Suchergebnisse

Statistically designed sample surveys have enabled pollsters to gauge PO on a wide range of issues. In such surveys, selection of a data collection technique is generally based on 4 criteria: (1) cost; (2) completion time; (3) response rate; & (4) response bias. Typically, more weight is placed on the 1st 3 factors &, as a result, adequate attention has not been given to the latter consideration. The study described looks at one type of response bias-that which results from the use of a specific data collection method. More specifically, this res uses a controlled exp'al design in order to determine whether responses given in a PO polling are influenced by the method used to collect the data. 3 methods are investigated: (1) a mail questionaire (2) a telephone interview; & (3) personal interview. The major finding from the study is that responses are not always independent of the method used to collect the data & that this form of response bias is likely to be a problem in telephone & personal interviews whenever the question being asked is one for which there exists a soc'ly undesirable response. AA.

The use of magnitude estimation in identifying utility functions of housing is explored. Magnitude estimation involves measuring the subjective, or perceived, magnitude of 'real' variables. This approach is demonstrated via the variables which a group of white Mc Coll students would take into account in choosing where to live (N=23 S's aged 18-27). S's were asked to assume that a standard housing situation consisting of 2 persons in a 2-bedroom apartment, costing $80 per month per person, in good condition & located 1 mile from the campus, was to be rated 100 on an arbitrary scale of desirability. Then they were asked to estimate the overall desirability of each of 25 hyp'al situations in which all 5 variables were allowed to vary independently. The values of the regression coefficients obtained suggest that type & quality of housing are of primary importance to the students, followed fairly closely by space & price. Location emerged as a negligible factor in these evaluations, & commuting costs, either in time or money, were not regarded as very important by the R's. A regression using the logarithms of all variables was then performed. Type, space, price, & quality emerged all as about equally important, with location rather more important than was indicated by the linear regression, but still relatively much less signif than the other variables. Non-linear regression is considered preferable for various reasons, eg, it explains almost 33% of the linear equation's unexplained variance. The method of magnitude estimation should be tested further & more studies are needed in order to describe utility functions for all important groups in the society. 1 Table, 4 Figures. M. Maxfield.

A discussion of methodological & theoretical problems re the concept of status crystallization as developed by G. Lensky & E. Jackson (See SA 1019-B3526). The authors' analytical statements often refer to SR's but their measures of crystallization lack a relational perspective. The relational meaning of status crystallization as discussed by G. Homans is discussed in contrast with the individual-oriented theories of Lenski & Jackson. Also, only one dimension, ethnicity, accounts for the variations in rates under study. It is suggested that 'Lenski & Jackson develop their summary measures along more traditional lines: that they add their several dimensions together, one at a time, & then indicate that no single dimension or no single combination of dimensions adequately explains what appears to be an underlying theme of low status crystallization.' This raises the question of whether people participate more or less in groups in reaction to stresses supposedly arising from low status crystallization. G. Jacobs.

A major task of methodology is to analyze why theories succeed or fail. The numerous theories of social change proposed since WWII have largely failed. The following four commonly held postulates of social change theories explain this failure: a coherent social structure; nomological regularity in social phenomena; a meaningful concept of social structure as an intrinsic feature of society; & the identification of main determinants of social change. All of these postulates represent a realistic rather than nominalistic approach to typology. Abandonment of this realistic approach is necessary if a workable scientific theory of social change is to be achieved. W. H. Stoddard.

A SCHEMA FOR CATEGORIZING OBJECTS OF INVESTIGATION IS DEVELOPED & USED AS A BASIS FOR ANALYZING SOME FUNDAMENTAL METHODOLOGICAL ISSUES OF INTEREST TO SOCIAL SCIENTISTS & INTELLECTUAL HISTORIANS. THESE INCLUDE THE IDEA OF A SCIENCE OF SCIENCE, THE METHODOLOGIES OF SCIENCE & HISTORY, & THEIR RELATIONSHIP TO THE SOCIAL SCIENCES. SOCIAL & NATURAL PHENOMENA CAN BE CONSIDERED FROM 3 ASPECTS OF INQUIRY: (1) SUBJECTIVE--THE PERSON THINKING, DOING, OR SAYING SOMETHING, (2) OBJECTIVE--THE RESULTING ACTION, IDEA, ETC, & (3) UTILITY--A SUBJECTIVE OR OBJECTIVE JUDGMENT OF THE PHENOMENA. SOME PHILOSOPHERS OR HISTORIANS OF SCIENCE HAVE PROFESSED A SEPARATION OF SCIENCE FROM THE CONTEXT OF THE TIMES IN WHICH THE SCIENTIFIC THEORY HAS BEEN DEVELOPED; IS THIS VALID DESPITE THE CRITICISM OF THE SOCIAL SCIENCES FOR FRAGMENTATION DUE TO INTERDEPENDENCY WITHIN THE DISCIPLINE? NONFLEXIBLE EMPIRICAL METHODOLOGY CANNOT ACCOUNT FOR CHANGING SOCIAL BACKGROUNDS. THE EFFICACY OF VARIOUS APPROACHES SHOULD BE JUDGED ACCORDING TO THEIR RELEVANCE TO A PARTICULAR STUDY RATHER THAN BASED UPON METHODOLOGICAL DOGMA. 1 TABLE. MODIFIED AA.

In an effort to demonstrate the advantages of certain parametric statistical techniques, J. P. Gibbs' data (see SA 1636/ D5394) on the deterrent efficacy of punishment were re-analyzed using multiple regression & analysis of covariance procedures. Assumptions about the data maintained by other authors were rejected on the basis of standard tests for normality, linearity, & additivity. The results of the regression & covariance analyses suggest that variation in homicide rates among states can be attributed substantially to variability in proportion black in the total pop among states. In addition, the variables measuring legal reactions to crime, severity & certainty of punishment, maintain their association to homicide in a direction consistent with the deterrence hypothesis even after the etiological factor of proportion black is controlled. AA.

The purpose was to test the hypothesis that: response-& acquiescence-set when answering questions concerning personal values are much more powerful among low educated than among highly educated. Since in the developing countries, the gap between high & low educated groups is esp large, the hypothesis should be clearly confirmed in such a setting. By admin'ing a 7-item California F-Scale twice (original form & with items reversed) & in diff orders to groups of U students, We mothers & housemaids, the hypothesis was thoroughly confirmed. The importance of subject motivation to take the test & variation between items to elicit response set were also investigated. The study was conducted in Chile. The implication of the study is to question the wisdom of utilizing diff's in scores on value surveys employing questions of the F-Scale type in order to arrive at conclusions about possible real diff's in values between groups differing in educ. AA.