Respondents' Intrusion Upon the Situation: The Problem of Interviewing Subjects with Special Qualities*

Researchers have long been aware that when they become part of the system they are investigating, their own behavior may affect the behavior they wish to study. Little data has been collected on the effect of respondents on the researcher and the consequences of this on the research process. The problem is described and analyzed in this paper. We contend that respondents who have special qualities may intrude upon the researcher's role and make it difficult for him to collect necessary data. An illustration is presented from the authors' study of stripteasers.     

Adaptive estimation of non-linear regression models

This paper summarizes from an econometric perspective recent work by statisticians on adaptive estimation. It also presents new findings concerning the adaptive estimability of non-linear regression models.     

Managers' uses of models

This paper argues that management scientists need to understand how models are used by managers and correspondingly apply appropriate criteria to their own efforts. Types of models used by managers are discussed. Then an example is presented of how such understanding can be used to evaluate the worth of formal analysis. The evidence points to patterns of quite different inquiry and response by managers than have typically been envisioned by management scientists.     

Using performance standards to evaluate social programs with incomplete outcome data. General issues and application to a higher education block grant program

The idea of program evaluation is both simple and appealing. Program outcomes are measured and compared to some minimum performance standard or threshold. In practice, however, evaluation is difficult. Two fundamental problems of outcome measurement must be addressed. The first, which we call the problem of auxiliary outcomes, is that we do not observe outcome of interest. The second, which we call the problem of counterfactual outcomes, is that we do not observe the threshold standard. This article examines how performance standard should be set and applied in the face of these problems in measuring outcomes. The central message is that the proper way to implement standards varies with the prior information an evaluator can credibly bring to bear to compensate for incomplete outcome data. By combining available data with credible assumptions on treatments and outcomes, the performance of a program may be deemed acceptable, unacceptable, or indeterminate.     

A management science application in the courtroom: A jury selection model

This paper reports the results of a study involving the application of management science techniques to the area of jury research. A public opinion survey was conducted, entered in evidence in a motion for change of venue, and the survey data used to develop a jury selection model used by the defense in a second-degree murder trial. The case had involved considerable pre-trial publicity. From a defense perspective, the defendant received the least undesirable guilty verdict. Management science techniques employed involved statistical sampling, linear discriminant function, and regression analysis. Suggestions for other applications in jury research are offered.     

Management science models for evaluating regional government policies

This paper discusses the use of econometric models in evaluating alternative courses of action for public investment and governmental programs. Included are an extensive statewide and regional input-output analysis of the Texas economy, a simulation model to be used by government officials for fiscal policy-making and a model which simulates the demand for and use of water resources. These models provide a means whereby government planners and policy makers can plan and understand the consequences of investing limited resources in various public programs.     

Nonparametric Estimation of Bounded Survival Functions with Censored Observations

Stochastic ordering of survival functions is a useful concept in many areas of statistics, especially in nonparametric and order restricted inferences. In this paper we introduce an algorithm to compute maximum likelihood estimates of survival functions where both upper and lower bounds are given. The algorithm allows censored survival data. In a simulation study, we found that the proposed estimates are more efficient than the unrestricted Kaplan-Meier product limit estimates both with and without censored observations.