Suchergebnisse

A formal analysis of the theories of the spiral of silence & of pluralistic ignorance. Five hypotheses about the relationships among the variables in each theory of how individuals perceive PO are tested, & the points of success & failure for each noted. Finally, a general argument is made linking the spiral of silence & theories of the perception of PO to the analysis of the general class of social choice situations where people's expectations influence the outcome. The spiral of silence provides a number of different tools for analyzing the buildup of expectations in such settings. 6 Tables, 2 Figures. AA.

Most of the interesting theories of PO change can be thought of as statistical models that make formal predictions about percentages or percent differences. Statistical criteria are proposed based on the percent difference models of James A. Davis ('Analyzing Contingency Tables with Linear Flow Graphs' in Sociological Methodology, San Francisco: Jossey Bass, 1976) & James E. Grizzle, C. Frank Starmer, & G. Koch ('Analysis of Categorical Data by Linear Models,' Biometrics, 1969, 25, 489-504). This approach is illustratively applied to some examples from previously published studies. Certain substantive theories of change are more parsimonious than others in terms of their statistical predictions; these theories should be accepted or rejected first. 4 Tables, 1 Figure. Modified HA.

This essay looks at the various uses which have been made of public opinion data in explaining and predicting patterns of racial residential segregation. Public opinion is central to understanding the process of housing discrimination because housing market behavior is difficult to monitor and police in any official or bureaucratic fashion. Current poll results show that racial attitudes or preferences for segregated neighborhoods by whites (or by blacks) are not the central stumbling blocks in desegregating American neighborhoods. Rather, shifts in the nature of the housing market in an area reflect responses to a broader, more complex set of neighborhood processes. Factors causing neighborhood change vary greatly in their effects depending on the kind of neighborhood "at risk" and the amount of integration which is likely to occur. Simple calculations of the future of integrated neighborhoods or the amount of segregation due to preferences are in error unless they take these complexities into account.

PO is central to understanding the housing-discrimination process because the housing market is difficult to monitor or police in any official or bureaucratic fashion. A look at various uses of PO data in explaining & predicting residential segregation patterns shows that racial attitudes or preferences for segregated neighborhoods by whites or blacks are not the major problem in desegregating American neighborhoods. A broader, more complex set of processes is the shift in the housing market. Depending on the neighborhood, the type of integration that is likely to occur & the factors that cause change may vary greatly. These complexities must be taken into account in calculating the future of integration & the amount of segregation due to preference. Hope of change depends on: diffusing accurate information about the market in American neighborhoods; taking steps to ensure that neighborhoods will not be damaged by desegregation; & providing an affirmative, moral leadership at the national level. 3 Figures. Modified HA.

As part of the 1974 Amalgam Survey, the National Opinion Research Center (NORC) conducted an experiment to check the ability of a R to self-code his or her occupation accurately. 3 forms of the self-coding question were used, X,Y,Z; randomly distributed in equal thirds, in the multistaged probability sample of the noninstitutional, continental US population. In each case, the R's were asked to code their own occupation into 1 of 10 standard census categories. As a check, the R's were also asked to describe in writing what they did. Those responses were coded into 3-digit occupational codes (US Census, CLASSIFIED INDEX OF INDUSTRIES AND OCCUPATIONS, Washington DC, GPO, 1971). Self-coding is cheaper & quicker than other methods, so to test reliability, forms X, Y, & Z were designed to be useful in analyzing 4 accuracy factors: form of the question, certainty of the R, education of the R, & sex of the R. X & Y are similar, except that Y breaks the 10 choices down into 2 large categories--white collar, & blue collar & farm. Z is more detailed, listing typical jobs in each code. Cross-classifying the coded responses vs the NORC/Census codes yielded 3 10 X 10 contingency tables. Overall, X showed 75% of the cases within 1 cell of the diagonal, Y 70%, & Z 83%. Collapsing 2 cells, operative & labor, increased the accuracy; a problem arising from terminology. Collapsing the data into white vs blue collar groups will produce a 80-90% accuracy level. X & Y versions of the question contained random errors, while the errors in the Z version tended to result from systematic upgrading of occupation levels. Quasi-independence was tested using a X2 test for interdependence between those cells not on the diagonal. Analysis of the certainty of response question indicated that the response was usually correct, with those who felt 'absolutely certain' responding quite accurately. The tendency to upgrade one's occupational code was not correlated to education. F's were more accurate than M's & some errors were sex specific. Self-coding proved to be accurate enough for most uses, though there are systematic sources of error to be considered in choosing a specific format. 7 Tables, 1 Appendix. S. Lupton.

Different wordings of a question to ascertain religious affiliation are shown to produce different results on the same sample questioned at 2 points in time. Several hypotheses are set forth in an attempt to determine under what conditions people respond differently to the different stimuli. Asking for present religion rather than for religious preference is a more precise way of determining actual religious affiliation. AA.