Publication Type: Paper/Poster Proposal
Abstract: The objective of the BRFSS is to collect uniform, state specific data on preventive health practices and risk behaviors that are linked to chronic diseases, injuries, and preventable infectious diseases in the adult population. Data are collected from a random sample of adults (one per household) through a telephone survey.

This study evaluates demographic non-response biases in the BRFSS, assesses the effect of the BRFSS design weights and post-stratification on the survey’s representativeness, and examines the results of different weighting strategies on calculated estimates of risk factors and prevalences. The analysis examines the effect of the current design weights and post-stratification in the BRFSS for the years 2000-2003 to evaluate the effect of the weights on the estimates produced and their variance. The project also evaluates different potential post-stratification methodologies and their effect on the same estimates. Data will be examined at the state level for trends over the four survey years.

The results of this analysis are being compared to external population and public health data sources as a measure of accuracy of the survey and appropriateness of the current and proposed weights. The risk factors analyzed include smoking, overweight/obesity, health insurance, health status, physical activity, drinking, and nutrition. Demographic characteristics included in the analysis are race/ethnicity, age, sex, educational attainment, income, and employment. This analysis is invaluable for determining the success of the data collection operations of the BRFSS and for evaluating and potentially reducing demographic bias in the survey.

Publication Type: Poster
Abstract: Exploring the relationship between personality and gender needs to be done on a deeper level than is currently occurring. The current study examines relationships between the big five personality factors and conformity to gender norms in men and women to better understand within-gender variances.

Publication Type: Conference Paper/Unpublished Manuscript
Abstract: The ever closer collaboration between many intergovernmental organizations (IGOs) and non-governmental organizations (NGOs) is empirically well described but poorly theorized. The aim of this paper is to deliver a general theoretical framework for analyzing emergent patterns of cooperation between IGOs and NGOs, which may be used to generate hypotheses or guide comparative studies. The starting point is a rationalist conception of organizational actors as purposeful but resource-dependent. The paper then combines a ‘resource exchange perspective’ from organizational sociology with the model of a policy cycle from comparative politics. The result is an analytical framework that allows us to identify incentives for, as well as obstacles to, IGO-NGO cooperation along all phases of the policy cycle. In a concluding section the limits of this model and the underlying rationalist assumptions are discussed.

Publication Type: Conference Paper/Unpublished Manuscript
Review Method: Peer Reviewed
Abstract: Selecting the number of factors to include in a factor analysis model is a fundamental and long-standing question in psychometrics and other fields that utilize related models. In political science, this question of dimensionality arises in factor analysis models of survey responses, item-response theory models of legislative and judicial behavior, classifications of welfare state regimes, assigning democracy ratings, and other situations. However, these models unfortunately require the number of dimensions to be specified in advance, rather than left unspecified and subsequently inferred from the estimates. The estimator presented in this paper remedies this problem by going back to theoretical ideas that were put forward in the literature before computers were invented but were impossible to implement at the time. With substantively minor modifications, these ideas are feasible today and allow researchers to consistently estimate a factor analysis model whose primary purpose is to infer the number of factors from the output of the model. From there, the researcher can then estimate a different model that takes the dimensionality as given. An empirical example is given that illustrates all the steps in the algorithm.

Publication Type: Conference Paper/Unpublished Manuscript
Abstract: Family factors have been considered important in the explanation of juvenile delinquency. The majority of research findings have identified a range of family factors such as poor parental supervision, harsh and inconsistent parenting, broken home or separation, high conflict homes as well as large family size as ‘high’ risk factors. A number of these findings have been replicated in several contexts in mainly western countries, especially in the United States and the United Kingdom. The consistency in findings regarding some of these family factors in western countries raised an important question as to the universal importance or generality of these factors. In other words, does culture or context matter in the determination of risk factors? This paper presents results of a study investigating the relationship between family factors and delinquency in Ghana. It also compares key findings with previous results obtained in the Cambridge Study of Delinquency Development (CSDD) and Pittsburgh Youth Study (PYS).