arXiv subjects

Suzanne Thornton

Publications and source records attributed to Suzanne Thornton.

7 recordsLinked to original sources

arXiv2025

Atomic clock frequency ratios with fractional uncertainty $\leq 3.2 \times 10^{-18}$

We report high-precision frequency ratio measurements between optical atomic clocks based on $^{27}$Al$^+$, $^{171}$Yb, and $^{87}$Sr. With total fractional uncertainties at or below $3.2 \times 10^{-18}$, these measurements meet an important milestone criterion for redefinition of the second in the International System of Units. Discrepancies in $^{87}$Sr ratios at approximately $1\times10^{-16}$ and the Al$^+$/Yb ratio at $1.6\times10^{-17}$ in fractional units compared to our previous measurements underscore the importance of repeated, high-precision comparisons by different laboratories. A key innovation in this work is the use of a common ultrastable reference delivered to all clocks via a 3.6 km phase-stabilized fiber link between two institutions. Derived from a cryogenic single-crystal silicon cavity, this reference improves comparison stability by a factor of 2 to 3 over previous systems, with an optical lattice clock ratio achieving a fractional instability of $1.3 \times 10^{-16}$ at 1 second. By enabling faster comparisons, this stability will improve sensitivity to non-white noise processes and other underlying limits of state-of-the-art optical frequency standards.

physics.atom-ph↗

arXiv2024

Ethical considerations for data involving human gender and sex variables

The inclusion of human sex and gender data in statistical analysis invokes multiple considerations for data collection, combination, analysis, and interpretation. These considerations are not unique to variables representing sex and gender. However, considering the relevance of the ethical practice standards for statistics and data science to sex and gender variables is timely, with results that can be applied to other sociocultural variables. Historically, human gender and sex have been categorized with a binary system. This tradition persists mainly because it is easy, and not because it produces the best scientific information. Binary classification simplifies combinations of older and newer data sets. However, this classification system eliminates the ability for respondents to articulate their gender identity, conflates gender and sex, and also obscures potentially important differences by collapsing across valid and authentic categories. This approach perpetuates historical inaccuracy, simplicity, and bias, while also limiting the information that emerges from analyses of human data. The approach also violates multiple elements in the American Statistical Association (ASA) Ethical Guidelines for Statistical Practice. Information that would be captured with a nonbinary classification could be relevant to decisions about analysis methods and to decisions based on otherwise expert statistical work. Statistical practitioners are increasingly concerned with inconsistent, uninformative, and even unethical data collection and analysis practices. This paper presents a historical introduction to the collection and analysis of human gender and sex data, offers a critique of a few common survey questioning methods based on alignment with the ASA Ethical Guidelines, and considers the scope of ethical considerations for human gender and sex data from design through analysis and interpretation.

stat.AP↗

arXiv2023

The Mastery Rubric for Statistics and Data Science: promoting coherence and consistency in data science education and training

Consensus based publications of both competencies and undergraduate curriculum guidance documents targeting data science instruction for higher education have recently been published. Recommendations for curriculum features from diverse sources may not result in consistent training across programs. A Mastery Rubric was developed that prioritizes the promotion and documentation of formal growth as well as the development of independence needed for the 13 requisite knowledge, skills, and abilities for professional practice in statistics and data science, SDS. The Mastery Rubric, MR, driven curriculum can emphasize computation, statistics, or a third discipline in which the other would be deployed or, all three can be featured. The MR SDS supports each of these program structures while promoting consistency with international, consensus based, curricular recommendations for statistics and data science, and allows 'statistics', 'data science', and 'statistics and data science' curricula to consistently educate students with a focus on increasing learners independence. The Mastery Rubric construct integrates findings from the learning sciences, cognitive and educational psychology, to support teachers and students through the learning enterprise. The MR SDS will support higher education as well as the interests of business, government, and academic work force development, bringing a consistent framework to address challenges that exist for a domain that is claimed to be both an independent discipline and part of other disciplines, including computer science, engineering, and statistics. The MR-SDS can be used for development or revision of an evaluable curriculum that will reliably support the preparation of early e.g., undergraduate degree programs, middle e.g., upskilling and training programs, and late e.g., doctoral level training practitioners.

