Analyzing and Interpreting Rating Scale Data from Questionnaires

Questionnaires comprising Likert rating scale items are probably the most commonly used instrument for data collection in education and educational research, yet there is much less guidance available concerning how to analyze and interpret such rating scales than there is on how to design them. Likert scale questionnaire data is most commonly analyzed using means, although this assumes that the Likert scale descriptors are equally spaced points on a continuum. Alternative ways to analyze rating scale data include percentage frequencies, medians and modes. If means are used, there are several different approaches to interpretation depending on how concrete, intuitive or unbiased the interpretations should be. This paper presents the options available in analyzing and interpreting Likert scale data and shows how the choice between options depends on the purposes of the research.