A Comparison of Bayesian and Sampling Theory Inferences in a Probit Model

HE PROBIT MODEL IS A POPULAR DEVICE for explaining binary choice decisions in econometrics. It has been used to describe choices such as labor force participation, travel mode, home ownership, and type of education. These and many more examples can be found in papers by Amemiya (1981) and Maddala (1983). Given the contribution of economics towards explaining such choices, and given the nature of data that are collected, prior information on the relationship between a choice probability and several explanatory variables frequently exists. Bayesian inference is a convenient vehicle for including such prior information. Given the increasing popularity of Bayesian inference it is useful to ask whether inferences from a probit model are sensitive to a choice between Bayesian and sampling theory techniques. Of interest is the sensitivity of inference on coefficients, probabilities, and elasticities. We consider these issues in a model designed to explain choice between fixed and variable interest rate mortgages. Two Bayesian priors are employed: a uniform prior on the coefficients, designed to be noninformative for the coefficients, and an inequality restricted prior on the signs of the coefficients. We often know, a priori, whether increasing the value of a particular explanatory variable will have a positive or negative effect on a choice probability. This knowledge can be captured by using a prior probability density function (pdf) that is truncated to be positive or negative. Thus, three sets of results are compared:those from maximum likelihood (ML) estimation, those from Bayesian estimation with an unrestricted uniform prior on the coefficients, and those from Bayesian estimation with a uniform prior truncated to accommodate inequality restrictions on the coefficients.

Does workforce participation empower women? Micro-level evidence from urban Bangladesh

Empirical studies on the impact of women’s paid jobs on their empowerment and welfare in the Bangladesh context are rare. The few studies on the issue to date have all been confined to the garment workers only although studies indicate that women’s workforce participation in Bangladesh has increased across-the-board. Besides, none of these studies has made an attempt to control for the non-working women and/or applied any statistical technique to control for the effects of other pertinent determinants of women’s empowerment and welfare such as education, age, religion and place of living. This study overcomes these drawbacks and presents alternative assessments of the link between women’s workforce participation and empowerment on the basis of survey data from the two largest cities in Bangladesh. While the generic assessment indicates that women’s paid jobs have positive implications for women’s participation in decisions on fertility, children’s education and healthcare as well as their possession and control of resources, the econometric assessment negates most of these observations. Women’s education, on the other hand, appears to be more important than their participation in the labour force. The study underlines the fact that by omitting other relevant explanatory variables from the analysis, the previous studies might have overestimated the impact of women’s paid work on their empowerment. Among other things, the paper also highlights the importance of women’s job category, religion and regional differences for women’s empowerment.

The Limits of Rational Choice: New Institutionalism in the Test Bed of Central Banking Politics in Australia

This paper tests the explanatory capacities of different versions of new institutionalism by examining the Australian case of a general transition in central banking practice and monetary politics: namely, the increased emphasis on low inflation and central bank independence. Standard versions of rational choice institutionalism largely dominate the literature on the politics of central banking, but this approach (here termed RC1) fails to account for Australian empirics. RC1 has a tendency to establish actor preferences exogenously to the analysis; actors' motives are also assumed a priori; actor's preferences are depicted in relatively static, ahistorical terms. And there is the tendency, even a methodological requirement, to assume relatively simple motives and preference sets among actors, in part because of the game theoretic nature of RC1 reasoning. It is possible to build a more accurate rational choice model by re-specifying and essentially updating the context, incentives and choice sets that have driven rational choice in this case. Enter RC2. However, this move subtly introduces methodological shifts and new theoretical challenges. By contrast, historical institutionalism uses an inductive methodology. Compared with deduction, it is arguably better able to deal with complexity and nuance. It also utilises a dynamic, historical approach, and specifies (dynamically) endogenous preference formation by interpretive actors. Historical institutionalism is also able to more easily incorporate a wider set of key explanatory variables and incorporate wider social aggregates. Hence, it is argued that historical institutionalism is the preferred explanatory theory and methodology in this case.

