My research is total population sampling technique. Total population is 147 company. But now only have 107 company answer the questionnaire. Izit possible to proceed data analysis?
Sample size, Effect size, Significance level and Power - they all need to be in equilibrium.
In any survey, non-response is expected. In household diet surveys, the non-response of 25-30% is common. To manage this, generally, the selected sample size is raised accordingly and then the survey is carried out.
In your case, whole population is 147 and you got the response of 107 that is about 72% response rate which is not different from what is expected. However, an important point to note is that the distribution of non-respondents with respect to variables under study should not be differing much from those who have responded. For example, the age distribution of non-respondents may be different from those who responded. Generally, speaking if non-respondents forms a special group and expected to behave differently from the respondents then the results obtained from the sample should be interpreted carefully.
Try to assess in what way the missingness is or may be related (directly or indirectly) to the research question. This allows you to make some educated guess about the possible bias. This should all be discussed when presenting the results.
Izit I need attach any past study that also same situation with me to support when presenting?
Sample size, Effect size, Significance level and Power - they all need to be in equilibrium.
Though I may not be clearly understanding your question, it seems that you had a sample size of 147 study participants of which 107 participated in the study. Are you asking if you proceed with your data analysis? should this be the question, then about 73% of your sample size responded to your questionnaire. Mostly, expert opinion advocates for a response rate of above 90%. My suggestion is that you may go back to the field and meet the suggested response rate before you analyze your data so that your sample may be a true representative of the population. Thank you.
What I would do in this case is include the 107 participants only and perform the statistical analyses. You can exclude those who did not participate in the study. Anyway, the point of the your study is to extract relevant information and reach conclusions. So as long as the sample size is larger than 20, you're fine.
- Badges
- Science topic