There is a certain way of reporting statistics to make it standardised and accessible for everyone. Scientific reporting is everyone's responsibility!
When it comes to writing statistics, it is common to use the APA format, regardless of whichever reporting standard you have used elsewhere in your paper (it is okay to use another style for your writing and APA for your statistics). For more information about different formatting styles, please see our Citing and Referencing page.
When writing statistical symbols which are not Greek letters, they should be italicised, e.g. p value, z score, SD, etc.
The tabs of this guide will support you in reporting the results of your statistics. The sections are organised as follows:
When reporting p values, you also need to include your level of significance (your α value), so the reader can have an understanding of the threshold at which you reject or fail to reject the null hypothesis. Your p values need to be reported as your software has calculated it as, either to two or to three significant figures, for example:
p = .25
p = .981
p < .001
There may be some instances where your software has not calculated the p value directly, and instead only indicated if the value is less than .1, .05 or .01. In this case, you can simply report that your results were/were not significant to the appropriate significant level.
Note that, since p can only be a value between 0 and 1, there is no need to write a 0 before the decimal point.
Recall that non-Greek mathematical symbols should be italicised, e.g. p value, z score, SD, etc.
To report the results of a correlation test such as Pearson's r or Spearman's Rank, you need the test statistic (r for Pearson, ρ for Spearman) along with its interpretation, the degrees of freedom and the p value. For example:
We observed a significant moderate negative correlation between the variables, r(82) = -.55, p = .002.
To report the results of a Chi-Squared test of Independence, you require the test statistic χ², degrees of freedom, sample size, p value and the interpretation of your results. For example:
The relation between the variables was not significant, χ²(2, N = 205) = 23.81, p = .341
Reporting the results of a Chi-Squared Test of Goodness of Fit is the same as for Independence: you need the test statistic χ, degrees of freedom, sample size, p value and the interpretation of your results.
By the results of the Chi-Squared Goodness of Fit test, we can see that the groups were equally significant, χ²(2, N = 90) = 23.81, p < .001
Recall that non-Greek mathematical symbols should be italicised, e.g. p value, z score, SD, etc.
When reporting a one-sample t-test, you should report the mean and standard deviation of your sample as well as your compared results, as well as the test statistic t, degrees of freedom, p value of the test and the interpretation of your results. For example:
"Our variable (M = 31.09, SD = 1.48) differed significantly from the compared results (M=45.22, SD = 2.33), t(70) = 1.88, p = .012."
or, perhaps:
"According to the t-test, our variable (31.09 ± 1.48) was significantly different to the standard results (45.22 ± 2.33), t(70) = 1.88, p = .012."
It is a good idea not only to report the differences in mean and standard deviation between your groups but also indicate which group was greater/higher/more than the other. This can be part of your results interpretation.
For a paired t-test, you need the means of both groups as well as their respective standard deviations, along with the test statistic t, degrees of freedom, p value and the interpretation of your results. For example:
The mean of the pre- measurement was 139.89 (SD = 3.91) and of the post- measurement was 132.40 (SD = 5.99). Based on the results of the paired t-test, t(210) = 4.09, p = .503, there is not enough evidence to say that the variable had an effect on our sample.
Remember to not only to report the differences in mean and standard deviation between your groups but also indicate which group was greater/higher/more than the other, as part of your results interpretation.
For an independent samples t-test, you need the means of both of your samples as well as their respective standard deviations, along with the test statistic t, degrees of freedom, p value and the interpretation of your results. For example:
"According to the independent samples t-test, t(55) = 1.52, p = .073, the variables (M = 59.0, SD = 22.1 and M=48.2, SD = 20.38) did not differ significantly ."
or, perhaps:
"Our variables (59.0 ± 22.1) and (48.2 ± 20.38) showed no significant difference , t(55) = 1.52, p = .073."
Remember to not only to report the differences in mean and standard deviation between your groups but also indicate which group was greater/higher/more than the other, as part of your results interpretation.
To report an ANOVA you require your test statistic F, degrees of freedom (both between groups and within groups) and p value, and of course your interpretation of results. For example:
"The one-way ANOVA showed that the effect of our variable was significant, F(2, 109) = 5.91, p = .021."
When you have a significant ANOVA result, you should follow it up with a post-hoc test to show you where the difference in the groups actually lie. You should name the post-hoc test alongside your results. For example:
"The Tukey HSD test showed that the mean value of the dependent variable was significantly different between Group A and Group B (the mean difference being 10.62, p = .014, 95% CI = (8.19, 13.05)). No significant difference was found between any other group."
In doing this, you are more successfully answering your question in indicating where the differences in the groups exist, and you can successfully determine which of your groups were statistically different than the other(s) if a significant value was found. Remember to report not only the means and standard deviations of your groups but also indicate which are greater/higher/more than others.
Recall that non-Greek mathematical symbols should be italicised, e.g. p value, z score, SD, etc.
Regressions are a type of F test. What you need to report them include the R² value, the test statistic F, the degrees of freedom (for regression and for residual) and the p value, as well as the interpretation of what your results mean. It is also a good idea to report the regression equation. For example:
"Overall, according to the simple linear regression test performed, X had a significant effect on Y, R² = .304, F(1, 37) = 1.01, p = .033. The R² value showed that X explained 30.4% of the variance of the dependent variable, and the regression equation was:
Y = -3.56 + 14.05*X
which indicated that for each incremental increase in X, Y was predicted to increase by 14.05 [units]"
When you have a lot of numerical data to communicate, it is often acceptable to report them in a table. If you are using Microsoft Word to write your paper, build a table in APA format directly into Word using the 'tables' feature. You can have a look at what tables should look like in the APA Guidelines site.
To report the results of many statistical tests, you should report:
Make sure to use the same font in your table as you do in the rest of your paper, including any notes you wish to include underneath the table. Always align tables flush against the left margin of the page.
It is generally unadvised to use colour for printing and publishing your projects, due to the high cost involved in colour printing, so if you have used colour, transform it into greyscale to check it will still be legible when printed in black and white.
However, for classroom projects and coursework, it is more acceptable to use colour as long as the project delivery supports it. Make sure you use colours with high contrast ratios to make your work accessible, not only to those who are colour-blind or who do not see colour in a typical way, but also is understandable if your work is printed in greyscale. You can transform your work to greyscale or use a contrast checker to make sure this is adequate.
