Generating Variable That Depicts Increase

[Nicholas McPherson]

My dataset comes from a survey with 5 waves that has a general score (variable GScore). I am trying to do the following:

1) I am trying to generate a new variables that shows the General Score increase from each wave to the next (so that a decrease in scores will be a negative value). What code would work best for this? I am currently trying to use the following unsuccessfully:

bysort id: gen GScore_increase = GScore - GScore[_n-1]

2) What code would calculate the overall mean increase (or potential decrease) in the GScore across all waves for all participants?

3) Is there an efficient code I could use to determine how many participants decreased or maintained the same GScore across all waves (i.e. how many never increased their Gscore across all the waves)?

[Joseph Coveney]

A first shot at your first question would be something like

bysort id (wave): generate double GScore_increase = GScore - GScore[_n-1] if _n > 1

but it assumes a lot about your dataset.

For your other two, it will be better to follow the forum's FAQ.

[Nicholas McPherson]

A first shot at your first question would be something like

bysort id (wave): generate double GScore_increase = GScore - GScore[_n-1] if _n > 1

but it assumes a lot about your dataset.

For your other two, it will be better to follow the forum's FAQ.

Unfortunately this did not work...does anyone else have any other ideas as to how to address this?

[Hemanshu Kumar]

It's hard to troubleshoot and come up with precise strategies in the absence of a data extract. As the Statalist FAQ advise (see esp. no. 12), please use the dataex command to post an extract.

[Rich Goldstein]

in addition, the FAQ also explains why "this did not work" is not helpful - more generally, please read the FAQ to get better answers