Confirming your purchase

Why these six assessments — the research

Each assessment in the suite is here because the personnel-selection literature supports the test type. The research justifies choosing these test types; TestMyself’s own scoring bands remain provisional while population norms are gathered.

Intelligence — “Can they think?”
Why: Reasoning under time pressure feeds every other work skill.
Evidence: Schmidt & Hunter’s 85-year meta-analysis ranked general mental ability the top predictor of job performance (≈ .51) and training success; a 2022 re-analysis argues ≈ .31 under stricter corrections — the ranking holds either way.
Sources: Schmidt & Hunter (1998), Psychological Bulletin · Sackett et al. (2022), Journal of Applied Psychology

General Knowledge — “Do they understand the world?”
Why: Accumulated knowledge is the substrate new problems get solved with, and it compounds — the more you know, the faster you learn the next thing.
Evidence: Gottfredson’s work on cognitive ability in everyday life ties knowledge and reasoning to handling information complexity; job-knowledge measures sit near the top of Schmidt & Hunter’s predictor hierarchy.
Sources: Gottfredson (1997), Intelligence · Schmidt & Hunter (1998), Psychological Bulletin

Cognitive Flexibility & Control — “How fast do they learn?”
Why: Dropping a stale rule, holding a new one in mind, and switching cleanly is what “fast learner” means in observable behavior.
Evidence: Miyake et al. (2000) showed shifting, working-memory updating, and inhibition are separable, measurable executive functions — which is why this pillar uses timed performance tasks for exactly those three, not self-description. Pulakos et al. built the standard taxonomy of adaptive performance and found ability/adaptability measures predict supervisor ratings of it.
Sources: Miyake et al. (2000), Cognitive Psychology · Pulakos et al. (2000), Journal of Applied Psychology

AI & Tool Adaptability — “Can they handle what’s next?”
Why: The differentiator with unfamiliar and AI tools is judgment — fluency plus knowing when the output is wrong.
Evidence: In the HBS/BCG field experiment (758 consultants, GPT-4), AI users finished 12.2% more tasks, 25.1% faster, at 40%+ higher quality — but were 19 points less likely to be right on a task outside the AI’s frontier. Both halves are one measurable skill.
Sources: Dell’Acqua et al. (2023), HBS Working Paper 24-013 · SSRN

Moral Compass — “Can they be trusted?”
Why: The costliest hiring failures are character failures, and everyone interviews honest.
Evidence: Ones, Viswesvaran & Schmidt (1993; 665 coefficients, 500k+ people) found integrity measures predict job performance (≈ .41) and, distinctively, counterproductive behavior. Van Iddekinge et al. (2012) reported smaller estimates (≈ .15 performance / .32 counterproductive) — still positive, which is why this pillar reports a stated-values profile from forced-choice dilemmas, not a pass/fail honesty score.
Sources: Ones, Viswesvaran & Schmidt (1993), Journal of Applied Psychology · Van Iddekinge et al. (2012), Journal of Applied Psychology

Determination — “Will they finish?”
Why: Follow-through shows in behavior before it shows in references.
Evidence: Duckworth et al. (2007) introduced grit as a predictor of retention and achievement beyond talent; Credé, Tynan & Harms (2017) sharpened it — the active ingredient is perseverance of effort specifically — so this pillar measures sustained effort directly (including a quittable persistence task) instead of trusting a questionnaire.
Sources: Duckworth et al. (2007), J. of Personality and Social Psychology · Credé, Tynan & Harms (2017), JPSP · Dugan et al. (2019), J. of Personal Selling & Sales Management

What we store when you take an assessment

Your answers, per-question timings, device class, and tab-switch/away signals — tied to a one-way hash of your purchase session, never to your name or email (we don’t collect those).

Why: to score your assessment server-side (answer keys never leave our server), to enforce one scored attempt per purchase, and to let employers verify your result. That measurement is the product: it’s what makes a result credible instead of self-reported. What we don’t do: no ad-tech, no third-party trackers, no pixels, no data brokers. Nothing is sold or shared with advertisers.

Anonymous calibration data

When you submit an assessment, we record anonymous per-question right/wrong results, response timing, and device type. We use this to improve question difficulty, reliability, and fairness for future test-takers — calibrated norms are what let an employer trust a score. We do not collect your name, email, or identity for calibration, and we do not use advertising trackers or third-party scripts.