Annals of Applied Statistics

About Annals of Applied Statistics

About the Annals of Applied Statistics

The Annals of Applied Statistics is a peer-reviewed scientific journal published by the Institute of Mathematical Statistics (IMS), one of the world's leading professional societies for statistics and probability. Established in 2007, the journal was created to give applied statistics a dedicated flagship outlet within the IMS family of journals, complementing the more theoretically oriented Annals of Statistics. Its remit is deliberately broad: the journal covers all areas of statistics, featuring papers in the applied half of this range — research in which statistical methodology is developed, refined, or deployed to solve genuine problems arising in science, technology, medicine, business, and public policy.

From its founding, the Annals of Applied Statistics has pursued a distinctive editorial philosophy: statistical work is judged by the depth of its engagement with real data and real scientific questions. A paper in this journal is expected to do more than prove theorems or propose an algorithm in the abstract; it should demonstrate how statistical thinking changes the analysis of a substantive problem. That orientation has made the journal a home for many of the most influential applied methodology papers of the past two decades, including work on random forests, Bayesian regression trees, regularization paths, and Bayesian computation that is now cited thousands of times across disciplines.

The journal publishes quarterly and is hosted on Project Euclid, the digital publishing platform for mathematics and statistics literature, where its full archive from the inaugural 2007 volume is available. With an ISSN of 1932-6157 in print and 1941-7330 online, and a 2022 Journal Citation Reports impact factor of 1.8, the Annals of Applied Statistics stands as a top-tier venue for statisticians whose work bridges methodology and application.

History and Founding

The Annals of Applied Statistics was founded in 2007 with Bradley Efron, professor emeritus of statistics at Stanford University, as its founding editor-in-chief. Efron — one of the most decorated statisticians of the modern era, known for the bootstrap and empirical Bayes methodology — shaped the journal's founding vision: a venue that would demonstrate the real-world power of statistical ideas and give applied statistical research the same prestige that theoretical advances enjoyed in the existing Annals journals.

The timing of the journal's creation was significant. The mid-2000s saw an explosion of complex, high-dimensional data across genomics, neuroscience, finance, and the emerging internet economy, and statisticians were increasingly called upon not merely to analyze data but to build the inferential machinery that entire scientific fields depended on. The IMS created the Annals of Applied Statistics to capture this moment — to publish research that, in the words used by the journal's editorial leadership, demonstrates the real-world applications of statistics from econometrics to epidemiology and beyond.

Since its launch, the journal has grown into one of the most cited applied statistics journals in the world. Aggregator-reported figures indicate it published roughly 156 articles in 2025, reflecting a healthy and selective publication program. Its archive now spans nearly two decades of methodological development, documenting the rise of machine learning-adjacent statistics, high-dimensional inference, causal inference, Bayesian computation, and the interface between statistics and artificial intelligence.

Editorial Leadership

The current editor-in-chief of the Annals of Applied Statistics is Lexin Li, professor of biostatistics at the UC Berkeley School of Public Health, appointed by the Institute of Mathematical Statistics for a term running from January 2025 through December 2027. Li's appointment was announced in 2024 and reflects the journal's forward-looking direction: his research focuses on statistical methodology and data analysis for neuroimaging, computational biology, and biomedical research, with contributions to dimension reduction, tensor regression, and the analysis of complex biomedical data.

Li was elected a fellow of the Institute of Mathematical Statistics in 2021 and a fellow of the American Statistical Association in 2017, and his editorial program explicitly embraces the rapid advancement of artificial intelligence. Upon taking the editorship, he wrote that the moment presents "a unique and timely opportunity to foster stronger connections between statistics and AI," committing the journal to remain at the forefront of integrating the two disciplines and promoting innovative research that enhances the application of statistical methodologies across numerous domains.

Under Li's editorship, the journal issued a special-issue call for papers on "Statistics and Artificial Intelligence" (announced September 2025), inviting submissions that identify core problems motivated by significant real-world applications where the integration of AI and statistical principles can produce novel and impactful solutions. Topics span modern AI-related study design, deep learning, reinforcement learning, representation learning, generative AI, large language models, and the alignment, interpretation, and uncertainty quantification of AI models — with applications ranging across the biological, medical, neural, environmental, physical, social, and political sciences.

Editorial policy under the current leadership emphasizes application- and data-centered papers: manuscripts should be motivated by a substantive problem and demonstrate how statistical methodology advances its solution. The journal's editorial board, drawn from the international statistics community, oversees a rigorous peer-review process consistent with IMS standards.

Key Facts at a Glance

Aims and Scope

The Annals of Applied Statistics publishes papers of the highest quality across the full spectrum of applied statistics. Its scope encompasses statistical methodology motivated by problems in the biological and biomedical sciences, econometrics, epidemiology, environmental science, finance, genetics and genomics, neuroscience, public health, social sciences, and engineering, as well as the computational and theoretical innovations that make such applications possible.

