News and Updates

Categories

Inferring Cause and Effect in Time Series

December 15, 2015

One of the greatest challenges in social science research is to validly identify cause-and-effect relations. A recent interactive graphic in the New York Times highlights a prospective, quasi-experimental approach to linking increased gun sales to specific social and political events.

Read More

Within- and Between-Group Effects

December 9, 2015

An important concept to understand with multilevel data is the distinction between between-group and within-group effects. For instance, larger animal species (elephants) tend to live longer than smaller species (ducks); this is the between-group effect of size on life expectancy. Within a species, however, larger individuals (big ducks) tend to live less long than smaller individuals (small ducks); this is the within-group effect of size on life expectancy. Recently, Shankar Vedantan described another example on NPR: people report being happier making $50 for a project if their co-workers are making $40 versus making $60 for the same project if their co-workers are making $70. See the full story.

Read More

Using Listservs to get Advice

December 7, 2015

Although a bit old school, listservs such as SEMNET (for latent variable models) and the Multilevel Discussion List remain useful resources when grappling with quantitative modeling issues. You can search the archives to see if your question has come up before or post a new question to obtain fresh feedback / opinions from list members. Quick tip: use the “receive digest” setting to avoid having your inbox flooded with listserv posts. (more…)

Read More

Cross-classified item and person effects

November 22, 2015

Something that has come up often recently: When participants provide responses to a set of items or prompts, such as words, faces, or pictures, these items may represent a sample from a broader universe of possible items (e.g., a sampling of words from the lexicon). Responses may then best viewed as reflecting two crossed random factors, items and persons. A couple of nice papers on how to fit cross-classified random effects models to this kind of data are Baayen et al (2007) and Locker et al (2007).

Read More