Suchergebnisse
Filter
4 Ergebnisse
Sortierung:
Nonparametric Differentially Private Confidence Intervals for the Median
In: Journal of survey statistics and methodology: JSSAM, Band 10, Heft 3, S. 804-829
ISSN: 2325-0992
Abstract
Differential privacy is a restriction on data processing algorithms that provides strong confidentiality guarantees for individual records in the data. However, research on proper statistical inference, that is, research on properly quantifying the uncertainty of the (noisy) sample estimate regarding the true value in the population, is currently still limited. This article proposes and evaluates several strategies to compute valid differentially private confidence intervals for the median. Instead of computing a differentially private point estimate and deriving its uncertainty, we directly estimate the interval bounds and discuss why this approach is superior if ensuring privacy is important. We also illustrate that addressing both sources of uncertainty—the error from sampling and the error from protecting the output—simultaneously should be preferred over simpler approaches that incorporate the uncertainty in a sequential fashion. We evaluate the performance of the different algorithms under various parameter settings in extensive simulation studies and demonstrate how the findings could be applied in practical settings using data from the 1940 Decennial Census.
In Search of Meaning: Does the Fortune Reputation Survey Alter Performance Expectations?
In: Canadian journal of administrative sciences: Revue canadienne des sciences de l'administration, Band 20, Heft 3, S. 187-195
ISSN: 1936-4490
AbstractOur study theoretically and empirically examines performance antecedents and consequences of the Fortune annual Survey of Corporate Reputation. Accounting‐and market‐based measures of performance are used to predict the raings, and investor reactions to the publication of the ratings are predicted to be associated with the extent to which the ratings diverge from antecedent predictions. Lower‐than‐predicted ratings should generate a negative response while higher‐than‐predicted ratings should generate a positive response. Contrary to expectations, we found a negative relationship. In addition, this negative relationship was only for the lower‐than‐predicted ratings. For higher‐than‐predicted ratings the relationship with investor reaction was insignificant.RésuméNotre étude consiste en un examen théorigue et empirique des facteurs influençant le classement annuel du magazine Fortune et des conséquences de ce classement sur la performance des firmes évaluées. Nous utilisons des mesures comptables et financières pour examiner le lien entre la performance et la réputation de la firme. La façon dont les investisseurs réagissent à ces révaluations doit en principe être proportionnelle au degré de divergence par rapport aux prédictions antérieures. Théoriquement, les évaluations qui sont moins élevées que prévues entraînent une réaction négalive des investisseurs, tandis que les évaluations qui sont plus élevées que prévues entraínent une réponse positive des investisseurs. Mais dans la réalité, on observe plutôt une relation inverse, en l'occurrence dans le can des évaluations qui sont moins élevées que prévues. Les évaluations qui sont plus élevées que prévues n'ont qu'un impact limité sur la réaction des investisseurs.
Discussion
In: Proceedings of the Academy of Political Science in the City of New York, Band 6, Heft 1, S. 185