Encoding dissimilarity data for statistical model building

We summarize, review and comment upon three papers which discuss the use of discrete, noisy, incomplete, scattered pairwise dissimilarity data in statistical model building. Convex cone optimization codes are used to embed the objects into a Euclidean space which respects the dissimilarity information while controlling the dimension of the space. A “newbie” algorithm is provided for embedding new objects into this space. This allows the dissimilarity information to be incorporated into a smoothing spline ANOVA penalized likelihood model, a support vector machine, or any model that will admit reproducing kernel Hilbert space components, for nonparametric regression, supervised learning, or semisupervised learning. Future work and open questions are discussed. The papers are:     

Gambling and Adverse Life Events Among Urban Adolescents

This study explored the cross sectional association between adverse life events and gambling in a sample of 515 urban adolescents (average age 17, 55% male, 88% African American). Approximately half of the sample had gambled in the past year (51%); 78% of the gamblers gambled monthly and 39% had a gambling-related problem. On the other hand, 88% of the sample had experienced at least one life event in the past year, and those experiencing events tended to live in more disadvantaged neighborhoods. The mere acknowledgement of experiencing a stressful life event in the past year (yes/no) was not associated with an increase in odds of being a gambler, with gambling more frequently, or with having a gambling problem. However, when the context of the event was considered, an association was found between directly experiencing threatening and deviant/violent types of events and frequent gambling (OR > 2). Additionally, the probability of being a gambler increased as the number of events experienced increased (aOR = 1.07, 95% CI = 1.01, 1.13, P = 0.013), but problems among gamblers were not associated with the number of events experienced (aOR = 1.01, 95% CI = 0.92, 1.11, P = 0.876). During adolescence, life events appear to be connected more with the frequency of gambling rather than with problems related to gambling.     

Make Time for Physical Activity or You May Spend More Time Sick!

This work estimates the association between various types of physical activities and physical health, mental health and health service utilization. Specifically, we consider participation in housework, manual, walking, sports/exercise and total physical activities to a moderate level. We view this as an important contribution given that governments usually recommend a total level of activity that their citizens should achieve to be healthy, rather than a particular activity per se. Our results suggest that participation in any of these activities predicts a lower probability of poor health roughly to the same level. For mental health, our estimates highlight that those who participate in (and do more of) any activity have better mental health when compared to those who do not. The associations are higher for women. The associations are also higher for manual activities and sports/exercise when compared to other activities. The analysis also reveals some lower utilization of in-patient, out-patient visits and GP services among individuals who are moderately physically active.     

A multi-index model for quantile regression with ordinal data

In this paper, we propose a quantile approach to the multi-index semiparametric model for an ordinal response variable. Permitting non-parametric transformation of the response, the proposed method achieves a root-n rate of convergence and has attractive robustness properties. Further, the proposed model allows additional indices to model the remaining correlations between covariates and the residuals from the single-index, considerably reducing the error variance and thus leading to more efficient prediction intervals (PIs). The utility of the model is demonstrated by estimating PIs for functional status of the elderly based on data from the second longitudinal study of aging. It is shown that the proposed multi-index model provides significantly narrower PIs than competing models. Our approach can be applied to other areas in which the distribution of future observations must be predicted from ordinal response data.     

Multivariate thin plate spline estimates for the posterior probabilities in the classification problem

A nonparametric estimate for the posterior probabilities in the classification problem using multivariate thin plate splines is proposed. This method presents a nonpararnetric alternative to logistic discrimination as well as to survival curve estimation. The degree of smoothness of the estimate is determined from the data using generalized crossvalidation.     

The Quality of Life in Taiwan

The AsiaBarometer survey of 1,006 respondents shows that in Taiwan, people have access to modern utilities and digital media, signs of materialistic achievement, and yet are more concerned with physical security and financial safety than with personal growth. Regardless of their demographic backgrounds and value priorities, the Taiwanese, like other Confucian publics, are most satisfied with the interpersonal life sphere and least satisfied with the public life sphere. Their satisfaction levels concerning various life domains affect their sense of well-being more than does their prioritization of values. Assessments of material and nonmaterial life domains contribute to their sense of well-being more than those of interpersonal and public domains do. Access to modern utilities and a high income, however, detract from their sense of well-being. As is the case elsewhere, a better set of objective life circumstance does not necessarily make for a greater quality of life in Taiwan. The government, therefore, should seek to provide more than economic goods and services.     

The Impact of Temporary Labor Migration on Mexican Children's Educational Aspirations and Performance1

This article examines how temporary U.S. labor migration by family members and by students affects the educational aspirations and performance of those same students growing up in Mexican migrant communities. Labor migration affects these children in two ways. First it brings remitted U.S. earnings into the household which allows parents to provide more education for their children and reduce the need for children's labor. Higher incomes are also associated with numerous factors that improve the general well‐being of children, as reflected in various indicators including higher school grades. Labor migration also has negative impacts on children. In addition to family stress and behavioral problems with adolescents due to parental and sibling absence, migration provides an example of an alternative route to economic mobility. Children growing up in migrant households have access to information and social networks that reduce their likelihood of migration failure should they choose this alternative to the Mexican labor market. We analyze a unique data set from a stratified random sample of 7600 grammar, junior high, and high school‐level students in a state capital, a large town, and 25 rural communities in a Mexican migrant‐sending state. We find that high levels of U.S. migration are associated with lower aspirations to attend a university at all academic levels. We find, however, a positive relationship between U.S. migration and grades. We conclude that while U.S. migration provides financial benefits that allow children to continue schooling and perform well, it may also reduce the motivation to attain above‐average years of schooling.     

The Role of Social Contact in the Experience of Discrimination: Implications for Subjective Health Among Marriage Migrant Women in South Korea

Several studies have found that various social contexts could serve a protective role for minorities facing discrimination. One of these contexts is frequency of social contact, which is known to be positively associated with health outcomes. The purpose of this study was to examine the frequency of social contact with various social groups as a protective factor against the negative association between perceived discrimination and subjective health. Using data from the 2009 Korean National Survey of Multicultural Families, a total of 27,083 marriage migrant women were analyzed via hierarchical multiple regression. After controlling for socio‐demographic factors and Korean proficiency, the experience of discrimination was found to be significantly associated with lower levels of subjective health (B = ?0.22; P < 0.001). More frequent meetings with one's family of origin (B = 0.02; P < 0.01) and in‐laws (B = 0.06; P < 0.001) were associated with higher levels of subjective health, while more frequent meetings with neighbors were negatively associated (B = ?0.02; P < 0.001). More frequent contact with one's family of origin (B = 0.03; P < 0.05), rather than with other groups, such as friends from their country of origin, in‐laws, and neighbors, seemed to reduce the negative influence of discrimination on subjective health.