Courses
Primary Courses taught by Miratrix
The following are the primary courses taught by Luke Miratrix. S022 and S043/Stat151 are typically offered yearly. The others are upon occasion.
EDU S-022: Introduction to Statistical Computing and Data Science in Education
This course teaches a synergetic blend of statistical computing and re-sampling (permutation and bootstrap) methods. Statistical computing allows more flexible investigation of data, such as generating customized visualizations and summarizations or custom-tailoring an analysis. Re-sampling methods can often allow for principled data analysis in circumstances where, for example, the parametric assumptions behind more traditional analyses such as linear regression are held in doubt or the sample sizes are too small for asymptotics to hold. They can also be used when ones estimands and estimators of interest are too complex for theoretical approximations. This course teaches how to program in R, a widely adopted statistical computing platform, and uses case studies and projects to give students hands-on experience. This is an applied course in that the goal is to learn contemporary methods that can immediately be applied to one's own work.
EDU S-043/Stat-151: Multilevel and Longitudinal Models
Data often have structure that needs to be modeled explicitly. For example, when investigating students' outcomes we need to account for the fact that students are nested inside classes that are in turn nested inside schools. If we are watching students develop over time, we need to account for the dependence of measurements across time. If we do not, our inferences will tend to be overly optimistic and wrong. The course provides an overall framework, the multilevel and generalized multilevel (hierarchical) model, for thinking about and analyzing these forms of data. We will focus on specific versions of these tools for the most common forms of longitudinal and clustered data. This course will focus on applied work, using real data sets and the statistical software R. R will be specifically taught and supported. While the primary focus will be on the linear model with continuous outcomes (i.e., the classic regression framework) we will also discuss binary, categorical, and ordinal outcomes. We will emphasize how to think about the applicability of these methods, how they might fail, and what one might do to protect oneself in such circumstances. Applications of hierarchical (multi-level) models will include the canonical specific cases of random-slope, random-intercept, mixed effect, crossed effect, marginal, and growth-curve models.
EDU S-049M: Simulation Design in R
This course focuses on how to write Monte Carlo simulations in R. This course is therefore primarily designed for students interested in writing simulations for the "Simulation" section of an academic paper. We will also touch on other uses of simulation, but this purpose is the main driver of the course. Along the way we will deepen our R programming skills and statistical intuition.
EDU S071: Doctoral Workshop on the Analysis of Complex Data
The use of statistical tools in education and the social sciences is benefiting from new technologies, increased computational power, and innovative approaches to solving analytic problems. At the same time, these advancements have opened the door to types of data and a volume of data not easily tackled using classic approaches. This workshop is dedicated to learning the statistical methodology relevant for those areas where the classic methods fall short. Building our course around students’ own research puzzles and questions, we will investigate and apply viable solutions to challenges faced by empirical researchers in education and related fields.
Selected Long Ago Courses
- Stat 328: Bayesian Nonparametrics
- Stat 242: Permutation and Resampling Based Statistical Methods
- Stat 240: Matched Sampling and Study Design (joint with Don Rubin)
- Stat 100: Introduction to Quantitative Methods for the Social Sciences and Humanities