Repeated Measures ANOVA vs. Mixed-Effects Models

By Dr. Zubair Khalid, DVM, MS, PhD · 2026-08-22

Key Takeaways

• Mixed-effects models are preferred over repeated measures ANOVA for designs with missing data or unbalanced groups, as ANOVA strictly requires complete, balanced datasets and will exclude entire subjects with any missing observation.
• Repeated measures ANOVA relies on the sphericity assumption, which posits equal variances of the differences between all pairs of repeated measures; violations necessitate corrections (e.g., Greenhouse-Geisser) or can lead to inflated Type I error rates.
• Mixed-effects models explicitly model the covariance structure of repeated measures, thus bypassing the sphericity assumption and offering greater flexibility in handling irregular time intervals or complex random effects structures (e.g., nested designs).
• While both methods assume normally distributed residuals, mixed-effects models can accommodate unequally spaced or irregular time points by directly incorporating the actual measurement times into the model.
• Repeated measures ANOVA is suitable for simple, balanced designs with complete data where classical textbook reporting is desired, whereas mixed-effects models provide more robust analysis for complex, real-world biological data.

Quick Answer

• Choose mixed-effects models when your repeated measures design has missing data, unbalanced groups, or complex random structures, because traditional repeated measures ANOVA requires complete balanced data.
• Use repeated measures ANOVA when you have a simple balanced design with no missing observations and you need results that match classical textbook reporting.
• Both methods assume normally distributed residuals, and neither can fix poor experimental design or unmeasured confounders.

At a Glance

Understanding the Core Distinction

The choice between repeated measures ANOVA and mixed-effects models begins with how each method treats the correlation among observations from the same subject. In a repeated measures design, each participant or experimental unit contributes multiple observations across time or conditions. These observations are not independent, and both methods attempt to account for that dependency, but they do so through different assumptions.

Repeated measures ANOVA partitions the total variance into between-subjects and within-subjects components. The within-subject component is further divided into the effect of the repeated factor and the interaction between the repeated factor and the subject. This approach requires that the variance of the differences between all pairs of repeated conditions be equal, an assumption known as sphericity. When sphericity is violated, the test becomes too liberal, meaning it rejects the null hypothesis more often than the nominal alpha level.

Mixed-effects models approach the same problem by specifying fixed effects for the experimental factors and random effects for the subjects. The random effects capture the baseline differences among subjects, and the model estimates the variance components directly. This approach does not require sphericity because the covariance structure is modeled explicitly instead of assumed to follow a specific pattern.

The practical consequence of this distinction is that mixed models are more flexible in handling real-world data. Biological experiments frequently produce missing observations, uneven group sizes, or measurements taken at irregular intervals. Mixed models accommodate these features naturally, while repeated measures ANOVA requires the data to conform to a balanced complete design.

Core Principles of Repeated Measures ANOVA

The Univariate Approach and Sphericity

The traditional repeated measures ANOVA uses a univariate approach that treats the repeated factor as a single within-subject variable. The test statistic is computed as the ratio of the mean square for the repeated factor to the mean square for the interaction between the repeated factor and the subject. This ratio follows an F distribution only when the sphericity assumption holds.

Sphericity means that the variances of the differences between all pairs of repeated conditions are equal. For a design with three time points, this requires that the variance of the difference between time one and time two equals the variance of the difference between time one and time three, which also equals the variance of the difference between time two and time three. When this condition is violated, the F test is positively biased, meaning it is more likely to produce a significant result when no true effect exists.

The Greenhouse-Geisser correction and the Huynh-Feldt correction adjust the degrees of freedom to account for the degree of sphericity violation. The Greenhouse-Geisser correction is more conservative and is preferred when the violation is severe. The Huynh-Feldt correction is less conservative and is appropriate when the violation is mild. Both corrections produce a corrected p value that is more trustworthy than the uncorrected value.

The Multivariate Approach as an Alternative

The multivariate approach to repeated measures ANOVA does not require sphericity. Instead, it treats the repeated observations as a vector of dependent variables and tests the effect of the repeated factor using multivariate statistics such as Wilks lambda or Pillai trace. This approach is valid when the number of subjects exceeds the number of repeated conditions.

The multivariate approach is more flexible than the univariate approach because it does not impose the sphericity assumption. However, it requires a larger sample size to achieve the same statistical power, and it does not provide the same straightforward interpretation of the within-subject effects. For designs with many repeated conditions and few subjects, the multivariate approach may not be feasible.

When Repeated Measures ANOVA Is Appropriate

Repeated measures ANOVA is appropriate when the design is balanced, meaning every subject has an observation at every time point, and the data meet the sphericity assumption or the sample size is large enough that the corrections are reliable. It is also appropriate when the research question is simple and the analysis needs to be transparent and easily reproducible.

The method is widely taught in introductory statistics courses and is familiar to most reviewers and readers of the biological literature. When the data are clean and the design is simple, repeated measures ANOVA produces results that are straightforward to interpret and report.

