I have all the items on 5-point likert scale (1-strongly disagree-5-strongly agree), but one construct is having items on 5 -point likert scale (0- not applicable-4- severe). Can I run all the items together on EFA. Will different range of point scaling cause any problem in factor loadings or the EFA result?
The actual values (1 to 5 or 0 to 4) don't really matter. The important thing is the number of items on the scale. There's no reason why you couldn't change your 0-4 construct to 1-5 to make it easier to compare to the others.
Dear Kirti,
While Stuart Cunningham is correct in that the scale itself does not matter from a quantitative viewpoint, it does from a semantic viewpoint. The fact that you are using different descriptors with different meanings (agreement in one and applicability in another) does have implications. First, it is best not to mix the two response types in the survey itself to lessen the chance of confusing the respondents. Second, the two response types should be analyzed separately. I hope this is helpful and good luck with your project.
Jim
Dear James and Long,
In one question I am seeking the level of agreement or disagreement while in other, the level of impact intensity not applicability (NA, MILD, MODERATE, SEVERE, VERY SEVERE). Not applicable is given to help the respondent opt for that category in case they have not experienced a particular situation for which I intend to measure the impact. First, I am not sure if we can use Not applicable as a response category in a likert scale.@ Long She@James Secondly, I am going to put all my items together in EFA. How can the two be analysed separately?
Kirti Goyal, just quickly: I think it is a really good idea to include a not applicable response option in scales, partly because many respondents might use the midpoint of an SA to SD option range for exactly that purpose - and that will muddy the data. However, responses on that option should not be included when forming composite scores or in such procedures as EFA. I would choose to do the EFA with pairwise (not the default of listwise) deletion of items.
As James McLean cautioned, I would be VERY reluctant to pool all of your items into a single EFA. You can do it, but I'm pretty sure it would be a case of GIGO: garbage in, garbage out.
All the best with your research.
- Badges
- Science topic
- Similar topics
- Linguistics
- Composition Studies
- Publication