Having promoted this study, I’m now suspicious about p-hacking.
^ study will be out in April @“Cardinal Fang”
@“Cardinal Fang”
I don’t see error bars. Within the 12 GPA-major subgroups(3 majors x 4 GPA brackets), the only result that was statistically significant between genders at the author’s chosen threshold(p=0.01 level) was math majors with A GPAs. So for 11 out of 12 of the GPA-major subgroups, the between gender differences were not statistically significant at the author’s chosen p-value threshold.
Here’s why I’m suspicious about p-hacking in this male vs. female resume study.
@roethlisburger reports that in the full study, men had a 14% callback rate, and women had a 11.9% callback rate. For a study with 2106 resumes sent out, assuming about half the resumes were “male,” this is not a statistically significant result, not even at the lowly p=.05 standard that nobody should use because it’s far too permissive.
So, faced with a result that was not statistically significant, the authors looked for significance in subgroups. This is the very art and essence of p-hacking.
When I read “Of the applications she submitted from equally high-achieving male and female personas, men received calls for further discussion twice as often as did women with equal grades,” I interpreted that to mean that at all achievement levels, men were called back twice as often as women. Nope-- the authors sliced and diced until they found a big difference in certain subgroups.
I’m sorry I brought this study up. I should have been more suspicious.
Now, the initial study, about the income trajectory for different groups, is completely different. The population studied was enormous, and the results are clearly statistically significant, whatever their explanation.
Nowadays it is difficult to conduct a meaningful study like this by just sending out fake resumes. You really need to create full-blown online personas. You may not get called if you do not have a LikedIn account.
It is also important to know where she sent these resumes and for what positions. Customer-facing positions are hired based on likability. Women in Tech are hot commodities so for some mundane positions recruiters may prefer less accomplished candidates.
I disagree. I think the authors undertook a meaningful study, but then p-hacked the result.
Okay, I googled p-hacking. Seems to me that one could find plenty of data proving inequity without having to resort to that.
But I did learn a new word. 
There is plenty of data proving inequity, but unfortunately this study does not seem to be part of that data.
One final word about the gender study: GPAs between 2.5 and 2.83 got coded as C+/B-. If GPAs that low were getting callbacks and not eliminated by a minimum GPA threshold(ex. 3.0), that tells you something about bias in what jobs were being applied for.
IF this study had discovered an effect, then it wouldn’t matter about whether there was a GPA threshold. After all, someone is taking the jobs that are being offered, and it would be interesting that the hiring companies were favoring men over women.
But this study did not discover an effect. It is bogus.
I don’t know how many of you have screened resumes, but a lot of it is variable. Could depend on everything from how many people are screening to what their breakfast was that morning to how many are being looked at at the time. The same resume could be accepted and rejected depending on the day.
IF the study had discovered an effect, then “resume-screening is variable” would not explain that effect. Resume-screening is variable for both genders.
I’m thinking Jim-Bob and John-Boy are also unacceptable names.
@NPKR01 I agree!
By the way, people certainly SHOULDN’T discriminate based on someone’s name, but giving a child a highly unusual or yes, even a foreign name, can handicap that child later in life, because people don’t always act as they should. My son was friends with a number of green-card Asians in high school, and they all adopted typical American names when they arrived.
I was reading another discussion of the unusual name issue and some of the examples were truly surprising. My favorite in the unusual name category was “Y’Alljealous”. That would be an interesting one to test on a resume. As someone who hires people I’d be simultaneously horrified and fascinated all at once. Might even invite her in for an interview just because I wanted to meet her.
The names used in the study were not highly unusual. They weren’t even unusual at all. There are many, many Jamals, Rasheeds, Keishas and Latonyas. If you saw the resume of a Jamal or Keisha, you would not think, “I have never heard of that name before.” You would think, “This candidate is African-American.”
As usual Asians do better than everyone, but somehow white people get called out for some reason. Where I work all of top management is AA one white guy.
All I need to look at is med school admissions to see the curve is skewed and it is not to Asians or white people. I agree there are many racist people in America still even in 2018, but I think that racisms goes both ways. If you bring up race in every single conversation you perpetuate it you feed it.
The successful AA men I know rarely bring up race the people that aren’t doing so well seem to bring it up all the time. To me it seems the ones that point “you are a racist” are the most racist of all.
Hopefully it gets better.
Many decades ago I came across a book by Thomas Sowell in my local library. In one chapter he was looking into accusations of sexism and racism (because blacks and women earned less with comparable education). So he looked at folks with PHDs and sure enough, white males earned more than white females and white folks earned more than black folks.
When he looked deeper into it, he noticed the problem mostly disappeared if majors were factored in. White males simply majored in subjects such as engineering or computer science much more so than blacks or women, driving the discrepancy.
He then looked into Asian PHDs who have higher income as a group than even the whites to test his hypothesis and sure enough, they have an even higher percentage in these hot but difficult majors. (The only other possibility is that whites love Asian even more than they themselves. Even I don’t believe that).
This is what I call critical thinking.
While on the topic, Eric Schmidt. former CEO of Google, recently said that computer science attracts the brightest students at Princeton, not physics and math anymore. To me this is an interesting development that I did not see coming, tho I should.
@EarlVanDorn You don’t see a problem with this?
Sure some people might be comfortable with this but not all.
Frankly, people shouldn’t have to change so that society accepts them. This just stinks of racism and discrimination.
I, for one, REFUSE to change my name so that “society” will accept me.
The recent study @“Cardinal Fang” referenced may be flawed however it doesn’t change the results from the other study where identical resumes were sent out and the only thing that was changed was the name. Clearly there is a racial bias when POC apply for jobs, obviously that bias varies between industries and male/females.