CLARENDON COLLEGE

Roxas, Oriental Mindoro

Tel fax: (043)289-7056 / admin@clarendonph.com

Name: Mark Brian M. Mariño

Subject Advanced Statistics
Year Level 2nd Year – BSED Math

Solve the following problems. Show your solutions.

1. A graduate student performed a pilot study for his dissertation. He wanted to examine the
effects of animal companionship on elderly males. He selected 10 male participants from
a nursing home. Then he used an ABAB research design where A represented a week

with the absence of a cat and B represented a week with the presence of a cat. At the end
of each week, he administered a 20-point survey to measure quality of life satisfaction.
The survey results are presented in Table 5.9.
Use a Friedman test to determine if one or more of the groups are significantly
different. Since this is pilot study, use a = 0.10. If a significant difference exists, use
Wilcoxon signed ranks tests to identify which groups are significantly different. Use
the Bonferroni procedure to limit the Type I error rate. Report your findings.

1. STATE THE NULL AND REEARCH HYPOTHESES

The null hypothesis is
HO = θW1= θW2 = θW3 = θW4

The research hypothesis is

HA = One or both participant’s strategies will have the effects on the animal
companionship on elderly males.

2. SET THE LEVEL OF OF RISK (OR THE LEVEL OF SIGNIFICANCE)

ASSOCIATED WITH THE NULL HYPOTHESIS
The level of risk, also called an alpha (a), is frequently set at 0.10. We will use an alpha
of 0.10 in our example. In other words, there is a 90% chance that any observed statistical
difference will be real and not due to chance.

3. CHOOSE THE APPROPRIATE TEST STATISTICS

The data are obtained from four dependent or related conditions, the four samples are
small with some violation or our assumptions of normality. Since we are comparing
four dependent conditions, we will use the Friedman test.
4. COMPUTE THE TEST STATISTICS.

TABLE 5.9
PARTICIPANT WEEK 1 WEEK 2 WEEK 3 WEEK 4
S
1 2 1 3 4
2 3 2 4 1
3 1 4 2 3
4 2 1 3 4
5 1 2 3 4
6 1 4 3 2
7 2 4 1 3
8 4 2 1 3
9 3 2 4 1
10 1 4 2 3

WEEK 1 = 2+3+1+2+1+1+2+4+3+1 = 20
WEEK 2 = 1+2+4+1+2+4+4+2+2+4 = 26
WEEK 3 = 3+4+2+3+3+3+1+1+4+2 = 26
WEEK 4 = 4+1+3+4+4+2+3+3+1+3 = 28
5. DETERMINE THE VALUE FOR REJECTION OF THE NULL
HYPOTHESIS USING THE APPROPRIATE TABLE OF CRITICAL VALUES
FOR THE PARTICULAR STATISTICS
k= 4
n= 10
a= 0.10
Table B.5 returns a critical value for the Friedman Test of 6.360

6. COMPARE THE OBTAINED VALUE TO THE CRTITICAL VALUE

The critical value for rejecting the null hypothesis is 6.360 and the obtained value is Fr =
2.16. If the critical value is less than or equal to the obtained value, we must reject the
null hypothesis. If instead, the critical value exceeds the obtained value, we do not reject
the null hypothesis. Since the critical value exceeds the obtained value, we do not reject
the null hypothesis.

7. INTERPRET THE RESULTS

We did not reject the null hypothesis, suggesting that no significant difference exists
between any of the four conditions. Therefore, no further comparisons are necessary with
these data.`

8. REPORTING THE RESULTS

The reporting of results for the Friedman test should include such information as the
number of subjects, the Fr statistic, degrees of freedom, and p-value’s relation to a.For
this example, the frequencies of participants s were compared over four conditions. The
Friedman test was not significant (Fr = 2.16 p > 0.10 ). Therefore, we can state that the
data do not support to examine the effects of animal companionship on elderly males.
2. A physical education teacher conducted an action research project to examine a strength
and conditioning program. Using 12 male participants, she measures the number of curl-
ups they could do in 1 min. She measured their performance before the programs. Then,
she measured their performance at 1-month intervals. Table 5.10 presents the
performance results.

Use a Friedman test with a = 0.05 to determine if one or more of the groups are
significantly different. The teacher is expecting performance gains, so if a significant
difference exists, use one-tailed Wilcoxon signed ranks tests to identify which groups are
significantly different. Use the Bonferroni procedure to limit the Type I error rate. Report
your findings.