SW Statistics

Problem Set on Hypothesis Testing

Answer all questions using both the Test of Significance Approach (using the indicated
alpha level/s) and the P-value Approach (using the conventional alpha levels):
(Independent T-test)
1. A school administrator would like to know whether students tend to select college majors that
are consistent with their skills. A sample of 12 Psych students and a sample of 12 English
majors are selected. Each student completed a verbal abilities test when he/she entered college.
The scores of these 2 samples are as follows:

Psych 18 15 14 21 18 13
Majors 17 16 20 19 21 14
English 19 20 23 18 16 15
Majors 22 19 19 16 21 23
a. Which t-test is appropriate for this problem – independent samples t test or a paired samples t-
test?
Explain.
b. Use the appropriate t-test to determine whether these data indicate any significant difference in
verbal ability for English and Psych majors. Test at 5% alpha.

(Chi-square)
2. The results of a random sample of patients with a certain disease treated with a drug or with a
placebo are shown in the Xtab below. At α = .05 and 0.10 can you conclude that the treatment
is related to the result? Based on the results, would you recommend using the drug as part of a
treatment for the disease?

Result
DRUG PLACEBO
Improvement 39 25
No change 54 70
(ANOVA)
3. An experiment is performed to determine whether 5 rice varieties differ in terms of yield. Four
plots of land are assigned to each variety and the yields (in sacks) are as shown below.
Assuming the plots are of similar fertility and that the varieties are assigned to the plots at
random.

VARIETY 1 VARIETY 2 VARIETY 3 VARIETY 4 VARIETY 5

20 10 23 15 21
12 14 16 17 14
15 12 18 16 17
19 15 14 12 18
a. Use your calculator to get the ⅀X and ⅀X^2 for each variety.
b. Determine whether there is a difference between the yields at 5% level of significance.
Interpret.
c. If appropriate, conduct the HSD given that q(.05) equals 4.65 and q(.01) equals 6.14.
Interpret.

(Test of means)
4. A random sample of 33 individuals was selected and the hours per week that each one
watches TV was determined. The data are as follows:
9 8 7 5 8 6 8 7 10 8 8
6 5 8 4 6 8 8 5 5 10 7
4 6 6 5 6 7 8 9 8 7 7
 Test the claim that people spend more than 6 hours per week watching TV at 5% and 1%
alpha.
(Paired sample t-test)
5. A sleep disorder specialist wants to test the effectiveness of a new drug that is reported to
increase the number of hours of sleep patients get during the night. To do so, the specialist
randomly selected 10 patients and records the number of hours of sleep each gets with and
without the new drug. The results are as follows:

PATIENT Hrs of Sleep Hrs of sleep with D D^2

without the drug the new drug
A 1.8 3 -1.2 1.44
B 2 3.6 -1.6 2.56
C 3.4 4 -0.6 0.36
D 3.5 4.4 -0.9 0.81
E 3.7 4.5 -0.8 0.64
F 3.9 5.5 -1.6 2.56
G 4.9 6.3 -1.4 1.96
H 5.1 6.6 -1.5 2.25
I 5.2 7.8 -2.6 6.76
J 3.9 5.7 -1.8 3.24
Is the drug effective in increasing sleep among patients who take it? Test at 5% and 1%
alpha.