Are scores on high-stakes tests primarily a function of socioeconomic status? Do mandatory seat belt laws save lives? In this course, students will learn how to use a set of quantitative methods referred to as the general linear model--regression, correlation, analysis of variance, and analysis of covariance--to address these and other questions that arise in educational, psychological, and social research. [Syllabus]
Our strategy involves learning statistical analysis by doing statistical analysis. During the semester, we address a variety of substantive research questions by analyzing dozens of data sets and fitting increasingly sophisticated regression models. As we learn how to use regression models in practice, we willl discuss their:
- Purpose. What types of research questions is the model most useful for addressing?
- Mathematical representation. How does the model algebraically capture the relationship(s) we’re trying to examine?
- Assumptions. What assumptions need we make so that we can fit the model to data? How do we determine whether these assumptions hold? What should we do when they don’t?
- Implementation. How do we get the computer to do the calculations?
- Interpretation. How do we interpret the computer results? What inferences may we make? What inferences shouldn’t we make?
- Presentation. How should we present results to a technical audience? To a non-technical audience?
- Relationship to other statistical methods. How is regression similar to and different from other methods you’ve learned or read about?
- Implications for research design. How should the next study be designed so that we’d be in better shape to address our research questions?
- Limitations. What cautions and caveats should we be aware of, and how should we convey these concerns to technical and non-technical audiences?
By the end of the semester, students' statistical skills should be sufficiently developed that they can critically examine other people’s research and carefully perform some of their own.