Science AI Journal
Mathematics · Research topic

Open research questions in Statistical Methods and Bayesian Inference

133 unresolved questions extracted from the limitations and future-work sections of 996 Statistical Methods and Bayesian Inference papers in our library. Each links back to the study that raised it.

What the literature leaves open

  • Traditional methods for identifying auxiliary variables do not perform well when missingness follows a nonlinear functional form. The methodological literature recommends including as many potential auxiliary variables as possible, but in practice, that is often infeasible.

    Tutorial: Using Random Forest Analysis to Identify Auxiliary Variables of Missing Data · 2026 · DOI

  • The study identifies a gap in the existing literature on the evaluation of imputation methods and forecasting models for MNAR data. The study highlights the need for more research on the effectiveness of different imputation methods and forecasting models for air pollution data.

    Analysis of Imputation Methods for Missing Not at Random (MNAR) Data: A Comparative Study of Air Pollution Data in Bangkok, Thailand · 2026 · DOI

  • The methodology tested imputation at missing rates up to 70%, but whether KNN maintains effectiveness at even higher missing rates (>70%) remains unexplored.

    Analysis of Imputation Methods for Missing Not at Random (MNAR) Data: A Comparative Study of Air Pollution Data in Bangkok, Thailand · 2026 · DOI

  • The method may not perform well when data comes from simulation conditions that result in items being binary, nonnormal, or not varying enough. The study only investigates a limited number of item correlation structures. The study uses a small sample size.

    Bayesian Estimation of Coefficient Alpha Using a Normal Posterior: Non‑normal Distributions · 2026 · DOI

  • Investigate the method's performance using a larger sample size. Investigate the method's performance using different item correlation structures. Investigate the method's performance using different types of data.

    Bayesian Estimation of Coefficient Alpha Using a Normal Posterior: Non‑normal Distributions · 2026 · DOI

  • Further evaluation of the adapted tree-based imputation methods in real-world datasets. Exploration of other types of missingness mechanisms and more complex data structures. Development of new imputation methods that can handle multilevel data with non-normal distributions.

    Adapting tree-based multiple imputation methods for multilevel data? A simulation study · 2026 · DOI

  • There is a need for effective imputation methods for multilevel data that account for dependencies between observations. Standard imputation methods assume independence between observations, limiting their applicability to multilevel data.

    Adapting tree-based multiple imputation methods for multilevel data? A simulation study · 2026 · DOI

  • Existing maximum likelihood estimation procedures for the negative binomial distribution may fail or produce unstable estimates when the true data-generating process is Poisson. The phenomenon of over-dispersion is frequently observed, and a negative binomial distribution with two parameters is more appropriate.

    From Poisson observations to fitted negative binomial distribution · 2026 · DOI

  • Existing methods cannot accommodate participants with unknown survival status or those who drop out for reasons unrelated to mortality. There is a need for a method that can jointly estimate causal effects on multivariate outcomes and account for different types of incomplete data.

    Bayesian Inference for Cluster‐Randomized Trials With Multivariate Outcomes Subject to Both Truncation by Death and Missingness · 2026 · DOI

  • The lack of a protocol to address heterogeneity in cluster sizes and structures in cluster-randomised trials. The limitations of currently recommended methods in addressing this challenge.

    Cluster trials inference with CARE · 2026 · DOI

  • The increasing availability of large-scale epidemiological studies, such as the UK Biobank, poses important modelling challenges, including mixed data types, high dimensionality, and structured missingness.

    MIDFA: Scalable Bayesian Factor Analysis for Mixed and Incomplete Data · 2026 · DOI

  • Assessing model fit for multiply imputed data is challenging due to the difference in fit across imputed datasets. The study faces the challenge of deriving a relationship between the proposed estimators and existing methods. The performance of the estimators under MNAR is a challenge that requires further research.

    Estimators of the AIC and BIC in multiply imputed data · 2026 · DOI

  • Future studies can investigate the performance of the estimators under MNAR. The derivation of the relationship between the proposed estimators and existing methods can be generalized to models for categorical data. Further research can explore the application of the new estimators in various fields.

    Estimators of the AIC and BIC in multiply imputed data · 2026 · DOI

  • Future research should investigate the performance of record linkage algorithms in different scenarios. Future research should develop new methods for record linkage that take into account the linkage mechanism and the amount of overlap between the files.

    Analysis of Linked Files: A Missing Data Perspective · 2026 · DOI

  • Further research can be conducted to apply the proposed methods to other survival models. The study can be extended to include more complex models and larger datasets.

    Jackknife-based diagnostics for non-monotonic hazard survival model with interval-censored data · 2026 · DOI

  • Traditional approaches often fail to perform well under right or interval censoring. There is a need for effective model diagnostics for the TBPR model with interval-censored data.

