01, *** p<0.001

[do-file]

To change the defaults for the contents of the first row of e(dc) use the c() option (for continuous variables) and the b() option (for binary variables). For example, to tabulate the marginal effects for continuous variables and the 0 to 1 change effects for binary variables (see the helpfile for the list of available effects types), type:

. estadd prchange, c(margefct) replace

logit: Changes in Probabilities for lfp

      min->max      0->1     -+1/2    -+sd/2  MargEfct
  k5   -0.6361   -0.3499   -0.3428   -0.1849   -0.3569
k618   -0.1278   -0.0156   -0.0158   -0.0208   -0.0158
 age   -0.4372   -0.0030   -0.0153   -0.1232   -0.0153
  wc    0.1881    0.1881    0.1945    0.0884    0.1969
  hc    0.0272    0.0272    0.0273    0.0133    0.0273
 lwg    0.6624    0.1499    0.1465    0.0865    0.1475
 inc   -0.6415   -0.0068   -0.0084   -0.0975   -0.0084

         NotInLF     inLF
Pr(y|x)   0.4222   0.5778

            k5     k618      age       wc       hc      lwg      inc
   x=  .237716  1.35325  42.5378  .281541  .391766  1.09711   20.129
sd_x=  .523959  1.31987  8.07257  .450049  .488469  .587556  11.6348

added scalars:
            e(predval) =  .57779419
              e(delta) =  1
           e(centered) =  1

added matrices:
                 e(dc) :  6 x 7      (main, min->max, 0->1, -+1/2, -+sd/2, Marg
> Efct)
            e(pattern) :  1 x 7
                  e(X) :  4 x 7      (X, SD, Min, Max)

first row in e(dc) contains:

  01 change for binary variables
  margefct for continuous variables

. esttab, aux(dc) nopar wide

-----------------------------------------
                      (1)                
                      lfp                
-----------------------------------------
lfp                                      
k5                 -1.463***       -0.357
k618              -0.0646         -0.0158
age               -0.0629***      -0.0153
wc                  0.807***        0.188
hc                  0.112          0.0272
lwg                 0.605***        0.148
inc               -0.0344***     -0.00840
_cons               3.182***             
-----------------------------------------
N                     753                
-----------------------------------------
dc in second column
* p<0.05, ** p<0.01, *** p<0.001

[do-file]

Alternatively, if you want to tabulate the different effect types computed by prchange separately, address the rows in e(dc) using dc[#] where # is the row number or dc[name] where name is the row name. Example:

. spex binlfp2
(Data from 1976 PSID-T Mroz)

. quietly logit lfp k5 k618 age wc hc lwg inc, nolog

. estadd prchange

logit: Changes in Probabilities for lfp

      min->max      0->1     -+1/2    -+sd/2  MargEfct
  k5   -0.6361   -0.3499   -0.3428   -0.1849   -0.3569
k618   -0.1278   -0.0156   -0.0158   -0.0208   -0.0158
 age   -0.4372   -0.0030   -0.0153   -0.1232   -0.0153
  wc    0.1881    0.1881    0.1945    0.0884    0.1969
  hc    0.0272    0.0272    0.0273    0.0133    0.0273
 lwg    0.6624    0.1499    0.1465    0.0865    0.1475
 inc   -0.6415   -0.0068   -0.0084   -0.0975   -0.0084

         NotInLF     inLF
Pr(y|x)   0.4222   0.5778

            k5     k618      age       wc       hc      lwg      inc
   x=  .237716  1.35325  42.5378  .281541  .391766  1.09711   20.129
sd_x=  .523959  1.31987  8.07257  .450049  .488469  .587556  11.6348

added scalars:
            e(predval) =  .57779419
              e(delta) =  1
           e(centered) =  1

added matrices:
                 e(dc) :  6 x 7      (main, min->max, 0->1, -+1/2, -+sd/2, Marg
> Efct)
            e(pattern) :  1 x 7
                  e(X) :  4 x 7      (X, SD, Min, Max)

first row in e(dc) contains:

  01 change for binary variables
  sd change for continuous variables

. esttab, cells("dc[2] dc[3] dc[4] dc[5] dc[6]")

-----------------------------------------------------------------------------
                      (1)                                                    
                      lfp                                                    
                    dc[2]        dc[3]        dc[4]        dc[5]        dc[6]
-----------------------------------------------------------------------------
k5              -.6360998    -.3498737    -.3427888    -.1848925    -.3568748
k618            -.1277862    -.0156047    -.0157506    -.0207876    -.0157519
age             -.4372017     -.002954     -.015336    -.1231976    -.0153371
wc               .1880592     .1880592     .1944887     .0884042     .1969329
hc               .0271984     .0271984     .0272506     .0133135     .0272572
lwg              .6624324     .1499499       .14648     .0864619     .1475137
inc             -.6415044    -.0068042     -.008403    -.0974665    -.0084031
-----------------------------------------------------------------------------
N                     753                                                    
-----------------------------------------------------------------------------

. esttab, cells("dc[min->max] dc[0->1] dc[-+1/2] dc[-+sd/2] dc[MargEfct]")

-----------------------------------------------------------------------------
                      (1)                                                    
                      lfp                                                    
                 min->max         0->1        -+1/2       -+sd/2     MargEfct
-----------------------------------------------------------------------------
k5              -.6360998    -.3498737    -.3427888    -.1848925    -.3568748
k618            -.1277862    -.0156047    -.0157506    -.0207876    -.0157519
age             -.4372017     -.002954     -.015336    -.1231976    -.0153371
wc               .1880592     .1880592     .1944887     .0884042     .1969329
hc               .0271984     .0271984     .0272506     .0133135     .0272572
lwg              .6624324     .1499499       .14648     .0864619     .1475137
inc             -.6415044    -.0068042     -.008403    -.0974665    -.0084031
-----------------------------------------------------------------------------
N                     753                                                    
-----------------------------------------------------------------------------

[do-file]

Tabulating results from prvalue or asprvalue

The procedure to prepare results from prvalue for tabulation is to first collect a series of predictions by repeated calls to estadd prvalue and then apply estadd prvalue post to rearrange results and post them in e(). Use the label() option to label the single predictions. Example:

. spex binlfp2
(Data from 1976 PSID-T Mroz)

. logit lfp k5 k618 age wc hc lwg inc, nolog

Logistic regression                               Number of obs   =        753
                                                  LR chi2(7)      =     124.48
                                                  Prob > chi2     =     0.0000
Log likelihood = -452.63296                       Pseudo R2       =     0.1209

------------------------------------------------------------------------------
         lfp |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
          k5 |  -1.462913   .1970006    -7.43   0.000    -1.849027   -1.076799
        k618 |  -.0645707   .0680008    -0.95   0.342    -.1978499    .0687085
         age |  -.0628706   .0127831    -4.92   0.000    -.0879249   -.0378162
          wc |   .8072738   .2299799     3.51   0.000     .3565215    1.258026
          hc |   .1117336   .2060397     0.54   0.588    -.2920969     .515564
         lwg |   .6046931   .1508176     4.01   0.000     .3090961    .9002901
         inc |  -.0344464   .0082084    -4.20   0.000    -.0505346   -.0183583
       _cons |    3.18214   .6443751     4.94   0.000     1.919188    4.445092
------------------------------------------------------------------------------

. estadd prvalue, x(age=35 k5=2 wc=0 hc=0 inc=15) label(family type 1)

logit: Predictions for lfp

Confidence intervals by delta method

                                95% Conf. Interval
  Pr(y=inLF|x):       0.1318   [ 0.0556,    0.2081]
  Pr(y=NotInLF|x):    0.8682   [ 0.7919,    0.9444]

           k5       k618        age         wc         hc        lwg
x=          2  1.3532537         35          0          0  1.0971148

          inc
x=         15

added matrices:
    e(_estadd_prvalue) :  1 x 12
  e(_estadd_prvalue_x) :  1 x 7

. estadd prvalue, x(age=50 k5=0 k618=0 wc=1 hc=1) label(family type 2)

logit: Predictions for lfp

Confidence intervals by delta method

                                95% Conf. Interval
  Pr(y=inLF|x):       0.7166   [ 0.6333,    0.7999]
  Pr(y=NotInLF|x):    0.2834   [ 0.2001,    0.3667]

           k5       k618        age         wc         hc        lwg
x=          0          0         50          1          1  1.0971148

          inc
x=  20.128965

updated matrices:
    e(_estadd_prvalue) :  2 x 12
  e(_estadd_prvalue_x) :  2 x 7

. estadd prvalue, label(average family)

logit: Predictions for lfp

Confidence intervals by delta method

                                95% Conf. Interval
  Pr(y=inLF|x):       0.5778   [ 0.5392,    0.6164]
  Pr(y=NotInLF|x):    0.4222   [ 0.3836,    0.4608]

           k5       k618        age         wc         hc        lwg
x=   .2377158  1.3532537  42.537849   .2815405  .39176627  1.0971148

          inc
x=  20.128965

updated matrices:
    e(_estadd_prvalue) :  3 x 12
  e(_estadd_prvalue_x) :  3 x 7

. estadd prvalue post

scalars:
                  e(N) =  753

macros:
             e(depvar) : "lfp"
                e(cmd) : "estadd_prvalue"
              e(model) : "logit"
         e(properties) : "b"

matrices:
                  e(b) :  1 x 6      (predictions)
                 e(se) :  1 x 6      (standard errors)
                 e(LB) :  1 x 6      (lower CI bounds)
                 e(UB) :  1 x 6      (upper CI bounds)
           e(Category) :  1 x 6      (outcome values)
                  e(X) :  7 x 3      (k5, k618, age, wc, hc, lwg, inc)

. esttab, ci wide nostar ///
>     keep(inLF:) eqlabels(none) varwidth(15)

---------------------------------------------------
                         (1)                       
                         lfp                       
---------------------------------------------------
family type 1          0.132         [0.0556,0.208]
family type 2          0.717          [0.633,0.800]
average family         0.578          [0.539,0.616]
---------------------------------------------------
N                        753                       
---------------------------------------------------
95% confidence intervals in brackets

[do-file]

The procedure for asprvalue is analogous (however, note that asprvalue does not provide standard errors or confidence intervals).

Tabulating differences in predictions from prvalue or asprvalue

If you want to tabulate differences in predictions, first apply prvalue (or asprvalue) with the save option and then estadd prvalue (or estadd asprvalue) with the diff option. Example:

. spex binlfp2
(Data from 1976 PSID-T Mroz)

. quietly logit lfp k5 k618 age wc hc lwg inc, nolog

. quietly prvalue, x(k5=0 wc=0) save

. estadd  prvalue, x(k5=0 wc=1) label(k5 = 0) brief diff

logit: Change in Predictions for lfp

                     Current     Saved    Change   95% CI for Change
  Pr(y=inLF|x):       0.7758    0.6069    0.1689  [ 0.0830,   0.2549]
  Pr(y=NotInLF|x):    0.2242    0.3931   -0.1689  [-0.2549,  -0.0830]

added matrices:
    e(_estadd_prvalue) :  1 x 12
  e(_estadd_prvalue_x) :  1 x 7

. quietly prvalue, x(k5=1 wc=0) save

. estadd  prvalue, x(k5=1 wc=1) label(k5 = 1) brief diff

logit: Change in Predictions for lfp

                     Current     Saved    Change   95% CI for Change
  Pr(y=inLF|x):       0.4449    0.2633    0.1815  [ 0.0763,   0.2868]
  Pr(y=NotInLF|x):    0.5551    0.7367   -0.1815  [-0.2868,  -0.0763]

updated matrices:
    e(_estadd_prvalue) :  2 x 12
  e(_estadd_prvalue_x) :  2 x 7

. quietly prvalue, x(k5=2 wc=0) save

. estadd  prvalue, x(k5=2 wc=1) label(k5 = 2) brief diff

logit: Change in Predictions for lfp

                     Current     Saved    Change   95% CI for Change
  Pr(y=inLF|x):       0.1565    0.0764    0.0801  [ 0.0156,   0.1445]
  Pr(y=NotInLF|x):    0.8435    0.9236   -0.0801  [-0.1445,  -0.0156]

updated matrices:
    e(_estadd_prvalue) :  3 x 12
  e(_estadd_prvalue_x) :  3 x 7

. estadd  prvalue post

scalars:
                  e(N) =  753

macros:
             e(depvar) : "lfp"
                e(cmd) : "estadd_prvalue"
              e(model) : "logit"
         e(properties) : "b"

matrices:
                  e(b) :  1 x 6      (predictions)
                 e(se) :  1 x 6      (standard errors)
                 e(LB) :  1 x 6      (lower CI bounds)
                 e(UB) :  1 x 6      (upper CI bounds)
           e(Category) :  1 x 6      (outcome values)
                  e(X) :  7 x 3      (k5, k618, age, wc, hc, lwg, inc)

. esttab, keep(inLF:) ci wide nostar ///
>     mtitle("wc=1 - wc=0")

------------------------------------------------
                      (1)                       
              wc=1 - wc=0                       
------------------------------------------------
inLF                                            
k5 = 0              0.169         [0.0830,0.255]
k5 = 1              0.182         [0.0763,0.287]
k5 = 2             0.0801         [0.0156,0.145]
------------------------------------------------
N                     753                       
------------------------------------------------
95% confidence intervals in brackets

[do-file]

Tabulating bootstrap confidence intervals from prvalue

The confidence bounds computed by prvalue are saved by estadd prvalue post in e(LB) and e(UB). The following example illustrates how to tabulate these results:

. spex binlfp2
(Data from 1976 PSID-T Mroz)

. quietly logit lfp k5 k618 age wc hc lwg inc, nolog

. estadd prvalue, x(age=35 k5=2 wc=0 hc=0 inc=15) ///
>     label(family type 1) bootstrap

logit: Predictions for lfp

Bootstrap confidence intervals using percentile method
(1000 of 1000 replications completed)

                                95% Conf. Interval
  Pr(y=inLF|x):       0.1318   [ 0.0629,    0.2220]
  Pr(y=NotInLF|x):    0.8682   [ 0.7780,    0.9371]

           k5       k618        age         wc         hc        lwg
x=          2  1.3532537         35          0          0  1.0971148

          inc
x=         15

added matrices:
    e(_estadd_prvalue) :  1 x 12
  e(_estadd_prvalue_x) :  1 x 7

. estadd prvalue, x(age=50 k5=0 k618=0 wc=1 hc=1) ///
>     label(family type 2) bootstrap

logit: Predictions for lfp

Bootstrap confidence intervals using percentile method
(1000 of 1000 replications completed)

                                95% Conf. Interval
  Pr(y=inLF|x):       0.7166   [ 0.6305,    0.7994]
  Pr(y=NotInLF|x):    0.2834   [ 0.2006,    0.3695]

           k5       k618        age         wc         hc        lwg
x=          0          0         50          1          1  1.0971148

          inc
x=  20.128965

updated matrices:
    e(_estadd_prvalue) :  2 x 12
  e(_estadd_prvalue_x) :  2 x 7

. estadd prvalue, label(average family) bootstrap 

logit: Predictions for lfp

Bootstrap confidence intervals using percentile method
(1000 of 1000 replications completed)

                                95% Conf. Interval
  Pr(y=inLF|x):       0.5778   [ 0.5389,    0.6205]
  Pr(y=NotInLF|x):    0.4222   [ 0.3795,    0.4611]

           k5       k618        age         wc         hc        lwg
x=   .2377158  1.3532537  42.537849   .2815405  .39176627  1.0971148

          inc
x=  20.128965

updated matrices:
    e(_estadd_prvalue) :  3 x 12
  e(_estadd_prvalue_x) :  3 x 7

. estadd prvalue post

scalars:
                  e(N) =  753

macros:
             e(depvar) : "lfp"
                e(cmd) : "estadd_prvalue"
              e(model) : "logit"
         e(properties) : "b"

matrices:
                  e(b) :  1 x 6      (predictions)
                 e(se) :  1 x 6      (standard errors)
                 e(LB) :  1 x 6      (lower CI bounds)
                 e(UB) :  1 x 6      (upper CI bounds)
           e(Category) :  1 x 6      (outcome values)
                  e(X) :  7 x 3      (k5, k618, age, wc, hc, lwg, inc)

. esttab, cells("b LB UB") ///
>     keep(inLF:) eqlabels(none) varwidth(15)

------------------------------------------------------
                         (1)                          
                         lfp                          
                           b           LB           UB
------------------------------------------------------
family type 1       .1318369      .062898     .2219745
family type 2       .7166017     .6304579     .7994323
average family      .5777942     .5389454     .6205298
------------------------------------------------------
N                        753                          
------------------------------------------------------

[do-file]

Logit and probit

Logit/probit and fitstat

. spex binlfp2
(Data from 1976 PSID-T Mroz)

. quietly logit lfp k5 k618 age wc hc lwg inc, nolog

. estadd fitstat

Measures of Fit for logit of lfp

Log-Lik Intercept Only:       -514.873   Log-Lik Full Model:           -452.633
D(745):                        905.266   LR(7):                         124.480
                                         Prob > LR:                       0.000
McFadden's R2:                   0.121   McFadden's Adj R2:               0.105
ML (Cox-Snell) R2:               0.152   Cragg-Uhler(Nagelkerke) R2:      0.204
McKelvey & Zavoina's R2:         0.217   Efron's R2:                      0.155
Variance of y*:                  4.203   Variance of error:               3.290
Count R2:                        0.693   Adj Count R2:                    0.289
AIC:                             1.223   AIC*n:                         921.266
BIC:                         -4029.663   BIC':                          -78.112
BIC used by Stata:             958.258   AIC used by Stata:             921.266

added scalars:
                e(dev) =  905.26592
             e(dev_df) =  745
               e(lrx2) =  124.48049
            e(lrx2_df) =  7
             e(lrx2_p) =  8.923e-24
              e(r2_mf) =  .12088461
           e(r2_mfadj) =  .1053468
              e(r2_ml) =  .15237143
              e(r2_cu) =  .20445312
              e(r2_mz) =  .2171939
              e(r2_ef) =  .15493519
            e(v_ystar) =  4.2026603
            e(v_error) =  3.2898681
              e(r2_ct) =  .69322709
           e(r2_ctadj) =  .28923077
               e(aic0) =  1.2234607
              e(aic_n) =  921.26592
               e(bic0) =  -4029.6627
              e(bic_p) =  -78.112037
           e(statabic) =  958.25844
           e(stataaic) =  921.26592
              e(n_rhs) =  7
             e(n_parm) =  8

. eststo logit

. quietly probit lfp k5 k618 age wc hc lwg inc, nolog

. estadd fitstat

Measures of Fit for probit of lfp

Log-Lik Intercept Only:       -514.873   Log-Lik Full Model:           -452.695
D(745):                        905.390   LR(7):                         124.356
                                         Prob > LR:                       0.000
McFadden's R2:                   0.121   McFadden's Adj R2:               0.105
ML (Cox-Snell) R2:               0.152   Cragg-Uhler(Nagelkerke) R2:      0.204
McKelvey & Zavoina's R2:         0.247   Efron's R2:                      0.154
Variance of y*:                  1.328   Variance of error:               1.000
Count R2:                        0.687   Adj Count R2:                    0.274
AIC:                             1.224   AIC*n:                         921.390
BIC:                         -4029.539   BIC':                          -77.988
BIC used by Stata:             958.382   AIC used by Stata:             921.390

added scalars:
                e(dev) =  905.38993
             e(dev_df) =  745
               e(lrx2) =  124.35648
            e(lrx2_df) =  7
             e(lrx2_p) =  9.471e-24
              e(r2_mf) =  .12076418
           e(r2_mfadj) =  .10522638
              e(r2_ml) =  .15223182
              e(r2_cu) =  .2042658
              e(r2_mz) =  .24703499
              e(r2_ef) =  .15420358
            e(v_ystar) =  1.328083
            e(v_error) =  1
              e(r2_ct) =  .68658699
           e(r2_ctadj) =  .27384615
               e(aic0) =  1.2236254
              e(aic_n) =  921.38993
               e(bic0) =  -4029.5387
              e(bic_p) =  -77.988025
           e(statabic) =  958.38245
           e(stataaic) =  921.38993
              e(n_rhs) =  7
             e(n_parm) =  8

. eststo probit

. esttab, scalars(r2_mf r2_mfadj r2_ml r2_cu) wide mtitles

----------------------------------------------------------------------
                      (1)                          (2)                
                    logit                       probit                
----------------------------------------------------------------------
lfp                                                                   
k5                 -1.463***      (-7.43)       -0.875***      (-7.70)
k618              -0.0646         (-0.95)      -0.0386         (-0.95)
age               -0.0629***      (-4.92)      -0.0378***      (-4.97)
wc                  0.807***       (3.51)        0.488***       (3.60)
hc                  0.112          (0.54)       0.0572          (0.46)
lwg                 0.605***       (4.01)        0.366***       (4.17)
inc               -0.0344***      (-4.20)      -0.0205***      (-4.30)
_cons               3.182***       (4.94)        1.918***       (5.04)
----------------------------------------------------------------------
N                     753                          753                
r2_mf               0.121                        0.121                
r2_mfadj            0.105                        0.105                
r2_ml               0.152                        0.152                
r2_cu               0.204                        0.204                
----------------------------------------------------------------------
t statistics in parentheses
* p<0.05, ** p<0.01, *** p<0.001

. eststo clear

[do-file]

Logit/probit and listcoef

. spex binlfp2
(Data from 1976 PSID-T Mroz)

. quietly logit lfp k5 k618 age wc hc lwg inc, nolog

. estadd listcoef, std

logit (N=753): Unstandardized and Standardized Estimates 

 Observed SD: .49562951
   Latent SD: 2.0500391

  Odds of: inLF vs NotInLF

-------------------------------------------------------------------------------
         lfp |      b         z     P>|z|    bStdX    bStdY   bStdXY      SDofX
-------------+-----------------------------------------------------------------
          k5 |  -1.46291   -7.426   0.000  -0.7665  -0.7136  -0.3739     0.5240
        k618 |  -0.06457   -0.950   0.342  -0.0852  -0.0315  -0.0416     1.3199
         age |  -0.06287   -4.918   0.000  -0.5075  -0.0307  -0.2476     8.0726
          wc |   0.80727    3.510   0.000   0.3633   0.3938   0.1772     0.4500
          hc |   0.11173    0.542   0.588   0.0546   0.0545   0.0266     0.4885
         lwg |   0.60469    4.009   0.000   0.3553   0.2950   0.1733     0.5876
         inc |  -0.03445   -4.196   0.000  -0.4008  -0.0168  -0.1955    11.6348
-------------------------------------------------------------------------------

added matrices:
               e(b_xs) :  1 x 7      (bStdX)
               e(b_ys) :  1 x 7      (bStdY)
              e(b_std) :  1 x 7      (bStdXY)
              e(b_sdx) :  1 x 7      (SDofX)

