anonymous posted on Saturday, December 17, 2005 - 11:53 am
my question is if is possible for LGC models in Mplus to estimate the effects of a predictor variable which is constant for all individuals at each time point, but changes across time points? My model covers individual attitudes towards government policies measured by a five-wave panel study and I now want to estimate the effects of unemployment rate on the slope of the LGC for individual attitudes. Would that be possible?
bmuthen posted on Saturday, December 17, 2005 - 1:12 pm
You can do this as type = twolevel analysis, using cluster = id, so that variation across time is on the within level and variation across individuals is on the between level. The predictor variable should then be put on the within = list so that no between-level variance (variance across individuals) is estimated for this variable.
anonymous posted on Saturday, December 17, 2005 - 2:11 pm