How to Read a Scientific Figure
By Dr. Zubair Khalid, DVM, MS, PhD ·
Quick Answer
- Read the axes first: identify what is measured, the units, and the scale type before interpreting any pattern in the data.
- Locate the legend and figure caption to understand group definitions, statistical tests, and error bar meaning before drawing conclusions.
- Check the source and reporting standards: figures from studies following official reporting guidelines are more likely to show complete methods and transparent data handling.
Why Misreading Figures Leads to Wrong Conclusions
A scientific figure condenses complex data into a visual argument. When readers skip the axes, units, or legend, they fill gaps with assumptions. Those assumptions often produce conclusions the authors never intended. A graph showing a steep upward trend may represent a trivial change if the y-axis starts at a nonzero value or uses a logarithmic scale. A bar chart with error bars may show overlapping uncertainty that makes two groups statistically indistinguishable, even when the bar heights look different.
The practical problem is universal across animal science, laboratory research, and clinical medicine. A farmer reading a growth trial figure may conclude a feed additive works because the final weight bars look taller. A researcher scanning a journal article may cite a figure as evidence for an effect that the statistical test did not support. A laboratory staff member may adjust a protocol based on a figure that used a different unit than the one in their own assay. Each mistake traces back to the same root cause: reading the visual pattern before reading the figure's metadata.
This article provides a universal checklist for decoding any scientific graph. The checklist works for figures in journal articles, government reports, extension publications, and product datasheets. It also explains how reporting standards, data-sharing policies, and publication ethics shape what a figure can and cannot tell you.
At a Glance: The Five-Step Figure Reading Checklist
| Step | What to Examine | Key Question | Common Mistake |
|---|---|---|---|
| 1 | Figure caption and legend | What did the authors measure and under what conditions? | Drawing conclusions without reading the caption |
| 2 | Axes and units | What is plotted on each axis and in what unit? | Ignoring units or assuming they match your own data |
| 3 | Scale type | Is the axis linear, logarithmic, or transformed? | Interpreting a log scale as if it were linear |
| 4 | Error and variability indicators | Do error bars show standard deviation, standard error, or confidence intervals? | Treating overlapping error bars as proof of no difference |
| 5 | Statistical annotations | What test was used and what do the significance markers mean? | Assuming asterisks mean the same thing across all papers |
The Role of Reporting Guidelines in Figure Quality
What Reporting Guidelines Require
Reporting guidelines are checklists that tell authors what information to include when publishing a study. The EQUATOR Network maintains a comprehensive collection of these guidelines for different study types, including randomized trials, observational studies, diagnostic accuracy studies, and animal research. When authors follow a reporting guideline, their figures are more likely to include complete axis labels, clear unit definitions, and explicit descriptions of statistical methods.
For a reader, the presence of a reporting guideline citation is a signal. It means the journal asked the authors to document their methods in a structured way. That documentation extends to figures. A figure from a study that followed a reporting guideline should have a caption that explains the experimental conditions, the number of observations, and the statistical approach. When that information is missing, the figure is harder to interpret and the study is harder to evaluate.
How to Check Whether a Study Followed a Guideline
Look for a statement in the methods section or near the figure caption that names a specific guideline, such as CONSORT for randomized trials or ARRIVE for animal research. The EQUATOR Network provides a searchable library of these guidelines, so you can look up what a particular guideline requires. If the paper does not mention a guideline, you can still apply the underlying principles by asking whether the figure caption provides enough information to understand the data.
Core Principles for Reading Any Scientific Figure
Principle 1: The Caption Is Part of the Data
The figure caption is not decorative text. It is the authors' formal statement of what the figure shows. A complete caption identifies the experimental groups, the measurement method, the number of replicates, the time point, and the statistical test. Read the caption before you look at the visual pattern. If the caption is vague, treat the figure with caution.
Principle 2: Axes Define the Claim
Every axis has three components: the variable name, the unit, and the scale. A figure that plots body weight in kilograms tells a different story than one that plots body weight in pounds. A figure that plots bacterial growth on a logarithmic scale shows relative changes, not absolute differences. Check all three components on both axes before interpreting any trend.
Principle 3: Error Bars Have Specific Meanings
Error bars are not decoration. They quantify uncertainty, but the type of uncertainty depends on what the authors chose to show. Standard deviation describes the spread of individual observations. Standard error describes the precision of the sample mean. Confidence intervals describe the range within which the true population value likely falls. The caption or methods section should state which one is shown. Without that information, you cannot know whether overlapping error bars mean the groups are similar or whether the study simply had low statistical power.
Principle 4: Statistical Significance Is Not the Same as Practical Importance
A figure may show a statistically significant difference with a small p-value, but the actual difference between groups may be too small to matter in practice. Conversely, a figure may show a large visual difference that is not statistically significant because the sample size was small or the variability was high. The figure alone cannot tell you whether a result matters. You need the effect size, the confidence interval, and the context of the study.
