help for ^mkdat^ ^holrein^

SOEP - Retrivals ----------------

^mkdat^ varlist ^using^ pathname, ^w^aves^(^wavelist^)^ ^f^iles^(^filetype^) > ^ [^k^eep^(^ppfadvars^)^ ^n^etto^(^#,[#] ... [#]^)^ ^uc^ ^hhgr^ ^clear > ^]

^holrein^ varlist ^using^ pathname, ^w^aves^(^wavelist^)^ ^f^iles^(^filetype > ^)^ [^uc^]

Description ----------- ^mkdat^ and ^holrein^ are tools for SOEP-Retrivals. They merge variables from varlist to ppfad. mkdat is to generate a new file and holrein to merge new variables to a file generated with mkdat. Balanced (default) and various unbalanced designs could be specified.

Consider an example where individual gross and net income should be combined with information concerning the household income using 1991 and 1992 data (see Table 4.8a of the Desktop Companion). To make that file with mkdat and holrein, one may type interactivly

^mkdat hp5401 ip5401 hp5402 ip5402 using c:\soep, files(p) waves(h i)^ ^holrein hh49 ih49 using c:\soep, files(h) waves(h i)^

Note that ^varlist^ should be specified like the Item Correspondence List and should only contain variables from filetypes defined by the option ^files()^. After selecting waves, for which one wants to do the retrieval, one specifies the first variable of the retrival. After this follows the specification of the corresponding variable from the other waves. Then go on with the second variable. The specific form of the varlist could be easily seen, if one wrote the mkdat command above in the following form (an a do-file for example).

#delimit ; mkdat ^hp5401 ip5401^ ^hp5402 ip5402^ using c:\soep, files(p) waves(h i) ;

mkdat and holrein are especially constructed for use together with soepinfo. Consider one wants to observe political interest and party identification from 1984 to 1989. While working with soepinfo, one founds the variables and wrotes the Item-Correspondence list to a file. Let's see, how the file looks:

----------------------------------------------------------- 1984 |1985 |1986 |1987 |1988 |1989 ----------------------------------------------------------- Politik Politisches Interesse - |BP75 |CP75 |DP84 |EP73 |FP89 Politik Allgemeine Parteienpraeferenz AP5601 |BP7901 |CP7901 |DP8801 |EP7701 |FP9301 Politik Parteienidentifikation AP5602 |BP7902 |CP7902 |DP8802 |EP7702 |FP9302

After cutting all lines and headings and maybe change uppercase to lowercase (but also note ^uc^ mentioned below). That leads to

^- bp75 cp75 dp84 ep73 fp89^ ^ap5601 bp7901 cp7901 dp8801 ep7701 fp9301^ ^ap5602 bp7902 cp7902 dp8802 ep7702 fp9302^

which is exactly the structure varlist should have. Before one can put that table into the mkdat (or holrein) command, there is still one thing which should be changed:

^ - ^ bp75 cp75 dp84 ep73 fp89

Note that there is no variable name for political interest in 1984. mkdat will not accept a varlist which contains single dashes. Note also, that mkdat would not accept a varlist which is not completly rectangular. One have to change single dashes from the soepinfo outfile into a specific variable name. The new variable name should start with ^_^ and shouldn't be used twice in varlist. You may use _X1, _X2 and so on. However, the names do not appear in the generated file .

The mkdat command therefore becomes

#delimit ; mkdat ^_x1 bp75 cp75 dp84 ep73 fp89^ ^ap5601 bp7901 cp7901 dp8801 ep7701 fp9301^ ^ap5602 bp7902 cp7902 dp8802 ep7702 fp9302^ using c:\soep, files(p) waves(a b c d e f) ;

Options -------

^waves()^ is not optional. List letters of all waves you wish to observe inside the parantheses. You can not expand the wavelist from mkdat with holrein (But you may choose fewer waves in holrein). The wavelist should be in the same order as variables appear in the varlist. Consider the first example above:

^mkdat hp5401 ip5401 hp5402 ip5402 using c:\soep, wave(i p) files(p)^

Note that the proper specification of wave() is wave(^h i^) not wave(i h). Note also that there has to be a ^{blank}^ between the letters of wavelist.