. eststo logit

. quietly probit lfp k5 k618 age wc hc lwg inc, nolog

. estadd listcoef

probit (N=753): Unstandardized and Standardized Estimates 

 Observed SD: .49562951
   Latent SD: 1.1524248

-------------------------------------------------------------------------------
         lfp |      b         z     P>|z|    bStdX    bStdY   bStdXY      SDofX
-------------+-----------------------------------------------------------------
          k5 |  -0.87471   -7.703   0.000  -0.4583  -0.7590  -0.3977     0.5240
        k618 |  -0.03859   -0.953   0.340  -0.0509  -0.0335  -0.0442     1.3199
         age |  -0.03782   -4.971   0.000  -0.3053  -0.0328  -0.2649     8.0726
          wc |   0.48831    3.604   0.000   0.2198   0.4237   0.1907     0.4500
          hc |   0.05717    0.461   0.645   0.0279   0.0496   0.0242     0.4885
         lwg |   0.36563    4.165   0.000   0.2148   0.3173   0.1864     0.5876
         inc |  -0.02053   -4.297   0.000  -0.2388  -0.0178  -0.2072    11.6348
-------------------------------------------------------------------------------

added matrices:
               e(b_xs) :  1 x 7      (bStdX)
               e(b_ys) :  1 x 7      (bStdY)
              e(b_std) :  1 x 7      (bStdXY)
              e(b_sdx) :  1 x 7      (SDofX)

. eststo probit

. esttab, aux(b_std) nopar wide mtitles

----------------------------------------------------------------------
                      (1)                          (2)                
                    logit                       probit                
----------------------------------------------------------------------
lfp                                                                   
k5                 -1.463***       -0.374       -0.875***       -0.398
k618              -0.0646         -0.0416      -0.0386         -0.0442
age               -0.0629***       -0.248      -0.0378***       -0.265
wc                  0.807***        0.177        0.488***        0.191
hc                  0.112          0.0266       0.0572          0.0242
lwg                 0.605***        0.173        0.366***        0.186
inc               -0.0344***       -0.195      -0.0205***       -0.207
_cons               3.182***                     1.918***             
----------------------------------------------------------------------
N                     753                          753                
----------------------------------------------------------------------
b_std in second column
* p<0.05, ** p<0.01, *** p<0.001

. eststo clear

[do-file]

Logit and listcoef: factor and percent change

. spex binlfp2
(Data from 1976 PSID-T Mroz)

. quietly logit lfp k5 k618 age wc hc lwg inc, nolog

. estadd listcoef, quietly std

added matrices:
               e(b_xs) :  1 x 7      (bStdX)
               e(b_ys) :  1 x 7      (bStdY)
              e(b_std) :  1 x 7      (bStdXY)
              e(b_sdx) :  1 x 7      (SDofX)

. estadd listcoef, quietly fact nosd

added matrices:
             e(b_fact) :  1 x 7      (e^b)
            e(b_facts) :  1 x 7      (e^bStdX)

. estadd listcoef, quietly per nosd

added matrices:
              e(b_pct) :  1 x 7      (%)
             e(b_pcts) :  1 x 7      (%StdX)

. esttab, cell("b_std b_facts b_pcts b_sdx")

----------------------------------------------------------------
                      (1)                                       
                      lfp                                       
                    b_std      b_facts       b_pcts        b_sdx
----------------------------------------------------------------
k5              -.3738985     .4646334    -53.53666      .523959
k618            -.0415725     .9183055    -8.169451     1.319874
age             -.2475695     .6019823    -39.80177     8.072574
wc               .1772225     1.438086      43.8086     .4500494
hc               .0266231     1.056095      5.60953     .4884694
lwg              .1733095     1.426596     42.65962     .5875564
inc             -.1954974     .6697992    -33.02008      11.6348
----------------------------------------------------------------
N                     753                                       
----------------------------------------------------------------

[do-file]

Logit/probit and prchange

. spex binlfp2
(Data from 1976 PSID-T Mroz)

. quietly logit lfp k5 k618 age wc hc lwg inc, nolog

. estadd prchange

logit: Changes in Probabilities for lfp

      min->max      0->1     -+1/2    -+sd/2  MargEfct
  k5   -0.6361   -0.3499   -0.3428   -0.1849   -0.3569
k618   -0.1278   -0.0156   -0.0158   -0.0208   -0.0158
 age   -0.4372   -0.0030   -0.0153   -0.1232   -0.0153
  wc    0.1881    0.1881    0.1945    0.0884    0.1969
  hc    0.0272    0.0272    0.0273    0.0133    0.0273
 lwg    0.6624    0.1499    0.1465    0.0865    0.1475
 inc   -0.6415   -0.0068   -0.0084   -0.0975   -0.0084

         NotInLF     inLF
Pr(y|x)   0.4222   0.5778

            k5     k618      age       wc       hc      lwg      inc
   x=  .237716  1.35325  42.5378  .281541  .391766  1.09711   20.129
sd_x=  .523959  1.31987  8.07257  .450049  .488469  .587556  11.6348

added scalars:
            e(predval) =  .57779419
              e(delta) =  1
           e(centered) =  1

added matrices:
                 e(dc) :  6 x 7      (main, min->max, 0->1, -+1/2, -+sd/2, Marg
> Efct)
            e(pattern) :  1 x 7
                  e(X) :  4 x 7      (X, SD, Min, Max)

first row in e(dc) contains:

  01 change for binary variables
  sd change for continuous variables

. eststo logit

. quietly probit lfp k5 k618 age wc hc lwg inc, nolog

. estadd prchange

probit: Changes in Probabilities for lfp

      min->max      0->1     -+1/2    -+sd/2  MargEfct
  k5   -0.6441   -0.3380   -0.3320   -0.1778   -0.3422
k618   -0.1221   -0.0150   -0.0151   -0.0199   -0.0151
 age   -0.4274   -0.0031   -0.0148   -0.1190   -0.0148
  wc    0.1844    0.1844    0.1892    0.0858    0.1911
  hc    0.0223    0.0223    0.0224    0.0109    0.0224
 lwg    0.6649    0.1450    0.1423    0.0839    0.1431
 inc   -0.6425   -0.0068   -0.0080   -0.0932   -0.0080

         NotInLF     inLF
Pr(y|x)   0.4218   0.5782

            k5     k618      age       wc       hc      lwg      inc
   x=  .237716  1.35325  42.5378  .281541  .391766  1.09711   20.129
sd_x=  .523959  1.31987  8.07257  .450049  .488469  .587556  11.6348

added scalars:
            e(predval) =  .57816368
              e(delta) =  1
           e(centered) =  1

added matrices:
                 e(dc) :  6 x 7      (main, min->max, 0->1, -+1/2, -+sd/2, Marg
> Efct)
            e(pattern) :  1 x 7
                  e(X) :  4 x 7      (X, SD, Min, Max)

first row in e(dc) contains:

  01 change for binary variables
  sd change for continuous variables

. eststo probit

. esttab, aux(dc) nopar wide mtitles

----------------------------------------------------------------------
                      (1)                          (2)                
                    logit                       probit                
----------------------------------------------------------------------
lfp                                                                   
k5                 -1.463***       -0.185       -0.875***       -0.178
k618              -0.0646         -0.0208      -0.0386         -0.0199
age               -0.0629***       -0.123      -0.0378***       -0.119
wc                  0.807***        0.188        0.488***        0.184
hc                  0.112          0.0272       0.0572          0.0223
lwg                 0.605***       0.0865        0.366***       0.0839
inc               -0.0344***      -0.0975      -0.0205***      -0.0932
_cons               3.182***                     1.918***             
----------------------------------------------------------------------
N                     753                          753                
----------------------------------------------------------------------
dc in second column
* p<0.05, ** p<0.01, *** p<0.001

. eststo clear

[do-file]

Logit and prvalue

. spex binlfp2
(Data from 1976 PSID-T Mroz)

. quietly logit lfp k5 k618 age wc hc lwg inc, nolog

. estadd prvalue, x(k5=0 wc=0) label(k5 = 0) brief

logit: Predictions for lfp

                                95% Conf. Interval
  Pr(y=inLF|x):       0.6069   [ 0.5567,    0.6570]
  Pr(y=NotInLF|x):    0.3931   [ 0.3430,    0.4433]

added matrices:
    e(_estadd_prvalue) :  1 x 12
  e(_estadd_prvalue_x) :  1 x 7

. estadd prvalue, x(k5=1 wc=0) label(k5 = 1) brief

logit: Predictions for lfp

                                95% Conf. Interval
  Pr(y=inLF|x):       0.2633   [ 0.1932,    0.3335]
  Pr(y=NotInLF|x):    0.7367   [ 0.6665,    0.8068]

updated matrices:
    e(_estadd_prvalue) :  2 x 12
  e(_estadd_prvalue_x) :  2 x 7

. estadd prvalue, x(k5=2 wc=0) label(k5 = 2) brief

logit: Predictions for lfp

                                95% Conf. Interval
  Pr(y=inLF|x):       0.0764   [ 0.0258,    0.1271]
  Pr(y=NotInLF|x):    0.9236   [ 0.8729,    0.9742]

updated matrices:
    e(_estadd_prvalue) :  3 x 12
  e(_estadd_prvalue_x) :  3 x 7

. estadd prvalue, x(k5=3 wc=0) label(k5 = 3) brief

logit: Predictions for lfp

                                95% Conf. Interval
  Pr(y=inLF|x):       0.0188   [-0.0014,    0.0390]
  Pr(y=NotInLF|x):    0.9812   [ 0.9610,    1.0014]

updated matrices:
    e(_estadd_prvalue) :  4 x 12
  e(_estadd_prvalue_x) :  4 x 7

. estadd prvalue post NoCollege

scalars:
                  e(N) =  753

macros:
             e(depvar) : "lfp"
                e(cmd) : "estadd_prvalue"
              e(model) : "logit"
         e(properties) : "b"

matrices:
                  e(b) :  1 x 8      (predictions)
                 e(se) :  1 x 8      (standard errors)
                 e(LB) :  1 x 8      (lower CI bounds)
                 e(UB) :  1 x 8      (upper CI bounds)
           e(Category) :  1 x 8      (outcome values)
                  e(X) :  7 x 4      (k5, k618, age, wc, hc, lwg, inc)

results stored as NoCollege

. estadd prvalue, x(k5=0 wc=1) label(k5 = 0) brief replace

logit: Predictions for lfp

                                95% Conf. Interval
  Pr(y=inLF|x):       0.7758   [ 0.7077,    0.8439]
  Pr(y=NotInLF|x):    0.2242   [ 0.1561,    0.2923]

added matrices:
    e(_estadd_prvalue) :  1 x 12
  e(_estadd_prvalue_x) :  1 x 7

. estadd prvalue, x(k5=1 wc=1) label(k5 = 1) brief

logit: Predictions for lfp

                                95% Conf. Interval
  Pr(y=inLF|x):       0.4449   [ 0.3331,    0.5567]
  Pr(y=NotInLF|x):    0.5551   [ 0.4433,    0.6669]

updated matrices:
    e(_estadd_prvalue) :  2 x 12
  e(_estadd_prvalue_x) :  2 x 7

. estadd prvalue, x(k5=2 wc=1) label(k5 = 2) brief

logit: Predictions for lfp

                                95% Conf. Interval
  Pr(y=inLF|x):       0.1565   [ 0.0582,    0.2548]
  Pr(y=NotInLF|x):    0.8435   [ 0.7452,    0.9418]

updated matrices:
    e(_estadd_prvalue) :  3 x 12
  e(_estadd_prvalue_x) :  3 x 7

. estadd prvalue, x(k5=3 wc=1) label(k5 = 3) brief

logit: Predictions for lfp

                                95% Conf. Interval
  Pr(y=inLF|x):       0.0412   [-0.0021,    0.0845]
  Pr(y=NotInLF|x):    0.9588   [ 0.9155,    1.0021]

updated matrices:
    e(_estadd_prvalue) :  4 x 12
  e(_estadd_prvalue_x) :  4 x 7

. estadd prvalue post College

scalars:
                  e(N) =  753

macros:
             e(depvar) : "lfp"
                e(cmd) : "estadd_prvalue"
              e(model) : "logit"
         e(properties) : "b"

matrices:
                  e(b) :  1 x 8      (predictions)
                 e(se) :  1 x 8      (standard errors)
                 e(LB) :  1 x 8      (lower CI bounds)
                 e(UB) :  1 x 8      (upper CI bounds)
           e(Category) :  1 x 8      (outcome values)
                  e(X) :  7 x 4      (k5, k618, age, wc, hc, lwg, inc)

results stored as College

. quietly prvalue, x(k5=0 wc=0) save

. estadd  prvalue, x(k5=0 wc=1) label(k5 = 0) brief diff replace

logit: Change in Predictions for lfp

                     Current     Saved    Change   95% CI for Change
  Pr(y=inLF|x):       0.7758    0.6069    0.1689  [ 0.0830,   0.2549]
  Pr(y=NotInLF|x):    0.2242    0.3931   -0.1689  [-0.2549,  -0.0830]

added matrices:
    e(_estadd_prvalue) :  1 x 12
  e(_estadd_prvalue_x) :  1 x 7

. quietly prvalue, x(k5=1 wc=0) save

. estadd  prvalue, x(k5=1 wc=1) label(k5 = 1) brief diff

logit: Change in Predictions for lfp

                     Current     Saved    Change   95% CI for Change
  Pr(y=inLF|x):       0.4449    0.2633    0.1815  [ 0.0763,   0.2868]
  Pr(y=NotInLF|x):    0.5551    0.7367   -0.1815  [-0.2868,  -0.0763]

updated matrices:
    e(_estadd_prvalue) :  2 x 12
  e(_estadd_prvalue_x) :  2 x 7

. quietly prvalue, x(k5=2 wc=0) save

. estadd  prvalue, x(k5=2 wc=1) label(k5 = 2) brief diff

logit: Change in Predictions for lfp

                     Current     Saved    Change   95% CI for Change
  Pr(y=inLF|x):       0.1565    0.0764    0.0801  [ 0.0156,   0.1445]
  Pr(y=NotInLF|x):    0.8435    0.9236   -0.0801  [-0.1445,  -0.0156]

updated matrices:
    e(_estadd_prvalue) :  3 x 12
  e(_estadd_prvalue_x) :  3 x 7

. quietly prvalue, x(k5=3 wc=0) save

. estadd  prvalue, x(k5=3 wc=1) label(k5 = 3) brief diff

logit: Change in Predictions for lfp

                     Current     Saved    Change   95% CI for Change
  Pr(y=inLF|x):       0.0412    0.0188    0.0224  [-0.0037,   0.0485]
  Pr(y=NotInLF|x):    0.9588    0.9812   -0.0224  [-0.0485,   0.0037]

updated matrices:
    e(_estadd_prvalue) :  4 x 12
  e(_estadd_prvalue_x) :  4 x 7

. estadd prvalue post Difference

scalars:
                  e(N) =  753

macros:
             e(depvar) : "lfp"
                e(cmd) : "estadd_prvalue"
              e(model) : "logit"
         e(properties) : "b"

matrices:
                  e(b) :  1 x 8      (predictions)
                 e(se) :  1 x 8      (standard errors)
                 e(LB) :  1 x 8      (lower CI bounds)
                 e(UB) :  1 x 8      (upper CI bounds)
           e(Category) :  1 x 8      (outcome values)
                  e(X) :  7 x 4      (k5, k618, age, wc, hc, lwg, inc)

results stored as Difference

. esttab, se nostar nonumber noobs mtitles ///
>     keep(inLF:) eqlabels(none)

---------------------------------------------------
                NoCollege      College   Difference
---------------------------------------------------
k5 = 0              0.607        0.776        0.169
                 (0.0256)     (0.0348)     (0.0439)

k5 = 1              0.263        0.445        0.182
                 (0.0358)     (0.0570)     (0.0537)

k5 = 2             0.0764        0.157       0.0801
                 (0.0259)     (0.0502)     (0.0329)

k5 = 3             0.0188       0.0412       0.0224
                 (0.0103)     (0.0221)     (0.0133)
---------------------------------------------------
Standard errors in parentheses

. eststo clear

[do-file]

Probit and prvalue

. spex binlfp2
(Data from 1976 PSID-T Mroz)

. quietly probit lfp k5 k618 age wc hc lwg inc, nolog

. estadd prvalue, x(k5=0 wc=0) label(k5 = 0) brief

probit: Predictions for lfp

                                95% Conf. Interval
  Pr(y=inLF|x):       0.6055   [ 0.5563,    0.6547]
  Pr(y=NotInLF|x):    0.3945   [ 0.3453,    0.4437]

added matrices:
    e(_estadd_prvalue) :  1 x 12
  e(_estadd_prvalue_x) :  1 x 7

. estadd prvalue, x(k5=1 wc=0) label(k5 = 1) brief

probit: Predictions for lfp

                                95% Conf. Interval
  Pr(y=inLF|x):       0.2719   [ 0.2017,    0.3421]
  Pr(y=NotInLF|x):    0.7281   [ 0.6579,    0.7983]

updated matrices:
    e(_estadd_prvalue) :  2 x 12
  e(_estadd_prvalue_x) :  2 x 7

. estadd prvalue, x(k5=2 wc=0) label(k5 = 2) brief

probit: Predictions for lfp

                                95% Conf. Interval
  Pr(y=inLF|x):       0.0692   [ 0.0140,    0.1244]
  Pr(y=NotInLF|x):    0.9308   [ 0.8756,    0.9860]

updated matrices:
    e(_estadd_prvalue) :  3 x 12
  e(_estadd_prvalue_x) :  3 x 7

. estadd prvalue, x(k5=3 wc=0) label(k5 = 3) brief

probit: Predictions for lfp

                                95% Conf. Interval
  Pr(y=inLF|x):       0.0092   [-0.0065,    0.0249]
  Pr(y=NotInLF|x):    0.9908   [ 0.9751,    1.0065]

updated matrices:
    e(_estadd_prvalue) :  4 x 12
  e(_estadd_prvalue_x) :  4 x 7

. estadd prvalue post NoCollege

scalars:
                  e(N) =  753

macros:
             e(depvar) : "lfp"
                e(cmd) : "estadd_prvalue"
              e(model) : "probit"
         e(properties) : "b"

matrices:
                  e(b) :  1 x 8      (predictions)
                 e(se) :  1 x 8      (standard errors)
                 e(LB) :  1 x 8      (lower CI bounds)
                 e(UB) :  1 x 8      (upper CI bounds)
           e(Category) :  1 x 8      (outcome values)
                  e(X) :  7 x 4      (k5, k618, age, wc, hc, lwg, inc)

results stored as NoCollege

. estadd prvalue, x(k5=0 wc=1) label(k5 = 0) brief replace

probit: Predictions for lfp

                                95% Conf. Interval
  Pr(y=inLF|x):       0.7752   [ 0.7070,    0.8434]
  Pr(y=NotInLF|x):    0.2248   [ 0.1566,    0.2930]

added matrices:
    e(_estadd_prvalue) :  1 x 12
  e(_estadd_prvalue_x) :  1 x 7

. estadd prvalue, x(k5=1 wc=1) label(k5 = 1) brief

probit: Predictions for lfp

                                95% Conf. Interval
  Pr(y=inLF|x):       0.4527   [ 0.3477,    0.5578]
  Pr(y=NotInLF|x):    0.5473   [ 0.4422,    0.6523]

updated matrices:
    e(_estadd_prvalue) :  2 x 12
  e(_estadd_prvalue_x) :  2 x 7

. estadd prvalue, x(k5=2 wc=1) label(k5 = 2) brief

probit: Predictions for lfp

                                95% Conf. Interval
  Pr(y=inLF|x):       0.1602   [ 0.0547,    0.2658]
  Pr(y=NotInLF|x):    0.8398   [ 0.7342,    0.9453]

updated matrices:
    e(_estadd_prvalue) :  3 x 12
  e(_estadd_prvalue_x) :  3 x 7

. estadd prvalue, x(k5=3 wc=1) label(k5 = 3) brief

probit: Predictions for lfp

                                95% Conf. Interval
  Pr(y=inLF|x):       0.0309   [-0.0135,    0.0752]
  Pr(y=NotInLF|x):    0.9691   [ 0.9248,    1.0135]

updated matrices:
    e(_estadd_prvalue) :  4 x 12
  e(_estadd_prvalue_x) :  4 x 7

. estadd prvalue post College

scalars:
                  e(N) =  753

macros:
             e(depvar) : "lfp"
                e(cmd) : "estadd_prvalue"
              e(model) : "probit"
         e(properties) : "b"

matrices:
                  e(b) :  1 x 8      (predictions)
                 e(se) :  1 x 8      (standard errors)
                 e(LB) :  1 x 8      (lower CI bounds)
                 e(UB) :  1 x 8      (upper CI bounds)
           e(Category) :  1 x 8      (outcome values)
                  e(X) :  7 x 4      (k5, k618, age, wc, hc, lwg, inc)

results stored as College

. quietly prvalue, x(k5=0 wc=0) save

. estadd  prvalue, x(k5=0 wc=1) label(k5 = 0) brief diff replace

probit: Change in Predictions for lfp

                     Current     Saved    Change   95% CI for Change
  Pr(y=inLF|x):       0.7752    0.6055    0.1696  [ 0.0839,   0.2554]
  Pr(y=NotInLF|x):    0.2248    0.3945   -0.1696  [-0.2554,  -0.0839]

added matrices:
    e(_estadd_prvalue) :  1 x 12
  e(_estadd_prvalue_x) :  1 x 7

. quietly prvalue, x(k5=1 wc=0) save

. estadd  prvalue, x(k5=1 wc=1) label(k5 = 1) brief diff

probit: Change in Predictions for lfp

                     Current     Saved    Change   95% CI for Change
  Pr(y=inLF|x):       0.4527    0.2719    0.1808  [ 0.0803,   0.2814]
  Pr(y=NotInLF|x):    0.5473    0.7281   -0.1808  [-0.2814,  -0.0803]

updated matrices:
    e(_estadd_prvalue) :  2 x 12
  e(_estadd_prvalue_x) :  2 x 7

. quietly prvalue, x(k5=2 wc=0) save

. estadd  prvalue, x(k5=2 wc=1) label(k5 = 2) brief diff

probit: Change in Predictions for lfp

                     Current     Saved    Change   95% CI for Change
  Pr(y=inLF|x):       0.1602    0.0692    0.0910  [ 0.0217,   0.1604]
  Pr(y=NotInLF|x):    0.8398    0.9308   -0.0910  [-0.1604,  -0.0217]

updated matrices:
    e(_estadd_prvalue) :  3 x 12
  e(_estadd_prvalue_x) :  3 x 7

. quietly prvalue, x(k5=3 wc=0) save

. estadd  prvalue, x(k5=3 wc=1) label(k5 = 3) brief diff

probit: Change in Predictions for lfp

                     Current     Saved    Change   95% CI for Change
  Pr(y=inLF|x):       0.0309    0.0092    0.0216  [-0.0090,   0.0523]
  Pr(y=NotInLF|x):    0.9691    0.9908   -0.0216  [-0.0523,   0.0090]

updated matrices:
    e(_estadd_prvalue) :  4 x 12
  e(_estadd_prvalue_x) :  4 x 7

