General question- does combining the 2013 and 2014 survey make sense, given that we expect a lot of overlap (same people surveyed)?
Also, why treat EA as a latent variable when it was directly surveyed? Shouldn’t we just combine by saying if you answered Yes to an EA questions, you are EA?
does combining the 2013 and 2014 survey make sense, given that we expect a lot of overlap (same people surveyed)?
Yes, because the surveys have (since the 2012 survey, IIRC) asked if you took a previous survey.
So I dropped anyone in the 2014 data who answered yes (s2014 <- subset(survey2014, PreviousSurveys!="Yes", select=...); as long as everyone was honest, there should be zero overlap/double-counting. I also included 2013 vs 2014 response as a covariate to allow estimating some of the heterogeneity.
Also, why treat EA as a latent variable when it was directly surveyed? Shouldn’t we just combine by saying if you answered Yes to an EA questions, you are EA?
If you answer both yes and also claim to go to EA events, aren’t you probably more of an EA than someone who says yes but doesn’t attend? In any event, it doesn’t make a difference. I just like using latent variables when I can.
General question- does combining the 2013 and 2014 survey make sense, given that we expect a lot of overlap (same people surveyed)?
Also, why treat EA as a latent variable when it was directly surveyed? Shouldn’t we just combine by saying if you answered Yes to an EA questions, you are EA?
Yes, because the surveys have (since the 2012 survey, IIRC) asked if you took a previous survey.
So I dropped anyone in the 2014 data who answered yes (
s2014 <- subset(survey2014, PreviousSurveys!="Yes", select=...); as long as everyone was honest, there should be zero overlap/double-counting. I also included 2013 vs 2014 response as a covariate to allow estimating some of the heterogeneity.If you answer both yes and also claim to go to EA events, aren’t you probably more of an EA than someone who says yes but doesn’t attend? In any event, it doesn’t make a difference. I just like using latent variables when I can.