How to Compare Results and Conclusions in a Scientific Paper

By Dr. Zubair Khalid, DVM, MS, PhD · 2026-08-27

Quick Answer

• Compare each conclusion in a paper directly against the specific results that are supposed to support it, checking for numeric agreement, statistical significance, and logical fit.
• Use a structured checklist to map every conclusion to its source result, and flag any conclusion that goes beyond what the data actually show.
• A conclusion can be technically correct but still overstate the findings, so always check for scope, generalizability, and alternative explanations before accepting it.

Why Comparing Results and Conclusions Matters

Readers of scientific papers face a common problem. The results section reports what the researchers observed. The conclusions section interprets what those observations mean. Between those two sections, a gap can open. Sometimes the gap is small and honest. Sometimes it is large enough to mislead a careful reader.

For farmers, laboratory staff, and life-science professionals, the ability to spot that gap has direct practical value. A feed trial conclusion that claims a product improves weight gain needs to be checked against the actual weight measurements. A vaccine study conclusion that claims protection needs to be checked against the infection rates in the data. A genetics paper conclusion that claims a marker predicts disease needs to be checked against the predictive values in the results.

The skill is not about distrusting researchers. It is about verifying that the evidence supports the claim. Peer review reduces errors, but it does not eliminate them. The Committee on Publication Ethics Core Practices describe the responsibilities of journals and authors, including handling data and responding to concerns, but those practices do not guarantee that every published conclusion is perfectly matched to its results. The reader retains responsibility for critical evaluation.

This article gives you a concrete method for comparing results and conclusions. You will learn a checklist, see a worked example, and understand the common ways conclusions drift away from the data.

At a Glance: The Conclusion-to-Result Mapping Table

The table below summarizes the core comparison tasks you will perform when reading any scientific paper.

Core Principles of Critical Paper Reading

Results Are Observations, Conclusions Are Interpretations

A results section presents measurements, counts, percentages, test statistics, and confidence intervals. It describes what happened in the study. A conclusions section interprets those observations. It states what the authors believe the findings mean.

The distinction matters because interpretation requires judgment. Two researchers can look at the same results and draw different conclusions. One may emphasize the practical importance of a small but significant effect. Another may emphasize the limitations that make the effect unreliable. Neither is necessarily wrong. Your job as a reader is to decide which interpretation the data support.

The Burden of Proof Rests on the Conclusion

A conclusion must earn its claim. The results section is the evidence. The conclusion is the verdict. A verdict without evidence is unsupported. A verdict that exceeds the evidence is overstated. A verdict that contradicts the evidence is wrong.

When you read a paper, treat every conclusion as a hypothesis about what the results mean. Then test that hypothesis against the results. This is the same logic you would apply to a farm record. If a record shows that a treatment group gained 5 percent more weight than a control group, a claim that the treatment is responsible requires more than the raw difference. It requires checking whether the groups were comparable, whether the difference is consistent, and whether other factors could explain it.

Reporting Guidelines Help You Know What to Expect

Reporting guidelines describe the minimum information that should appear in a paper for a given study type. The EQUATOR Network maintains a collection of these guidelines and helps researchers select the right one for their work. When a paper follows a reporting guideline, the results section should contain the information needed to evaluate the conclusions. When it does not, missing information becomes a reason for caution.

For example, a randomized trial reporting guideline typically requires information about how participants were assigned to groups, how many completed the study, and how missing data were handled. If a paper claims a treatment effect but does not report how many animals dropped out of each group, you cannot fully evaluate whether the conclusion is sound.

Practical Workflow for Comparing Results and Conclusions

Step 1: Read the Conclusions First

Start by reading the abstract and the conclusions section. Write down each distinct claim the authors make. Number them. This gives you a list of claims to verify.

A claim might be specific, such as "Treatment A reduced mortality by 15 percent compared with the control." Or it might be general, such as "These findings suggest that early intervention improves long-term outcomes." Both types need checking, but they need different kinds of checking.

