Education 255C - Statistical Methods for School-Based Intervention Studies

Referred to as Educ 255C SEM - Data Analysis
on the UCLA web site

Winter 2006
January 9 - March 15

Instructor:  Bengt Muthen (bmuthen@ucla.edu)
Course Times: Mon-Wed 3-5, Moore Hall 2120
Office Hours: Mon-Wed 2-3, Moore Hall 2023

new page
TA: Karen Nylund (knylund@ucla.edu) Shaunna Clark (shclark@ucla.edu)
Optional Lab Session: Monday 5:15 - 7:00PM, Moore Hall 3140
(*Note: Because of the holiday on Jan. 16, the first lab session will be Wed., Jan 11)

Syllabus, Handouts & Suggested Readings

The course will discuss a collection of statistical analyses, designs, and applications related to intervention studies based in schools as well as in other settings. A major focus will be placed on randomized studies, reflecting the increasing emphasis on such studies by the US Department of Education, but quasi-experimental studies will also be covered. The course gives a broad and application-oriented overview of related statistical techniques presented as non-technically as possible. This includes topics such as logistic regression, mediation analysis involving risk and protective factors, latent class cluster analysis techniques, growth and growth mixture modeling of developmental trajectories, multilevel modeling and value-added techniques, survival analysis, propensity score techniques, randomized trial designs and analysis techniques, principal stratification, counterfactual approaches to causal inference including complier-average causal effect estimation, and missing data-attrition modeling. Applications include evaluation of reading achievement programs, growth modeling for school accountability, studies of reading failure, high-school dropout, as well as NIH-oriented preventive interventions including school-based preventive interventions related to conduct disorder, ADHD, teenage suicide, alcohol abuse and dependence. Brief overviews will also be given for related techniques in risk and prevention studies related to analysis of twin and family data, candidate gene linkage and association analysis, and gene-environment susceptibility analysis. Assignments are both applied and analytical, involving reporting and discussions of currently ongoing and planned studies as well as computer analyses where students can choose packages such as SAS, SPSS, Splus, Stata, and Mplus. Prerequisite: linear regression analysis.