Hi Statalist,

I am trying to generate a change score that captures changes in labour force participation. I have three survey waves. Using binary variable cremp2 (employment status, Yes/No) I run the following command:

Code:
bysort Findid (round): gen change = cremp2-cremp2[_n-1]
recode change -1=2
label define chnnge 0 "No change" 1 "Unemployed - employed" 2 "employed - unemployed"
label values change chnnge
I am not sure if this is the correct way to generate the variable I need as it only produces three categories. I thought I would get the following categories:

No change
Unemployed – employed – unemployed
Employed – unemployed – unemployed
Unemployed – unemployed – employed
Employed – employed – unemployed
Unemployed – employed – employed

And so on – to basically captures all possible changes in LFP over the three waves. Do I understand this correctly?

Below is an example of my dataset, you can see the change variable produces two observations for each panel ID at T=2012 (change from 2006-2012) and T=2018 (change from 2012-2018). Whereas I am trying to produce one observation at T=2018 for each individual in my panel dataset (i.e. change from 2006 - 2012 - 2018)

Code:
* Example generated by -dataex-. To install: ssc install dataex
clear
input double Findid byte(round cremp2_) float change
601002402  6 0 .
601002402 12 0 0
601002402 18 0 0
601004603  6 0 .
601004603 12 0 0
601004603 18 0 0
601005902  6 1 .
601005902 12 1 0
601005902 18 1 0
601011602  6 0 .
601011602 12 0 0
601011602 18 0 0
601012406  6 0 .
601012406 12 0 0
601012406 18 0 0
601019209  6 0 .
601019209 12 0 0
601019209 18 1 1
601020203  6 1 .
601020203 12 1 0
601020203 18 0 2
601024002  6 1 .
601024002 12 0 2
601024002 18 0 0
601025503  6 0 .
601025503 12 0 0
601025503 18 0 0
601025702  6 0 .
601025702 12 0 0
601025702 18 0 0
601030704  6 0 .
601030704 12 0 0
601030704 18 0 0
601031403  6 1 .
601031403 12 1 0
601031403 18 1 0
601032102  6 1 .
601032102 12 1 0
601032102 18 0 2
601032309  6 0 .
601032309 12 0 0
601032309 18 0 0
601033103  6 0 .
601033103 12 0 0
601033103 18 0 0
601036602  6 0 .
601036602 12 0 0
601036602 18 0 0
601037505  6 0 .
601037505 12 0 0
601037505 18 0 0
601038302  6 0 .
601038302 12 0 0
601038302 18 0 0
601045003  6 0 .
601045003 12 0 0
601045003 18 0 0
601048702  6 0 .
601048702 12 0 0
601048702 18 0 0
601050804  6 0 .
601050804 12 0 0
601050804 18 0 0
601051404  6 0 .
601051404 12 0 0
601051404 18 0 0
601054402  6 0 .
601054402 12 0 0
601054402 18 0 0
601055607  6 0 .
601055607 12 0 0
601055607 18 1 1
601056302  6 0 .
601056302 12 0 0
601056302 18 0 0
601058302  6 0 .
601058302 12 1 1
601058302 18 1 0
601058602  6 0 .
601058602 12 0 0
601058602 18 0 0
601059402  6 0 .
601059402 12 0 0
601059402 18 1 1
601060002  6 0 .
601060002 12 0 0
601060002 18 1 1
601060202  6 0 .
601060202 12 0 0
601060202 18 0 0
601060302  6 0 .
601060302 12 0 0
601060302 18 0 0
601061602  6 0 .
601061602 12 0 0
601061602 18 0 0
601061802  6 1 .
601061802 12 1 0
601061802 18 1 0
601062002  6 0 .
end
label values round year
label def year 6 "2006", modify
label def year 12 "2012", modify
label def year 18 "2018", modify
label values change chnnge
label def chnnge 0 "No change", modify
label def chnnge 1 "Unemployed - employed", modify
label def chnnge 2 "employed - unemployed", modify