Perbandingan Skala Kontinu dan Skala Diskret

Abstract

In social studies, researchers often measure the characteristics of coutinuously measured objects by using discrete scales. So far the discrete scale are considered among the most popular methods used in the study because it is easy and simple. For this purpose, we need scale which can put respondents answers into a particular option, but has a relatively small bias. This study focused on the measurement of the magnitude of the bias arising from the multiple scale number associated with the number of samples, the distribution of the data and the correlation between variables. We than simulated using a continuous scale and then transformed into a discrete scale with six categories, namely scale 2 - 7. For each category, the magnitude of the sample size did not significantly affect the bias. The magnitude of the bias due to the categorization of the scale in this study indicates that the more categories used the smaller the bias. Category scale that has the smallest of bias was 7 with the average value of the bias 3.32%.