. estadd prvalue post Difference

scalars:
                  e(N) =  753

macros:
             e(depvar) : "lfp"
                e(cmd) : "estadd_prvalue"
              e(model) : "probit"
         e(properties) : "b"

matrices:
                  e(b) :  1 x 8      (predictions)
                 e(se) :  1 x 8      (standard errors)
                 e(LB) :  1 x 8      (lower CI bounds)
                 e(UB) :  1 x 8      (upper CI bounds)
           e(Category) :  1 x 8      (outcome values)
                  e(X) :  7 x 4      (k5, k618, age, wc, hc, lwg, inc)

results stored as Difference

. esttab, se nostar nonumber noobs mtitles ///
>     keep(inLF:) eqlabels(none)

---------------------------------------------------
                NoCollege      College   Difference
---------------------------------------------------
k5 = 0              0.606        0.775        0.170
                 (0.0251)     (0.0348)     (0.0437)

k5 = 1              0.272        0.453        0.181
                 (0.0358)     (0.0536)     (0.0513)

k5 = 2             0.0692        0.160       0.0910
                 (0.0282)     (0.0539)     (0.0354)

k5 = 3            0.00922       0.0309       0.0216
                (0.00800)     (0.0226)     (0.0157)
---------------------------------------------------
Standard errors in parentheses

. eststo clear

[do-file]

Complementary log-log regression

cloglog and fitstat/listcoef

. spex binlfp2
(Data from 1976 PSID-T Mroz)

. quietly cloglog lfp k5 k618 age wc hc lwg inc, nolog

. estadd fitstat

Measures of Fit for cloglog of lfp

Log-Lik Intercept Only:       -514.873   Log-Lik Full Model:           -448.471
D(745):                        896.943   LR(7):                         132.804
                                         Prob > LR:                       0.000
McFadden's R2:                   0.129   McFadden's Adj R2:               0.113
ML (Cox-Snell) R2:               0.162   Cragg-Uhler(Nagelkerke) R2:      0.217
Efron's R2:                      0.160                              
Count R2:                        0.687   Adj Count R2:                    0.274
AIC:                             1.212   AIC*n:                         912.943
BIC:                         -4037.986   BIC':                          -86.435
BIC used by Stata:             949.935   AIC used by Stata:             912.943

added scalars:
                e(dev) =  896.9429
             e(dev_df) =  745
               e(lrx2) =  132.80351
            e(lrx2_df) =  7
             e(lrx2_p) =  1.631e-25
              e(r2_mf) =  .1289672
           e(r2_mfadj) =  .11342939
              e(r2_ml) =  .16168879
              e(r2_cu) =  .21695524
              e(r2_ef) =  .15960051
              e(r2_ct) =  .68658699
           e(r2_ctadj) =  .27384615
               e(aic0) =  1.2124076
              e(aic_n) =  912.9429
               e(bic0) =  -4037.9857
              e(bic_p) =  -86.435051
           e(statabic) =  949.93542
           e(stataaic) =  912.9429
              e(n_rhs) =  7
             e(n_parm) =  8

. estadd listcoef

cloglog (N=753): Unstandardized and Standardized Estimates 

 Observed SD: .49562951

-------------------------------------------------------------
         lfp |      b         z     P>|z|    bStdX      SDofX
-------------+-----------------------------------------------
          k5 |  -1.00288   -7.101   0.000  -0.5255     0.5240
        k618 |  -0.05225   -1.197   0.231  -0.0690     1.3199
         age |  -0.04036   -5.047   0.000  -0.3258     8.0726
          wc |   0.41893    2.877   0.004   0.1885     0.4500
          hc |   0.05546    0.408   0.683   0.0271     0.4885
         lwg |   0.58236    4.781   0.000   0.3422     0.5876
         inc |  -0.02493   -4.157   0.000  -0.2900    11.6348
-------------------------------------------------------------

added matrices:
               e(b_xs) :  1 x 7      (bStdX)
              e(b_sdx) :  1 x 7      (SDofX)

. esttab, cell("b b_xs b_sdx") scalars(r2_mf r2_mfadj r2_ml r2_cu)

---------------------------------------------------
                      (1)                          
                      lfp                          
                        b         b_xs        b_sdx
---------------------------------------------------
lfp                                                
k5              -1.002878     -.525467      .523959
k618            -.0522477    -.0689604     1.319874
age             -.0403616    -.3258222     8.072574
wc               .4189326     .1885404     .4500494
hc               .0554553     .0270882     .4884694
lwg              .5823638     .3421716     .5875564
inc             -.0249275    -.2900264      11.6348
_cons            1.554071                          
---------------------------------------------------
N                     753                          
r2_mf               0.129                          
r2_mfadj            0.113                          
r2_ml               0.162                          
r2_cu               0.217                          
---------------------------------------------------

[do-file]

cloglog and prvalue

. spex binlfp2
(Data from 1976 PSID-T Mroz)

. quietly cloglog lfp k5 k618 age wc hc lwg inc, nolog

. estadd prvalue, x(age=35 k5=2 wc=0 hc=0 inc=15) ///
>     label(family type 1) brief

cloglog: Predictions for lfp

                                95% Conf. Interval
  Pr(y=inLF|x):       0.1716   [ 0.0931,    0.2500]
  Pr(y=NotInLF|x):    0.8284   [ 0.7500,    0.9069]

added matrices:
    e(_estadd_prvalue) :  1 x 12
  e(_estadd_prvalue_x) :  1 x 7

. estadd prvalue, x(age=50 k5=0 k618=0 wc=1 hc=1) ///
>     label(family type 2) brief

cloglog: Predictions for lfp

                                95% Conf. Interval
  Pr(y=inLF|x):       0.6862   [ 0.5920,    0.7803]
  Pr(y=NotInLF|x):    0.3138   [ 0.2197,    0.4080]

updated matrices:
    e(_estadd_prvalue) :  2 x 12
  e(_estadd_prvalue_x) :  2 x 7

. estadd prvalue, label(average family) brief

cloglog: Predictions for lfp

                                95% Conf. Interval
  Pr(y=inLF|x):       0.5608   [ 0.5225,    0.5991]
  Pr(y=NotInLF|x):    0.4392   [ 0.4009,    0.4775]

updated matrices:
    e(_estadd_prvalue) :  3 x 12
  e(_estadd_prvalue_x) :  3 x 7

. estadd prvalue post

scalars:
                  e(N) =  753

macros:
             e(depvar) : "lfp"
                e(cmd) : "estadd_prvalue"
              e(model) : "cloglog"
         e(properties) : "b"

matrices:
                  e(b) :  1 x 6      (predictions)
                 e(se) :  1 x 6      (standard errors)
                 e(LB) :  1 x 6      (lower CI bounds)
                 e(UB) :  1 x 6      (upper CI bounds)
           e(Category) :  1 x 6      (outcome values)
                  e(X) :  7 x 3      (k5, k618, age, wc, hc, lwg, inc)

. esttab, ci wide nostar ///
>     keep(inLF:) eqlabels(none) varwidth(15)

---------------------------------------------------
                         (1)                       
                         lfp                       
---------------------------------------------------
family type 1          0.172         [0.0931,0.250]
family type 2          0.686          [0.592,0.780]
average family         0.561          [0.522,0.599]
---------------------------------------------------
N                        753                       
---------------------------------------------------
95% confidence intervals in brackets

[do-file]

Ordered logit/probit

ologit and brant

. spex ordwarm2
(77 & 89 General Social Survey)

. quietly ologit warm yr89 male white age ed prst

. estadd brant

Brant Test of Parallel Regression Assumption

    Variable |      chi2   p>chi2    df
-------------+--------------------------
         All |     49.18    0.000    12
-------------+--------------------------
        yr89 |     13.01    0.001     2
        male |     22.24    0.000     2
       white |      1.27    0.531     2
         age |      7.38    0.025     2
          ed |      4.31    0.116     2
        prst |      4.33    0.115     2
----------------------------------------

A significant test statistic provides evidence that the parallel
regression assumption has been violated.

added scalars:
         e(brant_chi2) =  49.181219
           e(brant_df) =  12
            e(brant_p) =  1.944e-06

added matrix:
              e(brant) :  2 x 6      (chi2, p>chi2)

. esttab, cell("b t brant[chi2] brant[p>chi2]") ///
>     scalars(brant_chi2 brant_df brant_p) ///
>     eqlabels(none)

----------------------------------------------------------------
                      (1)                                       
                     warm                                       
                        b            t         chi2       p>chi2
----------------------------------------------------------------
yr89             .5239025     6.557071     13.01311     .0014936
male            -.7332997    -9.343457      22.2379     .0000148
white           -.3911595    -3.304247     1.267856      .530504
age             -.0216655    -8.777619     7.383264     .0249313
ed               .0671728     4.204878     4.310353     .1158828
prst             .0060727     1.844178     4.331991     .1146358
cut1            -2.465362    -10.31909                          
cut2             -.630904     -2.70408                          
cut3             1.261854     5.392123                          
----------------------------------------------------------------
N                    2293                                       
brant_chi2          49.18                                       
brant_df               12                                       
brant_p        0.00000194                                       
----------------------------------------------------------------

[do-file]

Ordered logit/probit and fitstat

. spex ordwarm2
(77 & 89 General Social Survey)

. quietly ologit warm yr89 male white age ed prst, nolog

. estadd fitstat

Measures of Fit for ologit of warm

Log-Lik Intercept Only:      -2995.770   Log-Lik Full Model:          -2844.912
D(2284):                      5689.825   LR(6):                         301.716
                                         Prob > LR:                       0.000
McFadden's R2:                   0.050   McFadden's Adj R2:               0.047
ML (Cox-Snell) R2:               0.123   Cragg-Uhler(Nagelkerke) R2:      0.133
McKelvey & Zavoina's R2:         0.127                              
Variance of y*:                  3.768   Variance of error:               3.290
Count R2:                        0.432   Adj Count R2:                    0.093
AIC:                             2.489   AIC*n:                        5707.825
BIC:                        -11982.891   BIC':                         -255.291
BIC used by Stata:            5759.463   AIC used by Stata:            5707.825

added scalars:
                e(dev) =  5689.8246
             e(dev_df) =  2284
               e(lrx2) =  301.71628
            e(lrx2_df) =  6
             e(lrx2_p) =  3.508e-62
              e(r2_mf) =  .05035704
           e(r2_mfadj) =  .04735281
              e(r2_ml) =  .12329214
              e(r2_cu) =  .13304665
              e(r2_mz) =  .12682954
            e(v_ystar) =  3.7677272
            e(v_error) =  3.2898681
              e(r2_ct) =  .4317488
           e(r2_ctadj) =  .09324983
               e(aic0) =  2.4892388
              e(aic_n) =  5707.8246
               e(bic0) =  -11982.891
              e(bic_p) =  -255.29058
           e(statabic) =  5759.4631
           e(stataaic) =  5707.8246
              e(n_rhs) =  6
             e(n_parm) =  9

. eststo ologit

. quietly oprobit warm yr89 male white age ed prst, nolog

. estadd fitstat

Measures of Fit for oprobit of warm

Log-Lik Intercept Only:      -2995.770   Log-Lik Full Model:          -2848.611
D(2284):                      5697.222   LR(6):                         294.319
                                         Prob > LR:                       0.000
McFadden's R2:                   0.049   McFadden's Adj R2:               0.046
ML (Cox-Snell) R2:               0.120   Cragg-Uhler(Nagelkerke) R2:      0.130
McKelvey & Zavoina's R2:         0.136                              
Variance of y*:                  1.158   Variance of error:               1.000
Count R2:                        0.429   Adj Count R2:                    0.089
AIC:                             2.492   AIC*n:                        5715.222
BIC:                        -11975.494   BIC':                         -247.893
BIC used by Stata:            5766.861   AIC used by Stata:            5715.222

added scalars:
                e(dev) =  5697.222
             e(dev_df) =  2284
               e(lrx2) =  294.31886
            e(lrx2_df) =  6
             e(lrx2_p) =  1.349e-60
              e(r2_mf) =  .0491224
           e(r2_mfadj) =  .04611816
              e(r2_ml) =  .12045924
              e(r2_cu) =  .12998962
              e(r2_mz) =  .1363472
            e(v_ystar) =  1.1578727
            e(v_error) =  1
              e(r2_ct) =  .42913214
           e(r2_ctadj) =  .08907446
               e(aic0) =  2.4924649
              e(aic_n) =  5715.222
               e(bic0) =  -11975.494
              e(bic_p) =  -247.89316
           e(statabic) =  5766.8605
           e(stataaic) =  5715.222
              e(n_rhs) =  6
             e(n_parm) =  9

. eststo oprobit

. esttab, scalars(r2_mf r2_mfadj r2_ml r2_cu) wide eqlabels(none) mtitles

----------------------------------------------------------------------
                      (1)                          (2)                
                   ologit                      oprobit                
----------------------------------------------------------------------
yr89                0.524***       (6.56)        0.319***       (6.80)
male               -0.733***      (-9.34)       -0.417***      (-9.16)
white              -0.391***      (-3.30)       -0.227**       (-3.26)
age               -0.0217***      (-8.78)      -0.0122***      (-8.47)
ed                 0.0672***       (4.20)       0.0387***       (4.15)
prst              0.00607          (1.84)      0.00328          (1.71)
cut1               -2.465***     (-10.32)       -1.429***     (-10.29)
cut2               -0.631**       (-2.70)       -0.361**       (-2.63)
cut3                1.262***       (5.39)        0.768***       (5.60)
----------------------------------------------------------------------
N                    2293                         2293                
r2_mf              0.0504                       0.0491                
r2_mfadj           0.0474                       0.0461                
r2_ml               0.123                        0.120                
r2_cu               0.133                        0.130                
----------------------------------------------------------------------
t statistics in parentheses
* p<0.05, ** p<0.01, *** p<0.001

. eststo clear

[do-file]

Ordered logit/probit and listcoef

. spex ordwarm2
(77 & 89 General Social Survey)

. quietly ologit warm yr89 male white age ed prst, nolog

. estadd listcoef, std

ologit (N=2293): Unstandardized and Standardized Estimates 

 Observed SD: .9282156
   Latent SD: 1.9410634

-------------------------------------------------------------------------------
        warm |      b         z     P>|z|    bStdX    bStdY   bStdXY      SDofX
-------------+-----------------------------------------------------------------
        yr89 |   0.52390    6.557   0.000   0.2566   0.2699   0.1322     0.4897
        male |  -0.73330   -9.343   0.000  -0.3658  -0.3778  -0.1885     0.4989
       white |  -0.39116   -3.304   0.001  -0.1287  -0.2015  -0.0663     0.3290
         age |  -0.02167   -8.778   0.000  -0.3635  -0.0112  -0.1873    16.7790
          ed |   0.06717    4.205   0.000   0.2123   0.0346   0.1094     3.1608
        prst |   0.00607    1.844   0.065   0.0880   0.0031   0.0453    14.4923
-------------------------------------------------------------------------------

added matrices:
               e(b_xs) :  1 x 6      (bStdX)
               e(b_ys) :  1 x 6      (bStdY)
              e(b_std) :  1 x 6      (bStdXY)
              e(b_sdx) :  1 x 6      (SDofX)

. eststo ologit

. quietly oprobit warm yr89 male white age ed prst, nolog

. estadd listcoef

oprobit (N=2293): Unstandardized and Standardized Estimates 

 Observed SD: .9282156
   Latent SD: 1.0760449

-------------------------------------------------------------------------------
        warm |      b         z     P>|z|    bStdX    bStdY   bStdXY      SDofX
-------------+-----------------------------------------------------------------
        yr89 |   0.31881    6.805   0.000   0.1561   0.2963   0.1451     0.4897
        male |  -0.41703   -9.156   0.000  -0.2080  -0.3876  -0.1933     0.4989
       white |  -0.22650   -3.260   0.001  -0.0745  -0.2105  -0.0693     0.3290
         age |  -0.01222   -8.471   0.000  -0.2051  -0.0114  -0.1906    16.7790
          ed |   0.03872    4.153   0.000   0.1224   0.0360   0.1137     3.1608
        prst |   0.00328    1.705   0.088   0.0476   0.0031   0.0442    14.4923
-------------------------------------------------------------------------------

added matrices:
               e(b_xs) :  1 x 6      (bStdX)
               e(b_ys) :  1 x 6      (bStdY)
              e(b_std) :  1 x 6      (bStdXY)
              e(b_sdx) :  1 x 6      (SDofX)

. eststo oprobit

. esttab, aux(b_std) nopar wide eqlabels(none) mtitles

----------------------------------------------------------------------
                      (1)                          (2)                
                   ologit                      oprobit                
----------------------------------------------------------------------
yr89                0.524***        0.132        0.319***        0.145
male               -0.733***       -0.188       -0.417***       -0.193
white              -0.391***      -0.0663       -0.227**       -0.0693
age               -0.0217***       -0.187      -0.0122***       -0.191
ed                 0.0672***        0.109       0.0387***        0.114
prst              0.00607          0.0453      0.00328          0.0442
cut1               -2.465***                    -1.429***             
cut2               -0.631**                     -0.361**              
cut3                1.262***                     0.768***             
----------------------------------------------------------------------
N                    2293                         2293                
----------------------------------------------------------------------
b_std in second column
* p<0.05, ** p<0.01, *** p<0.001

. eststo clear

[do-file]

Ordered logit/probit and prchange

. spex ordwarm2
(77 & 89 General Social Survey)

. eststo ologit: quietly ologit warm yr89 male white age ed prst, nolog

. eststo oprobit: quietly oprobit warm yr89 male white age ed prst, nolog

. estadd prchange male age prst: *

. esttab, main(dc) nostar not mtitles

--------------------------------------
                      (1)          (2)
                   ologit      oprobit
--------------------------------------
Avg|Chg|                              
male               0.0896       0.0819
age                0.0447       0.0404
prst               0.0108      0.00938
--------------------------------------
1SD                                   
male               0.0746       0.0810
age                0.0360       0.0390
prst             -0.00870     -0.00905
--------------------------------------
2D                                    
male                0.105       0.0827
age                0.0533       0.0417
prst              -0.0130     -0.00972
--------------------------------------
3A                                    
male              -0.0814      -0.0622
age               -0.0401      -0.0301
prst              0.00977      0.00702
--------------------------------------
4SA                                   
male              -0.0979       -0.101
age               -0.0492      -0.0506
prst               0.0119       0.0117
--------------------------------------
N                    2293         2293
--------------------------------------
dc coefficients

. eststo clear

[do-file]

Ordered logit/probit and prchange: selected outcome

. spex ordwarm2
(77 & 89 General Social Survey)

. quietly ologit warm yr89 male white age ed prst, nolog

. estadd prchange male age prst, outcome(2)

ologit: Changes in Probabilities for warm

Outcome: 2 (2D)

        Min->Max        0->1       -+1/2      -+sd/2    MargEfct
male   .10462105   .10462105   .10556346   .05362016   .10812605
 age    .1862759   .00289795   .00319454   .05328724   .00319461
prst  -.06301633  -.00080028  -.00089544  -.01297233  -.00089543

               1SD         2D         3A        4SA
Pr(y|x)  .11125716  .32816544  .39936733  .16121005

          yr89     male    white      age       ed     prst
   x=  .398604  .464893  .876581  44.9355  12.2181  39.5853
sd_x=  .489718  .498875  .328989   16.779  3.16083  14.4923

added scalars:
            e(predval) =  .32816544
            e(outcome) =  2
              e(delta) =  1
           e(centered) =  1

added matrices:
                 e(dc) :  6 x 3      (main, Min->Max, 0->1, -+1/2, -+sd/2, Marg
> Efct)
            e(pattern) :  1 x 3
                  e(X) :  4 x 6      (X, SD, Min, Max)

first row in e(dc) contains:

  01 change for binary variables
  sd change for continuous variables

. eststo ologit

. quietly oprobit warm yr89 male white age ed prst, nolog

. estadd prchange male age prst, outcome(2)

oprobit: Changes in Probabilities for warm

Outcome: 2 (2D)

        Min->Max        0->1       -+1/2      -+sd/2    MargEfct
male   .08271328   .08271328   .08378038    .0423308   .08521085
 age    .1476199   .00280122   .00249711   .04172876   .00249716
prst  -.04764727  -.00058544  -.00067082  -.00971937  -.00067081

               1SD         2D         3A        4SA
Pr(y|x)  .11177309  .32895118  .39563131   .1636444

          yr89     male    white      age       ed     prst
   x=  .398604  .464893  .876581  44.9355  12.2181  39.5853
sd_x=  .489718  .498875  .328989   16.779  3.16083  14.4923

added scalars:
            e(predval) =  .32895118
            e(outcome) =  2
              e(delta) =  1
           e(centered) =  1

added matrices:
                 e(dc) :  6 x 3      (main, Min->Max, 0->1, -+1/2, -+sd/2, Marg
> Efct)
            e(pattern) :  1 x 3
                  e(X) :  4 x 6      (X, SD, Min, Max)

first row in e(dc) contains:

  01 change for binary variables
  sd change for continuous variables

. eststo oprobit

. esttab, main(dc) nostar not stats(predval outcome) mtitles

--------------------------------------
                      (1)          (2)
                   ologit      oprobit
--------------------------------------
male                0.105       0.0827
age                0.0533       0.0417
prst              -0.0130     -0.00972
--------------------------------------
predval             0.328        0.329
outcome                 2            2
--------------------------------------
dc coefficients

. eststo clear

[do-file]

ologit and prchange: split option

. spex ordwarm2
(77 & 89 General Social Survey)

. quietly ologit warm yr89 male white age ed prst, nolog

. estadd prchange male age prst, split

ologit: Changes in Probabilities for warm

male
        Avg|Chg|         1SD          2D          3A         4SA
0->1   .08961766   .07461427   .10462105  -.08137083  -.09786447

age
            Avg|Chg|         1SD          2D          3A         4SA
Min->Max   .18319855   .18012119    .1862759  -.17905769  -.18733941
   -+1/2   .00266841   .00214228   .00319454  -.00240716  -.00292964
  -+sd/2    .0446563   .03602537   .05328724   -.0401054   -.0492072
MargEfct   .00266844   .00214226   .00319461  -.00240723  -.00292964

prst
            Avg|Chg|         1SD          2D          3A         4SA
Min->Max   .05186236  -.04070839  -.06301633   .04440692   .05931778
   -+1/2   .00074795  -.00060046  -.00089544   .00067475   .00082116
  -+sd/2   .01083777  -.00870322  -.01297233   .00977433    .0119012
MargEfct   .00074795  -.00060046  -.00089543   .00067473   .00082116

               1SD         2D         3A        4SA
Pr(y|x)  .11125716  .32816544  .39936733  .16121005

          yr89     male    white      age       ed     prst
   x=  .398604  .464893  .876581  44.9355  12.2181  39.5853
sd_x=  .489718  .498875  .328989   16.779  3.16083  14.4923

added scalars:
            e(predval) =  .11125716
            e(outcome) =  1
              e(delta) =  1
           e(centered) =  1

added matrices:
                 e(dc) :  6 x 3      (main, Min->Max, 0->1, -+1/2, -+sd/2, Marg
> Efct)
            e(pattern) :  1 x 3
                  e(X) :  4 x 6      (X, SD, Min, Max)

first row in e(dc) contains:

