1
Independent-Sample t-test (Part 2) (DR SEE KIN HAI)
(Comparing 2 groups of unrelated/uncorrelated scores) 1. It is used to determine whether there is a significant difference in the mean scores of 2 independent samples. 2. Fulfill the assumptions in Part 1 Practice 1 You have administered a test to a group of 60 Year 11 students of which 30 are boys and 30 are girls. You wanted to find out whether there is a significant difference in the mean scores between the boys and the girls. In order to save time you have randomly selected 20 boys and 20 girls from this group of 60 students. The scores are shown below.
1. Enter the data into SPSS 20 using Gender ( 1= Male, 2= Female) and Scores as below.
2. Select [Analyze] on the menu bar then click on [Compare Means] then [Independent-samples t test] to open the dialogue box 3. Select [Scores] then move to [Test Variables] box and [Gender] then move to [Grouping Variable] box.
4. Click on [Define Groups] to open the sub-dialogue box. Type [1] in [Group 1] box then type [2] in [Group 2] box. Click [Continue], then [OK]
3 Interpreting the output
This table shows, the total no of students for each gender, the mean, SD and Std Error. The mean for Male = 72.95 which is higher than the Female (mean = 65.25)
T = mean Diff/ Std Error Diff = 7.7/2.41 = 3.195
If p = 0.922 which is > 0.05 for Levenes test , use the information on the first row. If p < 0.05 where thee variances are significantly different, then use the information on the second row. In this case use the first row.
For equal variance, t = 3.195, DF = 38 is significant at p = 0.003 for two-tailed level. For one-tailed level p = 0.003/2 = 0.0015 = 0.002 (3 dp)
Reporting the results 1. Levenes test for equality of variance shows a probability of p = 0.922. This value is greater than 0.05. Therefore there is no significant difference in the variances of the test scores between the boys and the girls. Therefore the data came from a homogenous population sample.
4 2. As the significant level > 0.05 use the first row where t = 3.195 and the variances for boys and girls are equal because p > 0.05 is not statistically significant. The data fulfilled the assumption that the data is drawn from a normally distributed population.
3. Conclusion and Report: The result shows that there is a significant difference in the mean scores between the boys and the girls at p < 0.025 (for 2-tailed test) (Here p = 0.003). The boys (mean = 72.95) perform significantly better than the girls (mean = 65.25) in the test. COURSEWORK A pretest was administered to 22 Year 10 students from SOAS prior to a course in History on certain subject, and then again after the course was complete as posttest. The students are classified as 1=male, 2=female. You are interested to answer the following research question. Does the understanding of History on certain subject before and after the course differ amongst the boys and girls? 1=Male, 2=Female Tips: Move [Pretest] and [Posttest] into [Test Variables] box and [Sex] into [Grouping Variable]