Research Findings – Definition, Types and Examples

Research findings are the evidence-based outcomes identified after research data have been collected, processed, and analyzed. They show what the study discovered in relation to its questions, objectives, or hypotheses. Findings may consist of numerical estimates, differences, relationships, themes, patterns, explanations, or integrated insights, depending on the research design.

Introduction

Research findings are the part of a study that tells readers what the investigation actually discovered. They connect the research questions to the evidence and provide the foundation for the discussion, conclusions, recommendations, and future research.

However, “findings” is not used in exactly the same way in every discipline. In many quantitative journal articles, the term is largely interchangeable with “results.” In qualitative research, findings may include analytically developed themes and a degree of interpretation. Some dissertations separate findings and discussion, while others combine them.

This guide explains what research findings are, how they differ from related parts of a study, which types may be reported, and how to write quantitative, qualitative, and mixed-methods findings accurately.

Key Takeaways

• Research findings are outcomes supported by analyzed evidence, not personal opinions.
• Findings should address the study’s research questions, objectives, or hypotheses.
• Results and findings are often interchangeable, but disciplinary conventions differ.
• Quantitative findings should report estimates and uncertainty, not p-values alone.
• Qualitative findings should connect themes to a documented analysis and supporting evidence.
• Null, negative, unexpected, and inconclusive findings should not be hidden.

What Are Research Findings?

Definition of research findings

Research findings are the patterns, relationships, differences, themes, estimates, explanations, or other outcomes identified through the systematic analysis of research data.

A finding is more than a piece of raw information. It is an analytically supported statement about what the evidence shows.

For example, a spreadsheet containing 500 survey responses is raw data. A calculation showing that respondents who received one form of instruction had higher average scores than another group is a result. After checking the size, uncertainty, design, and relevance of that difference, the researcher may state it as a key finding.

In qualitative research, interview transcripts are raw data. Codes developed from the transcripts are analytical labels. A recurring and well-supported theme identified across participants may become a research finding.

Why research findings matter

Findings serve several purposes:

• They answer the research questions.
• They show whether the collected evidence supports the hypotheses.
• They provide the factual or analytical basis for the discussion.
• They allow readers to assess the study’s claims.
• They contribute evidence that may be compared or synthesized with other studies.
• They may inform theory, policy, practice, or future investigation.

A paper with strong conclusions but poorly reported findings asks readers to trust claims without seeing an adequate evidential foundation.

Research Findings Versus Data, Results, Discussion, and Conclusion

The terms surrounding research findings are related but not identical.

Term	Meaning	Typical example
Raw data	Information collected before full analysis	Survey responses, recordings, measurements, field notes
Analysis	Procedures used to organize, test, compare, model, code, or interpret data	Regression, thematic analysis, content analysis
Result	A direct output of analysis	A mean difference, confidence interval, coefficient, or identified theme
Finding	A result or group of results stated in relation to the research question	The intervention group improved more than the comparison group
Discussion	Interpretation of what the findings mean and how they relate to other knowledge	Possible explanations, implications, comparisons with previous studies
Conclusion	A synthesized answer based on the findings, discussion, and study limitations	The intervention appears promising but requires evaluation in broader samples

Are findings and results the same?

Findings and results are often used interchangeably, especially in quantitative research. However, “results” may refer more narrowly to analytical outputs, while “findings” may refer to the most meaningful evidence-based outcomes selected and explained in relation to the research questions.

This is not a universal rule. Researchers should follow the conventions of their discipline, institution, journal, and methodology.

Findings versus discussion

The findings section primarily tells readers what the analysis showed. The discussion explains what those findings may mean, why they matter, how they compare with previous research, and what limitations affect their interpretation.

The boundary is not equally strict in every field. Experimental papers often maintain a clear separation. Ethnographies, case studies, and some qualitative dissertations may combine evidence and interpretation because meaning is developed through the analytical narrative.

Findings versus conclusions

A finding is a specific evidence-based outcome. A conclusion is a broader judgment or answer formed by considering multiple findings, the research objectives, limitations, and existing knowledge.

A study can have several findings but one overarching conclusion.

