Suchergebnisse

Summary This paper reviews research situations in medicine, epidemiology and psychiatry, in psychological measurement and testing, and in sample surveys in which the observer(rater or interviewer) can be an important source of measurement error. Moreover, most of the statistical literature in observer variability is surveyed with attention given to a notational unification of the various models proposed. In the continuous data case, the usual analysis of variance (ANOVA) components of variance models are presented with an emphasis on the intraclass correlation coefficient as a measure of reliability. Other modified ANOVA models, response error models in sample surveys, and related multivariate extensions are also discussed. For the categorical data case, special attention is given to measures of agreement and tests of hypotheses when the data consist of dichotomous responses. In addition, similarities between the dichotomous and continous cases are illustrated in terms of intraclass correlation coefficients. Finally, measures of agreement, such as kappa and weighted‐kappa, are discussed in the context of nominal and ordinal data. A proposed unifying framework for the categorical data case is given in the form of concluding remarks.

3-wave panel data collected during the 1968 presidential election campaign in Fla & NC are analyzed on the basis of the partial candidate preference information available for some R's & the complete information obtained from others. A supplemented marginals model based on the general linear approach to categorical data analysis is extended to the situation involving incomplete time data. The final statistical models show no politically significant effects associated with the campaign (time) on the level of citizens' preferences for Hubert Humphrey, Richard Nixon, or George Wallace for 7 race & occupation classifications. Even though statistically significant time effects occurred among several classes & states, they had no important political consequences. Black voters overwhelmingly supported Humphrey, but whites supported either Nixon or Wallace. Nixon derived support from white collar professionals & retired workers; Wallace gained support from all other classes (small businessmen, farmers, skilled, & unskilled blue collar workers). There was only one case where the campaign made a politically significant difference: among unskilled blue-collar workers in NC Wallace lost his plurality to Nixon. With one limited exception, the models did not show significant state effects; evidently, different state "political cultures" are not important for the stability of voter preferences. The study determined the level of choice for each major candidate, not the proportion of the actual vote & of course there is no constant relationship between voting & preference for different occupational & racial groups. 6 Tables. H. Dorian.

The authors report levels of intercommunity variation in rates of cigarette smoking observed during the course of an evaluation of the influence of three mass media campaigns on the initi ation of adolescent cigarette smoking. Despite careful design and statistical adjustment for a set of 10 known sociodemographic and personality correlates of adolescent cigarette smoking, high intercommunity variation precluded the detection of media campaign effects in this study. The authors discuss the general implications of high intercommunity variation for studies using the individual as the unit of analysis when the unit of randomization is the community.

Once a health promotion program has proven to be effective in one or two initial settings, attempts may be made to transfer the program to new settings. One way to conceptualize the transference of health promotion programs from one locale to another is by considering the programs to be innovations that are being diffused. In this way, diffusion of innovation theory can be applied to guide the process of program transference. This article reports on the development of six questionnaires to measure the extent to which health promotion programs are successfully disseminated: Organizational Climate, Awareness-Concern, Rogers's Adoption Variables, Level of Use, Level of Success, and Level of Institutionalization. The instruments are being successfully used in a study of the diffusion of health promotion/tobacco prevention curricula to junior high schools in North Carolina. The instruments, which measure the four steps of the diffusion process, have construct validity since they were develóped within existing theories and are derived from the work of previous researchers. No previous research has attempted to use instruments like these to measure sequentially the stages of the diffusion process.

A survey of 2,056 12- to 14-year-olds from 10 southeastern U.S. cities showed that Blacks and girls spent more time with television and radio than Whites and boys did; television use declined with age, whereas radio use increased. Access to cable television and videocassette recorders was not related to frequency of television or radio use. Lack of access to parents, either because the mother was employed or because no father was in the home, generally increased the time that adolescents spent with radio and television.

ABSTRACTThe purpose of this study was to investigate cross‐sectional and longitudinal associations between hearing acuity and tooth loss in 1156 US veterans taking part in the Veterans Affairs' Normative Aging (NAS) and Dental Longitudinal (DLS) Studies in the Boston, MA, area. The mean age was 48 years (SD = 8.9), 5.3% were edentulous, and 15.4% had < 17 teeth at baseline. Hearing acuity was determined by puretone, air‐ and bone‐conduction audiometry, and speech discrimination tests at triennial examinations over a 20‐year follow‐up period. Hearing decline was defined as a change from baseline in the average puretone air‐conduction thresholds of ≥ 20 dB at 0.25, 0.5, 1, 2, 3, 4, 6, and 8 kHz. The explanatory variables of interest were change since baseline in dentate status (cut points at < 1, < 17, and < 20 teeth), and in the number of teeth lost (linear). Linear and logistic regression models—which controlled for baseline audiological status, age, air‐bone gap, and otoscopic examination at current visit—showed that subjects who went from having ≥ 17 to < 17 teeth had 1.64 times (95% CI, 1.24‐2.17) as high odds of having hearing decline as those with no change in their dentate status. For every tooth lost since baseline, there was a 1.04 times as high odds (95% CI, 1.02‐1.06) for hearing decline, when additional baseline and time‐varying covariates were taken into account in the model.