Search results

The purpose of this paper is to discuss and assess the structural characteristics (conceptual utility) of the most popular classification and predictive techniques employed in customer relationship management and customer scoring and to evaluate their classification and predictive precision.

A sample of customers' credit rating and socio‐demographic profiles are employed to evaluate the analytic and classification properties of discriminant analysis, binary logistic regression, artificial neural networks, C5 algorithm, and regression trees employing Chi‐squared Automatic Interaction Detector (CHAID).

With regards to interpretability and the conceptual utility of the parameters generated by the five techniques, logistic regression provides easily interpretable parameters through its logit. The logits can be interpreted in the same way as regression slopes. In addition, the logits can be converted to odds providing a common sense evaluation of the relative importance of each independent variable. Finally, the technique provides robust statistical tests to evaluate the model parameters. Finally, both CHAID and the C5 algorithm provide visual tools (regression tree) and semantic rules (rule set for classification) to facilitate the interpretation of the model parameters. These can be highly desirable properties when the researcher attempts to explain the conceptual and operational foundations of the model.

Most treatments of complex classification procedures have been undertaken idiosyncratically, that is, evaluating only one technique. This paper evaluates and compares the conceptual utility and predictive precision of five different classification techniques on a moderate sample size and provides clear guidelines in technique selection when undertaking customer scoring and classification.

This paper aims to investigate the efficiency and effectiveness of alternative credit‐scoring models for consumer loans in the banking sector. In particular, the focus is upon the financial risks associated with both the efficiency of alternative models in terms of correct classification rates, and their effectiveness in terms of misclassification costs (MCs).

A data set of 630 loan applicants was provided by an Egyptian private bank. A two‐thirds training sample was selected for building the proposed models, leaving a one‐third testing sample to evaluate the predictive ability of the models. In this paper, an investigation is conducted into both neural nets (NNs), such as probabilistic and multi‐layer feed‐forward neural nets, and conventional techniques, such as the weight of evidence measure, discriminant analysis and logistic regression.

The results revealed that a best net search, which selected a multi‐layer feed‐forward net with five nodes, generated both the most efficient classification rate and the most effective MC. In general, NNs gave better average correct classification rates and lower MCs than traditional techniques.

By reducing the financial risks associated with loan defaults, banks can achieve a more effective management of such a crucial component of their operations, namely, the provision of consumer loans.

The use of NNs and conventional techniques in evaluating consumer loans within the Egyptian private banking sector utilizes rigorous techniques in an environment which merits investigation.

The purpose of this paper is to validate the unobtrusive knowledge test (UKT) in a minority population, and examine its potential for limiting stereotype threat.

Study One: (convergent validity): UKT and Wonderlic Personnel Test (WPT) scores were correlated for 131 students. Study Two: (stereotype threat) 202 minority students were placed into one of four groups based on whether or not they were given instructions to elicit stereotype threat, and whether they took the Excellence scale of the UKT or the WPT.

Correlations provided evidence of convergent validity between the Excellence subscale of the UKT and the WPT. The stereotype threat study was inconclusive, with no differences being seen in the threat/non‐threat conditions for the WPT, and higher scores in the threat condition than the non‐threat condition for the UKT.

Unreliability of some scales and low correlations of others with the WPT, lessened the overall UKT's convergent validity.

The need to develop measures of intelligence not subject to adverse impact is clear, and the results of the current research provide justification for further research establishing the properties of the UKT as a selection tool.

This paper offers new evidence of the usefulness of the UKT as a measure of cognitive ability for minority populations, and raises questions about the impact of stereotype threat on the UKT test.

This chapter reviews some of our most recent research as to whether the cognitive performance of reading disabled and poor readers can be separated under dynamic assessment procedures. We describe results related to junior high school students (mean chronological age of 12 years) with reading disabilities, poor readers, and skilled readers. Students were administered intelligence, reading and math tests, and working memory (WM) measures (presented under static and dynamic testing conditions). The results thus far show that: (1) dynamic assessment measures (maintenance scores) contributed unique variance to predicting reading; and (2) poor readers and skilled readers were more likely to change and maintain gains under the dynamic testing conditions than students with reading disabilities. Some discussion was given to developing a valid classification of reading disabilities.

To provide educationalists with an understanding of the key quality issues relating to multiple‐choice tests, and a set of guidelines for the quality assurance of such tests.

The discussion of quality issues is structured to reflect the order in which those issues naturally arise. It covers the design of individual multiple‐choice questions, issues relating to the question bank as a whole, choice of test format, and what can be learned through post‐test analysis. The paper offers practical advice, with an emphasis on maximising test reliability.

It is recognised that considerable expertise and effort is required to undertake a thorough post‐test statistical analysis, but pre‐test quality assurance is relatively straightforward, if labour‐intensive. The question of which is best amongst the various alternative test formats is left open.

The general issue of quality assurance of multiple‐choice tests is surely an important one, yet the author is not aware of any other publication that deals directly with this topic.