  01 change for binary variables
  sd change for continuous variables

results for outcome 1 stored as ologit_1
results for outcome 2 stored as ologit_2
results for outcome 3 stored as ologit_3
results for outcome 4 stored as ologit_4

. esttab, main(dc) nostar not stats(predval outcome) ///
>     mtitles nonumbers

----------------------------------------------------------------
                      1SD           2D           3A          4SA
----------------------------------------------------------------
male               0.0746        0.105      -0.0814      -0.0979
age                0.0360       0.0533      -0.0401      -0.0492
prst             -0.00870      -0.0130      0.00977       0.0119
----------------------------------------------------------------
predval             0.111        0.328        0.399        0.161
outcome                 1            2            3            4
----------------------------------------------------------------
dc coefficients

. eststo clear

[do-file]

oprobit and prchange: split option

. spex ordwarm2
(77 & 89 General Social Survey)

. quietly oprobit warm yr89 male white age ed prst, nolog

. estadd prchange male age prst, split

oprobit: Changes in Probabilities for warm

male
        Avg|Chg|         1SD          2D          3A         4SA
0->1   .08185343   .08099359   .08271328  -.06222743  -.10147941

age
            Avg|Chg|         1SD          2D          3A         4SA
Min->Max   .16764789   .18767587    .1476199  -.14207253  -.19322326
   -+1/2   .00241081   .00232452   .00249711  -.00180408  -.00301754
  -+sd/2   .04038233   .03903589   .04172876  -.03013682  -.05062783
MargEfct   .00241084   .00232452   .00249716  -.00180413  -.00301755

prst
            Avg|Chg|         1SD          2D          3A         4SA
Min->Max   .04505957  -.04247186  -.04764727   .03203639   .05808274
   -+1/2   .00064762  -.00062443  -.00067082   .00048462   .00081059
  -+sd/2   .00938461  -.00904983  -.00971937   .00702184   .01174738
MargEfct   .00064762  -.00062443  -.00067081   .00048464    .0008106

               1SD         2D         3A        4SA
Pr(y|x)  .11177309  .32895118  .39563131   .1636444

          yr89     male    white      age       ed     prst
   x=  .398604  .464893  .876581  44.9355  12.2181  39.5853
sd_x=  .489718  .498875  .328989   16.779  3.16083  14.4923

added scalars:
            e(predval) =  .11177309
            e(outcome) =  1
              e(delta) =  1
           e(centered) =  1

added matrices:
                 e(dc) :  6 x 3      (main, Min->Max, 0->1, -+1/2, -+sd/2, Marg
> Efct)
            e(pattern) :  1 x 3
                  e(X) :  4 x 6      (X, SD, Min, Max)

first row in e(dc) contains:

  01 change for binary variables
  sd change for continuous variables

results for outcome 1 stored as oprobit_1
results for outcome 2 stored as oprobit_2
results for outcome 3 stored as oprobit_3
results for outcome 4 stored as oprobit_4

. esttab, main(dc) nostar not stats(predval outcome) ///
>     mtitles nonumbers

----------------------------------------------------------------
                      1SD           2D           3A          4SA
----------------------------------------------------------------
male               0.0810       0.0827      -0.0622       -0.101
age                0.0390       0.0417      -0.0301      -0.0506
prst             -0.00905     -0.00972      0.00702       0.0117
----------------------------------------------------------------
predval             0.112        0.329        0.396        0.164
outcome                 1            2            3            4
----------------------------------------------------------------
dc coefficients

. eststo clear

[do-file]

ologit and prvalue

. spex ordwarm2
(77 & 89 General Social Survey)

. quietly ologit warm yr89 male white age ed prst, nolog

. estadd prvalue, x(yr89=0 male=1 prst=20 age=64 ed=16) ///
>     brief label(type1)

ologit: Predictions for warm

                                95% Conf. Interval
     Pr(y=1SD|x):     0.2317   [ 0.1776,    0.2857]
     Pr(y=2D|x):      0.4221   [ 0.3942,    0.4500]
     Pr(y=3A|x):      0.2723   [ 0.2249,    0.3198]
     Pr(y=4SA|x):     0.0739   [ 0.0523,    0.0954]

added matrices:
    e(_estadd_prvalue) :  1 x 24
  e(_estadd_prvalue_x) :  1 x 6

. estadd prvalue, x(yr89=1 male=0 prst=80 age=30 ed=24) ///
>     brief label(type2)

ologit: Predictions for warm

                                95% Conf. Interval
     Pr(y=1SD|x):     0.0164   [ 0.0106,    0.0222]
     Pr(y=2D|x):      0.0781   [ 0.0554,    0.1008]
     Pr(y=3A|x):      0.3147   [ 0.2636,    0.3658]
     Pr(y=4SA|x):     0.5908   [ 0.5143,    0.6673]

updated matrices:
    e(_estadd_prvalue) :  2 x 24
  e(_estadd_prvalue_x) :  2 x 6

. estadd prvalue, x(yr89=0) brief label(type3)

ologit: Predictions for warm

                                95% Conf. Interval
     Pr(y=1SD|x):     0.1336   [ 0.1176,    0.1496]
     Pr(y=2D|x):      0.3577   [ 0.3348,    0.3806]
     Pr(y=3A|x):      0.3737   [ 0.3517,    0.3957]
     Pr(y=4SA|x):     0.1349   [ 0.1195,    0.1504]

updated matrices:
    e(_estadd_prvalue) :  3 x 24
  e(_estadd_prvalue_x) :  3 x 6

. estadd prvalue, x(yr89=1) brief label(type4)

ologit: Predictions for warm

                                95% Conf. Interval
     Pr(y=1SD|x):     0.0837   [ 0.0711,    0.0963]
     Pr(y=2D|x):      0.2802   [ 0.2571,    0.3032]
     Pr(y=3A|x):      0.4277   [ 0.4046,    0.4507]
     Pr(y=4SA|x):     0.2085   [ 0.1855,    0.2315]

updated matrices:
    e(_estadd_prvalue) :  4 x 24
  e(_estadd_prvalue_x) :  4 x 6

. estadd prvalue post

scalars:
                  e(N) =  2293

macros:
             e(depvar) : "warm"
                e(cmd) : "estadd_prvalue"
              e(model) : "ologit"
         e(properties) : "b"

matrices:
                  e(b) :  1 x 16     (predictions)
                 e(se) :  1 x 16     (standard errors)
                 e(LB) :  1 x 16     (lower CI bounds)
                 e(UB) :  1 x 16     (upper CI bounds)
           e(Category) :  1 x 16     (outcome values)
                  e(X) :  6 x 4      (yr89, male, white, age, ed, prst)

. esttab, nostar unstack ///
>     coeflabels(type1 "old working class men 1977"   ///
>                type2 "young prestigious women 1989" ///
>                type3 "average individual 1977" ///
>                type4 "average individual 1989") ///
>     wrap varwidth(18)

----------------------------------------------------------------------
                            (1)                                       
                           warm                                       
                            1SD           2D           3A          4SA
----------------------------------------------------------------------
old working class         0.232        0.422        0.272       0.0739
men 1977                 (8.40)      (29.67)      (11.25)       (6.72)

young prestigious        0.0164       0.0781        0.315        0.591
women 1989               (5.56)       (6.74)      (12.07)      (15.13)

average individual        0.134        0.358        0.374        0.135
1977                    (16.37)      (30.57)      (33.30)      (17.09)

average individual       0.0837        0.280        0.428        0.208
1989                    (13.00)      (23.81)      (36.29)      (17.77)
----------------------------------------------------------------------
N                          2293                                       
----------------------------------------------------------------------
t statistics in parentheses

[do-file]

oprobit and prvalue

. spex ordwarm2
(77 & 89 General Social Survey)

. quietly oprobit warm yr89 male white age ed prst, nolog

. estadd prvalue, x(yr89=0 male=1 prst=20 age=64 ed=16) ///
>     brief label(type1)

oprobit: Predictions for warm

                                95% Conf. Interval
     Pr(y=1SD|x):     0.2370   [ 0.1821,    0.2918]
     Pr(y=2D|x):      0.4006   [ 0.3750,    0.4261]
     Pr(y=3A|x):      0.2931   [ 0.2844,    0.3017]
     Pr(y=4SA|x):     0.0693   [ 0.0450,    0.0936]

added matrices:
    e(_estadd_prvalue) :  1 x 24
  e(_estadd_prvalue_x) :  1 x 6

. estadd prvalue, x(yr89=1 male=0 prst=80 age=30 ed=24) ///
>     brief label(type2)

oprobit: Predictions for warm

                                95% Conf. Interval
     Pr(y=1SD|x):     0.0088   [ 0.0040,    0.0136]
     Pr(y=2D|x):      0.0870   [ 0.0754,    0.0985]
     Pr(y=3A|x):      0.3338   [ 0.3083,    0.3593]
     Pr(y=4SA|x):     0.5704   [ 0.4977,    0.6432]

updated matrices:
    e(_estadd_prvalue) :  2 x 24
  e(_estadd_prvalue_x) :  2 x 6

. estadd prvalue, x(yr89=0) brief label(type3)

oprobit: Predictions for warm

                                95% Conf. Interval
     Pr(y=1SD|x):     0.1378   [ 0.1213,    0.1544]
     Pr(y=2D|x):      0.3534   [ 0.3262,    0.3805]
     Pr(y=3A|x):      0.3746   [ 0.3605,    0.3886]
     Pr(y=4SA|x):     0.1342   [ 0.1182,    0.1502]

updated matrices:
    e(_estadd_prvalue) :  3 x 24
  e(_estadd_prvalue_x) :  3 x 6

. estadd prvalue, x(yr89=1) brief label(type4)

oprobit: Predictions for warm

                                95% Conf. Interval
     Pr(y=1SD|x):     0.0794   [ 0.0660,    0.0929]
     Pr(y=2D|x):      0.2872   [ 0.2615,    0.3128]
     Pr(y=3A|x):      0.4180   [ 0.3990,    0.4370]
     Pr(y=4SA|x):     0.2154   [ 0.1917,    0.2391]

updated matrices:
    e(_estadd_prvalue) :  4 x 24
  e(_estadd_prvalue_x) :  4 x 6

. estadd prvalue post

scalars:
                  e(N) =  2293

macros:
             e(depvar) : "warm"
                e(cmd) : "estadd_prvalue"
              e(model) : "oprobit"
         e(properties) : "b"

matrices:
                  e(b) :  1 x 16     (predictions)
                 e(se) :  1 x 16     (standard errors)
                 e(LB) :  1 x 16     (lower CI bounds)
                 e(UB) :  1 x 16     (upper CI bounds)
           e(Category) :  1 x 16     (outcome values)
                  e(X) :  6 x 4      (yr89, male, white, age, ed, prst)

. esttab, nostar unstack ///
>     coeflabels(type1 "old working class men 1977"   ///
>                type2 "young prestigious women 1989" ///
>                type3 "average individual 1977" ///
>                type4 "average individual 1989") ///
>     wrap varwidth(18)

----------------------------------------------------------------------
                            (1)                                       
                           warm                                       
                            1SD           2D           3A          4SA
----------------------------------------------------------------------
old working class         0.237        0.401        0.293       0.0693
men 1977                 (8.47)      (30.74)      (66.40)       (5.59)

young prestigious       0.00879       0.0870        0.334        0.570
women 1989               (3.59)      (14.72)      (25.67)      (15.37)

average individual        0.138        0.353        0.375        0.134
1977                    (16.33)      (25.49)      (52.31)      (16.44)

average individual       0.0794        0.287        0.418        0.215
1989                    (11.57)      (21.95)      (43.16)      (17.83)
----------------------------------------------------------------------
N                          2293                                       
----------------------------------------------------------------------
t statistics in parentheses

[do-file]

Multinomial logit/probit

Multinomial logit and fitstat

. spex nomocc2
(1982 General Social Survey)

. quietly mlogit occ white ed exper, nolog

. estadd fitstat

Measures of Fit for mlogit of occ

Log-Lik Intercept Only:       -509.844   Log-Lik Full Model:           -426.800
D(321):                        853.601   LR(12):                        166.087
                                         Prob > LR:                       0.000
McFadden's R2:                   0.163   McFadden's Adj R2:               0.131
ML (Cox-Snell) R2:               0.389   Cragg-Uhler(Nagelkerke) R2:      0.409
Count R2:                        0.501   Adj Count R2:                    0.253
AIC:                             2.628   AIC*n:                         885.601
BIC:                         -1014.646   BIC':                          -96.246
BIC used by Stata:             946.722   AIC used by Stata:             885.601

added scalars:
                e(dev) =  853.60095
             e(dev_df) =  321
               e(lrx2) =  166.08716
            e(lrx2_df) =  12
             e(lrx2_p) =  3.010e-29
              e(r2_mf) =  .16288035
           e(r2_mfadj) =  .13149821
              e(r2_ml) =  .38911114
              e(r2_cu) =  .40895353
              e(r2_ct) =  .50148368
           e(r2_ctadj) =  .25333333
               e(aic0) =  2.627896
              e(aic_n) =  885.60095
               e(bic0) =  -1014.6457
              e(bic_p) =  -96.246162
           e(statabic) =  946.72228
           e(stataaic) =  885.60095
              e(n_rhs) =  3
             e(n_parm) =  16

. eststo mlogit

. esttab, wide scalars(r2_ct r2_ctadj aic0 aic_n) mtitles

-----------------------------------------
                      (1)                
                   mlogit                
-----------------------------------------
Menial                                   
white              -1.774*        (-2.35)
ed                 -0.779***      (-6.79)
exper             -0.0357*        (-1.98)
_cons               11.52***       (6.23)
-----------------------------------------
BlueCol                                  
white              -0.538         (-0.67)
ed                 -0.878***      (-8.74)
exper             -0.0309*        (-2.15)
_cons               12.26***       (7.35)
-----------------------------------------
Craft                                    
white              -1.302*        (-2.01)
ed                 -0.685***      (-7.67)
exper            -0.00797         (-0.63)
_cons               10.43***       (6.87)
-----------------------------------------
WhiteCol                                 
white              -0.203         (-0.23)
ed                 -0.426***      (-4.62)
exper            -0.00106         (-0.07)
_cons               5.280**        (3.14)
-----------------------------------------
Prof                                     
o.white                 0             (.)
o.ed                    0             (.)
o.exper                 0             (.)
o._cons                 0             (.)
-----------------------------------------
N                     337                
r2_ct               0.501                
r2_ctadj            0.253                
aic0                2.628                
aic_n               885.6                
-----------------------------------------
t statistics in parentheses
* p<0.05, ** p<0.01, *** p<0.001

. eststo clear

[do-file]

Multinomial logit and listcoef

. spex nomocc2
(1982 General Social Survey)

. quietly mlogit occ white ed exper, nolog

. estadd listcoef, gt adjacent

mlogit (N=337): Factor Change in the Odds of occ 

Variable: white (sd=.27642268)

Odds comparing    |
Alternative 1     |
to Alternative 2  |      b         z     P>|z|     e^b   e^bStdX
------------------+---------------------------------------------
BlueCol -Menial   |   1.23650    1.707   0.088   3.4436   1.4075
Craft   -BlueCol  |  -0.76416   -1.208   0.227   0.4657   0.8096
WhiteCol-Craft    |   1.09904    1.343   0.179   3.0013   1.3550
Prof    -WhiteCol |   0.20292    0.233   0.815   1.2250   1.0577
----------------------------------------------------------------

Variable: ed (sd=2.9464271)

Odds comparing    |
Alternative 1     |
to Alternative 2  |      b         z     P>|z|     e^b   e^bStdX
------------------+---------------------------------------------
BlueCol -Menial   |  -0.09942   -0.972   0.331   0.9054   0.7461
Craft   -BlueCol  |   0.19324    2.494   0.013   1.2132   1.7671
WhiteCol-Craft    |   0.25934    2.773   0.006   1.2961   2.1471
Prof    -WhiteCol |   0.42569    4.616   0.000   1.5307   3.5053
----------------------------------------------------------------

Variable: exper (sd=13.959364)

Odds comparing    |
Alternative 1     |
to Alternative 2  |      b         z     P>|z|     e^b   e^bStdX
------------------+---------------------------------------------
BlueCol -Menial   |   0.00472    0.271   0.786   1.0047   1.0681
Craft   -BlueCol  |   0.02296    1.829   0.067   1.0232   1.3779
WhiteCol-Craft    |   0.00691    0.495   0.621   1.0069   1.1013
Prof    -WhiteCol |   0.00106    0.073   0.941   1.0011   1.0148
----------------------------------------------------------------

added matrices:
              e(b_raw) :  1 x 12     (b)
               e(b_se) :  1 x 12     (se)
                e(b_z) :  1 x 12     (z)
                e(b_p) :  1 x 12     (P>|z|)
             e(b_fact) :  1 x 12     (e^b)
            e(b_facts) :  1 x 12     (e^bStdX)
              e(b_sdx) :  1 x 12     (SDofX)

. esttab , cell("b_raw b_fact b_facts b_sdx") varwidth(14)

------------------------------------------------------------------
                        (1)                                       
                        occ                                       
                      b_raw       b_fact      b_facts        b_sdx
------------------------------------------------------------------
BlueCol-Menial                                                    
white              1.236504     3.443553     1.407476     .2764227
ed                -.0994247     .9053581     .7460612     2.946427
exper              .0047212     1.004732     1.068126     13.95936
------------------------------------------------------------------
Craft-BlueCol                                                     
white             -.7641602     .4657249     .8095869     .2764227
ed                 .1932401     1.213174      1.76715     2.946427
exper              .0229626     1.023228     1.377875     13.95936
------------------------------------------------------------------
WhiteCol-Craft                                                    
white              1.099042     3.001288     1.354998     .2764227
ed                 .2593423     1.296077     2.147132     2.946427
exper              .0069121     1.006936     1.101296     13.95936
------------------------------------------------------------------
Prof-WhiteCol                                                     
white              .2029212     1.224976     1.057695     .2764227
ed                 .4256943     1.530653     3.505304     2.946427
exper               .001055     1.001056     1.014837     13.95936
------------------------------------------------------------------
N                       337                                       
------------------------------------------------------------------

[do-file]

mlogit and mlogtest

. spex nomocc2
(1982 General Social Survey)

. quietly mlogit occ white ed exper, nolog

. estadd mlogtest, wald lr set(white exper)

**** Likelihood-ratio tests for independent variables (N=337)

 Ho: All coefficients associated with given variable(s) are 0.

             |       chi2   df   P>chi2
-------------+-------------------------
       white |      8.095    4    0.088
          ed |    156.937    4    0.000
       exper |      8.561    4    0.073
-------------+-------------------------
      set_1: |     16.452    8    0.036
       white |
       exper |
---------------------------------------

**** Wald tests for independent variables (N=337)

 Ho: All coefficients associated with given variable(s) are 0.

             |       chi2   df   P>chi2
-------------+-------------------------
       white |      8.149    4    0.086
          ed |     84.968    4    0.000
       exper |      7.995    4    0.092
-------------+-------------------------
      set_1: |     15.773    8    0.046
       white |
       exper |
---------------------------------------

added scalars:
     e(wald_set1_chi2) =  15.773146
       e(wald_set1_df) =  8
        e(wald_set1_p) =  .0457446
   e(lrtest_set1_chi2) =  16.451934
     e(lrtest_set1_df) =  8
      e(lrtest_set1_p) =  .03634985

added matrices:
               e(wald) :  3 x 3      (chi2, df, p)
             e(lrtest) :  3 x 3      (chi2, df, p)

. estout, cell("wald[chi2] wald[df] wald[p]")       ///
>     stat(wald_set1_chi2 wald_set1_df wald_set1_p, ///
>         layout("@ @ @") label("white&exper"))     ///
>     mlabel(none)

---------------------------------------------------
                     chi2           df            p
---------------------------------------------------
white            8.149203            4     .0862631
ed               84.96817            4     1.54e-17
exper            7.994939            4     .0917638
---------------------------------------------------
white&exper      15.77315            8     .0457446
---------------------------------------------------

. estout, cell("lrtest[chi2] lrtest[df] lrtest[p]")       ///
>     stat(lrtest_set1_chi2 lrtest_set1_df lrtest_set1_p, ///
>         layout("@ @ @") label("white&exper"))           ///
>     mlabel(none)

---------------------------------------------------
                     chi2           df            p
---------------------------------------------------
white            8.095408            4     .0881451
ed               156.9372            4     6.63e-33
exper            8.560953            4      .073061
---------------------------------------------------
white&exper      16.45193            8     .0363498
---------------------------------------------------

. estout, cell(" wald[p](label(P>Wald) fmt(4)) lrtest[p](label(P>LR))")   ///
>     stat(wald_set1_p lrtest_set1_p, layout("@ @") label("white&exper")) ///
>     mlabel(none)

--------------------------------------
                   P>Wald         P>LR
--------------------------------------
white              0.0863       0.0881
ed                 0.0000       0.0000
exper              0.0918       0.0731
--------------------------------------
white&exper        0.0457       0.0363
--------------------------------------

[do-file]

Multinomial logit/probit and prchange

. spex nomocc2
(1982 General Social Survey)

. quietly mlogit occ white ed exper, nolog

. estadd prchange

mlogit: Changes in Probabilities for occ

white
        Avg|Chg|      Menial     BlueCol       Craft    WhiteCol        Prof
0->1   .11623582  -.13085523   .04981799  -.15973434   .07971004    .1610615

ed
            Avg|Chg|      Menial     BlueCol       Craft    WhiteCol
Min->Max   .39242268  -.13017954  -.70077323  -.15010394   .02425591
   -+1/2   .05855425  -.02559762  -.06831616  -.05247185   .01250795
  -+sd/2    .1640657  -.07129153  -.19310513  -.14576758   .03064777
MargEfct   .05894859  -.02579097  -.06870635  -.05287415   .01282041

                Prof
Min->Max   .95680079
   -+1/2   .13387768
  -+sd/2   .37951647
MargEfct   .13455107

exper
            Avg|Chg|      Menial     BlueCol       Craft    WhiteCol
Min->Max   .12193559  -.11536534  -.18947365   .03115708   .09478889
   -+1/2   .00233425  -.00226997  -.00356567   .00105992    .0016944
  -+sd/2   .03253578  -.03167491  -.04966453   .01479983   .02360725
MargEfct   .00233427  -.00226997  -.00356571   .00105992   .00169442

                Prof
Min->Max   .17889298
   -+1/2   .00308132
  -+sd/2   .04293236
MargEfct   .00308134

            Menial    BlueCol      Craft   WhiteCol       Prof
Pr(y|x)  .09426806  .18419114  .29411051  .16112968  .26630062

         white       ed    exper
   x=  .916914   13.095  20.5015
sd_x=  .276423  2.94643  13.9594

added scalars:
           e(predval1) =  .09426806
           e(predval2) =  .18419114
           e(predval3) =  .29411051
           e(predval4) =  .16112968
           e(predval5) =  .26630062
              e(delta) =  1
           e(centered) =  1

added matrices:
                 e(dc) :  6 x 18     (main, Min->Max, 0->1, -+1/2, -+sd/2, Marg
> Efct)
            e(pattern) :  1 x 18
                  e(X) :  4 x 3      (X, SD, Min, Max)

first row in e(dc) contains:

