4
$\begingroup$

I have data in which the response variable (attitudes towards tourism) is scaled in nature ranging from -10 to +10 (calculated from the summation of scores of a few questions related to tourism). The independent variables are continuous, categorical (3-4 categories), and binary (Yes/NO) in nature. I can categorize the response (attitude) into Negative, Neutral, and Positive and run an ordinal logistic regression (not very sure even this is the correct way). But I have come across some literature relevant to my study where they have used multiple linear regression even when their independent variable was Categorical (Negative, Neutral and Positive) and some cases when the independent variable was scaled like mine. Can I use multiple linear regression in this case?? IF not, what about ordinal logistic regression?

$\endgroup$
4
$\begingroup$

With a response variable on an ordinal scale of -10 to +10 this is a valid reason for treating it as numeric and running a standard multivariable regression. You will need to inspect residual plots in order to assess whether they are plausibly normally distributed if you are going to make certain inferences.

This will likely be much better than categorizing into 3 levels as this will result in a lot of information loss.

| cite | improve this answer | |
$\endgroup$
  • $\begingroup$ Thank you for the prompt response to my query. So, now I should test the normality to ensure that I can use standard multivariate regression. Can you help to get access to reading materials related to my query so that I reference that in my method section? $\endgroup$ – kabindra shahi Jul 17 at 14:33
  • $\begingroup$ Don't test normality. Just assess it with residual plots. You can use multivariable regression, not multivariate regression. $\endgroup$ – Robert Long Jul 17 at 14:38
  • $\begingroup$ Yes, it answers my query. $\endgroup$ – kabindra shahi Jul 21 at 15:31
  • $\begingroup$ Done. Sorry, I was unaware of this. $\endgroup$ – kabindra shahi Jul 21 at 23:12
  • $\begingroup$ After checked the residuals plots what I found is that the lower-left points are not along the line (are below the line). I also did the shapiro.test (I am not sure if this is the right way of seeing normality in this case) and found value less than .05. So, I guess I have to opt for ordinal logistic regression now. $\endgroup$ – kabindra shahi Jul 23 at 17:03

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.