For instructors
Using this book in a course
What the book assumes
Graduate students in the social sciences, including those who have not taken a methods course recently and who do not identify as methodologists. No programming background is assumed by the text itself; the exercises require some, and this site’s Getting started page is designed to close that gap in the first week.
Course fit
The book is built for a semester-length graduate methods course on computational or AI-assisted social science research. It also works as:
- a research design course with computational examples (emphasize the theory chapters);
- a methods practicum where students build one project across the term (emphasize the exercise chapters, which chain into a single pipeline);
- a supplementary text in a data science course that needs the social-science research-design framing its main text lacks.
The structural logic
Chapters alternate between concept and practice. Theory chapters (1, 2, 3, 5, 7, 9, 11, 12) develop the reasoning; exercise chapters (4, 6, 8, 10) apply it to a dataset each student builds from their own field’s literature.
That last design choice matters for teaching. Because each student works with their own corpus, the exercises cannot be completed by copying, and students arrive at the seminar with genuinely different results to compare. The cost is that you cannot debug all submissions the same way; the setup and primer pages on this site absorb much of that support load.
A possible fifteen-week shape
| Weeks | Chapters | Deliverable |
|---|---|---|
| 1–2 | 1, 2 | Project environment set up; research direction drafted |
| 3–4 | 3, 4 | Corpus retrieved, screened, documented; codebook and ERD |
| 5–7 | 5, 6 | Coding pipeline: dictionary, LLM, classifier; validity audit |
| 8–10 | 7, 8 | Network, prediction, and embedding analyses |
| 11–12 | 9, 10 | Integrative artifact with an uncertainty layer |
| 13–14 | 11 | Ethics review of the student’s own pipeline |
| 15 | 12 | Final presentation |
Assessment
Every exercise names a deliverable, which makes them directly assignable. The Reflective Exercises work well as seminar discussion prompts or short written responses; they ask students to interpret what they did rather than to report that they did it, which is harder to outsource to a model.
On that point: since the book teaches AI-assisted methods, forbidding AI in assessment would be incoherent. The alternative it supports is requiring documentation. Ask students to submit the promptbook and promptlog alongside the deliverable. A student who used a model well can show you exactly how; one who used it thoughtlessly cannot.
Instructor materials
SAGE commissions PowerPoint slides and test materials for the book; these are distributed through SAGE’s instructor site rather than here. Everything else is on this site and free:
- Chapter materials: one hub per chapter, which is the page to give students at the start of each unit. Each states the deliverable, lists prerequisites, and carries the notebooks and data.
- CSS Empirical Studies Database: 122 annotated studies, filterable by design stage. The four theory-chapter hubs pre-filter it to that chapter’s Practice assignment, so you can send students straight to the right slice.
- Worked notebooks and data: pre-executed, with a sample network dataset for the Chapter 8 exercises. Useful when a student’s own corpus is not ready in time.
- Getting started and the primers: the pages that absorb the first-week support load.
A note on the studies database
Its coverage is uneven by design stage and thins after 2023. Both facts are stated on the page, and both are usable in teaching: asking students why the communication lens has the fewest annotations is a better discussion prompt than any question with a settled answer. Suggestions for additions are welcome through the errata page.
Engaging the book as scholarship rather than teaching from it? See For researchers.