latent class analysis (lca) and latent profile.DOCVIP

Latent Class Analysis (LCA) and Latent Profile Analysis (LPA) Latent Class Analysis (LCA) and Latent Profile Analysis (LPA) can extend growth modeling when these are done as mixture models. We will review both LCA for binary indicators and LPA for continuous indicators. In addition to clarifying the vocabulary for subsequent sections, one outcome of this section will be a time varying covariate which will be a class variable. Conceptual Example: Developing a Typology of Parenting Style. There are many indicators we could use for parenting style. A few of them are: How closely the father and mother supervise or monitor the adolescent How supportive the father and mother are How closely the father and mother control the adolescent’s activities How often the father and mother praise the child. We can have a large number of families for which we have scores on each of these dimensions that are organized by the couple as the unit of analysis. The first five couples might look like the following, if scores were quantitative Couple His Super-vision Her Super-vision His Support Her Support His Control Her Control His Praise Her Praise A 9 9 9 9 9 9 8 9 B 3 4 9 9 4 4 9 9 C 2 3 1 3 8 8 2 3 D 1 9 1 8 1 9 1 9 E 6 2 8 1 6 1 9 1 and our data might look like this if scores were binary. Here we have assigned a 1 for any score above the threshold value of 4. Couple His Super-vision Her Super-vision His Support Her Support His Control Her Control His Praise Her Praise A 1 1 1 1 1 1 1 1 B 0 0 1 1 0 0 1 1 C 0 0 0 0 1 1 0 0 D 0 1 0 1 0 1 0 1 E 1 0 1 0 1 0 1 0 LPA and LCA attempt to find subsets of families who share similar patterns of responses. Couple A is high on all dimensions and this is true for both the husband and the wife. Couple B is supportive and praising their adolescent, but do little monitoring—joint permissive. Families conforming to this pattern might have very different adolescent outcomes than families in the first type. The third couple is hig