12 Mixed Effects Multilevel Models

Lecture slides: Mixed Models

Today’s exercise: Web Excercise 12

A primary object of sociological inquiry is how social environments, ranging from the family and the neighborhood to the state, shape human action. An enthusiasm for such context effects motivates many of the greatest examples of sociological thought—from and his classical investigation of anomie in Le Suicide to and his recent portrayal of the Great American City. In recent decades, scholars have increasingly examined context effects using quantitative statistical analysis, thanks in part to a growing availability of suitable data. Suuch analyses pose statistical challenges because observations belonging to the same contextual unit or cluster (e.g., a school, census district, or coun- try) tend to be more similar than two randomly chosen observations, violating the independence assumption of traditional regression analysis. Statisticians and practitioners in sociology and other disciplines have developed several distinct approaches for addressing this complication, subsumed under the term “multilevel modeling techniques.”

Today I will briefly introduce you to the prevalent approach in sociology: Mixed-effects multilevel models.