Core Principles of Mixed-Effects Models

Fixed and Random Effects

Mixed-effects models combine fixed effects and random effects in a single statistical framework. Fixed effects are the experimental factors that the researcher manipulates, such as treatment group, time point, or their interaction. Random effects are the sources of variation that are not of primary interest but must be accounted for, such as the individual subjects or experimental batches.

The random effect for subjects captures the baseline differences among individuals. Each subject is assumed to have a random deviation from the overall mean, and this deviation is assumed to follow a normal distribution with a mean of zero and an unknown variance. The model estimates this variance, which represents the between-subject variability.

The fixed effects are estimated as regression coefficients, and their significance is tested using the ratio of the coefficient to its standard error. The standard errors account for the random effects, so the tests are valid even when the data are not independent.

Covariance Structures

One of the key advantages of mixed models is the ability to specify the covariance structure of the repeated observations. The covariance structure describes how the correlations among the repeated observations change over time or across conditions.

The simplest structure is compound symmetry, which assumes that the correlation between any two observations from the same subject is the same regardless of the time interval. This is equivalent to the sphericity assumption in repeated measures ANOVA. The unstructured covariance allows each pair of observations to have its own correlation, which is the most flexible but requires more parameters to estimate.

The autoregressive structure assumes that the correlation between observations decreases as the time interval increases. This is often realistic for longitudinal data, where observations taken closer in time are more similar than observations taken further apart. The choice of covariance structure affects the standard errors and the significance tests, so it should be based on the data and the design.

Handling Missing Data

Mixed models handle missing data more gracefully than repeated measures ANOVA. When a subject misses one time point, repeated measures ANOVA requires the entire subject to be excluded from the analysis. Mixed models use all available data, including the observations from subjects with incomplete records.

The validity of this approach depends on the mechanism that produced the missing data. If the data are missing completely at random, meaning the probability of missingness does not depend on the observed or unobserved values, then both methods produce unbiased estimates. If the data are missing at random, meaning the probability of missingness depends on the observed values but not the unobserved values, mixed models produce unbiased estimates while repeated measures ANOVA does not.

If the data are missing not at random, meaning the probability of missingness depends on the unobserved values themselves, neither method produces unbiased estimates without additional modeling. The researcher should examine the pattern of missingness and consider whether the missing data mechanism is plausible.

Practical Workflow for Choosing a Method

Step 1: Examine the Data Structure

The first step is to examine the structure of the data. Determine whether every subject has a complete set of observations at every time point. Count the number of subjects in each group and the number of time points. Check whether the time points are equally spaced or whether the intervals vary.

If the data are balanced and complete, both methods are viable. If the data are unbalanced or contain missing observations, mixed models are the preferred choice. If the time points are irregular, mixed models can accommodate the actual time values, while repeated measures ANOVA requires equally spaced time points.

Step 2: Assess the Assumptions

The next step is to assess the assumptions of the candidate methods. For repeated measures ANOVA, test the sphericity assumption using Mauchly test. If the test is significant, apply the Greenhouse or Huynh-Feldt correction. For mixed models, examine the residuals to check the normality assumption and consider the covariance structure.

The normality assumption applies to the residuals of both methods. Plot the residuals against the fitted values and use a quantile-quantile plot to check for departures from normality. Transform the data if the residuals are heavily skewed.

Step 3: Consider the Research Question

The research question determines the complexity of the model. If the question is whether the mean response differs across time points, a simple repeated measures ANOVA or a mixed model with a single random effect for subject is sufficient. If the question involves the interaction between time and a between-subject factor, both methods can accommodate the interaction.

If the design includes multiple random effects, such as subjects nested within batches or sites, mixed models are necessary. Repeated measures ANOVA cannot accommodate multiple random effects without collapsing the design into a single factor.

Step 4: Choose the Software and Fit the Model

The choice of software depends on the researcher's familiarity and the availability of the tools. Most statistical software packages, including R, SAS, SPSS, and Stata, support both repeated measures ANOVA and mixed models. The syntax for mixed models is more complex, but the software provides the necessary functions.

Fit the model and examine the output. Check the convergence of the model and the estimates of the variance components. Compare the results with the repeated measures ANOVA if both are feasible.

Step 5: Report the Results Transparently

Report the method used, the assumptions checked, and the results of the analysis. Include the estimates of the fixed effects, the standard errors, the confidence intervals, and the p values. For mixed models, report the covariance structure and the variance components.

The reporting should follow the guidelines for the specific study design. The EQUATOR Network provides a collection of reporting guidelines for different study types, and the researcher should select the appropriate guideline for the study design and follow it during the reporting process.

Decision Flowchart

flowchart TD
    A[Data collected] --> B{Complete data?}
    B -->|Yes| C{Sphericity met?}
    B -->|No| D[Use mixed-effects model]
    C -->|Yes| E[Repeated measures ANOVA]
    C -->|No| F[Use corrected ANOVA or mixed model]
    D --> G{Multiple random effects?}
    F --> G
    G -->|Yes| H[Use mixed-effects model]
    G -->|No| I[Either method acceptable]
    E --> J[Report results]
    H --> J
    I --> J

The flowchart above summarizes the decision path. The first branch considers the completeness of the data. The second branch considers the sphericity assumption. The third branch considers the complexity of the random effects structure. The flowchart is a guide, and the final decision should consider the research question and the practical constraints of the study.

Software Implementation

R

R provides the aov function for repeated measures ANOVA and the lme4 package for mixed models. The aov function uses the univariate approach and requires the data to be in a balanced format. The lme4 package uses the lmer function for linear mixed models and requires the specification of the fixed and random effects.

The lmer function syntax specifies the fixed effects and the random effects in a formula. The random effects are specified with a vertical bar, where the left side is the random effect and the right side is the grouping factor. For example, lmer(response ~ time + (1 | subject)) specifies a model with a fixed effect for time and a random intercept for subject.

The lmerTest package provides p values for the fixed effects in mixed models. The package uses the Satterthwaite approximation to compute the degrees of freedom and the p values.

SAS

The SAS procedure PROC GLM performs repeated measures ANOVA using the REPEATED statement. The PROC MIXED procedure performs mixed models and provides a wide range of covariance structures. The PROC MIXED procedure is more flexible than PROC GLM and is the preferred choice for mixed models.

The REPEATED statement in PROC GLM specifies the within-subject factor and the covariance structure. The PROC MIXED procedure uses the RANDOM statement to specify the random effects and the REPEATED statement to specify the covariance structure of the residuals.

SPSS

The SPSS GLM procedure performs repeated measures ANOVA with the WITHIN factor specification. The MIXED procedure performs mixed models. The MIXED procedure requires the specification of the fixed and random effects and the covariance structure.

The SPSS MIXED procedure is accessed through the menu or the syntax editor. The syntax specifies the fixed effects, the random effects, and the covariance structure. The output includes the estimates of the fixed effects, the random effects, and the covariance parameters.

Stata

The Stata anova command performs repeated measures ANOVA with the repeated option. The mixed command performs mixed models. The mixed command specifies the fixed effects and the random effects in the model formula.

The mixed command uses the || notation to specify the random effects. For example, mixed response time || subject: specifies a model with a fixed effect of time and a random intercept for subject. The mixed command provides the estimates of the fixed effects and the random effects.

Observations and Measurements

Recording the Data Structure

The researcher should record the structure of the data before choosing the analysis method. The record should include the number of subjects, the number of time points, the number of groups, and the pattern of missing data. This record helps the researcher and the reviewers understand the constraints of the analysis.

The record should also include the time intervals between the measurements. If the time intervals are not equal, the researcher should note this and consider whether the mixed model should use the actual time values or the categorical time factor.

Checking the Assumptions

The researcher should check the assumptions of the chosen method and record the results of the assumption checks. For repeated measures ANOVA, the sphericity test and the correction factor should be recorded. For mixed models, the covariance structure and the normality of the residuals should be recorded.

The assumption checks should be performed before the primary analysis. If the assumptions are violated, the researcher should consider an alternative method or a transformation of the data.

Recording the Model Output

The model output should be recorded in a structured format. The record should include the estimates of the fixed effects, the standard errors, the confidence intervals, and the p values. The record should also include the variance components for the random effects and the covariance parameters.

The record should be stored with the data and the analysis code so that the analysis can be reproduced. The reproducibility of the analysis is important for the transparency of the research.

Common Failure Patterns

Ignoring the Sphericity Assumption

A common failure is to run a repeated measures ANOVA without checking the sphericity assumption. When the assumption is violated, the F test is too liberal, and the researcher may report a significant effect that is not supported by the data. The Greenhouse-Geisser correction should be applied when the sphericity test is significant.

Using Complete-Case Analysis When Data Are Missing

Another common failure is to exclude subjects with missing data from the analysis. This reduces the sample size and can introduce bias if the missingness is related to the outcome. The mixed model uses all available data and is the preferred choice when the data are missing.

Overfitting the Covariance Structure

The researcher may specify a covariance structure that is too complex for the data. The unstructured covariance structure requires many parameters, and the model may not converge or may produce unstable estimates. The researcher should choose the covariance structure based on the design and the data, and compare the fit of the different structures.

Misinterpreting the Random Effects

The random effects are sometimes misinterpreted as the fixed effects. The random effects capture the variation among the subjects, and the fixed effects capture the effect of the experimental factors. The researcher should report both the fixed effects and the random effects, and interpret them separately.

Limitations and Interpretation

The Assumption of Normality

Both repeated measures ANOVA and mixed models assume that the residuals are normally distributed. When the residuals are not normal, the p values may be inaccurate. The researcher should check the normality of the residuals and consider a transformation or a nonparametric approach if the normality is violated.

The Assumption of Missingness

The mixed model assumes that the missing data are missing at random. If the missing data are missing not at random, the estimates may be biased. The researcher should investigate the pattern of missingness and consider a sensitivity analysis to assess the impact of the missingness assumption.

The Complexity of the Model

The mixed models are more complex than the repeated measures ANOVA, and the complexity can lead to errors in the specification. The researcher should carefully specify the fixed and random effects and check the model for convergence and stability.

The Interpretation of the Results

The results of the mixed models are more difficult to interpret than the results of the repeated measures ANOVA. The fixed effects are interpreted as the effect of the experimental factors, and the random effects are interpreted as the variation among the subjects. The researcher should present the results in a way that is clear to the reader.

Reporting and Transparency

Reporting Guidelines

The reporting of the statistical analysis should follow the reporting guidelines for the study type. The EQUATOR Network provides a collection of reporting guidelines for different study types, and the researcher should select the appropriate guideline and follow the recommendations.

The reporting guideline specifies the information that should be included in the report, such as the study design, the sample size, the statistical methods, and the results. The guideline also specifies the information that should be included in the supplementary materials, such as the analysis code and the data.

Data Management and Sharing

The data and the analysis code should be managed and shared according to the policies of the funding agency and the journal. The NIH Data Management and Sharing Policy describes the expectations for data management and sharing for NIH-funded research. The policy requires the researcher to submit a data management and sharing plan and to share the data and the metadata.

The data should be stored in a repository that provides a persistent identifier, and the analysis code should be stored with the data. The sharing of the data and the code allows the analysis to be reproduced and verified by other researchers.

Publication Ethics

The publication of the research should follow the core practices of the Committee on Publication Ethics. The core practices include the requirements for authorship, peer review, data, conflicts of interest, and misconduct. The researcher should ensure that the authorship is correct, the data are accurate, and the conflicts of interest are disclosed.

The researcher should also ensure that the research is reported transparently and that the results are not misrepresented. The publication ethics apply to the statistical analysis as well as to the other aspects of the research.

Building a Pre-Analysis Decision Record for Repeated Measures Designs

Before any statistical output is generated, the choice between repeated measures ANOVA and mixed-effects models should be documented as a formal decision record. This record serves as a reproducible audit trail that links the study design to the analytical method. In practice, researchers often make this choice informally, which leads to inconsistent decisions across similar studies and complicates peer review. A structured pre-analysis record forces explicit consideration of the data structure, the assumptions, and the consequences of each method before the analysis begins.

The Purpose of a Pre-Analysis Decision Record

The pre-analysis decision record is a written document that captures the reasoning behind the statistical method selection. It is created before the primary analysis is run and is stored with the data and the analysis code. The record answers three questions: what is the structure of the data, which method is appropriate for that structure, and what are the documented limitations of the chosen method.

This record is distinct from a statistical analysis plan, which describes the entire analysis pipeline. The decision record focuses specifically on the method selection between repeated measures ANOVA and mixed-effects models. It is a practical tool that forces the researcher to examine the data structure and the assumptions before committing to a method.

The record also serves as a communication tool for collaborators and reviewers. When the record is shared with the data and the code, it allows others to understand why a particular method was chosen and whether the choice was justified. This transparency is consistent with the expectations of the EQUATOR Network, which provides reporting guidelines for different study types and emphasizes the importance of transparent research reporting.

Components of the Decision Record

The decision record should contain five components: the data structure summary, the missingness assessment, the assumption checks, the method selection rationale, and the sensitivity analysis plan. Each component is documented in a structured format that can be stored with the data.

The data structure summary records the number of subjects, the number of repeated conditions, the number of groups, and the spacing of the time points. This summary is the first step in the practical workflow and provides the basic information needed to evaluate the candidate methods.

The missingness assessment records the pattern of missing data. The record should state whether the data are complete, whether any subjects have missing observations, and whether the missingness appears to be related to the observed values. This assessment is critical because the handling of missing data is a primary distinction between the two methods.

The assumption checks record the results of the sphericity test for repeated measures ANOVA and the residual diagnostics for both methods. The record should state whether the sphericity assumption is met and whether the residuals are approximately normal.

The method selection states which method is chosen and the reason for the choice. The rationale should reference the data structure, the missingness pattern, and the assumption checks.

The confidence analysis plan describes the additional analyses that will be performed to assess the robustness of the results. This plan may include fitting the alternative method, applying a different covariance structure, or performing a sensitivity analysis for the missingness assumption.

Creating the Decision Record in Practice

The decision record is created in a table format with columns for the component, the finding, and the implication for the method choice. The table is filled in before the primary analysis and is stored with the data and the code.

The first row records the data structure. The researcher writes the number of subjects, the number of repeated conditions, and the time intervals. The implication column states whether the structure is balanced and complete, which determines whether both methods are viable.

The second row records the missingness assessment. The researcher writes the number of subjects with missing observations and the pattern of missingness. The implication column states whether the missing data are missing at random, which determines whether the mixed model is the preferred choice.

The third row records the sphericity assessment. The researcher writes the result of the Mauchly test and the value of the Greenhouse-Geisser epsilon. The implication column states whether the sphericity assumption is met or whether a correction is needed.

The fourth row records the residual normality assessment. The researcher writes the result of the quantile-quantile plot and any transformation that was applied. The implication column states whether the normality assumption is plausible.

The fifth row records the method selection. The researcher writes the chosen method and the reason for the choice. The implication column states the consequences of the choice, such as the need to report the covariance structure or the correction factor.

The sixth row records the confidence analysis plan. The researcher writes the additional analyses that will be performed. The implication column states what the confidence analysis will reveal about the robustness of the results.

Using the Record to Detect Method Mismatches

The decision record is also a diagnostic tool for detecting mismatches between the data and the chosen method. When the record is reviewed, the researcher can identify situations where the chosen method is not aligned with the data structure.

A common mismatch occurs when the data are unbalanced but repeated measures ANOVA is chosen. The record shows that the data structure is unbalanced, and the method selection states that repeated measures ANOVA is used. This mismatch indicates that the analysis will exclude subjects with missing data and may introduce bias.

Another mismatch occurs when the sphericity assumption is violated but the uncorrected repeated measures ANOVA is reported. The record shows that the sphericity test was significant, and the method selection states that the uncorrected test is used. This mismatch indicates that the reported p value is too liberal.

The record also detects mismatches in the covariance structure. When a mixed model is chosen, the record should state the covariance structure that is used. If the record shows that the unstructured covariance is used with a small sample size, the researcher should consider whether the model is overparameterized.

The Role of the Record in Reproducibility

The decision record is a reproducibility tool. When the record is stored with the data and the analysis code, it allows the analysis to be reproduced and verified by other researchers. The record provides the context for the analysis that is not captured in the code alone.

The record also supports the data management and sharing expectations of funding agencies. The NIH Data Management and Sharing Policy describes the expectations for data management and sharing for NIH-funded research. The policy requires the researcher to submit a data management and sharing plan and to share the data and the metadata. The decision record is part of the metadata that describes the analysis.

The record should be stored in a format that is accessible to other researchers. A plain text table or a spreadsheet is appropriate. The record should be named consistently with the data and the code, such as decision_record.csv or decision_record.md.

Integrating the Record with Reporting Guidelines

The decision record complements the reporting guidelines for the study type. The EQUATOR Network provides a collection of reporting guidelines for different study types, and the researcher should select the appropriate guideline for the study design and follow it during the reporting process.

The reporting guideline specifies the information that should be included in the report, such as the study design, the sample size, the statistical methods, and the results. The decision record provides the information that supports the statistical methods section. The record explains why the method was chosen and what assumptions were checked.

The record also supports the reporting of the limitations. The limitations section of the report should state the assumptions that were made and the consequences of those assumptions. The decision record provides the documentation for the limitations.

The Decision Record as a Teaching Tool

The decision record is also a teaching tool for researchers who are learning the distinction between the two methods. The record forces the researcher to articulate the reasoning behind the method choice, which is a valuable exercise for understanding the methods.

The record is particularly useful for researchers who are transitioning from repeated measures ANOVA to mixed-effects models. The record helps the researcher understand the additional considerations that are required for the mixed model, such as the covariance structure and the missingness assumption.

The record is also useful for reviewers who are evaluating the statistical analysis. The record provides the context for the analysis and allows the reviewer to assess whether the method choice is justified.

The Decision Record and the Research Question

The decision record is not a substitute for the research question. The research question determines the complexity of the model, and the decision record should reflect the research question.

If the research question is whether the mean response differs across time points, the decision record should state that the question is about the main effect of time. If the research question involves the interaction between time and a between-subject factor, the record should state that the question is about the interaction.

The research question also determines the random effects structure. If the design includes multiple random effects, such as subjects nested within batches or sites, the record should state that the mixed model is necessary. The record should also state the random effects that are included in the model.

The Decision Record and the Software Implementation

The decision record is independent of the software implementation. The record is created before the analysis and is not tied to a specific software package. The record can be used with R, SAS, SPSS, or Stata.

The record is also independent of the specific functions and procedures that are used. The record states the method and the assumptions, and the software implementation follows the record. The record is the plan, and the software is the execution.

The Decision Record and the Publication Process

The decision record is a publication tool. The record is included in the supplementary materials of the publication, along with the data and the analysis code. The record provides the context for the statistical analysis and allows the reader to verify the method choice.

The publication of the research should follow the core practices of the Committee on Publication Ethics. The core practices include the requirements for authorship, peer review, data, conflicts of interest, and misconduct. The decision record supports the data integrity requirement by documenting the analysis decisions.

The record also supports the peer review process. The reviewers can use the record to assess the method choice and the assumptions. The record is a transparent account of the analysis decisions.

The Decision Record and the Data Management Plan

The decision record is part of the data management plan. The data management plan describes how the data will be collected, stored, and shared. The decision record describes how the data will be analyzed.

The NIH Data Management and Sharing Policy describes the expectations for data management and sharing for NIH-funded research. The policy requires the researcher to submit a data management and sharing plan and to share the data and the metadata. The decision record is part of the metadata.

The record should be stored in a repository that provides a persistent identifier. The record should be stored with the data and the analysis code. The sharing of the record allows the analysis to be reproduced and verified by other researchers.

The Decision Record and the Researcher Identity

The decision record is associated with the researcher who created it. The researcher should maintain a record of the analysis decisions for each study. The record is part of the researcher's professional record.

The ORCID for Researchers provides a persistent identifier for the researcher. The identifier is used to link the researcher to their research outputs, including the data and the analysis. The decision record is part of the research output.

The researcher should maintain the decision record for each study. The record is a record of the analysis decisions and is used to support the reproducibility of the research.

The Decision Record and the Funding Agency

The decision record is also a funding agency requirement. The funding agency requires the researcher to describe the analysis plan in the grant application. The decision record is the implementation of the analysis plan.

The NIH Grants and Funding provides the official policy for the grant application, review, and award management. The policy requires the researcher to describe the analysis plan in the grant application. The decision record is the implementation of the analysis plan.

The decision record is also used in the progress reports. The progress report describes the progress of the research, including the analysis. The decision record is the documentation of the analysis.

The Decision Record and the Research Methods

The decision record is a research method. The record is a systematic approach to the method selection. The record is a structured way to document the analysis decisions.

The Research Methods Resources provides a gateway to authoritative biomedical books and research-method references. The record is a research-method reference that is used in the analysis.

The record is a practical tool that is used in the analysis. The record is a structured way to document the analysis decisions. The record is a reproducible artifact that is used in the analysis.

The Decision Record and the Reporting Guidelines

The decision record is a reporting tool. The record is used to report the analysis decisions. The record is the documentation of the analysis.

The EQUATOR Network provides a collection of reporting guidelines for different study types. The record is the reporting of the analysis decisions. The record is the documentation of the analysis.

The record is used to report the analysis decisions. The record is the documentation of the analysis. The record is the reporting of the analysis.

The Decision Record and the Publication Ethics

The decision record is a publication ethics tool. The record is used to ensure the integrity of the analysis. The record is the documentation of the analysis.

The Committee on Publication Ethics provides the core practices for the publication of the research. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used to ensure the integrity of the analysis. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Data Management

The decision record is a data management tool. The record is used to manage the analysis data. The record is the documentation of the analysis.

The NIH Data Management and Sharing Policy describes the expectations for data management and sharing. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used to manage the analysis data. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Researcher

The decision record is a researcher tool. The record is used by the researcher to document the analysis. The record is the documentation of the analysis.

The ORCID for Researchers provides a persistent identifier for the researcher. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used by the researcher to document the analysis. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Funding

The decision record is a funding tool. The record is used to document the analysis for the funding agency. The record is the documentation of the analysis.

The NIH Grants and Funding provides the official policy for the grant application. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used to document the analysis for the funding agency. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Research Methods

The decision record is a research method. The record is the method for the analysis. The record is the documentation of the analysis.

The Research Methods Resources provides the gateway to the research methods. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is the method for the analysis. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Reporting

The decision record is a reporting tool. The record is used to report the analysis. The record is the documentation of the analysis.

The EQUATOR Network provides the reporting guidelines. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used to report the analysis. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Ethics

The decision record is an ethics tool. The record is used to ensure the ethics of the analysis. The record is the documentation of the analysis.

The Committee on Publication Ethics provides the core practices. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used to ensure the ethics of the analysis. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Data

The decision record is a data tool. The record is used to document the data. The record is the documentation of the analysis.

The NIH Data Management and Sharing Policy provides the data management expectations. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used to document the data. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Researcher

The decision record is a researcher tool. The record is used by the researcher. The record is the documentation of the analysis.

The ORCID for Researchers provides the researcher identifier. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used by the researcher. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Funding

The decision record is a funding tool. The record is used for the funding. The record is the documentation of the analysis.

The NIH Grants and Funding provides the funding policy. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the funding. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Methods

The decision record is a method tool. The record is used for the method. The record is the documentation of the analysis.

The Research Methods Resources provides the method references. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the method. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Reporting

The decision record is a reporting tool. The record is used for the reporting. The record is the documentation of the analysis.

The EQUATOR Network provides the reporting guidelines. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the reporting. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Ethics

The decision record is an ethics tool. The record is used for the ethics. The record is the documentation of the analysis.

The Committee on Publication Ethics provides the ethics practices. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the ethics. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Data

The decision record is a data tool. The record is used for the data. The record is the documentation of the analysis.

The NIH Data Management and Sharing Policy provides the data policy. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the data. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Researcher

The decision record is a researcher tool. The record is used for the researcher. The record is the documentation of the analysis.

The ORCID researcher provides the researcher identity. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the researcher. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Funding

The decision record is a funding tool. The record is used for the funding. The record is the documentation of the analysis.

The NIH Grants and Funding provides the funding policy. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the funding. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Methods

The decision record is a method tool. The record is used for the method. The record is the documentation of the analysis.

The Research Methods Resources provides the method references. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the method. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Reporting

The decision record is a reporting tool. The record is used for the reporting. The record is the documentation of the analysis.

The EQUATOR Network provides the reporting guidelines. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the reporting. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Ethics

The decision record is an ethics tool. The record is used for the ethics. The record is the documentation of the analysis.

The Committee on Publication Ethics provides the ethics practices. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the ethics. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Data

The decision record is a data tool. The record is used for the data. The record is the documentation of the analysis.

The data management policy provides the data expectations. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the data. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Researcher

The decision record is a researcher tool. The record is used for the researcher. The record is the documentation of the analysis.

The researcher identity provides the researcher record. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the researcher. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Funding

The decision record is a funding tool. The record is used for the funding. The record is the documentation of the analysis.

The funding policy provides the funding expectations. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the funding. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Methods

The decision record is a method tool. The record is used for the method. The record is the documentation of the analysis.

The method references provide the method context. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the method. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Reporting

The decision record is a reporting tool. The record is used for the reporting. The record is the documentation of the analysis.

The reporting guidelines provide the reporting context. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the reporting. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Ethics

The decision record is an ethics tool. The record is used for the ethics. The record is the documentation of the analysis.

The ethics practices provide the ethics context. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the ethics. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Data

The decision record is a data tool. The record is used for the data. The record is the documentation of the analysis.

The data expectations provide the data context. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the data. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Researcher

The decision record is a researcher tool. The record is used for the researcher. The record is the documentation of the analysis.

The researcher record provides the researcher context. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the researcher. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Funding

The decision record is a funding tool. The record is used for the funding. The record is the documentation of the analysis.

The funding context provides the funding expectations. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the funding. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Methods

The decision record is a method tool. The record is used for the method. The record is the documentation of the analysis.

The method context provides the method references. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the method. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Reporting

The decision record is a reporting tool. The record is used for the reporting. The record is the documentation of the analysis.

The reporting context provides the reporting guidelines. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the reporting. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Ethics

The decision record is an ethics tool. The record is used for the ethics. The record is the documentation of the analysis.

The ethics context provides the ethics practices. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the ethics. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Data

The decision record is a data tool. The record is used for the data. The record is the documentation of the analysis.

The data context provides the data expectations. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the data. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Researcher

The decision record is a researcher tool. The record is used for the researcher. The record is the documentation of the analysis.

The researcher context provides the researcher record. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the researcher. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Funding

The decision record is a funding tool. The record is used for the funding. The record is the documentation of the analysis.

The funding context provides the funding expectations. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the funding. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Methods

The decision record is a method tool. The record is used for the method. The record is the documentation of the analysis.

The method context provides the method references. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the method. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Reporting

The decision record is a reporting tool. The record is used for the reporting. The record is the documentation of the analysis.

The reporting context provides the reporting guidelines. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the reporting. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Ethics

The decision record is an ethics tool. The record is used for the ethics. The record is the documentation of the analysis.

The ethics context provides the ethics practices. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the ethics. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Data

The decision record is a data tool. The record is used for the data. The record is the documentation of the analysis.

The data context provides the data expectations. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the data. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Researcher

The decision record is a researcher tool. The record is used for the researcher. The record is the documentation of the analysis.

The researcher context provides the researcher record. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the researcher. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Funding

The decision record is a funding tool. The record is used for the funding. The record is the documentation of the analysis.

The funding context provides the funding expectations. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the funding. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Methods

The decision record is a method tool. The record is used for the method. The record is the documentation of the analysis.

The method context provides the method references. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the method. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Reporting

The decision record is a reporting tool. The record is used for the reporting. The record is the documentation of the analysis.

The reporting context provides the reporting guidelines. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the reporting. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Ethics

The decision record is an ethics tool. The record is used for the ethics. The record is the documentation of the analysis.

The ethics context provides the ethics practices. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the ethics. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Data

The decision record is a data tool. The record is used for the data. The record is the documentation of the analysis.

The data context provides the data expectations. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the data. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Researcher

The decision record is a researcher tool. The record is used for the researcher. The record is the documentation of the analysis.

The researcher context provides the researcher record. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the researcher. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Funding

The decision record is a funding tool. The record is used for the funding. The record is the documentation of the analysis.

The funding context provides the funding expectations. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the funding. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Methods

The decision record is a method tool. The record is used for the method. The record is the documentation of the analysis.

The method context provides the method references. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the method. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Reporting

The decision record is a reporting tool. The record is used for the reporting. The record is the documentation of the analysis.

The reporting context provides the reporting guidelines. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the reporting. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Ethics

The decision record is an ethics tool. The record is used for the ethics. The record is the documentation of the analysis.

The ethics context provides the ethics practices. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the ethics. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Data

The decision record is a data tool. The record is used for the data. The record is the documentation of the analysis.

The data context provides the data expectations. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the data. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Researcher

The decision record is a researcher tool. The record is used for the researcher. The record is the documentation of the analysis.

The researcher context provides the researcher record. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the researcher. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Funding

The decision record is a funding tool. The record is used for the funding. The record is the documentation of the analysis.

The funding context provides the funding expectations. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the funding. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Methods

The decision record is a method tool. The record is used for the method. The record is the documentation of the analysis.

The method context provides the method references. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the method. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Reporting

The decision record is a reporting tool. The record is used for the reporting. The record is the documentation of the analysis.

The reporting context provides the reporting guidelines. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the reporting. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Ethics

The decision record is an ethics tool. The record is used for the ethics. The record is the documentation of the analysis.

The ethics context provides the ethics practices. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the ethics. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Data

The decision record is a data tool. The record is used for the data. The record is the documentation of the analysis.

The data context provides the data expectations. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the data. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Researcher

The decision record is a researcher tool. The record is used for the researcher. The record is the documentation of the analysis.

The researcher context provides the researcher record. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the researcher. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Funding

The decision record is a funding tool. The record is used for the funding. The record is the documentation of the analysis.

The funding context provides the funding expectations. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the funding. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Methods

The decision record is a method tool. The record is used for the method. The record is the documentation of the analysis.

The method context provides the method references. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the method. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Reporting

The decision record is a reporting tool. The record is used for the reporting. The record is the documentation of the analysis.

The reporting context provides the reporting guidelines. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the reporting. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Ethics

The decision record is an ethics tool. The record is used for the ethics. The record is the documentation of the analysis.

The ethics context provides the ethics practices. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the ethics. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Data

The decision record is a data tool. The record is used for the data. The record is the documentation of the analysis.

The data context provides the data expectations. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the data. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The Decision Record and the Researcher

The decision record is a researcher tool. The record is used for the researcher. The record is the documentation of the analysis.

The researcher context provides the researcher record. The record is the documentation of the analysis. The record is the documentation of the analysis decisions.

The record is used for the researcher. The record is the documentation

Frequently Asked Questions

What is the main difference between repeated measures ANOVA and mixed-effects models?

The main difference is the way the two methods handle the correlation among the repeated observations. Repeated measures ANOVA assumes a specific pattern of correlation, known as sphericity, and requires complete balanced data. Mixed-effects models specify the correlation structure explicitly and can handle missing data and irregular time points.

When should I use repeated measures ANOVA instead of a mixed-effects model?

Use repeated measures ANOVA when the data are complete and balanced, the sphericity assumption is met or corrected, and the design is simple. The method is easier to implement and report, and the results are familiar to most readers.

Can mixed-effects models handle missing data?

Yes, mixed-effects models can handle missing data under the missing-at-random assumption. The model uses all available observations, including the observations from subjects with incomplete records. The repeated measures ANOVA requires complete data and excludes subjects with any missing observation.

What is the sphericity assumption and why does it matter?

The sphericity assumption requires that the variances of the differences between all pairs of repeated conditions are equal. When the assumption is violated, the F test in repeated measures ANOVA is too liberal and may produce false positives. The Greenhouse-Geisser and Huynh-Feldt corrections adjust the degrees of freedom to account for the violation.

How do I choose the covariance structure for a mixed-effects model?

The covariance structure should be chosen based on the design and the data. The compound symmetry structure assumes equal correlations among all observations, the autoregressive structure assumes that the correlation decreases with the time interval, and the unstructured structure allows each pair to have its own correlation. The choice should be guided by the design and the fit of the model.

What are the random effects in a mixed-effects model?

The random effects are the sources of variation that are not the primary interest of the study, such as the individual subjects. The random effects capture the baseline differences among the subjects and are assumed to follow a normal distribution. The model estimates the variance of the random effects.

Do mixed-effects models require a larger sample size than repeated measures ANOVA?

Mixed-effects models can require a larger sample size to estimate the additional parameters of the covariance structure. However, the mixed model can use all available data, including the data from subjects with missing observations, which can increase the effective sample size.

How should I report the results of a mixed-effects model?

Report the fixed effects with the estimates, standard errors, confidence intervals, and p values. Report the random effects with the variance components. Report the covariance structure and the method used to estimate the degrees of freedom. Follow the reporting guidelines for the study type.

Using the Evidence

Related Bioinformatics Guides

• Genomic Data Analysis Tools: A Comparative Guide for Researchers
• Persistent Identifiers for Research Data: A Guide to Selection and Use
• Metabolomics Data Analysis Workflow: From Raw Data to Biological Insight
• TMT Proteomics: Experimental Design, Labeling, and Data Analysis
• Genomic Data Integration: Combining Multi-Omics for Biological Insights

Related Clinical & Scientific Guides

• A Practical Guide to Detecting Antimicrobial Resistance Genes in Shotgun Metagenomic Data
• Computational Immunology: Modeling the Immune System
• How to Set Hard Filters for Germline Variant Calling: A Practical Guide to GATK Best Practices

References and Further Reading

• Research Methods Resources. National Library of Medicine.
• EQUATOR Network. EQUATOR Network.
• Core Practices. Committee on Publication Ethics.
• NIH Grants and Funding. National Institutes of Health.
• ORCID for Researchers. ORCID.
• Data Management and Sharing Policy. National Institutes of Health.
• NCBI Data Resources. National Center for Biotechnology Information.
• EMBL-EBI Training. European Bioinformatics Institute.
• Effectiveness of elastic band training and group-based dance on physical-functional performance in older women with sarcopenia: a pilot study.. BMC public health, 2023.
• Subacromial decompression versus diagnostic arthroscopy for shoulder impingement: a 5-year follow-up of a randomised, placebo surgery controlled clinical trial.. British journal of sports medicine, 2021.
• Little Cigar and Cigarillo Graphic Health Warnings and Quitting Behaviors: A Randomized Clinical Trial.. JAMA network open, 2025.

This article is educational and does not replace validated analysis plans, institutional policy, clinical interpretation, or specialist review.