Several characteristics define what the journal looks for in a submission:

• Substantive motivation: Papers should be driven by a genuine scientific or practical problem. Methodological novelty alone, without a compelling application, is a poor fit.
• Data-centered analysis: The editorial leadership explicitly encourages application- and data-centered papers — work in which real datasets play a central role in motivating, illustrating, and validating the methodology.
• Methodological rigor: Applications must be supported by sound statistical reasoning, appropriate uncertainty quantification, and honest assessment of limitations.
• Breadth of readership: Because the journal serves the entire applied statistics community, papers should be written accessibly enough that statisticians outside the immediate application area can appreciate the contribution.

The journal also publishes special issues on emerging themes — most recently the intersection of statistics and artificial intelligence — positioning itself at the frontier where classical statistical inference meets modern machine learning. This makes the Annals of Applied Statistics an especially relevant outlet for researchers working on uncertainty quantification for AI models, statistical foundations of deep learning, and the design of experiments and observational studies in the age of massive data.

Landmark Papers

Few applied statistics journals can claim a roster of landmark papers as influential as this one's. The following are among the most-cited papers ever published in the Annals of Applied Statistics, with citation counts from Crossref and DOIs verified as resolving:

1. Random survival forests (2008)

Authors: Hemant Ishwaran, Udaya B. Kogalur, Eugene H. Blackstone
Citation: Annals of Applied Statistics, 2008, Volume 2, Issue 3, pages 841–860
DOI: 10.1214/08-AOAS169

This is the most-cited paper in the journal's history, with over 2,500 Crossref citations. It introduced random survival forests, an extension of random forests to right-censored survival data, and became a cornerstone of modern survival analysis in biostatistics. Its influence extends across medicine, epidemiology, and reliability engineering, and it exemplifies the journal's mission: a methodological innovation built around a pressing applied problem — predicting patient survival.

2. A weakly informative default prior distribution for logistic and other regression models (2008)

Authors: Andrew Gelman, Aleks Jakulin, Maria Grazia Pittau, Yu-Sung Su
Citation: Annals of Applied Statistics, 2008, Volume 2, Issue 4, pages 1360–1383
DOI: 10.1214/08-AOAS191

This paper addressed the problem of separation in logistic regression — when maximum likelihood estimates blow up — by proposing a weakly informative default prior (a Cauchy distribution) that stabilizes estimation. With over 1,600 citations, it reshaped default Bayesian practice in applied regression modeling and remains one of the most practically influential Bayesian methodology papers of its era.

3. BART: Bayesian additive regression trees (2010)

Authors: Hugh A. Chipman, Edward I. George, Robert E. McCulloch
Citation: Annals of Applied Statistics, 2010, Volume 4, Issue 1, pages 266–298
DOI: 10.1214/09-AOAS285

BART introduced a Bayesian sum-of-trees model that has become one of the most widely used nonparametric regression methods in applied statistics, with over 1,500 citations. Its success in prediction, causal inference, and uncertainty quantification made it a signature achievement of the journal and a staple of modern applied Bayesian analysis.

4. Pathwise coordinate optimization (2007)

Authors: Jerome Friedman, Trevor Hastie, Holger Höfling, Robert Tibshirani
Citation: Annals of Applied Statistics, 2007, Volume 1, Issue 2, pages 302–332
DOI: 10.1214/07-AOAS131

Published in the journal's inaugural volume, this paper presented the coordinate-descent algorithm for computing regularization paths in penalized regression (lasso, elastic net, and generalizations). With over 1,300 citations, it underpins the computational engine of high-dimensional statistics and much of modern machine learning practice.

5. Measuring reproducibility of high-throughput experiments (2011)

Authors: Qunhua Li, James B. Brown, Haiyan Huang, Peter J. Bickel
Citation: Annals of Applied Statistics, 2011, Volume 5, Issue 3, pages 1752–1779
DOI: 10.1214/11-AOAS466

This paper introduced the irreproducible discovery rate (IDR) framework for assessing the reproducibility of high-throughput genomic experiments such as ChIP-seq. With over 1,000 citations, it became a standard tool in genomics data analysis and demonstrates the journal's impact at the interface of statistics and the life sciences.

