Outreachy won't (directly) solve all our diversity problems

[Myrle Krantz <my...@apache.org>]

Hey all,

You already know this, but I ran some example numbers, and I figured you'd
appreciate them.

We're too big to use Outreachy to significantly impact our diversity
numbers.  Just to take one dimension of one segment of our committer
pipeline, and some simplified numbers as an example: if we have 8000
committers across all of our projects, and 5% are women, we'd need to add
444 women committers to raise the percentage of women committers to 10%.
(8000*(.1 -.05)/(1 - .1))

If all of the Outreachy interns were women, and 90% of the Outreachy
interns were to become project committers, we'd need to employ 494
Outreachy interns.

Outreachy interns cost the org 6.5 thousand dollars, work for 6 weeks, and
require roughly 5 hours of mentoring per week.  *So raising participation
of women to 10% would cost more than 3.2 million dollars, and nearly 15
thousand mentor hours, if done solely via Outreachy.*

Of course several of my assumptions are downright silly.  So, it would
likely cost much more.  But even that amount is currently well beyond our
reach.  And the participation of minorities in our projects is probably
worse than the participation of women, so continuing this calculation along
those dimensions is just going to get depressing.

Direct impact on our diversity statistics can't be our goal in
participating in Outreachy.

Instead, one of the hopes that have been expressed with respect to hosting
an Outreachy intern is that we'll be able to follow their progress and
learn from the problems they encounter (following the "Switch" model (1)),
so that we can figure out what's keeping people out of our communities and
tackle those underlying problems.

Naomi's recent e-mail on this is super exciting.  I am really looking
forward to watching the work she's describing.

Best Regards,
Myrle

1.)
https://www.amazon.com/Switch-change-things-when-hard-ebook/dp/B005TKD512/

Re: Outreachy won't (directly) solve all our diversity problems

[Daniel Gruno <hu...@apache.org>]

On 6/24/19 10:19 PM, Sage Sharp wrote:
> I would suggest taking a look at the OpenStack diversity reports:
> 
> https://superuser.openstack.org/articles/2018-gender-diversity-report/
> 
> "Newcomers who enter the community through a mentorship program boast a
> higher retention rate than other newcomers, at 13 percent versus less than
> 10 percent, respectively. In addition, these mentees contribute to the
> community for an average of 15 months, far longer than their three-month
> program duration. This report provides a springboard for additional
> research into mentorship programs, with the aim to identify and share best
> practices."
> 
> One of the concerns raised was whether the ASF is "too big" for diversity
> efforts to have an impact. It was mentioned that there are 8,000 ASF
> contributors. OpenStack has 89,000 contributors. If they can improve their
> diversity numbers, I'm sure ASF member projects can as well. :)

How is a contributor defined in this argument?

We have around 7,500 *committers*, of which roughly 3,700 can be said to 
be active (have committed code in the past year). We also have around 
74,800 *contributors* (depending on how you define a contributor, it's a 
grey area to say the least - my number is defined as "having interacted 
with the development process in some way that leaves behind a trace"), 
of which 11,500 are actively contributing code/docs material, expected 
to increase to 15,000 active within two years time.

Needless to say, we need to also ask ourselves "what are we trying to 
improve the diversity of?" within this discussion, re the very different 
numbers above.

I haven't caught up on all the threads here, but I believe at least 
Myrle's point was that items like Outreachy cannot stand alone if we are 
to achieve even a slight dent in our demographics. I do however believe 
that valuable insight could be learned, and then applied in a greater 
scope afterwards.

With regards,
Daniel.

Re: Outreachy won't (directly) solve all our diversity problems

[Sage Sharp <sh...@otter.technology>]

I would suggest taking a look at the OpenStack diversity reports:

https://superuser.openstack.org/articles/2018-gender-diversity-report/

"Newcomers who enter the community through a mentorship program boast a
higher retention rate than other newcomers, at 13 percent versus less than
10 percent, respectively. In addition, these mentees contribute to the
community for an average of 15 months, far longer than their three-month
program duration. This report provides a springboard for additional
research into mentorship programs, with the aim to identify and share best
practices."

One of the concerns raised was whether the ASF is "too big" for diversity
efforts to have an impact. It was mentioned that there are 8,000 ASF
contributors. OpenStack has 89,000 contributors. If they can improve their
diversity numbers, I'm sure ASF member projects can as well. :)

It will take several different programs and years of cultural change.
Having mentorship programs like Outreachy will be vital towards
understanding what changes need to be made. But to do nothing is to accept
the status quo.

(I'll also note that the Outreachy internships are 13 or 14 weeks long, not
6 weeks long.)

Sage Sharp
Outreachy Organizer

On Mon, Jun 24, 2019 at 8:57 AM Naomi S <no...@tumbolia.org> wrote:

