Search

The paucity of published material on the coding of open-ended questions material, and particularly on the relative advantages of field coding and office coding, may justify this report of a very simple and limited experiment that bears on two aspects of the coding problem. The decision on whether to field code or office code a particular open-ended question is affected by a number of factors. If the aim is only to measure the frequency with which a few previously determined answers come up in a population, one may start with the presumption that field coding may be more efficient. If, on the other hand, one starts out with no specific aim other than to get a complete picture of the responses a question evokes, it is obviously necessary to build the code only when all the returns are in the office. The discussion in the report assumes that accuracy of field and office coding are equal.

This article focuses on proceedings of the organization American Association for Public Opinion Research. The sixth annual conference of the Association was held between June 19-25, 1951 in Princeton, New Jersey. While the outpouring of public opinion researchers may have been partly due to the convenient location and the other attractions of Princeton, and to the efforts of an unusually hard working committee on arrangements, it is also true that the community of public opinion researchers is rapidly growing, and the Association is becoming more and more the common meeting ground for all the practitioners. Chairman of National Production Authority, David Wallace, in his introductory remarks, pointed out that for purposes of analysis, the process of communication might be broken into six component parts. They are, the nature of the communicator, the climate of the communications situation, the content of the communication, the method of communication, the interpretation of the communicatee and lastly the response of the communicatee.

This article focuses on plant-community public relations. Although corporations have become increasingly concerned with the development of favorable community attitudes toward plants, little has been known about the audience toward which a public relations program should be addressed. In this article, the authors point to the small segment of a community population which is actively interested in corporation activities and suggest that public relations efforts, to be effective, should be addressed to this group. The concern which corporations are developing for local public relations has, of course, an important ethical basis. Executives are citizens themselves and are interested in being part of an organization that not only plays fair with the local community but also assumes some of the obligations for civic improvement. A man does not like to work for an organization in which he does not believe, his ego is bound up to some extent with the corporation's reputation. A reputation for good corporate citizenship is a real factor in organization morale.

In this article, the author produces a history of opinion research, by tracking down present trends into the future, and simultaneously predicting the things to come. The article is based on an address delivered to the American association for Public Opinion Research. The author is inspired from a document from professor Mcquilkin DeGrange of Dartmouth College, entitled, "A History of the Future," a compound of social commentary and science fiction. The author propose to engage in a few serious and a few not-so-serious predictions confined to the field of public opinion surveying, with the serious and non-serious so carefully mixed. In 1951, human researchers were emerging from the great controversy over whether the probability sample was a revolutionary development in polling science, doomed immediately to drive out all baser unquotable forms of sampling, but hardly hope to reach with the primitive tools then available. One result of the sampling ferment in the early 1950's was a type of approach that was dubbed possibility sampling by one of its supporters. A great methodological of the early fifties was of interviewer bias. Everyone experimented with new methods of evaluating interviewer performance and at the same time the personnel testers were called in to help devise better methods for selecting interviewer material.

The prospect that opinion surveys may be able to assist the peoples of different countries to communicate with and to understand each other is one that has fired the imaginations of many students of public opinion. It is generally acknowledged, however, that this goal is as difficult so achieve, as it is inspiring to contemplate. The symposium discusses some of the methodological problems which arose when research firm Elmo Roper and the periodical "Time" attempted to obtain comparable data simultaneously in ten countries. With developments after World War II, the problem of measuring public opinion cross-culturally has reached urgent dimensions. For an essential of lasting international understanding is full and accurate knowledge of what the people in partner countries think. It is regrettable that the methodology needed for accurate appraisal of international opinion remains largely to be developed. In the execution of a polling operation covering several countries a great many difficulties arise which do not usually occur in studies done in one single country. The points taken up in this article are some of the hurdles encountered in coordinating studies in four European countries namely Sweden, France, Switzerland and Italy, in the Time survey. It should not necessarily be assumed that any of the solutions the surveyors decided to use could be considered as permanent. The purpose of this report is merely to point out some of the problems and to promote discussion on the ways to overcome them. Three difficulties mentioned in this article are: a clear description of the universe with which the surveys deal, the problems of translation of questionnaires, and the element of timing.

The article focuses on coding of answers to open-ended questions, which is one of the important processes in market and public opinion research. Coding is a process often involving some fairly sophisticated judgments, but at the same time it is a routine operation that rapidly loses interest for many people. Few studies on coding reliability have been reported. The study reported here is a small one, and it should be viewed as little more than a first step in research on coding reliability. It was carried on in the coding room of the Elmo Roper office during the intervals between peak load activity, and was designed to give us some feel for the magnitude of coding errors in two or three more or less typical coding situations. A more thorough-going study would extend the range of questions covered as well as increase the number of different coders involved. Three different open-ended questions were used in the experiment. These coding experiments are of course, not sufficient to be conclusive. They encourage us, however, to feel that coding error is at least under some sort of control, and that the amount of time devoted to train the coders and to instruct them in detail with regard to each job is paying dividends.

To design and carry out government opinion research in such a way as to make it bear upon problems of the administrator is not easy, especially under the conditions of war-time work in Washington. There has been much fumbling, considerable waste of time and money, and not a little essentially useless research done while social scientists were learning how to behave in the new ways circumstances require of them. Opinion research people in the government find a gradual but steadily increasing demand for their services, which is encouraging and helps promote their morale. At the same time, however, they find certain aspects of the Washington research environment continually frustrating, and they have many moments when a return to academic life seems irresistibly attractive. The expected research findings need not be, and seldom are, completely determining in relation to the decision between alternatives but that decision must be affected significantly by them; otherwise there is an operational problem but no corresponding research problem.