Search

A conference paper pertaining to the role of representative sampling in the field of marketing is presented. According to the author, there are two breakdowns that affect the adequacy of a sample: how many different ways people are classified and how many different answers to each question there are.

The article discusses a number of problems related to the reader-interest surveys. The author notes that most surveys of this kind consist of an interviewer going into the field and administering a questionnaire. According to the author, an appropriate sample of the population being tested should be constructed, interviewers should be carefully trained to assure reliability and control questions should be included to ensure soundness of data collected. Also noted are problems with the method of interviewers pointing out sections of periodicals to respondents.

A great deal of methodological study in the field of survey research deals with the many possible sources of bias to be found in sampling surveys. It seems that the problem of bias is inherent in research upon human individuals and groups. It therefore becomes important to advance beyond the current accumulation of evidence that such bias does exist to an investigation of the more important question of the significance of such bias and the problems in the analysis of data that are known to contain bias. This paper illustrates this point with an analysis of bias resulting from sampling mortality. Many studies on bias in mail questionnaires have shown that those individuals who return their questionnaires at a later date are different from those who respond early. Other studies of refusals to be interviewed show the nonrespondents to differ in important respects from those who do respond. Panel studies reveal differences between those who continue as panel members and those who drop out. Thus, it is generally accepted that sampling mortality due to nonresponse is a serious source of bias in survey research and every effort should be made to secure as high a percentage of returns as possible.

The article reports on mail surveys. The author focuses on identifying methods that can be employed to improve the reliability of surveys by examining a related study. The study, which focused on data retrieved from recipients of a men's wear survey, featured a sample size of 2980. After interpreting the results of the study, the author asserts that mail questionnaires will produce efficient samples if the creators follow a set of criteria: 1.) Creators must understand the product 2.) Creators must compose an attractive questionnaire. 3.) Pretesting must be performed. 4.) Mailed reminders must follow the original questionnaire.

The article presents information about preparing questionnaires and conducting interviews for marketing research. The author describes the process of formulating questions and collecting data for market analysis. Details are presented related to selecting a method of obtaining information, the structural design of the selected method, and various forms of interviewing. Methods such as the mail questionnaire, telephone interview, and qualitative interview are outlined. Details related to the marketing strategies obtained from each form of interviewing are presented.

The article presents information on marketing research techniques used to measure consumers' attitudes. The author looks at marketing research in areas such as population distribution, income, trading areas, and retail outlets and how it has affected business enterprises in the United States. Details related to a new form of questionnaire to be used in consumer research is presented. Profiles of different types of consumers are presented. Several charts are presented showing consumers' preferences for a variety of commercial products.

It is informed that a number of individuals can be expected not to respond to any particular survey, especially if it is conducted through the mail. Therefore, it is necessary to take the problem of nonresponse into account in the survey design or the results may be biased. The purpose of this article is to explore the assumptions underlying the use of statistical weighting to adjust for survey nonresponse and the implications of using this procedure. The particular procedures appropriate for estimating population values from sample results depend upon the nature of the sample and the population. An important distinction between types of samples is that between judgment and probability samples. Among probability samples a distinction is made between simple random sampling and random sampling methods that involve sampling within subgroups of the population. When sampling is done within subgroups the number of individuals drawn from each subgroup may be proportionate to the relative size of those subgroups. "Thus a stratum containing 10 percent of the population will also receive 10 percent of the sample." When proportionate random sampling is employed every individual within the population has an equal chance of being selected.

While a number of surveys, on sex attitudes and habits have been conducted in the U.S., they are far rarer in other countries. This article outlines the results of a study along these lines in Great Britain. A national cross-section was asked for their attitudes on sex education, divorce, birth control, prostitution and related matters and a more restricted sample was asked about their personal sex habits. With some notable exceptions, results are consistent with the report given by researcher Alfred Kinsey with the help of his Mass-Observation survey carried out in the U.S. and Great Britain. The report was subjected to much criticism in Great Britain. Expert counsel was felt to be particularly necessary in a sex survey and a board of assessors was therefore assembled to give general advice throughout. Funds available for the survey were strictly limited and of the many ideas put forward for the survey clearly only a few could be implemented. Thus, it was finally decided that a national cross-section questionnaire was an essential background for comparison and therefore a direct interview was prepared and asked of a sample of 2,000 people.

The article explores some ways of controlling bias in mail surveys. Mail surveys present a peculiar sampling problem, the questionnaire is returned by only a fraction of all recipients and one must know whether this fraction constitutes an unbiased sub-sample of the original mailing list. So far, the only known way of proving that no bias exists is the solicitation of additional response waves, sometimes supported by personal interviewing of a sample of those who did not respond to any of the successive mailings. It is based on the idea that the difference between respondents and nonrespondents to the first mailing of a questionnaire is reproduced in miniature in the replies of the early respondents as compared with those of the later ones. The importance of timing accurately the delay from date of receipt to date of mailing cannot be overemphasized to those who would use the proposed technique, for this reason as any error in timing serves to distort the true rank order, which reduces the possibility of finding the bias for which people are testing, and this in turn may result in a verdict of no bias and unwarranted complacency.

The authors look at the problem of response error in research. Some of the work that has been done on response error are reviewed, and areas which have not received the proper attention are pointed out. For a number of years research people have been interested in the problem of estimating sampling errors. For a time, the attainment of reasonable accuracy in this area seemed to be sufficient, but gradually it was realized that errors arising in the non sampling area were equally important. One of the pioneering studies on memory error was that done in the early 1930's by Read Bain at Miami University. He was interested in getting data from his students which might be of value in the discussion of vocational problems. This stability of response is not a problem limited solely to formalized types of questionnaires, but also occurs in the area of observation. The Institute of Human Relations at Yale University has used motion pictures to study technical problems which cannot be solved adequately by data from life situations.

This article focuses on the socio-economic status and political socialization patterns of families of student activists. Two themes appear to dominate the literature on student activism. First of all, the majority of studies were conducted between 1964 and 1966, a time when student protest was beginning and was confined to a few "elite" universities. Secondly, these studies are remarkably consistent in their findings. It is only recently that data have become available on activists at "non-elite" universities. Although the results of these studies appear to contradict much of the earlier research, the paucity of research in this area renders the conclusions rather tenuous. The purpose of the study was to determine whether student activists in the given sample are recruited from liberal, affluent family backgrounds, as earlier studies have indicated, or, if the growth of activism has effectively eliminated the class base found earlier. The study was conducted on the campus of Central Michigan University during the fall semester of 1970. The sampling procedure consisted of both random and non-random samples, comprising 212 students. The data were gathered through the use of written questionnaires.