Hi,
I am analyzing a study and I would like to see if individual differences interact with situation factors (i.e. individual and situation = IV) regarding their impact on another factor (=DV).
IV1: Individual differences are numeric / interval variables (e.g. attitudes - score from 1 to 7).
IV2: Situation are nominal (e.g. Accepted vs Rejected) variables.
DV (e.g. meta-perception) is interval (score from 1 to 7).
And I would like to know if Attitudes and Acceptance (vs rejection) interact to predict meta-perception level.
So something like: Meta-perception ~ Attitudes * Acceptance.
This would normally be a regression but the situation variables are either fully repeated measures (meaning participants did all the conditions) or mixed (participants did 2 conditions on 4).
Do you know which test this could be? And do you have any resources to recommend about it that present a similar example? If you know resources using R it will be even better !
Thanks for your help,
Is your meta-perceptions variable represented by a single item, or is it a multi-item set that could be combined into a scale?
It sounds like this could be addressed with a multilevel (random coefficient) regression model. Since you have a repeated measures design, you have observations (time points, Level 1) nested within persons (Level 2). In a multilevel regression, your DV would be regressed on a situation dummy (0; 1) variable at Level 1. Your Level 1 intercept and slope could be specified as random coefficients (variables) that are themselves regressed on your person-level continuous IV (attitude) at Level 2. When you combine the Level 1 and Level 2 equations, you automatically obtain a product term dummy*attitude that represents the interaction effect that you are interested in (referred to as a "cross-level interaction"), similar to what you have in regular moderated regression analysis.
I believe several R packages are available for multilevel analysis.
The meta-perception measure is a 1-item scale.
And I will look into the multilevel regression model, it sounds interesting.
Thanks
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Statistical Software
- R