for example there are variables A, B,C, and D. Each variable has its own questionnaire and different likert scales. A is based on 5-likert scale, B 6, C and d 7. Will it be okay to combine all in one and there is now issue as far as likert scale is concerned or I have to standardize all the variable to a similar likert scale?
CTT solution- bin into equivalent response categories.
IRT solution- self-evident with potential for nominal response model to estimate response biases associate with different number of categories.
Hope this helps,
Matt
It is not necessary you combine the variables into thesame or similar scale since they are measuring different constructs. Once they have been appropriately sectionalized and the items in the different sections have been content validated, it's okay. However, the reliabily coefficient of each of the sections in line with the various variables under investigation should be separately calculated and results of the data collected from the use of the Questionnaire should be reported separately in line with the variables and research objectives.
Your may use POMP to get the same scale for all the variables:
Article The Problem of Units and the Circumstance for POMP
If A, B, C and D are the same variable measured in disparate questionnaires with different Likert-type scales, then you may "POMP-them" and then you could compare the scores. For example, testing the hypothesis that the four generated scores come from the same distribution or not.