Conditions for the successful integration of Human and Organizational Factors (HOF) in the nuclear safety analysis

This communication presents some elements which come from the experience feedback at CEA about the conditions for the successful integration of HOF in the nuclear safety analysis. To point out some of these conditions, one of the concepts proposed by Edgar Morin to describe the functioning of "complex" systems: the dialogical principle has been used. The idea is to look for some dialogical pairs. The elements of this kind of pair are both complementary and antagonist to one another. Three dialogical pairs are presented in this communication. The first two pairs are related to the organization of the HOF network and the last one is related to the methods which are used to analyse the working situations. The three pairs are: specialist - non-specialist actors of the network, centralized - distributed human resources in the network and microscopic - macroscopic levels of HOF methods to analyse the working situations. To continuously improve these three dialogical pairs, it is important to keep the differences which exist between the two elements of a pair and to find and maintain a balance between the two elements of the pairs.     

Spatial and genetic distribution of a north American termite, <Emphasis Type="Italic">Reticulitermes flavipes</Emphasis>, across the landscape of Paris

Urbanization often negatively impacts biological diversity. Some organisms, however, have traits that are preadapted to urban environments and thus may thrive. Reticulitermes flavipes is one such organism. Indigenous to the Eastern US, it has been introduced into multiple countries. In France, R. flavipes causes major damage to buildings. Although Paris is one of the country’s most infested cities, the factors determining R. flavipes’ distribution and propagation are poorly understood. Using data on termite occurrence, termite genetics, and environmental variables, our study aims to identify factors that explain the distribution and spread of R. flavipes within Paris. First, we explored the association between several environmental variables and the termite’s distribution pattern in Paris; since 2000, a total of 2106 infestations have been recorded. Second, we inferred termite population and colony genetic structure at 66 sample locations using 9 microsatellite loci. Third, we used least-cost models and partial Mantel tests to study the effects of environmental variables on the termite’s population genetic structure. Our analyses revealed that building-related variables were significantly associated with termite infestations and could thus help explain the termite’s spatial distribution pattern. Furthermore, railway networks also explain termite propagation and genetic patterns. Additionally, we found that the termite’s spread is likely driven by budding dispersal, which may be constrained by buildings and roads. Even if budding dispersal could facilitate R. flavipes’ spread in urban areas, it cannot explain the termite’s distribution in Paris all on its own. Indeed, termite propagation seems to be significantly driven by anthropogenic activities.     

Joint covariate selection and joint subspace selection for multiple classification problems

We address the problem of recovering a common set of covariates that are relevant simultaneously to several classification problems. By penalizing the sum of ℓ ₂ norms of the blocks of coefficients associated with each covariate across different classification problems, similar sparsity patterns in all models are encouraged. To take computational advantage of the sparsity of solutions at high regularization levels, we propose a blockwise path-following scheme that approximately traces the regularization path. As the regularization coefficient decreases, the algorithm maintains and updates concurrently a growing set of covariates that are simultaneously active for all problems. We also show how to use random projections to extend this approach to the problem of joint subspace selection, where multiple predictors are found in a common low-dimensional subspace. We present theoretical results showing that this random projection approach converges to the solution yielded by trace-norm regularization. Finally, we present a variety of experimental results exploring joint covariate selection and joint subspace selection, comparing the path-following approach to competing algorithms in terms of prediction accuracy and running time.     

Robust cointegration testing in the presence of weak trends,with an application to the human origin of global warming

Standard tests for the rank of cointegration of a vector autoregressive process present distributions that are affected by the presence of deterministic trends. We consider the recent approach of Demetrescu et al. (2009 Demetrescu, M., Lütkepohl, H., Saikkonen, P. (2009). Testing for the cointegrating rank of a vector autoregressive process with an uncertain deterministic trend term. Econometrics Journal 12:414–35.[Crossref], [Web of Science ®] , [Google Scholar]) who recommend testing a composite null. We assess this methodology in the presence of trends (linear or broken) whose magnitude is small enough not to be always detectable at conventional significance levels. We model them using local asymptotics and derive the properties of the test statistics. We show that whether the trend is orthogonal to the cointegrating vector has a major impact on the distributions but that the test combination approach remains valid. We apply of the methodology to the study of cointegration properties between global temperatures and the radiative forcing of human gas emissions. We find new evidence of Granger Causality.     

Preparative study regarding the implementation of a muscular fatigue model in a virtual task simulator

Muscle fatigue is considered as one of the major risk factors for developing musculoskeletal disorders. The aim of this project was to select an adequate fatigue assessment model for an implementation in Dassault Systemes digital human modeling software. A review of existing MET models has been done resulting in a decision to use the extended Ma's model (2010). In this project, only shoulder and elbow joints have been tested and more subjects will be necessary for further validation. The model has been compared to several endurance time (ET) static studies. Two dynamic experiments were also performed by two different subjects. The results showed that because of the inter-individual variability, a simple prediction curve or value, can't well predict individual measured ET (or task failure). There is a need for a chart representation which also shows standard deviation (SD) range. Considering the SD range, the results were included in the prediction. Thus, this range may help the human factors expert to nuance the prediction results while considering environment factors and some realities specific to the industry.     

