Social surveys are often used to estimate various indicators of social and health characteristics, while survey non-response is a common phenomenon which may cause bias to the estimates. This study used a unique dataset on knowledge of child injury to analyze the pattern of non-response and its possible effects on the estimates. The relational pattern between the characteristics of respondents and the outcomes of interview has direct effect on bias in the survey results. In an effort to find ways to increase response rates and therefore improve the quality of the data in future surveys, this study examines the characteristics associated with survey non-response.

The aim of this research is to identify the pattern and determinants of non-response in the context of Bangladesh and analyze with the above methods to have in-depth knowledge about the non-response affect on used data which is one of the most important topics for the statisticians. The findings of the study show that non-response is higher in the households with persons having higher education and a service as occupation. The effects of non-response are analyzed on statistical models; however, the effect was found to be statistically insignificant might be because of overall low awareness of the people of all levels. But such a high non-response of a selected group may produce significant bias in other social studies where indicators are more strongly related to education or occupation.