June 27-28, 2013 -- Boulder, Colorado
The purpose of this two-day course is to provide an introduction to multilevel modeling (also called hierarchical linear modeling). Multilevel are used when the units of observation are grouped (or "nested") within clusters; the paradigm examples are organizational settings (e.g., in which students are nested within schools) and longitudinal research (in which repeated observations are nested within individual subjects). The intended audience for this course is anyone who has completed an introductory course on regression, and wishes to become familiar with the concepts and procedures of multilevel models, possibly as preparation for more advanced study. After a brief but vigorous review of the core principles and assumptions of regression, we will discuss the motivations for multilevel models, the concepts of fixed and random effects, and the various types of multilevel models (including variance component models, random intercept models, and random coefficient or mixed models). We will consider a variety of examples, both organizational and longitudinal, and will focus on the connections between statistical models and substantive research questions. In addition, we will use Stata to specify, estimate, and interpret a variety of models; to get the most out of the course, participants should have Stata installed on their laptops.
- The course is open to current graduate students, postdoctoral fellows, and pre-tenure faculty from any institution. Participation in the course will be based on a competitive selection process. To apply, please send a CV and a one-page cover letter specifying your interest in the course to Brian Houle at email@example.com
- If desired, students may enroll in the course for 1 credit, and the CUPC will cover tuition expenses. Credit-seeking students will be expected to complete a final paper.