Announcements
- 2:53 AM, December 14, 2010
- I posted
the answers to the homework problems due last week to the
homework solutions web page.
The grace period for the last extra credit parts ends at 10:15 PM on today (when the exam ends) because a
university regulation doesn't allow any extra work to be submitted after the final exam. *** I posted some corrections (found by one of you students) for p. 346 of the
text on the
web page listing errors found in the text.
- 6:32 AM, December 7, 2010
- I posted information pertaining to the final exam on the
web page pertaining to the final exam. (Please note that the exam is open book and notes, and you can use a calculator
but not a computer.) *** I posted some extra credit problems on the
homework web page. If you've been doing great on the homework all semester there is no need to do these extra credit problems.
If StatXact won't give you exact p-values, then do Monte Carlo approximations (but let me know that is what you did).
*** I added the answers to the homework problems due last week to the
homework solutions web page.
*** I posted some info about using StatXact to do Ch. 12 tests on the
web page pertaining to using StatXact and other software.
*** I listed some more corrections on the
web page listing errors found in the class notes (for pages 10-20 and 12-15).
*** Please try a permutation test (using the raw data values instead of any scores) on
this data and e-mail me what you get for a p-value (from a two-tailed test). Please ... everyone do this.
- 2:48 AM, December 1, 2010
- I posted
some more homework (based on Ch. 10 and Ch. 11, and due on 12/7) on the
homework web page. (Note: Next week I'll post some extra credit problems based on the parts of Ch. 12 we can easily
do using StatXact, but many of you don't have to worry about getting any extra credit points (and so you can feel free to skip these extra credit problems).)
*** I posted some info about using StatXact to do Ch. 11 tests on the
web page pertaining to using StatXact and other software.
*** I corrected some minor typos in this example. (I twice indicated a sample size of 6 where it should have been 8.
Also in one place I changed a M-W statistic value from 57.7 to 57.5.)
*** I listed some more corrections on the
web page listing errors found in the class notes.
*** Please try a permutation test (using the raw data values instead of any scores) on
this data and e-mail me what you get for a p-value (from a two-tailed test). Please ... everyone do this.
- 6:09 AM, November 23, 2010
- I posted
some more homework (based on Ch. 10, and due on 12/7) on the
homework web page. (Note: Next week I'll post some more problems some Ch. 10, and perhaps some from
Ch. 11. These will also be due of 12/7. (So you'll just have one week to do the problems I post next week, instead of the usual two weeks.)
I may post some simple extra credit problems focused on using StatXact to do some of the tests I cover during the last lecture. These would
be due sometime prior to the final exam. Students who have done excellent work on the homework throughout the semester should not feel obligated to do
any of these last extra credit problems.)
*** I added the answers to the homework problems due last week to the
homework solutions web page. ***
Also, I posted some info about using StatXact to do Ch. 10 tests on the
web page pertaining to using StatXact and other software. *** Please not that I'll be traveling and away from
easy access to e-mail from 11/24 to 11/28. *** Finally, when doing any of the homework problems, if you encounter ties when assigning scores, be sure to use
the midrank method unless I explicitly instruct you to do otherwise for a particular part.
- 6:42 AM, November 17, 2010
- I listed a correction for p. 9-1 on the
web page listing errors found in the class notes.
- 8:56 AM, November 16, 2010
- Despite the fact that I strongly suspect something is wrong with StatXact's implementation of the Conover test, I'll just keep it in the
homework. Although I think for some data sets the p-value can be way different from what it should be, I suspect that for the HW data the test can be viewed as
an approximate test. I recently forwarded you an e-mail exchange (with StatXact people) about bugs I found in the Wald-Wolfowitz test. I plan to
send them my complaints about the Conover test somewhat soon. (I want to do a Monte Carlo study to back up my claims about how their implementation of the
test does not give exact p-values.) I'll update you about this later.
*** I listed several more corrections on the
web page listing errors found in the class notes, and a couple of additional corrections on the
web page listing errors found in the text.
*** I added the answers to the homework problems due last week to the
homework solutions web page. *** I'll also suggest that doing Problems 9.1, 9.2, and 9.3 on p. 340 of the text might be
good preparation for some parts of the final exam. (You can compare answers with one another, or use StatXact to check your work.)
- 2:39 AM, November 10, 2010
- I changed the homework due dates on the
homework web page so that no problems will be due the Tues before Thanksgiving. This is because I didn't have time to
cover the normal scores test and some other tests in Ch. 8 ... it's not because I want to make it so you feel that coming to class is optional that night. So
a bunch of problems will be due on the 30th (Tue after Thanksgiving). I'll also announce here that I'll be away from e-mail from the 24th through the 28th.
