** user=**** ** password=**** ** package=spss ** project=les * UK LES 1997 : SPSS Example CAMSIS matching. get file= uk97e . * Current occ unit (soc90)=fj04. * (incomplete) Employment status info from fj02 and fj06. * Derive employment status indicator. compute status=fj02. recode status (1=2) (2, 4=3) (3, 5=7) (else=9) . if (fj06 ge 5 & fj02=1) status=1 . add value labels status 1 "Self-Employed gt 25 employees" 2 "Self-Employed with employees" 3 "Self-Employed no employees" 4 "Manager" 6 "Supervisor / foreman" 7 "Employee" 9 "Unknown" . variable label status "Employment status". fre var=status. * Derive key index var called occbst. compute occbst=-999. if (fj04 > 0 ) occbst=(fj04*10) + status . temp. select if (fj04 > 0). means tables=occbst by status. * Match on it with the CAMSIS file. sort cases by occbst. sav out= cstemp . get file="u:/camsis/csukes97.sav" . descriptives var=all. sort cases by occbst. match files table=* /in=csocc /file=cstemp /in=source /by=occbst. fre var=csocc source. cro csocc by source. select if (source=1). * (this removes those occbst values which are not * represented in the LES sample). cro source by db02. * Current values are within gender. * also can make a cross-gender CAMSIS variable : . compute cgentcs=-999. if (db02=1) cgentcs=htcs. if (db02=2) cgentcs=wtcs. compute cgenscs=-999. if (db02=1) cgenscs=hscs. if (db02=2) cgenscs=wscs. * Assess distribution of variables. descriptives var=occbst cgentcs cgenscs siops isei . temp. select if (fj04 > 0). descriptives var=occbst cgentcs cgenscs siops isei . temp. select if (fj04 > 0 & siops > 0). descriptives var=occbst cgentcs cgenscs siops isei . fre var=egp. * 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 status info. * Quick assessment (1): compare values by egp. temp. select if (db02=1 & fj04 > 0 & siops > 0). summarize tables=cgentcs cgenscs siops isei by egp /cells=count mean /statistics=anova . temp. select if (db02=2 & fj04 > 0 & siops > 0). summarize tables=cgentcs cgenscs siops isei by egp /cells=count mean /statistics=anova . * Quick assessment (2): compare values by a UK education recode. compute educ=et01. recode educ (-9,-1, 16, 17, 18, 19, 22=1) (2,3,4,5,6,7,8,9,10,11,12,13,14,15,20,21=2) (1=3) . add value labels educ 1 "Primary or less" 2 "Intermediate" 3 "High or Degree level " -999 "Unknown". variable label educ "education categories". fre var= educ . descriptives var= educ cgentcs cgenscs siops isei . temp. select if (fj04 > 0 & siops > 0 & db02=1). means tables= cgentcs cgenscs siops isei by educ /statistics=anova. temp. select if (fj04 > 0 & siops > 0 & db02=2). means tables= cgentcs cgenscs siops isei by educ /statistics=anova. /* For interest here are the Eta-squared statistics with education from above: Men: CAMSIS title only : 0. CAMSIS title by status : 0. SIOPS : 0. ISEI: 0. Women: CAMSIS title only : 0. CAMSIS title by status : 0. SIOPS : 0. ISEI: 0. */