** user=**** 
** password=**** 
** package=stata
** project=lis

#delimit;

* SWITZERLAND LIS 1992 : Example CAMSIS matching to CH 92 data;

use $ch92p;

* Current occ unit (btap minor groups)=pocc; 

* most Employment status info available on plfs;
gen empst=plfs;
recode empst 1 2=1 3 5 6=6 4=5 *=0;
label define stat 
   0 "Status unknown"
   1 "Self-employed (all)"
   2 "Self-Employed (principals)" 
   5 "Family Assistant" 
   6 "Employee";
label values empst stat;
label variable empst "Employment status";
tab empst psex;

* Key indexing variables are occ (btap) and empst;
gen btap=-999;
replace btap=pocc if (pocc > 0 );
tabulate empst if (pocc > 0 ), 
       summarize(btap) mean standard freq ;
sort btap empst;
save cstemp, replace;
clear;
infile btap empst mcam fcam siops isei 
      using $mydata\camsis\cschis92.dat;
summarize;
sort btap empst;
merge btap empst using cstemp;
gen occinfo=_merge;
tab occinfo; 
drop if (occinfo==1);
* (this removes those btap empst values which are not represented in the LES sample);
tab occinfo psex; 

* Current values are within gender; 
* also can make a cross-gender CAMSIS variable : ;
gen cgentcs=-999;
replace cgentcs=mcam if (psex==1);
replace cgentcs=fcam if (psex==2);

summarize btap empst cgentcs  siops isei ;
summarize btap empst cgentcs  siops isei if pocc > 0;
summarize btap empst cgentcs  siops isei if (pocc > 0 & siops > 0); 
summarize btap empst cgentcs  siops isei if (pocc > 0 & siops > 0 & cgentcs > 0); 

* A small number of occs don't have camsis or isei/siops 
* scores matched to them because their btap minor groups are not
* covered by the camsis index file;


mvdecode cgentcs, mv(-999);
* Declare some missing values;

***************************************;

* Quick assessment: compare values by an education recode;
di "Education";
* some recode of education ;
gen educ=peduc; 
recode educ 0 1=1 2 3 4 5 7 8 9=2 
  6 10 11 12 13 14 15=3 ;
label define edu  1 "Primary or less" 2 "Intermediate" 
 3 "High or Degree level "; 
label values educ edu;
label variable educ "education categories";
tab educ  ;
* note high level of missing data, 50 percent of cases;
* this is because this var is only collected for household heads;
tab psex;
* note a couple of missing psexes;
keep if (psex == 1 | psex ==2);
sort psex;
by psex:
  summarize educ 
      cgentcs siops isei if cgenscs > 0; 
by psex: 
 oneway cgentcs educ  if (pocc > 0), tabulate;
by psex: 
 oneway siops educ  if (pocc > 0 & siops > 0), tabulate;
by psex: 
 oneway isei educ  if (pocc > 0 & isei > 0), tabulate;


**************************************;