    Jackknife-based diagnostics for non-monotonic hazard survival model with interval-censored data · 2026 · DOI

  • The guarantees are conditional on the training fold and remain valid when (, ) are estimated by sample splitting. The results do not claim optimality of the test statistic or uniform tightness of the resulting lower bound across all regimes. Power depends on the choice of , the grid T, and the amount of information about latent heterogeneity in the sampling design.

    Exact one-sided inference for random-effects quantiles in negative binomial regression for ecological count data · 2026 · DOI

  • Investigation of the optimality of the test statistic and uniform tightness of the resulting lower bound across all regimes. Exploration of the use of other calibration approaches to obtain exact finite-sample valid tests. Application of the proposed method to other types of ecological count data.

    Exact one-sided inference for random-effects quantiles in negative binomial regression for ecological count data · 2026 · DOI

  • There is a need to compare the performance of different imputation methods for handling missing data in Bivariate Gamma-generated data. The choice of imputation method depends on the proportion of missing data.

    DATA IMPUTATION FOR BIVARIATE GAMMA-GENERATED DATA USING PREDICTIVE MEAN MATCHING AND RANDOM FOREST METHODS · 2026 · DOI

  • There are no recommendations regarding handling missing data values when using Scalelink. Scalelink cannot currently be used for real-world linkage due to lack of recommendations for handling missing data values.

    Handling missing data when using Goldstein et al.’s Scalelink method of data linkage · 2026 · DOI

  • Overdispersion in count data. The need for a new model that can address this issue. The complexity of estimating the parameters of the ZIPXG INAR(1) model.

    Zero Inflated Poisson Xgamma INAR(1) Model and Its Applications · 2026 · DOI

  • Missing data is common in healthcare due to various reasons. The primary goal in clinical practice is to make safe decisions by minimizing the risk of missing key events. There is a need for more clinically appropriate evaluation methods.

    Decision Framework Focused on Missing-Data Alarms in Healthcare · 2026 · DOI

  • There is a need for more clinically appropriate evaluation methods. Prior studies have focused on minimizing the root mean squared error (RMSE) of imputed missing data rather than decision sufficiency.

    Decision Framework Focused on Missing-Data Alarms in Healthcare · 2026 · DOI

  • The paper identifies a research gap in the lack of diverse subjective priors for parameters in the unit interval [0, 1]. It identifies a gap in the limited use of alternative prior distributions for binomial sampling models. The paper also identifies a gap in the need for a user-friendly tool for eliciting univariate distributions.

    Eliciting univariate priors for binomial sampling models: beyond the beta distribution · 2026 · DOI

  • There is a need for robust methods to extrapolate long-term survival beyond clinical trial duration. Standard parametric models may not accurately capture long-term survival patterns.

    A Simulation and Case Study to Evaluate the Extrapolation Performance of Flexible Bayesian Survival Models when Incorporating Real-World Data · 2026 · DOI

Most-cited papers in Statistical Methods and Bayesian Inference

Most recent work

Find a gap in your own Statistical Methods and Bayesian Inference sub-topic

This page shows what the Statistical Methods and Bayesian Inference literature already flags as unresolved. To narrow it to your specific question, run the guided finder — it searches the gap library on demand and checks candidates against 250M+ OpenAlex works.

Open the Research Gap Finder →

Related topics in Mathematics

Cognitive and developmental aspects of mathematical skills
807 open questions
Mathematics Education and Pedagogy
390 open questions
Statistics Education and Methodologies
379 open questions
History and Theory of Mathematics
322 open questions
Survey Sampling and Estimation Techniques
272 open questions
Mathematics Education and Programs
263 open questions
Advanced Causal Inference Techniques
262 open questions
Advanced Statistical Methods and Models
260 open questions
133 open questions have been extracted from the limitations and future-work passages of 996 Statistical Methods and Bayesian Inference papers in our library. Each one below links back to the study that raised it, so you can read the original claim in context.

Tools for your next paper

Gaps & Literature SearchFind open questions in any field — 120,000+ open gaps mined from our own library.Pre-CheckIs your paper ready? A Tier 1–5 acceptance probability in seconds.Journal FinderA ranked shortlist of target journals from a 17,500-venue index.Duplicate Publication CheckerCheck prior-publication & salami-slicing risk across 8 sources.Citation GeneratorFreeA DOI or title → APA, MLA, Vancouver, Chicago, BibTeX, RIS. No signup.Graphical Abstract MakerFreeTurn your findings into a colorblind-safe graphical abstract.

Compare the category — Honest roundups of the AI research tools, ours listed alongside the alternatives.

Command palette

Jump anywhere, run any action.