SPSS Lab: Repeated-Measures t-test

Exercise gets blood flowing to your brain, so with that in mind, you want to test whether exercise influences people’s scores on a statistics quiz. (Notice that even though you probably think exercise will increase blood to the brain and increase performance, it is actually stated as a two-tailed test: it says “influences”, not “increases”). To test your hypothesis, you get 10 statistics students. You give them a quiz before exercise and after exercise. (Because you want to avoid order effects, you use counterbalancing). The quiz you give them is worth 10 points, so these scores are out of 10 points total.

Use a two-tailed, alpha = .05 hypothesis test to see whether exercise affects quiz score.

Here are the data – PLEASE CALCULATE M1, M2, a column of D’s, and MD BY HAND:

Participant / Score on quiz before exercise / Score on quizafter exercise / Difference Scores (D)
1 / 5 / 7
2 / 9 / 10
3 / 10 / 8
4 / 8 / 10
5 / 5 / 10
6 / 8 / 9
7 / 4 / 6
8 / 10 / 10
9 / 7 / 10
10 / 6 / 6
M1= / M2= / MD=

SPSS instructions:

  1. Open IBM SPSS Statistics 21 from the desktop, or from the Start Menu if it doesn’t appear on your computer’s desktop.
  2. If it says something about “Unicode coding”, just click “Use Unicode coding”.
  3. In the first window that pops up, choose “Type in Data” (on the right hand side), then click “OK”.
  4. Under “Variable View” tab, create two variables: “NO_EXCERCISE” and “EXCERCISE”
  5. Under “Data View” tab, add the data. For a repeated-measures design, you pretty much just add the data exactly as they appear on this page (above). Before Exercise scores go in the NO_EXCERCISE column, and After Exercise scores go in the EXERCISE column.
  6. From the menu bar select: Analyze > Compare Means > Paired-Samples T test
  7. On the left it should show your two variables. Click “NO_EXCERCISE”. Hold either the Shift button or the Ctrl button down as you click the other variable, “EXCERCISE”. BOTH variables should now be highlighted.
  8. With both variables highlighted, click the right arrow to move these variables into the “Paired Variables” box. SPSS will now use these two variables to calculate the difference scores that are needed to perform a Repeated-Measures t-test.
  9. Click the “Options” button and make sure that SPSS is set to give you a 95% confidence interval. It probably already will be. If not, change it to 95%.
  10. Click “OK”.
  11. Find the Output window that pops up and use it to answer the following questions.

ANSWER QUESTIONS #1-10 (NUMBERED) ON LINED PAPER

Answer these from your SPSS output:

1. What is the n given for the Exercise scores? What is the n of the No Exercise scores? (both these numbers come from the “Paired Samples Statistics” box.)

2. What is the n given for the number of difference scores? (This comes from the “Paired Samples Correlations” box).

The remaining questions come from the “Paired Samples Test” box in the output:

3. What is the mean difference? (In other words, the mean of the difference scores)? On SPSS it is just called “Mean”.

4. What is the t-obtained? On SPSS, it is just called “t”.

5. What are the degrees of freedom?

6. What is the p-value? (Given as “Sig. (2-tailed)”).

7. Comparing the p-value with our alpha of .05, what decision do we make about the null?

8. What are the lower and upper bounds of the 95% confidence interval estimating μD?

9. Compare the confidence interval with zero. Is zero in the interval? Based on your answer to that, what decision do we make about the null?

10. Did your decisions about the null match whether you made them using the p-value or made them using the confidence interval?