{smcl} {* 26jun2025}{...} {cmd:help erepost}{...} {right:{browse "http://github.com/benjann/erepost/"}} {hline} {title:Title} {p 4 4 2}{hi:erepost} {hline 2} Repost the estimation results {title:Syntax} {p 8 16 2}{cmd:erepost} [{cmd:b =} {it:b}] [{cmd:V =} {it:V}] [{cmd:,} {cmd:cmd(}{it:string}{cmd:)} {cmdab:ren:ame} {cmd:se} {cmd:nov} {cmd:nob} {cmd:drop(}{it:namelist}{cmd:)} [{ul:{cmd:no}}]{cmdab:e:sample:}[{cmd:(}{it:varname}{cmd:)}] {it:ereturn_post_opts} ] {p 4 4 2} where {it:b} is a 1 x p coefficient vector (matrix) and {it:V} is a p x p covariance matrix. {title:Description} {p 4 4 2} {cmd:erepost} changes the {cmd:e(b)} or {cmd:e(V)} matrix of the current estimation results or changes the declared estimation sample. {cmd:erepost} is similar to {helpb ereturn repost}. However, {cmd:erepost} is allowed after estimation commands that do not post their results using {cmd:ereturn post} (e.g. {cmd:regress}) and {cmd:erepost} can be used outside of {cmd:eclass} programs (see help {helpb program}). {p 4 4 2} Technical notes: After applying {cmd:erepost} the original command (or {cmd:estimates replay}) may not be able to replay the output. Furthermore, {cmd:erepost} does not preserve hidden results. {title:Options} {p 4 8 2} {cmd:cmd(}{it:string}{cmd:)} (re)sets the {cmd:e(cmd)} macro. {p 4 8 2} {cmd:rename} causes the names from the coefficient vector to be used as the labels for both the coefficient vector and the covariance matrix in case of name conflicts. {p 4 8 2} {cmd:se} causes vector {cmd:e(se)} to be added to the results, containing the square roots of the diagonal elements of the variance matrix. {p 4 8 2} {cmd:nov} omits {cmd:e(V)} from the reposted results. {cmd:nov} has no effect if {it:V} is provided. {p 4 8 2} {cmd:nob} omits {cmd:e(b)} from the reposted results. {cmd:nob} has no effect if {it:b} is provided. {cmd:nob} implies {cmd:nov}. {p 4 8 2} {cmd:drop(}{it:namelist}{cmd:)} omits the specified macros, scalars, or matrices from the reposted results, where {it:namelist} is a space-separated list of the names of the elements to be omitted. The specified names may contain {cmd:*} and {cmd:?} wildcards. Note that {cmd:drop()} cannot be used to remove the coefficient vector {cmd:e(b)} or the variance matrix {cmd:e(V)}, and that macro {cmd:e(properties)} will be re-created automatically. {p 4 8 2} {cmd:esample(}{it:varname}{cmd:)} provides the name of the 0/1 variable indicating the observations involved in the estimation. The variable will be removed from the data but will be available for use as {hi:e(sample)}. {p 8 8 2} If option {cmd:esample()} is omitted, the existing {cmd:e(sample)} will be preserved. Specify {cmd:noesample} if you want to remove {cmd:e(sample)} from the estimation results. {p 4 8 2} {it:ereturn_post_opts} are any other options allowed with {cmd:ereturn post} (see help {helpb ereturn}). {title:Examples} {p 4 4 2} Say, you are estimating a regression model that contains all interactions between the regressors and a binary variable and you want to display the main effects and the interaction effects in two columns side by side. One approach is to modify the names in {cmd:e(b)} and {cmd:e(V)} so that there is a "main equation" and an "interaction equation" and then tabulate the model using {cmd:estout} or {cmd:esttab} with the {cmd:unstack} option. Example (you will need the {cmd:estout} package to run this example; type {net "describe http://fmwww.bc.edu/repec/bocode/e/estout.pkg":ssc describe estout}): {com}. sysuse auto {txt}(1978 Automobile Data) {com}. generate foreign_mpg = foreign*mpg {txt} {com}. generate foreign_weight = foreign*weight {txt} {com}. quietly regress price foreign_mpg foreign_weight foreign mpg weight {txt} {com}. matrix b = e(b) {txt} {com}. matrix coleq b = foreign foreign foreign main main main {txt} {com}. matrix colname b = mpg weight _cons mpg weight _cons {txt} {com}. erepost b=b, rename {txt} {com}. esttab , unstack order(main: foreign:) {res} {txt}{hline 44} {txt} (1) {txt} price {txt} main foreign {txt}{hline 44} {txt}mpg {res} 237.7 -257.5 {txt} {res} {ralign 12:{txt:(}1.90{txt:)}} {ralign 12:{txt:(}-1.66{txt:)}} {txt} {txt}weight {res} 4.415*** 0.741 {txt} {res} {ralign 12:{txt:(}5.18{txt:)}} {ralign 12:{txt:(}0.45{txt:)}} {txt} {txt}_cons {res} -13285.4* 8219.6 {txt} {res} {ralign 12:{txt:(}-2.58{txt:)}} {ralign 12:{txt:(}1.13{txt:)}} {txt} {txt}{hline 44} {txt}N {res} 74 {txt} {txt}{hline 44} {txt}t statistics in parentheses {txt}* p<0.05, ** p<0.01, *** p<0.001 {title:Author} {p 4 4 2}Ben Jann, University of Bern, ben.jann@unibe.ch {p 4 4 2}Thanks for citing this software as follows: {p 8 8 2}Jann, B. (2007). erepost: Stata module to repost the estimation results. Available from {browse "http://ideas.repec.org/c/boc/bocode/s456850.html"}. {title:Also see} {psee} Online: {helpb ereturn}, {help estcom}, {helpb estimates}