Publishing in the Annals of Applied Statistics

Researchers considering this journal should understand its selectivity and expectations. The Annals of Applied Statistics is a top-tier IMS journal, and manuscripts compete for space with submissions from leading statistics groups worldwide. Practical guidance:

• Fit the applied mission: The single most common reason for desk rejection is a mismatch with the journal's applied focus. Purely theoretical papers belong in the Annals of Statistics; papers whose central contribution is a data analysis belong here.
• Lead with the application: Successful submissions typically open with a substantive scientific problem, develop methodology in response to it, and return to the application with convincing empirical results.
• Computation matters: The journal's landmark papers are overwhelmingly computational-methodological. Reproducible code, careful simulation studies, and honest benchmarking against existing methods strengthen a submission.
• Submit through official channels: Manuscripts are submitted through the journal's online submission system, accessible from the IMS journal page and Project Euclid. Authors should consult the current author guidelines for formatting, reference style, and article-type requirements before submitting.
• Consider special issues: The journal periodically announces special issues — such as the 2025 call on statistics and artificial intelligence — which can be an excellent entry point for work aligned with the theme.

The journal does not officially publish its acceptance rate or article processing charges; authors should check the current IMS and Project Euclid pages for the latest publication policies.

Indexing and Visibility

The Annals of Applied Statistics is indexed in the major scholarly databases that matter for statistics research, including Scopus, the Web of Science (Science Citation Index Expanded), JSTOR, Mathematical Reviews / MathSciNet, and zbMATH. Its full text is available through Project Euclid, which provides open metadata and stable archival access. The journal's consistent presence in Journal Citation Reports, combined with its high H-index of 96, makes publications here highly visible within statistics and in the application disciplines its papers serve.

Research Themes and Signature Contributions

Beyond individual landmark papers, the Annals of Applied Statistics has shaped several broad research movements. High-dimensional and penalized regression found one of its most important early homes here, with the pathwise coordinate optimization paper establishing the computational paradigm that later powered the lasso, elastic net, and sparse modeling revolution. Bayesian nonparametrics and tree-based methods are another signature theme: BART and random survival forests both debuted in these pages, and the journal continues to publish leading work on Bayesian machine learning.

Genomics and computational biology represent a third pillar. From the irreproducible discovery rate framework to robust hyperparameter estimation for RNA sequencing data, the journal has consistently published methodology that working biologists and bioinformaticians actually adopt. Causal inference and policy evaluation, spatial and environmental statistics, and neuroimaging statistics round out a portfolio that mirrors the data challenges of contemporary science. The current editorial emphasis on statistics and artificial intelligence — including uncertainty quantification for large models and the statistical foundations of deep learning — signals where the journal's next decade of influence is likely to concentrate.

For early-career researchers, the journal's track record offers a useful lesson: the papers that endure are those that pair a sharp methodological idea with a genuine scientific need and demonstrate both with unusual care. A well-executed applied paper here can define a subfield.

Access, Archiving, and Scholarly Infrastructure

The Annals of Applied Statistics is published through the Institute of Mathematical Statistics' partnership with Project Euclid, a nonprofit digital library for mathematics and statistics jointly developed to provide sustainable, community-governed scholarly publishing. Every article from the inaugural 2007 volume onward is available through Project Euclid with persistent identifiers, stable citation metadata, and long-term digital preservation. This infrastructure matters for authors: papers published here remain discoverable and citable indefinitely, and the journal's complete backfile supports systematic reviews and meta-research.

As an IMS journal, the Annals of Applied Statistics operates under the governance of a scholarly society rather than a commercial publisher, which has historically kept the journal's focus on scientific quality rather than volume. Its editorial board and review standards follow IMS policies, and the journal participates in the broader IMS ecosystem of conferences, workshops, and early-career initiatives that support the statistics profession.

Frequently Asked Questions

Who publishes the Annals of Applied Statistics?

The journal is published by the Institute of Mathematical Statistics (IMS), based in the United States.

When was the journal founded, and by whom?

It was established in 2007, with Bradley Efron of Stanford University as founding editor-in-chief.

Who is the current editor-in-chief?

Lexin Li, professor of biostatistics at UC Berkeley, serving a term from January 2025 through December 2027.

What is the journal's impact factor?

The 2022 Journal Citation Reports impact factor is 1.8. The journal's H-index stands at 96.

How often is it published?

Quarterly — four issues per year.

What are the print and online ISSNs?

Print ISSN 1932-6157; online ISSN 1941-7330.

Where can I access the journal's articles?

The journal is hosted on Project Euclid, which carries the complete archive from 2007 onward.

Does the journal charge authors an article processing fee?

No article processing charge is officially published on the journal's pages; authors should consult the current IMS publication policies.

What topics does the journal cover?

All areas of statistics with emphasis on the applied half — methodology motivated by real problems in biomedicine, econometrics, epidemiology, environmental science, finance, genomics, neuroscience, and related fields, plus the emerging statistics–AI interface.

What is the journal's most-cited paper?

"Random survival forests" by Ishwaran, Kogalur, and Blackstone (2008, DOI 10.1214/08-AOAS169), with over 2,500 Crossref citations.