Psychological Assessment and Psychometrics
What this domain tests
This domain tests whether you can choose, administer, and interpret tests defensibly: reliability, validity, norms, base rates, and multi-method integration. It rewards candidates who understand what a score means and what it does not mean.
Core rules
- Reliability is consistency; validity is the accuracy of the inference you draw. A test can be highly reliable and still invalid for your purpose.
- Standard scores require the correct reference group: age, language, education, and country. Foreign norms on a translated test are a validity threat.
- Sensitivity and specificity belong to the instrument; positive and negative predictive value belong to the patient and shift with prevalence.
- Use multiple methods: clinical interview, collateral, records, self-report, and performance measures. Include feigning or symptom-validity measures when incentives exist.
- Adaptation of a test needs back-translation, piloting, local norms, and published evidence for the new population.
Worked example
A screen has sensitivity 0.90 and specificity 0.90, and prevalence in your clinic is 10%. In 1,000 patients: 100 true cases, of whom 90 test positive; 900 non-cases, of whom 90 test positive falsely. PPV = 90/180 = 50%. So a positive screen is a coin flip and must be confirmed by structured interview. Reasoning: predictive value depends on base rate, not on the test alone.
Traps that cost marks
Reading a percentile as percentage correct; confusing standard error of measurement with standard deviation; choosing a test because it is familiar; interpreting one elevated scale as a diagnosis; and ignoring language, literacy, and sensory factors during administration.