Reporting Statistics In-Text
When reporting statistics
in a Results section, you should include effect sizes and 95% confidence intervals.
Make sure to describe to readers what specific statistical models you used,
and report parameter estimates with 95% confidence intervals, e.g.:
We regressed students' love for Peruna onto their age.
Results indicated that older students loved Peruna less than did younger students (b = -0.03, 95% CI [-0.05, -0.01]).
Descriptive Statistics
Results sections oftentimes begin with a brief statement describing the means and standard deviations
observed in the sample. Oftentimes, this information is also included in a table.
Descriptive statistics and correlations for all study variables are included in Table 1.
As seen in Table 1, students generally felt very favorably about Peruna (M = 4.82, SD = 0.51). Men, however, loved Peruna less than did women (r = -.10, 95% CI [-.17, .03]).
Repeat Your Hypotheses
After reporting descriptive statistics, remind readers of your hypotheses, e.g.:
We predicted that older students would love Peruna less than would younger students.
Report Statistics and Briefly Interpret Them
Report your statistical tests and very briefly interpret them. For example:
We regressed students' love for Peruna onto age. Results revealed that there was no statistically significant
association between age and love for Peruna (β = 0.03, 95% CI [-0.09, 0.15]). Thus, it seems
that there is no difference between older and younger students in the extent to which they love Peruna.
Formatting Numbers
It is customary to round all numbers to two decimal places (e.g.,
M = 3.26 is correct, whereas
M = 3.2566 is not).
It is sometimes appropriate to round numbers to three decimal places (e.g., if your effect sizes are very small such as
b = 0.003).
For numbers that can range beyond
+1, always report numbers with leading zeros.
For example:
On a scale from 1 to 5, the average love for Peruna in our sample was 3.35, (SD = 0.86).
Notice that the reported standard deviation (
SD) is reported as 0.86, and
not as .86.
For numbers that are bounded at
+1 (such as correlations), do
not include leading zeros. For example:
Love for Peruna was negatively correlated with age (r = -.14, 95% CI [-.16, -.12]).
Common Statistics
Means are reported as
M (e.g.,
M = 3.26). Whenever you report a mean, always report the standard deviation, as well.
Standard deviations are reported as
SD (e.g.,
SD = 0.54).
Correlations range from -1 to +1 and indicate the extent to which two variables "go together." Correlations are reported as
r (e.g.,
r = .36, 95% CI [.26, .46]).
Regression coefficients are not bounded at +/-1 and are reported as a
b (e.g.,
b = 0.25, 95% CI [0.15, 0.35]).
Standardized regression coefficients are reported as
β (e.g.,
β = 0.14, 95% CI [0.10, 0.18]).
Standardized group differences are deported as
d (e.g.,
d = 0.36, 95% CI [0.30, 0.42]).
