Reply to comments on “artificial neural networks: an econometric perspective“

We briefly discuss the so called pseudo-GLS estimator in a standard linear regression model with nonsperical disturbances, and conclude that the potentiality for applications is higher than originally assumed by Fiebig Bartels and Krämer (1996).     

Crime and Punishment in Canada: A Statistical Test of the “Conventional Wisdom”*

Trois propositions fondamentales qui traduisent les impressions populaires sur le crime et ses chatiments sont mises a l'epreuve dans une analyse regressive a l'aide des moindres carres. Voici les resultats obtenus: (a) le taux general de criminalite n'a pas augmente durant les annees 1950-66 si on met de cote les infractions mineures; (b) durant la meme periode les jugements ne furent pas moins severes et les disparites dans les sentences ne se sont pas amenuisees; (c) 1'augmentation du taux des offenses mineures est directement associee a l'accroissement de la force policiere. On cherche a expliquer pourquoi les croyances populaires persistent meme si elles sont refutees par les faits. Three propositions which express the conventional wisdom about crime and punishment are tested with least squares regression analysis. The results are that the over-all crime rate has not increased between 1950 and 1966 (except for trivial infractions); sentences have not become more lenient, nor have disparities in sentencing decreased throughout the same period. Increases in the rates for the less serious offences are positively correlated with increases in the size of the police force. Reasons for the persistence of belief in the conventional wisdom in the face of contradictory evidence are offered.     

Sociology and social policy: from a love-hate relationship with the state to cynicism and pragmatism

Les sociologues québécois ont été au coeur de ce que Ton a appelé la Révolution tranquille. lis ont inspiré des réformes, animé des projets et joué le rôle d'éminences grises auprés du gouvernement. Cependant, au milieu des années '70, plusieurs sont devenus de plus en plus critiques de l'intervention de l'État en matiére de politiques sociales: non seulement L'État n'a-t-il pas réussi à redistribuer les ressources des plus riches vers les plus pauvres, mail il a aussi paralysé l'action des groupes communautaires et des autres mouvements sociaux. Le présent article recense les publications sur l'État et les politiques sociales depuis 1975. Les auteurs concluent que les sociologues québécois sont de moins en moins intéressés par l'État, concentrant davantage leur attention sur les groupes en besoin.
Quebec's sociologists have been at the heart of the so-called Revolution tranquille. They have inspired reforms, animated projects and served as 'grey eminences' to the government. Yet, by the mid 1970s, many had grown intensely critical of state intervention in social policies: not only was the state incapable of really redistributing resources from the rich to the poor, but it was also paralysing the action of community groups and other social movements. In this article, the francophone literature on the state and social policy since 1975 is reviewed. The authors conclude that Québec sociologists have become less fixated on the state, and are beginning to turn their attention more and more to specific social groups in need.     

1.5 Generation Internal Migration in the U.S.: Dispersion from States of Immigration?1

The issue of immigrant spatial concentration and the possibilities for immigrant dispersion through migration features in at least three interrelated debates about immigration. First, the ethnic enclave literature centers on the question of whether spatial concentration improves or harms the economic well‐being of immigrants. Second, spatial assimilation theory links immigrant relocation away from residential enclaves to socioeconomic gains. Although framed at an intra‐urban scale, we suggest that similar assimilation logics infuse thinking and expectations about immigrant settlement and spatial mobility at other scales. And third, immigrant clustering links to anxieties about the threats posed by non‐European origin newcomers to the traditional cultural fabric of the nation. In the current wave of immigration, research on questions of settlement geography and spatial mobility has so far been restricted to the first generation. But as the current wave of immigration matures there is a growing population of adults who are the children of immigrants. This article investigates the migration behavior of these adult children, specifically the 1.5 generation, seeking to answer the question of whether they will remain in the states in which their parent's generation settled or move on. It also assesses whether the out‐migration response of the 1.5 generation in states of immigrant concentration is similar to that of their parent's generation or the U.S.‐born population.     

Important Risk Factors in Home-Removal Decisions: Social Caseworker Perceptions

The current study identified factors that played a significant role in decisions to separate a child from his/her primary caretakers for 51 social service caseworkers. Participants rated and ranked the importance of 35 child risk/well-being factors used in recent child separation dispositions. Results indicated that boundary setting with a perpetrator of abuse and parental motivation played a significant role in decision-making. These preliminary findings suggest the need to prioritize services aimed at the promotion of the non-maltreating parent’s limit setting with the perpetrator of abuse, as well as caretaker motivation/cooperation.Virginia M. DeRoma is affiliated with The Citadel, Psychology Department, Capers Hall, 171 Moultrie, 29409, Charleston, SC, USA. Maria Lynn Kessler is affiliated with Oregon Institute of Technology, Klamath Falls, OR, USA. Ryan McDaniel is affiliated with New Hope for Children, Charleston, SC, USA. Cesar M. Soto is affiliated with Municipal Defense of Children and Adolescents, Chorrillos, Peru.Address correspondence to Virginia M. DeRoma, The Citadel, Psychology Department, Capers Hall, 171 Moultrie, Charleston, SC, 29409.     

Weighing the Evidence: A new method for research on restricted information

Greatly increased concern about the confidentiality of social records now jeopardizes the ability of social scientists to conduct research. A new confidential method is described that can facilitate the work of the researcher and also maintain the privacy of the researched. This method is called Weighing the Evidence or, more simply, the WEIGHT method. The WEIGHT method groups social records according to intervals of interest to the researcher and then uses a simple weighing procedure to estimate the incidence and distribution of cases. Two studies—one using simulated data and one using real data on handgun purchases—suggest that the WEIGHT method is characterized by high levels of both reliability and validity. Ten hypothetical cases are described in which the WEIGHT method would make possible research that would other wise be prevented by the confidential nature of the data involved.This project was assisted by a fellowship from the German Marshall Fund. The authors are indebted to the following people for helpful comments and advice: Donald T. Campbell, Elliot Aronson, Meredith Gould, David Harrington, John I. Kitsuse, Thomas F. Pettigrew, Alan Ritch, Robert Rosenthal, M. Brewster Smith, and Candace West.