  01 change for binary variables
  sd change for continuous variables

. eststo mlogit

. quietly mprobit occ white ed exper, nolog

. estadd prchange

mprobit: Changes in Probabilities for occ

white
        Avg|Chg|      Menial     BlueCol       Craft    WhiteCol        Prof
0->1   .11142595  -.13099539   .03923495  -.14756948   .07652746   .16280247

ed
            Avg|Chg|      Menial     BlueCol       Craft    WhiteCol
Min->Max   .39433155  -.13480758  -.63618958  -.21483173   .03136472
   -+1/2   .05591886  -.02555373  -.06636748  -.04787594   .00948766
  -+sd/2    .1577241   -.0714798  -.18783711  -.13499337   .02528359

                Prof
Min->Max   .95446413
   -+1/2   .13030948
  -+sd/2   .36902665

exper
            Avg|Chg|      Menial     BlueCol       Craft    WhiteCol
Min->Max     .123281  -.11218768  -.19601482   .01516485   .07135144
   -+1/2   .00229492  -.00215125  -.00358605   .00069633   .00136708
  -+sd/2   .03197956  -.02998788  -.04996105   .00971583   .01904152

                Prof
Min->Max    .2216862
   -+1/2   .00367388
  -+sd/2   .05119154

            Menial    BlueCol      Craft   WhiteCol       Prof
Pr(y|x)  .09325961  .18944861  .27852002  .15457167   .2842001

         white       ed    exper
   x=  .916914   13.095  20.5015
sd_x=  .276423  2.94643  13.9594

added scalars:
           e(predval1) =  .09325961
           e(predval2) =  .18944861
           e(predval3) =  .27852002
           e(predval4) =  .15457167
           e(predval5) =  .2842001
              e(delta) =  1
           e(centered) =  1

added matrices:
                 e(dc) :  5 x 18     (main, Min->Max, 0->1, -+1/2, -+sd/2)
            e(pattern) :  1 x 18
                  e(X) :  4 x 3      (X, SD, Min, Max)

first row in e(dc) contains:

  01 change for binary variables
  sd change for continuous variables

. eststo mprobit

. esttab mlogit, main(dc) nostar not unstack compress

----------------------------------------------------------------------
                 (1)                                                  
                 occ                                                  
            Avg|Chg|    Menial   BlueCol     Craft  WhiteCol      Prof
----------------------------------------------------------------------
white          0.116    -0.131    0.0498    -0.160    0.0797     0.161
ed             0.164   -0.0713    -0.193    -0.146    0.0306     0.380
exper         0.0325   -0.0317   -0.0497    0.0148    0.0236    0.0429
----------------------------------------------------------------------
N                337                                                  
----------------------------------------------------------------------
dc coefficients

. esttab mprobit, main(dc) nostar not unstack compress

----------------------------------------------------------------------
                 (1)                                                  
                 occ                                                  
            Avg|Chg|    Menial   BlueCol     Craft  WhiteCol      Prof
----------------------------------------------------------------------
white          0.111    -0.131    0.0392    -0.148    0.0765     0.163
ed             0.158   -0.0715    -0.188    -0.135    0.0253     0.369
exper         0.0320   -0.0300   -0.0500   0.00972    0.0190    0.0512
----------------------------------------------------------------------
N                337                                                  
----------------------------------------------------------------------
dc coefficients

. esttab, main(dc) nostar not mtitles

--------------------------------------
                      (1)          (2)
                   mlogit      mprobit
--------------------------------------
Avg|Chg|                              
white               0.116        0.111
ed                  0.164        0.158
exper              0.0325       0.0320
--------------------------------------
Menial                                
white              -0.131       -0.131
ed                -0.0713      -0.0715
exper             -0.0317      -0.0300
--------------------------------------
BlueCol                               
white              0.0498       0.0392
ed                 -0.193       -0.188
exper             -0.0497      -0.0500
--------------------------------------
Craft                                 
white              -0.160       -0.148
ed                 -0.146       -0.135
exper              0.0148      0.00972
--------------------------------------
WhiteCol                              
white              0.0797       0.0765
ed                 0.0306       0.0253
exper              0.0236       0.0190
--------------------------------------
Prof                                  
white               0.161        0.163
ed                  0.380        0.369
exper              0.0429       0.0512
--------------------------------------
N                     337          337
--------------------------------------
dc coefficients

. eststo clear

[do-file]

Multinomial logit/probit and prchange: selected outcome

. spex nomocc2
(1982 General Social Survey)

. quietly mlogit occ white ed exper, nolog

. estadd prchange, outcome(3)

mlogit: Changes in Probabilities for occ

Outcome: 3 (Craft)

         Min->Max        0->1       -+1/2      -+sd/2    MargEfct
white  -.15973434  -.15973434  -.17549489  -.05025825  -.18235627
   ed  -.15010394   .01737602  -.05247185  -.14576758  -.05287415
exper   .03115708     .001993   .00105992   .01479983   .00105992

            Menial    BlueCol      Craft   WhiteCol       Prof
Pr(y|x)  .09426806  .18419114  .29411051  .16112968  .26630062

         white       ed    exper
   x=  .916914   13.095  20.5015
sd_x=  .276423  2.94643  13.9594

added scalars:
            e(predval) =  .29411051
            e(outcome) =  3
              e(delta) =  1
           e(centered) =  1

added matrices:
                 e(dc) :  6 x 3      (main, Min->Max, 0->1, -+1/2, -+sd/2, Marg
> Efct)
            e(pattern) :  1 x 3
                  e(X) :  4 x 3      (X, SD, Min, Max)

first row in e(dc) contains:

  01 change for binary variables
  sd change for continuous variables

. eststo mlogit

. quietly mprobit occ white ed exper, nolog

. estadd prchange, outcome(3)

mprobit: Changes in Probabilities for occ

Outcome: 3 (Craft)

         Min->Max        0->1       -+1/2      -+sd/2
white  -.14756948  -.14756948  -.15760018  -.04463357
   ed  -.21483173   .01855293  -.04787594  -.13499337
exper   .01516485   .00139609   .00069633   .00971583

            Menial    BlueCol      Craft   WhiteCol       Prof
Pr(y|x)  .09325961  .18944861  .27852002  .15457167   .2842001

         white       ed    exper
   x=  .916914   13.095  20.5015
sd_x=  .276423  2.94643  13.9594

added scalars:
            e(predval) =  .27852002
            e(outcome) =  3
              e(delta) =  1
           e(centered) =  1

added matrices:
                 e(dc) :  5 x 3      (main, Min->Max, 0->1, -+1/2, -+sd/2)
            e(pattern) :  1 x 3
                  e(X) :  4 x 3      (X, SD, Min, Max)

first row in e(dc) contains:

  01 change for binary variables
  sd change for continuous variables

. eststo mprobit

. esttab, aux(dc) wide nopar stats(predval outcome) keep(Craft:) mtitles

----------------------------------------------------------------------
                      (1)                          (2)                
                   mlogit                      mprobit                
----------------------------------------------------------------------
Craft                                                                 
white              -1.302*         -0.160       -0.890          -0.148
ed                 -0.685***       -0.146       -0.472***       -0.135
exper            -0.00797          0.0148     -0.00778         0.00972
_cons               10.43***                     7.140***             
----------------------------------------------------------------------
predval             0.294                        0.279                
outcome                 3                            3                
----------------------------------------------------------------------
dc in second column
* p<0.05, ** p<0.01, *** p<0.001

. eststo clear

[do-file]

mlogit and prchange: split option

. spex nomocc2
(1982 General Social Survey)

. quietly mlogit occ white ed exper, nolog

. estadd prchange, split

mlogit: Changes in Probabilities for occ

white
        Avg|Chg|      Menial     BlueCol       Craft    WhiteCol        Prof
0->1   .11623582  -.13085523   .04981799  -.15973434   .07971004    .1610615

ed
            Avg|Chg|      Menial     BlueCol       Craft    WhiteCol
Min->Max   .39242268  -.13017954  -.70077323  -.15010394   .02425591
   -+1/2   .05855425  -.02559762  -.06831616  -.05247185   .01250795
  -+sd/2    .1640657  -.07129153  -.19310513  -.14576758   .03064777
MargEfct   .05894859  -.02579097  -.06870635  -.05287415   .01282041

                Prof
Min->Max   .95680079
   -+1/2   .13387768
  -+sd/2   .37951647
MargEfct   .13455107

exper
            Avg|Chg|      Menial     BlueCol       Craft    WhiteCol
Min->Max   .12193559  -.11536534  -.18947365   .03115708   .09478889
   -+1/2   .00233425  -.00226997  -.00356567   .00105992    .0016944
  -+sd/2   .03253578  -.03167491  -.04966453   .01479983   .02360725
MargEfct   .00233427  -.00226997  -.00356571   .00105992   .00169442

                Prof
Min->Max   .17889298
   -+1/2   .00308132
  -+sd/2   .04293236
MargEfct   .00308134

            Menial    BlueCol      Craft   WhiteCol       Prof
Pr(y|x)  .09426806  .18419114  .29411051  .16112968  .26630062

         white       ed    exper
   x=  .916914   13.095  20.5015
sd_x=  .276423  2.94643  13.9594

added scalars:
            e(predval) =  .09426806
            e(outcome) =  1
              e(delta) =  1
           e(centered) =  1

added matrices:
                 e(dc) :  6 x 3      (main, Min->Max, 0->1, -+1/2, -+sd/2, Marg
> Efct)
            e(pattern) :  1 x 3
                  e(X) :  4 x 3      (X, SD, Min, Max)

first row in e(dc) contains:

  01 change for binary variables
  sd change for continuous variables

results for outcome 1 stored as mlogit_1
results for outcome 2 stored as mlogit_2
results for outcome 3 stored as mlogit_3
results for outcome 4 stored as mlogit_4
results for outcome 5 stored as mlogit_5

. esttab, main(dc) nostar not scalars(predval outcome) noobs mtitles

-----------------------------------------------------------------------------
                      (1)          (2)          (3)          (4)          (5)
                   Menial      BlueCol        Craft     WhiteCol         Prof
-----------------------------------------------------------------------------
white              -0.131       0.0498       -0.160       0.0797        0.161
ed                -0.0713       -0.193       -0.146       0.0306        0.380
exper             -0.0317      -0.0497       0.0148       0.0236       0.0429
-----------------------------------------------------------------------------
predval            0.0943        0.184        0.294        0.161        0.266
outcome                 1            2            3            4            5
-----------------------------------------------------------------------------
dc coefficients

. eststo clear

[do-file]

mprobit and prchange: split option

. spex nomocc2
(1982 General Social Survey)

. quietly mprobit occ white ed exper, nolog

. estadd prchange, split

mprobit: Changes in Probabilities for occ

white
        Avg|Chg|      Menial     BlueCol       Craft    WhiteCol        Prof
0->1   .11142595  -.13099539   .03923495  -.14756948   .07652746   .16280247

ed
            Avg|Chg|      Menial     BlueCol       Craft    WhiteCol
Min->Max   .39433155  -.13480758  -.63618958  -.21483173   .03136472
   -+1/2   .05591886  -.02555373  -.06636748  -.04787594   .00948766
  -+sd/2    .1577241   -.0714798  -.18783711  -.13499337   .02528359

                Prof
Min->Max   .95446413
   -+1/2   .13030948
  -+sd/2   .36902665

exper
            Avg|Chg|      Menial     BlueCol       Craft    WhiteCol
Min->Max     .123281  -.11218768  -.19601482   .01516485   .07135144
   -+1/2   .00229492  -.00215125  -.00358605   .00069633   .00136708
  -+sd/2   .03197956  -.02998788  -.04996105   .00971583   .01904152

                Prof
Min->Max    .2216862
   -+1/2   .00367388
  -+sd/2   .05119154

            Menial    BlueCol      Craft   WhiteCol       Prof
Pr(y|x)  .09325961  .18944861  .27852002  .15457167   .2842001

         white       ed    exper
   x=  .916914   13.095  20.5015
sd_x=  .276423  2.94643  13.9594

added scalars:
            e(predval) =  .09325961
            e(outcome) =  1
              e(delta) =  1
           e(centered) =  1

added matrices:
                 e(dc) :  5 x 3      (main, Min->Max, 0->1, -+1/2, -+sd/2)
            e(pattern) :  1 x 3
                  e(X) :  4 x 3      (X, SD, Min, Max)

first row in e(dc) contains:

  01 change for binary variables
  sd change for continuous variables

results for outcome 1 stored as mprobit_1
results for outcome 2 stored as mprobit_2
results for outcome 3 stored as mprobit_3
results for outcome 4 stored as mprobit_4
results for outcome 5 stored as mprobit_5

. esttab, main(dc) nostar not scalars(predval outcome) noobs mtitles

-----------------------------------------------------------------------------
                      (1)          (2)          (3)          (4)          (5)
                   Menial      BlueCol        Craft     WhiteCol         Prof
-----------------------------------------------------------------------------
white              -0.131       0.0392       -0.148       0.0765        0.163
ed                -0.0715       -0.188       -0.135       0.0253        0.369
exper             -0.0300      -0.0500      0.00972       0.0190       0.0512
-----------------------------------------------------------------------------
predval            0.0933        0.189        0.279        0.155        0.284
outcome                 1            2            3            4            5
-----------------------------------------------------------------------------
dc coefficients

. eststo clear

[do-file]

mlogit and prvalue

. spex nomocc2
(1982 General Social Survey)

. quietly mlogit occ white ed exper, nolog

. levelsof ed, local(edlevels)
3 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

. foreach l of local edlevels {
  2.     quietly estadd prvalue, x(ed=`l' white=0) label(`l')
  3. }

. estadd prvalue post NonWhite

scalars:
                  e(N) =  337

macros:
             e(depvar) : "occ"
                e(cmd) : "estadd_prvalue"
              e(model) : "mlogit"
         e(properties) : "b"

matrices:
                  e(b) :  1 x 80     (predictions)
                 e(se) :  1 x 80     (standard errors)
                 e(LB) :  1 x 80     (lower CI bounds)
                 e(UB) :  1 x 80     (upper CI bounds)
           e(Category) :  1 x 80     (outcome values)
                  e(X) :  3 x 16     (white, ed, exper)

results stored as NonWhite

. foreach l of local edlevels {
  2.     quietly estadd prvalue, x(ed=`l' white=1) label(`l')
  3. }

. estadd prvalue post White

scalars:
                  e(N) =  337

macros:
             e(depvar) : "occ"
                e(cmd) : "estadd_prvalue"
              e(model) : "mlogit"
         e(properties) : "b"

matrices:
                  e(b) :  1 x 160    (predictions)
                 e(se) :  1 x 160    (standard errors)
                 e(LB) :  1 x 160    (lower CI bounds)
                 e(UB) :  1 x 160    (upper CI bounds)
           e(Category) :  1 x 160    (outcome values)
                  e(X) :  3 x 32     (white, ed, exper)

results stored as White

. esttab NonWhite White, b(4) se nostar wide ///
>     keep(Menial:) mtitles eqlabels(none) noobs

----------------------------------------------------------------
                      (1)                       (2)             
                 NonWhite                     White             
----------------------------------------------------------------
3                  0.2847     (0.2013)       0.1216     (0.0917)
6                  0.2987     (0.1578)       0.1384     (0.0680)
7                  0.2988     (0.1440)       0.1417     (0.0585)
8                  0.2963     (0.1312)       0.1431     (0.0487)
9                  0.2906     (0.1198)       0.1417     (0.0392)
10                 0.2814     (0.1100)       0.1366     (0.0308)
11                 0.2675     (0.1021)       0.1265     (0.0245)
12                 0.2476     (0.0956)       0.1104     (0.0212)
13                 0.2199     (0.0895)       0.0883     (0.0195)
14                 0.1832     (0.0821)       0.0632     (0.0175)
15                 0.1393     (0.0714)       0.0401     (0.0142)
16                 0.0944     (0.0566)       0.0228     (0.0102)
17                 0.0569     (0.0399)       0.0120     (0.0066)
18                 0.0310     (0.0250)       0.0060     (0.0039)
19                 0.0158     (0.0143)       0.0029     (0.0022)
20                 0.0077     (0.0077)       0.0014     (0.0012)
----------------------------------------------------------------
Standard errors in parentheses

. eststo clear

[do-file]

mprobit and prvalue

. spex nomocc2
(1982 General Social Survey)

. quietly mprobit occ white ed exper, nolog

. levelsof ed, local(edlevels)
3 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

. foreach l of local edlevels {
  2.     quietly estadd prvalue, x(ed=`l' white=0) label(`l')
  3. }

. estadd prvalue post NonWhite

scalars:
                  e(N) =  337

macros:
             e(depvar) : "occ"
                e(cmd) : "estadd_prvalue"
              e(model) : "mprobit"
         e(properties) : "b"

matrices:
                  e(b) :  1 x 80     (predictions)
                 e(se) :  1 x 80     (standard errors)
                 e(LB) :  1 x 80     (lower CI bounds)
                 e(UB) :  1 x 80     (upper CI bounds)
           e(Category) :  1 x 80     (outcome values)
                  e(X) :  3 x 16     (white, ed, exper)

results stored as NonWhite

. foreach l of local edlevels {
  2.     quietly estadd prvalue, x(ed=`l' white=1) label(`l')
  3. }

. estadd prvalue post White

scalars:
                  e(N) =  337

macros:
             e(depvar) : "occ"
                e(cmd) : "estadd_prvalue"
              e(model) : "mprobit"
         e(properties) : "b"

matrices:
                  e(b) :  1 x 160    (predictions)
                 e(se) :  1 x 160    (standard errors)
                 e(LB) :  1 x 160    (lower CI bounds)
                 e(UB) :  1 x 160    (upper CI bounds)
           e(Category) :  1 x 160    (outcome values)
                  e(X) :  3 x 32     (white, ed, exper)

results stored as White

. esttab NonWhite White, b(4) nostar not ///
>     keep(Menial:) mtitles eqlabels(none) noobs

--------------------------------------
                      (1)          (2)
                 NonWhite        White
--------------------------------------
3                  0.2446       0.1274
6                  0.2617       0.1421
7                  0.2654       0.1450
8                  0.2676       0.1462
9                  0.2679       0.1446
10                 0.2652       0.1389
11                 0.2577       0.1275
12                 0.2429       0.1098
13                 0.2185       0.0869
14                 0.1842       0.0623
15                 0.1429       0.0398
16                 0.1006       0.0225
17                 0.0636       0.0112
18                 0.0357       0.0049
19                 0.0178       0.0019
20                 0.0078       0.0006
--------------------------------------

. eststo clear

[do-file]

Stereotype logistic regression

slogit and fitstat/listcoef

. spex ordwarm2
(77 & 89 General Social Survey)

. quietly slogit warm yr89 male white age ed prst, nolog

. estadd fitstat

Measures of Fit for slogit of warm

Log-Lik Full Model:          -2845.595   D(2282):                      5691.189
Wald X2(6):                    185.448   Prob > X2:                       0.000
AIC:                             2.492   AIC*n:                        5713.189
BIC:                        -11966.051                              
BIC used by Stata:            5776.303   AIC used by Stata:            5713.189

added scalars:
                e(dev) =  5691.1894
             e(dev_df) =  2282
               e(lrx2) =  185.44833
            e(lrx2_df) =  6
             e(lrx2_p) =  2.361e-37
               e(aic0) =  2.4915785
              e(aic_n) =  5713.1894
               e(bic0) =  -11966.051
           e(statabic) =  5776.3032
           e(stataaic) =  5713.1894
              e(n_rhs) =  8
             e(n_parm) =  11

. estadd listcoef

slogit (N=2293): Factor Change in Odds 

  Odds of: 4SA vs 1SD

----------------------------------------------------------------------
        warm |      b         z     P>|z|    e^b    e^bStdX      SDofX
-------------+--------------------------------------------------------
        yr89 |   0.94405    6.179   0.000   2.5704   1.5878     0.4897
        male |  -1.25606   -8.274   0.000   0.2848   0.5344     0.4989
       white |  -0.63901   -2.973   0.003   0.5278   0.8104     0.3290
         age |  -0.03841   -8.541   0.000   0.9623   0.5249    16.7790
          ed |   0.10933    3.737   0.000   1.1155   1.4128     3.1608
        prst |   0.01148    1.983   0.047   1.0115   1.1810    14.4923
-------------+--------------------------------------------------------
      phi1_1 |   1.00000        .       .
      phi1_2 |   0.74885   13.840   0.000
      phi1_3 |   0.31837    6.397   0.000
-------------+--------------------------------------------------------
      theta1 |  -1.06006   -2.577   0.010
      theta2 |   0.13237    0.423   0.672
      theta3 |   0.62730    4.399   0.000
----------------------------------------------------------------------

added matrices:
             e(b_fact) :  1 x 6      (e^b)
            e(b_facts) :  1 x 6      (e^bStdX)
              e(b_sdx) :  1 x 6      (SDofX)

. esttab, cell("b b_fact b_facts") scalars(aic0 bic0) ///
>     eqlabels(none)

---------------------------------------------------
                      (1)                          
                     warm                          
                        b       b_fact      b_facts
---------------------------------------------------
yr89             .9440522     2.570376     1.587752
male            -1.256064     .2847727     .5343958
white           -.6390139     .5278126     .8103988
age             -.0384116     .9623168       .52492
ed               .1093335     1.115534     1.412815
prst             .0114819     1.011548     1.181044
phi1_1                  1                          
phi1_2           .7488452                          
phi1_3           .3183653                          
theta1          -1.060064                          
theta2           .1323735                          
theta3           .6272993                          
---------------------------------------------------
N                    2293                          
aic0                2.492                          
bic0             -11966.1                          
---------------------------------------------------

[do-file]

slogit and prchange

. spex ordwarm2
(77 & 89 General Social Survey)

. quietly slogit warm yr89 male white age ed prst, nolog

. estadd prchange male age prst

slogit: Changes in Probabilities for warm

male
        Avg|Chg|         1SD          2D          3A         4SA
0->1   .09093904   .07668686    .1051912  -.08178753  -.10009058

age
            Avg|Chg|         1SD          2D          3A         4SA
Min->Max   .19115428   .17745935    .2048492  -.18460661  -.19770195
   -+1/2   .00280473     .002321   .00328845  -.00249711  -.00311236
  -+sd/2   .04692763   .03891201   .05494326  -.04169974  -.05215551

prst
            Avg|Chg|         1SD          2D          3A         4SA
Min->Max   .05812308  -.04720601  -.06904018    .0497874   .06645873
   -+1/2   .00083838  -.00069379    -.000983    .0007464   .00093034
  -+sd/2   .01214788  -.01005406   -.0142417   .01081413   .01348165

