My study is relating to Islamic Banking Consumer knowledge and i want to develop a model with the help of CFA (by using AMOS). Actual problem is: According to prior studies, Subjective knowledge and Familiarity is developed on 5 points Likert Scale but Objective Knowledge is on Dichotomous Scale (yes, No, I dont know). I am confused like when i will be developing the model it will create problem and may not be loaded on desired variable because all constructs are not on same scale.
Should i convert objective knowledge (Dichotomous Scale) to Likert Scale?? if yes then how? Or i should continue with the Dichotomous Scale?
Dear Chaudhry,
It sounds like you are planning to turn the responses to your items into numbers and add them together for the items in the same scale. This is common practice but it is not the only way to cosntruct a scale variable based on a group of ordinal variables, and it may not be the most accurate method.
An alternative is to save the regression scores as the scale variable. This can be more accurate when the scores of the items on a factor differ in magnitude and it overcomes your problem.
Hello,
It is not essential that you scale equally with all the response formats. However, having variables with docitomic formats could cause problems in the modeling that you propose. My recommendation is that you transform the dichotomous responses to likert.
How? A very simple procedure is as follow: Recodify your dichotomous answer format to: No = 0; I Dont know = 1; Yes = 2. With that you get a small classic likert scale and you can solve future problems of negative variance.
I hope it will be useful.
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Materials Science
- Materials
- Inorganic Materials
- Ceramics