In my data set, I have variables coded (1-not worried, 2-Somewhat Worried, 3-Very Worried). I have six of these variables and I want to create a "Worry Index". One possible way to add the values for each case and divide the result for each case by 6 (number of variables). I also came across people just adding up the scores and keeps them as they are. When you don't dive them by 6, the new index variable has the values from 6-18 and I can reduce them 1-13 by recoding the variable. I am not sure that is appropriate and the best way. I will appreciate your comments and suggestions.
I think it would be correct to calculate how much information is contained in each variant of the answer (by the way, you can not encode them, but consider them as linguistic variables) that the Respondent who answered one way or another belongs to a particular group (class). This amount of information is a particular criterion that gives 0 (zero) if the answer does not contain any information about what interests us, gives >0 if the answer contains information about belonging to a class, and 0, если в ответе содержится информация о принадлежности к классу, и
- Badges
- Science topic
- Similar topics
- Indexes