Psychometrics certification - how computational is the exam vs conceptual?
I have a master's in educational research and I work in assessment development, but my quantitative background is more applied statistics than formal psychometrics. I'm considering the PC certification and trying to figure out whether my existing skills are close enough or whether I need to seriously upskill in IRT and factor analysis before attempting the exam.
My comfort level with classical test theory is solid - I work with reliability coefficients, item analysis, and standard error of measurement regularly in my current role. But item response theory is where I start to feel shaky. I understand the conceptual logic of 1PL/2PL/3PL models but the math underneath them isn't something I could reproduce from scratch under exam conditions.
I've been studying about 2 hours a day for the past 6 weeks using Crocker and Algina, NCME resources, and online IRT tutorials. My self-assessment scores on IRT content are hovering around 62%, which worries me. CTT content I'm scoring around 84% consistently.
Is the exam heavily computational or more conceptual? If I can understand when and why to use different models and interpret the outputs intelligently, is that sufficient - or does the exam expect you to derive the mathematics from first principles?
Factor analysis is tested more than people expect going in. Confirmatory vs exploratory, fit indices, modification indices - know those distinctions cold. In my estimate it's somewhere around 15-20% of the exam content and it catches people who under-prepared that section.
What helped me was working through scenario questions: given these item parameters, what does this tell you about the item's behavior, and what would you do with it? That applied reasoning approach matters more than formula derivation on this exam.
The exam is primarily conceptual and applied rather than computational. You need to understand what IRT parameters mean and how to interpret fit statistics, but you're not doing hand calculations under time pressure. Your educational research background is closer to sufficient than you're giving yourself credit for.
84% on CTT with 62% on IRT after 6 weeks is a predictable split for someone in your position. Most people with a similar background spend the last 4-6 weeks almost entirely on IRT, DIF analysis, and validity frameworks. Your CTT knowledge doesn't need more attention at this point.