The Defect of Polytomous Successive Integers in Likert Scale

The purpose of this paper is to point out the inherent defects of the Likert scale as a research tool. Research in social science is accomplished through survey questionnaire. The questionnaire is the instrument for the research. If the instrument is defective, the measurement would most likely be inaccurate. The accuracy in social science research is measured by means of bias determination. In addition to pointing out the structural defect of the Likert scale, this paper also provides a means for instrument calibration by calibrating the survey instrument to produce a 95% confidence interval yield. For a 95% confidence interval, the number of answer choice is 2 with a degree of freedom of 1. This paper presents trial runs which point to a conclusion that a discrete binomial data element is the most accurate and displays the least bias. According the survey instrument should be construct to solicit {yes|no} answers. A response subspace with successive integers in a Likert form should be avoided in order to eliminate artificial data bias