Bridging the gap between different statistical approaches: An integrated framework for modelling

This paper proposes a template for modelling complex datasets that integrates traditional statistical modelling approaches with more recent advances in statistics and modelling through an exploratory framework. Our approach builds on the well-known and long standing traditional idea of 'good practice in statistics' by establishing a comprehensive framework for modelling that focuses on exploration, prediction, interpretation and reliability assessment, a relatively new idea that allows individual assessment of predictions. The integrated framework we present comprises two stages. The first involves the use of exploratory methods to help visually understand the data and identify a parsimonious set of explanatory variables. The second encompasses a two step modelling process, where the use of non-parametric methods such as decision trees and generalized additive models are promoted to identify important variables and their modelling relationship with the response before a final predictive model is considered. We focus on fitting the predictive model using parametric, non-parametric and Bayesian approaches. This paper is motivated by a medical problem where interest focuses on developing a risk stratification system for morbidity of 1,710 cardiac patients given a suite of demographic, clinical and preoperative variables. Although the methods we use are applied specifically to this case study, these methods can be applied across any field, irrespective of the type of response.

Moderating effect of allocentrism on the pay referent comparison-pay level satisfaction relationship

Pay referent comparisons (comparisons of one's salary to that of others) such as other-inside (salary of other people in the organisation), other-outside (the market rate), and cost-of- living, have been shown to influence pay level satisfaction. Bordia and Blau (1998) identified family as another referent that had a significant effect on pay level satisfaction in a sample of public and private sector employees in India. The finding was interpreted in view of the importance of family in collectivistic cultures. In the study reported here, the moderating influence of an individual differences variable, allocentrism-idiocentrism (the individual level conceptualisation of collectivism-individualism) on pay referent comparison-pay level satisfaction relationship was investigated. A sample of 146 employees from three public sector organisations in India participated in the study. In line with the predictions, results showed that after controlling for age, tenure, and pay level, pay referent comparisons explained more variance in pay level satisfaction for allocentrics than for idiocentrics. Family and pay level were stronger explanatory variables of pay level satisfaction for allocentrics and idiocentrics, respectively, while cost of living was a significant explanatory variable for both sub-groups.

A Bayesian hierarchical model for categorical longitudinal data from a social survey of immigrants

The paper investigates a Bayesian hierarchical model for the analysis of categorical longitudinal data from a large social survey of immigrants to Australia. Data for each subject are observed on three separate occasions, or waves, of the survey. One of the features of the data set is that observations for some variables are missing for at least one wave. A model for the employment status of immigrants is developed by introducing, at the first stage of a hierarchical model, a multinomial model for the response and then subsequent terms are introduced to explain wave and subject effects. To estimate the model, we use the Gibbs sampler, which allows missing data for both the response and the explanatory variables to be imputed at each iteration of the algorithm, given some appropriate prior distributions. After accounting for significant covariate effects in the model, results show that the relative probability of remaining unemployed diminished with time following arrival in Australia.

Testing alternative models for the conservation of koalas in fragmented rural-urban landscapes

Predicting the various responses of different species to changes in landscape structure is a formidable challenge to landscape ecology. Based on expert knowledge and landscape ecological theory, we develop five competing a priori models for predicting the presence/absence of the Koala (Phascolarctos cinereus) in Noosa Shire, south-east Queensland (Australia). A priori predictions were nested within three levels of ecological organization: in situ (site level) habitat (< 1 ha), patch level (100 ha) and landscape level (100-1000 ha). To test the models, Koala surveys and habitat surveys (n = 245) were conducted across the habitat mosaic. After taking into account tree species preferences, the patch and landscape context, and the neighbourhood effect of adjacent present sites, we applied logistic regression and hierarchical partitioning analyses to rank the alternative models and the explanatory variables. The strongest support was for a multilevel model, with Koala presence best predicted by the proportion of the landscape occupied by high quality habitat, the neighbourhood effect, the mean nearest neighbour distance between forest patches, the density of forest patches and the density of sealed roads. When tested against independent data (n = 105) using a receiver operator characteristic curve, the multilevel model performed moderately well. The study is consistent with recent assertions that habitat loss is the major driver of population decline, however, landscape configuration and roads have an important effect that needs to be incorporated into Koala conservation strategies.

