A Statistical Framework for Hypothesis Testing in Real Data Comparison Studies

In computational sciences, including computational statistics, machine learning, and bioinformatics, it is often claimed in articles presenting new supervised learning methods that the new method performs better than existing methods on real data, for instance in terms of error rate. However, these claims are often not based on proper statistical tests and, even if such tests are performed, the tested hypothesis is not clearly defined and poor attention is devoted to the Type I and Type II errors. In the present article, we aim to fill this gap by providing a proper statistical framework for hypothesis tests that compare the performances of supervised learning methods based on several real datasets with unknown underlying distributions. After giving a statistical interpretation of ad hoc tests commonly performed by computational researchers, we devote special attention to power issues and outline a simple method of determining the number of datasets to be included in a comparison study to reach an adequate power. These methods are illustrated through three comparison studies from the literature and an exemplary benchmarking study using gene expression microarray data. All our results can be reproduced using R codes and datasets available from the companion website http://www.ibe.med.uni-muenchen.de/organisation/mitarbeiter/020_professuren/boulesteix/compstud2013.     

Transitions to adulthood in contexts of economic crisis and post-recession. The case of Argentina

The possibilities that are available to youth in contemporary Western societies are highly sophisticated and related mostly to the transition between secondary school and higher education, the first job placement, independent housing, lifestyle construction and the consumption of goods produced by cultural industries in a globalist world. Secondary education represents a main factor in terms of socialisation, training, access to employment, citizenship and youth identity construction, making it a key for public policies aimed at youth in Latin America. The article presents the initial results of a research project on transition processes from secondary school to employment in Argentina. The findings are based on statistical information and the empirical analysis of data gathered in two very different economic and social contexts: the first was one of the worst social and economic crisis in the country's history (1999–2003), and the second was a long period of economic growth and increased social welfare (2004–2012). The data are presented in response to the following question: How do different economic contexts influence the transitions of youth in different social class contexts? By answering this question, the significant role of public policies in generating better living conditions for youth becomes clear.     

Cost of Living Variation,Nonmetropolitan America,and Implications for the Supplemental Poverty Measure

Population Research and Policy Review - Poverty scholarship in the United States is increasingly reliant upon the Supplemental Poverty Measure (SPM) as opposed to the Official Poverty Measure of...     

University Students’ Attitudes Toward Same-Sex Parenting and Gay and Lesbian Rights in Portugal

The purpose of this study was to explore university students’ attitudes toward same-sex parenting and toward gay and lesbian rights. A total of 292 participants, aged between 18 and 27 (M = 21) responded to a questionnaire measuring attitudes toward parenting by gay men and lesbians, gay and lesbian rights, and beliefs about the etiology of homosexuality. Results revealed that the majority of students were against gay and lesbian parenting, gay and lesbian equal rights, and believed that homosexuality has a social/environmental basis. It was found that sexual prejudice is highly prevalent in Portuguese university students, and implications of these findings are discussed.     

Materialism in Pedagogy Students in Chile

Many studies have used Richins and Dawson’s (J Consum Res 19: 303–316, 1992) Material Values Scale (MVS), applying it to different types of populations that exhibit a particular psychometric behavior, and showing little stability in their factorial structure. In the present study, 1,070 pedagogy students from the northern, central and southern regions of Chile answered the MVS. This sample was randomly divided in two. Using the first sub-sample (N = 539), an exploratory factorial analysis was carried out, from which a structure of nine items was grouped into two factors called “Social Success” and “Personal Happiness”, which presented adequate reliability. Later, with the second sub-sample (N = 531), the factorial structure indicated above was put to the test through a confirmatory factorial analysis. The data from the model show that the scale contains 8 items in total, grouped into two dimensions. The factorial loads are significant at the level of 1 %, which indicates that the 2-factor structure can be confirmed. Finally—using the proposed structure—the presence of the students’ material values was evaluated.     

Life Domain Satisfactions as Predictors of Overall Life Satisfaction Among Workers: Evidence from Chile

This article examines the subjective antecedents of life satisfaction of workers. Adopting a ‘bottom-up’ perspective, we assessed the unique influence that satisfaction with multiple life domains have on evaluative judgments of overall life satisfaction. Based on a nationwide sample of 530 Chilean workers, we simultaneously tested the effects of seven life domain satisfactions that have been consistently included in extant models of life satisfaction and subjective well-being. These were satisfaction with health, financial situation, social relationships, one’s self-worth, leisure-time, family, and work. Having controlled for age and gender, results showed that satisfaction with one’s financial situation was the dominant predictor of overall life satisfaction of workers, with a weight of .36. Satisfaction with family, work, and health had effects of .25, .14, and .14, respectively. Interestingly, satisfaction with one’s self-worth, leisure-time, and social relationships did not have statistically significant effects on life satisfaction, although the first two showed t values near the critical value.