James A. Davis

The NORC General Social Survey: A User's Guide.

Introduction Content Study Design Sample Design and Weighting Field Procedures Data Preparation and Distribution Data Analysis

A Partial Coefficient for Goodman and Kruskal's Gamma

Abstract Following Goodman and Kruskals interpretation of their coefficient, γA,B, a partial coefficient, γA, B|C is defined as “how much more probable it is to get like than unlike orders in measures A and B when pairs of individuals differing on A and on B and tied on C but unselected on any other measure are chosen at random from the population.” It is shown that this coefficient is a weighted sum of the values of γ in the various strata defined by categories of C, where the weight in stratum i is its proportion of the total pairs which differ on A and B and are tied on C. An empirical example illustrates the calculation of the co-efficient.

New money, and old man/lady and ‘Two's company’: Subjective Welfare in the NORC General Social Surveys, 1972–1982

The nine surveys, dozens of variables, and more than ten thousand cases in the NORC General Social Surveys, 1972–1982, allow one to test a variety of cross-sectional and over-time hypotheses about Subjective Welfare (Happiness). I used discrete multivariate analyses to test five hypotheses: (1) the Economists prediction that Happiness is a function of income; the Sociologists hypotheses that Happiness is a function of (2) rank on various evaluated dimensions and (3) number of social ties; and the Psychologists hypotheses that Happiness is affected by (4) social comparisons and (5) adaptation. None of the five is supported impressively, but three variables emerge as good cross-sectional predictors — Race (Blacks are less Happy, but not necessarily because of discrimination), Marital Status (all categories of nonmarried are less happy) and Financial Change (those whose finances are improving are happier, those who finances have turned for the worse are less happy). When Marital Status and Recent Financial Change are used in a year-to-year social indicator model, fluctuations in the predictors produce significant but small changes in Happiness.

Nonparametric path analysis—The multivariate structure of dichotomous data when using the odds Ratio or Yule's Q

The recent paperback text, Elementary Survey Analysis (1971, James A. Davis, Ed., Prentice-Hall, Englewood Cliffs, N.J.), gives a set of intuitive rules for path analysis of dichotomous variables. In this paper a statistical model using odds ratios is used to evaluate this system of rules more precisely. The major conclusions are as follows. (1) With nonrecursive systems and no higher-order interaction effects, the rules are quite valid and can even be simplified a little. (2) In the presence of interactions and strong skews on the variables, the rules are not correct. (3) The rules, as stated, are not valid for recursive systems.

Locals and Cosmopolitans in American Graduate Schools

ALTHOUGH the American educational ladder proceeds in continuous graded steps from pre-school (kindergarten and nursery school) through the twelve grades of primary and high school and four years of undergraduate college, to post-college &dquo;graduate&dquo; and &dquo;professional&dquo; studies, there is a distinct break between high school and college, in terms of social organization. In particular, faculty recruitment and training are very different in the two layers. The elementary and high school classroom teacher typically &dquo;majors in education&dquo; and has received considerably more training in educational methods and techniques than in the subject matter which will be taught. Conversely, the college and graduate school teacher (save, of course, teachers of education) has almost universally completed a graduate degree (master’s or Doctor’s) in a particular field (English, Chemistry, French, Economics, etc.) with little or

THE PARENTAL FAMILIES OF AMERICANS IN BIRTH COHORTS 1890 - 1955: A CATEGORICAL,LINEAR EQUATION MODEL ESTIMATED FROM THE NORC GENERAL SOCIAL SURVEY*

A linear, categorical statistical model with five variables, Fathers Education, Fathers occupation, Size of Place at Age 16, Mothers Employment, and Total Siblings, is estimated with data from the 1972–1978 NORC General Surveys to describe the parental families of Americans in the birth cohorts 1890 to 1955. The major substantive conclusions are: (1) The educational level of American parents increased at an accelerating rate from 1890 to 1955, (2) The proportion growing up in farm homes declined steadily. Farm fathers were less well educated but the educational difference grew steadily smaller, (3) The proportion of Americans growing up in cities of 50 000 or larger increased steadily, the trend being similar in both educational levels, (4) Metropolitan families increased at an accelerating rate, the acceleration being due to the acceleration in education attainment, (5) Farm families decreased at an approximately constant rate because two opposite trends — acceleration in Education and declining association between Education and Farm cancelled each other out, (6) Town Families —non farm families living in cities under 50 000 increased throughout the period, but faster before 1930 than afterwards, (7) Metropolitan families had consistently more children and more employment of mothers than Town families; farm families were slower in experiencing the trend toward working wives; farm families were about the same as town families in decreasing rates of fertility, so the urban/farm gap in fertility remained constant, (8) at the turn of the century higher status mothers were more likely to have small families and less likely to work. After 1910 the pattern changed, as better educated families opted for the pattern of working mothers and fewer children. By the birth cohort of 1955 the education difference in fertility had grown considerably while the education difference in maternal employment had reversed.

A generation of attitude trends among US householders as measured in the NORC General Social Survey 1972–2010

Two generations (1972-1976 and 2006-2008) are compared using 43 replicated attitudes in the NORC General Social Survey. The report describes the generational changes (primarily liberal), weighs the causal impact of rising educational levels (liberal), cohort replacement (liberal) and period effects (mildly conservative). It argues that this long term causal mechanism is slowly eroding.