The most important table you’ve been ignoring
In addition to blogging about recent publications coming out of our lab, we’ll also be sharing others’ publications and resources that we may create to help people interested in these topics. In today’s post, we discuss Dr. Murphy’s 2021 publication on the value of descriptive statistics, which inspired my own approach to teaching introductory statistics.
In organizational research, “Table 1” is often the quietest table in the paper — dutifully reporting means, standard deviations, and correlations, and just as quickly ignored. But what if Table 1 is actually the most important part of a study?
That’s the case Kevin Murphy makes in his 2021 focal article for Industrial and Organizational Psychology: Perspectives on Science and Practice (IOP). As part of the journal’s “focal article – commentary” format, Murphy’s piece is followed by over a dozen commentaries from scholars who weigh in, critique, or extend the core argument. (I was lucky enough to contribute one of these myself, on the need to teach data visualization early in statistics training.)
Murphy’s central message is simple, but powerful: We can easily overcomplicate organizational science. In our rush toward increasingly sophisticated inferential analyses, we’ve undervalued the very statistics that should anchor our work: the descriptives. “Table 1,” Murphy writes, “is the Cinderella of tables. It is often overlooked, but it can be of immense value” (p. 473).
Three Problems, One Solution
The article opens with an audit of quantitative articles in JAP and JOM, showing the wide variety of analytical methods being used: from OLS to SEM to multilevel and latent variable models. Murphy identifies three core issues emerging from this complexity arms race:
Misapplication and misunderstanding of advanced methods. Even seasoned researchers regularly miscalculate degrees of freedom or misinterpret interaction terms.
Overreliance on statistical significance, often without reporting or interpreting effect sizes.
Poor communication with practitioners and general audiences due to convoluted methods and jargon-laden interpretations.
His solution? Re-center our scientific reasoning on Table 1. Instead of treating descriptive statistics as a formal requirement to breeze past, we should use them to critically examine whether our hypotheses are even plausible before moving on to complex modeling.
“Feasibility Testing” with Descriptives
One of Murphy’s most useful contributions is a practical framework for evaluating common hypothesis types using Table 1 alone:
Intervention studies: Are group means in the predicted direction? Is the effect size meaningful?
Moderator studies: Are the correlations among variables such that moderation is statistically feasible (e.g., large enough sample, low multicollinearity)?
Mediator studies: Do the proposed causal pathways hold up and match theory? Are the correlations between predictor to mediator, and mediator to outcome, stronger than the direct correlation of predictor to outcome?
These tests won’t prove your hypothesis, but they can flag when it’s not even reasonable to proceed. And as Murphy points out, they could save researchers from drawing “reverse” conclusions or chasing statistical ghosts that simple plots or correlations would’ve shown to be unlikely from the start.
A Call for Simplicity
The article closes with a motto borrowed from the designers of the F-5 fighter jet: “Add simplicity and lightness.” Murphy suggests we apply this philosophy to our statistical storytelling. Descriptive statistics offer a first-order approximation of the world, one that is often more interpretable and more persuasive than any three-level, cross-lagged SEM.
For those of us who teach statistical methods, Murphy’s piece is also a challenge. I’ve found — as I’m sure many other instructors have as well — that explaining a simple concept in a meaningful way can actually be as hard as explaining more complex ideas. What if we trained students not just in estimating coefficients but in reading descriptive tables — slowly, carefully, contextually?
That was the spirit of my own commentary in response — which I hope will be reflected in my upcoming Intro Stats class at CMC: we should be teaching data visualization, data cleaning, and descriptive analysis early and often, not as “nice to have” supplements, but as core tools of scientific reasoning.
Read More:
Ahmad, A. S., & Zhou, S. (2021). Spreading the word: Equipping IO students to use descriptive statistics for effective data visualization. Industrial and Organizational Psychology, 14(4), 510-513. https://doi.org/10.1017/iop.2021.115
Murphy, K. R. (2021). In praise of Table 1: The importance of making better use of descriptive statistics. Industrial and Organizational Psychology, 14(4), 461-477. https://doi.org/10.1017/iop.2021.90
