Psychological testing Study Guide

📖 Core Concepts Psychological testing – Administration of tests by trained evaluators; scores reflect differences in the construct being measured. Psychometrics – Scientific study of test construction, reliability, and validity. Latent variable – An unobserved construct (e.g., intelligence, anxiety) inferred from test items. Validity – Evidence that a test measures what it claims to measure. Reliability – Consistency of test scores across forms, administrations, or items (e.g., test‑retest, internal consistency). Standardization – Uniform procedures for test delivery, scoring, and interpretation across all examinees. Norms – Representative sample data that define high, low, and average scores for a specific population. Sample of behavior – The limited set of tasks/items used to represent a larger domain of behavior. 📌 Must Remember A test must be both valid and reliable to be useful. Standardization minimizes examiner subjectivity; objectivity reduces scoring bias. Discrimination: good tests separate extreme groups (e.g., patients vs. healthy controls). Norm‑referenced scores compare to a population; criterion‑referenced scores assess mastery of a defined content set. Reliability coefficients (e.g., \(r = .80\) or higher) indicate acceptable consistency. Cohen’s κ is used to assess inter‑rater reliability for direct observation. Projective tests have limited validity despite being used to explore unconscious material. 🔄 Key Processes Test Development Define construct → generate items → pilot → assess reliability & validity → standardize administration → norm on representative sample. Psychological Assessment Workflow Gather data (tests, inventories, collateral info) → integrate across sources → interpret using norms & validity evidence → formulate conclusions/recommendations. Scoring & Interpretation Follow objective scoring rules → compute raw scores → convert to standardized scores (e.g., T‑scores, percentiles) → compare to relevant norms (age, grade, etc.). 🔍 Key Comparisons Achievement vs. Aptitude Tests Achievement: measures learned knowledge; often criterion‑referenced. Aptitude: measures potential to learn or perform; can be specific (clerical) or general (intelligence). Norm‑referenced vs. Criterion‑referenced Norm‑referenced: “How does this person compare to peers?” (percentiles). Criterion‑referenced: “Has the person mastered the target content?” (pass/fail, mastery). Questionnaire/Interview Scales vs. Psychoeducational Tests Questionnaire/Interview: assess typical behavior or attitudes. Psychoeducational: assess maximum performance (e.g., IQ, achievement). ⚠️ Common Misunderstandings Reliability ≠ Validity – A test can be highly consistent yet measure the wrong construct. High scores always mean pathology – Clinical tests are normed; a score 1 SD above the mean may be typical, not diagnostic. Projective tests are definitive – Evidence shows limited validity; they should not be sole diagnostic tools. Norms are universal – Norms are population‑specific; using adult norms for children yields invalid conclusions. 🧠 Mental Models / Intuition “The Test as a Mirror” – Think of a test as a mirror that reflects the underlying construct only if the glass (validity) is clear and the frame (reliability) is sturdy. “Sampling the Ocean” – A small, well‑chosen sample of behavior can reliably estimate a huge domain, provided the sample is representative. 🚩 Exceptions & Edge Cases Direct observation may require Cohen’s κ ≥ 0.70 for acceptable inter‑rater reliability; lower values indicate poor agreement. Projective tests: despite low validity, may be used for exploratory purposes in research settings where no better instrument exists. Test security breaches: If items become public, norms and reliability degrade; tests must be withdrawn or revised. 📍 When to Use Which Diagnosing psychopathology → Clinical symptom scales (e.g., PHQ‑9, PCL‑5) with established cut‑offs. Career counseling → Interest inventories (e.g., Holland codes) and occupational depression inventories. Assessing learning disabilities → Achievement tests (norm‑referenced) combined with ability tests. Evaluating treatment change → Short screening scales (K6/K10) or weekly worry questionnaires for repeated measures. Measuring attitudes → Likert‑type attitude scales when a unidimensional favorability construct is needed. 👀 Patterns to Recognize “Norm‑referenced + Standardized score + Percentile” → Typical for large‑scale personality or clinical tests. “Criterion‑referenced + Pass/Fail” → Common in certification or mastery exams. “Multiple‑item self‑report + Likert 1–5” → Characteristic of attitude, affect (PANAS), and stress scales. “Observer‑report + Cohen’s κ” → Signals direct observation methodology. 🗂️ Exam Traps Choosing “projective test” for a high‑stakes diagnosis – tempting because it sounds “deep”; wrong because validity is limited. Confusing reliability coefficient with validity evidence – a high \(r\) does not guarantee the test measures the intended construct. Applying adult norms to a child’s score – yields misleading percentile; always verify age‑matched norms. Assuming a “high” raw score automatically indicates pathology – need to reference norm‑based cut‑offs (e.g., > 1 SD). Over‑relying on a single source of information – assessment requires integration of multiple instruments and collateral data.