The final exam in this class is comprehensive. That is, material from the entire semester may be involved. However, it will tend to focus on issues that arose since the midterm.
As you prepare for an exam, review the syllabus and ask yourself, "What was really important in each section of the class? What did we spend a lot of time on (either in class, or on homework assignments)?" With such questions in mind, let's review where we've been. After each major section, I have listed a number of questions that you can ask yourself to check your understanding. These questions do not represent an exhaustive list of material that might be covered on the exam.
Prior to the midterm exam, we had covered various aspects of multiple regression. Topics included simple matrix algebra, implementation and interpretation of multiple regression models, inference for regression, diagnostics and transformations, sequential regression, stepwise regression, and regression models involving categorical predictors and mixtures of categorical and continuous predictors. At the time of the midterm exam, we had recently covered interactions between continuous and categorical predictors that allow regression models to differ for different groups. You should remain comfortable with that material; probably at least one item on the final exam will relate to it.
Immediately following the midterm exam, we discussed regression models with interactions between continuous predictors. Some related skills you should have are involved in the following questions:
Next, we moved on to consider traditional path analysis. We saw how to create path diagrams involving measured variables, and dealt with different methods for estimating path coefficients. We defined saturated and unsaturated path models, as well as recursive and nonrecursive models. We learned how to estimate the coefficients of saturated models using sequential regression, and spent a little time on the issue of how to estimate the coefficients of unsaturated models algebraically. Some questions that might help you prepare for related exam items include:
Two years ago, I also included the following:
I have decided that this was excessively sadistic; you need not prepare for such questions.
After introducing path analysis, we considered exploratory and confirmatory factor analysis (EFA and CFA). We saw how to conduct an exploratory factor analysis in SAS and in Mplus. We discussed interpretation of factors, as well as factor rotation (both oblique and orthogonal). We noted that whereas in EFA, the data inform us about latent structure, in CFA, we posit a particular latent structure and assess whether the data fit that structure. Some questions that you should consider as you prepare for the exam include:
Next, we moved on to full-fledged structural equation modeling, examining path models that incorporate both regression-like aspects and factor analytic aspects. We spent time discussing the logic of such models, and learning to interpret related Mplus output. Some questions relevant to exam preparation include:
Our final topic for the semester was hierarchical linear modeling (HLM). We applied this first in the context of nested data structure such as kids nested within classrooms and classrooms nested within schools. We introduced the idea of randomly varying regression coefficients, along with the idea of variance components. We implemented and interpreted several examples of such analyses. Then we turned to a particularly useful application of HLM: longitudinal analysis. We examined several examples of growth-curve models, treating repeated measures across time as observations nested withing the individual subject. As you prepare for the exam, you should ask yourself if you can:
One other broad topic may arise on the exam: explanation of program syntax. Whereas you would never be expected to write a SAS or Mplus program on the spot during an exam, it is possible that you would be expected to provide annotations explaining each line of a simple program.