Assessment and Diagnosis
Reliability and validity
Reliability is consistency (test–retest, interrater, internal consistency/split-half). Validity is whether the test measures what it claims (content, criterion/predictive, construct, face). Reliability is necessary but not sufficient for validity-a scale can be consistently wrong.
Score metrics
IQ: mean 100, SD 15. Subtests: mean 10, SD 3. T score: mean 50, SD 10. z score: mean 0, SD 1. A confidence interval equals the obtained score ± (z × SEM); the 95% interval uses 1.96 × SEM. Low reliability inflates SEM and widens the band.
Diagnosis
DSM-5-TR dropped the multiaxial system. Know operational anchors for frequently tested disorders (duration, functional impairment, rule-outs), and remember that a diagnosis requires clinically significant distress or impairment and exclusion of substance and medical causes. Sensitivity identifies true positives; specificity identifies true negatives. Positive predictive value rises with base rate-a "good" test used in a low-prevalence population still yields many false positives.
Inventories
MMPI validity scales (L, F, K and related) index response style, not psychopathology. Projective and objective measures answer different questions; know which is which.
Culture
Assess acculturation, language proficiency, and the population on which norms were built before calling a difference a deficit. Test bias is an empirical question, not an assumption.
Worked example
A client scores 85 with an SEM of 5. The 95% CI is 85 ± 1.96(5) ≈ 75–95. Report the interval rather than treating the point estimate as exact.