Test-retest, Does the instrument yield consistent results over time?, For constructs assumed to stay constant (e.g., IQ, stable personality traits)., Subject to memory effects (practice bias) or maturation (actual change in the trait)., Pearson correlation (r) - consistency over time, r ≥ 0.80 (Stability), Parallel-forms, Do two equivalent versions of a test produce similar scores?, To prevent "testing threats" (cheating/memory) in pre-test and post-test designs., Extremely difficult and expensive to create two truly equal forms., Pearson correlation (r) between Form A and Form B., r ≥ 0.80 (Equivalence), Inter-rater, Do different observers or raters agree on their assessments?, Observational research, subjective grading, or medical diagnoses., Human judgment is inherently subjective; requires intensive training and clear rubrics., Cohen’s Kappa (κ), k > 0.60 (Agreement), Internal Consistency, Do all items in the scale measure the same underlying construct?, Multi-item surveys or Likert scales that measure a single construct. , High values may indicate redundancy rather than quality; doesn't prove unidimensionality., Cronbach’s Alpha (α), a ≥ .70 (internally consistent).

Group sort

автор:
Друкувати
Вбудувати

Список переможців

Візуальний стиль

Параметри

Обрати інший шаблон

Відновити автоматично збережене: ?