Statistics – understand independent t-test
In this example I will make a t-test to test if males and females mean values of salary differs significantly from each other. To do a t-test the salary should be normal distributed for male resp female. Let say that it is pretty normal, have a look below on the 2 histograms of males and females salaries.
So in this case I will call it rather normal distributed, and will do a t-test to compare male and female.
The t-test you find under the Analyze menu:
Here is how you do the command:
As you see you have to define the codes for the grouping variable, here 1 and 2. (1 for male and 2 for female).
Here is the result:
The difference is not so big, male has 19 365 in salary and female has 18 944 in mean salary.
- Before you check the results of the t-test you must know if the variances are different in the 2 Groups, that’s why there is Levene’s test fo the left. As the significance value is 0.586 (above 0.05) we can see that the variances are not different from each other. Then you check the result on the first row (equal variances assumed)
- So have a look in the first row’s t-test result and you can see that the significance value for t-test is 0.128, so that means no significant differences between males and females salaries.
- The difference between the 2 mean values is 421 (19 365-18 944).
Thanks for watching, and good luck with the statistics!