               1SD         2D         3A        4SA
Pr(y|x)  .11714774  .32349858  .39201239  .16734134

          yr89     male    white      age       ed     prst
   x=  .398604  .464893  .876581  44.9355  12.2181  39.5853
sd_x=  .489718  .498875  .328989   16.779  3.16083  14.4923

added scalars:
           e(predval1) =  .11714774
           e(predval2) =  .32349858
           e(predval3) =  .39201239
           e(predval4) =  .16734134
              e(delta) =  1
           e(centered) =  1

added matrices:
                 e(dc) :  5 x 15     (main, Min->Max, 0->1, -+1/2, -+sd/2)
            e(pattern) :  1 x 15
                  e(X) :  4 x 6      (X, SD, Min, Max)

first row in e(dc) contains:

  01 change for binary variables
  sd change for continuous variables

. esttab, main(dc) unstack nostar not

-----------------------------------------------------------------------------
                      (1)                                                    
                     warm                                                    
                 Avg|Chg|          1SD           2D           3A          4SA
-----------------------------------------------------------------------------
male               0.0909       0.0767        0.105      -0.0818       -0.100
age                0.0469       0.0389       0.0549      -0.0417      -0.0522
prst               0.0121      -0.0101      -0.0142       0.0108       0.0135
-----------------------------------------------------------------------------
N                    2293                                                    
-----------------------------------------------------------------------------
dc coefficients

[do-file]

slogit and prchange: selected outcome

. spex ordwarm2
(77 & 89 General Social Survey)

. quietly slogit warm yr89 male white age ed prst, nolog

. estadd prchange age ed prst, outcome(2)

slogit: Changes in Probabilities for warm

Outcome: 2 (2D)

        Min->Max        0->1       -+1/2      -+sd/2
 age    .2048492   .00294244   .00328845   .05494326
  ed  -.17157121  -.00616026  -.00935909   -.0295499
prst  -.06904018  -.00090283    -.000983   -.0142417

               1SD         2D         3A        4SA
Pr(y|x)  .11714774  .32349858  .39201239  .16734134

          yr89     male    white      age       ed     prst
   x=  .398604  .464893  .876581  44.9355  12.2181  39.5853
sd_x=  .489718  .498875  .328989   16.779  3.16083  14.4923

added scalars:
            e(predval) =  .32349858
            e(outcome) =  2
              e(delta) =  1
           e(centered) =  1

added matrices:
                 e(dc) :  5 x 3      (main, Min->Max, 0->1, -+1/2, -+sd/2)
            e(pattern) :  1 x 3
                  e(X) :  4 x 6      (X, SD, Min, Max)

first row in e(dc) contains:

  01 change for binary variables
  sd change for continuous variables

. esttab, cell("dc[Min->Max] dc[-+1/2] dc[-+sd/2]") stats(predval outcome)

---------------------------------------------------
                      (1)                          
                     warm                          
                 Min->Max        -+1/2       -+sd/2
---------------------------------------------------
age              .2048492     .0032884     .0549433
ed              -.1715712    -.0093591    -.0295499
prst            -.0690402     -.000983    -.0142417
---------------------------------------------------
predval          .3234986                          
outcome                 2                          
---------------------------------------------------

[do-file]

slogit and prchange: split option

. spex ordwarm2
(77 & 89 General Social Survey)

. quietly slogit warm yr89 male white age ed prst, nolog

. estadd prchange male age prst, split

slogit: Changes in Probabilities for warm

male
        Avg|Chg|         1SD          2D          3A         4SA
0->1   .09093904   .07668686    .1051912  -.08178753  -.10009058

age
            Avg|Chg|         1SD          2D          3A         4SA
Min->Max   .19115428   .17745935    .2048492  -.18460661  -.19770195
   -+1/2   .00280473     .002321   .00328845  -.00249711  -.00311236
  -+sd/2   .04692763   .03891201   .05494326  -.04169974  -.05215551

prst
            Avg|Chg|         1SD          2D          3A         4SA
Min->Max   .05812308  -.04720601  -.06904018    .0497874   .06645873
   -+1/2   .00083838  -.00069379    -.000983    .0007464   .00093034
  -+sd/2   .01214788  -.01005406   -.0142417   .01081413   .01348165

               1SD         2D         3A        4SA
Pr(y|x)  .11714774  .32349858  .39201239  .16734134

          yr89     male    white      age       ed     prst
   x=  .398604  .464893  .876581  44.9355  12.2181  39.5853
sd_x=  .489718  .498875  .328989   16.779  3.16083  14.4923

added scalars:
            e(predval) =  .11714774
            e(outcome) =  1
              e(delta) =  1
           e(centered) =  1

added matrices:
                 e(dc) :  5 x 3      (main, Min->Max, 0->1, -+1/2, -+sd/2)
            e(pattern) :  1 x 3
                  e(X) :  4 x 6      (X, SD, Min, Max)

first row in e(dc) contains:

  01 change for binary variables
  sd change for continuous variables

results for outcome 1 stored as slogit_1
results for outcome 2 stored as slogit_2
results for outcome 3 stored as slogit_3
results for outcome 4 stored as slogit_4

. esttab, main(dc) nostar not stats(predval outcome) ///
>     mtitles nonumbers

----------------------------------------------------------------
                      1SD           2D           3A          4SA
----------------------------------------------------------------
male               0.0767        0.105      -0.0818       -0.100
age                0.0389       0.0549      -0.0417      -0.0522
prst              -0.0101      -0.0142       0.0108       0.0135
----------------------------------------------------------------
predval             0.117        0.323        0.392        0.167
outcome                 1            2            3            4
----------------------------------------------------------------
dc coefficients

. eststo clear

[do-file]

slogit and prvalue

. spex ordwarm2
(77 & 89 General Social Survey)

. quietly slogit warm yr89 male white age ed prst, nolog

. estadd prvalue, x(yr89=0 male=1 prst=20 age=64 ed=16) ///
>     brief label(type1)

slogit: Predictions for warm

  Pr(y=1SD|x):        0.2341
  Pr(y=2D|x):         0.4333
  Pr(y=3A|x):         0.2646
  Pr(y=4SA|x):        0.0680

added matrices:
    e(_estadd_prvalue) :  1 x 24
  e(_estadd_prvalue_x) :  1 x 6

. estadd prvalue, x(yr89=1 male=0 prst=80 age=30 ed=24) ///
>     brief label(type2)

slogit: Predictions for warm

  Pr(y=1SD|x):        0.0112
  Pr(y=2D|x):         0.0737
  Pr(y=3A|x):         0.3993
  Pr(y=4SA|x):        0.5158

updated matrices:
    e(_estadd_prvalue) :  2 x 24
  e(_estadd_prvalue_x) :  2 x 6

. estadd prvalue, x(yr89=0) brief label(type3)

slogit: Predictions for warm

  Pr(y=1SD|x):        0.1412
  Pr(y=2D|x):         0.3548
  Pr(y=3A|x):         0.3656
  Pr(y=4SA|x):        0.1384

updated matrices:
    e(_estadd_prvalue) :  3 x 24
  e(_estadd_prvalue_x) :  3 x 6

. estadd prvalue, x(yr89=1) brief label(type4)

slogit: Predictions for warm

  Pr(y=1SD|x):        0.0860
  Pr(y=2D|x):         0.2738
  Pr(y=3A|x):         0.4236
  Pr(y=4SA|x):        0.2167

updated matrices:
    e(_estadd_prvalue) :  4 x 24
  e(_estadd_prvalue_x) :  4 x 6

. estadd prvalue post

scalars:
                  e(N) =  2293

macros:
             e(depvar) : "warm"
                e(cmd) : "estadd_prvalue"
              e(model) : "slogit"
         e(properties) : "b"

matrices:
                  e(b) :  1 x 16     (predictions)
                 e(se) :  1 x 16     (standard errors)
                 e(LB) :  1 x 16     (lower CI bounds)
                 e(UB) :  1 x 16     (upper CI bounds)
           e(Category) :  1 x 16     (outcome values)
                  e(X) :  6 x 4      (yr89, male, white, age, ed, prst)

. esttab, nostar not unstack ///
>     coeflabels(type1 "old working class men 1977"   ///
>                type2 "young prestigious women 1989" ///
>                type3 "average individual 1977" ///
>                type4 "average individual 1989") ///
>     varwidth(28) compress

--------------------------------------------------------------------
                                   (1)                              
                                  warm                              
                                   1SD        2D        3A       4SA
--------------------------------------------------------------------
old working class men 1977       0.234     0.433     0.265    0.0680
young prestigious women 1989    0.0112    0.0737     0.399     0.516
average individual 1977          0.141     0.355     0.366     0.138
average individual 1989         0.0860     0.274     0.424     0.217
--------------------------------------------------------------------
N                                 2293                              
--------------------------------------------------------------------

[do-file]

Conditional logit

clogit and fitstat/listcoef

. spex travel2
(Greene & Hensher 1997 data on travel mode choice)

. quietly clogit choice train bus time invc, group(id) nolog

. estadd fitstat

Measures of Fit for clogit of choice

Log-Lik Intercept Only:       -166.989   Log-Lik Full Model:            -80.961
D(148):                        161.922   LR(4):                         172.056
                                         Prob > LR:                       0.000
McFadden's R2:                   0.515   McFadden's Adj R2:               0.491
ML (Cox-Snell) R2:               0.678   Cragg-Uhler(Nagelkerke) R2:      0.762
Count R2:                        0.875                              
AIC:                             1.118   AIC*n:                         169.922
BIC:                          -581.612   BIC':                         -151.960
BIC used by Stata:             186.412   AIC used by Stata:             169.922

added scalars:
                e(dev) =  161.92227
             e(dev_df) =  148
               e(lrx2) =  172.05587
            e(lrx2_df) =  4
             e(lrx2_p) =  3.786e-36
              e(r2_mf) =  .51517105
           e(r2_mfadj) =  .49121738
              e(r2_ml) =  .67759491
              e(r2_cu) =  .76229428
              e(r2_ct) =  .875
               e(aic0) =  1.1179097
              e(aic_n) =  169.92227
               e(bic0) =  -581.61205
              e(bic_p) =  -151.96034
           e(statabic) =  186.41224
           e(stataaic) =  169.92227
              e(n_rhs) =  3
             e(n_parm) =  4

. estadd listcoef

clogit (N=456): Factor Change in Odds 

  Odds of: 1 vs 0

--------------------------------------------------
      choice |      b         z     P>|z|    e^b  
-------------+------------------------------------
       train |   2.67124    5.895   0.000  14.4579
         bus |   1.47233    3.674   0.000   4.3594
        time |  -0.01915   -7.812   0.000   0.9810
        invc |  -0.04817   -4.030   0.000   0.9530
--------------------------------------------------

added matrices:
             e(b_fact) :  1 x 4      (e^b)

. estadd listcoef, percent

clogit (N=456): Percentage Change in Odds 

  Odds of: 1 vs 0

--------------------------------------------------
      choice |      b         z     P>|z|      %  
-------------+------------------------------------
       train |   2.67124    5.895   0.000   1345.8
         bus |   1.47233    3.674   0.000    335.9
        time |  -0.01915   -7.812   0.000     -1.9
        invc |  -0.04817   -4.030   0.000     -4.7
--------------------------------------------------

added matrices:
              e(b_pct) :  1 x 4      (%)

. esttab, cell("b b_fact b_pct") scalars(r2_mf r2_mfadj r2_ml r2_cu)

---------------------------------------------------
                      (1)                          
                   choice                          
                        b       b_fact        b_pct
---------------------------------------------------
choice                                             
train            2.671238     14.45786     1345.786
bus              1.472335     4.359401     335.9401
time            -.0191453     .9810368    -1.896319
invc            -.0481658     .9529758    -4.702424
---------------------------------------------------
N                     456                          
r2_mf               0.515                          
r2_mfadj            0.491                          
r2_ml               0.678                          
r2_cu               0.762                          
---------------------------------------------------

[do-file]

clogit and asprvalue: alternative-specific variables only

. spex travel2
(Greene & Hensher 1997 data on travel mode choice)

. quietly clogit choice train bus time invc, group(id) nolog

. quietly asprvalue, x(time=643.4 674.6 578.3) rest(asmean) ///
>     cat(train bus) base(car) save

. estadd asprvalue, x(time=653.4 674.6 578.3) rest(asmean) ///
>     cat(train bus) base(car) label(time train + 10 min) brief diff

clogit: Predictions for choice

          Current       Saved        Diff
train   .38845369   .43478274  -.04632905
  bus   .16446434   .15200497   .01245937
  car   .44708198   .41321227   .03386971

added matrices:
    e(_estadd_asprval) :  1 x 3

. estadd asprvalue, x(time=643.4 684.6 578.3) rest(asmean) ///
>     cat(train bus) base(car) label(time bus + 10 min) brief diff

clogit: Predictions for choice

          Current       Saved        Diff
train   .44661152   .43478274   .01182878
  bus   .12893429   .15200497  -.02307068
  car   .42445418   .41321227   .01124191

updated matrices:
    e(_estadd_asprval) :  2 x 3

. estadd asprvalue, x(time=643.4 674.6 588.3) rest(asmean) ///
>     cat(train bus) base(car) label(time car + 10 min) brief diff

clogit: Predictions for choice

          Current       Saved        Diff
train   .46851528   .43478274   .03373253
  bus   .16379826   .15200497   .01179329
  car   .36768648   .41321227  -.04552579

updated matrices:
    e(_estadd_asprval) :  3 x 3

. estadd asprvalue post

scalars:
                  e(N) =  456

macros:
             e(depvar) : "choice"
                e(cmd) : "estadd_asprvalue"
              e(model) : "clogit"
         e(properties) : "b"

matrices:
                  e(b) :  1 x 9      (predictions)

. esttab, unstack not nostar varwidth(20)

-----------------------------------------------------------
                              (1)                          
                           choice                          
                            train          bus          car
-----------------------------------------------------------
time train + 10 min       -0.0463       0.0125       0.0339
time bus + 10 min          0.0118      -0.0231       0.0112
time car + 10 min          0.0337       0.0118      -0.0455
-----------------------------------------------------------
N                             456                          
-----------------------------------------------------------

[do-file]

clogit and asprvalue: alternative-specific and case-specific variables

. spex travel2
(Greene & Hensher 1997 data on travel mode choice)

. gen busXhinc = bus*hinc

. gen trainXhinc = train*hinc

. gen busXpsize = bus*psize

. gen trainXpsize = train*psize

. quietly clogit choice busXhinc busXpsize bus trainXhinc trainXpsize train ///
>     time invc, group(id) nolog

. quietly asprvalue, x(psize=1) rest(asmean) base(car) save

. estadd asprvalue,  x(psize=2) rest(asmean) base(car) label(_cons) brief diff

clogit: Predictions for choice

          Current       Saved        Diff
  bus   .13919763   .21251462  -.07331699
train   .44040644   .40365174    .0367547
  car   .42039591   .38383365   .03656226

added matrices:
    e(_estadd_asprval) :  1 x 3

. estadd asprvalue post

scalars:
                  e(N) =  456

macros:
             e(depvar) : "choice"
                e(cmd) : "estadd_asprvalue"
              e(model) : "clogit"
         e(properties) : "b"

matrices:
                  e(b) :  1 x 3      (predictions)

. esttab, b not nostar eqlabels(none) ///
>      mtitle("psize=2 - psize=1") modelw(20)

---------------------------------
                              (1)
                psize=2 - psize=1
---------------------------------
bus                       -0.0733
train                      0.0368
car                        0.0366
---------------------------------
N                             456
---------------------------------

[do-file]

Alternative-specific multinomial probit

asmprobit and asprvalue

. spex travel2
(Greene & Hensher 1997 data on travel mode choice)

. quietly asmprobit choice time invc, case(id) alternatives(mode) nolog

. estadd asprvalue, label(at means)

asmprobit: Predictions for choice

            prob
Train  .76511556
  Bus  .09945779
  Car  .13547295

alternative-specific variables

          Train        Bus        Car
time  632.10965  632.10965  632.10965
invc  33.951754  33.951754  33.951754

added matrices:
    e(_estadd_asprval) :  1 x 3
e(_estadd_asprval_asv) :  1 x 6
e(_estadd_asprval_csv) :  1 x 2

. estadd asprvalue, rest(asmean) label(at asmeans)

asmprobit: Predictions for choice

            prob
Train  .42618456
  Bus  .12483268
  Car  .44901165

alternative-specific variables

          Train        Bus        Car
time  643.44079  674.61842  578.26974
invc  48.618421  33.144737  20.092105

updated matrices:
    e(_estadd_asprval) :  2 x 3
e(_estadd_asprval_asv) :  2 x 6
e(_estadd_asprval_csv) :  2 x 2

. estadd asprvalue post, swap

scalars:
                  e(N) =  456

macros:
             e(depvar) : "choice"
                e(cmd) : "estadd_asprvalue"
              e(model) : "asmprobit"
         e(properties) : "b"

matrices:
                  e(b) :  1 x 6      (predictions)
                e(asv) :  2 x 6      (time, invc)
                e(csv) :  2 x 2      (hinc, psize)

. esttab, unstack not nostar nomtitle nonumber

--------------------------------------
                 at means   at asmeans
--------------------------------------
Train               0.765        0.426
Bus                0.0995        0.125
Car                 0.135        0.449
--------------------------------------
N                     456             
--------------------------------------

[do-file]

Rank-ordered logit

rologit and fitstat/listcoef

. spex wlsrnk
(1992 Wisconsin Longitudinal Study data on job values)

. label variable value1 "est"

. label variable value2 "var"

. label variable value3 "aut"

. label variable value4 "sec"

. case2alt, casevars(fem hn) rank(value) case(id) alt(hashi haslo) gen(rank)
(note: variable _altnum used since altnum() not specified)

ranks indicated by: rank
case identifier: id
case-specific interactions: est* var* aut* sec*
alternative-specific variables: hashi haslo

. rologit rank estXfem estXhn est varXfem varXhn var ///
>     autXfem autXhn aut hashi haslo, group(id) reverse nolog

Rank-ordered logistic regression                Number of obs      =     12904
Group variable: id                              Number of groups   =      3226

Ties handled via the exactm method              Obs per group: min =         4
                                                               avg =      4.00
                                                               max =         4

                                                LR chi2(11)        =   1947.39
Log likelihood = -6127.559                      Prob > chi2        =    0.0000

------------------------------------------------------------------------------
        rank |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     estXfem |  -.1497926   .0783025    -1.91   0.056    -.3032626    .0036774
      estXhn |   .1375338   .0394282     3.49   0.000      .060256    .2148116
         est |  -1.017202    .054985   -18.50   0.000     -1.12497   -.9094331
     varXfem |  -.1640212   .0728306    -2.25   0.024    -.3067666   -.0212759
      varXhn |   .2590404   .0370942     6.98   0.000     .1863372    .3317437
         var |    .528224   .0525894    10.04   0.000     .4251506    .6312974
     autXfem |  -.1401769   .0718325    -1.95   0.051     -.280966    .0006123
      autXhn |   .2133866   .0361647     5.90   0.000      .142505    .2842682
         aut |  -.1516741   .0510374    -2.97   0.003    -.2517055   -.0516427
       hashi |   .1780449   .0374744     4.75   0.000     .1045965    .2514934
       haslo |  -.2064148   .0425829    -4.85   0.000    -.2898758   -.1229539
------------------------------------------------------------------------------

. estadd fitstat

Measures of Fit for rologit of rank

Log-Lik Full Model:          -6127.559   D(12893):                    12255.119
                                         LR(11):                       1947.390
                                         Prob > LR:                       0.000
AIC:                             0.951   AIC*n:                       12277.119
BIC:                       -109780.899   BIC':                        -1843.272
BIC used by Stata:           12359.237   AIC used by Stata:           12277.119

added scalars:
                e(dev) =  12255.119
             e(dev_df) =  12893
               e(lrx2) =  1947.3899
            e(lrx2_df) =  11
             e(lrx2_p) =  0
               e(aic0) =  .95141963
              e(aic_n) =  12277.119
               e(bic0) =  -109780.9
              e(bic_p) =  -1843.2717
           e(statabic) =  12359.237
           e(stataaic) =  12277.119
              e(n_rhs) =  10
             e(n_parm) =  11

. estadd listcoef

rologit (N=12904): Factor Change in Odds 

  Odds of: ranked ahead vs ranked behind

--------------------------------------------------
        rank |      b         z     P>|z|    e^b  
-------------+------------------------------------
     estXfem |  -0.14979   -1.913   0.056   0.8609
      estXhn |   0.13753    3.488   0.000   1.1474
         est |  -1.01720  -18.500   0.000   0.3616
     varXfem |  -0.16402   -2.252   0.024   0.8487
      varXhn |   0.25904    6.983   0.000   1.2957
         var |   0.52822   10.044   0.000   1.6959
     autXfem |  -0.14018   -1.951   0.051   0.8692
      autXhn |   0.21339    5.900   0.000   1.2379
         aut |  -0.15167   -2.972   0.003   0.8593
       hashi |   0.17804    4.751   0.000   1.1949
       haslo |  -0.20641   -4.847   0.000   0.8135
--------------------------------------------------

added matrices:
             e(b_fact) :  1 x 11     (e^b)

. estadd listcoef, percent replace

rologit (N=12904): Percentage Change in Odds 

  Odds of: ranked ahead vs ranked behind

--------------------------------------------------
        rank |      b         z     P>|z|      %  
-------------+------------------------------------
     estXfem |  -0.14979   -1.913   0.056    -13.9
      estXhn |   0.13753    3.488   0.000     14.7
         est |  -1.01720  -18.500   0.000    -63.8
     varXfem |  -0.16402   -2.252   0.024    -15.1
      varXhn |   0.25904    6.983   0.000     29.6
         var |   0.52822   10.044   0.000     69.6
     autXfem |  -0.14018   -1.951   0.051    -13.1
      autXhn |   0.21339    5.900   0.000     23.8
         aut |  -0.15167   -2.972   0.003    -14.1
       hashi |   0.17804    4.751   0.000     19.5
       haslo |  -0.20641   -4.847   0.000    -18.7
--------------------------------------------------

added matrices:
              e(b_pct) :  1 x 11     (%)

. esttab, cell("b b_fact b_pct") scalars(aic0 aic_n bic0 bic_p)

---------------------------------------------------
                      (1)                          
                     rank                          
                        b       b_fact        b_pct
---------------------------------------------------
estXfem         -.1497926     .8608865    -13.91135
estXhn           .1375338      1.14744     14.74405
est             -1.017202     .3616054    -63.83946
varXfem         -.1640212      .848724     -15.1276
varXhn           .2590404     1.295686     29.56862
var               .528224     1.695918     69.59177
autXfem         -.1401769     .8692045    -13.07955
autXhn           .2133866     1.237863     23.78631
aut             -.1516741     .8592683    -14.07317
hashi            .1780449     1.194879      19.4879
haslo           -.2064148     .8134955    -18.65045
---------------------------------------------------
N                   12904                          
aic0                0.951                          
aic_n             12277.1                          
bic0            -109780.9                          
bic_p             -1843.3                          
---------------------------------------------------

[do-file]

rologit and asprvalue

. spex wlsrnk
(1992 Wisconsin Longitudinal Study data on job values)

. label variable value1 "est"

. label variable value2 "var"

. label variable value3 "aut"

. label variable value4 "sec"

. case2alt, casevars(fem hn) rank(value) case(id) alt(hashi haslo) gen(rank)
(note: variable _altnum used since altnum() not specified)

ranks indicated by: rank
case identifier: id
case-specific interactions: est* var* aut* sec*
alternative-specific variables: hashi haslo

