#psychometrics

Establish psychometric properties of research instruments. Reliability assessment: 1. Internal consistency: Cronbach's α > 0.70 for research, > 0.90 for clinical decisions. 2. Test-retest: correlation between administrations 2-4 weeks apart (r > 0.80). 3. Inter-rater reliability: agreement between observers (ICC > 0.75, κ > 0.60). 4. Split-half: correlation between odd/even items, Spearman-Brown correction. Validity assessment: 1. Face validity: instrument appears to measure what it claims. 2. Content validity: expert panel review of item relevance (I-CVI > 0.78). 3. Construct validity: factor analysis confirms hypothesized structure. 4. Criterion validity: concurrent (correlates with gold standard) and predictive (predicts future outcomes). Advanced techniques: 1. Item Response Theory (IRT) for item-level analysis. 2. Generalizability theory for multiple sources of error. 3. Structural equation modeling for latent constructs. Report all reliability and validity evidence in methods section.

Create reliable and valid survey instruments. Design process: 1. Literature review to identify existing validated scales. 2. Define constructs clearly, create item pool (3-5 items per construct). 3. Expert review panel (5-7 subject matter experts) for content validity. 4. Pilot testing with 30-50 participants for clarity and comprehension. 5. Main validation study (minimum 10 participants per item, 200+ total). Analysis: 1. Exploratory Factor Analysis (EFA) to identify factor structure. 2. Confirmatory Factor Analysis (CFA) to test model fit (CFI > 0.95, RMSEA < 0.08). 3. Internal consistency reliability (Cronbach's α > 0.70). 4. Test-retest reliability over 2-week period (r > 0.80). 5. Discriminant and convergent validity testing. Use software: R lavaan, SPSS, or Mplus.