I am trying to measure destination image but the established scales varies for each variable some are 5 point likert scale, some 10 point likert and another is a 7 point likert scale. On pilot survey, it was visible that the respondents are facing difficulty due to different types of scale in the same questionnaire. Would it be wise to modify all the scales in simple 5 point likert scale. The nature of the scale remains similar to the original scales such as middle of the scale represents neutral, left represents negativity and right represent positivity. A pilot survey was conducted with this scale and the cronchbach alpha value is high. Can this modification be adopted.
This is an interesting question...perhaps you could contact the scale's creator/creators and ask if they will allow modification to the scale? Perhaps you might want to share with them the reliability coefficients that you obtained after modification. Let us know, please, your response from them - if you do contact them.
Ruth
You would need to compute alpha or the equivalent for the revised scale, but you would probably do that any way.
In general, I believe that there would be a preference for the longer (7-point) versus shorter versions. Here is an article by two well-known authors who discussion this sort of thing.
How many points are in a scale should be the result of whether there is value in dividing the answers more or less. The original authors presumably has reasons for using five, seven, or ten points. You need to ponder whether using the same number of points for all sections will give you the differentiation of answers you need for the kind of analysis you will do. It may be that some instruments provide better results when divided into more choices, but it may also be that there is no benefit.
In your paper, you will need to cite the original instruments and explain the logic leading you to change the scales. I think you should say more than "to reduce confusion to respondents."
No, the interest of proof questionary is to know how respondents understand questions, Scale would be improved after this
Usually before applying a scale for a country or region should do their validation, although in the same language some expressions can have different meanings. For example in validation we are doing in older adults in Colombia,There is a question that says:Do you use the phone mobile?It was changed to: Do you use celular?. for better understanding.
The assessment or quantification of the question should not be changed if the validation is not done.
Standardization of scales is wise idea, and I would go for a 7 point scale as a 5 point scale always display enough variability among constructs. But this will mean you are adapting scale rather then adoption. As a result you must validate your scales through different reliability and validity procedures.