DI‐II.
I found four studies to assess a range of measurement properties of the BDI‐II in general population adults, without comorbid conditions. There was weak evidence in support of internal consistency—many studies did not calculate Cronbach's alpha for each subscale separately. However, all studies showed support for the internal consistency of the BDI‐II total score with acceptable alphas above (.7). There was weak evidence in support of test‐retest reliability with one fair study (as it was unclear how missing items were handled), with a high alpha (.89). There was strong evidence for content validity in one methodologically excellent study of the BDI‐II in a non-English speaking Kenyan sample. There was moderate evidence in support of structural validity from the two studies. Both studies showed fair evidence for a single factor solution. Evidence for hypothesis testing was moderate—the BDI‐II showed acceptable correlations with other depression measures (r > .57). There was weak evidence for cross-cultural validity as there were weaknesses in the quality of the translations (only one forward/backwards translation), or failure to pretest the items in a sample for interpretability and cultural relevance. There was moderate evidence for criterion validity. The BDI‐II showed adequate sensitivity (>.7) and specificity (>.8) in determining Major Depressive Episodes with clinician ratings used as the criterion.
PHQ‐9. Six studies were found that explored a range of measurement properties of the PHQ‐9 in general population adults. There was moderate evidence in support of internal consistency with adequate Cronbach's alphas (>.7) for the unidimensional measure (confirmed using IRT methods and factor analytic methods). There was moderate evidence in support of test‐retest reliability with correlations >.7. There was moderate evidence for structural validity showing consistent evidence for a one-factor solution (using factor analysis). There was moderate evidence for hypothesis testing; the PHQ‐9 correlated strongly with other measures of similar constructs (e.g., the BDI), and support was found for consistent factor structure across time points and subgroups. There was moderate evidence for criterion validity, with acceptable sensitivity and specificity (>.79) in detecting clinical diagnosis of depressive disorder.
Comments
Post a Comment
Your Thoughts?