Re: What is the best way to analyze Likert type items?

It all depends on your original resarch question. What is your
hypothesis? If certain questions measure a similar underlying concept,
than you could create a new variable by calculating the total score of
several questions. This way you create a continuous variable and could
use multiple linear regression. If you think that the the variable is
not continuous (for example because there are only a few categories)
you could use ordinal regression.
An other option would be to use logistic regression but you would have
to create a cut off point. Again this all depends on what you want to
know.

.

Quantcast