Understanding Estimators of Treatment Effects in Regression Discontinuity Designs

In this paper, we propose two new estimators of treatment effects in regression discontinuity designs. These estimators can aid understanding of the existing estimators such as the local polynomial estimator and the partially linear estimator. The first estimator is the partially polynomial estimator which extends the partially linear estimator by further incorporating derivative differences of the conditional mean of the outcome on the two sides of the discontinuity point. This estimator is related to the local polynomial estimator by a relocalization effect. Unlike the partially linear estimator, this estimator can achieve the optimal rate of convergence even under broader regularity conditions. The second estimator is an instrumental variable estimator in the fuzzy design. This estimator will reduce to the local polynomial estimator if higher order endogeneities are neglected. We study the asymptotic properties of these two estimators and conduct simulation studies to confirm the theoretical analysis.     

A Rejoinder to Forman on AHP,with Emphasis on the Requirements of Composite Ratio Scales*

In an earlier paper [11], the problems of rank reversals and invalid composite priorities in AHP were addressed by modifications to the AHP procedure. That solution was subsequently criticized [5]. In this paper, we rebut these criticisms, and we show how rank reversals in AHP can arise merely from the process of normalizing local priorities.     

PREFERENCE MAPPING OF PERCEIVED INFORMATION STRUCTURE: IMPLICATIONS FOR DECISION SUPPORT SYSTEMS DESIGN

In a recent issue of this journal, Watkins [13] presented an approach for discovery of decision-maker perceptions of the complexity (dimensionality) of information items that might be supplied by a decision support system. Through use of multidimensional scaling and cluster analysis, relatively homogeneous groups of decision makers, sharing common perceptions of various information items, were formed. This prior research was referred to as a first step in suggesting that information reports could be tailored to groups of decision makers classified on the basis of common perceptions of information. The current research extends the prior study by evaluating decision maker preferences for information in a variety of decision-making scenarios in relation to the previously identified perceptions of the information. Based on the results of the study, conclusions are made which suggest that the tailoring of information to groups of decision makers should be based on both perceptions and preferences for information. Even so, it is demonstrated that the decision tasks have an impact on the preferences for information which may affect the attempt to tailor information to groups of decision makers.     

Confidence interval estimation of the difference between paired AUCs based on combined biomarkers

In many diagnostic studies, multiple diagnostic tests are performed on each subject or multiple disease markers are available. Commonly, the information should be combined to improve the diagnostic accuracy. We consider the problem of comparing the discriminatory abilities between two groups of biomarkers. Specifically, this article focuses on confidence interval estimation of the difference between paired AUCs based on optimally combined markers under the assumption of multivariate normality. Simulation studies demonstrate that the proposed generalized variable approach provides confidence intervals with satisfying coverage probabilities at finite sample sizes. The proposed method can also easily provide P-values for hypothesis testing. Application to analysis of a subset of data from a study on coronary heart disease illustrates the utility of the method in practice.     

Some Cautions on the Measurement of User Information Satisfaction*

User information satisfaction (UIS) is important because of its potential effects on MIS department goals, quality of user work life, and extent of voluntary usage of systems. Reliable measurement of UIS is important for providing evaluative information for both researchers and practitioners. This study used 92 managers and executives as subjects to compare the test/retest reliability of a widely used, 13-scale UIS instrument together with four summary questions under experimental and control conditions. The summary questions behaved more reliably than the detailed questions for all groups, perhaps because of problems with scale units and origins and with item heterogeneity. This suggests that researchers need more reliable measures of UIS and that practitioners need to exercise caution when collecting and interpreting UIS scores.     

Ambiguous Criteria Weights in AHP: Consequences and Solutions*

Two related streams of criticism of the analytic hierarchy process (AHP) have not yet been satisfactorily resolved, although both date from the early 1980s. The first relates to ambiguity in the meaning of the relative importance of one criterion as compared to another. The second is concerned with reversals of rank alleged to be possible when new options are introduced in an AHP problem. Both proponents and critics of AHP agree that rank reversals occur, but disagree on the legitimacy of such reversals. This paper shows that there is a necessary correspondence between the manner in which criteria importances are interpreted and computed and the manner in which the weights of the options under each criterion are normalized. In general, if this relationship is ignored, incorrect weights are generated for options under consideration regardless of whether new options are added or deleted. A rank reversal on the addition of an option is merely symptomatic of this fact, and such reversals do not occur when the correspondence condition is met.     

ON MODELING ADVERTISING PULSING DECISIONS

Based on empirical findings in the literature, sales response to advertising pulsing policy (APP) is modeled mathematically. The implications for APP are discussed. This policy is compared with an alternative policy of uniform expenditures (UAP), a commonly used strategy. The results of the research indicate that substantial savings in advertising budget or an increase in sales revenues may be achieved for a firm using APP.     

Locally optimal designs for the simple death process

We demonstrate the existence and uniqueness of optimal observation schedules for discrete observation of the simple death process. Optimality is framed in the sense of observing the process at times that maximise the Fisher information for the death rate. We examine the relationship of our designs with those obtained by other authors for the simple birth process. A number of interesting properties are uncovered. Practical considerations in the application of these designs in an experimental setting are investigated, and we examine the performance of some approximately optimal designs that are usually simpler to implement. We will show that our optimal designs are highly robust to mis-specification of the death rate.     

Time Series Models for Count or Qualitative Observations

Time series sometimes consist of count data in which the number of events occurring in a given time interval is recorded. Such data are necessarily nonnegative integers, and an assumption of a Poisson or negative binomial distribution is often appropriate. This article sets ups a model in which the level of the process generating the observations changes over time. A recursion analogous to the Kalman filter is used to construct the likelihood function and to make predictions of future observations. Qualitative variables, based on a binomial or multinomial distribution, may be handled in a similar way. The model for count data may be extended to include explanatory variables. This enables nonstochastic slope and seasonal components to be included in the model, as well as permitting intervention analysis. The techniques are illustrated with a number of applications, and an attempt is made to develop a model-selection strategy along the lines of that used for Gaussian structural time series models. The applications include an analysis of the results of international football matches played between England and Scotland and an assessment of the effect of the British seat-belt law on the drivers of light-goods vehicles.     

Optimal estimation of a finite population mean under generalized random permutation models

Employing certain generalized random permutation models and a general class of linear estimators of a finite population mean, it is shown that many of the conventional estimators are “optimal” in the sense of minimum average mean square error. Simple proofs are provided by using a well-known theorem on UMV estimation. The results also cover certain simple response error situations.