I think this is not usually done for Likert scale anchors. The “other” answers take those responses out of your ranked numeric scoring for the questions, so it will have the effect of making missing data for these questions when scoring. It is probably better to carefully word the items to avoid “double barreled“ or otherwise ambiguous questions, pilot test your questions with a good number of test takers who can give you feedback, and make sure that each written item is actually something where frequency is the important characteristic you wish to measure. This is key. If frequency is not the correct measurement, then your Likert scale will be confusing to responders and your answers will be less valid.

The simple answer is 'no'. What would you be expecting people to answer? If your item is asking about frequency, then this should be covered by the Likert options you have given (unless it could be 'all the time' in which case you could add this to your options). I agree with Kathryn, if your item is asking about something as well as frequency, then you are probably asking two questions in one. You may need to rethink your items. You could have a 'not applicable' option if that is appropriate.

Thank you Kathryn and Nutmeg, your response are helpful.

Glad to hear it, Yumna!

I would totally agree. Likert scales typically represent ordinal level variables which is where "the order of the values is what’s important and significant, but the differences between each one is not really known". "Other" does not fit into a more or less type of scale. So agree the answer is no.