Non‐parametric Estimation of Extreme Risk Measures from Conditional Heavy‐tailed Distributions

In this paper, we introduce a new risk measure, the so‐called conditional tail moment. It is defined as the moment of order a ≥ 0 of the loss distribution above the upper α‐quantile where α ∈ (0,1). Estimating the conditional tail moment permits us to estimate all risk measures based on conditional moments such as conditional tail expectation, conditional value at risk or conditional tail variance. Here, we focus on the estimation of these risk measures in case of extreme losses (where α ↓0 is no longer fixed). It is moreover assumed that the loss distribution is heavy tailed and depends on a covariate. The estimation method thus combines non‐parametric kernel methods with extreme‐value statistics. The asymptotic distribution of the estimators is established, and their finite‐sample behaviour is illustrated both on simulated data and on a real data set of daily rainfalls.     

Prevalence of Sexual Abuse in Childhood: Some Critical Methodological Reflections

This article attempts to analyze the prevalence rates of sexual abuse in childhood reported in comparative empirical studies, giving an overview of a selection of representative studies reporting high rates of abuse and comparing it to an overview of studies reporting lower rates. Extremely discrepant rates of childhood sexual abuse are reported in a number of empirical studies conducted in this field over the past 26 years, particularly those looking beyond prevalence to study the links between this type of trauma and future motherhood. To shed light on the disparities in the reported prevalence rates, the article reviews the principal methodological biases that have contaminated the findings. The authors then offer several suggestions for improvement that might be implemented in future studies.     

Parental Divorce and Romantic Attachment in Young Adulthood: Important Role of Problematic Beliefs

In the current study 45 university students with either divorced or continuously married parents were surveyed about their romantic attachment, positive emotionality, depressive symptomology, self-esteem, and, when applicable, their retrospective beliefs about their parents’ marital dissolution. Findings revealed that parental divorce did not predict attachment insecurity, depression, or low self-esteem. In fact, adult children of divorced parents (ACDP) reported increased compassion, awe, enthusiasm, and perspective taking. Among ACDP, a composite factor representing increased fear of abandonment, peer rejection, and maternal blame was positively associated with adult attachment anxiety, even while controlling for parental conflict and divorce-related socioenvironmental disruption. Results are discussed in terms of their support of a complex understanding of the long-term effects of parental divorce, and in their inconsistency with a purely pathogenic model of parental divorce.     

Nonverbal Social Sensing in Action: Unobtrusive Recording and Extracting of Nonverbal Behavior in Social Interactions Illustrated with a Research Example

Nonverbal behavior coding is typically conducted by “hand”. To remedy this time and resource intensive undertaking, we illustrate how nonverbal social sensing, defined as the automated recording and extracting of nonverbal behavior via ubiquitous social sensing platforms, can be achieved. More precisely, we show how and what kind of nonverbal cues can be extracted and to what extent automated extracted nonverbal cues can be validly obtained with an illustrative research example. In a job interview, the applicant’s vocal and visual nonverbal immediacy behavior was automatically sensed and extracted. Results show that the applicant’s nonverbal behavior can be validly extracted. Moreover, both visual and vocal applicant nonverbal behavior predict recruiter hiring decision, which is in line with previous findings on manually coded applicant nonverbal behavior. Finally, applicant average turn duration, tempo variation, and gazing best predict recruiter hiring decision. Results and implications of such a nonverbal social sensing for future research are discussed.     

The Effects of Misconceptions on the Properties of Friedman's Test

Friedman's test is a widely used rank-based alternative to the analysis of variance (ANOVA) F-test for identifying treatment differences in a randomized complete block design. Many texts provide incomplete or misleading information about when Friedman's test may be appropriately applied. We discuss the assumptions needed for the test and common misconceptions. We show via simulation that when the variance or skew of the treatment distributions differ, application of Friedman's test to detect differences in treatment location can result in Type I error probabilities larger than the nominal α, and even when α is unaffected, the power of the test can be less than expected.     

Deaths from Alcohol and Violence in Moscow: Socio-economic Determinants

Objective: To examine the association between accidental, violent and alcohol related adult mortality in the Russian capital and socio-economic status characteristics such as educational status, occupational group and marital status.Data and methods: individual death records for Moscow City for the years 1994 and 1995, for 86121 deaths between the ages of 20–59. Proportional mortality analysis was used to compare trends for alcohol related deaths and accidental and violent deaths (representing 5 and 28% of all deaths in this age group), with multiple controls consisting of deaths from cancer at various sites.Results: The probability of death from alcohol related diseases increased as education level decreased, with those men failing to complete secondary education over two and a half times as likely to die from these causes than men with higher education. Blue collar workers were also much more likely to die from these causes than white collar workers. Marriage had a marked protective effect for both men and women.Conclusion: Despite 75 years of official egalitarian ideology, there are marked socio-economic differentials in mortality in Russia. For the causes of death analysed, socio-economic mortality differentials were greater for women than for men.     

Estimation of the Weibull tail-coefficient with linear combination of upper order statistics

We present a new family of estimators of the Weibull tail-coefficient. The Weibull tail-coefficient is defined as the regular variation coefficient of the inverse failure rate function. Our estimators are based on a linear combination of log-spacings of the upper order statistics. Their asymptotic normality is established and illustrated for two particular cases of estimators in this family. Their finite sample performances are presented on a simulation study.     

Risk Target: An Interactive Context Factor in Risk Perception

The effect of specification of the target on risk evaluation was examined. A whole set of hazards, covering most of the domains, were considered: common individual hazards, outdoor activities, medical care, public transportation, energy production, pollutants, sex, deviance, and addictions. Three human targets were introduced: personal health risk (including personal risk of death), health risk for people in the country, and health risk for people in the world. The basic design was a between-subjects design. The first hypothesis was that risk judgments made in the "world" condition should be higher than risk judgments made in the "country" condition, and risk judgments made in this condition should be higher than risk judgments made in the "personal" condition. This is what was observed. The second hypothesis was that the target effect should differ as a function of the kind of hazards considered. This also is what was observed. In two domains--pollutants, and deviance, sex, and addictions--the target effect was important. It corresponded to about one-tenth of the response scale. In the four remaining domains, the target effect was unimportant or absent.     

Weibull tail-distributions revisited: A new look at some tail estimators

In this paper, we propose to include Weibull tail-distributions in a more general family of distributions. In particular, the considered model also encompasses the whole Fréchet maximum domain of attraction as well as log-Weibull tail-distributions. The asymptotic normality of some tail estimators based on the log-spacings between the largest order statistics is established in a unified way within the considered family. This result permits to understand the similarity between most estimators of the Weibull tail-coefficient and the Hill estimator. Some different asymptotic properties, in terms of bias, rate of convergence, are also highlighted.