I85PREP
⌘ K
EPPP Part 2 Skills study materials
EPPP Part 2 Skills study resources

EPPP Part 2 Skills: Assessment and Diagnosis

EPPP Part 2 Skills · Work through your exam, one topic at a time.

All guide sections

Assessment and Diagnosis

Assessment stems test whether you can choose an instrument, interpret a score correctly, and reach a defensible diagnosis. The most common failure is an interpretation that sounds plausible but ignores psychometric base rates or the limits of the standardization sample.

Reliability and validity

  • Reliability types: test-retest (stability over time), internal consistency (coefficient alpha), inter-rater (agreement between judges). Reliability is necessary but not sufficient for validity.
  • Validity: content (coverage of the construct), criterion (concurrent and predictive), construct (convergent and discriminant).
  • The standard error of measurement quantifies the band around an observed score; report confidence intervals, not single numbers, as the true score.

Interpreting scores

  • Wechsler FSIQ and index scores: mean 100, SD 15; subtests: mean 10, SD 3. T scores: mean 50, SD 10. z scores: mean 0, SD 1.
  • Sensitivity is the proportion of true cases detected; specificity is the proportion of non-cases correctly cleared. Positive predictive value depends on prevalence.

Diagnostic reasoning

  • DSM-5-TR requires ruling out substance and medical causes, establishing impairment, and considering cultural explanations of the presentation.
  • MMPI-2 validity scales: L (naive virtue), F (infrequency/overreporting), K (defensiveness), plus VRIN and TRIN for inconsistency.
  • Treat multiscale elevations and code types as hypotheses to corroborate, not as diagnoses.

Worked example

A screener has sensitivity .90 and specificity .90. In a clinic where 1% of referrals have the disorder, 1,000 referrals include 10 cases: 9 are true positives, and among the 990 non-cases 99 are false positives. Positive predictive value is 9/108, about 8%. Nine of ten positive screens are false. Reasoning: strong sensitivity and specificity do not guarantee a useful positive result when prevalence is low, so confirmatory assessment is required. This is why screening instruments rule out rather than diagnose.

Traps

  • Equating a high reliability coefficient with validity.
  • Reporting an IQ or T score as a fixed point without a confidence interval.
  • Using an instrument outside its normative age range and interpreting the resulting floor effect.
  • Reading a personality profile as a diagnosis without behavioral corroboration.