Statistics and data science: A broad perspective

Course goals

1. Refresh or acquire essential data-analysis techniques—such as ANOVA, linear regression, and logistic regression—to build a solid foundation for graduate-level quantitative research and data science coursework.
2. Interpret the outputs produced by applying these techniques to data (e.g., parameter estimates, uncertainty intervals, effect sizes, model fit, and predictions), and communicate what they imply in context.
3. Explain the key ideas and assumptions behind additional methods that may be new to you, including hierarchical/multilevel models, resampling and bootstrapping, mediation, structural equation modeling, clustering, and introductory machine-learning approaches.
4. Evaluate when particular techniques are (and are not) appropriate for a given research aim and data situation (e.g., explanation vs prediction, causal claims vs association, limited sample sizes, model assumption violations).
5. Summarize and explain statistical findings in a way that is accurate, transparent about uncertainty and limitations, and understandable to applied researchers and other non-specialist stakeholders.
6. Use R to implement an essential set of data-analysis workflows covered in the course (data preparation, model fitting, diagnostics, and basic reporting), and reproduce analyses from code.
7. Differentiate between explanatory and predictive research questions and reflect on the methodological value and implications of each for study design, modeling choices, and interpretation.
8. Explain and reason about the bias–variance tradeoff across different modeling contexts (e.g., simpler vs more flexible models, small vs large samples, regularization, resampling/validation).
9. Formulate and communicate statistical consulting advice to non-experts by clarifying goals, diagnosing common pitfalls, recommending analyses, and explaining tradeoffs and limitations.
10. Apply responsible research practices, including reproducible workflows, clear reporting, data/code documentation, and core open science principles appropriate to the discipline and dataset.
11. Develop a critical understanding of statistical and data-science reasoning, including the ability to assess claims made from data, identify common sources of bias, and recognize limits of inference.
12. Identify how specialized methods/topics covered here connect to later courses in the programme, and articulate what additional depth you would need to pursue them.

Content

This course introduces you to graduate-level methodology, statistics, and data science. The aim is to place your previous knowledge of techniques such as statistical tests or regression within a broader perspective in which you begin to see connections among models, techniques, and ways of reasoning about data. Our perspective will be structured around common themes, especially the bias-variance tradeoff and prediction versus explanation. After this course, you will be prepared for deeper dives into more specialized topics, such as those offered by the other courses within the programme.