Response rate

from Wikipedia, the free encyclopedia

Response rate or response rate , occasionally exhausting , (Engl. Response rate ) referred to in the quantitative social research , the ratio of solid sampling units actually reached units.

High response rates are usually interpreted as an indication that the sample is highly representative . However, this is not a necessary condition for a representative sample, provided the low response rates are the result of random failures. A non-response can also cause a non-response bias ( non-response bias result). Since empirical research strives for the answers to be highly representative, this should be excluded. At least the response rate is an indicator of the general data quality.

The recruitment affects the number of all possible interviews. Every non-response subsequently influences the response rate. This can have very different causes, for example if only certain information is missing or contact was not possible.

meaning

Response rates play a role especially in survey research . Here, the response rate corresponds to the ratio of completed interviews to attempted interviews: if an interviewer tries to survey 150 people, but only manages to complete the survey for 30 of them (e.g. because most respondents refuse to cooperate), the response rate is 20%. The situation is similar with written surveys: If the market researcher receives 100 of the 1,000 questionnaires sent back, the response rate is 10%.

In general, the highest possible level of exploitation should be aimed for, since failures are often not clear as to whether they are systematic failures (for example, a particularly large number of supporters of a certain party refuse to take part in the survey) or non-distorting, non-systematic failures (e.g. in a household sample advised business premises).

The different types of failure are also taken into account when calculating the response rates. Non-systematic failures are not taken into account for the calculation of net utilization .

In general, however, the response rates for written surveys are significantly lower than for oral surveys. For written examinations - unless they involve special surveys with a defined group of people - response rates of more than 15 percent are already remarkably high.

Measures to increase the response rate

Many steps can be taken to increase the response rate. These measures basically have different starting points.

  • On the one hand, an attempt can be made to increase the general willingness to participate in a survey. In traditional market research , inter alia, interviewers can be trained in order to be more persuasive in the interview initiation situation. Written advance information can also have a positive effect on the study participants. Many potential study participants are often unclear about the purpose of a study; this can be counteracted. In the online research other strategies need to be applied. Althoff and MacElroy refer to the definition of a suitable dispatch date for invitation e-mails, the possibility of further personalization of invitation e-mails (personal address through known contacts) or the use of the Anita effect .
  • On the other hand, measures can be taken to minimize the risk of the interview being broken off. The termination of the interview correlates very strongly with the length of the questionnaire and the interview time. Here a trained interviewer can counteract a termination in a paper-pencil or CAPI interview . For online interviews , Althoff and MacElroy refer to the use of animations ( surveytainment ), the layout design of the questionnaires (use of images) and the PRD technique (a procedure in which a respondent is shown the previous ratings) as possible instruments to minimize interview interruptions.

calculation

In practice there are different methods of calculating the response rate. This is because there are several ways to divide the number of completed interviews by the number of "attempts". If you divide the number of completed interviews by the total target sample, you get a lower response rate than if you divide by the sample adjusted for neutral failures (address errors, target person not part of the population ). The members of scientific associations involved in market and opinion research are required to adhere to agreements on the calculation of the various response rates in order to avoid manipulation and misleading customers and the public.

The American Association for Public Opinion Research gives 6 different types of response rates:

  • Response rate 1 (RR1): Minimum response rate, the number of complete interviews divided by the total number of interviews. The total number includes both complete and partial interviews, as well as non-interviews (rejections or abandonments, non-contacts, etc.) and all other reasons
  • Response rate 2 (RR2): like RR1, in addition, interviews are sometimes evaluated as answers,
  • Response rate 3 (RR3): like RR1, plus an estimate of the proportion of unknown cases (estimated impermissible cases are removed from the denominator),
  • Response rate 4 (RR4): like RR3, but takes into account the complete and partial interviews,
  • Response rate 5 (RR5): like RR3, assumes that the number of unknown cases is zero,
  • Response rate 6 (RR6): like RR5, but takes into account the complete and partial interviews (maximum response rate).

literature

  • Robert M. Groves, Lars E. Lyberg: An overview of nonresponse issues in telephone surveys. In: Telephone survey methodology. 1988, ISBN 0-471-62218-4 , pp. 191-212.
  • Larry E. Miller, Keith L. Smith: Handling nonresponse issues. In: Journal of extension. 21, 5, 1983, pp. 45-50.
  • Donald B. Rubin: Multiple imputation for nonresponse in surveys. John Wiley & Sons, 2004, ISBN 0-471-65574-0 .

Individual evidence

  1. ^ Arnold Hinz: Psychology of the time. Waxmann Verlag, 2000, p. 158.
  2. ↑ Jürgen Schupp, Christof Wolf (Ed.): Nonresponse Bias: Quality assurance of social science surveys. Springer-Verlag, 2015, p. 13.
  3. ↑ Stefan Althoff: Selection process in market, opinion and emotional social research . 1st edition. Centaurus-Verlagsgemeinschaft, Pfaffenweiler 1993, ISBN 3-89085-899-6 , p. 113 .
  4. ↑ Should an online interview be fun? In: Quirks Marketing Research Media. Retrieved June 10, 2016 .
  5. ↑ PRD technology. In: www.marktforschung.de. Retrieved September 19, 2016 .
  6. ↑ Should an online interview be fun? In: Quirks Marketing Research Media. Retrieved June 10, 2016 .
  7. ↑ Standard Definitions ( Memento of November 3, 2012 in the Internet Archive ), American Association for Public Opinion Research, 2012.
  8. ↑ AAPOR Standard Definitions - PDF, 2015. S52ff.