In: Statistics and Probability
A weight-lifting coach wanted to know whether weight-lifters can change their strength by taking a certain supplement. To answer this question, the coach randomly selected 8 athletes and gave them a strength test using a bench press. Thirty days later, after regular training using the supplement, they were tested again. The results were listed below. A test was conducted to determine whether weight-lifters can change their strength by taking a certain supplement. Assume the populations are normally distributed.
Athlete 1 2 3 4 5 6 7 8 Mean SD
Before 215 240 188 212 275 260 225 200 226.88 29.72
After 225 245 188 210 282 275 230 195 231.25 34.55
Difference -10 -5 0 2 -7 -15 -5 5 -4.38 6.55
1. What is the null hypothesis for this test?
2. What is the value of the standardized test statistic?
3. The p-value was 0.05. Suppose this test was conducted at α = 0.01. What can you conclude?
a. There is insufficient evidence to conclude that weight-lifters can change their strength by taking a certain supplement.
b. There is sufficient evidence to conclude that weight-lifters can change their strength by taking a certain supplement.
c. None of the above; this test is invalid.
4. Suppose the coach had incorrectly performed a two-sample t-test. In comparison to the correct analysis, which of the following regarding the test statistic would be true?
a. The test statistic is closer to 0
b. The test statistic is farther from 0
c. The test statistic does not change.
Solution:
Here, we have to use paired t test.
The null and alternative hypotheses for this test are given as below:
1. What is the null hypothesis for this test?
Null hypothesis: H0: The weight-lifters cannot change their strength by taking a certain supplement.
Alternative hypothesis: Ha: The weight-lifters can change their strength by taking a certain supplement.
H0: µd = 0 versus Ha: µd ≠ 0
This is a two tailed test.
2. What is the value of the standardized test statistic?
Test statistic for paired t test is given as below:
t = (Dbar - µd)/[Sd/sqrt(n)]
From given data, we have
Dbar = -4.38
Sd = 6.55
n = 8
df = n – 1 = 7
α = 0.01
t = (Dbar - µd)/[Sd/sqrt(n)]
t = (-4.38 – 0)/[6.55/sqrt(8)]
t = -1.8914
3. The p-value was 0.05. Suppose this test was conducted at α = 0.01. What can you conclude?
P-value = 0.05 > α = 0.01, so we do not reject the null hypothesis
a. There is insufficient evidence to conclude that weight-lifters can change their strength by taking a certain supplement.
4. Suppose the coach had incorrectly performed a two-sample t-test. In comparison to the correct analysis, which of the following regarding the test statistic would be true?
a. The test statistic is closer to 0
...because difference between two sample means will be close to zero