Thank you, you've been a great class. Please fill out an evaluation, put it in the white envelope (circulating, or by HW folder) The Envelope and blank forms are outside my door. Please don't be a nonresponder; please do one. Be assured there is no chance I will look in the envelope until mid - June. Even if I wanted to, I don't have time.
Compare with your neigbors:
your answers to the homework.
Discuss the choices you made
for problems 74, 75, 76, and the conclusions you came to.
**Final Exam Wednesday afternoon, May 16. Comprehensive, with special attention on Ch.7. You may bring one sheet of paper with notes (both sides). The exam will be similar in style to the midterms, a mix of multiple choice, computation, written answers. About 1 1/2 to 2 times the length of a midterm. You should not need the whole 3 hours but you may have it if you like. If you plan to start late, please let me know ahead of time so I don't worry about you.
My availability:
Monday: Help session, in the classroom, 1-2. (If nobody
comes by 1:10, I'll return to my office)
Tuesday: Help session, in the classroom, 1-2.
""
""
I'll be on campus: If I'm not in my office (Mac 102), I'm
probably in Mac 101 (working with CS103 students) Find me there.
This afternoon, till about 4.
Monday 9:30 - 4
Tuesday 9:30 - 2
Wednesday 11- 4
Mary West (assistant)
Monday till 4:30
Tuesday by mutual arrangement (mwest@wells.edu).
(She is less sure about the material in 7.2; ask
me about that part)
~~~~~~~~~~~~~~~~~~~~~
Review:
[Data Analysis: description and exploration]
[Data Production: Sampling, Designing
Experiments]
[Statistical Inference: formal Estimating and Testing--quantifying our
uncertainty and satisfying the skeptic]
p. 424, # 7.68
p. 424, # 7.69, part b.
p. 426, middle--problems 74, 75, 76.
The above problems are quite good for review.
I encourage you to do them, if you haven't.
Analysis of usefulness of Sample Test. Did
working with the sample test improve the grade?
There was a weak positive relationship, mostly
because of one very low grade, low use of the sample test, pulling
the line down on the left.
When we did a scatterplot
of "used sample test" vs. (exam 3-exam2) we got a stronger relationship,
still not very strong. This was done to look at "improvement",
eliminate the lurking variable of the level you would have had anyway.
This doesn't entirely do it: the low
left outlier student told me she didn't use the sample exam because she
wasn't prepared enough to benefit from it. So instead of "Using
the sample test heavily" causing a better grade, maybe Using the sample
test heavily was really the effect of being well-prepared already.
Another lurking variable was nonresponse: 7 students didn't hand in the Used Sample Test info. Comparing the grades of the students who did with the students who didn't showed that students who did tended to do better on the exam. But a two-sample, one-sided t-test gave P= about .12, not enough to publish. (Done by hand, df = 7-1 =6. The machine computation in SPSS would have given a bigger df; I might have gotten P under .10. BUT:)
Sub-lurker: Of the 7 who didn't hand the
info in, 5 were absent, 2 were present. So lumping them is probably
a bad idea anyhow?.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Optional problems to study with: bottom of Day
41
Good luck!
SRS
| Sievers home | Math151-Sp01/Day42.htm | 11am | 5/11/01 |