Types of Research Findings

Research findings can be classified in more than one way. The most useful classifications concern the form of evidence, analytical purpose, and relationship to prior expectations.

Findings by Research Method

Quantitative findings

Quantitative findings are based on numerical data and statistical analysis. They may describe a sample, estimate a population value, compare groups, examine relationships, test a model, or evaluate an intervention.

Examples include:

• The average score in a sample.
• A difference between treatment and comparison groups.
• A correlation between two variables.
• A regression coefficient.
• An estimated risk ratio.
• A confidence interval around an effect.
• A change over time.

Good quantitative findings normally report more than whether a test was “significant.” They describe the direction and size of the result, the uncertainty surrounding it, the relevant sample, and any assumptions or limitations that affect interpretation.

Qualitative findings

Qualitative findings are analytically developed patterns, themes, categories, narratives, concepts, or explanations derived from non-numerical material such as interviews, observations, documents, images, or field notes.

Examples include:

• A theme describing how participants experience a service.
• A process model explaining how people make a decision.
• A set of categories describing organizational responses.
• A narrative showing how an identity changes over time.
• A comparison of experiences across contexts.

A qualitative finding should be traceable to the analytic process. A memorable quotation alone is not a finding. Quotations, observations, or document extracts provide evidence that illustrates and supports the researcher’s analysis.

Mixed-methods findings

Mixed-methods findings result from the purposeful integration of quantitative and qualitative evidence.

Integration may show that:

• The two forms of evidence converge.
• Qualitative evidence explains a numerical pattern.
• One strand expands the scope of the other.
• The strands contradict each other.
• One method reveals an issue that the other method missed.

Simply placing a survey result beside an interview theme is not necessarily integration. The researcher must explain what is learned by considering them together.

Findings by Analytical Purpose

Descriptive findings

Descriptive findings summarize what was observed without claiming an explanation or causal effect.

Examples include frequencies, percentages, means, medians, distributions, participant characteristics, recurring categories, and descriptions of a setting.

Comparative findings

Comparative findings identify similarities or differences between groups, cases, time periods, conditions, or settings.

A difference may be descriptively important even when statistical uncertainty remains high. Conversely, a statistically detectable difference may be too small to matter practically.

Relational or associational findings

These findings describe how variables or concepts are related.

A correlation or regression association does not automatically show that one variable caused the other. Reverse causation, confounding, selection effects, measurement error, or model assumptions may provide alternative explanations.

Causal findings

Causal findings claim that a change in one factor produced a change in another.

Such claims require a design and analysis capable of supporting causal inference. Randomization is one strong approach, but causal inference may also use carefully justified natural experiments, quasi-experimental designs, longitudinal methods, instrumental variables, regression discontinuity, or other identification strategies.

Causal language should match the strength of the design. “Was associated with” is more appropriate than “caused” when causal assumptions cannot be defended.

Exploratory findings

Exploratory findings identify unanticipated patterns, relationships, categories, or questions that were not specified as primary tests in advance.

They can be valuable, but researchers should label them as exploratory rather than presenting them as preplanned confirmation.

Synthesis findings

Systematic reviews and evidence syntheses generate findings by combining or comparing evidence from multiple studies.

Findings may include:

• A pooled effect estimate.
• The direction and consistency of evidence.
• Heterogeneity across studies.
• Themes synthesized from qualitative studies.
• Gaps in the available literature.
• An assessment of confidence or certainty in the evidence.

Findings by Relationship to Expectations

Confirmatory findings

These are outcomes from analyses designed in advance to evaluate a stated hypothesis or prediction.

A confirmatory finding does not “prove” the hypothesis. It reports how compatible the evidence is with the prediction under the design and analytical assumptions.

Null findings

A null finding occurs when the analysis does not provide clear evidence of the anticipated difference, relationship, or effect.

A null finding does not necessarily prove that no effect exists. The estimate may be imprecise, the sample may be small, the measurement may be weak, or the plausible effect may be smaller than the study could detect.

Report the estimate and uncertainty rather than merely stating “there was no significant effect.”

Negative findings

“Negative finding” may describe a result in the opposite direction from the hypothesis or a study that did not support the anticipated outcome. Because the term is ambiguous, the exact result should be stated directly.

