[OpenStack Foundation] [Diversity] re: Diversity Workgroup APAC 2015-08-27

Rochelle Grober rochelle.grober at huawei.com
Sat Sep 12 01:44:17 UTC 2015

Thank you Roland, for doing this.

The current discussion got me thinking about some of the underlying issues that categories like religion and geolocation encompass.

One of those underlying issues is calendar based availabity.  Perhaps we should ask questions around this specifically.  Here is what I mean:

Observant Jews are not available from Friday sunset to Saturday sunset
Observant Muslims are not available from Friday sunrise.
Observant Christians are not available Christmas or Easter, plus....you get the idea.

Chinese have holidays around both traditional religious holidays and various state holidays but also usually work the last Saturday of the month.
Most countries have Independence Days or special days for their hereditary leaders or celebrations of various winning battles.

All of these sorts of calendar restrictions feed into subtle or not so subtle discrimination, especially when significant community events are scheduled over significant holidays.

Food for thought and thought for whether something along these lines should be part of the current or future survey.


From: Roland Chan [mailto:roland at aptira.com]
Sent: Friday, September 11, 2015 4:59 AM
To: Eoghan Glynn; Geoff O'Callaghan
Cc: Foundation at lists.openstack.org
Subject: Re: [OpenStack Foundation] [Diversity] re: Diversity Workgroup APAC 2015-08-27

Regarding the risk to decision quality:

The diversity working group will be tabling a report of the survey results at the board meeting in Tokyo. This report will be freely available afterwards. Response rates are definitely the biggest risk to the validity of the data, and indeed our entire ability to measure success, and this will be reported upon.

Interpretation of the data will occur in public in the meetings of the working group and on this list. To the extent conclusions are be drawn, they will be drawn in public. Similarly decisions on the execution of diversity programs will be made in the open and all comments on those are welcome.

I agree that there is risk from low response rates. I disagree that this increases the risk of poor decision being made in the future because we are cognisant of the risk. Having said that a diverse range of views can only increase decision quality and I welcome criticism.

Regarding the scope of the survey:

Originally, I was thinking along similar lines, that we should survey only the first phase priorities. However, over time I changed my views. Firstly, and least importantly, a single survey is more efficient.  Secondly, if we are to establish a baseline, it behooves us to establish it across all the areas we have identified are being of concern. The data may show something unexpected and we should alter our actions to suit.


On Fri, 11 Sep 2015 9:17 PM Eoghan Glynn <eglynn at redhat.com<mailto:eglynn at redhat.com>> wrote:

> Hi all
> Is it just me or does anyone else find this a bit disturbing.  Surely the
> goal is to structure the openstack community to be inclusive rather than
> spending time prying into people's unrelated private lives so we can use
> the data as part of a science experiment.
> My religion or sexuality has nothing to do with openstack.

Hi Geoff,

I sympathize somewhat with your view. I argued up-thread for removing
the question on religion, and FWIW I think the survey should concentrate
on a small number of characteristics that are both innate to the person
and also outwardly visible, such as:

 * gender presentation
 * race
 * geographical location

IMO people are unlikely to object to sharing such information about
themselves, since it's quite obviously apparent anyway in their daily

Also these are the kinds of characteristics on which discrimination/
exclusion, whether conscious or unconscious, is most likely to be based
(because these characteristics are so outwardly visible).

I'd be a bit concerned about interpreting the data if we include more
intrusive/personal questions *and* many of respondents decline to answer
those specific questions, while answering other questions. There's a
non-response rate above which those data become unreliable (i.e. we
can't necessarily assume that the data gathered from the respondents
are suitable for accurately extrapolating across the community, as the
non-respondents may skew disproportionately one way or the other, since
they're essentially a self-selected sub-group).


Foundation mailing list
Foundation at lists.openstack.org<mailto:Foundation at lists.openstack.org>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/foundation/attachments/20150912/f02a3a36/attachment.html>

More information about the Foundation mailing list