54. Chasing the (Causal) Dragon: Intermediate Quantitative Data Analysis for Sociologists
Sociologists and other social scientists are often interested in understanding causal and dynamic social processes such as:
“How do the places we live, work, and play get under the skin and affect health and well-being across the life course?”
“Does upward social class mobility change one’s political attitudes?”
“What social currents are responsible for changes in support for same-sex marriage across historical time?”
“Are long-standing racial inequalities declining, persisting, or increasing in recent years?”
Many of these questions are methodologically difficult to answer with observational (non-experimental) data, and they require that we get a handle on the study of change, context, and causality. You likely have learned how to answer questions like these with standard OLS (linear) regression techniques and cross-sectional data, which remain useful tools in social scientists’ methodological toolbox. But these techniques are also quite limited, and impose strict assumptions that do not allow us to meet many of our goals, adequately answer our questions, or provide stringent tests of our theories and hypotheses.
In this course, we’ll pick up where introductory statistics courses leave off, and get an introduction to more advanced statistical methods for observational data, including but not limited to: regression for categorical dependent variables, fixed and random effects models, and hierarchical linear modeling. This course will be a mix of seminar and lecture, where we will be focused on understanding how we can use these methods to better meet our goals and answer our research questions. Put differently, this course is less focused on going “under the hood” and more focused on “how to drive”—specifically, we will interrogate the assumptions and use of these statistical methods in the social sciences and learn how to implement these methods using STATA. This will include: discussion of core methodological assumptions and limitations, how to apply these statistical methods in different settings, and learning when specific methods are appropriate tools and when they are not. We will explore these issues through student-led discussions, hands-on data analysis, and dissecting the application of these methods in academic journal articles. As part of this course, you will be exposed to (and critique) a wide range of sociological research published in our major disciplinary journals. The course will culminate in an independent research project where students will analyze data and use the one or more of the modeling techniques discussed during the term to answer a sociological research question of their choosing. SOCY 10 or equivalent and a basic understanding of STATA is required to enroll in this course. Dist: QDS. Houle.