[FoRK] Harley & Google's battle of the sexes

Gregory Alan Bolcer greg at bolcer.org
Fri Aug 11 08:59:29 PDT 2017


I guess you could just temporarily self-identify.

But getting to the real question.  Is Google institutionally sexist 
because of the low number of women engineers?

They call this the legal theory of disparate impact liability.  For 
Google in particular, I don't think they are.  Here's my reasoning, I 
don't think the numbers by themselves are sufficient to prove 
discrimination.

This is the whole correlation versus causation argument.  I use the same 
analogy I learned in my very first machine learning class every single 
time.

The grass is wet implies it is raining.
There is smoke implies there is a fire.
17% of engineers at Google are female, but 28% of computer science 
degrees are female implies there exists institutional sexism at Google.

The sprinklers are on;
My neighbor is making smoked salmon, (yes, you can have wood that smokes 
heavily w/o igniting, aka no fire);

So what about the third?  There are around 40,000 engineers who work at 
Google spread across r&d, operations, and other things.  60 women are 
currently considering suing Google for sexism, aka learning less despite 
equal qualifications (with no mention of duties, btw).

If you want a confidence Interval of 95%, you take a sample size of 60 
w/ a population 40000, to be 50% confident.  But working backwards, you 
probably need 104 to join the lawsuit.

The bottom line is, and I quote from a source I read recently: "If a 
significant fraction of employees think that diversity is important, 
then it is important. No further justification is needed for this."

That's just how business and business culture works--or works against....

Greg

On 8/11/2017 7:14 AM, Stephen D. Williams wrote:
> And how will you/he be certain of everyone's current and historical gender?
> 
> sdw
> 
> On 8/11/17 4:27 AM, Gary Stock wrote:
>> I will refrain from commenting on the previous substance of this thread
>> until at least five women have posted.
>>
>> Further, all new threads should be prohibited until Rohit convinces a
>> significant number of women to join FoRK.  (Alternately, he could drop a
>> significant number of men, although this may raise deeper questions of
>> equity.)
>>
>> Any application of statistics in response to my comment should focus on
>> establishing the "significant" number, what professions the added 
>> women (or
>> dropped men) must have, what their average salary must be, and so forth.
>>
>> If Rohit questions any of these priorities, Google should be made 
>> aware of
>> his persistent, blatant sexism.
>>
>> Hmmm...  There may be some problems down this path...
>>
>> -----------------------------------------------------------
>> Gary Stock                              gstock at nexcerpt.com
>> Nexcerpt    http://www.nexcerpt.com/    269.624.1140 direct
>>
>> On Thu, Aug 10, 2017 at 9:55 AM, Gregory Alan Bolcer <greg at bolcer.org>
>> wrote:
>>
>>> Insurance company A sells motorcycle insurance. Only 14% of motorcycle
>>> riders are women according to the Motorcycle Industry Council which 
>>> is far
>>> less than the 28% of women w/ computer science degrees. Does that 
>>> make the
>>> insurance company sexist?  Do they charge the same for both? Why aren't
>>> they doing more to encourage more women to ride motorcycles? Should the
>>> government step in and force them to discount their insurance to women
>>> motorcycle riders? (Yes, Harley is Rob from his fbook post, not 
>>> Davidson).
>>>
>>> There are groups that are trying to prove systemic racism and
>>> institutional racism.  I'm sure systemic sexism and institutional sexism
>>> are next.   Basically the institutional part means that anywhere you 
>>> find a
>>> numeric discrepancy over a large enough sample set, and here's where it
>>> gets iffy, that it must be the cause of institutional -isms even if none
>>> are observed.
>>>
>>> So do the statistics lie?
>>>
>>> Greg
>>>
>>> -- 
>>> greg at bolcer.org, http://bolcer.org, c: +1.714.928.5476
>>> _______________________________________________
>>> FoRK mailing list
>>> http://xent.com/mailman/listinfo/fork
>>>
>> _______________________________________________
>> FoRK mailing list
>> http://xent.com/mailman/listinfo/fork
> 
> 


More information about the FoRK mailing list