A fast Monte Carlo expectation–maximization algorithm for estimation in latent class model analysis with an application to assess diagnostic accuracy for cervical neoplasia in women with atypical glandular cells

In this article, we use a latent class model (LCM) with prevalence modeled as a function of covariates to assess diagnostic test accuracy in situations where the true disease status is not observed, but observations on three or more conditionally independent diagnostic tests are available. A fast Monte Carlo expectation–maximization (MCEM) algorithm with binary (disease) diagnostic data is implemented to estimate parameters of interest; namely, sensitivity, specificity, and prevalence of the disease as a function of covariates. To obtain standard errors for confidence interval construction of estimated parameters, the missing information principle is applied to adjust information matrix estimates. We compare the adjusted information matrix-based standard error estimates with the bootstrap standard error estimates both obtained using the fast MCEM algorithm through an extensive Monte Carlo study. Simulation demonstrates that the adjusted information matrix approach estimates the standard error similarly with the bootstrap methods under certain scenarios. The bootstrap percentile intervals have satisfactory coverage probabilities. We then apply the LCM analysis to a real data set of 122 subjects from a Gynecologic Oncology Group study of significant cervical lesion diagnosis in women with atypical glandular cells of undetermined significance to compare the diagnostic accuracy of a histology-based evaluation, a carbonic anhydrase-IX biomarker-based test and a human papillomavirus DNA test.     

Private Polls and Presidential Policymaking: Reagan as a Facilitator of Change

Did the Reagan administration disregard majority will when craftingits policy initiatives? Did it cater to a narrow partisan constituencyinstead? The answers to these questions will help with an assessmentof Jacobs and Shapiro’s (2000b) hypothesis that presidentssince the late 1970s have used private White House survey researchas a tool to manipulate or assuage centrist public opinion whilemeeting the policy demands of their partisan core supporters,resulting in a decline in presidential responsiveness to majoritywill. Using the actual surveys administered by Richard Wirthlin(Reagan’s pollster) between 1981 and 1983, this articlewill demonstrate the level of consistency between majority opinionon 129 policy issues and Reagan’s behavior through 1984,and it will explore the conditions under which the presidentwas more or less likely to respond to public preferences. Thedata reveal that the Reagan administration was constrained bythe popular will in predicable ways: if the policy issues wereabout domestic concerns, highly popular, and visible in themedia, then the administration acted in line with public preferencesmore than 70 percent of the time. Further, Reagan and his adviserswere selective in responding to party activists: they championedissues drawn from their conservative ideological agenda thatfit with the current tide in public opinion, while sidesteppingother issues dear to party activists that encountered strongmajority resistance. While I do not contest Jacobs and Shapiro’s(2000b) important observation that presidents often use surveyresearch to "craft talk" in an attempt to channel the publicdebate, the evidence here highlights how the president nonethelessremains constrained by the popular will, at least on domesticissues.