While it can be accepted that we as human beings are capable of knowing and assessing ourselves, it is not clear whether such self-assessment information can be accurately reported and used in a predictive role. Self-assessment information was requested of 85 applicants before and after an assessment centre selection program. Those applicants em ployed provided self-assessments of job performance after 12 months on the job. Self-assessments were com pared with organizational evaluations of performance at selection and after 12 months employment. The moderating effects of age, gender, intelligence, social desirability, work experience, and education were considered for each relationship. Resu Its showed that general ly self-assessment was unrelated to organizational assessments. Post-assessment centre selfassessments were significantly lower than self-assessments requested prior to the centre. Dave George is an In dustrial/Organizational PsyctiologisI in Itie Manage ment Services Department of Peat Mar%vick, in Wcllmgtot\,Nortti, New Zealand. He has published articles in the areas ot personnel selection and as sessment. 1-te has lectured in Industrial /Organizational Psychotogy al Massey Univer sity and has been involved in the design of selection and ap praisal programs, sur\'ey and research projects for a range of government and private sector organizations. continuiä A number of recent review^s have shown that self-assessment is a valid source of predictor information. Shrauger and Osberg (1981) considered those studies which compared self-a.ssessment with other assessment pro cedures in a range of settings and concluded that other assessment methods did not show clearly superior validity when compared to self-assessment. Mabe and West (1982) focused on the effects of individual characteristics and measurement conditions on the relationship between various ability and performance measures and self-assessment. Potential individual char acteristics associated with accurate self-evaluation included relatively higher intelligence, high achievement status, and internal locus of control. In general terms, accurate self-evaluator tended to be the more intelligent subjects, individuals high in achievement status, and those with an internal locus of control. In addition, Mabe and West (1982) showed that when certain favorable conditions of measurement were present there was a substantial increase in the mean validity coefficient. The measurement conditions which were the main contributors to explained variance were relative self-evaluation, expectations of validation, anonymity instruction and self-evaluation experience. The increase from .00 (no measurement conditions present), to .64 (any four measurement conditions) implied that many of the finding reported in the literature may be affected by the way in which .self-assessment data were collected.A recent review by Harris and Schaubroek (1988) focused on the relationship between self, peer, and supervisory assessments in mainly work settings. They reported a relative...

Il existe une controverse sur la valeur de l'autoévaluation dans la sélection professionnelle. Peu d‘études appliquées sont actuellement disponibles. Cette recherche s'intéresse à l'amélioration que l'autoévaluation procure à la validité prédictive des actuelles procédures de sélection. On a pris en considération diverses variables susceptibles d'intervenir (le sexe, l’âge, l'expérience professionnelle, l'information sur l'emploi, le niveau culturel, l'expérience de l'autoévaluation et les jugements absolus ou non). Un échantillon de 176 candidate á un travail saisonnier ont fourni des autoévaluations de leur performance au travail qui furent comparées à trois critères: le niveau de performance, la rotation et l'absentéisme. Des hypothèses particulières relatives à l'influence des jugements absolus ou non et des différents niveaux de l'information sur l'emploi n'ont pas élé validées. L‘âge intervient significativement dans les critères “niveau de la performance” et “absentéisme”. Sur ces deux dimensions, les candidats les plus âgés (30 ans et plus) ont réagi aux autoévaluations différemment des plus jeunes. Les autoévaluations étaient généralement optimistes. Controversy exists regarding the value of self‐assessment in personnel selection with few studies reporting research in applied settings. This investigation considers the addition to predictive validity that self‐assessment offers an existing selection procedure. A number of potential moderating effects (gender, age, work experience, education, job information, self‐assessment experience, and comparative versus non‐comparative ratings) were considered. A sample of 176 applicants for seasonal work completed self‐assessments of work performance which were compared to three types of work performance criteria: appraised performance, turnover and absence. Specific hypotheses examining the effect of comparative versus non‐comparative ratings and the effect of different levels of job information were not proved. Age demonstrated the only significant moderating effect in relation to appraised performance and absence criteria. For both these criteria older applicants (30+ years) treated self‐assessments differently from younger applicants. Self‐assessment ratings were generally inflated.

Personnel practitioners use a variety of methods to improve the accuracy of selection decisions amongst which the weighted application form is one of the best available. Predictive studies using application form information involve either differential item weighting, which ignores relationships between items, or multivariate techniques, which weight items according to the unique contribution they make to prediction. In the absence of evidence indicating which approach is better, the present study compared the efficiency of the horizontal percent method and linear discriminant analysis in an applied setting. Application forms completed by 159 nurse trainees were weighted using both methods to predict success in the first year of a training course leading to comprehensive registration. Results showed that linear discriminant analysis was more accurate than the horizontal percent method in classifying cases in the validation sample, but on cross validation there was no significant difference in classification accuracy.

Interoperability among the various departments of an organization and the ancillary units has been geographically distributed and the coordination has become more complicated. Web-based services with service-oriented paradigm enable flexible and dynamic interoperation of autonomous software units. Service oriented architecture can be realized with web services as the low level building blocks and higher level services can be composed with the Business Process Execution Language. In this paper, effective interaction among the various departments of an organization and the ancillary units is studied and a solution based on composite web services is proposed. Intricacies involved, issues connected with this type of composition will be studied with reference to engineering job outsourcing environment. This study provides orchestration mechanism which leads to composition of new higher level abstract services.

This study investigated the computerization of the Weighted Application Form (WAF). A discriminant analysis was performed on the application form responses of 281 nursing course applicants. Computer software was then developed incorporating the classification function from the discriminant analysis procedure. In the second part of the study 45 subjects were divided into three groups and entered eight sets of nursing course application form data. T h e ' computerized ' group received instructions on how to do this from the screen, the ' written ' group from a manual and the ' verbal ' group verbally from the experimenter. Time taken to complete the task and the number of errors made were recorded. Results indicated that group influenced time takcn on the task and that group did not exert an influence on errors made on each question. As a result of the experimental work and responses to a questionnaire, it was concluded that the ' computerized ' version of the software was the option that should be adopted for computer software involving the WAF.