Confidence Intervals for Current Status Data

Abstract.  The likelihood ratio statistic for testing pointwise hypotheses about the survival time distribution in the current status model can be inverted to yield confidence intervals (CIs). One advantage of this procedure is that CIs can be formed without estimating the unknown parameters that figure in the asymptotic distribution of the maximum likelihood estimator (MLE) of the distribution function. We discuss the likelihood ratio-based CIs for the distribution function and the quantile function and compare these intervals to several different intervals based on the MLE. The quantiles of the limiting distribution of the MLE are estimated using various methods including parametric fitting, kernel smoothing and subsampling techniques. Comparisons are carried out both for simulated data and on a data set involving time to immunization against rubella. The comparisons indicate that the likelihood ratio-based intervals are preferable from several perspectives.     

COMMENT ON "MONITORING AND BUDGET MAXIMIZATION IN THE CONTROL OF POLLUTION"

Lee's model of optimal pollution taxation and costly monitoring is examined. It is argued that his conclusions regarding the nature of the optimal tax and monitoring levels need to be modified to account for the likelihood of a corner solution to his first order conditions. Additionally, the predictions of Lee's version of the budget-maximizing pollution control bureaucracy are compared to some (more or less) inconsistent facts.     

MAIL SURVEYS FOR ELECTION FORECASTING?: AN EVALUATION OF THE COLUMBUS DISPATCH POLL

Because of slow turnaround time and typically low response rates,mail surveys have generally been considered of little valuein election forecasting. However, statewide mail surveys conductedby the Columbus Dispatch newspaper since 1980 have made remarkablyaccurate forecasts of Ohio election outcomes. In comparisonto statewide surveys by two other organizations employing conventionaltelephone interview methods, the mail surveys were consistentlymore accurate and were generally less susceptible to sourcesof inaccuracy such as high rolloff and low publicity. The mailsurvey's advantage is attributable at least in part to largersample sizes, sampling and response procedures that yieldedmore representative samples of voters, lack of the need to allocateundecided respondents, and superior questionnaire design. Thesefindings suggest that mail surveys not only may be viable alternativesto telephone surveys but may actually be superior to them undersome conditions. Further-more, these results demonstrate thatsurveys with low response rates are not necessarily low in validity.     

INFLUENCE OF AN INVITATION TO ANSWER BY TELEPHONE ON RESPONSE TO CENSUS QUESTIONNAIRES

An experimental test was made of whether completion rates fora census questionnaire could be improved by offering the optionof calling a toll-free number and providing the requested informationto an interviewer as an alternative to mailing it back. Conductedon a national probability sample of house holds, five treatmentpanels were created to examine the effect of delivering, indifferent ways, the invitation to respond by telephone or mail.Offering such invitations did not improve completion rates exceptwhen included with a follow-up letter that did not also includea replacement questionnaire.     

Weighted Likelihood for Semiparametric Models and Two-phase Stratified Samples, with Application to Cox Regression

Abstract.  We consider semiparametric models for which solution of Horvitz–Thompson or inverse probability weighted (IPW) likelihood equations with two-phase stratified samples leads to consistent and asymptotically Gaussian estimators of both Euclidean and non-parametric parameters. For Bernoulli (independent and identically distributed) sampling, standard theory shows that the Euclidean parameter estimator is asymptotically linear in the IPW influence function. By proving weak convergence of the IPW empirical process, and borrowing results on weighted bootstrap empirical processes, we derive a parallel asymptotic expansion for finite population stratified sampling. Several of our key results have been derived already for Cox regression with stratified case–cohort and more general survey designs. This paper is intended to help interpret this previous work and to pave the way towards a general Horvitz–Thompson approach to semiparametric inference with data from complex probability samples.     

Comparative Evaluation of the Local Police

Gallup data indicating presidential popularity from 1950 to1974 are analyzed in a distributed lag model which tests theimportance of key economic influences. Results indicate thatinflation and military expenditures are statistically significantand politically important influences on presidential popularity.The success of this model in revealing hitherto undetected economicinfluence calls into question the nonfindings of previous simplelinear analyses and suggests the need for more sophisticatedmodels which provide, through the use of time lags and cumulativeimpacts, a more complex pattern of economic influence on presidentialpopularity.     

French connections: the networking strategies of French highly skilled migrants in London

Although the migration studies literature often takes social networks for granted, these social ties are not spontaneous but require effort and nurturing. There has been insufficient research on the actual process of networking, especially among highly skilled migrants. Our understanding of why and how migrants form networks with particular characteristics is still poor. In this article, we argue that it is necessary to consider both the structure and content of networks – the nature of the relationships as well as the flow of resources within various social ties. Drawing on qualitative data from a study of highly skilled French migrants in London's business and financial sector, we use a microanalysis of network‐making processes. In the context of London as a dynamic and highly competitive financial centre, we examine the importance of opportunities, skills and shared interests in building new social relationships from scratch. In addition, we also assess how mobility and proximity, virtual communication and co‐presence impact on geographically dispersed networks and why some long distance relationships endure while others fade over time. By bringing together classic literature on professional networking and wider discussions on how relationships are managed across time and space, our work contributes to a fuller understanding of why and how highly skilled migrants form networks with particular characteristics.     

EFFECTS OF QUESTIONNAIRE LENGTH, RESPONDENT-FRIENDLY DESIGN, AND A DIFFICULT QUESTION ON RESPONSE RATES FOR OCCUPANT-ADDRESSED CENSUS MAIL SURVEYS

An experimental study of alternatives to the current U.S. decennialcensus questionnaire shows that shortening the questionnaireand respondent-friendly questionnaire design improve response,whereas asking a potentially difficult and/or objectionablequestion, that is, social security number, lowers response.This national study of 17,000 household addresses also demonstratesthat relatively high mail survey response can be achieved withoutaddressing correspondence to individual names of residents.     

INCENTIVES AND BEHAVIOR IN ENGLISH, DUTCH AND SEALED-BID AUCTIONS

The Pareto optimality and price behavior of English and Dutch oral auctions, and First-Price and Second-Price sealedbid auctions are compared under various procedures for assigning valuations among cash motivated bidders. The Vickrey propositions with respect to the mean and variance of prices under the English, Dutch and Second-Price auctions are not falsified by the data. Individual behavior and prices in the First-Price auction deviates considerably from Vickrey's Nash postulate. Behaviorly, the English and Second-Price auctions appear to be isomorphic, but the Dutch and First-Price auctions may not be isomorphic.