Reporting Correlation Analysis

Reporting Correlation Analyses in Research Manuscripts

Correlation analysis is one of the essential statistical methods used to investigate the relationships among continuous or ordinal variables.

Reporting Correlation Analyses

Recent Post

Introduction

Correlation analysis is one of the essential statistical methods used to investigate the relationships among continuous or ordinal variables. Whether we are examining the correlation between clinical outcomes, financial variables, or psychological measures, identifying the proper ways to report correlation analyses needs to occur to create transparency, reproducibility of analyses, and facilitate proper interpretation of our findings. In this article, we will describe fundamental concepts to report correlation analyses in research manuscripts and provide examples from the literature to illustrate how these concepts can be incorporated in a variety of research disciplines, domains, or areas of study.

State the Purpose of the Analysis

The first step to report a correlation analysis is to clearly state the intent of the analysis. First, why did you conduct the correlation? That is, what were you seeking to find from the analysis? The intent of the correlation needs to be expressed well for the reader to understand the context and derivation for the statistical exploration.

Example:

  • “To analyse how cloud technology adoption relates to the efficiency of accounting functions.”
  • “To analyse how sleep time and depression are predictors of cognitive development in adolescents. “

In Georgiev (2025) the study was to establish the relationship of cloud technology adoption to improved IFRS compliance and accounting efficiency. The authors set out to determine if the technology brought forth a transformation, and the extent to which this transformational technology impacted the overall accounting function performance in organizations. This was a primary goal used to assess the transformational implications of digital systems in financial reporting.

Similarly, Cao (2025) also undertook correlation analysis to quantify sleep time and symptoms of depression as predictors of cognitive development in adolescents without explicit mention; however, to understand the indirect implications for long-term and underlying influence on adolescent learning and development.

Summarize the Variables Involved

Once the purpose is established, the next step is to summarize the variables under study in an explicit manner. The variables can include descriptions of descriptive statistics (mean, standard deviation, range and median) as well as measure of scale (continuous, categorical, continuous), distribution of measure or the description of the range of variation, etc.

Example:

  • “Cloud technology adoption (the independent variable) was scored on a scale of one to five (1- no adoption, 5- full adoption). Average processing time per financial transaction was the dependent variable (accounting efficiency).”

In Daver et al. (2025), data points containing descriptive statistics (survival rate and remission duration) related to leukaemia treatment were presented to provide context for interpreting the survey research aimed at determining how combination therapy impacted treatment outcomes. Descriptive data points also allowed the reader to consider the baseline, such as age and medical history of participants, which is essential to understanding the way treatment relates to survival outcomes.

It is important to determine the type of correlation coefficient used in your analysis because it informs how you can think about relationships among variables – dependent on the data type.

  • Pearson’s r is used on a linear relationship where the data are continuous, and the data are directly distributed normal.
  • Spearman’s rho is used for ordinal data or when data are not normally distributed.
    Rico-Bordera et al. (2025), decided to use Pearson’s correlation because they looked at the relationship of the Dark triad personality traits and inter-rater agreement on those traits, which were based on continuous and normally distributed data.

In contrast, Peng et al. (2025), used Spearman’s correlation coefficient since they examined relationships between physical abuse, internet addiction, and anxiety in adolescents by assessing ordinal variables (e.g., severity of abuse) and not normally distributed data.

Prior to conducting the correlation analysis, it is important to check that the statistical assumptions of the analysis method are met. For example, Pearson’s correlation requires data to be linear and have a normal distribution while Spearman’s correlation is a non-parametric analysis and thus has no assumptions.

If the assumptions of normality or linearity are not met, it is appropriate to justify the use of non-parametric analyses or to apply transformations. This demonstrates transparency in your analysis, which is paramount to analysing the data correctly and confirming your conclusions are warranted.

Example:

  • Fu et al. (2025) determined that the data was normally distributed prior to applying Pearson’s r to the multiplex graph diffusion network analysis. They applied the Shapiro-Wilk test to check for normality and subsequently determined if the data were not normally distributed, appropriate transformations were applied.
One of the most vital considerations when reporting correlation analyses is to report the statistical significance of Bills, 2002, and other sources. Simple state which alpha level is being used (typically the threshold alpha level of  =0.05 or 0.01, or something else if it was selected/acclaimed) and report the correlation coefficient value (e.g., “r = 0.62”). You can also add a 95% confidence interval (CI) to present an estimate of the precision of the correlation coefficient.

Example:

  • “r = 0.62, p < 0.05 , 95% CI: 0.45–0.76”

In Wei et al., (2025), the researchers stated oil pollution and community changes in microorganisms of ecosystems, including specific statistical significance and confidence intervals of the correlation. This provided clarity in the communication of strength and confidence of the finding in their study.

The use of weak, moderate, or strong also has ambiguity unless using established alpha thresholds. The weakness or strength of a correlation may depend on the researchers where the variables were located on a continuum of rankings of larger distributions, where empirical studies often occur.

