Multilevel analysis is quite complicated - and obviously the goal of this course is not to provide complete clarity but to introduce this as an approach to dealing with the complex contextual factors that we have been discussing throughout this course. Essentially, we want to make sure you all are aware of the analytic approaches you can use to move beyond the "black box" or "chronic disease" epidemiology and to do analysis, when appropriate, that takes into account the socioecological model.
In response to your discussion of SES, another option is to treat it as a "confounder". This is complex, as we discussed in class, since SES does not cause race, but there can be reasons to control for it in this way.
In terms of maternal depression, multi level effects you could consider could include neighborhood (i.e. presence of parks, maternal support), policy level factors (i.e. maternal leave), etc.