Contradiction of quantum mechanics with local hidden variables for quadrature phase amplitude measurements

We demonstrate a contradiction of quantum mechanics with local hidden variable theories for continuous quadrature phase amplitude (position and momentum) measurements. For any quantum state, this contradiction is lost for situations where the quadrature phase amplitude results are always macroscopically distinct. We show that for optical realizations of this experiment, where one uses homodyne detection techniques to perform the quadrature phase amplitude measurement, one has an amplification prior to detection, so that macroscopic fields are incident on photodiode detectors. The high efficiencies of such detectors may open a way for a loophole-free test of local hidden variable theories.

The Modified Card Sorting Test: Test-retest stability and relationships with demographic variables in a healthy older adult sample

Objectives. To investigate the test-retest stability of a standardized version of Nelson's (1976) Modified Card Sorting Test (MCST) and its relationships with demographic variables in a sample of healthy older adults. Design. A standard card order and administration were devised for the MCST and administered to participants at an initial assessment, and again at a second session conducted a minimum of six months later in order to examine its test-retest stability. Participants were also administered the WAIS-R at initial assessment in order to provide a measure of psychometric intelligence. Methods. Thirty-six (24 female, 12 male) healthy older adults aged 52 to 77 years with mean education 12.42 years (SD = 3.53) completed the MCST on two occasions approximately 7.5 months (SD = 1.61) apart. Stability coefficients and test-retest differences were calculated for the range of scores. The effect of gender on MCST performance was examined. Correlations between MCST scores and age, education and WAIS-R IQs were also determined. Results. Stability coefficients ranged from .26 for the percent perseverative errors measure to .49 for the failure to maintain set measure. Several measures were significantly correlated with age, education and WAIS-R IQs, although no effect of gender on MCST performance was found. Conclusions. None of the stability coefficients reached the level required for clinical decision making. The results indicate that participants' age, education, and intelligence need to be considered when interpreting MCST performance. Normative studies of MCST performance as well as further studies with patients with executive dysfunction are needed.

Stereotypic explanatory bias: Implicit stereotyping as a predictor of discrimination

Two experiments examined whether a measure of implicit stereotyping based on the tendency to explain Black stereotype-incongruent events more often than Black stereotype-congruent events (Stereotypic Explanatory Bias or SEB) is predictive of behavior toward a partner in an interracial interaction. In Experiment I SEB predicted White males' choice to ask stereotypic questions of a Black female (but not a White male or White female) in an interview. In Experiment 2 the type of explanation (internal or external attribution) made for stereotype-inconsistency was examined. Results showed that White participants who made internal attributions for Black stereotype-incongruent behavior were rated more positively and those who made external attributions were rated more negatively by a Black male confederate. These results point to the potential of implicit stereotyping as an important predictor of behavior in an interracial interaction. (C) 2002 Elsevier Science (USA). All rights reserved.

A comparison of lipid variables as predictors of cardiovascular disease in the Asia Pacific Region

PURPOSE: Many guidelines advocate measurement of total or low density lipoprotein cholesterol (LDL), high density lipoprotein cholesterol (HDL), and triglycerides (TG) to determine treatment recommendations for preventing coronary heart disease (CHD) and cardiovascular disease (CVD). This analysis is a comparison of lipid variables as predictors of cardiovascular disease. METHODS: Hazard ratios for coronary and cardiovascular deaths by fourths of total cholesterol (TC), LDL, HDL, TG, non-HDL, TC/HDL, and TG/HDL values, and for a one standard deviation change in these variables, were derived in an individual participant data meta-analysis of 32 cohort studies conducted in the Asia-Pacific region. The predictive value of each lipid variable was assessed using the likelihood ratio statistic. RESULTS: Adjusting for confounders and regression dilution, each lipid variable had a positive (negative for HDL) log-linear association with fatal CHD and CVD. Individuals in the highest fourth of each lipid variable had approximately twice the risk of CHD compared with those with lowest levels. TG and HDL were each better predictors of CHD and CVD risk compared with TC alone, with test statistics similar to TC/HDL and TG/HDL ratios. Calculated LDL was a relatively poor predictor. CONCLUSIONS: While LDL reduction remains the main target of intervention for lipid-lowering, these data support the potential use of TG or lipid ratios for CHD risk prediction. (c) 2005 Elsevier Inc. All rights reserved.