Dear everyone,
Could you please tell me if my sample size calculation is correct?
Following the practical recommendation on sample size for structural equation modeling made by Kenny (2015), a 5 to 1 ratio of sample size to the number of free parameters is suggested (Bentler & Chou, 1987). The total number of elements in the initial covariance matrix is k(k+1)/2, where k equals the number of observed variables in the matrix. In this case, the total number of parameters = (27*28)/2 = 378. The parameters that have to be estimated in the model contains 27 observed variable variances, 20 loadings, 7 latent variable variances, 14 regression paths, and 7 latent variable covariances, which results in 75 free parameters. Note that 378-75 = 203 degrees of freedom. Given that we compare two groups, the degrees of freedom is 606 and the minimum sample size equals to 5 (ratio)*75 (number of free parameters)*2 (number of groups) = 750.
http://davidakenny.net/cm/fit.htm
R code:
modelSDO
Hello Robin,
As well, you can try this web-based calculator for SEM models: https://www.danielsoper.com/statcalc/calculator.aspx?id=89
Good luck with your work.
Dear David,
Thank you for this suggestion. I tested it and based on the anticipated effect size, the recommended minimum sample size ranges from 170 (anticipated effect of .3) to 783 (anticipated effect of .15).
- Badges
- Science topic
- Similar topics
- Philosophy
- Moral Philosophy
- Free Will