In: Statistics and Probability
We are given the two tests were similar. So, it means we have to use a matched pair test.
1. Consider (Test A - Test B). Use a 0.01
(a) What test method should be used?
Matched Pairs is the correct test method because the data is similar in nature.
Load the data into Excel.
Go to Data>Megastat.
Select the option Hypothesis tests and go to Paired Observations.
Select the Group 1 and Group 2 as Test A and Test B data set respectively.
Click OK.
The output obtained will be as follows:
0.000 | hypothesized value | |
100.200 | mean Test A | |
102.100 | mean Test B | |
-1.900 | mean difference (Test A - Test B) | |
2.558 | std. dev. | |
0.809 | std. error | |
10 | n | |
9 | df | |
-2.349 | t | |
.0217 | p-value (one-tailed, lower) | |
-2.8214 | t Critical one-tail | |
-4.529 | confidence interval 99.% lower | |
0.729 | confidence interval 99.% upper | |
2.629 | margin of error |
The hypothesis for testing is:
Paired T hypothesis test:
μD = μ1 - μ2: Mean of the
difference between Test A and Test B
H0 : μD = 0
People do good on both tests.
HA : μD < 0
People do better on the second test.
(b) The test statistic from the output is -2.349.
(c) The critical value from the output is -2.8214.
(d) Since |tstatistic| < tcritical, we fail to reject the null hypothesis.
So, there is not sufficient evidence to support the claim that people do better on the second test.
Therefore, the correct answer is No.
2. Construct a 99% confidence interval.
The 99% confidence interval from the output is between -4.529 and 0.729.