| dc.description.abstract | A reliability test is often performed on survey data. The goal is to determine the data characteristics both in terms of consistency and accuracy. The reliability test has a condition that to be satisfied, i.e., the data must be continuous (ratio), while the data obtained from a survey is commonly an ordinal type. This paper aims to measure the magnitude of the bias related to reliability test and the patterns of alleged bias related to the number of categories and the number of data samples surveyed. It was shown in this work, as the number of categorized ordinal data and the number of sample data increase, the amount of bias decreases. It was also shown that by assigning 6 categories ordinal data scales and the number of data sample was 400, the level of bias would be not more than 5%. | en |
