Correlation vs Causation: Common Mistakes in Student Research
Princeton Journal of Pre-Collegiate Research

Within the context of pre-collegiate and early undergraduate research methodology, an area of growing scientific importance given its implications for the integrity of published findings, the conflation of correlation with causation represents one of the most consequential and frequently observed errors in student-authored manuscripts. The distinction between these two concepts is not merely semantic; it is foundational to the validity of any empirical claim. Manuscripts submitted to peer-reviewed journals, including those specifically designed to publish pre-collegiate work, are routinely returned for revision or rejected outright when authors fail to maintain analytical precision regarding the directional and mechanistic claims their data can legitimately support. A thorough understanding of the conceptual boundary separating correlation from causation, as well as the specific methodological and rhetorical mistakes that arise from its neglect, is therefore essential for any student researcher seeking to produce credible, publishable scholarship.
Defining the Distinction: Correlation vs Causation in Student Research
Correlation, in its statistical formulation, describes a quantifiable relationship between two or more variables, such that changes in one variable are associated with systematic changes in another. This relationship is expressed through coefficients such as Pearson's r or Spearman's rho, and it carries no inherent implication regarding the mechanism, directionality, or independence of the observed association. Causation, by contrast, asserts that one variable directly produces or influences change in another, a claim that requires substantially more rigorous evidentiary support than correlation alone can provide.
The criteria most widely accepted for establishing causal inference in empirical research include temporal precedence, whereby the proposed cause must precede the proposed effect; covariation, whereby the two variables must be statistically associated; and the elimination of plausible alternative explanations, typically achieved through experimental control, randomization, or advanced statistical modeling techniques such as instrumental variable analysis or propensity score matching. Observational and cross-sectional study designs, which constitute the majority of research conducted by pre-collegiate students due to practical and ethical constraints, are inherently limited in their capacity to satisfy these criteria, a limitation that must be explicitly acknowledged in any manuscript submitted for peer review.
Common Mistakes in Student Research: Overstating Causal Claims
The most prevalent error observed in student manuscripts is the use of causal language to describe findings derived exclusively from correlational data. Verbs such as "causes," "produces," "leads to," "results in," and "determines" carry explicit causal implications and are therefore inappropriate in the discussion or conclusion sections of studies that employed observational, survey-based, or retrospective methodologies. When a student researcher reports, for example, that "increased screen time causes elevated anxiety levels in adolescents" on the basis of a cross-sectional survey, the claim exceeds what the data can support. The appropriate formulation would instead assert that screen time and anxiety levels were found to be positively correlated within the sampled population, with the acknowledgment that the directionality and mechanism of this association remain to be established through longitudinal or experimental investigation.
This pattern of overstating causal claims is not confined to verbal formulations. Students frequently embed causal assumptions into the structural logic of their hypotheses, their variable selection procedures, and their interpretation of regression coefficients. A regression coefficient, regardless of its magnitude or statistical significance, does not establish that the independent variable causes variation in the dependent variable; it establishes only that the two are linearly associated after controlling for the variables included in the model. Omitted variable bias, reverse causality, and measurement error can all produce statistically significant regression coefficients in the absence of any true causal relationship, a point that must be addressed in the limitations section of any observational study.
Students seeking foundational guidance on the statistical reporting conventions relevant to these distinctions are encouraged to consult the detailed discussion available in the resource addressing common statistical mistakes in student papers, which addresses related errors in the interpretation and presentation of quantitative findings.
Confounding Variables and the Spurious Correlation Problem
A particularly instructive category of error involves the failure to account for confounding variables, which are third variables that are causally related to both the independent and dependent variables under examination. When a confounding variable is present but uncontrolled, the observed correlation between the primary variables of interest may be entirely spurious, meaning that it reflects the shared influence of the confounder rather than any direct relationship between the variables themselves. Classic examples from the epidemiological literature, such as the positive correlation between ice cream sales and drowning rates, which is explained by the confounding variable of ambient temperature and seasonal outdoor activity, are frequently cited in methods courses precisely because they illustrate the danger of inferring causation from correlation with clarity and memorability.
In student research, confounding is especially problematic in studies examining behavioral, social, or biological outcomes across demographically heterogeneous samples. A student investigating the relationship between dietary habits and academic performance, for instance, must contend with the possibility that socioeconomic status, sleep duration, parental education level, and school resource availability are all simultaneously associated with both the dietary predictor and the academic outcome. Failure to measure and statistically control for these potential confounders renders any observed correlation between diet and academic performance uninterpretable from a causal standpoint.
The analytical resources available through the discussion of correlation vs causation in student research provide additional structured guidance for researchers navigating these methodological challenges within pre-collegiate and early collegiate research contexts.
Mistakes in Data Visualization That Imply Causation
Beyond the textual formulation of claims, student researchers frequently introduce causal implications through the visual presentation of their data. Scatter plots with fitted regression lines, bar charts comparing group means, and time-series graphs are all subject to interpretive distortion when presented without appropriate contextual annotation. A regression line fitted to a scatter plot, for example, is a descriptive statistical tool that summarizes the linear association between two variables; it does not represent a causal pathway, and labeling the axes as "cause" and "effect" rather than "predictor" and "outcome" is a substantive methodological error.
