This dissertation illustrates how to detect the rater centrality effect in a simulation study that approximates data collected in large scale performance assessment settings. It addresses three research questions that: (1) which of several centrality-detection indices are most sensitive to the difference between effect raters and non-effect raters; (2) how accurate (and inaccurate), in terms of Type I error rate and statistical power, each centrality-detection index is in flagging effect raters; and (3) how the features of the data collection design (i.e., the independent variables including the level of centrality strength, the double-scoring rate, and the number of raters and ratees) influence the accuracy of rater classifications by these centrality-detection indices. The results reveal that the measure-residual correlation, the expected-residual correlation, and the standardized deviation of assigned scores perform better than the point-measure correlation. The mean-square fit statistics, traditionally viewed as potential indicators of rater centrality, perform poorly in terms of differentiating central raters from normal raters. Along with the rater slope index, the mean-square fit statistics did not appear to be sensitive to the rater centrality effect. All of these indices provided reasonable protection against Type I errors when all responses were double scored, and that higher statistical power was achieved when responses were 100% double scored in comparison to only 10% being double scored. With a consideration on balancing both Type I error and statistical power, I recommend the measure-residual correlation and the expected-residual correlation for detecting the centrality effect. I suggest using the point-measure correlation only when responses are 100% double scored. The four parameters evaluated in the experimental simulations had different impact on the accuracy of rater classification. The results show that improving the classification accuracy for non-effect raters may come at a cost of reducing the classification accuracy for effect raters. Some simple guidelines for the expected impact of classification accuracy when a higher-order interaction exists summarized from the analyses offer a glimpse of the "pros" and "cons" in adjusting the magnitude of the parameters when we evaluate the impact of the four experimental parameters on the outcomes of rater classification. / Ph. D.

Made available in DSpace on 2016-03-15T19:32:48Z (GMT). No. of bitstreams: 1 Hugo Crespi Junior.pdf: 1552051 bytes, checksum: 34b1ca02581c6a8240a8bd4702d15a71 (MD5) Previous issue date: 2014-02-14 / One of the most important ways to finance consumers in the Brazilian market is the consumer credit offered in store. Provided by independent or captive finance companies, the consumer credit is normally granted or denied without taking into account its effect on the retailer s profitability. Denying credit to finance a high profit margin product is more damaging to the companies profits than if such refusal is about the sale of lower margin product. This suggests that there is an opportunity to improve the profitability in this sales channel. The objective of this research was to check the possibility to increase the Brazilian retailers profitability by introducing the retail products operating margin as an additional parameter of the consumer credit analysis. Simulations using tangents to ROC curves, as proposed by Stein (2005), made possible to confirm, through a new balance of type I and type II errors, that the maximization of corporate earnings occurs when using operating retail margins for determining cutoffs in consumer credit models. / Uma das mais importantes ferramentas de financiamento no varejo brasileiro é o crédito direto ao consumidor oferecido nos estabelecimentos por ocasião da compra. Operado através de financeiras cativas ou parceiras, o crédito é, normalmente, concedido ou negado sem que se leve em conta seu efeito na lucratividade do varejista. Quando se recusa o financiamento para um produto de grande margem de lucro, o efeito sobre os ganhos da empresa é evidentemente maior do que quando essa recusa inibe a venda de produto de menor margem, o que sugere haver ineficácia no processo. Esta pesquisa visou verificar se há espaço para aumentar a rentabilidade no varejo brasileiro, introduzindo a margem operacional nos critérios utilizados para concessão de crédito direto ao consumidor. Utilizando a curva ROC e a abordagem oferecida por Stein (2005), construíram-se simulações em torno de valores reais praticados no mercado, que permitiram confirmar, através de um novo balanceamento dos erros tipo I e tipo II, que ocorre a maximização de ganhos empresarias quando as margens operacionais do varejo são consideradas para a determinação de pontos de corte em modelos de crédito direto ao consumidor.