Professor Rick Wash
The replication crisis in psychology has caused many people to question much of the research in the social sciences. While there are many reasons for this crisis, a number of people are pointing at inadequate and poorly conducted statistical analysis – and particularly the emphasis on Null Hypothesis Testing – as one of the major problems. This course teaches students the foundations of many common statistical techniques in the social sciences. It emphasizes statistical and analytical techniques that hold up under scrutiny and promote high quality social scientific inquiry.
While most of the time in the course is spent conducting and interpreting common statistics like t-tests and regressions, this class focuses attention not on which statistic is being used, but instead on how the statistics provide insight into the world and can be used to make strong arguments about social scientific concepts. I emphasize concepts like thinking about how good your measurements are, understanding the units of your variables, thinking about effect sizes instead of statistical significance, using statistics to answer “how much?” questions instead of “yes/no” questions, how statistical models provide control, and how to ensure that your analysis and results can be reproduced and replicated. In doing so, I will discuss a variety of statistical tools that are used in the social sciences, including t-tests, chi-squared tests, anova, OLS regression, logistic regression, multi-level models, statistics to make causal claims, and structural equation modeling.
After completing this course, students will have a strong foundation in statistics that will allow them to conduct high-quality statistical analyses of social scientific research. This foundation should provide them with an understanding of what statistical analysis in the social sciences will be like in their careers in the future. This foundation will also prepare them to learn more advanced analytical techniques, such as big data analysis, structural equation modeling, causal analysis, time series analysis, bayesian statistics, machine learning, and econometrics.