What Did the Study Actually Find Before the Authors Interpreted It?

What did the study actually find?

When you read a research paper, the authors usually do more than report results. They explain those results, connect them with previous studies, propose mechanisms, discuss implications, and eventually tell you what they think the study means.

That interpretation may be thoughtful and well supported. But it can also influence how you perceive the evidence before you have examined the evidence yourself.

A useful critical-reading habit is therefore surprisingly simple: before asking what the authors think their study means, ask a more basic question. What actually happened in the data?

This means reconstructing the main findings as plainly as possible before moving into explanations, implications, recommendations, or conclusions.

Identify the findings before deciding what they mean

How to reconstruct a study’s findings before interpretation

Start with the question the study was designed to answer

A result only makes sense in relation to what the researchers actually investigated. Before extracting findings, briefly return to the research question, hypotheses, outcomes, variables, or objectives specified in the paper.

This matters because papers can contain many results. Some correspond directly to the primary question, while others may be secondary outcomes, subgroup analyses, sensitivity analyses, exploratory findings, or observations that emerged after the main analysis.

If you begin with whatever result seems most interesting, you can accidentally give an ancillary finding more importance than the study design originally assigned to it.

Read the Results section as evidence, not as a storyline

In a well-reported quantitative study, the Results section should tell you what was observed in the sample and what the analyses produced. Depending on the design, this may include participant numbers, descriptive statistics, group differences, associations, effect estimates, confidence intervals, outcome events, model coefficients, sensitivity analyses, and other relevant results.

Major reporting frameworks make a similar distinction. For example, STROBE asks observational studies to report outcome data, main estimates and their precision, and additional analyses in the Results section, while interpretation is addressed separately in the Discussion. CONSORT 2025 likewise asks randomized trials to report results for primary and secondary outcomes, including effect sizes and precision, and separately calls for an interpretation consistent with those results.

Your first task as a reader is therefore not to decide whether the result is exciting, important, convincing, or useful. It is to establish what was actually observed.

Reduce the central result to a neutral statement

Try rewriting the main finding in language that contains as little interpretation as possible.

Suppose a paper reports that students using an instructional intervention obtained a mean score of 78.4, compared with 75.9 in the comparison group, for an estimated difference of 2.5 points with a 95% confidence interval from 0.4 to 4.6.

A relatively neutral reconstruction would be:

The intervention group scored an estimated 2.5 points higher than the comparison group, with a 95% confidence interval of 0.4 to 4.6 points.

Notice what this statement does not yet say. It does not call the intervention effective, transformative, educationally meaningful, cost-effective, or responsible for the difference under every possible design. Those are interpretive claims that require further reasoning.

Look for magnitude and uncertainty, not only statistical significance

If the study is quantitative, reconstructing the findings usually requires more than recording whether a p-value crossed a conventional threshold. The size and direction of an estimated effect often matter, as does its precision.

CONSORT 2025, for example, calls for reporting the estimated effect size and its precision, such as a 95% confidence interval, for trial outcomes. A confidence interval provides information about the range of effect values compatible with the observed data under the model and assumptions used.

Consequently, “statistically significant” is rarely an adequate summary of what a study found. A small but precisely estimated difference and a large but highly uncertain difference can tell quite different evidential stories even when attention is drawn to a p-value.

The question of whether a statistically detectable result is important in practical terms should be considered separately.

Distinguish observations from explanations

One of the easiest ways to identify interpretation is to look for statements that answer why rather than what.

For example, “participants in Group A reported lower anxiety scores than participants in Group B” describes a result. “The intervention reduced anxiety because it improved participants’ sense of control” goes further. It introduces a causal interpretation and a proposed mechanism.

The second statement might ultimately be justified. The important point is that it is not identical to the first.

Tables and figures can help you bypass narrative emphasis

Authors must decide which findings to emphasize in prose. Tables and figures can sometimes give you a broader view of the reported evidence, including estimates that receive little attention in the narrative.