Step 2: Locate the Supporting Results for Each Claim

Go to the results section and find the data that supposedly support each claim. Look for the table, figure, or text passage that reports the relevant measurements. Mark it clearly.

If you cannot find any result that corresponds to a conclusion, that is a serious problem. A conclusion without a supporting result is unsupported.

Step 3: Check the Numbers

Compare the numbers in the conclusion with the numbers in the results. They should match. Watch for rounding that changes the meaning. A result of 14.7 percent should not become "about 20 percent" in the conclusion. A confidence interval that includes zero should not be described as a clear effect.

Step 4: Check the Statistical Language

Pay attention to words like significant, tended, suggested, and indicated. These words carry statistical meaning. A result that is not statistically significant should not be described as a proven effect. A result that is marginally significant should not be described as definitive.

The National Library of Medicine Bookshelf provides access to research-methods references that explain statistical concepts in detail. If you are unsure what a statistical term means, look it up before judging the conclusion.

Step 5: Check the Scope

Ask whether the conclusion stays within the boundaries of the study. Consider the population studied, the conditions of the study, the duration, and the measurements used. A conclusion that extends beyond those boundaries is an overgeneralization.

For example, a study of dairy cows on one farm in one season cannot support a conclusion about all dairy cows in all climates. A study that measured weight gain over 30 days cannot support a conclusion about lifetime productivity.

Step 6: Check for Alternative Explanations

Ask what else could explain the results. Could the effect be due to chance? Could it be due to a confounding factor? Could it be due to the way measurements were taken? A conclusion that ignores plausible alternative explanations is weaker than one that addresses them.

Step 7: Record Your Assessment

Write down your judgment for each claim. Note whether it is fully supported, partially supported, or unsupported by the results. Note the specific reason for your judgment. This record becomes useful when you compare multiple papers on the same topic.

Worked Example: Mapping Conclusions to Results

Consider a hypothetical study of a new mineral supplement for growing pigs. The paper reports the following results:

• Pigs receiving the supplement gained an average of 0.85 kg per day over 60 days.
• Pigs receiving the control diet gained an average of 0.80 kg per day.
• The difference was not statistically significant (p = 0.12).
• Mortality was 2 percent in the supplement group and 4 percent in the control group.
• The mortality difference was not statistically significant (p = 0.30).
• Feed conversion ratio was 2.8 in the supplement group and 2.9 in the control group.

Now examine the conclusions:

Conclusion 1: "The supplement improved weight gain in growing pigs."

Check the numbers. The supplement group gained 0.85 kg per day versus 0.80 kg per day for the control. The direction of the difference matches the conclusion. But the difference was not statistically significant. The conclusion overstates the finding. A more accurate conclusion would say the supplement was associated with numerically higher weight gain, but the difference was not statistically significant in this study.

Conclusion 2: "The supplement reduced mortality."

Check the numbers. Mortality was 2 percent in the supplement group and 4 percent in the control group. The direction matches. But the difference was not statistically significant, and the study likely had few animals, making the mortality estimate unstable. The conclusion overstates the finding.

Conclusion 3: "The supplement improved feed efficiency."

Check the numbers. Feed conversion ratio was 2.8 versus 2.9. The direction matches. But no statistical test was reported for this outcome. Without a statistical test, you cannot know whether the difference is likely due to the supplement or to chance. The conclusion is partially supported at best.

Conclusion 4: "These findings support the use of the supplement in commercial pig production."

Check the scope. The study was conducted under specific conditions. The results do not show statistically significant benefits. A conclusion that recommends commercial use goes beyond what the data support. This is an overgeneralization.

This example shows the full range of problems you will encounter. Some conclusions match the direction of the results but ignore statistical limitations. Others extend beyond the study scope. Your checklist catches all of these.

Common Failure Patterns in Results and Conclusions

The Overstated Effect

The conclusion claims a stronger effect than the results show. This is the most common pattern. It often appears when a difference is not statistically significant but the conclusion describes it as a real effect. It also appears when a small effect is described in dramatic language.

