Ordered (ordinal) logistic regression (ologit, or oprobit), or multinomial logistic regression (mlogit, or mprobit).

OLS regression is only appropriate when the assumptions of OLS regression are met (normally distributed errors, etc).

As mentioned, if you have scales built from the sum of several 5-point questions, you might be able to get away with using OLS.  You still need to check model assumptions for the inference tests to be valid.

If you are focusing on single 5-point responses, or your scales end up with only a few different values, ordinal regression is a good choice.  It is rather easy to perform in R, if that's your poison.  ( http://rcompanion.org/handbook/G_11.html )

In theory, even a scale made up several 5-point Likert items is not appropriate for OLS.  One issue is that the data are necessarily limited to the left (all people answer "1") and to the right (all people answer "5").  Another issue is that with OLS, you are assuming that the responses can be summed to mean in a meaningful way.  That it, you are assuming that an answer of "3" and an answer of "5" average out to an answer of "4".  For 5-point scales, this is probably not true.  (E.g. "agree" might be closer to "neutral" for most people than it is to "strongly agree").

In practice, if Likert scale data will pretty well meet OLS assumptions, people will often use it as the best option.

So, the best advice to understand the assumptions of the method you will use, and be able to defend your choice.

I wrote a little bit here on the topic: http://rcompanion.org/handbook/E_01.html .

No. Adopting OLS in this case will produce estimates that are biased and inconsistent, which is against the assumptions/ properties of OLS. It will be better if you adopt a multinomial or ordered probit/logit, as the estimators here are obtained by maximum likelihood.

Hope this helps?

Thanks for all of you and these are helpful advice.

I prefer using RII, that is relative importance index for likert scale data

OLS? Nooo.

Using multinomial logistics will be a good idea