# How to Read a Table in a Scientific Paper ## Quick Answer - Read the title, column headers, row labels, and units before examining any number so you know what the table actually measures. - Check footnotes and table notes for definitions, statistical symbols, and abbreviations because they explain how the data were collected and analyzed. - Compare values within the context of the study design and sample size, since a table alone cannot tell you whether a difference is meaningful. ## Why Tables Are Hard to Read Scientific tables pack a large amount of information into a small space. A reader who jumps straight to the numbers often misses the structure that gives those numbers meaning. The title tells you the topic, the column headers tell you the variables, the row labels tell you the groups, and the footnotes tell you the conditions. Each part works together, and skipping any one of them can lead to a wrong interpretation. Tables in research papers follow conventions that are not always obvious to a new reader. Some tables report raw data, others report summary statistics, and still others report the results of statistical tests. The same column header can mean different things depending on the journal style, the study design, and the reporting guidelines the authors followed. The [EQUATOR Network](https://www.equator-network.org/) maintains a collection of reporting guidelines that help authors present their methods and results clearly, and these guidelines often specify how tables should be structured. A practical reading strategy treats a table as a complete document instead of a collection of numbers. You read the title, then the headers, then the row labels, then the footnotes, and only then do you look at the data. This order prevents you from assigning meaning to a number before you know what that number represents. ## The Structure of a Scientific Table ### Title and Caption The title or caption sits above the table and tells you what the table contains. A good title states the variables, the population, and sometimes the time period. For example, a title that reads "Mean daily weight gain in broiler chickens fed three experimental diets" tells you the outcome, the species, and the treatment groups before you look at any data. The caption may also contain information about the study design, such as the number of animals, the duration of the trial, or the location. This context matters because a table from a small pilot study carries different weight than a table from a large field trial. The [National Library of Medicine](https://www.ncbi.nlm.nih.gov/books) hosts research methods references that explain how study design affects the interpretation of reported results. ### Column Headers Column headers tell you what each vertical column contains. They usually include the variable name and the unit of measurement. A column labeled "Weight (kg)" means the values are in kilograms, while a column labeled "Weight (lb)" means the values are in pounds. The same number can mean very different things depending on the unit. Some tables use multi-level headers. The top level may name a treatment group, and the second level may name the time point or the measurement. In this case, you need to read both levels together to understand what each number represents. ### Row Labels Row labels identify the groups or categories being compared. They may be treatment groups, time points, or demographic categories. The row label tells you which population the number in that row describes. In animal research, row labels often list the experimental groups, such as "Control," "Diet A," and "Diet B." In clinical research, row labels may list patient groups, such as "Treatment" and "Placebo." The row label is the anchor for the entire row, and reading it first prevents you from assigning a value to the wrong group. ### Footnotes and Table Notes Footnotes appear below the table and contain information that does not fit in the main body. They may define abbreviations, explain the statistical tests used, or state the level of significance. A footnote that says "Values are means plus or minus standard deviation" tells you how the numbers in the table should be interpreted. Footnotes also contain the limitations of the data. A table may include a footnote that says "Data were missing for 5 percent of participants" or "Values were adjusted for age and sex." These notes change the meaning of the numbers, and ignoring them can lead to an incorrect conclusion. ## How to Read a Table Step by Step ### Step 1: Read the Title Start with the title. Ask yourself what the table is about, what population it describes, and what variables it reports. Write down the answer in one sentence. If you cannot answer these questions after reading the title, you are not ready to read the numbers. ### Step 2: Read the Column Headers Read every column header from left to right. Note the units and the measurement scale. If a column reports a percentage, check whether the percentage is a proportion of the row total or the column total. If a column reports a rate, check the denominator, such as per 1,000 animals or per 100,000 people. ### Step 3: Read the Row Labels Read every row label. Identify the groups being compared and the relationship between them. If the rows represent time points, note the interval between measurements. If the rows represent treatment groups, note the number of groups and the presence of a control group. ### Step 4: Read the Footnotes Read every footnote before you look at the data. The footnotes contain