Statistics: Lab #4

Chapters 9 - 10

You will have two files for this lab, they are separated below as parts A and B.

PART A, Hypothesis Tests (1 population)

Hypothesis testing one population.

Data: We will be using some data from the Fall 2015 Student Database to conduct hypothesis tests.

Preparing the data: Start a new file, put your name(s) in cell A1 and today’s date in A2, and then copy the Female, and Alcohol columns of data from the database into columns B and C.  Now copy the Greek column and Donate column of data into that worksheet and save it as Stat Lab 4 TESTS, yourname. This will serve as your sample of college students. Remember a woman is labeled with a 1 and a man is labeled with a 0. A student who says they will donate money back to the college is labeled with a 1, a student who has not is given a 0. A student in a Greek organization is identified with a 1, and a student who is not is given a 0. The Alcohol column is the number of alcoholic drinks the student claims to consume in a typical week.

Note: In your separate lab report, set up and solve each of the following three problems, submitting with your lab. Do not delete observations that have missing data. You may assume that you are sampling from an infinite population.

1. You will need to create a table of descriptive statistics for all four variables so that you know the sample mean, sample standard deviation and the number of responses. Put this output at the top of the worksheet. (4 points)

2a. Using the techniques covered in class, conduct a hypothesis test, including both a test statistic and a p-value that tests this research question: Is the average drinking of Hanover College students significantly greater than 3 drinks per week? Show all of your work in your lab report, including the null and alternative hypotheses, and be sure to specify a level of confidence. Thoroughly explain your results. (4 points)

2b. Now we’ll do this using the Ztest formula in Excel. The output of the Z-test is the p-value of the hypothesis test that a sample mean is greater than a hypothesized population mean. Somewhere near the descriptive statistics, create a label called Test 1. In the cell below this label type =ztest and then click on the function key ƒx . A window opens and you need to input the range of the alcohol data, the hypothesized population mean and, if it is known, the population standard deviation. If this last part is left blank, Excel will substitute the sample standard deviation so you can leave it blank. Using the result produced by the formula, thoroughly explain the results of this hypothesis test. (3 points)

3a. Using the techniques covered in class, conduct a hypothesis test, including both a test statistic and a p-value that tests this research question: is the population proportion of students at Hanover College who are in a Greek organization less than .50? Show all of your work, including the null and alternative hypotheses, and be sure to specify a level of confidence. Thoroughly explain your results. (4 points)

3b.  Now repeat this test using the Ztest formula in Excel. Create a label called Test 2. Put the Ztest formula in below this label. You may need to modify the output of the test to compute the accurate p-value. Using the result produced by the formula, thoroughly explain the results of this hypothesis test. (3 points)

4a. Using the techniques covered in class, conduct a hypothesis test, including both a test statistic and a p-value that tests this research question: is the population proportion of students at Hanover College who plan to donate back to the college different from .50? Show all of your work, including the null and alternative hypotheses, and be sure to specify a level of confidence. Thoroughly explain your results. (4 points)

4b.  Now repeat this test using the Ztest formula in Excel. Create a label called Test 3. Put the Ztest formula below the label. You may need to modify the output of the test to compute the accurate p-value. Using the result produced by the formula, thoroughly explain the results of this hypothesis test. (3 points)

 

Part B: Hypothesis testing two populations.

 

Sort the Data: We would like to do some hypothesis tests comparing women and men and Greek and non-Greek students.

Copy the Greek and Alcohol columns of data and re-paste them in a second worksheet in the same Excel file. Click at the top of the Greek column and click on the button that says Sort and Filter.  Select largest to smallest (Z to A). You should now see all of the data from the Greek students (Greek=1) arranged from the top of the spreadsheet down to the first non-Greek student (Greek =0).

For the Alcohol data, do descriptive statistics so that you have the separate means, variances and number of observations for both the Greek and non-Greek students. Clearly label these somewhere near the top of the worksheet. We will assume that the sample variances are equal to the known population variances. (2 points)

1a. Do Hanover Greek students drink significantly more in a week than Hanover non-Greek students? In your separate lab report, conduct a hypothesis test, including both a test statistic and a p-value that tests this research question. Show all of your work and be sure to specify a level of confidence. Thoroughly explain your results. (4 points)

1b. Now use Excel to conduct the very same hypothesis test. In the Data Analysis menu, choose “z-test: two sample for means”. Variable 1 range is your range of Greek drinking. Don’t check the Labels box. The variable 2 range is the range of non-Greek drinking. The hypothesized mean difference is zero. Enter the variance for each from the formulas you created earlier. Select the same level of alpha that you used when you did this by hand. Select a location for the output. (4 points)

1c. Using the result produced by Excel, thoroughly explain the outcome of this hypothesis test. What can you conclude about the mean drinking between Greek and non-Greek students? Be as specific as possible. (3 points)

 

Now copy and re-paste the Female and Donate columns in a third worksheet in the same Excel file. Sort the Female column so that we see all of the female students at the top, followed by the male students and their corresponding decision to donate. We will assume that the Female students are population 1.

For the Donate data, do descriptive statistics so that you have the separate sample proportions and number of observations for both the female and male students. Clearly label these somewhere near the top of the worksheet. (2 points)

2a. Is there a difference between the proportion of female and male students that plan to donate back to HC? In your lab report, conduct a hypothesis test, including both a test statistic and a p-value, which tests this research question.  Be sure to use the appropriate techniques for this situation. Show all of your work and be sure to specify a level of confidence. Thoroughly explain your results.  (4 points)

2b. Now use Excel to conduct the very same hypothesis test. This time choose the “t-test: Two Sample Assuming Unequal Variances” form of the test. Using the result produced by Excel, thoroughly explain the results of this hypothesis test. What can you conclude about the proportion that plan to donate between these two groups of students? Be as specific as possible. (4 points) 

3. For either of the two tests above, would your conclusions have changed if you selected different values for alpha? Explain and please be specific. (2 points)

 Save your file.

OK, you’ve carefully completed both parts A and B to this lab. You’ve double-checked that you have all printouts and your typed responses to questions. Bring printouts and typed comments to class on the due date. Your emailed files must also be sent by the beginning of class on the due date. When you email your file, make sure your name is in the filename. Late assignments or those sent incorrectly will be penalized. Check the syllabus or ask for help if you’re uncertain of how to do any of this.