Unexpected findings

Unexpected findings were not anticipated by the hypothesis, theory, or previous evidence.

They should be reported honestly and then examined for possible explanations such as chance variation, data quality problems, subgroup differences, model misspecification, contextual factors, or a genuinely new pattern.

Inconclusive findings

A finding is inconclusive when the available evidence does not support a sufficiently precise or defensible answer.

Calling a result inconclusive is often more accurate than forcing it into a positive-versus-negative category.

How Are Research Findings Produced?

Research findings do not appear automatically when data collection ends. They emerge through a chain of decisions:

1. The researcher defines questions, objectives, outcomes, or phenomena of interest.
2. A design is selected to generate relevant evidence.
3. Data are collected and documented.
4. Data quality is checked.
5. An analytical method is applied.
6. Analytical outputs are examined for patterns and uncertainty.
7. Results are evaluated in relation to the research questions.
8. Key evidence-based outcomes are selected as findings.
9. The findings are communicated with enough detail for readers to assess them.

Weakness at any point in this chain can affect the findings. Excellent writing cannot repair an inappropriate design, invalid measurement, poor coding process, or unsuitable statistical model.

How to Write Research Findings

1. Revisit the research questions and objectives

Place each research question, objective, hypothesis, or predefined outcome beside the relevant analysis.

Ask:

• Which analysis addresses this question?
• What was the main outcome?
• Which evidence supports the finding?
• How certain is the estimate or interpretation?
• Was the analysis planned or exploratory?

This alignment prevents the findings section from becoming a list of unrelated outputs.

2. Verify the data and analysis

Before drafting, check:

• Data-cleaning decisions.
• Missing-data handling.
• Coding consistency.
• Statistical assumptions.
• Model specifications.
• Outliers and influential observations.
• Transcription accuracy.
• Theme development.
• Changes from the original analysis plan.
• Reproducibility of tables and figures.

A polished finding based on an unchecked error is still wrong.

3. Distinguish analytical output from a meaningful finding

Statistical software can produce hundreds of values. Qualitative software can display extensive code lists. Not every output is a finding.

Select outcomes that:

• Address the research questions.
• Are supported by adequate evidence.
• Are necessary for understanding the study.
• Represent important expected or unexpected patterns.
• Affect interpretation of the main results.

Secondary analyses should be identified as secondary or exploratory.

4. Choose a logical organizing principle

Common structures include:

• Research question by research question.
• Hypothesis by hypothesis.
• Primary outcome followed by secondary outcomes.
• Theme and subtheme.
• Case or setting.
• Chronological stage.
• Quantitative strand, qualitative strand, and integration.
• Convergent and divergent evidence.

The structure should help readers connect each finding to the purpose of the study.

5. Begin each subsection with the finding

State the main outcome before presenting every supporting number or extract.

A useful sequence is:

1. Direct finding statement.
2. Supporting evidence.
3. Reference to a table, figure, or extract.
4. Brief clarification of scope or uncertainty.
5. Transition to the next finding.

6. Report the evidence required to evaluate the finding

For quantitative research, this may include:

• Sample size.
• Group summaries.
• Effect estimate.
• Confidence or credible interval.
• Test statistic and degrees of freedom where relevant.
• Exact p-value where required.
• Sensitivity or adjusted analysis.
• Missing-data information.

For qualitative research, this may include:

• Theme or category description.
• Supporting extracts or observations.
• Variation within the theme.
• Divergent cases.
• Context required to understand participants’ accounts.
• Explanation of how themes relate to the research question.

7. Communicate uncertainty

All findings have limits.

Quantitative uncertainty may arise from sampling variability, measurement error, missing data, model assumptions, or bias.

Qualitative uncertainty may involve the adequacy and diversity of the dataset, interpretive choices, contextual dependence, researcher positioning, and the extent to which alternative interpretations were examined.

Do not write with more certainty than the evidence supports.

8. Include null, contradictory, and unexpected findings

Selective reporting distorts the study.

Include relevant findings even when they:

• Do not support the hypothesis.
• Conflict with previous research.
• Complicate the expected explanation.
• Show substantial uncertainty.
• Reveal no clear difference.
• Differ across methods or subgroups.