> for convenience, my thread is here:
>
>
> https://lists.apache.org/thread.html/dbba6f4fd358737fddcb3da4b78df4f87fe384d59a35469a0d131545@%3Cdev.diversity.apache.org%3E
>
>
> On Mon, 24 Jun 2019 at 17:06, Myrle Krantz <my...@apache.org> wrote:
>
> > Hey all,
> >
> > You already know this, but I ran some example numbers, and I figured
> you'd
> > appreciate them.
> >
> > We're too big to use Outreachy to significantly impact our diversity
> > numbers.  Just to take one dimension of one segment of our committer
> > pipeline, and some simplified numbers as an example: if we have 8000
> > committers across all of our projects, and 5% are women, we'd need to add
> > 444 women committers to raise the percentage of women committers to 10%.
> > (8000*(.1 -.05)/(1 - .1))
> >
> > If all of the Outreachy interns were women, and 90% of the Outreachy
> > interns were to become project committers, we'd need to employ 494
> > Outreachy interns.
> >
> > Outreachy interns cost the org 6.5 thousand dollars, work for 6 weeks,
> and
> > require roughly 5 hours of mentoring per week.  *So raising participation
> > of women to 10% would cost more than 3.2 million dollars, and nearly 15
> > thousand mentor hours, if done solely via Outreachy.*
> >
> > Of course several of my assumptions are downright silly.  So, it would
> > likely cost much more.  But even that amount is currently well beyond our
> > reach.  And the participation of minorities in our projects is probably
> > worse than the participation of women, so continuing this calculation
> along
> > those dimensions is just going to get depressing.
> >
> > Direct impact on our diversity statistics can't be our goal in
> > participating in Outreachy.
> >
> > Instead, one of the hopes that have been expressed with respect to
> hosting
> > an Outreachy intern is that we'll be able to follow their progress and
> > learn from the problems they encounter (following the "Switch" model
> (1)),
> > so that we can figure out what's keeping people out of our communities
> and
> > tackle those underlying problems.
> >
> > Naomi's recent e-mail on this is super exciting.  I am really looking
> > forward to watching the work she's describing.
> >
> > Best Regards,
> > Myrle
> >
> > 1.)
> >
> https://www.amazon.com/Switch-change-things-when-hard-ebook/dp/B005TKD512/
> >
>

Re: Outreachy won't (directly) solve all our diversity problems

[Naomi S <no...@tumbolia.org>]

for convenience, my thread is here:

https://lists.apache.org/thread.html/dbba6f4fd358737fddcb3da4b78df4f87fe384d59a35469a0d131545@%3Cdev.diversity.apache.org%3E


On Mon, 24 Jun 2019 at 17:06, Myrle Krantz <my...@apache.org> wrote:

> Hey all,
>
> You already know this, but I ran some example numbers, and I figured you'd
> appreciate them.
>
> We're too big to use Outreachy to significantly impact our diversity
> numbers.  Just to take one dimension of one segment of our committer
> pipeline, and some simplified numbers as an example: if we have 8000
> committers across all of our projects, and 5% are women, we'd need to add
> 444 women committers to raise the percentage of women committers to 10%.
> (8000*(.1 -.05)/(1 - .1))
>
> If all of the Outreachy interns were women, and 90% of the Outreachy
> interns were to become project committers, we'd need to employ 494
> Outreachy interns.
>
> Outreachy interns cost the org 6.5 thousand dollars, work for 6 weeks, and
> require roughly 5 hours of mentoring per week.  *So raising participation
> of women to 10% would cost more than 3.2 million dollars, and nearly 15
> thousand mentor hours, if done solely via Outreachy.*
>
> Of course several of my assumptions are downright silly.  So, it would
> likely cost much more.  But even that amount is currently well beyond our
> reach.  And the participation of minorities in our projects is probably
> worse than the participation of women, so continuing this calculation along
> those dimensions is just going to get depressing.
>
> Direct impact on our diversity statistics can't be our goal in
> participating in Outreachy.
>
> Instead, one of the hopes that have been expressed with respect to hosting
> an Outreachy intern is that we'll be able to follow their progress and
> learn from the problems they encounter (following the "Switch" model (1)),
> so that we can figure out what's keeping people out of our communities and
> tackle those underlying problems.
>
> Naomi's recent e-mail on this is super exciting.  I am really looking
> forward to watching the work she's describing.
>
> Best Regards,
> Myrle
>
> 1.)
> https://www.amazon.com/Switch-change-things-when-hard-ebook/dp/B005TKD512/
>

Re: Outreachy won't (directly) solve all our diversity problems

[Ted Dunning <te...@gmail.com>]

I would also hope for some virality such that each intern brings some
number of others.



On Mon, Jun 24, 2019, 8:06 AM Myrle Krantz <my...@apache.org> wrote:

> Hey all,
>
> You already know this, but I ran some example numbers, and I figured you'd
> appreciate them.
>
> We're too big to use Outreachy to significantly impact our diversity
> numbers.  Just to take one dimension of one segment of our committer
> pipeline, and some simplified numbers as an example: if we have 8000
> committers across all of our projects, and 5% are women, we'd need to add
> 444 women committers to raise the percentage of women committers to 10%.
> (8000*(.1 -.05)/(1 - .1))
>
> If all of the Outreachy interns were women, and 90% of the Outreachy
> interns were to become project committers, we'd need to employ 494
> Outreachy interns.
>
> Outreachy interns cost the org 6.5 thousand dollars, work for 6 weeks, and
> require roughly 5 hours of mentoring per week.  *So raising participation
> of women to 10% would cost more than 3.2 million dollars, and nearly 15
> thousand mentor hours, if done solely via Outreachy.*
>
> Of course several of my assumptions are downright silly.  So, it would
> likely cost much more.  But even that amount is currently well beyond our
> reach.  And the participation of minorities in our projects is probably
> worse than the participation of women, so continuing this calculation along
> those dimensions is just going to get depressing.
>
> Direct impact on our diversity statistics can't be our goal in
> participating in Outreachy.
>
> Instead, one of the hopes that have been expressed with respect to hosting
> an Outreachy intern is that we'll be able to follow their progress and
> learn from the problems they encounter (following the "Switch" model (1)),
> so that we can figure out what's keeping people out of our communities and
> tackle those underlying problems.
>
> Naomi's recent e-mail on this is super exciting.  I am really looking
> forward to watching the work she's describing.
>
> Best Regards,
> Myrle
>
> 1.)
> https://www.amazon.com/Switch-change-things-when-hard-ebook/dp/B005TKD512/
>