** user=**** ** password=**** ** package=spss ** project=lis * Austrian LIS 1995 : Example CAMSIS matching . * This example on the person file, change var names to use hhld file. * Occ var index is isco88, used is pocc with isco88 4 digit . * Austrian data has employment status 0 - unknown, * 1 - self-employed, 6 - employee . get file= at95p . * Employment status info from ptypewk. compute empst=ptypewk. recode empst (1,2=1) (3,4,5,6,7,8=6) (else=0). add value labels empst 0 "Unknown" 1 "Self-employed" 6 "Employee" variable label empst "Employment status". cro empst by psex. compute isco88 =-999. if (pocc > 0 ) isco88=pocc . temp. select if (pocc > 0). means tables=isco88 by empst /cells=min max count . sort cases by isco88 empst. sav out= cstemp . get file="u:/camsis/csatis95.sav". descriptives var=all. sort cases by isco88 empst. match files table=* /in=csocc /file=cstemp /in=source /by=isco88 empst. fre var=csocc source. select if (source=1). * (this removes those occbst values which are not * represented in the LIS sample). cro source by psex. * Current values are within gender. * also can make a cross-gender CAMSIS variable : . compute cgentcs=-999. if (psex=1) cgentcs=mcam. if (psex=2) cgentcs=fcam. descriptives var=isco88 empst cgentcs siops isei . temp. select if (pocc > 0). descriptives var=isco88 empst cgentcs siops isei . temp. select if (pocc > 0 & siops > 0). descriptives var=isco88 empst cgentcs siops isei . * note that the last summary shows us that a handful of cases * have occs and camsis scores but have missing siops * scores by the ISMF translation. * note also that a lot more cases are missing on egp, because of * insufficient empst info. * Quick assessment : compare values by an education recode. compute educ=peduc. recode educ (1=1) (2,3,4=2) (5,6,7,8=3) (else=-9). add value labels educ 1 "Elementary" 2 "Intermediate" 3 "Higher level" -9 "Unknown". variable label educ "education categories". fre var= educ . descriptives var= educ cgentcs siops isei . temp. select if (pocc > 0 & siops > 0 & psex=1 & educ > 0). means tables= cgentcs siops isei by educ /statistics=anova. temp. select if (pocc > 0 & siops > 0 & psex=2 & educ > 0). means tables= cgentcs siops isei by educ /statistics=anova. /* For interest here are the Eta-squared statistics with education from above: Men: CAMSIS : 0.383 SIOPS : 0.288 ISEI: 0.325 Women: CAMSIS : 0.396 SIOPS : 0.307 ISEI: 0.337 */