Explain their possible meaning later in the discussion unless the reporting convention combines findings and interpretation.

9. Use tables, figures, and quotations selectively

Use a table when readers need exact values or structured comparisons.

Use a figure when readers need to see a trend, distribution, model, process, or relationship.

Use participant quotations or observational extracts when they provide clear evidence for a qualitative theme.

Do not repeat every table entry in the text. The text should direct readers to the most important pattern and explain how the visual supports the finding.

10. Check the relevant reporting guideline

Reporting requirements differ by design and field. Before submission, identify the guideline expected by the journal, department, professional association, or funder.

A reporting checklist should improve completeness, not become a substitute for methodological judgment.

How to Report Quantitative Research Findings

Report the sample and analytical population

Readers should know how many observations were analyzed and whether that number differs from the original sample.

Report exclusions, attrition, missing observations, and the analytical population used for each major analysis when relevant.

Present descriptive statistics

Descriptive statistics establish what the data look like.

Depending on the variable, report:

• Frequencies and percentages.
• Means and standard deviations.
• Medians and interquartile ranges.
• Ranges.
• Distributional information.
• Baseline or group characteristics.

Choose summaries that match the data. A highly skewed variable may be described more meaningfully with a median and interquartile range than with a mean alone.

Report effect estimates

An effect estimate describes the size and direction of a difference, association, or change.

Examples include:

• Mean difference.
• Standardized mean difference.
• Risk difference.
• Risk ratio.
• Odds ratio.
• Correlation coefficient.
• Regression coefficient.
• Rate ratio.
• Hazard ratio.

The most appropriate estimate depends on the design, outcome, and analytical model.

Include uncertainty intervals

A confidence interval or credible interval shows the range of values compatible with the model and data under specified assumptions.

Intervals help readers assess precision. A wide interval may indicate that the evidence is compatible with both a meaningful effect and little or no effect.

Use p-values carefully

A p-value does not measure:

• The probability that the hypothesis is true.
• The size of an effect.
• The practical importance of a finding.
• The quality of the study.
• Whether the result will replicate.

When p-values are required, report them alongside effect estimates, uncertainty intervals, and relevant assumptions. Avoid reducing the entire finding to “significant” or “not significant.”

Report multiple and exploratory analyses transparently

Running many analyses increases the chance of obtaining apparently notable results by chance.

Identify:

• Primary analyses.
• Secondary analyses.
• Subgroup analyses.
• Sensitivity analyses.
• Exploratory analyses.
• Any adjustment for multiple comparisons.

Do not describe an exploratory result as though it were a prespecified confirmatory test.

Hypothetical quantitative example

The following figures are illustrative rather than results from a real study:

Students using a spaced-practice program achieved a higher mean assessment score than students receiving the standard revision schedule. The estimated mean difference was 5.3 points, with a 95% confidence interval from 2.1 to 8.5 points. The standardized difference was moderate. These results support a difference in this sample, although application to other courses or student populations would require further evidence.

This statement communicates the direction, magnitude, precision, scope, and limitation of the finding.

How to Report Qualitative Research Findings

Organize findings around analytical themes or concepts

Each major theme should represent an analytically meaningful pattern rather than a broad topic label.

A strong theme:

• Addresses the research question.
• Has a clear central idea.
• Is supported across relevant parts of the dataset.
• Acknowledges meaningful variation.
• Is distinct from other themes.
• Is explained, not merely named.

Explain the theme before adding quotations

Begin with an analytic statement that tells readers what the theme means.

Then use selected quotations, observations, or document extracts to show how that interpretation is grounded in the data.

Avoid presenting a sequence of quotations with little analytical explanation.

Preserve context

A quotation can be misleading when detached from the participant’s circumstances, the question asked, or the wider account.

Provide enough context for readers to understand why the extract supports the theme.

Include variation and divergent cases

Not every participant must express the same view for a theme to be meaningful.

Report:

• Common patterns.
• Important differences.
• Exceptions.
• Contradictory cases.
• Contexts in which the pattern changes.