Similarly, before-and-after visualizations, which display a measured outcome at two time points, are frequently misinterpreted as demonstrating that an intervening treatment or exposure produced the observed change. In the absence of a control group and randomization, such visualizations cannot rule out maturation effects, regression to the mean, or concurrent historical events as alternative explanations for the observed change. Students are therefore advised to accompany all visualizations with figure captions that explicitly characterize the displayed relationship as associational rather than causal, and to reserve causal language exclusively for findings derived from appropriately designed experimental studies.
A comprehensive examination of the specific presentational errors that introduce interpretive ambiguity into student manuscripts is available in the resource addressing common data visualization mistakes in student papers, which provides concrete guidance on figure design and annotation standards consistent with peer-reviewed publication requirements.
Correlation vs Causation: Common Mistakes in Hypothesis Formulation
The conflation of correlation with causation frequently originates not in the results or discussion sections of a manuscript but in the hypothesis formulation stage, where students construct directional, mechanistic predictions that their chosen research design is structurally incapable of testing. A hypothesis that states "exposure to bilingual educational environments causes improved executive function in elementary school students" presupposes a causal mechanism and implies that an experimental or quasi-experimental design has been employed to test it. When the study in question is a cross-sectional survey comparing executive function scores between students enrolled in bilingual versus monolingual programs, the hypothesis is misaligned with the methodology, and the resulting manuscript will contain a fundamental logical inconsistency that peer reviewers are trained to identify.
The appropriate reformulation of such a hypothesis would assert that students enrolled in bilingual educational environments are expected to demonstrate statistically higher executive function scores than their monolingual counterparts, with the understanding that this prediction describes an anticipated correlational pattern rather than a causal claim. This reformulation preserves the scientific content of the prediction while accurately representing the epistemological limits of the research design.
Students who are initiating their first independent research projects and who seek structured guidance on the full arc of the research process, from question formulation through manuscript preparation, are directed to the introductory resource on high school student research and how to get started, which situates methodological precision within the broader context of pre-collegiate scholarly practice.
The Role of Language Precision in Academic Manuscripts
The linguistic dimension of the correlation-causation distinction is sufficiently significant to warrant dedicated attention in the revision process. Academic writing conventions in the empirical sciences have developed a standardized vocabulary for characterizing the epistemic status of research findings, and student authors are expected to deploy this vocabulary with accuracy. Phrases such as "was associated with," "was positively correlated with," "predicted," "was a significant predictor of," and "was linked to" are appropriate for describing correlational findings. Phrases such as "caused," "produced," "led to," "resulted in," and "determined" are reserved for findings derived from experimental designs with adequate controls and randomization.
The distinction between "predicts" and "causes" in regression analysis deserves particular emphasis. In the context of multiple regression, a variable that "predicts" the outcome is one whose coefficient is statistically significant within the model; this terminology carries no causal implication and is consistent with the observational nature of most student research. Students who substitute "causes" for "predicts" in their discussion of regression results are making a claim that their data cannot support and that will be flagged by any competent peer reviewer.
Grammatical precision in this domain intersects with broader concerns about academic writing quality that are addressed in the resource examining common grammar mistakes in academic research papers, which provides detailed guidance on the syntactic and lexical conventions governing scientific prose.
Strengthening Observational Research Through Appropriate Framing
It is important to recognize that the limitation of observational research to correlational claims does not diminish the scientific value of such research. Correlational findings serve critical functions in the scientific enterprise, including the identification of associations that warrant further experimental investigation, the documentation of population-level patterns, and the generation of hypotheses regarding potential causal mechanisms. A student manuscript that accurately characterizes its findings as correlational, rigorously addresses the role of potential confounders, and articulates a clear rationale for why the observed association is theoretically meaningful makes a genuine and defensible contribution to the scholarly literature.
The framing of limitations is equally important. Peer-reviewed journals expect authors to explicitly acknowledge the constraints imposed by their research design and to discuss the specific threats to causal inference that are relevant to their study. A limitations section that identifies the cross-sectional design as precluding causal inference, names the specific confounders that could not be controlled, and recommends a longitudinal or experimental follow-up study demonstrates the kind of methodological self-awareness that distinguishes rigorous scholarship from superficial reporting.
Students who are preparing manuscripts for submission to pre-collegiate journals and who seek guidance on the publication process within specific national and regional contexts may find relevant information in the resources addressing publication pathways for high school researchers in various international settings, including those available for researchers in Germany, France, and Brazil, each of which addresses the institutional and procedural considerations specific to those research environments.
Conclusion
The distinction between correlation and causation constitutes one of the most fundamental principles of empirical research methodology, and the failure to maintain this distinction is among the most consequential mistakes observable in student-authored manuscripts. Accurate characterization of the epistemic status of research findings, appropriate use of statistical and linguistic conventions, rigorous identification and control of potential confounders, and transparent acknowledgment of design limitations are all essential components of scholarly practice at every level of the research enterprise. Pre-collegiate researchers who internalize these principles and apply them consistently throughout the research and writing process will produce manuscripts that are not only more likely to withstand peer review but that also make more credible and lasting contributions to the bodies of knowledge they seek to advance. The Princeton Journal of Pre-Collegiate Research remains committed to supporting student researchers in developing the methodological rigor and analytical precision that distinguish high-quality empirical scholarship from preliminary or speculative inquiry.
Have questions or want personalised guidance? Contact us and we'll be happy to help.
Read More

