At what sample size do latent variable correlations stabilize?

By André Kretzschmar, Gilles E. Gignac. Journal of Research in Personality, 2018

95 citations recorded by the source corpus.

We conducted a Monte-Carlo simulation within a latent variable framework by varying the following characteristics: population correlation (ρ = .10, .20, .30, .40, .50, .60, .70, .80, .90, and 1.00) and composite score reliability (coefficient omega: ω = .40, .50, .60, .70, .80, and .90). The sample sizes required to estimate stable measurement-error-free correlations were found to approach N = 490 for typical research scenarios (population correlation ρ = .20; composite score reliability ω = .70) and as high as N = 1,000+ for data associated with lower, but still sometimes observed, reliabilities (ω = .40 to .50). We encourage researchers to take into consideration reliability, when evaluating the sample sizes required to produce stable measurement-error-free correlations.

Subfields: Experimental and Cognitive Psychology, General Decision Sciences, Management Science and Operations Research

View Paper: https://www.semanticscholar.org/paper/7e4cedd06bd3479e7657b6b3d0f9cd21e3bfd311