It depends on your survey instrument (e.g., online 2% is a fine rate if your population of interest is the general population). The response rate itself does not mean anything about the quality of your survey or your data. What you need to analyze is the a.) non-response bias and the b.) representation of the population of interest. So your goal is not a high response rate, but the representation of the population of interest and low non-response bias (your entire data collection instrument affects this: population of interest, sampling, survey type, questions, benefit for the respondent, etc). Some offered 'panels' provide great response rate, I think most of the cases they are useless - lack of knowledge how they collected panel members.

a.) Non-response bias can be analyzed by a simple t-test. You compare early respondents and late respondents (assuming late respondent behave the same as non-respondents, please find the reference in my paper enclosed).

b.) You compare your population's demographics and your respondents demographics. If it is different, no problem, you need to apply the raking method (iterative proportional fitting) to assign a proper weight to each respondents. Please find my conference paper enclosed. You can find reference and application in my paper.

If you need support in your method design use Dillman Article The Dillman Total Design Survey Method

Conference Paper Manage sample and population differences by weighting - usin...

Article Public perception of bioenergy in North Carolina and Tennessee

Article Systematic Review of Bioenergy Perception Studies

Hi Ahmed,

I don't think that there are any absolute rates. Typical response rates range from 30% - 60%. The higher the better but it also depends on how your survey is conducted (mail, online). Here is a link that might be useful since it compares response rates:

https://www.uaf.edu/files/uafgov/fsadmin-nulty5-19-10.pdf

Thanks...Nick

It depends on your survey instrument (e.g., online 2% is a fine rate if your population of interest is the general population). The response rate itself does not mean anything about the quality of your survey or your data. What you need to analyze is the a.) non-response bias and the b.) representation of the population of interest. So your goal is not a high response rate, but the representation of the population of interest and low non-response bias (your entire data collection instrument affects this: population of interest, sampling, survey type, questions, benefit for the respondent, etc). Some offered 'panels' provide great response rate, I think most of the cases they are useless - lack of knowledge how they collected panel members.

a.) Non-response bias can be analyzed by a simple t-test. You compare early respondents and late respondents (assuming late respondent behave the same as non-respondents, please find the reference in my paper enclosed).

b.) You compare your population's demographics and your respondents demographics. If it is different, no problem, you need to apply the raking method (iterative proportional fitting) to assign a proper weight to each respondents. Please find my conference paper enclosed. You can find reference and application in my paper.

If you need support in your method design use Dillman Article The Dillman Total Design Survey Method

Conference Paper Manage sample and population differences by weighting - usin...

Article Public perception of bioenergy in North Carolina and Tennessee

Article Systematic Review of Bioenergy Perception Studies

Dear Ahmed,

I endorse previous answers. Regarding non-response bias, in addition to these fore montioned, you can compare those with non-completed questionnairs to those that were completed.

Regards

I totally agree with Robert. I face aspecific problem when researching companies in Croatia (right now exporters): the response rate ( early) is 8%, and they consider many questions to be getting into " business secrets".

Mr Robert's Response is very comprehensive and should be helpful to any one involved in research. Specific situations require specific solutions. In general terms, it is full answer. Thanks Robert.

I assume you want to publish the results eventually, so it is good to target for a response rates approximating 60% or more so you can meet expectation of the Journals editor.

The response rate in a survey depends on the size of the sample. If the response rate goes down a lot, it can affect the explanatory power of the sample.

Hello, Ahmed. The response rate for any given survey may be one of the least important factors in analyzing your results. There is no "accepted" response rate. It is accurate to say the higher the better, though. Telephone surveys historically had the highest response rates, but they were also the most expensive. Old fashioned snail mail surveys had the lowest response rates, but they were cheaper. Technology continues to change survey research methods. Moreover, your response rate may differ depending on your population and geographic area. You need a large enough N (useable returned surveys) to do reliable analyses. You also need representativeness; for example, if your responses are disproportionately from one sex, age group, income group, education group or race, you've got problems irrespective of your response rate.

The following sources should also be helpful:

• Baruch, Y. (1999). Response rate in academic studies - A comparative analysis. Human Relations, 52, 4, pp. 421-438.
• Baruch, Y. and Holtom, B. C. (2008). Survey response rate levels and trends in organizational research. Human relations, 61, 8, pp. 1139-1160.
• What is an acceptable response rate of questionnaires in Social Studies?: https://www.researchgate.net/post/What_is_an_acceptable_response_rate_of_questionnaires_in_Social_Studies

Low response rate should be expected when using a questionnaire sent by postal services (Dillman, 2011).

Its nice to see that no one is very adamant about response rates, but rather flexible. I have done one of the very few studies where non-respondents could actually be compared to responders in terms of responses to questionnaire scales. We found only very small effects, even though we had as little as 6% responders (see attached). My impression is that response and non-reponse is largely due to factors which have nothing to do with the responses, i.e. whether you respond or not is largely random.

In my Ph.D research,'Students and Radical Social Change'(Calcutta University,1991) it was 100%.But I think 80%could qualify as acceptable.

Basis of response rate is all respondant are active in the field of survey. By active I mean they know the well and have a thoughtful of the various dimensions of the issue, and are willing to voice the calculated opinion / information. Whereas, in factual situations, respondants picked through random may not always be familiar with issue. So important thing in survey in awareness of the issue and being thoughtful as well. Most of the random surveys about opinions of the people is with out criticality! Focus should be more of genuine and representative respondants. Sample size can vary from issue of survey. There cannot be general firmula as the cost and time are very important considerations as will.

Anything from 70% and above can be considered conventionally acceptable for a social science research.