Divergent cases may refine the finding and prevent an overly simple account.

Use counts cautiously

Counts can sometimes clarify qualitative patterns, but they should not automatically be treated as statistical prevalence estimates.

A statement such as “most participants mentioned…” should be used only when the analytic approach and sampling make that description appropriate. In many qualitative designs, importance is not determined by frequency alone.

Address researcher interpretation

Qualitative findings are shaped by analytical decisions. Researchers should explain the approach used to code, compare, interpret, and refine the data.

Reflexivity helps readers understand how the researcher’s position, assumptions, relationships, and decisions may have influenced the analysis.

Hypothetical qualitative example

In a hypothetical interview study of remote postgraduate learning, analysis might produce the following theme:

Flexibility created both control and disconnection. Participants valued being able to arrange study around employment and family responsibilities. At the same time, several accounts described reduced informal contact with peers and difficulty developing a sense of academic belonging. The theme therefore reflected a tension rather than a uniformly positive or negative experience.

This example states the analytical idea, describes supporting patterns, and preserves variation without inventing participant quotations.

How to Report Mixed-Methods Findings

Report each strand clearly

Readers should first understand the main quantitative and qualitative outcomes. Depending on the design, these may be reported sequentially or in parallel.

Integrate the strands

Integration asks what is learned from the combined evidence.

Common relationships include:

• Convergence: Both strands point to a similar conclusion.
• Complementarity: Each strand reveals a different aspect of the same issue.
• Explanation: Qualitative evidence helps explain a quantitative result.
• Expansion: One strand introduces an additional dimension.
• Divergence: The strands conflict or produce tension.
• Silence: An issue appears in one strand but not the other.

Use a joint display

A joint display places quantitative results, qualitative findings, and the integrated interpretation in one table or figure.

Research question	Quantitative result	Qualitative finding	Integrated inference
How did students experience the new platform?	Satisfaction scores increased, but continued-use rates varied	Convenience was valued, while technical interruptions reduced trust	Overall approval did not guarantee sustained use; reliability influenced whether positive attitudes translated into continued engagement

The integrated inference should provide knowledge that cannot be obtained by reading either strand alone.

How to Interpret Research Findings Responsibly

Interpretation belongs mainly in the discussion section unless the discipline uses a combined findings-and-discussion structure.

A responsible interpretation considers six questions.

1. What exactly did the study find?

State the main result without exaggeration.

2. How strong and precise is the evidence?

Consider effect size, uncertainty, consistency, data adequacy, credibility, and sensitivity to analytical choices.

3. What does the design permit the researcher to claim?

An observational association normally supports a different level of inference from a randomized experiment. A small contextual qualitative study supports depth and transferability judgments rather than automatic population-wide generalization.

4. What alternative explanations remain?

Consider confounding, bias, selection, measurement problems, model assumptions, contextual influences, chance variation, researcher interpretation, and competing theories.

5. How do the findings compare with previous evidence?

Identify agreement, contradiction, extension, or refinement. Do not force consistency when studies differ in population, design, measurement, or context.

6. What is the practical or theoretical importance?

A statistically detectable effect may be too small to matter in practice. A qualitative finding based on a small group may still be theoretically valuable if it reveals a process or experience that previous research overlooked.

Evidence, Interpretation, Implication, and Opinion

Level	Appropriate function	Example
Evidence	Reports what was observed or estimated	The estimated difference was 5.3 points
Interpretation	Explains what the evidence may mean	The program may have improved retention
Implication	Describes possible relevance	Course designers may consider spaced activities
Opinion or recommendation	Proposes an action or judgment	The university should adopt the program across all courses

Moving from evidence to recommendation requires increasingly broader reasoning. Each step should be justified, and uncertainty should remain visible.

Tables, Figures, and Quotations

Use a table when:

• Exact values matter.
• Multiple groups or outcomes must be compared.
• A large amount of structured information would be difficult to read in prose.
• A mixed-methods joint display is needed.

Use a figure when:

• A trend over time is central.
• A distribution must be shown.
• Readers need to see a relationship or interaction.
• A conceptual or process model communicates the finding better than prose.

