Math 251 Exam 1 Project

Due 3 pm Monday October 18, 1999, to me or under my door.

Your job is to find one or more data sets that interest you, and that are fairly rich. Then analyze the data, torturing it till it gives up its secrets, and write up what you have found out.

In terms of technical output: At least (others may be appropriate):

• One-variable data: stemplot and/or histogram, 5 number summary and boxplot (comparative boxplots if appropriate). Check for normality. Describe, especially outliers and any possible causes for them.
• Two variables (quantitative): Scatterplot, interesting points, correlation, linear regression (formula and fit line), residuals. If there is nonlinearity, a non-straight-line fit to the scatterplot.
• Data generation: How was the data generated? Experiment or sample? Critique the quality of the data generation, as best you can with the information you have. What would you like to know that you don’t?.
• (At least) One other: Time plot (if time was not a variable in your regression)

or Categorical data analysis
or Logarithm transformation of a variable (either to get a more "normal" distribution
of a right-skewed single variable, or to examine exponential-like growth)

Sources: These are a beginning. Any set with analyses attached will go into topics we haven’t learned. Just use our tools; don’t present anything you don’t understand.

EESEE Stories are good, and the questions can guide you to interesting analyses. Don’t be bound by their questions. The package should be on most of the computers in Mac101, under IPS-Student, in the Program files. Also, I found it will run from the CD just fine on the Macintoshes in Macmillan—you just have to go down enough folders to find the program file. On PC’s the thing that may keep it from running well from the CD is the needed type-font (Mpaltino.fon, in the EESEEPC folder). If this is copied to C:Windows\Fonts, it runs. If you don’t know how to do that, you can Install the CD (as described in the front of your book) and then if need be, Uninstall it. The font will be left in the right place. (Installing on the Wells networked computers may not actually get the programs installed, since many computers revert to the standard image every time you reboot. But the font will stay…)

IPS text, Appendix, has 7 interesting data sets.

IPS p.232 suggests some websites. I have not checked these out.

DASL (lib.stat.cmu.edu/DASL/DataArchive.html) is an on-line library of datafiles and stories that illustrate the use of basic statistical methods. These have some pre-analyzed "answers" but there is probably more to find out from the data than what they present. I looked through these—there are certainly interesting ones here.

Sharing the work: Use whatever method you like to share, from working together on everything to parceling out all the tasks separately. Everyone should read and be willing to stand behind the whole final thing. In an endnote tell me how you shared the work; who was responsible for what.

Pairings: 12 people= 6 teams of 2. I tried to match everyone with someone on their list. I didn't quite succeed. (4 of 6) It wasn’t easy! But I think the other pairings will work well too.

Lulu (Jinzhuo) Zhao – Michelle Trickey

Jocelyn Becraft—Carrie Gleasman

Jennifer Clark—Heather Legge

Jennifer Cuyle—Kristina Konechy

Yan Qing Xiao— Beth Kerstetter

Jessica Swanson —Wen Qiu Wang