My data for dependent variable was collected using likert. Now I like to perform binary logistic regression. For that I have converted likert scale data into binary form like I consider total score and calculate mid point of the total score, if the Respondent total score is higher than mid point then I consider it as positive and assigned it as 1 while for lower than mid point as negative and given 0.
Now is it appropriate to use.
Please help.
Thank you
Likert scale is an ordinal, then you can use it as a response variable for ordinal regression; or you can transform it to be a binary (and of course you then have less information) to use logistic regression, as you have done.
I would recommend ordinal logistic regression. As far as I know, nobody recommends transformation. See attached
As recommended by David Eugene Booth and Ayman Amin , it's preferably to deal with ordinal logistic regression, but be ware for sample size and convergence models difficulties.
Furthermore, in case of Items response data, I generaly suggest SEM techniques.
Best Luck
Hi Amir Sohail Choudhury ,
Technically, if you follow steven's categories of nominal, ordinal, interval, ratio, your Dependent Variable is ordinal, and should be analyzed with some form of ordinal logistic regression.
But the most common type (by far) of ordinal logistic regression is the proportional hazards model, which assumes proportional hazards. That assumption might be violated, in which case, you might want to use multinomial logistic. Collapsing the variable will only very rarely be correct. It throws away information, and that’s rarely a good thing to do.
- Badges
- Science topic
- Similar topics
- Linguistics
- Composition Studies
- Publication