Use quotations or extracts when:

• They clearly illustrate a qualitative theme.
• Participant language is analytically important.
• Variation or contradiction needs to be demonstrated.
• The extract is ethically appropriate and sufficiently de-identified.

Every visual should have an informative title, appropriate labels, units where relevant, and a caption or note explaining necessary abbreviations and analytical details.

Advantages of Clearly Reported Research Findings

Clear findings:

• Allow readers to assess whether conclusions follow from the evidence.
• Reduce ambiguity and misinterpretation.
• Help peer reviewers evaluate the study.
• Support replication, reanalysis, and evidence synthesis.
• Make null and unexpected outcomes visible.
• Improve communication with non-specialist audiences.
• Make later recommendations more defensible.

Clarity does not mean oversimplification. A clear findings section can still represent uncertainty, complexity, contradiction, and contextual variation.

Limitations of Research Findings

Research findings are not universal facts merely because they appear in a paper.

Their strength depends on:

• Research design.
• Sample selection.
• Measurement quality.
• Data completeness.
• Analytical assumptions.
• Researcher decisions.
• Context.
• Bias and confounding.
• Precision.
• Replication and consistency with other evidence.

Generalizability

Quantitative findings may not generalize beyond the population, setting, period, or conditions represented by the study.

Transferability

Qualitative researchers often provide contextual detail so readers can judge whether the findings may transfer to another setting. Transferability is not the same as statistical generalization.

Causal limitations

A finding that two variables are associated does not establish causality unless the research design and assumptions support causal inference.

Analytical dependence

Different defensible analytical choices may produce different estimates, models, codes, or themes. Sensitivity analysis, reflexivity, audit trails, and transparent reporting help readers understand this dependence.

Publication and reporting bias

A literature dominated by positive or statistically notable results can misrepresent the full evidence base. Null and contradictory findings therefore have scientific value.

Research Findings in Modern Research

Reporting Guidelines

Reporting guidelines help authors include the information readers need to understand and assess a study.

Common examples include:

Study type	Common reporting resource
Quantitative, qualitative, or mixed-methods psychology	APA Journal Article Reporting Standards
Randomized trial	CONSORT
Observational study	STROBE
Systematic review	PRISMA
Interview or focus-group study	COREQ
Broader qualitative research	SRQR
Animal research	ARRIVE
Mixed-methods research	JARS-Mixed, GRAMMS, or a field-specific mixed-methods guideline

A guideline should be selected according to the actual study design. PRISMA, for example, is designed for systematic reviews rather than ordinary narrative literature reviews.

Open Science and Transparent Findings

Modern reporting increasingly encourages researchers to make the path from question to finding visible.

Relevant practices include:

• Preregistering hypotheses and analysis plans.
• Distinguishing planned from exploratory analyses.
• Using registered reports where appropriate.
• Sharing de-identified data when ethical and lawful.
• Sharing analytical code and materials.
• Documenting exclusions and data-processing decisions.
• Reporting all predefined outcomes.
• Explaining deviations from the original plan.
• Providing persistent repository links.

Open practices do not eliminate bias or error, but they make research decisions easier to inspect and evaluate.

Digital Research Tools

Quantitative tools

R, Python, SPSS, Stata, SAS, JASP, jamovi, and spreadsheet software may be used for data preparation, statistical analysis, reproducible reporting, and visualization.

Qualitative tools

NVivo, ATLAS.ti, MAXQDA, Dedoose, and structured manual systems can help organize transcripts, codes, memos, cases, and relationships.

Software does not perform methodological reasoning automatically. Researchers remain responsible for the validity of the analysis and interpretation.

Reproducible workflows

Script-based analysis and dynamically generated tables can reduce transcription and copy-and-paste errors. Version control, documented code, data dictionaries, and research repositories can further strengthen transparency.

Artificial Intelligence and Research Findings

AI tools may assist with:

• Preliminary data organization.
• Code suggestions.
• Statistical programming support.
• Transcription correction.
• Searching large qualitative datasets.
• Drafting table captions.
• Language editing.
• Producing alternative visualizations.
• Checking consistency between reported numbers and tables.

AI should not be treated as an autonomous researcher or reliable source of findings.

Researchers must verify:

• Calculations.
• Statistical assumptions.
• Generated code.
• Theme labels.
• Summaries.
• Citations.
• Quotations.
• Confidentiality and data-protection requirements.
• Potential bias introduced by the tool.

Do not upload confidential participant data or unpublished manuscripts to an external AI system unless its privacy, security, contractual, and ethical conditions permit that use.

When AI contributes to the conduct or reporting of a study, follow the journal’s disclosure requirements. Record the tool, version or access date, purpose, relevant settings, and prompts when these details are necessary to understand or reproduce the work.

Human authors remain responsible for accuracy, originality, attribution, ethics, and final interpretation.

Common Mistakes When Reporting Research Findings

Reporting every output

A findings section should not reproduce the entire software output. Select and report evidence relevant to the research questions while making supplementary analyses available where appropriate.

Hiding null findings

Excluding an outcome because it did not cross a significance threshold creates a distorted account.

Claiming causation from association

Words such as “caused,” “led to,” and “resulted in” require causal justification.

Treating p-values as effect sizes

A smaller p-value does not necessarily indicate a larger or more important effect.

Ignoring uncertainty

Point estimates without intervals can create a false impression of precision.

Repeating tables in prose

Text should highlight the principal pattern, not restate every cell.

Using unsupported qualitative themes

A theme should be grounded in a transparent analytic process and supported by adequate evidence.

Using quotations as decoration

Each extract should perform a clear evidential or analytical function.

Treating frequency as qualitative importance

A frequently mentioned idea is not automatically the most meaningful finding.

Mixing findings and discussion unintentionally

The appropriate boundary depends on disciplinary convention, but the chosen structure should remain consistent.

Overgeneralizing

Claims should remain within the population, setting, design, and evidence represented by the study.

Using AI output without verification

AI-generated calculations, citations, code, or summaries may be inaccurate, incomplete, biased, or fabricated.

Research Findings Writing Template

Use the following structure for each major finding.

Finding heading

Direct finding: State the main evidence-based outcome in one or two sentences.

Research question or objective: Identify the question addressed.

Supporting evidence: Present the relevant estimate, interval, statistical test, theme, extract, observation, or joint display.

Scope: State which participants, cases, variables, period, or dataset the finding concerns.

Uncertainty or variation: Report imprecision, divergent cases, missing information, sensitivity to analysis, or other important qualification.

Visual reference: Direct readers to the relevant table or figure where applicable.

Status: Identify the analysis as primary, secondary, confirmatory, exploratory, or unexpected when this distinction matters.

Example subsection pattern

Finding 1: [Descriptive heading]

The analysis showed that [direct finding].

This finding addressed [research question or objective]. In [sample or context], [supporting numerical or qualitative evidence]. The estimate or interpretation was [precision, variation, or limitation]. Additional details are reported in [Table or Figure number].

Avoid adding broad explanations, literature comparisons, recommendations, or causal claims here unless the discipline deliberately combines findings and discussion.

Final Findings-Section Checklist

Before submitting, confirm that:

• Every major finding addresses a research question, objective, or predefined outcome.
• The sample or analytical dataset is clear.
• Findings are supported by reported evidence.
• Effect estimates and uncertainty are included where relevant.
• Null, negative, unexpected, and inconclusive findings are not hidden.
• Exploratory analyses are labelled.
• Qualitative themes are analytically explained and adequately supported.
• Mixed-methods strands are integrated rather than merely placed side by side.
• Tables, figures, and quotations add information rather than duplicate the text.
• Causal language matches the design.
• Generalizations remain within the evidence.
• Data, code, materials, and deviations from plans are documented where appropriate.
• AI-assisted work has been verified and disclosed according to applicable policies.
• The relevant reporting guideline and journal instructions have been checked.

Conclusion

Research findings are the evidence-based outcomes produced through systematic data analysis. Strong findings directly address the research questions, present sufficient supporting evidence, communicate uncertainty, and remain within the limits of the study design.

The most effective findings sections do not simply list numbers or themes. They show readers what was discovered, how the evidence supports that statement, how much confidence is justified, and where the boundaries of the finding lie.