SOA · Exam hub
← All exams Exam PA Predictive Analytics
✓ I passed this exam. Everything below is from the inside.
My honest verdict
“Unlike anything before it: a business communication exam wearing a data-science costume. The modeling is SRM material. The challenge is explaining decisions in writing, clearly, under time pressure, to a non-technical audience.”
How I’d study for PA today
Stop thinking “statistics exam.” Start thinking “consulting deliverable with a stopwatch.”
- Read model solutions first. Before writing anything, read three past projects and their published solutions. Notice how much of the credit is for justifying choices in plain English: why this variable, why this link function, why this validation split.
- Write, don’t just model. Practice producing the written sections under time. Most failing candidates ran out of writing time, not modeling ideas.
- Build a personal template. A reusable structure for data-cleaning notes, model comparison, and recommendations saves 30+ minutes on exam day.
The trap
Over-modeling. PA rewards a sensible model explained excellently over a fancy model explained poorly. When in doubt, simplify and explain.
Studying for PA? Weekly strategy in the newsletter
Exam tips, scholarship deadlines, and honest career advice. One email a week. No noise.