If you include variables from the specific ost-files in 1990 and 1991 the wavelist should include ^gost^ and ^host^ at the corresponding position. For example:

^mkdat gp4301 zp4101 hp5401 using c:\soep, wave(g gost h) files(p)^

^files()^ is not optional. Specify a letter for the type of files where the variables of varlist come from. Do not include variables from different filetypes in the varlist. Instead of this create a second holrein statement. You can choose the following filetypes:

Letter for Filetype ... for Variables from files ^p^ ap, bp, ... lp, ... xp ^pgen^ apgen, bpgen, ... lpgen, ... xpgen ^pbrutto^ apbrutto, bpbrutto, ... lpbrutto, ... xpbrutto ^pequiv^ apequiv, bpequiv, ... lpequiv, ... xpequiv ^kind^ apkind, bpkind, ... lpkind, ... xpkind ^pausl^ apausl, bpausl, ... lpausl, ... xpausl ^h^ ah, bh, ... lh, ... xh ^hgen^ ahgen, bhgen, ... lhgen, ... xhgen ^hbrutto^ bhbrutto, ... lhbrutto, ... xhbrutto

Use wave() for variables from gpost, ghost and hpost.

^keep()^ lets you specify variables from ppfad that should be in the generated file. The variables *netto, *hhnr, persnr and hhnr are in the the file without specifiing keep. Do not specify them in the list. Use the option only for additional variables.


mkdat hp5401 ip5401 using c:\soep, files(p) waves(h i) ^keep(sex gebjahr)^

^netto()^ (mkdat only) defines the design of the generated file. One can list the values of the specific "netto"-variables in ppfad which should be included. For example netto(1) leads to an balanced panel design (default). If one wants to include temporary drop-outs he may specify netto(1,4). With netto(-3,-2,0,1,2,3,4) all cases of ppfad will be included.

^uc^ changes an uppercased varlist into lowercase. You may usually select uc if you work with varlist generated by soepinfo.

^hhgr^ (mkdat only) can be used to calculate the number of persons in the actual household using the persons listed in ppfad.

Examples --------

1) Constructing Longitudinal Individual Records (see Table 4.5 of the Desktop Compagnion)

^mkdat gp109 zp6401 hp10901 ip10901 jp10901 using c:\soep, waves(g gost h i j) files(p)^

2) Constructing Cross-Sectional Householt Records

^mkdat hh49 using c:\soep, files(h) waves(h)^ ^sort hhhnr^ ^by hhnr: gen index=_n^ ^by hhnr: keep if index==_N^

3) Linking Household Data to Individuals (see Table 4.7 of the Desktop Compagnion)

^mkdat hp501 hp5402, waves(h) files(p)^ ^holrein hh48, waves(h) files(p)^

4) Linking Houshold Data to Individuals Across Waves (see Table 4.8 of the Desktop Compagnion)

^mkdat hp5401 ip5401 hp5402 ip5402 using c:\soep, files(p) waves(h i)^ ^holrein hh49 ih49 using c:\soep, files(h) waves(h i)^

5) Houshold Level Variables from Individual Data (see Table 4.9 of the Desktop Compagnion)

^mkdat hp07 hp15 using c:\soep, files(p) waves(h)^ ^gen ft=1 if hp15==1^ ^gen pt=1 if hp15==2^ ^gen unemp=1 if hp07==1^ ^gen noinf=1 if hp15==9^ ^#delimit ; ^ ^collapse (count) n_ft=ft^ ^ n_pt=pt^ ^ n_unemp=unemp^ ^ n_noinf=noinf^ ^ (mean) hhnr=hhnr, by(hhnrakt);^ ^#delimit cr ^ ^holrein htyphh1 htyphh2 using c:\soep, waves(h) files(hgen)^