9.3 Assessment Method Selection and Measurement Planning

Selecting assessment methods for defensible decisions

Assessment on Part 2 is an applied skill. The psychologist must decide what information is needed, how to obtain it, how reliable it is, and how to explain its limits. Because assessment and intervention is 33% of EPPP Part 2-Skills, the largest single domain, vignettes routinely require candidates to choose the next assessment step before treatment or diagnosis can be justified.

The referral question controls the method. A request for diagnostic clarification may require a clinical interview, symptom measures, records, and differential diagnosis. A request about cognitive functioning may require standardized testing with attention to norms, effort, language, sensory needs, and educational background. A request about workplace risk may require records, collateral data, structured risk factors, and careful limits on conclusions. The best answer is method-to-question fit.

Validity, reliability, and standard error

Validity is not a label a test owns forever; it belongs to an interpretation for a purpose. A depression inventory may be useful for screening and progress monitoring yet inadequate to answer a complex disability question. A cognitive test may have excellent psychometrics, yet interpretation is limited when the client was tested in a non-dominant language, without needed accommodations, or under conditions that invalidate standardized administration. Reliability sets a ceiling on validity, and the standard error of measurement reminds candidates that an obtained score is a band, not a point.

The exam rewards reporting confidence intervals and avoiding overinterpretation of small differences between subtest or index scores.

Part 2 frequently tests incremental data gathering. When a vignette contains inconsistent self-report, missing records, unclear medication effects, or a major cultural or language issue, the next best step is usually to gather more information rather than finalize a diagnosis. Collateral information can be valuable, but it must be obtained with proper authorization unless another legal or ethical basis applies. The exam favors respectful, relevant, and role-appropriate data collection over a sweeping battery.

Measurement planning and proportionate assessment

Measurement planning continues after the first assessment. If a psychologist begins treatment, baseline measures establish a starting point against which improvement is judged. Repeated measures can reveal change, nonresponse, deterioration, or emerging risk. Functional outcomes such as school attendance, sleep, work performance, parenting behavior, or social participation often matter as much as symptom scores. A plan without a monitoring strategy is difficult to defend.

Assessment selection checklist:

• State the decision the assessment must support.
• Choose methods with validity evidence for that decision.
• Check language, culture, disability, education, age, and setting factors.
• Use multiple data sources when the decision is high stakes or data conflict.
• Interpret results within the limits of administration conditions and available norms.
• Link findings to diagnosis, risk, recommendations, and treatment planning.

Part 2 answers must avoid both underassessment and overassessment. Underassessment appears when a psychologist concludes from too little data. Overassessment appears when the psychologist administers a broad battery unrelated to the question, delays urgent care, or creates unnecessary burden and cost. The better response is proportionate: enough valid information to answer the question safely and ethically. Sequencing matters too, since assessment results often determine whether further, more specialized testing is even warranted, and an iterative approach prevents both premature closure and wasteful over-testing.

Candidates should also keep matching theory in mind, because the exam often hides it inside scenarios. A criterion-referenced measure compares performance to a standard, useful for mastery and diagnostic cutoffs, while a norm-referenced measure compares a client to a reference sample, useful for relative standing such as intelligence or achievement percentiles. Sensitivity and specificity govern screening: a highly sensitive screen rarely misses true cases but produces false positives, so a positive screen warrants further evaluation rather than a diagnosis.

The positive predictive value of any screen falls as the base rate of the condition drops, which is why screening a low-prevalence population yields many false alarms. Test bias, item bias, and the appropriateness of the normative sample for the client's demographic group are recurring fairness concerns the exam expects candidates to flag.

Documentation is part of the skill. A well-formed report identifies the referral question, the methods used, relevant limitations, the findings, the reasoning that links data to conclusions, and the recommendations. It does not overstate certainty, reports scores with their confidence bands, and explains when further evaluation, consultation, referral, or monitoring is needed. That disciplined, transparent reasoning is exactly what Part 2 assessment vignettes are designed to test.