. quietly rologit rank estXfem estXhn est varXfem varXhn var ///
>     autXfem autXhn aut hashi haslo, group(id) reverse nolog

. estadd asprvalue, x(fem=1 hashi=0 haslo=0) base(sec) label(fem=1) brief save

rologit: Predictions for rank

          prob
est  .08876531
var  .41536212
aut  .21456315
sec  .28130943

added matrices:
    e(_estadd_asprval) :  1 x 4
e(_estadd_asprval_asv) :  1 x 8
e(_estadd_asprval_csv) :  1 x 2

. estadd asprvalue, x(fem=0 hashi=0 haslo=0) base(sec) label(fem=0) brief

rologit: Predictions for rank

          prob
est  .09200718
var  .43670157
aut   .2202711
sec  .25102016

updated matrices:
    e(_estadd_asprval) :  2 x 4
e(_estadd_asprval_asv) :  2 x 8
e(_estadd_asprval_csv) :  2 x 2

. estadd asprvalue, x(fem=0 hashi=0 haslo=0) base(sec) label(diff)  brief diff

rologit: Predictions for rank

        Current       Saved        Diff
est   .09200718   .08876531   .00324187
var   .43670157   .41536212   .02133945
aut    .2202711   .21456315   .00570795
sec   .25102016   .28130943  -.03028926

updated matrices:
    e(_estadd_asprval) :  3 x 4

. estadd asprvalue post, swap

scalars:
                  e(N) =  12904

macros:
             e(depvar) : "rank"
                e(cmd) : "estadd_asprvalue"
              e(model) : "rologit"
         e(properties) : "b"

matrices:
                  e(b) :  1 x 12     (predictions)
                e(asv) :  2 x 8      (hashi, haslo)
                e(csv) :  2 x 2      (fem, hn)

. esttab, not nostar unstack

---------------------------------------------------
                      (1)                          
                     rank                          
                    fem=1        fem=0         diff
---------------------------------------------------
est                0.0888       0.0920      0.00324
var                 0.415        0.437       0.0213
aut                 0.215        0.220      0.00571
sec                 0.281        0.251      -0.0303
---------------------------------------------------
N                   12904                          
---------------------------------------------------

[do-file]

Poisson/negative binomial regression

poisson/nbreg and fitstat

. spex couart2
(Academic Biochemists / S Long)

. eststo poisson: quietly poisson art fem mar kid5 phd ment, nolog

. eststo nbreg: quietly nbreg art fem mar kid5 phd ment, nolog

. estadd fitstat : *

. esttab, wide scalars(r2_mf r2_mfadj aic0 aic_n) mtitles

----------------------------------------------------------------------
                      (1)                          (2)                
                  poisson                        nbreg                
----------------------------------------------------------------------
art                                                                   
fem                -0.225***      (-4.11)       -0.216**       (-2.98)
mar                 0.155*         (2.53)        0.150          (1.83)
kid5               -0.185***      (-4.61)       -0.176***      (-3.32)
phd                0.0128          (0.49)       0.0153          (0.42)
ment               0.0255***      (12.73)       0.0291***       (8.38)
_cons               0.305**        (2.96)        0.256          (1.85)
----------------------------------------------------------------------
lnalpha                                                               
_cons                                           -0.817***      (-6.81)
----------------------------------------------------------------------
N                     915                          915                
r2_mf              0.0525                       0.0304                
r2_mfadj           0.0491                       0.0261                
aic0                3.622                        3.427                
aic_n              3314.1                       3135.9                
----------------------------------------------------------------------
t statistics in parentheses
* p<0.05, ** p<0.01, *** p<0.001

. eststo clear

[do-file]

poisson/nbreg and listcoef

. spex couart2
(Academic Biochemists / S Long)

. eststo poisson: quietly poisson art fem mar kid5 phd ment, nolog

. eststo nbreg: quietly nbreg art fem mar kid5 phd ment, nolog

. estadd listcoef fem ment: *

. estadd listcoef fem ment, percent nosd : *

. esttab, cell("b_facts b_pcts") keep(fem ment) mtitles

----------------------------------------------------------------
                      (1)                       (2)             
                  poisson                     nbreg             
                  b_facts       b_pcts      b_facts       b_pcts
----------------------------------------------------------------
fem              .8940439    -10.59561     .8976965    -10.23035
ment             1.274107     27.41066     1.317603     31.76034
----------------------------------------------------------------
N                     915                       915             
----------------------------------------------------------------

. eststo clear

[do-file]

poisson/nbreg and prchange: effects on rate

. spex couart2
(Academic Biochemists / S Long)

. eststo poisson: quietly poisson art fem mar kid5 phd ment, nolog

. eststo nbreg: quietly nbreg art fem mar kid5 phd ment, nolog

. estadd prchange: *

. esttab, aux(dc) nopar wide mtitles

----------------------------------------------------------------------
                      (1)                          (2)                
                  poisson                        nbreg                
----------------------------------------------------------------------
art                                                                   
fem                -0.225***       -0.359       -0.216**        -0.344
mar                 0.155*          0.244        0.150           0.235
kid5               -0.185***       -0.228       -0.176***       -0.216
phd                0.0128          0.0203       0.0153          0.0241
ment               0.0255***        0.391       0.0291***        0.443
_cons               0.305**                      0.256                
----------------------------------------------------------------------
lnalpha                                                               
_cons                                           -0.817***             
----------------------------------------------------------------------
N                     915                          915                
----------------------------------------------------------------------
dc in second column
* p<0.05, ** p<0.01, *** p<0.001

. eststo clear

[do-file]

poisson/nbreg and prchange: effects on probability of selected outcome

. spex couart2
(Academic Biochemists / S Long)

. eststo poisson0: quietly poisson art fem mar kid5 phd ment, nolog

. eststo nbreg0: quietly nbreg art fem mar kid5 phd ment, nolog

. estadd prchange, outcome(0) : *0

. eststo poisson1: quietly poisson art fem mar kid5 phd ment, nolog

. eststo nbreg1: quietly nbreg art fem mar kid5 phd ment, nolog

. estadd prchange, outcome(1) : *1

. esttab, main(dc) nostar not scalars(outcome predval) mtitles

----------------------------------------------------------------
                      (1)          (2)          (3)          (4)
                 poisson0       nbreg0     poisson1       nbreg1
----------------------------------------------------------------
fem                0.0725       0.0606       0.0431       0.0209
mar               -0.0506      -0.0424      -0.0289      -0.0141
kid5               0.0455       0.0377       0.0277       0.0133
phd              -0.00406     -0.00420     -0.00248     -0.00148
ment              -0.0777      -0.0769      -0.0473      -0.0270
----------------------------------------------------------------
N                     915          915          915          915
outcome                 0            0            1            1
predval             0.200        0.298        0.322        0.279
----------------------------------------------------------------
dc coefficients

. eststo clear

[do-file]

poisson/nbreg and prvalue

. spex couart2
(Academic Biochemists / S Long)

. eststo poisson: quietly poisson art fem mar kid5 phd ment, nolog

. estadd prvalue

poisson: Predictions for art

Confidence intervals by delta method

                                95% Conf. Interval
  Rate:               1.6101   [ 1.5265,    1.6937]
  Pr(y=0|x):          0.1999   [ 0.1832,    0.2166]
  Pr(y=1|x):          0.3218   [ 0.3116,    0.3320]
  Pr(y=2|x):          0.2591   [ 0.2538,    0.2643]
  Pr(y=3|x):          0.1390   [ 0.1290,    0.1491]
  Pr(y=4|x):          0.0560   [ 0.0490,    0.0629]
  Pr(y=5|x):          0.0180   [ 0.0149,    0.0212]
  Pr(y=6|x):          0.0048   [ 0.0037,    0.0059]
  Pr(y=7|x):          0.0011   [ 0.0008,    0.0014]
  Pr(y=8|x):          0.0002   [ 0.0001,    0.0003]
  Pr(y=9|x):          0.0000   [ 0.0000,    0.0001]

          fem        mar       kid5        phd       ment
x=  .46010929  .66229508  .49508197  3.1031093  8.7672131

added matrices:
    e(_estadd_prvalue) :  1 x 66
  e(_estadd_prvalue_x) :  1 x 5

. eststo nbreg: quietly nbreg art fem mar kid5 phd ment, nolog

. estadd prvalue

nbreg: Predictions for art

Confidence intervals by delta method

                                95% Conf. Interval
  Rate:                1.602   [ 1.4936,    1.7104]
  Pr(y=0|x):          0.2978   [ 0.2788,    0.3167]
  Pr(y=1|x):          0.2794   [ 0.2727,    0.2860]
  Pr(y=2|x):          0.1889   [ 0.1859,    0.1919]
  Pr(y=3|x):          0.1113   [ 0.1051,    0.1174]
  Pr(y=4|x):          0.0607   [ 0.0549,    0.0664]
  Pr(y=5|x):          0.0315   [ 0.0273,    0.0357]
  Pr(y=6|x):          0.0158   [ 0.0130,    0.0186]
  Pr(y=7|x):          0.0077   [ 0.0061,    0.0094]
  Pr(y=8|x):          0.0037   [ 0.0028,    0.0046]
  Pr(y=9|x):          0.0018   [ 0.0012,    0.0023]

          fem        mar       kid5        phd       ment
x=  .46010929  .66229508  .49508197  3.1031093  8.7672131

added matrices:
    e(_estadd_prvalue) :  1 x 66
  e(_estadd_prvalue_x) :  1 x 5

. estadd prvalue post, swap: *

. esttab, b(4) nostar ci wide compress ///
>     mtitles eqlabels(none)

----------------------------------------------------------------
                 (1)                        (2)                 
             poisson                      nbreg                 
----------------------------------------------------------------
mu            1.6101  [1.5265,1.6937]    1.6020  [1.4936,1.7104]
0             0.1999  [0.1832,0.2166]    0.2978  [0.2788,0.3167]
1             0.3218  [0.3116,0.3320]    0.2794  [0.2727,0.2860]
2             0.2591  [0.2538,0.2643]    0.1889  [0.1859,0.1919]
3             0.1390  [0.1290,0.1491]    0.1113  [0.1051,0.1174]
4             0.0560  [0.0490,0.0629]    0.0607  [0.0549,0.0664]
5             0.0180  [0.0149,0.0212]    0.0315  [0.0273,0.0357]
6             0.0048  [0.0037,0.0059]    0.0158  [0.0130,0.0186]
7             0.0011  [0.0008,0.0014]    0.0077  [0.0061,0.0094]
8             0.0002  [0.0001,0.0003]    0.0037  [0.0028,0.0046]
9             0.0000  [0.0000,0.0001]    0.0018  [0.0012,0.0023]
----------------------------------------------------------------
N                915                        915                 
----------------------------------------------------------------
95% confidence intervals in brackets

. eststo clear

[do-file]

Zero-truncated Poisson/negative binomial

ztp/ztnb and fitstat

. spex couart2
(Academic Biochemists / S Long)

. drop if art==0 // artificially truncated the data
(275 observations deleted)

. eststo ztp: quietly ztp art fem mar kid5 phd ment, nolog

. eststo ztnb: quietly ztnb art fem mar kid5 phd ment, nolog

. estadd fitstat : *

. esttab, wide scalars(r2_mf r2_mfadj aic0 aic_n) mtitles

----------------------------------------------------------------------
                      (1)                          (2)                
                      ztp                         ztnb                
----------------------------------------------------------------------
art                                                                   
fem                -0.229***      (-3.51)       -0.245*        (-2.52)
mar                0.0965          (1.32)        0.103          (0.95)
kid5               -0.142**       (-2.93)       -0.153*        (-2.12)
phd               -0.0127         (-0.41)     -0.00293         (-0.06)
ment               0.0187***       (8.22)       0.0237***       (5.54)
_cons               0.671***       (5.48)        0.355          (1.80)
----------------------------------------------------------------------
lnalpha                                                               
_cons                                           -0.603**       (-2.68)
----------------------------------------------------------------------
N                     640                          640                
r2_mf              0.0436                       0.0212                
r2_mfadj           0.0383                       0.0146                
aic0                3.394                        3.232                
aic_n              2172.1                       2068.6                
----------------------------------------------------------------------
t statistics in parentheses
* p<0.05, ** p<0.01, *** p<0.001

. eststo clear

[do-file]

ztp/ztnb and listcoef

. spex couart2
(Academic Biochemists / S Long)

. drop if art==0 // artificially truncated the data
(275 observations deleted)

. eststo ztp: quietly ztp art fem mar kid5 phd ment, nolog

. eststo ztnb: quietly ztnb art fem mar kid5 phd ment, nolog

. estadd listcoef fem ment: *

. estadd listcoef fem ment, percent nosd : *

. esttab, cell("b_facts b_pcts") keep(fem ment) mtitles

----------------------------------------------------------------
                      (1)                       (2)             
                      ztp                      ztnb             
                  b_facts       b_pcts      b_facts       b_pcts
----------------------------------------------------------------
fem              .8926405    -10.73595     .8855335    -11.44665
ment             1.213629     21.36292     1.277855     27.78551
----------------------------------------------------------------
N                     640                       640             
----------------------------------------------------------------

. eststo clear

[do-file]

ztp/ztnb and prchange: effects on rate

. spex couart2
(Academic Biochemists / S Long)

. drop if art==0 // artificially truncated the data
(275 observations deleted)

. eststo ztp: quietly ztp art fem mar kid5 phd ment, nolog

. eststo ztnb: quietly ztnb art fem mar kid5 phd ment, nolog

. estadd prchange: *

. esttab, aux(dc) nopar wide mtitles

----------------------------------------------------------------------
                      (1)                          (2)                
                      ztp                         ztnb                
----------------------------------------------------------------------
art                                                                   
fem                -0.229***       -0.463       -0.245*         -0.389
mar                0.0965           0.195        0.103           0.164
kid5               -0.142**        -0.216       -0.153*         -0.183
phd               -0.0127         -0.0257     -0.00293        -0.00466
ment               0.0187***        0.398       0.0237***        0.396
_cons               0.671***                     0.355                
----------------------------------------------------------------------
lnalpha                                                               
_cons                                           -0.603**              
----------------------------------------------------------------------
N                     640                          640                
----------------------------------------------------------------------
dc in second column
* p<0.05, ** p<0.01, *** p<0.001

. eststo clear

[do-file]

ztp/ztnb and prchange: effects on probability of selected outcome

. spex couart2
(Academic Biochemists / S Long)

. drop if art==0 // artificially truncated the data
(275 observations deleted)

. eststo ztp0: quietly ztp art fem mar kid5 phd ment, nolog

. eststo ztnb0: quietly ztnb art fem mar kid5 phd ment, nolog

. estadd prchange, outcome(0) : *0

. eststo ztp1: quietly ztp art fem mar kid5 phd ment, nolog

. eststo ztnb1: quietly ztnb art fem mar kid5 phd ment, nolog

. estadd prchange, outcome(1) : *1

. esttab, main(dc) nostar not scalars(outcome predval) mtitles

----------------------------------------------------------------
                      (1)          (2)          (3)          (4)
                     ztp0        ztnb0         ztp1        ztnb1
----------------------------------------------------------------
fem                0.0610       0.0662       0.0622       0.0207
mar               -0.0259      -0.0281      -0.0263     -0.00874
kid5               0.0278       0.0307       0.0292      0.00993
phd               0.00331     0.000781      0.00348     0.000253
ment              -0.0510      -0.0661      -0.0534      -0.0213
----------------------------------------------------------------
N                     640          640          640          640
outcome                 0            0            1            1
predval             0.129        0.315        0.264        0.270
----------------------------------------------------------------
dc coefficients

. eststo clear

[do-file]

ztp/ztnb and prvalue

. spex couart2
(Academic Biochemists / S Long)

. drop if art==0 // artificially truncated the data
(275 observations deleted)

. eststo ztp: quietly ztp art fem mar kid5 phd ment, nolog

. estadd prvalue

ztp: Predictions for art

                      Uncond     Cond
  Rate:               2.0507   2.3534
  Pr(y=0|x):          0.1286        .
  Pr(y=1|x):          0.2638   0.3028
  Pr(y=2|x):          0.2705   0.3104
  Pr(y=3|x):          0.1849   0.2122
  Pr(y=4|x):          0.0948   0.1088
  Pr(y=5|x):          0.0389   0.0446
  Pr(y=6|x):          0.0133   0.0152
  Pr(y=7|x):          0.0039   0.0045
  Pr(y=8|x):          0.0010   0.0011
  Pr(y=9|x):          0.0002   0.0003

          fem        mar       kid5        phd       ment
x=    .440625    .671875    .471875  3.1539765   10.14375

added matrices:
    e(_estadd_prvalue) :  1 x 77
  e(_estadd_prvalue_x) :  1 x 5

. eststo ztnb: quietly ztnb art fem mar kid5 phd ment, nolog

. estadd prvalue

ztnb: Predictions for art

                      Uncond     Cond
  Rate:               1.6097   2.3505
  Pr(y=0|x):          0.3152        .
  Pr(y=1|x):          0.2698   0.3940
  Pr(y=2|x):          0.1787   0.2609
  Pr(y=3|x):          0.1067   0.1559
  Pr(y=4|x):          0.0603   0.0881
  Pr(y=5|x):          0.0329   0.0481
  Pr(y=6|x):          0.0175   0.0256
  Pr(y=7|x):          0.0092   0.0134
  Pr(y=8|x):          0.0047   0.0069
  Pr(y=9|x):          0.0024   0.0035

          fem        mar       kid5        phd       ment
x=    .440625    .671875    .471875  3.1539765   10.14375

added matrices:
    e(_estadd_prvalue) :  1 x 77
  e(_estadd_prvalue_x) :  1 x 5

. estadd prvalue post, swap: *

. esttab, b(4) nostar not mtitles eqlabels(none)

--------------------------------------
                      (1)          (2)
                      ztp         ztnb
--------------------------------------
mu                 2.0507       1.6097
0                  0.1286       0.3152
1                  0.2638       0.2698
2                  0.2705       0.1787
3                  0.1849       0.1067
4                  0.0948       0.0603
5                  0.0389       0.0329
6                  0.0133       0.0175
7                  0.0039       0.0092
8                  0.0010       0.0047
9                  0.0002       0.0024
--------------------------------------
N                     640          640
--------------------------------------

. eststo clear

[do-file]

Zero-inflated Poisson/negative binomial

zip/zinb and fitstat

. spex couart2
(Academic Biochemists / S Long)

. eststo zip: quietly zip art fem mar kid5 phd ment, ///
>                     inf(fem mar kid5 phd ment) nolog

. eststo zinb: quietly zinb art fem mar kid5 phd ment, ///
>                     inf(fem mar kid5 phd ment) nolog

. estadd fitstat : *

. esttab, wide scalars(r2_mf r2_mfadj aic0 aic_n) mtitles

----------------------------------------------------------------------
                      (1)                          (2)                
                      zip                         zinb                
----------------------------------------------------------------------
art                                                                   
fem                -0.209***      (-3.30)       -0.196**       (-2.59)
mar                 0.104          (1.46)       0.0976          (1.16)
kid5               -0.143**       (-3.02)       -0.152**       (-2.80)
phd              -0.00617         (-0.20)    -0.000700         (-0.02)
ment               0.0181***       (7.89)       0.0248***       (7.10)
_cons               0.641***       (5.28)        0.417**        (2.90)
----------------------------------------------------------------------
inflate                                                               
fem                 0.110          (0.39)        0.636          (0.75)
mar                -0.354         (-1.11)       -1.499         (-1.60)
kid5                0.217          (1.10)        0.628          (1.42)
phd               0.00127          (0.01)      -0.0377         (-0.12)
ment               -0.134**       (-2.96)       -0.882**       (-2.79)
_cons              -0.577         (-1.13)       -0.192         (-0.14)
----------------------------------------------------------------------
lnalpha                                                               
_cons                                           -0.976***      (-7.21)
----------------------------------------------------------------------
N                     915                          915                
r2_mf              0.0444                       0.0372                
r2_mfadj           0.0373                       0.0292                
aic0                3.534                        3.416                
aic_n              3233.5                       3126.0                
----------------------------------------------------------------------
t statistics in parentheses
* p<0.05, ** p<0.01, *** p<0.001

. eststo clear

[do-file]

zip/zinb and listcoef

. spex couart2
(Academic Biochemists / S Long)

. eststo zip: quietly zip art fem mar kid5 phd ment, ///
>                     inf(fem mar kid5 phd ment) nolog

. eststo zinb: quietly zinb art fem mar kid5 phd ment, ///
>                     inf(fem mar kid5 phd ment) nolog

. estadd listcoef fem ment : *

. estadd listcoef fem ment, percent nosd : *

. esttab, cell("b_facts b_pcts") keep(fem ment) mtitles

----------------------------------------------------------------
                      (1)                       (2)             
                      zip                      zinb             
                  b_facts       b_pcts      b_facts       b_pcts
----------------------------------------------------------------
art                                                             
fem              .9009586    -9.904139     .9071068    -9.289321
ment             1.187247     18.72472     1.264998     26.49976
----------------------------------------------------------------
inflate                                                         
fem              1.056254     5.625355     1.373176     37.31758
ment             .2802993    -71.97007     .0002323    -99.97677
----------------------------------------------------------------
N                     915                       915             
----------------------------------------------------------------

. eststo clear

[do-file]

zip/zinb and prchange: effects on rate

. spex couart2
(Academic Biochemists / S Long)

. eststo zip: quietly zip art fem mar kid5 phd ment, ///
>                     inf(fem mar kid5 phd ment) nolog

. eststo zinb: quietly zinb art fem mar kid5 phd ment, ///
>                     inf(fem mar kid5 phd ment) nolog

. estadd prchange: *

. esttab, aux(dc) nopar wide mtitles

----------------------------------------------------------------------
                      (1)                          (2)                
                      zip                         zinb                
----------------------------------------------------------------------
art                                                                   
fem                -0.209***       -0.380       -0.196**        -0.331
mar                 0.104           0.258       0.0976           0.164
kid5               -0.143**        -0.226       -0.152**        -0.198
phd              -0.00617         -0.0106    -0.000700        -0.00116
ment               0.0181***        0.594       0.0248***        0.422
_cons               0.641***                     0.417**              
----------------------------------------------------------------------
inflate                                                               
fem                 0.110                        0.636                
mar                -0.354                       -1.499                
kid5                0.217                        0.628                
phd               0.00127                      -0.0377                
ment               -0.134**                     -0.882**              
_cons              -0.577                       -0.192                
----------------------------------------------------------------------
lnalpha                                                               
_cons                                           -0.976***             
----------------------------------------------------------------------
N                     915                          915                
----------------------------------------------------------------------
dc in second column
* p<0.05, ** p<0.01, *** p<0.001

. eststo clear

[do-file]

zip/zinb and prchange: effects on probability of selected outcome

. spex couart2
(Academic Biochemists / S Long)

. eststo zip0: quietly zip art fem mar kid5 phd ment, ///
>                     inf(fem mar kid5 phd ment) nolog

