MAT 2030 - Homework Assignments - Fall 2008

Experiment Question:  This entire question must be answered and turned in.

The National Institutes of Health is doing a clinical trial on the natural remedy Saint-John's-wort.  330 patients with mild depression have agreed to participate in the experiment to compare Saint-John's-wort with both a placebo and with Zoloft (a common prescription drug for depression).  At the end of the experiment, the subjects will rate their depression on a scale from 0 to 3.

1. Describe how you would set up a completely randomized design for this experiment.  That is, how many subjects should get each treatment, and how would they be assigned the treatments?

2. Explain why all 330 subjects should not be given Saint-John's-wort?

3. Explain what it means, in general, for an experiment to be double-blind.

4. Should this experiment be double-blind?  Explain why?

5. Suppose the 330 subjects were divided into males and females and then the treatments were randomly assigned within each gender.  What type of experimental design is this?  (Just give me the name.)

Journal

Th, 9/18

Your initial post is due before class on Tuesday, September 16th.  Your response to another student is due before class on Thursday, September 18th.  Please read the syllabus for what is expected of you in order to get full credit for a discussion.  Remember, these represent 10% of your final grade.  You will find the discussion forum under the Discussion Board button in Blackboard.

Note:  For # 20-22, you will find the data in Appendix B in the back of the book.

Note:  Data for # 13 & 14, 15 & 16, and 20 are in Blackboard under "Data."  Please print out and hand in the graphs for all these problems.

Test # 1 is on Thursday, September 25th, in the Martinetti HTM lab!!!!!!!!!

The test will cover all of Chapter 1, Gallup Polls, and Chapter 2, Sections 2-1, 2-2, and some of 2-4.  Anything in Homework # 2 - 5, 7 is fair game.  A list of topics is given below.

• Statistical terminology:  population, sample, parameter, statistic - be able to identify these in a given scenario

• Type of data: categorical vs. quantitative (discrete, continuous); nominal, ordinal, interval, ratio - be able to identify these

• leading questions

• critical thinking - alternative conclusions

• pictograms - why they should not be used

• sampling errors:  random sampling error - like in Gallup polls; bad sampling methods (voluntary response, convenience samples)

• nonsampling errors: processing errors, wording of questions, nonresponse, response error

• observational studies vs. experiments

• types of experimental design:  completely randomized, randomized block design

• blinding (single and double)

• types of samples: random (like simple random), cluster, stratified, systematic - be able to identify these

• types of studies: cross-sectional, retrospective, prospective

• analyzing a Gallup poll that you are given: you should know everything that was on the homework

• Creating pie charts, bar graphs, pareto charts, and line graphs using Excel

• Chapter 1, p. 34-35: # 1, 3, 4, 5, 6 (omit d), 7a
• Chapter 2, p. 69:  # 1

You will be allowed to bring one 3 by 5 index card to the test with whatever you want written on it (both sides).  You will also be allowed to bring the Excel instructions I provided you.  You should also bring a calculator.  Come to class with questions on Tuesday! 

Your initial post is due before class on Tuesday, September 23rd.  Your response to another student is due before the test on Thursday, September 25th.

• Section 2-3, p. 55: # 12, 13, 14  You are to create (in Excel) and print out histograms for each of these.  The data files needed are in Blackboard.  You have to define the bin range.  Remember to define your bins as the upper class limits of the frequency tables.  You should probably use the tables from the homework # 6 solutions, as you know these are 100% correct.  In addition:  for each problem, analyze the histogram (indicate the center, spread, shape, and any outliers).
• Section 2-4, p. 66:  # 6, 9  You are to make stem plots (by hand) for both of these.  In both cases, your stems should represent "10s" and your leaves should be "1s."  This is similar to the stemplot on p. 59.  It should be noted that a stem=8 would mean "80."  What would a stem=11 mean?

Journal

PS:  You do NOT have to calculate the midrange on any of the questions.

