[diversity] Best practice - increasing female participation in computer science

On 17/10/2017 11:02, Shane Kerr wrote:
> It looks like universities (well at least some universities) are 
> addressing the gender imbalance in their enrollment. This should
> start to resolve the "pipeline problem" (if indeed there is such a
> problem).

Agreed.

> If graduates start to look more diverse, do you think that the
> industry will automatically start to become more diverse?

That sounds like a plausible hypothesis.

> While this is one possible outcome, it also seems likely that we have
> similar structures in industry to the ones that were favoring men in
> academia. While the solutions may not be identical, certainly we can
> consider them as possibilities?

Very well, but that will require us to look at these approaches and
claims of bias critically, not simply to adopt those claims and
approaches without analysis.

> p.s. I note with interest that one of the solutions for increasing 
> women's enrollment in computer science was to make the criteria MORE 
> DIFFICULT for all applicants. Perhaps this can go some way to
> addressing the constant refrain that increasing diversity should not
> lower standards? :)

Are you referring to the following sentence from the article?

> The school made it harder to become a computer science major — as
> always applicants had to be good at math and science but now they
> also had to show they had leadership qualities.

Characterising this as "raising standards" or even as "make the criteria
MORE DIFFICULT" seems problematic.

I cannot see how this change of selection criteria would cause any shift
in the proportion of sexes (or any other characteristic) unless the
leadership score is being privileged at the expense of the previously
dominant math/science score. That's not a male/female thing nor any form
of comment on whether females are good at STEM; it applies regardless of
which characteristic "leadership skills" is attempting to proxy for (or,
indeed, in the absence of any such intent, if "leadership" is promoted
in its own right)

Allow me to explain by example. Consider the following candidates,
ranked in order of their STEM scores, also showing their leadership score.

   STEM    LEAD  TOTAL

A:  9       2      11
B:  9       1      10
C:  8       3      11
D:  7       2       9
E:  6       3       9
F:  5       5      10
G:  2       6       8

   (I wish to re-emphasise that *I* am not making any assertion about
   female capability in STEM or tech (or leadership); the above table
   is entirely sex and gender neutral)

With a class size of 5, under the old rules candidates A,B,C,D, and E
are selected and F is rejected; under the new rules F is selected
instead of either D or E. G is not selected under either ruleset.

What has occurred is simply that a candidate with lower capability in
STEM has replaced one with a higher capability, because they have a
higher capability in LEAD.

In other words, the rule change increases the priority given to
leadership scores compared with STEM. That is not to say that LEAD has
been given absolute priority over STEM (G is not selected under either
ruleset) but there has been a shift in priorities.

It is therefore appropriate to ask, is it desirable to de-emphasise STEM
skills as compared with leadership skills, as entry requirements for
admission onto an undergraduate computer science course?

When considering this question, it is impossible not to notice that the
college President's intentions were avowedly motivated: she addressed
this question not with the intention of sorting students into courses to
which they were best suited or most likely to succeed, but simply with
the intent of reducing the proportion of men on the course.

This is the kind of thing that leads some people to consider the
so-called diversity agenda simply divisive.

My final observation is that STEM scores are readily susceptible to
objective measurement, whereas leadership scores are not.

That is not to deny that STEM score measurements might be somewhat
affected by cultural biases in testing, or that at the extremes it will
be possible to achieve a fairly high level of agreement about whether a
candidate has exceptionally good or exceptionally poor leadership
skills. But in the round, it is highly probable that there will be a
materially higher level of objectivity in the measurement of  STEM
scores than of leadership.

In consequence, adopting "leadership skills" as an additional admission
criterion makes candidate selection a more subjective decision than
previously. In combination with the College President's lack of reported
interest in promoting leadership, and declared interest in promoting
female participation, an outside observer might fairly wonder whether
there was actually any sincere effort to broader the range of skills
reflected in candidates, or whether the college was simply looking for
an excuse to pick female candidates whose STEM scores wouldn't
previously permit their selection, without genuine reference to other
criteria.

That does little to, as you put it "address[] the constant refrain that
increasing diversity should not lower standards". I do not find it a
particularly attractive example for us to learn from, except as a
cautionary tale.


Kind Regards,

Malcolm.

-- 
            Malcolm Hutty | tel: +44 20 7645 3523
   Head of Public Affairs | Read the LINX Public Affairs blog
 London Internet Exchange | http://publicaffairs.linx.net/

                 London Internet Exchange Ltd
           Monument Place, 24 Monument Street London EC3R 8AJ

         Company Registered in England No. 3137929
       Trinity Court, Trinity Street, Peterborough PE1 1DA