Identification and classification of multiple outliers,high leverage points and influential observations in linear regression

Detection of multiple unusual observations such as outliers, high leverage points and influential observations (IOs) in regression is still a challenging task for statisticians due to the well-known masking and swamping effects. In this paper we introduce a robust influence distance that can identify multiple IOs, and propose a sixfold plotting technique based on the well-known group deletion approach to classify regular observations, outliers, high leverage points and IOs simultaneously in linear regression. Experiments through several well-referred data sets and simulation studies demonstrate that the proposed algorithm performs successfully in the presence of multiple unusual observations and can avoid masking and/or swamping effects.     

Optimal voting rules for two-member tenure committees

A tenure committee first votes on whether to hire a candidate; if it does, it receives an informative performance signal, and then votes on whether to tenure the candidate; rejection at either stage returns the committee to a candidate pool, endogenising the value of the outside option. A candidate’s fate depends only on the behaviour of two ‘weather-vane’ committee members. Committee members may vote against favoured candidates if the weather-vane is opposed; enthusiastic assessments by one of these weather-vanes may harm a candidate’s chances by increasing others’ thresholds for hiring him; sunk time costs may lead voters who voted against hiring to vote for tenuring him, even after a poor probationary performance. For two member committees that are patient and perceptive, the optimal voting rule is a (weak) majority at the hiring stage and unanimity at the tenure stage; when such committees are impatient or imperceptive, the double (weak) majority rule is optimal. Perversely, the performance of a patient, imperceptive committee improves as its perceptiveness further declines. Consistent with practice, falling threshold rules are not optimal. Results on optimal voting rules are also presented in limit cases as committee members’ beliefs become more correlated. Finally, we compare the model to a discrete-time European options model.     

Democratic assessment as scales of justice: the case of three Iranian high-stakes tests

High-stakes tests are often used as instruments of agenda-setting and control in developing areas to enable policy-makers to establish and embed education policy agendas throughout the education system and society. As an alternative method, Critical Language Testing assumes that tests are value-laden instruments and products of political, social, cultural, and ideologically driven educational agendas. This article critically examines three Iranian high-stakes tests (the National University Entrance Exam, the MA/MS Exam, and PhD examinations) to evaluate whether covert policies and agendas are being deployed in the Iranian context. It applies qualitative methods and Strauss and Corbin’s [1998. Basics of Qualitative Research: Techniques and Procedures for Developing Grounded Theory. London: Sage] constant comparative method to demonstrate the existence of clear patterns of domination. Democratic testing is therefore suggested as a way forward by which policy-makers can operationalize a just and fair exam whereby testing parties’ ideas, and intuitions are equally taken into account and their rights are protected.