{"id":148,"date":"1980-09-16T08:17:24","date_gmt":"1980-09-16T12:17:24","guid":{"rendered":"http:\/\/www.eprdev.com\/jag\/?p=148"},"modified":"2019-02-07T12:49:07","modified_gmt":"2019-02-07T17:49:07","slug":"a-three-factorial-experiment-on-response-rate-to-male-questionnaire","status":"publish","type":"post","link":"https:\/\/www.jagsheth.com\/consumer-behavior\/a-three-factorial-experiment-on-response-rate-to-male-questionnaire\/","title":{"rendered":"A Three Factorial Experiment On Response Rate To Male Questionnaire"},"content":{"rendered":"
While considerable experimental research exists on the rate of response in mail surveys as a function of a variety of manipulations, there is still very little agreement on the factors that are conducive to optimizing the rate of response (Champion and Sear, 1969; Erdos 1970; Gullahorn and Gullahorn 1963; Landy and Bates, 1973; Linsky 1965; Scott 1961). Our study is one more experimental effort to measure the main and the interaction effects of (a) questionnaire length, (b) follow- up procedures, and (c) geographical differences on the rate of return of questionnaires mailed to a large random sample of telephone customers. The study consisted of a 2x4x8 three factorial fixed effects field experiment study involving a total of 64 cells experimental design; and a follow-up interview with a subset of the sample to explore respondent attitudes toward mail surveys.<\/p>\n
The first experimental factor, questionnaire length, consisted of two levels: a four-page short questionnaire consisting of 23 items related to the socioeconomic-demographic profile of the respondent, and a six-page long questionnaire consisting of the sane 23 items plus an addition of 26 additional items measuring respondent attitudes toward the telephone as a product and a service.<\/p>\n
The second factor was follow-up methods. Four follow-up methods were selected based on cost and feasibility considerations. Following the mailing of a post card alert and the questionnaire, the four follow-up methods utilized were as follows:<\/p>\n
The third factor, geographical heterogeneity, consisted of eight distinct geographical regions of the country. The geographical areas were chosen based on a cluster analysis of one hundred geographical areas which represent the total population for the Bell System telephone customers. Each geographical cluster consisted of homogeneous customers with respect to their socioeconomic-demographic profile and their telephone behavior profile. The following eight geographical areas were chosen to represent each of the eight clusters: Southeast Massachusetts, San Jose, Arizona, Philadelphia, Eastern Wisconsin, Suburban Chicago, Fort Worth and Alabama.<\/p>\n
Within each of the eight markets, a representative random probability sample of 264 customers was drawn and equally divided among the eight experimental conditions created by the two questionnaire lengths and the four follow-up procedures. Thus, each experimental cell had a sample of 33 customers. The only exception was with respect to the eight experimental groups in the Southeast Massachusetts region where a sample of 296 customers was drawn and equally divided into the eight experimental conditions to compensate for the anticipated higher rate of disconnection of telephone services due to the end of the summer season.<\/p>\n
The experiment began with the mailing of an alert post card to the total sample of 2144 respondents (264×7+296) on a Monday of the first week of the experiment followed by the mailing of the questionnaire on the following Wednesday. The post card and telephone reminders were initiated on Friday of the second week if the questionnaire had not been received by that time. The mailing dates including those for the follow-up procedures were designed in such a way as to compensate for the regional variations in postal delays. A total of four attempts were made to contact each respondent for the telephone reminder and the telephone interview procedures. Finally, calls were made at various times of the day and evening which the local supervisor considered most appropriate in terms of receiving cooperation. In short, every possible effort was made to minimize the differences in situational factors (time, person and opportunity to reach) in the field experiment setting between the telephone and the mail interviews.<\/p>\n
As stated before, the total sample was equally divided into the 64 cells giving a cell-sample of 33 customers each except in the Southeast Massachusetts area. The questionnaires returned from the post office as undeliverable were subtracted from the sample base in each of the cells. Further, the additional mailings made in the Southeast Massachusetts were taken into account to modify the base for each experimental cell so that the number of returned questionnaires for each cell would be comparable.<\/p>\n
The experimental results are summarized in Table 1. The range of the rate of response varies from a low of 32.3 percent in one cell to a high of 87.9 percent in another cell.<\/p>\n