Dear experts,
becaus response rate of my survey was lower than expected I now put special attention to power analysis. I already learned that G*Power is the tool of choice for this task and I read the papers Faul et al 2007 and Faul et al 2009 about usage of G*Power 3. But, because I am new to smartPLS too, I am not sure if I transferred the right data from smartPLS to G*Power3 (v 3.1.5):
I have structural model with 3 exogenous latent variables pointing towards one endogenous latent variable. All latent variables have 2 to 4 reflective indicators.
When i use the G*Power calculator "Linear multiple regression: Fixed model, R² deviation from zero" to "compute implied alpha & power" I can click on "Determine =>" to caculate the input parameter effect size "From predictor correlations". G*Power needs therefore the number of predictors (in my case three) and I have to enter 2 matrices (1. Corr between predictors and outcome, 2. Corr between predictors)
Is it correct to use the numbers from "Latent Variable Correlations" from "PLS Algorithm" -> "Default Report" as input for the matrices in G*Power 3?
Correlations between endogenous variable and the exogenous variables for matrix 1 and correlations among exogenous variables for matrix 2.
Tilman