PSYC 480 -- Dr. King

Questions to Guide Your Reading on the Superiority of Women in Statistics

Answering these questions while you read will help you read these articles
the way a "professional" would. You do not need to turn this in. It is just
a guide.

1) Where and when were these articles published?


2) Who was the author of the first article and where was he from?



3) What is the basic claim of the author of the first paper?


4) What prompted this author to publish his observations?



5) According to Minton and Schneider (1985), why might women be expected to
   perform as well as men in stat classes, even if they don't in math classes?




6) On the average, how large were Dr. Brooks' stat classes? Do you think this
   might have influenced his results?



7) Brooks converted letter grades to numbers as A=4, B=3, C=2, D=1, F=0, and
   then calculated mean grades for his male and female students. Do you think
   this is a reasonable procedure?



8) Comment on Brooks' statement, "This superiority of women is statistically
   significant, t(320) = 5.30, p < .001."



9) Then he did a second significance test on the same data, a chi-square test.
   Why and how?



10) What were Brooks' explanations for his results?



11) Brooks' article is followed immediately by a reply from Buck. Do you
    think it was fair of the editor to allow Buck to reply in the same issue
    of the journal (and on the same page!) as Brooks' results were presented?



12) Where was Buck from? What are your thoughts on that?



13) Over all, whose students did better in their stat classes, Brooks'
    students or Buck's students? And BE CAREFUL about how you state your
    answer. Don't say anything that is statistically unjustified.




Answers

1) Teaching of Psychology, 1987
2) Charles T. Brooks of King's College, a small private college located in
   Wilkes-Barre, Pennsylvania.
3) Female students in his intro stat courses make higher grades than male
   students (to which he should have added "on the average").
4) A previous article by Buck (1985) stated that women do not perform better
   than men in stat courses and may even be starting at a disadvantage.
5) Because (they say) stat is based more on computation and accuracy than
   other math courses are, and women do not perform worse than men at these
   tasks. (I don't buy it myself. What do you think?)
6) Once a year for 10 years, so 10 classes consisting of a total of 154 + 168,
   or 322 students, so about 32 students per class. That's far smaller than
   the stat classes I had as an undergrad, but it's still fairly large for a
   stat class where people might need individual attention (in my opinion).
7) Reasonable or not, that's the way grades are usually handled, for example,
   in calculating GPAs.
8) I'm relieved that he said "this" and not "the." Still, a more reasonable
   statement would have been, "This difference in means was statistically
   significant..." The rest of it indicates he did a t-test with df=320 and
   got a t-value of 5.30 and a very small p-value. (Which means that the null
   hypothesis was rejected and that the male mean and female mean were found
   to be significantly different.)
9) The chi-square test is done when there are groups but the DV is categorical
   instead of numeric. The chi-square test cannot be done on the percentages.
   These must be converted to the observed frequencies. You probably discussed
   the chi-square test of independence in your first stat class.
10) Women were more likely to come to review sessions, attended class more
    often, and tended to sit more towards the front of the room when they did
    attend. They also came to office hours for help more othen than the men
    did. In other words, women appeared more motivated and worked harder at it.
11) Fair has nothing to do with it. It's done all the time. If you "refute"
    someone else's results in a published article, be prepared for a quick
    (and not always friendly) reply!
12) Dr. Buck was from Delaware State College (now Delaware State University),
    a "historically black university" (Wikipedia). I googled both these colleges
    and read a little about them. The confounds practically jump off the page at
    you! These two authors are squabbling over differences between two different
    populations of students. Buck finally manages to point that out in her reply
    but only after half a column of possible "psychological" explanations. Those
    explanations might be right, but there are no data bearing on that, and as
    Buck pointed out, controlled studies are needed. Meanwhile, the most obvious
    statistical explanation is mentioned only briefly in one sentence near the
    end of the article. Now, go answer question 13.
13) It appears overall that Buck's students got higher grades. It would not be
    correct to say "significantly higher" unless we actually have done a test
    which shows that. It's also hard to say that this means they "did better."
    We don't really know that. We just know their grades were higher on the
    average.