Include Visual Support

Means that in addition to the statistical text, there is also the visual support of something like scatter plots or correlation matrices for when they are sharing the findings. The visual would provide a graphical, and often visually appealing reference for how two variables relate and allow readers to interpret findings at a glance.

Example:

  • Fu et al. (2025) presented a scatter plot as an illustration of the relationships they examined in their multiplex graph diffusion networks. They included correlation coefficients, confidence intervals, sample size, and P-values in the visual representation.
By including appropriate visual support not only is helpful for the reader to interpret the study but also adds to the credibility of the manuscript as it helps readers visualize the information.

Specify the Statistical Software Used

  • A rather important piece of information authors should include in their methodology section is the software they used to complete the analysis. Some of the more common software include, SPSS, R, SAS, STATA, etc.
  • For example, in Cao (2025) study the authors used SPSS software was used to complete the correlation analysis and cross-lagged modeling to see predictive relationship between sleep, depression, and cognitive development.

    • Summary Checklist for Reporting Correlation Analyses

    Correlation Analysis Reporting Elements
    Reporting Element Required?
    Purpose of the correlation analysis Yes
    Descriptive statistics for each variable Yes
    Type of correlation coefficient stated Yes
    Assumptions confirmed Yes
    Alpha level and p-value reported Yes
    Correlation coefficient with 95% CI Yes
    Subjective labels used cautiously Recommended
    Scatter plot included Recommended
    Software identified Yes

    Conclusion

    By following these best practices ensure a correlation analysis that is scientifically valid, statistically significant, and suitable for publication. Adequate documentation of correlation analyses not only enhances the quality and reproducibility of the research work but also provides the peer reviewer with important context for assessing the reliability of the ultimate hypothesis being validated. Consequently, following these guidelines enables the researcher to reach ample clarity, conciseness, and context with systematic reporting of their correlation analysis which ultimately enhances the reliability of their academic work and public health potential.

    Require Professional Help Alongside Reporting Association Analyses for Your Biomedical Work?
    PhD Assistance Research Lab has dedicated services to help you report your association analyses accurately and transparently in your study.
    Contact us today to ensure that your study complies with the best statistical reporting practices to enhance the credibility and reproducibility of your research to successfully publish your work!

    References

    1. Cao, X. (2025). Sleep time and depression symptoms as predictors of cognitive development among adolescents: A cross-lagged study from China. Psychological Reports, 128(3), 1566-1587.

    2. Daver, N., Senapati, J., Kantarjian, H. M., Wang, B., Reville, P. K., Loghavi, S., … & Abbas, H. A. (2025). Azacitidine, Venetoclax, and Magrolimab in newly diagnosed and relapsed refractory acute myeloid leukemia: Phase Ib/II study and correlative analysis. Clinical Cancer Research, 31(12), 2386-2398. https://doi.org/10.1158/1078-0432.CCR-24-2904

    3. Fu, S., Peng, Q., He, Y., Du, B., Zou, B., Jing, X. Y., & You, X. (2025). Unsupervised multiplex graph diffusion networks with multi-level canonical correlation analysis for multiplex graph representation learning. Science China Information Sciences, 68(3), 132102. https://doi.org/10.1007/s11432-025-3378-1

    4. Georgiev, M. (2025). Examining the correlation between cloud technology, the improvement of IFRS development, and the efficiency of accounting functions. BizInfo Blace. https://doi.org/10.1016/j.bizinfo.2025.07.001

    5. Peng, J., Liu, Y., Wang, X., Yi, Z., Xu, L., & Zhang, F. (2025). Physical and emotional abuse with internet addiction and anxiety as a mediator and physical activity as a moderator. Scientific Reports, 15(1), 2305. https://doi.org/10.1038/s41598-025-07379-z

    6. Perčević, H., & Ercegović, M. (2025). Correlation between income and expenses from investment properties and net profit in insurance companies in Croatia. In The Multidisciplinary Conference on Intangibles (pp. 355-369). Springer, Cham. https://doi.org/10.1007/978-3-030-36093-0_33

    7. Rico-Bordera, P., Pineda, D., Galán, M., & Piqueras, J. A. (2025). Assessing the dark personality traits with observer reports: A meta-analysis of inter-rater agreement on the Dark Triad and Dark Tetrad traits. Personality and Mental Health, 19(1), e1639. https://doi.org/10.1002/pmh.1639

    8. Wei, Y., Zhu, Y., Yang, L., Chen, C., Yue, M., Mao, Z., … & Xue, W. (2025). Effects of oil pollution on the growth and rhizosphere microbial community of Calamagrostis epigejos. Scientific Reports, 15(1), 1278. https://doi.org/10.1038/s41598-025-12356-1

    9. Yahaya, P. D. O. A. (2025). Sustainability reporting quality in the face of the board of directors. Available at SSRN 5094168. https://ssrn.com/abstract=5094168