The effects of language on judgments of universal facial expressions of emotion

Because of the close connection between culture and language, a number of writers have suggested that bilinguals will differ in their behavior because of differences in the degree of assimilation of different cultures in the same individual. We tested this notion by obtaining data from bilingual (English and Hindi) college students in India using a well-studied cross-cultural research paradigm involving emotional perception. Subjects judged universal facial expressions of emotion in two separate sessions, one conducted entirely in English, the other in Hindi. In each session, they judged which emotion was being portrayed, and how intensely. Subjects recognized anger, fear, and sadness more accurately in English than in Hindi. They also attributed greater intensity to female photos of anger when rating in Hindi, but attributed greater intensity to female photos of sadness when rating in English. These findings were discussed in relation to the theoretical connection between culture and language.     

The World Fertility Survey: an appraisal of methodology

"In the first half of the article, a broad account of content and procedures is given. In conduct of individual surveys, the achievements of the World Fertility Survey were based on thoroughness rather than technical superiority. The later aspects of the program, including analysis, archiving, and data dissemination, were more innovative and represent models of excellence for similar future enquiries. In overall terms, the program is judged to be an expensive success. In the second half of the article, two methodological issues are discussed in more detail: the collection of retrospective birth histories and the translation of survey instruments into local languages." Comments by Anthony G. Turner (pp. 768-9), Kweku T. de Graft-Johnson (pp. 769-70), Burton Singer (pp. 771-2), and Joel E. Cohen (pp. 772-4) are appended.     

Complete censuses and samples

"Complete decennial censuses are needed for small areas and other domains. Sample surveys yield diverse and timely data. Censuses can also be combined with samples, and sometimes with data from registers, for diverse estimates that are detailed over both space and time, and hence are timely for small domains. Methods of 'postcensal estimates' for small domains are described. We note uses of censuses for improving samples and of samples for improving censuses, and propose a method for cumulating data from 'rolling' (or rotating) periodic (weekly, monthly or quarterly) samples specifically designed to cover the population in detail over designed spans (annual and quinquennial)."     

Asian immigrants in Canada: some findings from 1981 census

Immigrant characteristics in Canada are analyzed using data from the 1981 census. "The purpose of this paper is to focus on Asian immigrants and to compare them with immigrants born in the United Kingdom, and also to compare them with Canadian-born persons with respect to age, sex, marital status composition, educational attainment, labour force participation, class of worker, occupation and income." (summary in FRE, SPA)     

Boom and bust: housing equity withdrawal and consumption decisions and their impacts on household wealth

The authors examine whether the surge in housing refinance and equity withdrawal generated by the housing boom disproportionately affected older households in terms of decisions on whether to refinance, whether to withdraw equity, and how much to withdraw, and how in turn these decisions affected their household wealth. Using the Federal Reserve's Survey of Consumer Finances, the authors found the highest rates of refinancing and cashing out equity among older (although not necessarily the oldest) households. The authors also found that appreciation in house value, being a baby boomer, and having higher household income were the factors that most increased the probability of refinancing and withdrawing equity. Amounts cashed out were higher among older than among younger households. Ordinary least squares (OLS) estimates showed that amounts withdrawn were increased by price appreciation, financial assets, and income, along with being a "pre-boomer" or "early boomer." Older cohorts displayed larger wealth effects than younger ones, but their withdrawn equity was a smaller percentage of total net worth. A relatively small percentage of respondents' dollars was used for consumption, and most dollars were used for home improvement, other investment, or debt repayment. However, massive equity withdrawal and the subsequent collapse of housing prices have caused housing leverage to soar, putting some homeowners "underwater" and jeopardizing others' retirement preparedness. The ensuing housing finance crisis raises numerous policy issues, from reform of mortgage underwriting, securitization, servicing, and foreclosure procedures, to potential limitations on borrowing against unrealized housing equity gains, to enhanced government transfers for distressed older homeowners, to local and state revenue policy changes.     

A comparative analysis of school-to-work transitions in the European Union1

The present paper describes some aspects of ‘school-to-work’ transition by analyzing the employment situation of individuals as a function of the time elapsed since the completion of education or training. Our perspective is interdisciplinary, comparative and dynamic, with special focus on the patterns in southern European countries. In the literature, most of the studies have had the basic approach of constructing indicators based on retrospective information on the time of first leaving continuous education, and current information on status and characteristics of the person's economic activity – expressing the status of activity as a function of the time elapsed since leaving continuous education. In this approach, essentially cross-sectional (though in part retrospective) information is interpreted as if it pertains to real cohorts. Much of this comparative analysis of school-to-work transitions in EU countries has been based on the EU Labour Force Survey, the 2000 round of which incorporated a special module to collect information on the subject. Our basic approach is to use the longitudinal data from the European Community Household Panel to identify, at the time of each wave, the person's most recently completed education and training, and study this in relation to the person's current employment situation and other characteristics as a function of the time elapsed since that completion. Hence, in form at least, our approach is similar to that of earlier studies based on the LFS, though there are considerable differences in substantive content and statistical methodology resulting from the use of different types of data. We also demonstrate how data from a panel survey may be cumulated over time to obtain a more adequate sample size.     

A noise-based stabilizer for convolutional neural networks

Overfitting occurs when one tries to train a large model on small amount of data. Regularizing a neural network using prior knowledge remains a topic of research as it is not concluded how much prior information can be given to the neural network. In this paper, a novel algorithm is introduced which uses regularization to train a neural network without increasing the dataset. A trivial prior information of a class label is supplied to the model while training. Laplace noise is introduced to the intermediate layer for more generalization. The results show significant improvement in accuracy on the standard datasets for a simple Convolutional Neural Network (CNN). While the proposed method outperforms previous regularization techniques like dropout and batch normalization, it can also be applied with them for further improvement in the performance. On the variants of MNIST, proposed algorithm achieved an average 48% increment in the test accuracy.