Assignment 3, Inferential Statistics II

Student Name:

Grade:

[Please download a copy of this document and put all your answers in this document. Submit your assignment electronically through the assignment dropbox.]

[You must include the statistical software output that you used forobtaining the statistics to conclude your analysis.Please place them right after your answer for each part of the question.]

Part I

There were 18 subjects of similar physical conditions (similar in height, weight, gender, race, and health condition) participated in a study. They were randomly divided into three groups and each group took on a different diet and exercise program. To evaluate the outcomes from these 3 programs, BMI were recorded after two years participating in the program.

Program A / Program B / Program C
24.1 / 25.5 / 22.10
23.1 / 26.4 / 25.70
22.9 / 26.5 / 24.20
25.2 / 28.2 / 22.30
20.5 / 25.3 / 22.50
19.8 / 27.2 / 23.80

Perform a One-way ANOVA test to see if there is significant difference between the outcomes from three programs, at 5% level of significance, by answering the following questions.

a)Check necessary assumptions for the test and report your conclusions on these assumptions using p-values and with output from statistical software attached.

b)State the hypotheses in words or in math expressions. Report the ANOVA F-test result by reporting p-value, and conclude your analysis using p-value with output table from software attached.

Null hypothesis:

Alternative hypothesis:

Report p-value for the ANOVA test and conclude the analysis:

c)Use Tukey’s multiple comparisons method to identify and report the differencesin average BMI between programs and attach output table from software.

d)Perform Kruskal Wallis test and conclude the analysis using the p-value of the test.

Part II

In a case-control study for understanding the correlation between smoking and lung cancer, control group is the group of people who did not have the disease. Gender is considered as a possible confounder. Therefore, the data for each gender was collected and the cross-tabulated information is listed below. Exposure means subjects were exposed to the cigarette smoking environment.

Data for women:

Disease / Control
Exposure / 40 / 50
Unexposed / 50 / 1030

Data for men:

Disease / Control
Exposure / 100 / 30
Unexposed / 50 / 430

a)At 5% level of significance, use meta analysis (Mantel-Haenszel) approach to test to see if there is significant correlation between smoking and lung cancer using the data above.

b)Report the odds ratio of getting disease for those who were exposed to the smoking environment versus those who did not.

1