How to Report a Logistic Regression in APA 7

Logistic regression is reported differently from linear regression, and that's where many go wrong: the star is the odds ratio, not the B coefficient alone. Here's how to report it in APA 7 without getting flagged. The general guide is in how to report results in APA 7.

What the report must include

For each predictor: the B coefficient with its standard error, the Wald statistic with its p-value, and above all the odds ratio (OR, which is e raised to B) with its 95% confidence interval. For the model: a global fit measure (the omnibus χ² of the model with its degrees of freedom), a pseudo-R² (Nagelkerke or Cox-Snell), the Hosmer-Lemeshow goodness-of-fit test and, where appropriate, the correct classification rate.

Example in the text

"Anxiety significantly predicted the probability of dropout, B = 0.37, SE = 0.13, Wald = 8.10, p = .004, OR = 1.45, 95% CI [1.12, 1.88]. The overall model was significant, χ²(3) = 24.6, p < .001, and explained 18% of the variance (Nagelkerke R² = .18)."

Interpretation: an OR of 1.45 means that, for each one-unit increase in anxiety, the odds of dropout are multiplied by 1.45. Because the OR can exceed 1, it keeps the leading zero when it has one (1.45). As for formatting, B, SE, p and R are italicised; χ² (Greek) and "Wald" (a proper name) are not.

Common mistakes

The most penalised: giving the B without the OR; not reporting the OR's confidence interval (the thing reviewers ask for most); omitting the model fit measure; and interpreting the OR as if it were a relative risk, which it isn't. If your outcome is ordinal or a count, logistic regression isn't the right choice: you'd need ordinal logistic or Poisson regression.

That's your logistic regression properly reported. If you want to review the assumptions or the fit, we look at it in my statistical consulting for theses and papers.