HI Everyone, I'm conducting Actor-Partner Interdependence Model (APIM) analyses using AMOS21 (see http://tinyurl.com/k4du3pb). The analyses run without difficulty. However, I am wondering how to correctly interpret the correlation between the residuals at T2.
Here is what Laursen et al (2008) say: it "holds special significance because it describes increases or decreases in similarity from the first to the second measurement point. Put simply, c2 captures changes in dyadic similarity” (pp. 21).
However, here is what Popp et al (2008) say: “It is important to note that Time 2 residual similarity (c2) and Time 3 residual similarity (c3) do not assess changes in dyadic similarity over time." (pp. 1029-1030)
These seem to be saying completely contradictory things to me. If the former is correct, then I need to comment on it in the manuscript I'm preparing (and it would be a really interesting addition to the results). If it is not, then clearly I do not. My feeling is that Popp et al are correct, but I may be mistaken.
Many thanks in advance for your help.
Simon
References:
Laursen, B., Popp, D., Burk, W. J., Kerr, M., Stattin, H., & Card, N. A. (2008). Incorporating interdependence into developmental research: Examples from the study of homophily and homogeneity. In N. A. Card, J. P. Selig, & T. Little (Eds.), Modeling dyadic and interdependent data in the developmental and behavioral sciences (pp. 11-37). New York: Routledge.
Popp, D., Laursen, B., Kerr, M., Stattin, H., & Burk, W. K. (2008). Modeling homophily over time with an actor-partner interdependence model. Developmental Psychology, 44(4), 1028-1039.
Hi Simon, the variables at T1 cannot predict the variables at T2 to 100%. Thus, there is a rest, which is captured by the residuals. The residual at T2 from husband might covary with the residual from the wife, but I cannot see a meaningful interpretation of that, because it is not the covary between the variables itself, but of the rest. Thus, Popp is right.
Best, Peter
Thanks Peter. That was my understanding too. I contacted the first author, Brett Laurseb, about this and here's what he said too:
"Thanks for your interest in my research. You are correct in that the T1 and T2 correlations are not strictly comparable (although it is not always the case that the latter is lower than the former). For this reason, there are better ways to look at changes in similarity. I have a recent paper that uses a fairly simple correlational approach to the problem: http://scholar.google.com/citations?view_op=view_citation&hl=en&user=yVoiZ1kAAAAJ&pagesize=100&citation_for_view=yVoiZ1kAAAAJ:M05iB0D1s5AC
I hope this is helpful. Let me know if you have any additional questions."
The paper Brett cites is useful, I'd recommend it!
Simon
- Badges
- Science topic
- Similar topics
- Psychology
- Developmental Psychology