SRM is the data-science exam: generalized linear models, decision trees, time series, principal component analysis and clustering. The questions are more conceptual than on P or FM, more "which method fits here" than "compute this integral."
The SOA has published only 66 sample questions for it, which is the entire released set as of 2026 and far fewer than P or FM have. SRM has only been in its current form since 2018, so the SOA has had less time to build up released items. With so few, the set works best as a calibration check near the end of your prep, not as your main practice.
FreeFellow has all 66 free with worked solutions, next to an original SRM bank of 1,000-plus questions for the bulk of your practice. Both need a free account, though not a credit card, and you can preview the SRM samples without one.
What's in the set
| Topic area | Approx. sample count | What you'll see |
|---|---|---|
| Generalized linear models | ~20 | Logistic regression, Poisson regression, link functions, deviance, log-likelihood interpretation |
| Decision trees | ~12 | CART, splitting criteria (Gini, entropy), pruning, ensemble methods, bagging/boosting |
| Time series | ~12 | Stationarity tests, AR, MA, ARMA, model identification, forecasting |
| Principal component analysis | ~8 | Variance explained, loadings, biplot interpretation, application to dimension reduction |
| Cluster analysis | ~6 | K-means, hierarchical clustering, dendrogram interpretation |
| Model selection and validation | ~8 | Cross-validation, AIC/BIC, bias-variance tradeoff, train/test splitting |
GLM questions favor scenarios where you pick the link function from the structure of the response data (binary, count, continuous-positive). Time-series questions favor diagnosing stationarity and identifying the model from autocorrelation patterns. Tree questions favor comparing splitting criteria and ensemble logic.
When and how to use them
Read ISLR first. An Introduction to Statistical Learning (James, Witten, Hastie and Tibshirani) is the standard SRM textbook, free from its authors at statlearning.com, and the samples assume you've read chapters 3 through 10. If you haven't, you'll mostly find them frustrating.
Don't make the samples your main bank either. After one pass you'll start remembering answers, and they stop telling you much about where you are. The exam rewards understanding over pattern-matching, so more time in ISLR is a better use of your hours than re-drilling the same 66. SRM is also less calculation-heavy than P or FM, so raw volume matters less than it does on those exams.
Before you check a solution, write down why you chose the model you did. Then read the worked solution and grade yourself on the reasoning as well as the letter.
Spend extra time on GLM link functions. Of the 20 GLM items, roughly half hinge on picking the right link function (logit for binary, log for Poisson, identity for continuous-positive with constant variance), and getting this right earns you more points than the study time it takes.
Save all 66 for the last two weeks and work them timed at the exam's pace of about six minutes a question (the exam is 35 questions in 3.5 hours). Score 70 percent or higher and you're ready. Below 60 percent, schedule another week of topic review before you sit.
Three areas are thin in the set. Recent syllabus additions (2024+) like model interpretability (SHAP values, partial dependence plots) and fairness and bias metrics barely show up. The samples assume tidy data, while the exam can turn on spotting a data-quality problem. And newer ensemble methods (gradient boosting, random forests) get little coverage. The FreeFellow original bank is built to the current syllabus and covers the first and last of these better.
Where to get them
| Source | SRM Sample Questions | Format | Cost |
|---|---|---|---|
| FreeFellow | All 66, interactive | Same surface as topic practice | Free |
| ISLR (textbook) | None directly | Free PDF and printed book | Free |
| ASM SRM Manual | All 66 in study manual | Check current package price | |
| Coaching Actuaries Learn + Adapt SRM | All 66 + originals + video lessons | Interactive within Adapt | Check current package price |
| SOA only | All 66 | Static PDF on soa.org | Free |
Of the free resources for SRM, ISLR matters most, and it's written by the authors of the underlying statistical-learning material. The SOA samples and the FreeFellow original SRM bank are both free, so pair them with ISLR for the concepts. The Best Free Actuarial Exam Prep Resources (2026) roundup has the rest of the free options.