Infant Neurodevelopment is Affected by Prenatal Maternal Stress: The QF2011 Queensland Flood Study

Research shows that prenatal maternal stress (PNMS) negatively affects a range of infant outcomes; yet no single study has explored the effects of stress in pregnancy from a natural disaster on multiple aspects of infant neurodevelopment. This study examined the effects of flood‐related stress in pregnancy on 6‐month‐olds' neurodevelopment and examined the moderating effects of timing of the stressor in gestation and infant sex on these outcomes. Women exposed to the 2011 Queensland (Australia) floods in pregnancy completed surveys on their flood‐related objective and subjective experiences at recruitment and reported on their infants' neurodevelopment on the problem solving, communication, and personal–social scales of the Ages and Stages‐III at 6 months postpartum (N = 115). Interaction results showed that subjective flood stress in pregnancy had significantly different effects in boys and girls, and that at high levels of stress girls had significantly lower problem solving scores than boys. Timing of the flood later in pregnancy predicted lower personal–social scores in the sample, and there was a trend (p < .10) for greater objective flood exposure to predict lower scores. PNMS had no effect on infants' communication skills. In conclusion, differential aspects of maternal flood‐related stress in pregnancy influenced aspects of 6‐month‐olds' neurodevelopment.     

A characterization of the single-peaked domain

We provide in this paper two properties that are both necessary and sufficient to characterize the domain of single-peaked preference profiles. This characterization allows for a definition of single-peaked preference profiles without using an ad hoc underlying order of the alternatives and also sheds light on the structure of single-peaked profiles. Considering the larger domain of value-restricted preference profiles (Sen, Econometrica 34:491–499, 1966) we also provide necessary and sufficient conditions for a preference profile to be single-caved or group-separable. Our results show that for single-peaked, single-caved and group-separable profiles it is sufficient to restrict to profiles containing of either three individuals and three alternatives or two individuals and four alternatives.     

Application of ordinal correspondence analysis for submerged aquatic vegetation monitoring

The European Water Framework states that macrophyte communities (seaweeds and seagrass) are key indicators of the ecological health of lagoons. Furthermore, the restoration of these communities, especially the Zostera meadows, is one of the main objectives of the Berre lagoon restoration plan. Consequently, a monitoring programme of the main macrophyte species still present in the lagoon was initiated in 1996. This monitoring resulted in a sequence of 11 spatially structured annual tables consisting of the observed density of these species. These tables are processed in this study. First, we specify the principles of Beh's ordinal correspondence analysis (OCA), designed for ordered row/column categories, and compare this method to classical correspondence analysis (CA). Then, we show that OCA is straightforwardly adaptable for processing a sequence of ordered contingency tables like ours. Both OCA and CA are afterwards used to reveal and test the main patterns of spatio-temporal changes of two macrophyte species in the Berre lagoon: Ulva and Zostera. The results we obtained are compared and discussed.     

À l'école de la détention : quelques aspects de la socialisation professionnelle des surveillants de prison

Their first period of training in penitentiaries provided an opportunity for studying the occupational socialization of French prison wardens. Although the latter soon learn certain aspects of life behind the bars, their talk manifests a mistrust of both the training situation and prison in general. By combining class work outside the detention center and experiences with handling inmates inside, this training presents new recruits with the contradictory admonitions of trainers and corrections officers. A double bind thus complicates learning this trade: whether to follow suggestions about imposing order or to negotiate; whether to accept the penitentiary method or, like the others, consider it to be inefficient or even dangerous. This occupational socialization spawns dilemmas and contradictions; it does not peacefully transmit a predefined occupational role.     

Fully efficient estimation of coefficients of correlation in the presence of imputed survey data

Marginal imputation, that consists of imputing items separately, generally leads to biased estimators of bivariate parameters such as finite population coefficients of correlation. To overcome this problem, two main approaches have been considered in the literature: the first consists of using customary imputation methods such as random hot‐deck imputation and adjusting for the bias at the estimation stage. This approach was studied in Skinner & Rao 2002 . In this paper, we extend the results of Skinner & Rao 2002 to the case of arbitrary sampling designs and three variants of random hot‐deck imputation. The second approach consists of using an imputation method, which preserves the relationship between variables. Shao & Wang 2002 proposed a joint random regression imputation procedure that succeeds in preserving the relationships between two study variables. One drawback of the Shao–Wang procedure is that it suffers from an additional variability (called the imputation variance) due to the random selection of residuals, resulting in potentially inefficient estimators. Following Chauvet, Deville, & Haziza 2011 , we propose a fully efficient version of the Shao–Wang procedure that preserves the relationship between two study variables, while virtually eliminating the imputation variance. Results of a simulation study support our findings. An application using data from the Workplace and Employees Survey is also presented. The Canadian Journal of Statistics 40: 124–149; 2012 © 2011 Statistical Society of Canada