the definitions, the significance levels, and the limitations. A table without footnotes may still have a note in the caption or in the methods section of the paper. If the table does not explain how the values were calculated, look for that explanation in the methods section. ### Step 5: Read the Data Only after you understand the structure should you look at the numbers. Start with the control group or the baseline row if one exists. Then compare the other rows to that baseline. Look for patterns across the columns, such as a steady increase or decrease, and note any values that stand out as unusual. ### Step 6: Check the Significance If the table reports statistical significance, check the footnote for the significance level. A common convention is that a p value below 0.05 is considered statistically significant, but this threshold is not universal. The footnote will tell you what the authors used. If the table does not report significance, you cannot assume that a difference between two numbers is meaningful. ## At a Glance | Table Component | What It Tells You | What to Check | | --- | --- | --- | | Title | The topic, population, and variables | Whether the population matches your question | | Column headers | The variables and their units | Whether the units are consistent across columns | | Row labels | The groups or time points being compared | Whether a control group or baseline is present | | Footnotes | Definitions, significance, and limitations | Whether the data were adjusted or excluded | | Data values | The measured results | Whether the values are raw, means, or percentages | ## Common Mistakes When Reading Tables ### Skipping the Footnotes The most common mistake is reading the numbers without reading the footnotes. A table can report adjusted values, unadjusted values, or values that exclude certain participants. Without the footnote, you cannot know which type of value you are looking at. The [Committee on Publication Ethics](https://publicationethics.org/core-practices) core practices emphasize that authors should report their data accurately and transparently, and the footnotes are where that transparency appears. ### Confusing Units A table that reports weight in kilograms and another that reports weight in pounds will produce different numbers for the same animal. If you compare values across tables without checking the units, you can draw a false conclusion. Always check the unit in the column header before comparing any values. ### Ignoring the Sample Size A table that reports a mean from 5 animals is not comparable to a table that reports a mean from 500 animals. The sample size is often reported in the table, either as a column or in the footnote. If the sample size is not in the table, look for it in the methods section. The [NIH Grants and Funding](https://grants.nih.gov/) pages describe how research proposals are reviewed, and the review process places weight on the adequacy of the sample size for the research question. ### Treating Correlation as Causation A table can show that two variables move together, but it cannot show that one causes the other. The table reports the data, and the study design determines whether a causal claim is possible. A randomized controlled trial can support a causal claim, while an observational study cannot. The table alone does not tell you which type of study produced the data. ### Overlooking the Significance Level A difference between two numbers in a table is not automatically meaningful. The table may report a p value, a confidence interval, or a standard error. Without these statistics, you cannot know whether the difference is likely to be real or due to chance. The footnote or the methods section will tell you how the authors assessed significance. ## How to Use Tables in Your Own Work ### When You Are Reading a Paper Use the table to answer a specific question. Before you open the paper, write down what you want to know. Then find the table that addresses that question and read it using the steps above. This approach keeps you focused and prevents you from being overwhelmed by the volume of data. ### When You Are Writing a Paper If you are preparing a manuscript, the [EQUATOR Network](https://www.equator-network.org/) provides reporting guidelines that tell you what to include in your tables. These guidelines vary by study type, so you should select the guideline that matches your study design. The guidelines will tell you which variables to report, how to report the units, and what footnotes to include. ### When You Are Reviewing a Paper If you are a peer reviewer, the [Committee on Publication Ethics](https://publicationethics.org/core-practices) core practices describe the responsibilities of reviewers. A reviewer should check that the tables are clear, that the units are stated, that the footnotes explain the data, and that the numbers in the tables match the numbers in the text. A table that is inconsistent with the text is a sign that the manuscript needs revision. ## How to Build a Table That Is Easy to Read ### Choose the Right Columns A table should have a clear purpose. Each column should contain one variable, and each row should contain one group or one time point. If a table has too many columns, the reader cannot follow the comparisons. If a table has too many rows, the reader cannot see the patterns. ### State the Units in the Header The unit belongs in the column header, not in every cell. Write "Weight (kg)" instead of putting "kg" next to every number. This reduces clutter and makes the table