I am using the famous Muncy& Vitell(1992) scale for my study. The scale contains 4 ethically posed situations each having further vignettes. A total of 17 items as a whole. I have gathered 338 online valid responses (out of 350) and 388 self-administered survey responses (out of 410) in Pakistan. On conducting the EFA analysis we have received only two types of questionable activities out of four which I categorized as i) Harmful activities ii) Harmless Activities. This was further verified through 6 semi-structured interviews too. I wounder how much an internationally developed and tested scale can be presented or modified in this manner. Please guide...
Scales developed in context of a specific nation may not be applicable across others and hence it is possible to have certain components being changed or modified. However the same needs to be driven across samples to ensure stability
I agreed with Shweta because sometimes translation of a scale is not enough due to the context in which it will be aplplied.
In Brazilian environment everytime I use an international scale I submit it to a panel of theme specialists to see if the meaning of the translated version fits in the context of the theme I want to cover in local environment.
It is not unsual I have to adapt some terms and expressions otherwise the respondents will not understand what I want to say with this or that word.
Other question to consider is that when you use EFA with a software like SPSS, the onee I use the most, the purpose of the program is concentrate the answers in a minimum number of factors, which means that if you have a high correlation between the answers the tendency should be to have a concentration of them in a lower number of factors.
This can also indicates some level of collinearity of the items used to measure the constructs (depite the items are different probably they are measuring similar things) and when you perform a EFA the software put all items in a same factor.
Suggestions:
1) Run a correlation analysis of the answers and check if there are high values between items of different constructs. If there is a high correlation maybe you have to eliminate some of them because they are in different constructs but for the respondents they are measuring the same thing (clearly this is a problem related to the scale translation);
2) Try to do a EFA but forcing the number of factors for 4 and see if the items are addressed to the right ones. If this does not work, again there should be items related to different constructs but that for the respondents they are seen as measuring the same thing.
Hope I give you some sort of help.
Regards,
I agree with Alcides and recommend to perform CFA also after EFA.
Regards,
Hello. I agree with raising Alcides makes sense and is consistent with the logical in this field.
Due to cultural differences and different research setting, it is not a matter of wondering, I will prefer CFA with CB-SEM.
- Badges
- Science topic
- Similar topics
- Biological Science
- Bioecology
- Systems Ecology
- Landscape Ecology
- Connectivity