Validity and Reliability in Research

Cronbach's α > 0.70 is generally acceptable; α > 0.80 is good; α > 0.90 may indicate redundancy (too many similar items). α < 0.60 suggests poor internal consistency. However, α depends on the number of items and their correlations; consult field-specific standards. Always report α for your sample, not just for the original validation sample.

No. A highly controlled lab experiment may have excellent internal validity (strong evidence for causality) but poor external validity (findings may not apply to real-world populations or settings). Conversely, observational field studies may have high external validity but lower internal validity. Ideally, you want both; in practice, trade-offs occur.

Yes, but terminology differs. Qualitative research attains credibility (findings ring true), dependability (methods are consistent and transparent), transferability (findings may apply elsewhere), and confirmability (findings are grounded in data). These parallel quantitative validity and reliability. Rigorous qualitative research meets these criteria through careful methods, documentation, and reflexivity.

If the instrument is published but evidence is sparse, conduct a pilot study (n=30–50) to assess reliability in your population. Report limitations in the validity evidence for your sample. If the instrument is unpublished, treat it as newly developed and provide pilot validation data. Always acknowledge any gaps in evidence and their implications for interpreting results.