Axions and the Search for Dark Matter, Explained for Students
By
Princeton Journal of Pre-Collegiate Research
Read more

Axions and dark matter: 10 Research Questions and Experiment Ideas for Students
By
Princeton Journal of Pre-Collegiate Research
Read more

Axions and dark matter: 12 Questions Students Ask, Answered
By
Princeton Journal of Pre-Collegiate Research
Read more

Distance From the Sun to Earth (and Mars): Solar System Distances Explained
By
Princeton Journal of Pre-Collegiate Research
Read more

Distances in the solar system: 10 Research Questions and Experiment Ideas for Students
By
Princeton Journal of Pre-Collegiate Research
Read more

Distances in the solar system: 12 Questions Students Ask, Answered
By
Princeton Journal of Pre-Collegiate Research
Read more

Does Mars Have a Magnetic Field? What We Know and How We Know
By
Princeton Journal of Pre-Collegiate Research
Read more

The Martian magnetic field: 10 Research Questions and Experiment Ideas for Students
By
Princeton Journal of Pre-Collegiate Research
Read more

The Martian magnetic field: 12 Questions Students Ask, Answered
By
Princeton Journal of Pre-Collegiate Research
Read more

Rolling Friction Explained: Definition, Examples and Experiments for Students
By
Princeton Journal of Pre-Collegiate Research
Read more

Rolling friction: 10 Research Questions and Experiment Ideas for Students
By
Princeton Journal of Pre-Collegiate Research
Read more

Rolling friction: 12 Questions Students Ask, Answered
By
Princeton Journal of Pre-Collegiate Research
Read more

How to Find the Coefficient of Kinetic Friction: A Student Experiment Guide
By
Princeton Journal of Pre-Collegiate Research
Read more

The coefficient of kinetic friction: 10 Research Questions and Experiment Ideas for Students
By
Princeton Journal of Pre-Collegiate Research
Read more

The coefficient of kinetic friction: 12 Questions Students Ask, Answered
By
Princeton Journal of Pre-Collegiate Research
Read more

Harmonics and Standing Waves: What They Are and How to Test Them
By
Princeton Journal of Pre-Collegiate Research
Read more

Harmonics and standing waves: 10 Research Questions and Experiment Ideas for Students
By
Princeton Journal of Pre-Collegiate Research
Read more

Harmonics and standing waves: 12 Questions Students Ask, Answered
By
Princeton Journal of Pre-Collegiate Research
Read more

AM-GM Inequality: Proof, Examples and Olympiad Uses
By
Princeton Journal of Pre-Collegiate Research
Read more

For Educators: Supporting Student Research and Publication
By
Princeton Journal of Pre-Collegiate Research
Read more

How Teachers Can Support a Student’s First Publication
By
Princeton Journal of Pre-Collegiate Research
Read more