easier to scan. ### Use Footnotes for Explanations Put the definitions, the significance levels, and the limitations in the footnotes. The footnotes should be numbered or lettered and referenced in the table. A reader who wants the full explanation can find it without the table becoming unreadable. ### Report the Sample Size The sample size should be in the table or in the footnote. Without the sample size, the reader cannot judge the reliability of the values. The [NIH Data Management and Sharing Policy](https://sharing.nih.gov/data-management-and-sharing-policy) describes how researchers should plan for the sharing of their data, and a table that reports the sample size is part of that transparency. ### Keep the Format Consistent Use the same number of decimal places throughout a column. Use the same symbol for significance throughout the table. If you use an asterisk for a p value below 0.05, use that same symbol for every value that meets that threshold. ## How to Check Whether a Table Is Reliable ### Check the Source A table from a peer-reviewed journal is generally more reliable than a table from a website that does not disclose its methods. The [Committee on Publication Ethics](https://publicationethics.org/core-practices) core practices describe the standards for peer review and data integrity. If the journal follows these practices, the table is more likely to be accurate. ### Check the Methods Section The methods section should describe how the data were collected and analyzed. If the methods section does not match the table, the table may be wrong. For example, if the methods say that the study included 100 animals but the table reports data for 80, the discrepancy needs an explanation. ### Check the Numbers Add the numbers in a column to see if they match the total. Check the percentages to see if they add up to 100. Check the means to see if they are consistent with the raw data. These checks will not catch every error, but they will catch many common mistakes. ### Check the Funding and Conflicts The funding source and the conflicts of interest can affect how the data is presented. The [NIH Grants and Funding](https://grants.nih.gov/) page describes how grant applications are reviewed, and the review process includes an assessment of the investigators and their conflicts. A table from a study with a clear conflict of interest should be read with that conflict in mind. ## How to Read a Table in Different Types of Papers ### Tables in Clinical Trials Clinical trial tables often report the baseline characteristics of the participants, the outcomes, and the adverse events. The baseline table shows whether the treatment groups were similar at the start of the study. The outcome table shows the effect of the treatment. The adverse event table shows the safety profile. ### Tables in Animal Studies Animal study tables report the number of animals, the treatment groups, and the measured outcomes. The table may report the weight gain, the feed intake, or the mortality. The units are important because animal weights are often reported in grams or kilograms, and feed intake is often reported in grams per day. ### Tables in Laboratory Studies Laboratory tables report the results of experiments, such as the concentration of a substance or the activity of an enzyme. The units are often expressed in scientific notation, and the table may include the standard deviation or the standard error. The footnote will tell you how many replicates were performed. ### Tables in Observational Studies Observational study tables report the characteristics of the population and the associations between variables. The table may report the prevalence of a condition or the risk of a condition. The footnote will tell you whether the values were adjusted for confounding variables. ## How to Read a Table That Reports Statistical Results ### Understanding the P Value The p value is the probability of observing the data if the null hypothesis is true. A low p value suggests that the observed difference is unlikely to be due to chance. The footnote will tell you the threshold for significance, such as p below 0.05 or p below 0.01. ### Understanding the Confidence Interval The confidence interval is a range of values that is likely to contain the true value. A 95 percent confidence interval means that if the study were repeated many times, the interval would contain the true value 95 percent of the time. If the confidence interval does not include zero, the effect is statistically significant. ### Understanding the Standard Error The standard error is a measure of the variability of the mean. A smaller standard error means that the mean is more precise. The standard error is used to calculate the confidence interval and the p value. ### Understanding the Standard Deviation The standard deviation is a measure of the variability of the individual values. A larger standard deviation means that the values are more spread out. The standard deviation is often reported in the table as a plus or minus value. ## How to Use a Table to Make a Decision ### Making a Management Decision If you are a farmer or a laboratory manager, you may use a table to decide whether to change a practice. For example, a table that shows the mean weight gain of animals on different diets can help you choose a diet. The table will tell you the mean weight gain, but you also need to know the cost of the diet and the variability of the results. ### Making a Clinical Decision If you are a clinician, you may use a table to decide whether to