Principle 5: The Figure Is an Interpretation, Not the Raw Data
The authors chose what to plot, how to summarize the data, and which comparisons to highlight. Those choices reflect their judgment. A figure can be accurate and still be misleading if the authors selected a subset of data, used a different scale than a previous study, or omitted a comparison group. Cross-check the figure against the text and, when possible, against the underlying data.
The Role of Data Sharing in Figure Verification
What Data-Sharing Policies Mean for Readers
The National Institutes of Health requires researchers to include a data-management and sharing plan in grant applications and to follow that plan during the award period. The policy expects researchers to make scientific data available for sharing, with appropriate protections for privacy and confidentiality. For a reader, this policy creates an opportunity. When a study is funded by NIH, the underlying data may be available in a repository, which means you can check whether the figure accurately represents the data.
How to Use Shared Data to Verify a Figure
If a paper mentions a data availability statement, look for a repository name and an accession number. The NIH data-sharing page explains the expectations for data-management plans, including the types of data that should be shared and the repositories that can host them. When you access the shared data, you can reproduce the figure's summary statistics, check the number of observations, and test alternative visualizations. This verification step is especially valuable when a figure will inform a management decision or a research direction.
Limitations of Data Sharing
Not all data are shared. Privacy concerns, proprietary interests, and legal restrictions can limit access. The NIH policy acknowledges that some data cannot be shared and requires researchers to explain those limitations in their plans. When data are not available, you must rely on the figure and the methods description, which is another reason to read the caption and methods carefully.
Practical Workflow for Decoding a Figure
Step 1: Read the Caption and Legend
Start with the caption. Identify the experimental design, the groups, the measurement, and the time frame. Then locate the legend, which maps colors, symbols, and line styles to specific groups or conditions. If the legend is missing or unclear, you cannot reliably distinguish groups in the figure.
Step 2: Examine Both Axes
Write down the variable name, the unit, and the scale for each axis. Ask whether the axis starts at zero. A y-axis that starts at a nonzero value can exaggerate small differences. A logarithmic axis changes how you should read the slope of a line. If the axis label is missing a unit, look for the unit in the caption or methods.
Step 3: Identify the Type of Figure
Different figure types require different reading strategies. A scatter plot shows individual data points and their relationship. A line graph shows trends over time or across a continuous variable. A bar chart shows group means or totals. A box plot shows the distribution, including the median, quartiles, and outliers. A forest plot shows effect sizes and confidence intervals across multiple studies. Identify the type before interpreting the pattern.
Step 4: Locate Variability and Uncertainty Indicators
Look for error bars, shaded regions, or confidence bands. Determine what they represent by checking the caption or methods. If the figure does not show any variability indicator, ask why. A figure without error bars may be showing a single representative result, a summary of a subset, or a figure with too little information to assess reliability.
Step 5: Check Statistical Annotations
Find the asterisks, letters, or symbols that indicate statistical comparisons. Read the caption to learn what each symbol means. Some papers use asterisks for p-values below a threshold, while others use letters to indicate groups that differ significantly. The same symbol can mean different things in different papers, so never assume.
Step 6: Compare the Figure to the Text
Read the results section and find the sentences that describe the figure. The text should state the direction and size of the effect, the statistical test, and the p-value. If the text and the figure disagree, investigate why. The discrepancy may be a typo, a figure error, or a sign that the authors are emphasizing a pattern that the data do not support.
Step 7: Consider the Source and Funding
Check who funded the study and whether the authors have conflicts of interest. The Committee on Publication Ethics core practices require journals to address conflicts of interest and to have policies for handling misconduct. A conflict-of-interest statement does not prove a figure is wrong, but it tells you to read the figure with attention to whether the interpretation matches the data.
Common Failure Patterns in Figure Reading
Failure Pattern 1: Ignoring the Axis Origin
A y-axis that starts at 50 instead of zero makes a small change from 50 to 55 look like a 10 percent increase when it is actually a 10 percent increase relative to the starting value but a small absolute change. The visual impression depends on the axis range. Always check the axis limits before judging the size of an effect.
Failure Pattern 2: Misreading Logarithmic Scales
On a logarithmic axis, equal distances represent equal multiplicative factors, not equal additive amounts. A line that looks straight on a log scale represents exponential growth or decay. A reader who interprets the slope as a linear trend will misstate the relationship. Check the axis label for the word "log" or for tick marks that increase by factors of 10.
Failure Pattern 3: Confusing Standard Deviation and Standard Error
Standard deviation describes the variability of individual measurements. Standard error describes the precision of the mean. Standard error is always smaller than standard deviation for the same data, so error bars based on standard error make a figure look more precise. The caption must state which one is shown. If it does not, the figure is incomplete.
