In this section we will discuss how to present the results for an independent t test and a dependent (or paired) t test. They are very similar.

First, notice that so far I have used a lower case t. When report these tests you use "t" and NOT "T." Many word processors will try to capitalize the letter which can be frustrating to deal with. A "T" can mean something very different in math and statistics. The convention with t tests is "t."

Below is the standard form you use when there is a statistically significant finding.

t(df) = #.##, p < .##

And this is the standard form when there is NOT a statistically significant finding.

t(df) = #.##, p > .##

The only difference is the greater than or less than symbol. Many people find this counter intuitive, but you use a less than symbol (<) when you have a statistically significant result. You use a greater than symbol (>) when do not have a statistically significant result. The reason for this goes beyond my intentions here and requires an understanding of the logic behind statistical significance testing.

So what does all of this mean?

The df is the degrees of freedom for the test. With an independent samples t test is the total number of cases minus two. For a paired samples t test the degrees of freedom is the number of pairs minus one.

The red #.## is the obtained t-value. Only report this to two decimal places. If the obtain t-value is a decimal (0.53) remember to use a leading zero because t values can be greater than one.

The .## is the obtain p-value. Only report this to two decimal places. And you do not use a leading zero because p-values range between 1.00 and -1.00. Generally, you will only use one of two numbers here. Either .05 or .01. There really is only three options that should be reported here. Although you will occasionally see other numbers reported here it is best to stick to one of the three shown below until you are more proficient in statistics.

p < .05
This is used when the obtained p-value is less than .05, but equal to or greater than .01. This is a statistically significant finding.

p < .01
This is used when the obtained p-value is less than .01. This is a statistically significant finding.

p > .05
This is used when the obtained p-value is equal greater than .05. This is a non-significant finding.

So you may be asking what happens if the obtained p-value is exactly .05 or .01? Actually, this is very unlikely to happen. If your output shows .05 it is probably more likely something like .05000000001 or .049999999999 which has been rounded to fit nicely in your output table. Depending on your statistical program you can probably get more decimal places and report it appropriately. In the unlikely event that you have exactly .05 or .01 (and your software’s calculations are that accurate) you can use ≤ to be correct.

Example 1 Independent t Test

It is well-known that nursing can be a greatly rewarding job, but it can also be extremely stressful. As a result, there has been a lot of research on job burnout among nurses. For this example, let’s suppose that you gave a measure of job burnout to two different nurse specialties, hospice nurses and anesthetists nurses.

The burnout measure consisted of 16 statements that were rated on a 7-point scale with higher numbers indicating greater burnout. You computed the mean rating across the 16 statements to determine each nurse’s overall level of burnout. Using your favorite statistical program, you performed an independent samples t test to determine if the two types of nurses differed in their level of job burnout. Below are the results. I drew red boxes around the information that we will use.

 APAReportingStatsImage4.1

Below are a few ways that I would communicate the results.

A survey measuring job burnout was administered to two groups of specialty nurses. In total 29 nurses were surveyed consisting of 17 hospice nurses and 12 anesthetist nurses. The burnout survey contained 16 items measured on a 7-point scale with higher numbers indicating greater levels of burnout. The mean across the 16 items was computed to determine each nurse’s level of burnout. The results showed that hospice nurses reported greater levels of burnout (M = 5.90, SD = 0.69) than anesthetist nurses (M = 5.67, SD = 0.75). An independent samples t test should that this difference was not statistically significant, t(27) = 0.84, p > .05.

Even though we are discussing how to report a t test, this cannot be done in isolation. You can’t just report t(df) = xx.xx, p < > .xx. This alone doesn’t tell you much of anything useful. You must include enough information to give the reader a complete picture of the analysis, thereby allowing him or her to understand what the results may or may not mean. This requires giving some background information.

In this example we reported some basic information about the survey, the means, and standard deviations as well as the t test results. Most often the survey information will be reported in the method section of the report and the statistical results in the results section of the report. However, I included the information here so any reasonably knowledge reader could understand what was done.

Of course this is only one way to state the results. Below are a few other ways that assume we communicated the information about the survey earlier in the write up.

The results of an independent samples t test showed that hospice nurses tend to have higher levels of burnout (M = 5.90, SD = 0.69) than anesthetist nurses (M = 5.67, SD = 0.75), t(27) = 0.84, p > .05.

I found a statistically significant difference (t(27) = 0.84, p > .05) between the mean level of burnout reported by hospice nurses (M = 5.90, SD = 0.69) and anesthetist nurses (M = 5.67, SD = 0.75).

Note that APA style is okay with the use of personal pronouns (I and we). Just make sure if you use the right one. Don’t use we if you are the only one doing the research.

Hospice nurses had a mean burnout score of 5.90 (SD = 0.69) and anesthetist nurses had a mean burnout score of 5.67 (SD = 0.75). An independent samples t test indicated that this difference was statistically significant, t(27) = 0.84, p > .05.

Example 2 Dependent t Test

For our next example suppose we wanted to determine if self-esteem changes over time. We do some background research and find a 10-item measure of self-esteem that uses a 4-point scale. Overall self-esteem was taken as the mean of the 10 items with higher numbers indicating greater self-esteem.

We then gave it to 30 undergraduates who were between 18 and 20 years old. Here is the hard part. We wait 10 long years and contacted them again. Luckily all 30 agree to take the same self-esteem measure. Now we can look to see if their self-esteem has changed.  We enter the data into a statistical package and run a paired sample t test.  The results are shown below.

Paired Sample Statistics

  MeanNStd. DeviationS.E. Mean
Pair 1 Self-Esteem Tme 2 3.47 30 .43 .08
  Self-Esteem Time 1 3.19 30 .51 .09
Paired Samples Correlations
  NCorrelationSig.
Pair 1 Self-Esteem Tme 2 & Self-Esteem Time 1 30 -.19 .312
Paired Samples Test
  Paired Differences   
     95% Confidence Interval of the Difference   
  MeanStd. DeviationStd. Error MeanLowerUppertdfSig. (2-tailed)
Pair 1 Self-Esteem Tme 2 - Self-Esteem Time 1 .28 .73 .13 .01 .55 2.09 29 .046

The results of a paired sample t test showed that self-esteem significantly increased over the 10 year span, t(29) = 2.09, p < .05. The mean level of self-esteem at Time 1 when participants were between 18 and 20 years old was 3.19 (SD = 0.51). Ten years later at Time 2 the mean level of self-esteem was 3.47 (SD = 0.43).

At Time 1 participants were between 18 and 20 years old and the mean self-esteem was 3.19 (SD = 0.51). At Time 2, ten years later, the mean level of self-esteem was 3.47 (SD = 0.43). A dependent t test showed that this was a statistically significant increase in self-esteem, t(29) = 2.09, p < .05.

Test Yourself


1. How many formatting mistakes are in the written results below?

An independent sample t test showed a statistically significant difference in means, t(27)=2.40, p > .05.

One
Two
Five

 

2. What is the formatting mistake for this result?

t = 1.10, p > .05

The t should be in italic font.
There should be a zero before .05.