For a quantitative paper, compare the textual summary with the relevant tables and figures. Check the actual group values, effect estimates, confidence intervals, sample sizes, and results for the primary outcome. This makes it easier to separate the numerical results from the narrative built around them.

Do not assume that a table is somehow interpretation-free, however. Researchers still choose variables, models, reference categories, transformations, and analyses. The table reports outputs from those analytical decisions. Reading it directly simply reduces one additional layer of narrative framing.

Qualitative findings require a different kind of separation

The same principle applies to qualitative research, but “the data show” has a different meaning. Qualitative findings are not simply raw participant statements waiting to be counted. Analysis may involve coding, categorization, thematic development, comparison, interpretation, and theoretically informed abstraction.

For that reason, you should not pretend that a qualitative theme is equivalent to an uninterpreted numerical estimate. Instead, examine how the authors move from participant material or other qualitative evidence to codes, categories, themes, explanations, or conceptual claims.

Where quotations or other excerpts are provided, ask whether they plausibly illustrate the analytical claim being made and whether alternative readings remain possible. This requires carefully distinguishing participant accounts from the researcher’s interpretation without assuming that qualitative analysis can or should be interpretation-free.

Results and findings are already products of methodological choices

There is an important complication here. “What the study found” is not necessarily synonymous with untouched reality.

Researchers decide what to measure, how to operationalize constructs, which participants to include, how to handle missing data, what analytical model to use, which qualitative material to code, and how to define outcomes. The resulting findings are produced through those methodological and analytical choices.

So the purpose of separating findings from interpretation is not to imagine that findings are completely theory-free or judgment-free. It is to distinguish the evidence reported by the study from the additional claims the authors build from that evidence.

Primary, secondary, subgroup, and exploratory findings are not interchangeable

A study may report dozens of analyses. Before forming an overall impression, determine which findings correspond to the primary outcome or prespecified hypotheses and which came from additional analyses.

This distinction can be consequential. CONSORT 2025 asks authors to distinguish prespecified analyses from post hoc analyses, including subgroup and sensitivity analyses. An interesting subgroup result may deserve attention, but it should not quietly replace an unconvincing primary result as the apparent centerpiece of the study.

If the narrative concentrates on a particularly favorable subset of participants, examine whether the paper has shifted attention from the primary result to a subgroup finding.

Separate the evidential statement from the inferential statement

A useful reading technique is to formulate two separate sentences.

Evidence sentence: What specifically did the analysis produce?

Interpretation sentence: What do the authors infer from that result?

Keeping those sentences separate makes it easier to detect when a paper moves from association to causation, from uncertainty to certainty, or from a statistical result to a practical recommendation.

This is the broader distinction between what the data show and what the authors conclude. The two should connect, but they should not be treated as synonyms.

Strip a research claim back to the underlying result

This exercise does not require you to reject the authors’ interpretation. It simply prevents the interpretation from becoming indistinguishable from the result itself.

You can now evaluate the next layer of questions more carefully. Was the difference large enough to matter? Was the estimate precise? Were there missing data or protocol deviations? Does the design support a causal inference? Was engagement actually measured as a mechanism? Does the sample support the population-level claim?

Those are appraisal questions. The first step was simply establishing what happened.

Common ways interpretation gets mistaken for findings

Read the paper in two passes: evidence first, interpretation second

You do not need to distrust authors to read critically. The more useful habit is to temporarily delay their interpretation while you reconstruct the evidence.

After you have reconstructed the findings, return to the Discussion and Conclusion. At that point, the authors’ interpretation becomes something you can evaluate rather than something that automatically defines the study for you.

For consequential papers, you may even find it useful to form a provisional interpretation before reading the authors’ conclusion. The purpose is not to compete with the authors. It is to make your own inferential process visible enough that you can compare the two.

Can you state the findings without borrowing the authors’ conclusions?

Questions about separating findings from interpretation

Find the result before accepting the story around it

How to Cite This Guide

This guide is intended to be read, shared, and used in research, teaching, and academic work. If you draw on its ideas, explanations, or other content, please acknowledge the source by citing the guide. Doing so gives appropriate credit and helps your readers locate the original resource.