Part B, C, and D:
Part B
Use STATA where relevant. Interpret all of your results. You cannot use the same articles as your classmates nor can you use the same article for both 1 and 2-they must be TWO DIFFERENT ARTICLES.
1. Find a journal article that uses a chi-squared test (Pearson’s test) to test for
independence of two categorical variables (in a contingency table).
-
Write the reference to the article(hint:can use pubmed.gov or a particular medical journal and search for “chi-squared test” and a specific disease or
public health area that interests you),
-
State the appropriate hypotheses to test for independence of content of interest.
-
copy the contingency table into your homework,
-
Complete and reproduce the hypothesis test and their p-value (show the STATAoutput and do by hand expressing formulas appropriately).
-
What conclusion did the investigators draw Did they use the correct statistical test?If not, did they come to the right conclusion?
2. Find a journal article that uses a Fisher’s exact test to test for independence of two categorical variables (in a contingency table).
-
Write the reference to the article (hint:can use pubmed.gov or a particular medical journal and search for “chi-squared test” and a specific disease or public health area that interests you),
-
State the appropriate hypotheses to test for independence of content of interest.
-
Why did they use this test instead of the chi-squared test?
-
copy the contingency table into your homework, and
-
Complete and reproduce the hypothesis test and their p-value (show the STATAoutput and do by hand expressing formulas appropriately).
-
What conclusion did the investigators draw? Did they use the correct statistical test?If not, did they come to the right conclusion?
Part C
Use the Weight dataset.
First create three body mass index (BMI) groups: 30. Note that you will have to convert the data to the correct units.
1) Test the hypothesis that there is an overall mean difference in systolic blood pressure (sbp) levels among the three different BMI groups.
-
Write down the null and alternative hypotheses using appropriate notation
-
Check the assumptions of your test. Try with and withouttransformations on the data.
-
Conduct the test (i) with and (ii) without transformations on the data.
-
Perform pair wise comparisons of the means using a multiple comparison adjustment (i) with and (ii) without transformations on the data.
-
Interpret your results with and without transformations of the data and decide which is most appropriate.
Part D
Download Smoke data and complete the following questions in Stata.
1.Why are nonparametric methods well suited to study whether there is a difference between numbers of days abstinent from smoking for males vs females? Answer your question by checking the assumptions for normality, performing appropriate transformations and again checking for the normality.
2. Use the appropriate nonparametric method to test whether there is a difference between numbers of days abstinent from smoking for males vs females?
3. Suppose that there was a subset of individual smokers who had undergone treatment with the eventual goal to quit smoking completely by 6 months follow-up. Suppose we had information on the number of days abstained from the last cigarette at baseline and at a 2- month interim. How would you test whether the increase in days abstained from the last cigarette differed among males and females considering what you know about this dataset?