Thank you for allowing me to join this community!

I want to conduct a small TSSEM using individual participant data from six studies. Since I have the raw data for each study, I am running into a problem with multiple measures of single constructs. Essentially, I have many relationships, but I don’t have many studies. Here is a simplified example of four constructs in one study, using different measures:

Predictor1 – Measure A
Predictor1 – Measure B
Predictor2 – Measure A
Predictor2 – Measure B
Predictor2 – Measure C

Outcome1 – Measure A
Outcome1 – Measure B
Outcome1 – Measure C
Outcome2 – Measure A

In order to get my pooled correlations for TSSEM Step One, I need to obtain effect sizes for each study relationship. In the example above, I have six Predictor1-Outcome1 effect sizes but I need a single Predictor1-Outcome1 effect sizes in order to pool my data across studies.

From what I understand in Cheung’s (2015, p. 181) textbook is that I need to average these effect sizes. Since I have so few studies, I am hoping there is an alternative to this? Each study contains alternative measures of the same construct so I feel like I am losing a lot of information by potentially averaging them together.

Thank you for reading this! I appreciate any guidance, additional readings, or suggested search terms! Thank you!

Nicole