Multiple groups in one sample
Posted: Mon Apr 04, 2022 2:35 pm
Dear all
I have a methodological question regarding the mixing of samples. In this study somewhat over a hundred students were put into four groups and each group saw the same interaction with a different type of robot. All students then filled out the same questionnaire. The questions were all based on previous research and the questions came from a structural equation model. The original design was for group comparisons using ANOVAs (between-group).
I was now asked whether I could check whether the structural equation model holds up, as that would round off the publication well. For the sample size to even remotely be appropriate, I have to put all four groups into the main sample and run the structural equation model like that.
Am I allowed to do this?
Because the model is built for robots in general and the students all filled out the same questionnaires, I feel that it should be okay. If anything, I have the feeling it would make the finding more generalizable.
Would really appreciate your help.
Kind regards
David
I have a methodological question regarding the mixing of samples. In this study somewhat over a hundred students were put into four groups and each group saw the same interaction with a different type of robot. All students then filled out the same questionnaire. The questions were all based on previous research and the questions came from a structural equation model. The original design was for group comparisons using ANOVAs (between-group).
I was now asked whether I could check whether the structural equation model holds up, as that would round off the publication well. For the sample size to even remotely be appropriate, I have to put all four groups into the main sample and run the structural equation model like that.
Am I allowed to do this?
Because the model is built for robots in general and the students all filled out the same questionnaires, I feel that it should be okay. If anything, I have the feeling it would make the finding more generalizable.
Would really appreciate your help.
Kind regards
David