Common method bias assessment
Examine distortions caused by shared measurement procedures
Common method bias occurs when measurement procedures distort relationships among measured constructs. Shared respondents, item wording, context and response styles can create or suppress associations. Harman's single-factor test examines whether an unrotated factor solution is dominated by one factor, but it is an insensitive diagnostic and cannot establish absence of bias. Prevention and theory-informed modelling are more defensible than a single pass/fail test.
Use when important predictor and outcome measures share a source, occasion or response format and plausible method effects could affect the relationship being interpreted.
Strengths
- Encourages better measurement design before data collection
- Can separate substantive and method-related explanations
Limitations
- Statistical method factors require strong identification assumptions
- No single diagnostic rules out all possible method effects
Know the boundary
A first factor below a conventional percentage does not demonstrate that common method bias is absent.