I have two set of questionnaires (set A and B). Set A represent the relationships between independent and dependent variables while Set B denotes for moderating variables. Can I run hierarchical regressions in SPSS with two types of respondents?
Mohammad,
if set B consists of measures taken from different persons than the set A measures, this is not possible. Moderator variables are predictor variables, too, and they should have a relationship with the dependent variable. This is only possible if these measures are taken from the same persons.
Does this help?
Karin
Thanks Karin. I need a more clearer picture. My Independent variables towards dependent variables (IV to DV) answers are given by NGO clients (regarding NGO programmes towards clients' socio-economic development). This is questionnaire A (for NGO clients). I want to see the moderating impact of individual values of NGO managers (questionnaire B) which answer will be given by NGO managers. So, as you advise it will not work. So, pls give me advice as I want to see how values impact on this relationship. I don't have measures which will cover all these..IV, DV, and MV.
Aktar
Aktar, if you were able to relate managers to clients (say, 1 manager, 5 clients; another manager, 10 clients, etc.) you could perform a multilevel analysis with clients on level 1 and managers on level 2, or you could perform a multiple (or moderated) regression and use dummy variables in order to relate managers to clients. In moderated regression, you have to include the product of the predictor and the moderator variable as a third "predictor variable", and therefore it is necessary to relate each client to a manager, otherwise you would not be able to build the product.
HTH
Karin
The first thing to clarify is what you mean by "hierarchical regressions." When Karin mentions multi-level analysis, this is also known as hierarchical linear modeling. That is a different kind of analysis than looking for "moderating" effects, which would imply that the relationship between an IV and a DV differed, according to the types of organizations in which they occurred.
If your interest is indeed in moderating effects, then you do not have to get into the complexity of multi-level analysis. Instead, you could use the responses from the managers to create variables that characterize each organization, and then use the SPSS command Merge Files to add those organizational variables onto the data from the clients.
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis