Best Practices for Reporting Null Results in Scientific Research

Null results also referred to as non-significant or negative findings are a common outcome of rigorous scientific research, yet they are often underreported or poorly documented. Clear and transparent reporting of null results is essential for reducing publication bias, supporting reproducibility, and enabling accurate evidence synthesis.

This guide outlines best practices for reporting null results in scientific research, with a focus on statistical rigor, interpretability, and research integrity. Authors may also benefit from understanding how reviewers evaluate null results in research.

Why Null Results Matter in Scientific Research

Null results are findings that do not show a statistically significant effect or fail to support a hypothesized relationship, but are essential to cumulative science. When reported clearly and rigorously, they:

  • Reduce publication bias and selective reporting
  • Prevent redundant or misdirected research
  • Improve meta-analyses and theory refinement
  • Strengthen transparency and research integrity

This guide outlines best practices to help authors report null results in a way that is interpretable, reproducible, and maximally informative. When reported rigorously, null results improve meta-analyses, inform policy and practice, and strengthen evidence-based decision-making.

1. Start With a Clear Research Question

A null result is only interpretable when the underlying research question and hypothesis testing framework are clearly specified

Best practices:

  • Clearly state the research question(s) and hypotheses a priori
  • Distinguish between confirmatory and exploratory analyses
  • Explain the theoretical or practical motivation for the hypothesis, even if it was not supported

Avoid framing null results as “nothing happened.” Instead, specify what was tested and what range of effects can be ruled out.

This distinction is especially important for separating confirmatory hypothesis testing from exploratory data analysis

2. Describe the Study Design in Full

Transparent reporting of study design allows reviewers and readers to evaluate internal validity, reproducibility, and methodological rigor

Include:

  • Study type (e.g., randomized experiment, observational study, simulation)
  • Sampling strategy and inclusion/exclusion criteria
  • Randomization and blinding procedures (if applicable)
  • Any deviations from the original protocol

If the study was preregistered, explicitly reference the registration and note any departures.

3. Justify Sample Size and Statistical Power

Null results are most informative when readers can judge whether the study was capable of detecting meaningful effects.

Best practices:

  • Report how sample size was determined (power analysis, feasibility constraints, prior literature)
  • Specify the smallest effect size of interest (SESOI), when possible
  • Clarify whether the study was powered to detect that effect

If the study was underpowered, state this transparently and explain what can—and cannot—be concluded. Reporting statistical power and the smallest effect size of interest (SESOI) helps distinguish between inconclusive findings and evidence supporting a true null effect.

4. Report Results With Estimates, Not Just p-values

A non-significant p-value alone is insufficient.

Always report:

  • Effect size estimates
  • Confidence or credible intervals
  • Exact p-values (when used), not thresholds

Emphasizing effect size estimates and confidence or credible intervals provides more information than binary statistical significance testing

5. Use Appropriate Statistical Approaches for Evaluating Null Results

Null results benefit from methods that directly address evidence for absence of an effect. Certain statistical methods are better suited for assessing evidence for the absence of an effect rather than simply failing to reject the null hypothesis.

When appropriate, consider:

  • Equivalence testing
  • Bayesian analyses with clearly justified priors
  • Model comparison approaches

Explain why the chosen method is suitable for the research question and how results should be interpreted.

6. Distinguish “No Evidence of Effect” From “Evidence of No Effect”

Be precise in language.

  • No evidence of effect indicates insufficient data or sensitivity
  • Evidence of no effect indicates that effects of a meaningful size can be ruled out

Explicitly state which interpretation is supported and why. This distinction is central to accurate interpretation and peer-review of non-significant results in scientific research

7. Report All Analyses Transparently

Selective reporting undermines the value of null results.

Best practices:

  • Report all planned analyses, regardless of outcome
  • Clearly label exploratory or post hoc analyses
  • Avoid reframing hypotheses after seeing the results

Supplementary materials are encouraged for full model specifications, robustness checks, and alternative analyses.Comprehensive reporting supports research transparency, reproducibility, and compliance with modern reporting standards.

8. Contextualize the Null Result Within the Literature

Help readers understand what the result contributes.

Discuss:

  • How the findings relate to prior studies
  • Whether methodological differences might explain discrepancies
  • Implications for theory, practice, or future research

Avoid apologetic framing. A well-executed null result is a result. Contextualization helps prevent misinterpretation and improves cumulative evidence synthesis.

9. Share Data, Code, and Materials When Possible

Open research practices increase the long-term value of null results.

Recommended:

  • Share de-identified data, analysis code, and study materials
  • Use trusted repositories
  • Clearly document files so others can reproduce the analyses

If sharing is not possible, explain the constraints. Data and code sharing increase the long-term value of null results and support independent verification and reuse.

10. Write With Clarity and Neutrality

Tone matters.

  • Avoid language implying failure (e.g., “did not work,” “only null results”)
  • Use precise, neutral descriptions of outcomes
  • Focus on what was learned, not what was “missing”

Null results are a normal and expected part of rigorous research. Neutral scientific writing helps prevent outcome bias and reinforces the legitimacy of null findings.

Reporting Standards, Tools, and Author Resources

Authors reporting null results often rely on statistical software, reporting guidelines, and open research platforms to ensure accuracy and compliance. Familiarity with power analysis tools, Bayesian analysis frameworks, reporting checklists, and data repositories can improve the clarity and impact of null result publications

Final Checklist for Authors

Before submission, ensure that:

  • Research questions and hypotheses are clearly stated
  • Study design and sample size rationale are transparent
  • Effect sizes and uncertainty measures are reported
  • Statistical methods are justified and appropriate
  • Limitations are discussed without undermining valid conclusions
  • Data and code availability are addressed

Closing Note

Publishing null results is a commitment to honest, cumulative science. By following these best practices, authors help ensure that null findings are interpretable, credible, and valuable to the broader research community.

For insight into how these reporting choices are assessed, see our reviewer guide.

We welcome null results that are rigorous, transparent, and thoughtfully reported.

Scroll to Top