Browsing by Author "Reiter, JP"

Now showing 1 - 4 of 4
  • Thumbnail Image
    ItemOpen Access
    A note on Bayesian inference after multiple imputation
    (American Statistician, 2010-05-01) Zhou, X; Reiter, JP
    This article is aimed at practitioners who plan to use Bayesian inference on multiply-imputed datasets in settings where posterior distributions of the parameters of interest are not approximately Gaussian. We seek to steer practitioners away from a naive approach to Bayesian inference, namely estimating the posterior distribution in each completed dataset and averaging functionals of these distributions. We demonstrate that this approach results in unreliable inferences. A better approach is to mix draws from the posterior distributions from each completed dataset, and use the mixed draws to summarize the posterior distribution. Using simulations, we show that for this second approach to work well, the number of imputed datasets should be large. In particular, five to ten imputed datasets "which is the standard recommendation for multiple imputation" is generally not enough to result in reliable Bayesian inferences. © 2010 American Statistical Association.
  • Thumbnail Image
    ItemOpen Access
    Bayesian latent pattern mixture models for handling attrition in panel studies with refreshment samples
    (Annals of Applied Statistics, 2016-03-01) Si, Y; Reiter, JP; Hillygus, DS
    Many panel studies collect refreshment samples—new, randomly sampled respondents who complete the questionnaire at the same time as a subsequent wave of the panel. With appropriate modeling, these samples can be leveraged to correct inferences for biases caused by nonignorable attrition. We present such a model when the panel includes many categorical survey variables. The model relies on a Bayesian latent pattern mixture model, in which an indicator for attrition and the survey variables are modeled jointly via a latent class model.We allow the multinomial probabilities within classes to depend on the attrition indicator, which offers additional flexibility over standard applications of latent class models. We present results of simulation studies that illustrate the benefits of this flexibility. We apply the model to correct attrition bias in an analysis of data from the 2007–2008 Associated Press/Yahoo News election panel study.
  • Thumbnail Image
    ItemOpen Access
    Creating linked datasets for SME energy-assessment evidence-building: Results from the U.S. Industrial Assessment Center Program
    (Energy Policy, 2017-12-01) Dalzell, NM; Boyd, GA; Reiter, JP
    © 2017 Elsevier Ltd Lack of information is commonly cited as a market failure resulting in an energy-efficiency gap. Government information policies to fill this gap may enable improvements in energy efficiency and social welfare because of the externalities of energy use. The U.S. Department of Energy Industrial Assessment Center (IAC) program is one such policy intervention, providing no-cost assessments to small and medium enterprises (SME). The IAC program has assembled a wealth of data on these assessments, but the database does not include information about participants after the assessment or on non-participants. This study addresses that lack by creating a new linked dataset using the public IAC and non-public data at the Census Bureau. The IAC database excludes detail needed for an exact match, so the study developed a linking methodology to account for uncertainty in the matching process. Based on the linking approach, a difference in difference analysis for SME that received an assessment was done; plants that received an assessment improve their performance over time, relative to industry peers that did not. This new linked dataset is likely to shed even more light on the impact of the IAC and similar programs in advancing energy efficiency.
  • Thumbnail Image
    ItemOpen Access
    Releasing multiply-imputed synthetic data generated in two stages to protect confidentiality
    (Statistica Sinica, 2010-01-01) Reiter, JP; Drechsler, J
    To protect the confidentiality of survey respondents' identities and sensitive attributes, statistical agencies can release data in which confidential values are replaced with multiple imputations. These are called synthetic data. We propose a two-stage approach to generating synthetic data that enables agencies to release different numbers of imputations for different variables. Generation in two stages can reduce computational burdens, decrease disclosure risk, and increase inferential accuracy relative to generation in one stage. We present methods for obtaining inferences from such data. We describe the application of two stage synthesis to creating a public use file for a German business database.