*** I listed several more corrections on the
web page listing errors found in the class notes, and a couple of additional corrections on the
web page listing errors found in the text.
- 10:25 AM, November 9, 2010
- I added the answers to the homework problems due last week to the
homework solutions web page.
*** I listed a bunch more corrections at the bottom of the
web page listing errors found in the class notes.
- 7:11 AM, November 2, 2010
- I kept playing around with StatXact and firmed up the rest of the homework that will be due this month (exercises based on
Chapters 7, 8, and 9). These problems are now posted on the
homework web page. (Note: Be sure to give me exactly the number of significant digits I request.)
In December you'll have more homework due, starting with some exercises based on Ch. 10.
- 1:14 AM, November 2, 2010
- I posted
some more homework (based on Ch. 6, and due on 11/16) on the
homework web page. (Note: Be sure to give me exactly the number of significant digits I
request ... not more and not less. E.g., 0.023 and 0.40 have 2 significant digits, and 0.00301 and 0.0190 have 3 significant digits.)
*** I listed a minor correction (for p. 243 of the text) on the
web page listing errors found in the text. *** Also, I posted some info about using StatXact to do Ch. 6, Ch. 8, and Ch. 9 tests on the
web page pertaining to using StatXact and other software. If you carefully go over the example I posted
for the Ch. 6 tests you shouldn't have much trouble using StatXact for the new homework.
- 5:47 AM, October 29, 2010
- I added the answers to the homework problems due this week to the
homework solutions web page.
*** I listed two more corrections at the bottom of the
web page listing errors found in the class notes.
- 2:19 AM, October 26, 2010
- I posted
some more homework (that is due on 11/9) on the
homework web page.
*** I listed a comment and a corrections (for Ch. 7 of the text) on the
web page listing errors found in the text. *** Also, I posted some info about using StatXact to do Ch. 5 tests on the
web page pertaining to using StatXact and other software. If you carefully go over the example I posted
you shouldn't have much trouble using StatXact for the new homework.
- 1:31 PM, October 22, 2010
- I added the answers to the homework problems due this week to the
homework solutions web page.
- 5:58 AM, October 20, 2010
- I posted
some more homework (that is due on 11/2) on the
homework web page (just one problem, having two parts, to be turned in, along with a "just-for-fun" exercise not to be turned in).
*** I listed two more corrections (for Ch. 5 notes) at the bottom of the
web page listing errors found in the class notes.
Also, I posted a comment about p. 492 of the text on the
web page listing errors found in the text.
- 3:10 AM, October 8, 2010
- I created a new
homework solutions web page
and posted the answers to the homework turned in this week on it.
- 4:42 AM, October 6, 2010
- I listed three more corrections at the bottom of the
web page listing errors found in the class notes.
Also, I posted two errors I found in Ch. 5 of the text on the
web page listing errors found in the text.
- 7:06 AM, October 5, 2010
- I posted
some more homework problems (that are due on 10/26) on the
homework web page. I also edited the Ch. 4 info on the
web page pertaining to using StatXact and other software. I advise reading through all of the
Ch. 4 StatXact examples before trying the new homework. (The examples are very relevant.)
- 4:05 AM, September 29, 2010
- I posted
some more homework problems (that are due on 10/19) on the
homework web page. *** I'm also happy to announce (but at the same time puzzled about it) that I'm now
getting exact p-values for the data sets that I previously got exact p-values for using a previous version of StatXact. (I'm not sure why
over the past week or so StatXact refused to give exact p-values for these data sets.) *** With regard to preparing for the final exam
(which is of course a long ways off at this point in the semester), in addition to understanding the determination of the expected value of RVN on the
top half of p. 3-30 of the class notes, also make sure you understand how the exact null sampling distribution and p-values are obtained on pages
4-10 and 4-11, and how StatXact's Monte Carlo option works (see bottom half of p. 4-18 of the notes, which I'll cover next week). *** Finally, I
indicated two more corrections needed for the class notes on the
web page listing errors found in the class notes.
- 8:25 AM, September 27, 2010
- In the part of
this web page (linked to from the
web page pertaining to using StatXact and other software) pertaining to using
StatXact, I changed the instructions for
using StatXact's Monte Carlo option to do homework problems for this semester, now stipulating that 1,000,000 (instead of 1,000,000,000)
Monte Carlo trials should be used (since using a billion trials can take a rather long time for some of the tests with certain data sets).
(Note: For the homework I've assigned so far, there shouldn't be any need to use the Monte Carlo option, but it may be that the Monte Carlo option
should be used for some of the homework I'll assign in the next couple of days.)
