Dear all,

I hope I am in the correct forum and you can help me.

At the moment I pursue a survey study. In the survey I obtain items that shall describe one variable. E.g., trust is evaluated using several items. Each item is evaluated using a Likert scale, which is the same for all questions. The answers to the questions shall be used to obtain the construct trust, which is further used in regression analysis. My question: May I obtain the variable trust by taking the mean value of the items? Under which circumstances shall I use factor analysis and under which circumstances may I use mean values?

I would be very happy if you could help me further and also suggest some literature that points out the implications of the different approaches.
Regarding the regression analysis: I am planning to to probit/logit analyses. Does this have an influence on the selection of mean values (or factor analysis) for computing the construct?

Do I also have to consider anything special when I use the variable trust as dependent variable?

Kind regards