If my majority of the values in a variable are negative in pre test and while in the post test the values are positive.
How can I manage the negative values, Do I have to transform the data,
If I have to transform do I have to transform both pre test and post test values as we are going to compare between the variables
Dear Fayaz Khan
In Statistics, A common technique for handling negative values is to add a constant value to the data prior to applying the log transform. The transformation is therefore log(Y+a) where a is the constant. Some people like to choose a so that min(Y+a) is a very small positive number (like 0.001). Others choose a so that min(Y+a) = 1
Kind Regards
Qamar Ul Islam
I agree with Peter. Mathematically, the numerical values of your scale don't make a difference. However, you should consider that they could potentially influence answering tendencies and thus create systematic bias (especially if the scale is not validated well). Assuming tough it is, I personally would treat them the same way as a 0-7 scale. Like Peter already said, if you can calculate a total score depends on the questionnaire. But if it is possible and there are no inverse items (inverse items obviously need inversed scores for analysis) you can assume that an average lower score means average lower satisfaction.