You are on page 1of 8

CASE STUDYNO.

(TWO-SAMPLE TESTS: SEPARATE-VARIANCES T TEST FOR THE

DIFFERENCE BETWEEN TWO MEANS)

The results of Differential Calculus – Midterm Examination for random samples of

students taught by two self-proclaimed genius Instructors at CTU-DC are shown

below. Can you conclude that there is a difference in the mean Midterm Examination

scores for the students of the two instructors? Use 𝛼 = 0.01.

Note: Data were collected from Engineering Students at the Cebu Technological University-Danao Campus.
SOLUTION:

Step 1: State the null and alternative hypothesis.

The null hypothesis will be that the two methods are equally effective, we have

𝐻𝑜 : 𝜇𝐴 = 𝜇𝐵 . Since B is an experimental method a natural alternative to consider is

whether it improves on the standard method, A, so we choose the one-sided

alternative hypothesis 𝐻1 : 𝜇𝐴 = 𝜇𝐵 .

𝐻𝑜 : 𝜇1 = 𝜇2

𝐻𝑜 : there is no difference in the mean Midterm Examination scores for the students of

the two instructors

𝐻1 : 𝜇1 ≠ 𝜇2 (claim)

𝐻1 : there has difference in the mean Midterm Examination scores for the students of

the two instructors

Step 2: Choose the level of significance and the sample size.

𝛼 = 0.01

𝑛1 = 𝑃𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 1: 𝑆𝑎𝑚𝑝𝑙𝑒 𝑆𝑖𝑧𝑒 = 32

𝑛2 = 𝑃𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 2: 𝑆𝑎𝑚𝑝𝑙𝑒 𝑆𝑖𝑧𝑒 = 34

Step 3: Select the appropriate test statistic.

Because 𝜎1 and 𝜎2 is unknown, the samples are random and independent, and the

populations are normally distributed. We wish to determine if there is any difference in

mean Midterm Examination scores between the two instructors at CTU-DC. Figure3.1
shows a normal probability plot for the two samples of Differential Calculus–Midterm

Examination scores. The assumption of normality appears quite reasonable, but since

the slopes of the two straight lines are very different, itis unlikely that the population

variances are the same. So, you can use the separate-variance t test developed by

Satterthwaite.

SEPARATE-VARIANCE t TEST FOR THE DIFFERENCE BETWEEN TWO MEANS

̅̅̅1 − 𝑋
(𝑋 ̅̅̅2 ) − (𝜇2 − 𝜇1 )
𝑡𝑆𝑇𝐴𝑇 =
2 2
√𝑆1 + 𝑆2
𝑛1 𝑛2

COMPUTING DEGREES OF FREEDOM IN THE SEPARATE-VARIANCE t TEST

2
𝑆2 𝑆 2
( 𝑛1 + 𝑛2 )
1 2
𝑉= 2 2
𝑆2 𝑆 2
( 𝑛1 ) ( 𝑛2 )
1 2
+
𝑛1 − 1 𝑛2 − 1

Figure 3.1 Shows a normal probability plot for the two samples of Differential Calculus – Midterm
Examination scores.
Step 4: Determine the rejection region.

Using 𝛼 = 0.01, the upper and lower critical values for this two-tail test found in Table

E.3 are +1.6694 and -1.6694. Solving the critical values using t table:

Source: Extracted from Table E.3 of Basic Business Statistics: Concepts and Applications, 12th Ed.
by Berenson, Levine, and Krehbiel.

Alternate solution in solving the critical values using MS Excel:

2
𝑆2 𝑆 2
( 𝑛1 + 𝑛2 )
1 2
𝑉= 2 2
𝑆2 𝑆 2
( 𝑛1 ) ( 𝑛2 )
1 2
𝑛1 − 1 + 𝑛2 − 1

2
(12.0432)2 (13.1978)2
( + )
32 34
𝑉= 2 2
(12.0432)2 (13.1978)2
( ) ( )
32 34
+
32 − 1 34 − 1

Degrees of Freedom, 𝑉 ≈ 63

Lower Critical Value: =-TINV (0.1,63)


Results: ≈ −1.6694

Higher Critical Value: =TINV (0.1,63)

Results: ≈ 1.6694

From the t table as shown in previous sheet, the critical values are ±6694 .1.

The decision rule is:

Reject 𝐻𝑜 if 𝑡𝑆𝑇𝐴𝑇 < −1.6694 or if 𝑡𝑆𝑇𝐴𝑇 > +1.6694 , otherwise do not Reject 𝐻𝑜 .
Step 5: Collect the data and compute the value of the test statistic.

From MS Excel:
Figure 3.3 Separate-Variances t Test for the Difference between Two
means (assumes unequal population variances).

Figure 3.4 Solutions of figure 3.3


Alternate solution in solving 𝑡𝑆𝑇𝐴𝑇 :

̅̅̅1 − 𝑋
(𝑋 ̅̅̅2 ) − (𝜇2 − 𝜇1 )
𝑡𝑆𝑇𝐴𝑇 =
2 2
√𝑆1 + 𝑆2
𝑛1 𝑛2

(59.8438 − 59.0000) − (0)


𝑡𝑆𝑇𝐴𝑇 =
2 2
√(12.0432) + (13.1978)
32 34

𝑡𝑆𝑇𝐴𝑇 = 0.2715

Step 6: State the statistical decision and the managerial conclusion.

Because −1.6694 < 𝑡𝑆𝑇𝐴𝑇 < +1.6694, do not reject the null hypothesis.

There is no enough evidence at the 10% level of significance to support the claim that

the mean Differential Calculus–Midterm Examination scores for the students of the

two instructors are different.

(See also the solutions of Case Study No. 3 using MS Excel on the accompanied MS

Excel file of this document.)

You might also like