40pts Total

This is the second of two major assessments in this course.  You can think of it as a “project” or “take home exam.”  This is an individual assignment.  You may not work with other students or tutors.  You may only get help from Prof Krone or the TAs.  You can use any course resources, R, and the internet (but may not post directly to a message board or ask anyone online questions).

     I have scaffolded you well with the lecture-lab-assignment format.

Because this is one of your major assessments, I expect you to prioritize this assessment.  Therefore,         extensions will only be given under dire circumstances and must be requested before Sunday 12/11 at       5pm.  With the copious amount of office hours, I expect you to work ahead.  To succeed, you need to start early.

    You must show work and R code. In each of the following problems, show the set-up, show your

work, use appropriate notation, indicate the final numerical answer, and include any R code and output in your submission.  Provide us with the code from the “console” so we can see what      numbers you got.

    Due Wednesday 12/14 at 11:30am on GradeScope (within Canvas).

     Only typed assignments (ex. Made in Word or LaTeX) will be accepted.

    When you upload your assignment to GradeScope, you will have to identify where in your

scanned file the solutions to each problem are.

    Assignments are considered late if they are submitted after 11:30am without an extension o The late deadline is 11:30am on Friday 12/16

    Box, circle, or highlight your final answer for each question.

     If your final answer is not an integer, round it to four decimal places unless otherwise indicated.

You may ask questions regarding the wording of the problems or have us see if your code is doing what you think it should be doing.  We will not be checking your answers.  Please attend office hours.  Email Prof Krone (j[email protected]) if you have questions outside of office hours.  I cannot guarantee I will respond to emails sent after noon on Tuesday 12/13; this includes responses to threads started earlier.     Therefore, you need to start the project early to get help.

Problem 1 (6 points)

1.   You ran a study with two conditions and each observation only experienced one condition.

Import the “Project2Prob1” data, create a 97% confidence interval for the difference in group       means (group1 − group2), and interpret the confidence interval.  Make sure to do any previous steps you deem necessary. Note: you need to apply your understanding of confidence intervals to the two condition setting.  All the relevant information you need is provided.  Therefore, for         example, I do not want you to calculate a sample size.

Problem 2 (3 points)

2.   Using a test we covered in class, check to see if the “Project2Prob2” data comes from a gamma    distribution with a = 2 and F = 3.  Include a complete interpretation sentence.  Hint: running the test will be a single line of code.

Problem 3 (4 points)

3.   You are testing a new treatment and you take a before treatment measurement and an after        treatment measurement of each subject.  Import the “Project2Prob3” data, create a 98%             confidence interval for the difference in means (after − before), and interpret the confidence interval.  Make sure to do any previous steps you deem necessary.  All the relevant information you need is provided.  Therefore, for example, I do not want you to calculate a sample size.

Problem 4 (6 points)

You want to determine whether there is a difference in mean concertation of a pollutant in the soil of the  Midwest (mw) versus the East Coast (ec).  Based on the opinion of EPA officials, you believe the             variance of pollutant concertation will be 25.  It would be meaningful to the EPA to detect a difference of 6 with 99% power and a Type I error of 3%.

4.   Complete everything you need to do to run this study.  When you are ready to analyze data,     import the “Project2Prob4” data.  If the data is normal, find the power of your study to detect a difference of 7 and discuss why that power makes sense.  If the data is not normal, were the     assumptions of the test met?

Problem 5 (8 points)

You want to see if the average mean miles per gallon (mpg) are different for 3 experimental car models. You want a correct retention of 96% and sensitivity of 95% to detect a smallest difference of 2 with an  assumed standard deviation of 12.

5.   Complete everything you need to do to run this study.  When you are ready to analyze data,       import the “Project2Prob5” data.  If the data is normal and it is a balanced study design, find the actual power of your study (simply report the power, you do not need to interpret the power).  If the data is not normal or unbalanced, do not calculate power (because you can’t).

Problem 6 (10 points)

You are running a study to look at the effects of wind turbine height (Factor A: tall or short) and blade width (Factor B: thin or thick) on the amount of energy produced in rural Iowa (go Hawkeyes!).  You  want an alpha of 1%, power of 80%, to be able to detect a minimum difference of 10 for turbine height (assumed standard deviation of 4) and 5 for blade width (assumed standard deviation of 4).

6.   Complete everything you need to do to run this study.  When you are ready to analyze data,       import the “Project2Prob6” data.  If the data is normal and it is a balanced study design, find the actual power of your study (simply report the power, you do not need to interpret the power).  If the data is not normal or unbalanced, do not calculate power (because you can’t).