Hello,
I am using structural equation modeling (SEM) to analyze the effect of differents variables of social support on work-family interface. I would like to analyze if the coefficients change according to two unequal groups (G1 = 237, G2 = 561).
Please let me know which method you have used/recommend.
Thanks in advance
Within the SEM tradition, the most rigorous way to get at group differences is to test for factorial invariance using multi group CFA. Because your group sizes are quite unequal, you might also consider the MIMIC model approach. Brown discusses both these methods in his text: http://www.guilford.com/books/Confirmatory-Factor-Analysis-for-Applied-Research/Timothy-Brown/9781462515363.
Dear Felix
The analysis procedure depends on the type of your data ( scale, ordinal .... ) and the distribution of the data compared with the normal distribution. In the case of normal distribution and scale data, you can use T test, and in the case of the non-normal distribution or ordinal data you can use Mann-Whitney test.
Good Luck
I would go for a multi group SEM first, since MIMIC approaches make the assumption of strict measurement invariance (equal loadings, intercepts/thresholds and error variances). Test if these assumptions are tenable for your model. Probalby you find only partial invariance, stay within the MG framework. In addition, I prefer to test the equality of path coefficients using Chi-Square Difference testing or Wald-Tests.
HTH
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis