RESPONSE RATES AND QUALITY OF DATA: A RE-EXAMINATION OF THE MAIL QUESTIONNAIRE*

Traditionnellement, dans les sciences sociales on a été préoccupé par les techniques destinées à accoître les taux de réponse aux questionnaires distribués par la poste. Ce n'est que rarement qu'on a essayé de déterminer si des taux de réponse élevés avait une relation avec la qualité des données obtenues. Si on considère les taux de réponse d'envinon 60 pour cent obtenus dans les enquêtes faites par entrevues par les centres de recherche commerciaux, on doit réexaminer le problème de l'utilisation des questionnaires en voyés par la poste comme une alternative, même si on ne prend pas en considération la grande différence de coût par sujet entre les entrevues et les questionnaires envoyés par la poste. Dans la recherche discutée ici, le même questionnaire a été envoyé par la poste, dans des conditions différentes, à trois échantillons tirés au hasard. Le taux de réponse pour les trois Cchantillons variaient de 33 à 48 pour cent, mais les données obtenues dans chaque échantillon étaient pratiquement équivalentes. Pour controler ce resultat, un questionnaire semblable a été envoyéà un échantillon tiré d'une autre population. Le taux de réponse à ce questionnaire était juste inférieur à 90 pour cent, mais les données obtenues des premiers et des derniers sujets étaient généralement équivalentes. Traditionally, social scientists have been preoccupied with techniques for increasing response rates to mail questionnaires. Only rarely have there been attempts to ascertain whether high response rates are related to the quality of data obtained. Given current completion rates of about 60 per cent for interviews by commercial survey research centres, the issue of using mail questionnaires as an alternative data-gathering technique requires re-examination, even apart from any consideration of the wide differential between interviews and mail questionnaires in cost per respondent. In the current research, the same questionnaire was mailed under varying conditions to three independent random samples. The response rates for the three samples varied from 33 to 48 percent, but the data obtained from each sample were basically equivalent. As a further check on this finding, an overlapping questionnaire was sent to a sample from another population. The response rate to this questionnaire was just under 90 per cent, but the data from earlier and later respondents were generally equivalent.     

The Social Worker, the Client and the Social Anthropologist

Summary City Centre Project has been working with drifting and homelessyoung people and we considered it important to study not justthe clients, but also the project workers. The discipline andmethods of social anthropology appeared to offer unique advantagesin this attempt However, in the course of our fieldwork, weencountered situations which led us to question the validityof using participant observation with clients, and to explorethe tensions between social workers and academic researchers.We then altered our methods for gathering data, and realizedthat our experience had important implications for researchersconsidering the use of qualitative methods in a social worksetting. This article is our first attempt to generalize fromour particular experience, and to raise questions about powerrelations among the parties to research and action.     

The Objectives of the Probation Service

Summary Two kinds of objective are identified: an organizational objectiveconcerned with efficiency, and a functional objective concernedwith the effectiveness of the service. Increasing responsibilitiesin respect of social enquiry work, parole, after-care and prisonwelfare have been absorbed remarkably smoothly during the 1960s,and even the inevitable administrative changes have been lessdisruptive than might have been anticipated. Any discussionabout the 'effectiveness' of the probation service highlightsthe need to clarify its purposes and its functions. It is arguedthat, although most officers would accept that their primaryfunction is and has traditionally been that of exercising oversightover the offender on behalf of the court and the community,there are unmistakeable signs that pressure will be broughtto bear on the probation service to adopt a more treatment-orientatedapproach. The switch to such a correctional role is bound todisturb the organizational objectives of the service, and raisequestions about the place of casework theory in the trainingand supervision of probation officers     

Cross-validation Bandwidth Matrices for Multivariate Kernel Density Estimation

Abstract.  The performance of multivariate kernel density estimates depends crucially on the choice of bandwidth matrix, but progress towards developing good bandwidth matrix selectors has been relatively slow. In particular, previous studies of cross-validation (CV) methods have been restricted to biased and unbiased CV selection of diagonal bandwidth matrices. However, for certain types of target density the use of full (i.e. unconstrained) bandwidth matrices offers the potential for significantly improved density estimation. In this paper, we generalize earlier work from diagonal to full bandwidth matrices, and develop a smooth cross-validation (SCV) methodology for multivariate data. We consider optimization of the SCV technique with respect to a pilot bandwidth matrix. All the CV methods are studied using asymptotic analysis, simulation experiments and real data analysis. The results suggest that SCV for full bandwidth matrices is the most reliable of the CV methods. We also observe that experience from the univariate setting can sometimes be a misleading guide for understanding bandwidth selection in the multivariate case.