Charts and graphs are seen quite frequently in newspaper articles, magazines, books, etc.  There are pros and cons to using such visual representations.  Describe one pro for using a chart or graph.  Describe one con.

Your initial post is due before class on Tuesday, October 14th.  Your response to another student is due before class on Thursday, October 16th.

• The survey must be typed and well-written.
• Be sure to include brief directions.  Something like, "Please fill in the blank or circle ONE response for each question," should suffice.
• You should have 5 questions: at least two must be quantitative and at least two must be categorical.  The fifth question can be of either type.
  • For any quantitative questions, just leave a blank line for the participants to fill in.  The numerical questions are in essence open-ended so participants can put in any number they want.  For example,
  • How old are you?  __________ years
  • How many credits are you currently enrolled in?  ___________
  • For any categorical question, you must give options from which to choose.  You cannot have any open-ended categorical questions.  It wouldn't hurt to remind people to choose only one response again. You should give some sort of "other" and/or "none" option as well.  For example,
  • With what political party do you most associate?  Please circle only one.
    a) democrat      b) republican     c) independent   
    d)  progressive      e)  socialist     f) other:  __________      g) none
  • What is your favorite genre of music?  Please circle only one.
    a)  metal   b) pop   c) rap    d) jazz    e) hip hop   f) other:  ______________
• You should also turn in an edited copy of the informed consent form, which you can find in Blackboard under the project button.  Please fill in your name(s) and topic as appropriate.  Change all font to black (not red).  Do not change any other aspect of the letter.

Your initial post is due before class on Tuesday, October 21st. Your response to another student is due before class on Thursday, October 23rd.

• The survey must be typed and well-written.
• If you already turned in your informed consent form, you do not need to do so again.  Just make sure to make any changes to it that I suggested.
• If you have not turned in your set of questions to me, please email me with your survey and the informed consent form over the weekend so I can get it back to you so that you can turn in version two on Tuesday.

Your initial post is due before class on Tuesday, October 28th. Your response to another student is due before class on Thursday, October 30th.

Note:  For # 13 - 20, it may help to put the row and column totals on the table before you start.

Go to the website http://www.stat.sc.edu/~west/javahtml/LetsMakeaDeal.html.  Initially, read just the first paragraph.  What is your gut response to the question at the end of the first paragraph?  Read the rest of the webpage, play the game several times (I would suggest playing at least 30 times switching and 30 times staying), and discuss what happened.  Are you now convinced that you are actually twice as likely to win if you switch (versus if you stay)?  Why or why not?  Let’s argue!!

Your initial post is due before class on Tuesday, November 11th. Your response to another student is due before class on Thursday, November 13th.

**Note:  For questions 16, 20, 22 do the following.  Do NOT follow the instructions in the book.

• Make a scatter plot using Excel.  Also:
  1. Describe the overall pattern (if any) in the scatter plot: form (linear, curved, none), direction (+/-), and strength.
  2. Are there any outliers?
• Use Excel to calculate the linear correlation coefficient. Answer the following questions:
  1. Does the value of the linear correlation suggest that there is indeed a relatively strong linear relationship between the two variables?
  2. If yes, is it positive or negative, and why does that make sense?

• Empirical Rule:  Suppose that the heights of adult women are normally distributed with a mean of 65 inches and a standard deviation of 2.5 inches.  Answer the following questions using only the Empirical Rule. 
  1. Draw the corresponding bell curve including the number line.
  2. What percent of women are between 60 and 70 inches tall?
  3. What percent of women are less than 62.5 inches tall?
  4. What percent of women are more than 72.5 inches tall?
  5. The tallest 2.5% of women are at least how tall?
• Section 6-2, p. 257-258:  # 1, 3, 10, 14, 18, 22, 26, 34, 38, 40
  Journal

• The project must be typed in Word and well-written.
• You must attach a blank copy of the informed consent form and a blank copy of your survey.  If you don't, you lose 10 points.
• Don't forget that with your quantitative data, you can also use Excel to get Descriptive Statistics like the mean, median, std. dev., etc...
• JOURNAL:  If you are working with a partner(s), I only need to see one report, but I need each individual to do his/her own journal.
• DO NOT WAIT UNTIL THE LAST MINUTE!

