Test of Non-Response Bias
Test of non-response bias are examined to ensure that the non-response bias in mail surveys was not
debatable. The non-response bias testing procedure is evaluated by comparing between early and
late complete returned questionnaires, whereas the late responses represent to non-respondents
(Armstrong and Overton, 1977). In this research, all 89 received questionnaires are equally separated
into two groups. The 45 early respondents are in the first group and the 44 late respondents are in the
second. To test the non-response bias, both groups of respondents are compared with demographic
data. The result shows that no statistically significant difference between early and late respondents
indicating non-response bias between respondents and non-respondents in terms of demographics.
As a result, non-response bias is not a problem in this research.