Hello Alicia,

The answer depends on both your specific research question(s), and what, if anything, you do to the score associated with a response to a given statement by a respondent.

If you're using individual statement responses as variables in your work, then you should adopt an analysis which treats your scores as ordinal data points. Something as simple as, "82% of respondents agreed or strongly agreed with the statement, 'The income tax system favors the wealthy' " may be perfectly satisfactory as a summary of your data; again, it depends largely on your aims.

There are ways to combine multiple statement responses into scores which can more defensibly be treated as interval strength, but that of course minimally requires that it makes sense to combine them in the first place. That is, are the multiple statements all associated with the same attribute, characteristic, opinion, or construct being measured? Factoring the set of variables is one possible way to gain insight as to whether that common dimensionality exists.

Good luck with your work.

Hello

the analysis is dependent on what you want to do. For example, if you are looking to check the dimensionality of the scale, you must do an exploratory factor analysis

If you are looking to establish relationships and both your dependent and independent variables are Likert type, then Ordinal logistic regression will sort you out.

Alicia Silver, analysis at the question (item) level should focus on a descriptive analysis of the responses to each item. That is, provide a table or chart indicating the frequency or relative frequency of the responses to each of the agreement levels.

If you are looking for a single numerical representation of the responses to each item, then the statistical mode would be the measure of the center to employ with these ordinal-level data.

Upon completion of these descriptive analyses, then you can begin to make sense of these results in terms of degree of agreement or in terms of which items share similar results.

All of the above. The individual Likert scale values can be treated as categories, or as ordinal values.