FAQ About Survey Research
COMMON QUESTIONS ABOUT SURVEY RESEARCH

Why not do a mail survey to everyone in the City/Why is a phone survey best?

The simple answer is that a telephone survey allows greater control and confidence of the research results. With mail surveys, there are no guarantees of who responds, how many will respond, or when they respond. Telephone survey research allows the sample of respondents to be controlled, assuring a statistically valid representation of all residents. Further, in such telephone studies, the questionnaire is controlled and administered to assure that respondents understand the questions and address the appropriate issues. Of equal import, phone research enables better control of the timing and costs associated with the research effort.

Mailed surveys are the only solution in situations where you have no telephone access to the target market. In these cases, DMG can assist in the development of an effective mailed survey. Another consideration is the Internet. As Internet access has grown, in some cases, email lists can be purchased and respondents captured online. DMG has extensive experience with this type of survey. DMG principal Matt Disston teaches both market research and Internet courses at the University of California, Irvine.

Can't the surveys be conducted by volunteers or students to reduce the cost?

The use of inexperienced and unskilled people is tempting due to the perceived potential for cost savings. Such a perception many not fully appreciate the fact that accurate respondent selection, complete and correct data collection, as well as flawless recordation of answers constitute the most important links in the chain of tasks necessary to implement a valid consumer research effort. All three of these tasks are ultimately the responsibility of the interviewers administering the survey questionnaire. Based upon our experience, we have concluded that these tasks demand the use of trained and experienced professional interviewers in order to ensure data validity and maintain cost control. We bring all of our professional expertise to bear in the design and analysis of the research and feel that underestimating the importance of comparable expertise in the actual interview process jeopardizes the reliability of the research conclusions.

How many interviews are enough?

The volume of surveys required relates to two principal issues -- 1) the size of the error range, and 2) the number of universe sub-areas to be analyzed. As an example, a random sample of 400 surveys enables the City to be confident that 95% of the time, the responses of those sampled will be within + 5% of the average of all households in the City. With respect to analyzing the results by two or more sub-sets of the total sample, it should be recognized that the same test of confidence and error range is required for each of the sub-areas. Looking again at the example of 400 interviews in the City, a four-area breakdown yielding 100 questionnaires per sub-area would anticipate a + 10% error range. That is, 95% of the time, the responses of the 100 households sample in the sub-area will be within + 10% of the average of all households in the sub-area.

How long should the questionnaire be/What about refusals?

Refusals, question design and questionnaire length are interrelated aspects of survey design and implementation. The general population continues to be quite responsive to survey research, particularly when the issues are within their city. It has been our experience that the usage of professional interviewing staff significantly contributes to the level of cooperation. However, the questionnaire design must recognize respondents' tolerance for the time required to complete the survey. Moreover, the questionnaire design must consider respondents' ability to provide meaningful, reliable answers.

Can we accommodate non-English speaking respondents during our interviewing?

DMG Economics has the experience and ability to provide non-English speaking interviewers as well as questionnaire translation services.