The Hidden Scope Limitation

The conclusion describes the findings as applying to a broader population than the study actually examined. A study of healthy adult animals cannot support conclusions about young, old, or sick animals. A study conducted in one region cannot support conclusions about other regions.

The Missing Statistical Test

The conclusion claims a difference or effect, but the results section does not report a statistical test for that outcome. Without a test, you cannot distinguish a real effect from random variation.

The Correlation-to-Causation Leap

The results show an association between two variables. The conclusion states that one variable caused the other. Observational studies cannot establish causation because other factors may explain the association.

The Selective Emphasis

The results section reports multiple outcomes. Some show effects and others do not. The conclusion emphasizes only the outcomes that showed effects and ignores the null results. This gives a misleading impression of the overall findings.

The Unsupported Recommendation

The conclusion includes a practical recommendation, such as using a product or changing a practice. The results may support the underlying finding, but they do not support the recommendation because the recommendation depends on additional factors like cost, safety, and applicability to other settings.

Records and Measurements for Critical Reading

Building a Paper Comparison Log

Keep a log of the papers you read. For each paper, record the citation, the study design, the main results, and your assessment of each conclusion. This log becomes a personal reference that helps you track what the evidence actually shows.

A useful log entry includes:

• The full citation and a link to the paper
• The study design and population
• The main outcomes and their numeric results
• Each conclusion and your rating of its support
• Notes on limitations and alternative explanations

Using a Conclusion Rating Scale

Rate each conclusion on a simple scale. This forces you to make a judgment instead of passively accepting what you read.

Tracking Consistency Across Papers

When you read multiple papers on the same topic, compare their conclusions. Do they agree? If they disagree, check whether the results differ or whether the authors interpreted similar results differently. This comparison often reveals where the evidence is strong and where it is uncertain.

Quality Controls and Professional Escalation

When to Escalate Concerns

If you identify a serious mismatch between results and conclusions, consider whether to escalate your concern. The appropriate action depends on your role and the nature of the problem.

For a student or researcher reading a paper for background, a mismatch is a note in your analysis. You do not need to contact anyone. You simply account for the limitation in your own work.

For a peer reviewer, a mismatch is a required comment. Your review should identify the specific conclusion, the specific result, and the nature of the mismatch. The Committee on Publication Ethics Core Practices describe the responsibilities of peer reviewers and the importance of handling concerns about published work appropriately.

For a professional considering a product or practice based on a paper, a mismatch is a reason to withhold action. Do not adopt a practice or purchase a product based on a conclusion that the results do not support. Seek additional evidence or consult a specialist.

Handling Suspected Misconduct

If you suspect that a paper involves fabrication, falsification, or other research misconduct, the appropriate response is to raise your concern with the journal or the institution. The Committee on Publication Ethics Core Practices describe how journals should handle allegations of misconduct and how institutions should investigate concerns. Do not make public accusations without evidence. Follow the proper channels.

The Role of Data Sharing

Many funding agencies now require data management and sharing plans. The NIH Data Management and Sharing Policy describes expectations for sharing scientific data generated from NIH-funded research. When data are available, you can examine the underlying data yourself instead of relying only on the results as reported. This strengthens your ability to evaluate conclusions.

Limitations of the Comparison Method

You Cannot Verify Everything

Your comparison method checks whether conclusions match the reported results. It cannot verify whether the results themselves are accurate. A paper can have perfectly matched results and conclusions and still be wrong if the data were fabricated or the measurements were flawed.

You Depend on Complete Reporting

Your method depends on the paper reporting all relevant results. If the authors omitted unfavorable results, you cannot detect the omission from the paper alone. Reporting guidelines help by specifying what should be reported, but they cannot guarantee compliance.

Statistical Expertise Is Required

Some comparisons require statistical knowledge. You need to understand p-values, confidence intervals, and study design to judge whether a conclusion is supported. If you lack this knowledge, consult a statistician or a methods reference. The National Library of Medicine Bookshelf provides access to research-methods references that can help.

Context Matters

A conclusion that is unsupported in one context may be reasonable in another. A preliminary study with a small sample may appropriately use cautious language. A large confirmatory study may appropriately use stronger language. Your judgment must account for the study context.

Safety and Regulatory Context

Implications for Animal Products and Practices

When a paper concerns an animal product, feed additive, vaccine, or management practice, the conclusion has regulatory and safety implications. A conclusion that overstates benefits could lead to the adoption of an ineffective product. A conclusion that understates risks could lead to unsafe practices.

Regulatory agencies evaluate evidence before approving products for use. The standards for approval are typically higher than the standards for publication. A published conclusion that a product is effective does not mean the product is approved or safe for use. Check with the relevant regulatory authority before acting on a published conclusion.

The Role of Funding and Conflicts of Interest

Check the paper for funding sources and conflict-of-interest statements. The Committee on Publication Ethics Core Practices describe the importance of disclosing conflicts and the responsibilities of authors and journals in managing them. A conflict of interest does not automatically invalidate a paper, but it is a reason to read the conclusions with extra care.

Professional Judgment and Institutional Policy

Your response to a paper should account for your professional role and your institution's policies. A researcher may be required to follow specific guidelines for citing and using published work. A practitioner may be required to follow specific protocols for adopting new practices. Know the rules that apply to you.

Frequently Asked Questions

What is the difference between results and conclusions in a scientific paper?

Results report the measurements and observations from the study. Conclusions interpret what those measurements mean. Results answer the question of what happened. Conclusions answer the question of what the findings signify. The two sections serve different purposes, and a conclusion can be inaccurate even when the results are accurate.

How can I tell if a conclusion is supported by the results?

Check three things. First, verify that the numbers in the conclusion match the numbers in the results. Second, check whether the statistical language in the conclusion matches the statistical findings. Third, check whether the conclusion stays within the scope of the study population, conditions, and duration. A conclusion that passes all three checks is likely supported.

What should I do if a conclusion does not match the results?

Record the mismatch in your notes. Identify the specific conclusion and the specific result that conflict. Decide whether the mismatch changes your use of the paper. If you are a peer reviewer, report the mismatch in your review. If you are considering a product or practice, do not act on the unsupported conclusion.

Why do authors sometimes overstate their conclusions?

Authors may be enthusiastic about their findings. They may believe that their results are practically important even when the statistics do not show a clear effect. They may also face pressure to present their work favorably. Understanding these pressures helps you read critically without assuming bad faith.

Do reporting guidelines help readers evaluate conclusions?

Yes. Reporting guidelines specify the information that should appear in a paper for a given study type. When a paper follows a reporting guideline, the results section should contain the information needed to evaluate the conclusions. The EQUATOR Network maintains a collection of these guidelines and helps researchers select the appropriate one.

Can I check the underlying data for a paper?

Sometimes. Many funding agencies require data management and sharing plans. The NIH Data Management and Sharing Policy describes expectations for sharing scientific data from NIH-funded research. When data are available, you can examine them directly. When they are not, you must rely on the reported results.

How do conflicts of interest affect my reading of conclusions?

A conflict of interest is a reason for caution, not a reason for automatic rejection. Check the paper for funding sources and conflict-of-interest statements. If the authors have a financial interest in the product or practice being studied, read the conclusions with extra care and look for overstatement.

What is the best way to compare conclusions across multiple papers?

Create a comparison log. For each paper, record the study design, the main results, and your assessment of each conclusion. Then compare the conclusions across papers. Look for agreement and disagreement. When conclusions disagree, check whether the results differ or whether the authors interpreted similar results differently.

Related Articles

• Writing the Methods Section of a Scientific Paper: A Reproducibility-Focused Guide
• Scientific Data Provenance For Reproducible Research
• Choosing the Right Citation Style for Your Scientific Paper
• Writing the Methods Section for Qualitative Research Papers
• Research Figures: Making Methods and Results Easier to Evaluate

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 Literature Resources. National Center for Biotechnology Information.
• PubMed. National Library of Medicine.

This article is educational and does not replace institutional policy, professional advice, or applicable safety and regulatory requirements.