Scales with high Cronbach-α may still conceal several distinct latent factors.
Evidence
First, following Whitehead and Perry (2015), most sociologists assume that CN is not a latent but a visible factor. And they construct an additive 24-point scale combing all six five-point CN items together, treating these items as continuous indicators. The scale often comes with a high Cronbach-α score. But political scientist Davis (2022) questions the rationale of this scale’s construction. Quoting Gardner (1995), Davis contends that scales with high Cronbach-α scores may also measure several latent factors. — Ruiqian Li and Paul Froese, The Duality of American Christian…, 84
Categorization
- XREF: Speaks directly to the reader’s stated purpose: testing whether Christian nationalism is a distinct schema or collinear proxy — the Cronbach-α critique parallels psychometric concerns about measures that correlate strongly yet sever cleanly under factor analysis. — connects to: Long scales can show acceptable reliability while measuring many unrelated things; Unbalanced scales inflate both alphas and empirical correlations.; Long scales achieve acceptable reliability without valid unidimensional measurement (+3 more)
Broader Connections
- A high Cronbach’s alpha does not prove a scale is unidimensional.
#expands(0.95) - Long scales can show acceptable reliability while measuring many unrelated things
#expands(0.93) - Long scales achieve acceptable reliability without valid unidimensional measurement
#expands(0.90) - Symmetric two-sided scales are needed when processing may improve a signal
#expands(0.82) — David M. Howard, Jamie Angus, Acoustics and Psychoacoustics, 2395 - Unbalanced scales can show spurious significance without true construct validity
#expands(0.76) - Unbalanced scales inflate both alphas and empirical correlations.
#complicates(0.71)