prescribe a treatment. The table will show the outcomes of the treatment and the adverse events. You need to weigh the benefits against the risks, and the table provides the data for that weighing. ### Making a Research Decision If you are a researcher, you may use a table to decide whether to pursue a line of investigation. The table will show the results of a previous study, and you can use those results to design your own study. The table will also show the limitations of the previous study, and those limitations can inform your design. ## The Role of Reporting Guidelines in Table Quality The [EQUATOR Network](https://www.equator-network.org/) is a repository of reporting guidelines for health research. These guidelines specify what information should be included in a research paper, including the tables. A paper that follows a reporting guideline is more likely to have tables that are complete and transparent. The guidelines vary by study type. A randomized trial has a different guideline than an observational study, and a laboratory study has a different guideline than a clinical study. When you read a table, you can check whether the paper followed the relevant guideline. If the paper did not follow a guideline, the table may be missing important information. The [Committee on Publication Ethics](https://publicationethics.org/core-practices) core practices describe the responsibilities of authors, reviewers, and editors. These practices include the accurate reporting of data and the transparent handling of conflicts. A table that is accurate and transparent is the product of a publication process that follows these practices. ## The Role of Data Management in Table Quality The [NIH Data Management and Sharing Policy](https://sharing.nih.gov/data-management-and-sharing-policy) describes the expectations for the management and sharing of research data. The policy requires researchers to describe how they will manage and share their data. A table that is built on well-managed data is more likely to be accurate and reproducible. The data management plan should describe the data that will be collected, the standards that will be used, and the repository where the data will be deposited. The table in the paper is a summary of that data. If the data is well-managed, the table is more likely to be accurate. The policy also describes the sharing of data. When the data is shared, other researchers can check the numbers in the table and reproduce the analysis. This verification is a check on the accuracy of the table. ## The Role of the Researcher in a Table The [ORCID for Researchers](https://info.orcid.org/researchers) page describes the researcher identifier that links a researcher to their work. The ORCID record includes the researcher's publications, and the table in a paper is part of that record. When you read a table, you can check the ORCID record of the author to see their other work and their affiliations. The ORCID record also helps you to identify the author correctly. If two researchers have the same name, the ORCID distinguishes them. This distinction matters when you are trying to find the methods or the data behind a table. ## Common Failure Patterns in Table Reading ### Failure to Check the Denominator A table that reports percentages can be misleading if the denominator is not clear. A percentage of 50 percent can mean 50 out of 100 or 5 out of 10. The footnote or the column header should tell you the denominator. If it does not, the percentage is not interpretable. ### Failure to Check the Time Period A table that reports the incidence of a disease can be misleading if the time period is not clear. The incidence of 10 percent can mean 10 percent per year or 10 percent over five years. The column header or the footnote should tell you the time period. ### Failure to Check the Population A table that reports the mean weight of animals can be misleading if the population is not clear. The mean weight of 50 kg can mean the mean of all animals or the mean of a specific breed. The row label or the title should tell you the population. ### Failure to Check the Adjustment A table that reports the association between two variables can be misleading if the association is not adjusted. The unadjusted association can be different from the adjusted association. The footnote should tell you whether the association was adjusted for other variables. ## Limitations of Tables ### Tables Cannot Show the Full Picture A table is a summary of the data. It cannot show the distribution of the individual values, the outliers, or the missing data. The table shows the mean, but not the spread. The table shows the percentage, but not the number of events. ### Tables Cannot Show the Context A table cannot show the context of the study. It cannot show the methods, the limitations, or the conflicts of interest. The table is a part of the paper, and the paper provides the context. A table read in isolation can be misleading. ### Tables Cannot Show the Causality A table can show an association between two variables, but it cannot show that one causes the other. The causality is determined by the study design, not by the table. A table from an observational study cannot support a causal claim. ### Tables Cannot Show the Uncertainty A table can show the mean and the standard error, but it cannot show the full uncertainty of the estimate. The uncertainty is captured by the confidence interval, which is often reported in the text or in the footnote. The table alone does not show the full uncertainty. ## Professional