Answer Key on t-tests and F-tests
Instructions: For each of the following already calculated t-values and F-values use the information provided about the sample size to determine whether your should reject or accept the null hypothesis at a p < .05 significance level. Assume a non-directional hypothesis is being tested in each case.
| a. | t = +2.00 | N1 = 11, N2 = 11 | Accept or Reject null? |
| b. | t = -2.10 | N1 = 13, N2 = 14 | Accept or Reject null? |
| c. | F = 3.90 | N1 = 5, N2 = 5, N3 = 5 | Accept or Reject null? |
| d. | F = 4.00 | N1 = 3, N2 = 3, N3 = 3, N4 = 3 | Accept or Reject null? |
Instructions: Conduct a F-test for the following problem.
An employer hypothesizes that factory workers will differ in their productivity according to how much positive reinforcement they receive from their supervisors. So, the employer randomly assigns five employees to receive 5 minutes of positive reinforcement per day from their immediate supervisors. Five other employees are randomly assigned to receive 10 minutes of praise per day, and five more employees receive 15 minutes of praise per day. One month later the employer measures the employees’ productivity rates. The employer than conducts an F-test to determine if receiving different amounts of positive reinforcement matters. The F-value obtained is 5.00.
Does providing different amounts of reinforcement matter? And if so, how confident are you?
dfb = 2
dfw = 12Providing different amounts of reinforcement does make a difference (95% confident).
Instructions: You notice that studying with other people seems to improve you performance on exams. So, you decide to conduct a quick experiment to test your hypothesis. You manage to get 10 of your classmates to study alone and 10 of your classmates to study in groups. The average exam scores for those who studied alone was 80. On the other hand, the average exam score for those who studied in groups was 100.
Is this difference due to chance or due to the manipulation of the independent variable? Conduct a t-test to find out. (s diff = 10)
t = (100-80)/10 = 2
df = 18reject null hypothesis, Study groups do make a difference.