The rapid introduction of new technology and the various needs for human resource development are changing the priorities for vocational and technical training. While the fundamental goals for training remain, emphasis is shifting from a focus on content delivery to a recognition of the importance of a systems approach. As a consequence, professionals involved with training are increasingly aware of the need for more rigour in the process by which training is developed, implemented and evaluated. The chief purpose of this monograph is to introduce training and other interested personnel to the Instructional Systems Development (ISD) approach. Collected in this one document is the essential information on the ISD process for the development and conduct of efficient and effective performance‐based training programmes. For those planning a new training programme, this logical and organised approach provides a road map. Furthermore, the procedural steps presented are useful when modifying and revising existing programmes. Instructional Systems Development methodology is presented in five phases: analyse, design, develop, implement and control.

Reviews and describes the results of some psychological tests which have been used with entrepreneurs. The tests have been mainly employed either to explore the nature of the entrepreneur or to assess so‐called significant entrepreneurial characteristics. Key problems in the psychological testing of entrepreneurs relate to varying definitions of the entrepreneur, numerous entrepreneurial characteristics, uncertainty about the significance of entrepreneurial characteristics, and lack of rigour in test development. The results of well validated tests highlight many personality characteristics which could be correlated with characteristics of small firms and business owner‐managers. However, there have been specific measures of entrepreneurial characteristics developed in a search for a sensitive discriminating measure. Unfortunately, many of these tests have been poorly validated and this fails to supply useful information about the nature of entrepreneurs. Gives examples to illustrate some problems in psychological testing and discusses the importance of new approaches.

The purpose of this chapter is to review our findings related to the question “Do outcomes related to dynamic assessment on a cognitive measure predict reading growth?” Our discussion related to the predictive validity of such procedures focused on outcomes related to a battery of memory and reading measures administered over a three-year period to 78 children (11.6 years) with and without reading disabilities (RD). Working memory (WM) tasks were presented under initial, gain, and maintenance testing conditions. The preliminary results suggested that maintenance testing conditions were significant moderators of comprehension and vocabulary growth, whereas probe scores and gain testing conditions were significant moderators of nonword fluency growth. Overall, the results suggested that the dynamic assessment of WM added significant variance in predicting later reading performance.

This research complements the corporate reporting literature by exploring the different types of assurance, which companies are using to bolster the credibility of their integrated and sustainability reports. A composite quality measure is proposed and this study aims to provide evidence on how combined assurance quality (CAQ) varies among firms.

Content analysis is used to identify “elements” of combined assurance disclosed in integrated and sustainability reports and company webpages. Results are presented in tabular format and supported by non-parametric tests to evaluate differences in CAQ among firms in more detail.

Combined assurance is framed as a function of the responsibility of the board of directors to ensure accurate, complete and reliable reporting and the characteristics of different internal and external sources of assurance. Overall, combined assurance models are being designed conservatively. They focus mainly on specific disclosures and are guided by a limited number of assurance methodologies or frameworks instead of taking a more pluralistic approach to verification of integrated and sustainability reports as a whole.

The study is based on combined assurance practices by a sample of large listed companies in a single jurisdiction. An international comparison of combined assurance and the calibration of the proposed quality measure is deferred for future research.

Limitations in existing assurance practices are identified for the consideration of preparers and assurance providers. The quality schematic also offers practitioners, standard-setters and academics an easy-to-apply technique for examining the different elements of a company’s combined assurance model.

A better understanding of the quality of combined assurance is essential for users’ to place reliance on integrated and sustainability reports and for informing change to existing assurance practices.

The study is the first to examine the operation and quality of combined assurance. The method used to gauge assurance quality provides a useful basis for a more detailed empirical study on the relevance of combined assurance.

The general objective of this paper is to investigate the mathematics achievement of middle grade students in Pakistan. Specifically: to determine whether mathematics achievement varies systematically across students and schools; to what extent the mathematics curriculum and frameworks are implemented in schools; to what extent gender and location account for differences in mathematics achievement (at item and test scores levels); to what extent student demographics, home background, and homework variables predict mathematics achievement; to what extent schools' physical and academic resources predict mathematics achievement; and to what extent student‐ and school‐level variables interact to predict achievement indirectly.

The approach was to undertake a quantitative survey of 14,440 students from 770 schools across the country. The analytic strategy included item‐level Rasch analysis, DIF analysis across gender and regional location, and MLM analysis to test various student‐ and school‐level models.

Rasch analysis indicated that students were able to pass low‐rigour items requiring simple mathematical skills. The DIF analysis indicated that items favouring female students in either content domain belonged to knowledge of concepts to recall basic facts, terminologies, numbers, and geometric properties. Items favouring male students in either domain belonged to the problem solving level. MLM analysis revealed that at the student level, gender, location, and some home background and homework variables contributed towards mathematics achievement. At the school level, availability of learning resources and better physical facilities were found to be associated with increase in achievement scores.

Only a few major variables with policy and research implications were tested to keep the interpretations clear and simple. The next stage of this study could examine the more complex pattern of relationships and interactions among relationships for subgroups.

The study has implications for a review of the gender gap in school enrolment, the national curriculum for mathematics, homework policy, the role of regional languages as a medium of instruction in schools, the provision of school resources, and learning aids in schools.

The paper shows that the estimated models were successful in explaining the variation in average achievement in terms of proportion of variance explained and significance of estimates of the effects.