Escalation Criteria If you are reading a table and you cannot determine the units, the sample size, or the significance level, you should not use the table to make a decision. You should go back to the methods section of the paper to find the missing information. If the methods section does not provide the information, you should contact the author or consult a colleague with expertise in the field. If you are reading a table and the numbers do not match the text, you should flag the discrepancy. The discrepancy could be a typographical error, or it could be a sign of a more serious problem. You should contact the journal or the author to clarify the discrepancy. If you are reading a table and the data appears to be too good to be true, you should be cautious. The table may be based on a small sample, a biased sample, or a flawed analysis. You should check the methods section and the conflicts of interest before you use the table. If you are a peer reviewer and you find a table that is not transparent, you should recommend that the authors revise the table. The table should include the units, the sample size, and the footnotes. The table should be consistent with the text and the methods. ## Frequently Asked Questions ### What is the first thing I should look at in a table? Read the title first. The title tells you the topic, the population, and the variables. Without the title, you cannot know what the numbers represent. ### How do I know what the numbers in a table mean? Read the column headers and the row labels. The column headers tell you the variable and the unit. The row labels tell you the group or the time point. The footnotes tell you the definitions and the limitations. ### What is the difference between a standard deviation and a standard error? The standard deviation measures the variability of the individual values. The standard error measures the precision of the mean. The standard error is smaller than the standard deviation and is used to calculate the confidence interval. ### How do I know if a difference in a table is statistically significant? Look for the p value or the confidence interval in the footnote. A p value below the significance level, usually 0.05, means the difference is statistically significant. A confidence interval that does not include zero also means the difference is statistically significant. ### What should I do if the table does not have footnotes? Look for the explanation in the methods section of the paper. If the methods section does not explain the table, contact the author or ask a colleague. Do not guess the meaning of the numbers. ### Can I use a table from a paper to make a decision? You can use a table to inform a decision, but you should not use it alone. The table should be read in the context of the paper, and the paper should be read in the context of the field. The table is one piece of evidence. ### How do I know if a table is reliable? Check the study design, the sample size, and the methods. Check the numbers for internal consistency. Check the conflicts of interest. A table from a peer-reviewed journal that follows the reporting guidelines is more likely to be reliable. ### What is the best way to read a table in a paper? Read the title, then the headers, then the row labels, then the footnotes, and then the data. This order ensures that you understand the structure of the table before you interpret the numbers. ## Using the Evidence | Source | Best use in this topic | Important limitation | |---|---|---| | [Research Methods Resources](https://www.ncbi.nlm.nih.gov/books) | official guidance | Check the linked page for current local requirements | | [EQUATOR Network](https://www.equator-network.org/) | official guidance | Check the linked page for current local requirements | | [Core Practices](https://publicationethics.org/core-practices) | official guidance | Check the linked page for current local requirements | ## Related Articles - [How to Read a Scientific Paper Efficiently](/blog/careers/how-to-read-a-scientific-paper-efficiently) - [Choosing the Right Citation Style for Your Scientific Paper](/blog/guides/choosing-the-right-citation-style-for-your-scientific-paper) - [Writing the Methods Section of a Scientific Paper: A Reproducibility-Focused Guide](/blog/guides/writing-the-methods-section-of-a-scientific-paper-a-reproducibility-focused-guide) - [How to Write a Scientific Paper Abstract: A Practical Guide](/blog/guides/how-to-write-a-scientific-paper-abstract-a-practical-guide) - [Scientific Data Provenance For Reproducible Research](/blog/guides/scientific-data-provenance-for-reproducible-research) ## References and Further Reading - [Research Methods Resources](https://www.ncbi.nlm.nih.gov/books). National Library of Medicine. - [EQUATOR Network](https://www.equator-network.org/). EQUATOR Network. - [Core Practices](https://publicationethics.org/core-practices). Committee on Publication Ethics. - [NIH Grants and Funding](https://grants.nih.gov/). National Institutes of Health. - [ORCID for Researchers](https://info.orcid.org/researchers). ORCID. - [Data Management and Sharing Policy](https://sharing.nih.gov/data-management-and-sharing-policy). National Institutes of Health. - [NCBI Literature Resources](https://www.ncbi.nlm.nih.gov/). National Center for Biotechnology Information. - [PubMed](https://pubmed.ncbi.nlm.nih.gov/). National Library of Medicine. > This article is educational and does not replace institutional policy, professional advice, or applicable safety and regulatory requirements.