Failure Pattern 4: Overinterpreting Overlapping Error Bars
Overlapping error bars do not prove that two groups are the same. They may indicate that the study lacked the statistical power to detect a difference. Conversely, non-overlapping error bars do not always mean a statistically significant difference, especially if the error bars represent standard deviation instead of confidence intervals. The statistical test, not the visual overlap, determines significance.
Failure Pattern 5: Assuming Asterisks Have a Universal Meaning
Some journals use one asterisk for p less than 0.05, two for p less than 0.01, and three for p less than 0.001. Other journals use different thresholds or different symbols. The caption should define the symbols. If it does not, you cannot interpret the significance markers.
Failure Pattern 6: Skipping the Methods Section
The methods section explains how the data were collected, how many observations were made, and how the analysis was performed. A figure without a corresponding methods description is incomplete. Readers who skip the methods cannot know whether the figure shows raw data, normalized data, or a model prediction.
Records and Measurements for Figure Verification
What to Record When You Read a Figure
Keep a structured record of your figure reading so you can compare figures across studies and revisit your conclusions. For each figure, record the following items:
| Record Item | What to Write Down | Why It Matters |
|---|---|---|
| Citation | Full reference with DOI or URL | Enables you to find the paper again |
| Figure number | Figure 2, Figure 3, etc. | Avoids confusion when discussing multiple figures |
| Caption summary | Experimental groups, measurement, time point | Captures the authors' stated conditions |
| Axis variables and units | X-axis variable and unit, Y-axis variable and unit | Prevents unit conversion errors |
| Scale type | Linear, log, or other transformation | Determines how to interpret slopes and differences |
| Error bar type | SD, SE, CI, or none | Determines what the variability means |
| Statistical test | Test name and p-value or confidence interval | Determines whether the difference is significant |
| Your interpretation | One sentence stating what the figure shows | Forces you to articulate the conclusion |
How to Use Records for Decision Making
When a figure informs a management decision, such as choosing a feed additive or a treatment protocol, compare the figure against your own records. Ask whether the study population resembles your animals, whether the measurement units match your scale, and whether the effect size is large enough to justify a change. A figure from a study with a different breed, age group, or environment may not apply to your situation.
How to Identify Whether a Figure Comes From a Reliable Study
Check the Journal and the Peer-Review Process
The Committee on Publication Ethics core practices describe the responsibilities of journals, including peer review, handling of misconduct, and management of conflicts of interest. Journals that follow these practices are more likely to publish figures that accurately represent the underlying data. You can check whether a journal is a member of COPE by looking at its website or the COPE member list.
Check the Authors and Their Affiliations
The ORCID system provides a persistent identifier for researchers, which helps you verify that the author is who they claim to be and that their publication record is consistent. An ORCID record can show whether the author has published other work in the same field and whether they have corrected or retracted previous papers. This context helps you assess the reliability of the figure.
Check the Funding Source
The NIH grants and funding pages describe how NIH-funded research is reviewed and awarded. A funding source does not determine whether a figure is accurate, but it provides context. Industry-funded studies may have different incentives than government-funded studies. Read the conflict-of-interest statement and consider whether the funding source could influence the presentation of the data.
Check for a Data Availability Statement
A data availability statement tells you where the underlying data are stored and whether they are accessible. The NIH data-management and sharing policy expects researchers to share data with appropriate protections. When a paper includes a data availability statement with a repository link, you can verify the figure by examining the raw data. When the statement says data are not available, you must rely on the figure and the methods.
Welfare and Safety Context for Animal Research Figures
Why Welfare Context Matters for Figure Interpretation
Figures from animal studies often report outcomes such as body weight, feed intake, mortality, and clinical scores. These outcomes are influenced by the animals' welfare status. A figure that shows improved growth in a treatment group may be less meaningful if the study also reports higher mortality or more clinical disease in that group. Read the welfare-related outcomes in the text and in other figures before concluding that a treatment is beneficial.
How to Check Welfare-Related Information
Look for descriptions of housing, feeding, health monitoring, and humane endpoints in the methods section. The ARRIVE guideline, available through the EQUATOR Network, requires animal studies to report these details. If the methods do not describe welfare monitoring, the study may not meet current reporting standards, and the figures should be interpreted with caution.
Safety Considerations for Applying Figure Results
A figure may show that a treatment changes a physiological variable, but the figure does not tell you whether the change is safe. Safety requires information about adverse events, toxicity, and long-term outcomes, which are often reported in separate figures or tables. Before applying a figure's result to your own animals, check whether the study reported any safety concerns and whether the treatment is approved for your species and purpose.
Limitations of Figure Reading
A Figure Cannot Replace the Full Study
A figure is a summary. It does not show every data point, every statistical test, or every sensitivity analysis. The full study may include additional analyses that change the interpretation. Always read the results and discussion sections to understand how the authors interpreted the figure and whether alternative explanations were considered.
A Figure Cannot Tell You About External Validity
A figure shows what happened in the study population under the study conditions. It does not tell you whether the same result will occur in your population under your conditions. Differences in genetics, environment, management, and measurement methods can change the outcome. Use the figure to understand the study, but use your own records and local conditions to make decisions.
A Figure Cannot Reveal What Was Not Measured
A figure can only show the variables the authors chose to measure. Important outcomes that were not measured are invisible in the figure. A growth trial figure may show weight gain but not feed efficiency, carcass quality, or health outcomes. The absence of a variable in a figure does not mean the variable is unimportant.
A Figure Cannot Overcome Poor Study Design
A well-drawn figure cannot fix a poorly designed study. If the study lacked a control group, had too few replicates, or used an inappropriate statistical test, the figure will reflect those flaws. The figure reading checklist helps you identify what the figure shows, but it cannot tell you whether the study was designed well. For that, you need to evaluate the methods and the reporting quality.
Professional Escalation Criteria
When to Consult a Statistician or Methodologist
Seek professional help when a figure is central to a decision and you cannot determine the statistical method, the error bar type, or the scale transformation. A statistician can help you interpret the figure correctly and can assess whether the statistical analysis supports the authors' claims. This consultation is especially important when the figure will be used in a grant application, a regulatory submission, or a publication.
When to Contact the Authors
Contact the corresponding author when the figure caption is incomplete, the units are missing, or the statistical annotations are unclear. Authors are usually willing to clarify their figures. Keep a record of your question and the author's response. If the author cannot clarify the figure, treat the figure as unreliable.
When to Report a Suspected Error
If you find a discrepancy between a figure and the text, or if a figure appears to misrepresent the data, consider reporting the issue. The Committee on Publication Ethics core practices describe how journals should handle allegations of misconduct and errors. You can contact the journal editor with your specific concern. Provide the figure number, the discrepancy, and your evidence. Do not make a public accusation without first contacting the journal.
When to Avoid Using a Figure
Do not use a figure as the basis for a decision if the caption is missing, the axes are unlabeled, the error bars are undefined, or the statistical test is not described. These omissions indicate that the figure does not meet current reporting standards. Using such a figure to support a conclusion risks propagating an error.
Frequently Asked Questions
Why do I need to read the axes before looking at the data pattern?
The axes define what the figure actually shows. Without the variable names, units, and scale type, you cannot know whether the pattern represents a meaningful change or an artifact of the axis range. A figure with an unlabeled y-axis is uninterpretable, regardless of how clear the visual pattern appears.
What is the difference between standard deviation and standard error on a figure?
Standard deviation describes how much individual observations vary around the mean. Standard error describes how precisely the sample mean estimates the population mean. Standard error is always smaller than standard deviation for the same data, so error bars based on standard error make the results look more precise. The caption must state which one is shown.
How can I tell if a figure uses a logarithmic scale?
Check the axis label for the word "log" or for tick marks that increase by factors of 10, such as 1, 10, 100, and 1000. On a logarithmic axis, equal distances represent equal multiplicative factors. A straight line on a log scale represents exponential change, not linear change.
What should I do if the figure caption does not explain the error bars?
Treat the figure as incomplete. Look in the methods section for a description of the statistical analysis and the variability measure. If the methods do not explain the error bars, contact the corresponding author or avoid using the figure as evidence.
Do overlapping error bars mean there is no difference between groups?
No. Overlapping error bars do not prove that two groups are the same. The overlap may reflect low statistical power, high variability, or a small sample size. The statistical test, not the visual overlap, determines whether the difference is significant.
How do reporting guidelines help me read a figure?
Reporting guidelines, such as those collected by the EQUATOR Network, tell authors what information to include when publishing a study. When authors follow a guideline, their figures are more likely to have complete captions, defined error bars, and clear statistical annotations. The presence of a guideline citation is a signal that the figure meets a recognized reporting standard.
Can I verify a figure by looking at the raw data?
Sometimes. The NIH data-management and sharing policy expects researchers to share scientific data with appropriate protections. If the paper includes a data availability statement with a repository link, you can access the data and check whether the figure accurately represents it. If the data are not available, you must rely on the figure and the methods.
What should I record when I read a figure for a decision?
Record the citation, figure number, caption summary, axis variables and units, scale type, error bar type, statistical test, and your interpretation. This record allows you to compare figures across studies and to revisit your conclusions when new information becomes available.
Related Articles
- How to Read a Scientific Paper Efficiently
- Protein Synthesis Step By Step
- Protein Synthesis: A Step-by-Step Guide to Transcription and Translation
- Scientific Posters: Designing a Figure-First Research Narrative
- 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.