hello, every researcher,
If I want to study the relationship between an iv and a dv, and the moderation effect of gender, will the analytic plan shown below be appropriate?
1. MANOVA and Post hoc Bonferroni contrast tests
2.Pearson's r correlation
3.PROCESS macro in SPSS
Hi Tingting,
I may be biased because of my love for SEM, but i would do it in SEM :D
I do not know MANOVA very much, hence, this could work too (may be other experts can commend).
Just calculating a correlation is a weak test of your model.
The set up would be (and this fits a SEM appraoch as well as a moderated regression model in SPSS): you specify a autoregressive structure with the DV_t2 as the dependent variable and DV_t1 as one of the independent variables. Now you have the option to test for lagged effects by incorporating the IV_t1, or to test for synchronous effects by incorporating the IV_t2. Both variables test the change of the DV from T1 to T2.
Gender is included as a control variable. To test for moderation, you include a product term Gender*IV - either in the T1 or T2 "version".
By using a full blown SEM with latent variables, you can specify a latent product term which has tremendous advantages compared to the manifest/SPSS variant as it increases the statistical power of the product term.
If you apply SEM, please first understand the meaning of the terms "latent variables", "common factor" model. These are used all the time and factor models specifically are misapplied in many cases.
I collected a bunch of essential papers around all important topics:
*Latent variables*
Hayduk, L. A., & Littvay, L. (2012). Should researchers use single indicators, best indicators, or multiple indicators. BMC Medical Research Methodology, 12(159), 1-17. doi:10.1186/1471-2288-12-159
Hayduk, L. A. (2014). Seeing perfectly fitting factor models that are causally misspecified: Understanding that close-fitting models can be worse. Educational and Psychological Measurement, 1-22. doi:10.1177/0013164414527449
*Latent autoregressive moderation model*
Dormann, C., & Zapf, D. (1999). Social support, social stressors at work, and depressive symptoms: Testing for main and moderating effects with structural equations in a three-wave longitudinal study. Journal of Applied Psychology, 84(6), 874-884.
*latent interaction modeling*
Little, T. D., Bovaird, J. A., & Widaman, K. F. (2006). On the merits of orthogonalizing powered and product terms: Implications for modeling interactions among latent variables. Structural Equation Modeling, 13(4), 497-519.
Steinmetz, H., Davidov, E., & Schmidt, P. (2011). Three approaches to estimate latent interaction effects: Intention and perceived behavioral control in the theory of planned behavior. Methodological Innovations Online, 6(1), 95-110.
*Causal models / basic SEM*
Elwert, F. (2013). Graphical causal models. In S. L. Morgan (Ed.), Handbook of causal analysis for social research. (pp. 245-273). Dordrecht Heidelberg New York London: Springer.
Pearl, J., & MacKenzie, D. (2018). The book of why. New York: Basic books.
Best,
Holger
MODERATOR: How do you know that gender is a moderator? At this stage, you do not know whether it is a moderator. You need to test for it to confirm whether the effect is moderation or is it a mediator.
MANOVA: Multivariate analysis of variance deals with multiple dependent variates, i.e. 2 or more Ys. In your question, since the singular noun is used, it appears you are dealing with a single Y and a single X and gender is another potential intervening factor. If that is the case, MANOVA would not be applicable.
PEARSON CORRELATION: This is relevant only with quantitative Y and quantitative X case. Do you gave quantitative data for both X and Y?
What is/are you Ys? How many Y do you have? What kind of data?
What is/are your Xs? How many X do you have? What kind of data?
LINKS:
https://en.wikipedia.org/wiki/Multivariate_analysis_of_variance
Dear Paul Louangrath , if you allow me to chime in: She seems to have some theoretical assumptions of gender being a moderator. Hence, she plans to test this hypothesis with her design. (Scientific) business as usual.
Gender cannot be a mediator, thus, this option can be excluded beforehand.
And if it was another variable, it could be a mediator as well as a moderator (of the direct effect) but again - this would be based on her assumption of its causal role.
Best,
Holger
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS