Non-response is a form of survey research error that occurs when some participants don’t respond to the survey, thereby changing the representativeness of your sample. This blog concerns the bias or survey research error associated with this non-response.
Suppose you have a sampling procedure in which everyone in your sample must respond so that it represents the population. If some people do not respond, your sample is no longer representative of the population. Uneducated people may not be able to respond to the questions unless you conduct an in-person interview. Such an instance shows how your bias in the way you choose to collect data excludes people from the sample that need to be there to make it representative of the population. Suppose this lack of representation is causing demographic differences between your respondents and non-respondents. In that case, you have a large, inflated bias contributing to your survey research error which decreases the validity of your results.
Reasons behind non-response bias
- Inadequate survey design
One of the critical reasons for non-response bias is inadequate survey design. If the survey is excessively lengthy or difficult to comprehend, respondents may not answer, leading to non-response. Hence, it is imperative to keep surveys concise and user-friendly to maintain high response rates.
- Refusal to participate
In cases where respondents are busy or not interested enough to participate in the survey, it can lead to refusal and is fairly common. In that case, it is important to convince or motivate respondents to participate in the survey.
- Incorrect targetting
In some cases, even targeting could be a potential reason for non-response. They may not have the time to fill them out or be interested in supplying you with their personal information.
- Operational issues
Sometimes, non-response could also be because of operational issues in executing the surveys. In the case of face-to-face surveys, sometimes you cannot reach the right respondent. If you survey by email, it may end up in a spam section and never reach the intended recipient. It is still recorded as a non-response and does not contribute to the improvement of the survey.
Ways to minimize non-response
- Rigorous piloting and testing
Rigorous piloting and testing can sort out fundamental operational issues. It can help smooth the operation challenges and develop robust standard operating procedures (SOPs) for executing the surveys.
- Keeping it simple
One of the best pieces of advice is to keep it simple. If the survey is long, it might cause fatigue, and respondents may not complete their replies or even begin the survey.
- Finding the appropriate channel
One can also consider the efficacy and relevance of several survey distribution channels such as face-to-face interviews, emails, web, and links to determine the most successful and relevant survey.
- Clearly communicating about survey objective
Clear communication about survey benefits and use can help set realistic expectations about the survey. It also clarifies the potential benefit of the survey and why respondents shall participate. An excellent introductory message sets expectations about confidentiality and other consent conditions. One can also specify the survey’s purpose and length of completion. If the questions are sensitive, the consumer must be reassured that the survey will be anonymous
- Explaining the possible benefits
Inform your respondent how their participation will benefit them. It might be as easy as informing the consumer that their input would enhance the scheme’s functioning and, in the end, will benefit you.
- Persuading them gently
In case of non-response, don’t give up easily. There are several ways one can always follow up. In case of non-response, it is important to have a reason for non-response: if the reason is busy or away, one can always visit later. Further, if the respondent is not convinced, one can always find ways to convince respondents. One can always nudge gently by giving a follow-up message to let the respondent know that their feedback will be personally considered and methodically implemented to enhance the quality of the scheme, service, or product.
Analyzing non-response
Non-response is one of the most important issues in sample surveys. There are several methods for dealing with non-responses.
- Comparison with the population:
To adjust for non-response, survey averages are compared with population averages, particularly for demographic variables such as gender, marriage age, and income. Even though population figures may not be accessible in many instances, certain sources may occasionally supply population values for comparison. The influence of such a bias on the variables of interest is then examined by comparing the sample mean to the population mean and calculating the difference.
- Comparison to external estimate:
A second strategy for analyzing the impact of the non-response rate is to compare the survey estimate to an external estimate. However, the problem with this approach is that the difference could be attributable to several other factors, such as the survey length or how the questions were asked.
- Intensive post-sampling:
The intensive post-sampling strategy involves interviewing a sample of non-respondents to increase the sample size. However, it is not possible in most circumstances due to the associated expenditures and the issue of time overruns.
- Wave extrapolation:
Extrapolation techniques are founded on the premise that persons who answer less readily are comparable to non-respondents. Wave extrapolation is the most used form to analyze the non-response rate. This strategy incurs little marginal costs and is highly recommended.
Kultar Singh – Chief Executive Officer, Sambodhi
