10.4 Common APA-Style Errors
This section focuses on common errors students make when writing statistical results. Some errors are formatting issues. Others are interpretation issues. Both matter.
The following video details seven common errors that people make when reporting statistical results:
If you are reading the Word document version of this textbook, you can watch the video here.
Formatting Errors
Error 1: Forgetting Italics
In APA style, many statistical letters are italicized, including M, SD, p, t, F, r, d, and b.
Greek letters are not italicized. For example, α, β, χ², η², and ω² are not italicized.
| Incorrect | Correct |
|---|---|
| M = 10.25, SD = 2.11 | M = 10.25, SD = 2.11 |
| p = .032 | p = .032 |
| t(31) = 2.12 | t(31) = 2.12 |
Error 2: Using the Wrong Number of Decimal Places
A common guideline is to round most statistics to two decimal places and p-values to three decimal places.
| Statistic | Common reporting format |
|---|---|
| Mean | M = 10.25 |
| Standard deviation | SD = 2.11 |
| t statistic | t(31) = 2.12 |
| F statistic | F(2, 87) = 5.64 |
| p-value | p = .043 |
There may be exceptions depending on the statistic, journal, or assignment. When in doubt, check the APA Style website, the APA Manual, or your assignment instructions.
Error 3: Using Leading Zeros Incorrectly
Use a leading zero when the statistic can be greater than 1. Do not use a leading zero when the statistic cannot be greater than 1.
| Incorrect | Correct | Why |
|---|---|---|
| p = 0.043 | p = .043 | p-values cannot be greater than 1 |
| r = 0.42 | r = .42 | correlations cannot be greater than 1 |
| M = .75 | M = 0.75 | means can be greater than 1 |
| t(31) = .82 | t(31) = 0.82 | t values can be greater than 1 |
Error 4: Reporting p = .000 or p <= .001
Never report p = .000. A p-value may be very small, but it is not exactly zero.
| Incorrect | Correct |
|---|---|
| p = .000 | p < .001 |
| p <= .001 | p < .001 |
Error 5: Putting Parentheses Inside Parentheses
Do not place parentheses inside parentheses. If you need to put statistical information inside a parenthetical statement, use brackets for the inner parentheses.
| Incorrect | Correct |
|---|---|
| The result was statistically significant (t(31) = 2.12, p = .043). | The result was statistically significant (t[31] = 2.12, p = .043). |
You do not need brackets if the statistic is not already inside parentheses:
The result was statistically significant, t(31) = 2.12, p = .043.
Error 6: Mixing Up N and n
Use N for the full sample size and n for a subgroup or subsample.
| Example | Better reporting |
|---|---|
| This study examined differences in Anastasia’s class (N = 15) and Bernadette’s class (N = 15). | This study examined differences in Anastasia’s class (n = 15) and Bernadette’s class (n = 15). |
| This study examined differences in test performance in a sample of college students (n = 33). | This study examined differences in test performance in a sample of college students (N = 33). |
One way you can avoid this issue is by writing out the sample size in the text outside of parentheses. For example:
This study examined differences in test performance in 33 college students, 15 students in Anastasia’s class and 18 students in Bernadette’s class.
Error 7: Adding Spaces in the Wrong Places
There is no space between the statistical test symbol and the degrees of freedom.
| Incorrect | Correct |
|---|---|
| t (31) = 2.12 | t(31) = 2.12 |
| F (2, 87) = 5.64 | F(2, 87) = 5.64 |
| r (58) = .42 | r(58) = .42 |
For an F statistic, there is a comma and a space between the two degrees of freedom:
F(2, 87) = 5.64
Interpretation Errors
Formatting matters, but interpretation matters even more. Watch for these common interpretation problems.
Error 8: Saying the Hypothesis Was Proven
Statistical tests do not prove hypotheses.
| Weak wording | Stronger wording |
|---|---|
| The hypothesis was proven. | The results supported the hypothesis. |
| This proves that studying causes higher grades. | Students who studied more tended to have higher grades. |
Use especially careful language for correlational results. Correlation does not establish causation.
Error 9: Saying You Accepted the Null Hypothesis
In null hypothesis significance testing, we usually say we reject or fail to reject the null hypothesis.
| Weak wording | Stronger wording |
|---|---|
| We accepted the null hypothesis. | We failed to reject the null hypothesis. |
| The null hypothesis was true. | The result did not provide sufficient evidence to reject the null hypothesis. |
Error 10: Saying There Was No Difference When the Result Was Non-Significant
A non-significant result does not prove there is no difference. It means the test did not find sufficient evidence to reject the null hypothesis.
| Weak wording | Stronger wording |
|---|---|
| There was no difference between groups. | The difference between groups was not statistically significant. |
| The variables were unrelated. | The association between the variables was not statistically significant. |
Error 11: Reporting Significance Without Direction
Do not just say that a result was significant. Tell the reader what happened.
| Incomplete | Stronger |
|---|---|
| There was a significant difference between groups. | Anastasia’s students had significantly higher grades than Bernadette’s students. |
| There was a significant correlation. | Study time was positively correlated with exam scores. |
Error 12: Reporting Inferential Results Without Descriptives
The inferential test tells the reader whether the result was statistically significant. Descriptive statistics help the reader understand the pattern.
| Incomplete | Stronger |
|---|---|
| The groups differed significantly, t(31) = 2.12, p = .043. | Anastasia’s students (M = 74.53, SD = 9.00) had significantly higher grades than Bernadette’s students (M = 69.06, SD = 5.77), t(31) = 2.12, p = .043. |
Check Your Understanding
TipCheck Your Understanding
- How should you report a p-value that jamovi displays as .000?
- What is the difference between N and n?
- Why should you avoid saying that a non-significant result means there was “no difference”?
- What is wrong with the phrase “the p-value was significant”?
NoteAnswers
- Report it as p < .001.
- N refers to the full sample size; n refers to a subgroup or subsample.
- A non-significant result means there was not enough evidence to reject the null hypothesis. It does not prove that the groups are identical or that no effect exists.
- It is usually clearer to say that the result, test, association, or difference was statistically significant. The p-value is the evidence used to make that decision.