I have collected data during my field work for my research degree. I, with the use of this data, am trying to classify individuals in 4 levels of poverty: abject poor, very poor, poor and not poor. The data I have has continuous, ordered and nominal variable. Would you recommend using LCA or EFA in MPlus? Which version of Mplus should I buy for this analysis? Can I take into account missing data without loosing observations? Are there any examples available for this kind of mixture data using MPlus? Many thanks, Gaurav
LCA groups people. EFA groups variables. So it sounds like LCA would best meet your goals. You would need the Base Program and the Mixture Add-On to do this. Missing data estimation is available. Chapter 7 of the Mplus User's Guide contains example of LCA.