APA 7 asks for confidence intervals alongside point estimates and effect sizes: they show how precisely each was estimated. The format is fixed and short. This guide gives the rules and examples for the intervals you are most likely to report.
The APA 7 format template
[Statistic] = X.XX, 95% CI [X.XX, X.XX]
All 53 templates, ready to paste with the italics set: the Rescue Vault, $27 →
Formatting rules reviewers actually check:
- Square brackets around the limits, separated by a comma: [1.70, 10.90].
- State the confidence level (95% CI) the first time; later CIs in the same paragraph can be "95% CIs [...]" or just CI if the level does not change.
- CI is not italicized (it is an abbreviation, not a statistic).
- Same decimals as the statistic it belongs to.
- Leading zero follows the statistic: none for correlations and R², kept for d, b and mean differences.
- Use a minus sign for negative limits: [−12.67, −1.52].
Worked examples
Mean difference (t-test):
t(58) = 2.74, p = .008, d = 0.71, 95% CI [1.70, 10.90]
Correlation:
r(60) = .48, p < .001, 95% CI [.27, .64]
Effect size:
d = 0.95, 95% CI [0.09, 1.79]
Odds ratio:
OR = 3.14, 95% CI [1.50, 6.58]
Bootstrapped indirect effect:
ab = −0.22, 95% CI [−0.32, −0.14]
Say what the interval is for
"95% CI [1.70, 10.90]" after a t-test usually means the interval for the mean difference, in the outcome's units. If it is for the effect size instead, put it after d. When in doubt, say so: "95% CI of the difference [1.70, 10.90]".
Mistakes reviewers catch
- Round brackets or a dash: (1.70–10.90) is not APA style.
- No confidence level.
- Leading zeros inconsistent with the statistic: r = .48, 95% CI [0.27, 0.64].
- Interpreting "95% probability the true value is inside". Say instead that the interval gives the range of values compatible with the data.
Let KyroStat do the write-up for you
KyroStat reports a confidence interval with every estimate and effect size, formatted by these rules.
Frequently asked questions
Does an interval that excludes zero mean the result is significant? For a mean difference or a regression coefficient, a 95% CI that excludes 0 corresponds to p < .05. For an odds ratio, the reference value is 1.
Do I need a CI for every test? APA 7 recommends CIs for point estimates and effect sizes wherever possible. See how to report effect sizes.
What about CIs in tables? Give the limits in one column ("95% CI") as [LL, UL], or in two columns headed LL and UL.