- 1:35 AM, September 26, 2010
- I added more examples to the
web page pertaining to using StatXact and other software, but I haven't checked them to determine
if some minor changes have to be made due to the fact that that we're now using
StatXact9 instead of
StatXact8.
Also, I listed a minor error I found on p. 101 of the book on the
web page listing errors found in the text.
- 6:43 PM, September 21, 2010
- I added an extra credit exercise to the last set of exercises I posted on the
homework web page. Also, I added three more corrections to the
web page listing errors found in the class notes.
- 5:26 AM, September 21, 2010
- I posted
some more homework problems (that are due on 10/5) on the
homework web page. These problems are based on the material I plan to cover in class later
today. (I might add to them if I get farther along than I think I will.) For these homework problems you don't have to
show a lot of work. Just give the test statistic value, the p-value, and indicate how (e.g., using a certain software package,
using a certain table, etc.) you obtained the p-value. *** Also, I created a
new web page pertaining to using StatXact and other software. For the new homework
problems we won't be able to use StatXact a lot (we'll have to also use tables, and other software to piece together
test statistic values and get approximate p-values). But in the coming weeks we'll use StatXact a lot more.
- 12:57 AM, September 18, 2010
- I listed a minor error I found on p. 95 of the book on the
web page listing errors found in the text.
- 6:16 AM, September 15, 2010
- I posted two more corrections (for pages 3-12 and 3-14) on the
web page listing errors found in the class notes. ***
Also, I posted some more homework problems (that are due on 9/28) on the
homework web page. These problems should help you get a better insight into some of the
probability used in Chapters 2 and 3. Most of the remaining homework problems this semester will involve performing various
tests on various data sets, and reporting p-value. A lot of them (but not all of them) can be done using StatXact.
- 10:22 PM, September 8, 2010
- I posted one more correction (for p. 2-31) on the
web page listing errors found in the class notes. ***
I listed one more error I found in the book on the
web page listing errors found in the text. *** Finally, I posted some more homework problems (that are due on 9/21) on the
homework web page.
- 7:10 AM, September 1, 2010
- I created a
web page listing errors found in the class notes.
Please inform me of any errors you find in the text, or on my web site or in the class notes I give you.
*** Also, I forgot to write the quote from the Hollander and Wolfe book on the board as I said I would, and so I added it to the bottom of
this web page. *** Finally I posted the first HW problems on the
homework web page.
There are 10 points of regular problems and 3 points of extra credit problems. Because the extra credit problems contribute the numerator
but not the denominator in my HW grading scheme, they will generally take more work than the points may suggest. An important thing is to do any extra credit problems entirely on your own! Last fall I turned in 4 graduate students to the Honor Committee because they did
not respect my rule with regard to extra credit problems. (Some students don't realize how easy it is for experienced faculty members to
determine exactly which students work together on homework.) All in all, the extra credit problems shouldn't be that important for one's
grade and it's so stupid to risk one's grade and reputation by not respecting the rules. So please work alone on the extra credit problems,
and feel as though it's perfectly okay to not do them. One can get an A without doing a single extra credit problem.
- 4:35 AM, August 30, 2010
- I created a
web page listing errors found in the text.
Please inform me of any errors you find in the text, or on my web site or in the class notes I give you.
- 1:24 AM, August 30, 2010
- I finally got around to updating the web site for the course.
I plan to add a lot to it as the semester progresses. For now, please read the information on the
welcome page and the
syllabus before the first class meeting, as well as
Ch. 1 and the first half of Ch. 2 of the text.
(I checked the
bookstore on Wednesday, and they do have the text on the shelves.)
*** I advise you you allow for plenty of
time to find a parking spot during the first couple of weeks of classes since there are generally more cars on campus
during these weeks.
*** You might find the last post (see below) from my Fall 2008 offering of this course to be of interest. The grades were quite good on
the whole, and I felt great about the class because the students seemed to do the homework on their own (as opposed to treating it as a group
project) and they did not pester me with many questions about the homework (but instead took it upon themselves to work harder if they
were having problems). It should also be noted that 2 out of 7 students got nearly a perfect score on the final exam, indicating that it
was not an unreasonably hard test.
- 7:07 PM, December 11, 2008
- I just submitted the final grades (2 A+, 1 A, 1 A-, 2 B+, 1 B). The top scores on the final were 39.9 and 39.8
(out of 40). I want to thank you for being one of the best classes I have ever had ... not just in terms of doing good work,
but also because there was no begging for help or favors, or asking for unfair exceptions, and there was absolutely no complaining.