I have run a principal axis factor analysis and I want to form scales for each factor. To be more specific, after having run the FA, I would like to select all items that load in a specific factor and form a new variable (which will be the factor). For instance, if item 1 and item 5 load to factor1 which expresses distress, I would like to create a new variable under the name Distress, which will consist of item 1 + item 5. So how can this be done in SPSS? Apart from the compute command using SPSS syntax, is there an alternative way to form such scales in SPSS?

  • 2
    $\begingroup$ It is unclear what you mean saying to form scales for each factor. Either it is computation of factor scores (and then get under button "Save" in factor analysis dialog) or it is selection of items into the scale based on FA results. The latter is a psychometric task where FA is only one step. $\endgroup$ – ttnphns Aug 12 '15 at 12:42
  • 1
    $\begingroup$ Thank you for notifying me. Is the question clearer now? $\endgroup$ – Katigkou Aug 12 '15 at 13:02
  • $\begingroup$ I've seen the discussion of this under the name of "parcels" - perhaps google shows you some relevant discussions on this. (I understand you in that way, that the factor loadings give hints, which items to parcel together (for instance make a simple sum of them) to create some conceptionally unidimensional sub-scales from the whole scale.) I think I've also seen the term "clustering of variables" (maybe even hier in SSE) but can't say where... And to do this in SPSS I think you need to step into the matrix-language - at least there's no standard possibility. $\endgroup$ – Gottfried Helms Aug 14 '15 at 10:39

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.