stat.OT↗

arXiv2022

Approximate confidence distribution computing

Approximate confidence distribution computing (ACDC) offers a new take on the rapidly developing field of likelihood-free inference from within a frequentist framework. The appeal of this computational method for statistical inference hinges upon the concept of a confidence distribution, a special type of estimator which is defined with respect to the repeated sampling principle. An ACDC method provides frequentist validation for computational inference in problems with unknown or intractable likelihoods. The main theoretical contribution of this work is the identification of a matching condition necessary for frequentist validity of inference from this method. In addition to providing an example of how a modern understanding of confidence distribution theory can be used to connect Bayesian and frequentist inferential paradigms, we present a case to expand the current scope of so-called approximate Bayesian inference to include non-Bayesian inference by targeting a confidence distribution rather than a posterior. The main practical contribution of this work is the development of a data-driven approach to drive ACDC in both Bayesian or frequentist contexts. The ACDC algorithm is data-driven by the selection of a data-dependent proposal function, the structure of which is quite general and adaptable to many settings. We explore two numerical examples that both verify the theoretical arguments in the development of ACDC and suggest instances in which ACDC outperform approximate Bayesian computing methods computationally.

stat.ME↗

arXiv2022

Bridging Bayesian, frequentist and fiducial (BFF) inferences using confidence distribution

Bayesian, frequentist and fiducial (BFF) inferences are much more congruous than they have been perceived historically in the scientific community (cf., Reid and Cox 2015; Kass 2011; Efron 1998). Most practitioners are probably more familiar with the two dominant statistical inferential paradigms, Bayesian inference and frequentist inference. The third, lesser known fiducial inference paradigm was pioneered by R.A. Fisher in an attempt to define an inversion procedure for inference as an alternative to Bayes' theorem. Although each paradigm has its own strengths and limitations subject to their different philosophical underpinnings, this article intends to bridge these different inferential methodologies through the lenses of confidence distribution theory and Monte-Carlo simulation procedures. This article attempts to understand how these three distinct paradigms, Bayesian, frequentist, and fiducial inference, can be unified and compared on a foundational level, thereby increasing the range of possible techniques available to both statistical theorists and practitioners across all fields.

stat.ME↗

arXiv2021

Best Practices for Collecting Gender and Sex Data

The measurement and analysis of human sex and gender is a nuanced problem with many overlapping considerations including statistical bias, data privacy, and the ethical treatment of study subjects. Traditionally, human gender and sex have been categorized and measured with respect to an artificial binary system. The continuation of this tradition persists mainly because it is easy to replication and not, as we argue, because it produces the most valuable scientific information. Sex and gender identity data is crucial for many applications of statistical analysis and many modern scientists acknowledge the limitations of the current system. However, discrimination against sex and gender minorities poses very real privacy concerns when collecting and distributing gender and sex data. As such, extra thoughtfulness and care is essential to design safe and informative scientific studies. In this paper, we present statistically informed recommendations for the data collection and analysis of human subjects that not only respect each individual's identity and protect their privacy, but also establish standards for collecting higher quality data.

stat.ME↗

arXiv2018

An effective likelihood-free approximate computing method with statistical inferential guarantees

Approximate Bayesian computing is a powerful likelihood-free method that has grown increasingly popular since early applications in population genetics. However, complications arise in the theoretical justification for Bayesian inference conducted from this method with a non-sufficient summary statistic. In this paper, we seek to re-frame approximate Bayesian computing within a frequentist context and justify its performance by standards set on the frequency coverage rate. In doing so, we develop a new computational technique called approximate confidence distribution computing, yielding theoretical support for the use of non-sufficient summary statistics in likelihood-free methods. Furthermore, we demonstrate that approximate confidence distribution computing extends the scope of approximate Bayesian computing to include data-dependent priors without damaging the inferential integrity. This data-dependent prior can be viewed as an initial `distribution estimate' of the target parameter which is updated with the results of the approximate confidence distribution computing method. A general strategy for constructing an appropriate data-dependent prior is also discussed and is shown to often increase the computing speed while maintaining statistical inferential guarantees. We supplement the theory with simulation studies illustrating the benefits of the proposed method, namely the potential for broader applications and the increased computing speed compared to the standard approximate Bayesian computing methods.

stat.CO↗