. eststo zinb0: quietly zinb art fem mar kid5 phd ment, ///
>                     inf(fem mar kid5 phd ment) nolog

. estadd prchange, outcome(0) : *0

. eststo zip1: quietly zip art fem mar kid5 phd ment, ///
>                     inf(fem mar kid5 phd ment) nolog

. eststo zinb1: quietly zinb art fem mar kid5 phd ment, ///
>                     inf(fem mar kid5 phd ment) nolog

. estadd prchange, outcome(1) : *1

. esttab, main(dc) nostar not scalars(outcome predval) mtitles

----------------------------------------------------------------
                      (1)          (2)          (3)          (4)
                     zip0        zinb0         zip1        zinb1
----------------------------------------------------------------
art                                                             
fem                0.0609       0.0547       0.0439       0.0229
mar               -0.0624      -0.0277      -0.0110      -0.0112
kid5               0.0429       0.0324       0.0198       0.0138
phd               0.00156     0.000186      0.00136    0.0000842
ment               -0.173      -0.0752      0.00357      -0.0238
----------------------------------------------------------------
N                     915          915          915          915
outcome                 0            0            1            1
predval             0.258        0.269        0.236        0.278
----------------------------------------------------------------
dc coefficients

. eststo clear

[do-file]

zip/zinb and prvalue

. spex couart2
(Academic Biochemists / S Long)

. eststo zip: quietly zip art fem mar kid5 phd ment, ///
>     inf(fem mar kid5 phd ment) nolog

. estadd prvalue

zip: Predictions for art

  Expected y:         1.7032
  Pr(Always0|z):      0.1388
  Pr(y=0|x,z):        0.2580
  Pr(y=1|x):          0.2357
  Pr(y=2|x):          0.2331
  Pr(y=3|x):          0.1536
  Pr(y=4|x):          0.0760
  Pr(y=5|x):          0.0300
  Pr(y=6|x):          0.0099
  Pr(y=7|x):          0.0028
  Pr(y=8|x):          0.0007
  Pr(y=9|x):          0.0002

x values for count equation

          fem        mar       kid5        phd       ment
x=  .46010929  .66229508  .49508197  3.1031093  8.7672131

z values for binary equation

          fem        mar       kid5        phd       ment
z=  .46010929  .66229508  .49508197  3.1031093  8.7672131

added matrices:
    e(_estadd_prvalue) :  1 x 72
  e(_estadd_prvalue_x) :  1 x 5
 e(_estadd_prvalue_x2) :  1 x 5

. eststo zinb: quietly zinb art fem mar kid5 phd ment, ///
>     inf(fem mar kid5 phd ment) nolog

. estadd prvalue

zinb: Predictions for art

  Expected y:          1.701
  Pr(Always0|z):      0.0002
  Pr(y=0|x,z):        0.2687
  Pr(y=1|x):          0.2784
  Pr(y=2|x):          0.1987
  Pr(y=3|x):          0.1204
  Pr(y=4|x):          0.0665
  Pr(y=5|x):          0.0346
  Pr(y=6|x):          0.0172
  Pr(y=7|x):          0.0083
  Pr(y=8|x):          0.0039
  Pr(y=9|x):          0.0018

x values for count equation

          fem        mar       kid5        phd       ment
x=  .46010929  .66229508  .49508197  3.1031093  8.7672131

z values for binary equation

          fem        mar       kid5        phd       ment
z=  .46010929  .66229508  .49508197  3.1031093  8.7672131

added matrices:
    e(_estadd_prvalue) :  1 x 72
  e(_estadd_prvalue_x) :  1 x 5
 e(_estadd_prvalue_x2) :  1 x 5

. estadd prvalue post, swap: *

. esttab, b(4) nostar not wide compress ///
>     mtitles eqlabels(none)

------------------------------
                 (1)       (2)
                 zip      zinb
------------------------------
Ey            1.7032    1.7010
All0          0.1388    0.0002
0|xy          0.2580    0.2687
1|x           0.2357    0.2784
2|x           0.2331    0.1987
3|x           0.1536    0.1204
4|x           0.0760    0.0665
5|x           0.0300    0.0346
6|x           0.0099    0.0172
7|x           0.0028    0.0083
8|x           0.0007    0.0039
9|x           0.0002    0.0018
------------------------------
N                915       915
------------------------------

. eststo clear

[do-file]

Linear regression

regress and fitstat/listcoef

. spex regjob2
(Academic Biochemists / S Long)

. quietly regress job fem phd ment fel art cit

. estadd fitstat

Measures of Fit for regress of job

Log-Lik Intercept Only:       -567.512   Log-Lik Full Model:           -519.397
D(401):                       1038.793   LR(6):                          96.230
                                         Prob > LR:                       0.000
R2:                              0.210   Adjusted R2:                     0.198
AIC:                             2.580   AIC*n:                        1052.793
BIC:                         -1371.725   BIC':                          -60.162
BIC used by Stata:            1080.872   AIC used by Stata:            1052.793

added scalars:
                e(dev) =  1038.7933
             e(dev_df) =  401
               e(lrx2) =  96.229915
            e(lrx2_df) =  6
             e(lrx2_p) =  1.533e-18
             e(r2_adj) =  .19828803
               e(aic0) =  2.5803757
              e(aic_n) =  1052.7933
               e(bic0) =  -1371.7248
              e(bic_p) =  -60.162312
           e(statabic) =  1080.8722
           e(stataaic) =  1052.7933
              e(n_rhs) =  6
             e(n_parm) =  7

. estadd listcoef

regress (N=408): Unstandardized and Standardized Estimates 

 Observed SD: .97360294
 SD of Error: .8717482

-------------------------------------------------------------------------------
         job |      b         t     P>|t|    bStdX    bStdY   bStdXY      SDofX
-------------+-----------------------------------------------------------------
         fem |  -0.13919   -1.543   0.124  -0.0680  -0.1430  -0.0698     0.4883
         phd |   0.27268    5.529   0.000   0.2601   0.2801   0.2671     0.9538
        ment |   0.00119    1.692   0.091   0.0778   0.0012   0.0799    65.5299
         fel |   0.23414    2.469   0.014   0.1139   0.2405   0.1170     0.4866
         art |   0.02280    0.789   0.430   0.0514   0.0234   0.0528     2.2561
         cit |   0.00448    2.275   0.023   0.1481   0.0046   0.1521    33.0599
-------------------------------------------------------------------------------

added matrices:
               e(b_xs) :  1 x 6      (bStdX)
               e(b_ys) :  1 x 6      (bStdY)
              e(b_std) :  1 x 6      (bStdXY)
              e(b_sdx) :  1 x 6      (SDofX)

. esttab, aux(b_std) wide scalars(aic0 aic_n bic0 bic_p)

-----------------------------------------
                      (1)                
                      job                
-----------------------------------------
fem                -0.139       (-0.0698)
phd                 0.273***      (0.267)
ment              0.00119        (0.0799)
fel                 0.234*        (0.117)
art                0.0228        (0.0528)
cit               0.00448*        (0.152)
_cons               1.067***             
-----------------------------------------
N                     408                
aic0                2.580                
aic_n              1052.8                
bic0              -1371.7                
bic_p              -60.16                
-----------------------------------------
b_std in parentheses
* p<0.05, ** p<0.01, *** p<0.001

[do-file]

regress and prvalue

. spex regjob2
(Academic Biochemists / S Long)

. quietly regress job fem phd ment fel art cit

. estadd prvalue, x(ment=min)  label(ment=min)

regress: Predictions for job

                                95% Conf. Interval
  Predicted y:        2.1795   [ 2.0743,    2.2846]

          fem        phd       ment        fel        art        cit
x=  .38970588  3.2005637          0  .61764706  2.2769608  21.715686

added matrices:
    e(_estadd_prvalue) :  1 x 6
  e(_estadd_prvalue_x) :  1 x 6

. estadd prvalue, x(ment=mean) label(ment=mean)

regress: Predictions for job

                                95% Conf. Interval
  Predicted y:        2.2334   [ 2.1488,     2.318]

          fem        phd       ment        fel        art        cit
x=  .38970588  3.2005637  45.470584  .61764706  2.2769608  21.715686

updated matrices:
    e(_estadd_prvalue) :  2 x 6
  e(_estadd_prvalue_x) :  2 x 6

. estadd prvalue, x(ment=max)  label(ment=max)

regress: Predictions for job

                                95% Conf. Interval
  Predicted y:        2.8108   [ 2.1369,    3.4847]

          fem        phd       ment        fel        art        cit
x=  .38970588  3.2005637  531.99988  .61764706  2.2769608  21.715686

updated matrices:
    e(_estadd_prvalue) :  3 x 6
  e(_estadd_prvalue_x) :  3 x 6

. estadd prvalue post

scalars:
                  e(N) =  408

macros:
             e(depvar) : "job"
                e(cmd) : "estadd_prvalue"
              e(model) : "regress"
         e(properties) : "b"

matrices:
                  e(b) :  1 x 3      (predictions)
                 e(se) :  1 x 3      (standard errors)
                 e(LB) :  1 x 3      (lower CI bounds)
                 e(UB) :  1 x 3      (upper CI bounds)
           e(Category) :  1 x 3      (outcome values)
                  e(X) :  6 x 3      (fem, phd, ment, fel, art, cit)

. esttab, ci nostar wide eqlabels(none)

------------------------------------------------
                      (1)                       
                      job                       
------------------------------------------------
ment=min            2.179          [2.074,2.285]
ment=mean           2.233          [2.149,2.318]
ment=max            2.811          [2.137,3.485]
------------------------------------------------
N                     408                       
------------------------------------------------
95% confidence intervals in brackets

[do-file]

Censored normal regression

tobit and fitstat/listcoef

. spex tobjob2
(Academic Biochemists / S Long)

. quietly tobit jobcen fem phd ment fel art cit, ll(1) nolog

. estadd fitstat

Measures of Fit for tobit of jobcen

Log-Lik Intercept Only:       -604.850   Log-Lik Full Model:           -560.252
D(400):                       1120.504   LR(6):                          89.195
                                         Prob > LR:                       0.000
McFadden's R2:                   0.074   McFadden's Adj R2:               0.061
ML (Cox-Snell) R2:               0.196   Cragg-Uhler(Nagelkerke) R2:      0.207
McKelvey & Zavoina's R2:         0.205                              
Variance of y*:                  1.488   Variance of error:               1.182
AIC:                             2.786   AIC*n:                        1136.504
BIC:                         -1284.003   BIC':                          -53.128
BIC used by Stata:            1168.594   AIC used by Stata:            1136.504

added scalars:
                e(dev) =  1120.5042
             e(dev_df) =  400
               e(lrx2) =  89.195123
            e(lrx2_df) =  6
             e(lrx2_p) =  4.452e-17
              e(r2_mf) =  .0737333
           e(r2_mfadj) =  .06050687
              e(r2_ml) =  .19636934
              e(r2_cu) =  .20704523
              e(r2_mz) =  .20535921
            e(v_ystar) =  1.4875706
            e(v_error) =  1.1820843
               e(aic0) =  2.7855494
              e(aic_n) =  1136.5042
               e(bic0) =  -1284.0027
              e(bic_p) =  -53.12752
           e(statabic) =  1168.5943
           e(stataaic) =  1136.5042
              e(n_rhs) =  6
             e(n_parm) =  8

. estadd listcoef

tobit (N=408): Unstandardized and Standardized Estimates 

 Observed SD: .97360294
   Latent SD: 1.21966
 SD of Error: 1.087237

-------------------------------------------------------------------------------
      jobcen |      b         t     P>|t|    bStdX    bStdY   bStdXY      SDofX
-------------+-----------------------------------------------------------------
         fem |  -0.23685   -2.032   0.043  -0.1156  -0.1942  -0.0948     0.4883
         phd |   0.32258    5.047   0.000   0.3077   0.2645   0.2523     0.9538
        ment |   0.00134    1.514   0.131   0.0880   0.0011   0.0722    65.5299
         fel |   0.32527    2.656   0.008   0.1583   0.2667   0.1298     0.4866
         art |   0.03391    0.929   0.353   0.0765   0.0278   0.0627     2.2561
         cit |   0.00509    2.057   0.040   0.1683   0.0042   0.1380    33.0599
-------------------------------------------------------------------------------

added matrices:
               e(b_xs) :  1 x 6      (bStdX)
               e(b_ys) :  1 x 6      (bStdY)
              e(b_std) :  1 x 6      (bStdXY)
              e(b_sdx) :  1 x 6      (SDofX)

. esttab, aux(b_std) wide scalars(r2_mfadj r2_ml r2_cu r2_mz)

-----------------------------------------
                      (1)                
                   jobcen                
-----------------------------------------
model                                    
fem                -0.237*      (-0.0948)
phd                 0.323***      (0.252)
ment              0.00134        (0.0722)
fel                 0.325**       (0.130)
art                0.0339        (0.0627)
cit               0.00509*        (0.138)
_cons               0.685**              
-----------------------------------------
sigma                                    
_cons               1.087***             
-----------------------------------------
N                     408                
r2_mfadj           0.0605                
r2_ml               0.196                
r2_cu               0.207                
r2_mz               0.205                
-----------------------------------------
b_std in parentheses
* p<0.05, ** p<0.01, *** p<0.001

[do-file]

cnreg and fitstat/listcoef

. spex tobjob2
(Academic Biochemists / S Long)

. gen cens = -(jobcen<=1)

. quietly cnreg jobcen fem phd ment fel art cit, censored(cens) nolog

. estadd fitstat

Measures of Fit for cnreg of jobcen

Log-Lik Intercept Only:       -604.850   Log-Lik Full Model:           -560.252
D(400):                       1120.504   LR(6):                          89.195
                                         Prob > LR:                       0.000
McFadden's R2:                   0.074   McFadden's Adj R2:               0.061
ML (Cox-Snell) R2:               0.196   Cragg-Uhler(Nagelkerke) R2:      0.207
McKelvey & Zavoina's R2:         0.205                              
Variance of y*:                  1.488   Variance of error:               1.182
AIC:                             2.786   AIC*n:                        1136.504
BIC:                         -1284.003   BIC':                          -53.128
BIC used by Stata:            1168.594   AIC used by Stata:            1136.504

added scalars:
                e(dev) =  1120.5042
             e(dev_df) =  400
               e(lrx2) =  89.195123
            e(lrx2_df) =  6
             e(lrx2_p) =  4.452e-17
              e(r2_mf) =  .0737333
           e(r2_mfadj) =  .06050687
              e(r2_ml) =  .19636934
              e(r2_cu) =  .20704523
              e(r2_mz) =  .20535921
            e(v_ystar) =  1.4875706
            e(v_error) =  1.1820843
               e(aic0) =  2.7855494
              e(aic_n) =  1136.5042
               e(bic0) =  -1284.0027
              e(bic_p) =  -53.12752
           e(statabic) =  1168.5943
           e(stataaic) =  1136.5042
              e(n_rhs) =  6
             e(n_parm) =  8

. estadd listcoef

cnreg (N=408): Unstandardized and Standardized Estimates 

 Observed SD: .97360294
   Latent SD: 1.21966
 SD of Error: 1.087237

-------------------------------------------------------------------------------
      jobcen |      b         t     P>|t|    bStdX    bStdY   bStdXY      SDofX
-------------+-----------------------------------------------------------------
         fem |  -0.23685   -2.032   0.043  -0.1156  -0.1942  -0.0948     0.4883
         phd |   0.32258    5.047   0.000   0.3077   0.2645   0.2523     0.9538
        ment |   0.00134    1.514   0.131   0.0880   0.0011   0.0722    65.5299
         fel |   0.32527    2.656   0.008   0.1583   0.2667   0.1298     0.4866
         art |   0.03391    0.929   0.353   0.0765   0.0278   0.0627     2.2561
         cit |   0.00509    2.057   0.040   0.1683   0.0042   0.1380    33.0599
-------------------------------------------------------------------------------

added matrices:
               e(b_xs) :  1 x 6      (bStdX)
               e(b_ys) :  1 x 6      (bStdY)
              e(b_std) :  1 x 6      (bStdXY)
              e(b_sdx) :  1 x 6      (SDofX)

. esttab, aux(b_std) wide scalars(r2_mfadj r2_ml r2_cu r2_mz)

-----------------------------------------
                      (1)                
                   jobcen                
-----------------------------------------
model                                    
fem                -0.237*      (-0.0948)
phd                 0.323***      (0.252)
ment              0.00134        (0.0722)
fel                 0.325**       (0.130)
art                0.0339        (0.0627)
cit               0.00509*        (0.138)
_cons               0.685**              
-----------------------------------------
sigma                                    
_cons               1.087***             
-----------------------------------------
N                     408                
r2_mfadj           0.0605                
r2_ml               0.196                
r2_cu               0.207                
r2_mz               0.205                
-----------------------------------------
b_std in parentheses
* p<0.05, ** p<0.01, *** p<0.001

[do-file]

intreg and fitstat/listcoef

. spex tobjob2
(Academic Biochemists / S Long)

. gen jobcen0 = jobcen if jobcen>1
(99 missing values generated)

. intreg jobcen0 jobcen fem phd ment fel art cit, nolog

Interval regression                               Number of obs   =        408
                                                  LR chi2(6)      =      89.20
Log likelihood = -560.25209                       Prob > chi2     =     0.0000

------------------------------------------------------------------------------
             |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         fem |  -.2368486   .1165852    -2.03   0.042    -.4653513   -.0083458
         phd |   .3225846   .0639229     5.05   0.000      .197298    .4478712
        ment |   .0013436   .0008876     1.51   0.130     -.000396    .0030832
         fel |   .3252657   .1224575     2.66   0.008     .0852534     .565278
         art |   .0339053   .0365017     0.93   0.353    -.0376367    .1054474
         cit |     .00509   .0024752     2.06   0.040     .0002388    .0099412
       _cons |   .6854061   .2182717     3.14   0.002     .2576014    1.113211
-------------+----------------------------------------------------------------
    /lnsigma |   .0836397   .0428043     1.95   0.051    -.0002552    .1675346
-------------+----------------------------------------------------------------
       sigma |   1.087237   .0465384                      .9997449    1.182386
------------------------------------------------------------------------------

  Observation summary:        99  left-censored observations
                             309     uncensored observations
                               0 right-censored observations
                               0       interval observations

. estadd fitstat

Measures of Fit for intreg of jobcen0 jobcen

Log-Lik Intercept Only:       -604.850   Log-Lik Full Model:           -560.252
D(400):                       1120.504   LR(6):                          89.195
                                         Prob > LR:                       0.000
McFadden's R2:                   0.074   McFadden's Adj R2:               0.061
ML (Cox-Snell) R2:               0.196   Cragg-Uhler(Nagelkerke) R2:      0.207
McKelvey & Zavoina's R2:         0.160                              
Variance of y*:                  1.408   Variance of error:               1.182
AIC:                             2.786   AIC*n:                        1136.504
BIC:                         -1284.003   BIC':                          -53.128
BIC used by Stata:            1168.594   AIC used by Stata:            1136.504

added scalars:
                e(dev) =  1120.5042
             e(dev_df) =  400
               e(lrx2) =  89.195124
            e(lrx2_df) =  6
             e(lrx2_p) =  4.452e-17
              e(r2_mf) =  .0737333
           e(r2_mfadj) =  .06050687
              e(r2_ml) =  .19636935
              e(r2_cu) =  .20704524
              e(r2_mz) =  .16016798
            e(v_ystar) =  1.4075249
            e(v_error) =  1.1820845
               e(aic0) =  2.7855494
              e(aic_n) =  1136.5042
               e(bic0) =  -1284.0027
              e(bic_p) =  -53.127521
           e(statabic) =  1168.5943
           e(stataaic) =  1136.5042
              e(n_rhs) =  6
             e(n_parm) =  8

. estadd listcoef

intreg (N=408): Unstandardized and Standardized Estimates 

    LHS vars: jobcen0 jobcen
 Observed SD: .77904266
   Latent SD: .48211618
 SD of Error: .08363969

-------------------------------------------------------------------------------
             |      b         t     P>|t|    bStdX    bStdY   bStdXY      SDofX
-------------+-----------------------------------------------------------------
         fem |  -0.23685   -2.032   0.043  -0.1156  -0.4913  -0.2399     0.4883
         phd |   0.32258    5.046   0.000   0.3077   0.6691   0.6382     0.9538
        ment |   0.00134    1.514   0.131   0.0880   0.0028   0.1826    65.5299
         fel |   0.32527    2.656   0.008   0.1583   0.6747   0.3283     0.4866
         art |   0.03391    0.929   0.354   0.0765   0.0703   0.1587     2.2561
         cit |   0.00509    2.056   0.040   0.1683   0.0106   0.3490    33.0599
-------------------------------------------------------------------------------

added matrices:
               e(b_xs) :  1 x 6      (bStdX)
               e(b_ys) :  1 x 6      (bStdY)
              e(b_std) :  1 x 6      (bStdXY)
              e(b_sdx) :  1 x 6      (SDofX)

. esttab, aux(b_std) wide scalars(r2_mfadj r2_ml r2_cu r2_mz)

-----------------------------------------
                      (1)                
                  jobcen0                
-----------------------------------------
model                                    
fem                -0.237*       (-0.240)
phd                 0.323***      (0.638)
ment              0.00134         (0.183)
fel                 0.325**       (0.328)
art                0.0339         (0.159)
cit               0.00509*        (0.349)
_cons               0.685**              
-----------------------------------------
lnsigma                                  
_cons              0.0836                
-----------------------------------------
N                     408                
r2_mfadj           0.0605                
r2_ml               0.196                
r2_cu               0.207                
r2_mz               0.160                
-----------------------------------------
b_std in parentheses
* p<0.05, ** p<0.01, *** p<0.001

[do-file]

tobit/cnreg/intreg and prvalue

. spex tobjob2
(Academic Biochemists / S Long)

. eststo tobit: quietly tobit jobcen fem phd ment fel art cit, ll(1) nolog

. gen cens = -(jobcen<=1)

. eststo cnreg: quietly cnreg jobcen fem phd ment fel art cit, censored(cens) n
> olog

. gen jobcen0 = jobcen if jobcen>1
(99 missing values generated)

. eststo intreg: quietly intreg jobcen0 jobcen fem phd ment fel art cit, nolog

. estadd prvalue, x(ment=min)  label(ment=min)  : *

. estadd prvalue, x(ment=mean) label(ment=mean) : *

. estadd prvalue, x(ment=max)  label(ment=max)  : *

. estadd prvalue post : *

. esttab, se nostar eqlabels(none) mtitles

---------------------------------------------------
                      (1)          (2)          (3)
                    tobit        cnreg       intreg
---------------------------------------------------
ment=min            2.014        2.014        2.014
                 (0.0695)     (0.0695)     (0.0695)

ment=mean           2.075        2.075        2.075
                 (0.0563)     (0.0563)     (0.0563)

ment=max            2.729        2.729        2.729
                  (0.435)      (0.435)      (0.435)
---------------------------------------------------
N                     408          408          408
---------------------------------------------------
Standard errors in parentheses

. eststo clear

[do-file]

The three models are formally equivalent in this case and, therefore, yield identical predictions.