Designing a Likert-Type Scale Questionnaire and its Analyzing Procedure

Abstract

This article aims to provide comprehensive guidance on distinguishing between Likert-type and Likert-scale survey items, as well as the appropriate statistical approaches for analyzing their intended measurements. The study involved reviewing scholarly articles published in journals and other relevant documents pertaining to Likert-related issues. It is commonly observed that odd-type Likert items use a five-point response scale, while even-type Likert items utilize a four point response scale. Likert-type items consist of distinct and mutually exclusive questions that cannot be combined to form a scale, and therefore, need to be analyzed separately. On the other hand, Likert-scale items are comprised of questions that are logically sequenced, closely interrelated, and can be combined to create a composite index. When it comes to measuring the central tendency of Likert-type data, the mode and median are appropriate measures for ordinal scale data. For Likert-scale data, which should be in an interval scale, the standard deviation is a suitable measure, but it is important to calculate a composite score beforehand. In order to assess the internal consistency of Likert-type survey items, Cronbach's alpha is suitable for interval scale data, while ordinal alpha is appropriate for Likert survey items of an ordinal nature.