Time in a Box and Statistics

Kathleen's post, Time in a Box, about the oldest sewing patterns in her collection got me searching through mine to look for my oldies but goodies.

This is one of the oldest patterns that I bought new and kept. In my teens, I didn't buy many patterns. This is before patterns were sold as loss-leaders. They were pricey then.  (But, perhaps, better drafted and more thoroughly tested?)

I made the shorts several times and the pants once. The pants were a wadder, more due to my poor fabric choice than the pattern. But check out version one of the shorts!

I still remember that I paid $12/yard (1.5 yards) for the Hoffman cotton sateen, bought at Kaufman's Fabrics in downtown Berkeley.  I had to work 4 hours to pay for the fabric, notions and pattern at my job in the Chemistry lab, one of the highest-paying student jobs on campus.  I did gulp at the expenditure, but they became my favorite volleyball shorts.  I wore those giraffe shorts at least once a week for nearly a decade.

The giraffes were printed in Japan on American basecloth.  Hoffman is still selling cotton sateens for about the same price.  I wonder where they are made today?

Kaufman's closed years ago.  Willi Smith, the pattern designer, died of AIDS shortly after I made the shorts.

But I recently came back into contact with these friends/coworkers from the Chemistry lab through facebook.  I joined FB at my neighbor's suggestion.  I didn't know what to make of it--I still am not sure what it is about.

But, a week after I signed up, someone contacted me and asked me if I was the person he went to HS and college with.  I checked his pix and profile and recognized his friends as our friends.  So I friended him.  Then our former boss and the organizer of the Berkeley Chemistry Demo Lab group on FB found me because I was a friend of a friend.  I joined the group and he sent me a link to this photo from 1987.  I love the 1980s fashion vibe in the photo.

This photo then reminded me of the book, Statistics by David Freedman, Robert Pisani, and Roger Purves.  In the introduction, they discussed how statistics can be used to draw the wrong conclusions.  For example, the acceptance rate for women for graduate school at Berkeley was much lower than for men.  Did that mean bias?  A committee looked at acceptance rates for a half dozen major departments, broken down by gender.   It turned out, that acceptance rates varied by department, between ~30-80%.  In most departments, males and females were accepted at roughly comparable rates.  If anything, the data showed that women were more likely to be admitted on a department-wide basis.  The overall higher rejection rate for women reflected that they were more likely to apply to the most competitive graduate programs at Berkeley. 

Why did that picture remind me of the example in Statistics? Because everyone in the photo worked at the Chemistry Demo Lab (and stockroom) and there are two females and one male in the photo.  That was also the gender ratio of the workers during my tenure at the lab.  We were jokingly called Lonnie's angels by the male graduate students who used to hang out in the lab with us and get in our way.

Lonnie was called to task for hiring so many females.  If he drew his employees from students who had completed honors Freshman Chemistry and Quantitative Analysis,  and the population of that class was about 80-90% male, then why were the majority of his employees females?

Lonnie countered that he interviewed students that landed in the top third in theoretical grades and the top fifth in laboratory grades.  We made the stock solutions that the students used.  The students' quantitative analysis accuracy depended upon our accuracy.  It was very important to find the most careful workers.  Secondly, he wanted workers that understood what they were doing and why.

After the grade cutoff, he went by interview impressions.  We were interviewed by returning workers as well as Lonnie.  Basically, we had to be the type of person they wanted to work with.

So how did he hire 2/3 females from a class that was 7/8 male?  Because the female students who sign up for honors science classes are different than the males.  Women tend to underestimate their abilities and men tend to overestimate theirs.  The average female student who signed up for that class was stronger than the average male classmate.  Males and females were roughly evenly represented in the top third in theoretical scores, but the females had higher laboratory scores.  There was no bias.  Lonnie beat the rap.

No one told us, but we were smart enough to realize that, if we wore short skirts while performing the demos, the students were more likely to stay awake and pay attention.  And, thanks to FB, I learned that nearly all of my former coworkers went on to earn MDs and/or PhDs.  Lonnie certainly knew how to pick them.