What is one of the most important things you have learned in this class?  Explain why.

Your initial post AND your response to another student is due before YOUR final exam begins.

Final Exam

JO3 (10am section):  Thursday, December 18^th, from 8:00am to 10:00am, Martinetti HTM lab
JO4(11:30am section):  Tuesday, December 16^th, from 10:30am to 12:30pm, Martinetti HTM lab

Review Session:  Monday, December 15th, from 3:00 to 5:00pm in BENTLEY 313 (near my office)

Chapter 1: All; Gallup Polls; Chapter 2: Sections 2-3 and 2-4 (stemplots); all of Chapter 3; Chapter 4: Sections 4-1, 4-2, 4-3, and 4-7; Chapter 10:Section 10-2; Chapter 6: Sections 6-1, 6-2 (including Empirical Rule), and 6-3; Chapter 7: Sections 7-1 and 7-2 (calculating confidence intervals).Note: not all things from the first two tests will be on the final exam.  A list of topics is given below.  New topics (after Test # 2) are in red.

• Statistical terminology:  population, sample, parameter, statistic - be able to identify these in a given scenario

• Type of data: categorical vs. quantitative

• leading questions; critical thinking - alternative conclusions

• sampling errors:  random sampling error - like in Gallup polls; bad sampling methods (voluntary response, convenience samples)

• nonsampling errors: processing errors, wording of questions, nonresponse, response error

• observational studies vs. experiments

• types of experimental design:  completely randomized, randomized block design

• blinding (single and double)

• types of samples: random (like simple random), cluster, stratified, systematic - be able to identify these

• analyzing a Gallup poll that you are given: you should know everything that was on the homework

• Creating a pie chart or bar graph using EXCEL

• Interpreting a line graph you are given

• Making a histogram using EXCEL; interpreting a histogram: center, shape (skewed, symmetric, bimodal) and spread

• Interpreting a stem plot you are given:  min, max, outliers, center, spread

• Using your TI to calculate mean, median, mode, std. dev. and range

• Interpreting mean, median, std. dev., range

• How skewness affects the mean and std. dev.

• Calculating and interpreting z-scores

• Boxplots and Five-number summary:  interpreting one you are given

• Probability:  basic probability calculations using the classical approach - complement events; AND events; OR events; conditional probability; independence

• Calculating probabilities using a table - complement events; AND events; OR events; conditional probability; independence

• Counting: Fund. Counting Rule; Factorial; Permutations, and Combinations

• Creating a scatter plot and finding the correlation using EXCEL

• Interpreting a scatterplot (form, direction, outliers), recognizing different correlations, correlation does not equal causation!!

• Standard Normal Distribution:  using the Empirical Rule, calculating probabilities, percentiles

• Normal Distributions: calculating probabilities, percentiles

• Calculating 90, 95, or 99% confidence intervals

• Chapter 1, p. 34-35: # 1, 4cd, 5, 6 (omit d), 7a
• Chapter 2, p. 69:  # 2 (Excel), 5 (Excel)
• Chapter 3, p. 131: # 1 (omit d, j), 2, 4: (Put the data from #1 into Excel, to make the histogram.), 6, 8 
• Chapter 4, p. 192:  # 1-10, 16
• Chapter 6, p. 311: # 1ac, 2ac, 3abd
• Chapter 7, p. 377: # 1, 7 (Do a 99% conf. interval.)
• Chapter 10, p. 582: # 1 and 2 (For both of these, create a scatter plot using EXCEL: describe form, direction, etc.;  calculate r using EXCEL; does a linear model fit the data well yes or no?)

You will be allowed to bring one 4 by 6 index card to the test with whatever you want written on it (both sides).  You will also be allowed to bring the EXCEL and TI-83 instructions I provided you as well as the last purple handout on confidence intervals.  Bring your TI calculator!