On 8/19/25 19:48, Melissa Logan wrote:
The survey works well IMO - intuitive, simple, fast.
Does it allow us to slice and dice the data on the backend in different
ways? For example, if we want to understand how a certain cohort (e.g.
geography, age, gender) answered a particular question/set of questions, do
we have the ability to do this? It can help immensely with interpreting
data.
Yes, we can can filter/funnel in many ways. The most intuitive for a
birds eye view is done by simply clicking on the answers we want to dig
into, and the survey results will automatically filter to only show
respondents that picked that option (one can apply multiple filters).
We can also add business logic to the survey, meaning if someone for
instance ticked the "I am a professional developer" box in a survey, we
can route them to a special set of questions about OSS employers,
whereas if they ticked "I only work on Open Source as a hobby", we could
have them automatically skip that part.
If you'd like to look at the interactive view for survey results, I can
set up an account for you in the system, so you can take a look behind
the scenes.
On Fri, Aug 15, 2025 at 7:16 AM Jarek Potiuk <ja...@potiuk.com> wrote:
I think, as a baseline, we should be working with multiple teams prior
to any survey being published:
Very good idea indeed. Better safe than sorry.
It might make sense for us to have a privacy policy specifically for our
surveys in general
I think that would be a good idea for Niall, Christian and Bolke to chime
in on that one - whether to have a separate policy or whether to fold it in
the general policy. I know Christian, Niall and Bolke are working towards
making our privacy policy "official" and "approved by board" [1] - and one
of the important things raised by Christian in the discussion was that we
need to explicitly list (provide catalogue of) all the tools we are using
to collect potentially private data. We discuss Matomo (and the possibility
of gathering "product usage" information) in this thread but I think this
is fairly relevant as well. I linked directly to the message from
Christian, but the whole thread is interesting to read in this context I
think.
[1] The "Matomo thread "
https://lists.apache.org/thread/x4kjt81o1kxcy3wn79h25ghvsskgh912
I do not want to dominate the discussion here, so I will shut up, but I
thought it's worth mentioning.
J.
On Fri, Aug 15, 2025 at 1:43 PM Daniel Gruno <humbed...@apache.org> wrote:
On 8/15/25 13:30, Jarek Potiuk wrote:
The email one though is connected with keeping PII though ? (Personally
Identifiable Information) - so I guess that one would require at least
some
discussion with the privacy team ?
I think, as a baseline, we should be working with multiple teams prior
to any survey being published:
- privacy team for data privacy sanity checks
- M&P for messaging and promotion if/when needed
- Projects that have a significant interest in whatever survey is being
assembled.
It might make sense for us to have a privacy policy specifically for our
surveys in general, which we can then refer to at the beginning of the
survey - preferably before any data is entered. This policy could also
be a guideline for us on how to handle the data. For instance, we could
-- or rather, I think we should/must -- delete or otherwise obfuscate
any PII in surveys no later than 60 days after the survey has been
closed for submissions.
Let me know what people are thinking with regards to that, and I can
start putting together a privacy policy document for us in our wiki[1].
With regards,
Daniel.
[1] https://cwiki.apache.org/confluence/display/EDI/
On Fri, Aug 15, 2025 at 1:25 PM Daniel Gruno <humbed...@apache.org>
wrote:
On 8/15/25 13:14, Jarek Potiuk wrote:
One question - I understand the survey is anonymous, but is there a
way
to
distinguish several answers from the same person (or at the very
least
browser, due to anonymity). Just thinking of some "gaming" scenarios
-
ASF
people are smart and might get the idea to increase chances of their
choice
by responding to the survey several times (I just did with the low
rating
for poor pony choices ;) ) .
Great question, and thankfully one that I already asked myself as
well,
so I have some answers for you.
For surveys aimed at either a set group of people (like committers,
members, etc), we can make use of closed, individual links. We can
pre-generate these in advance to ensure only one submission per
person.
For wider, open audiences, we can use email verification to at least
limit responses to one per email address. This would require anyone
wanting to submit a survey to first verify their email address with
the
system through a callback link sent to their email address.
With the latter you could presumably also filter out obvious attempts
at
gaming any sort of survey, though I would hope we don't have to do
that.
On Fri, Aug 15, 2025 at 10:27 AM Daniel Gruno <humbed...@apache.org>
wrote:
On 8/15/25 09:53, Jarek Potiuk wrote:
This tool is great. Works flawlessly. But I think the choice of
ponies
was
very limited, so you should try harder :)
That's why we have the "how did we do?" questions at the end :-D
J.
On Fri, Aug 15, 2025 at 9:23 AM Daniel Gruno <humbed...@apache.org
wrote:
Hello again, wonderful D&I folks.
I have been experimenting with a self-hosted tool for conducting
surveys
and collecting feedback/reviews, and things are looking pretty
decent.
I'd love to get some feedback from the people on this list, so
that
we
can perhaps decide on using this tool or not and discuss some
basic
premises for conducting surveys in the future.
You can try a survey demo here:
https://surveys.diversity.apache.org/s/wcee4v8ygw8q0pih0nflnxl1
Those of you on the D&I committee[1] who wants to try their hands
at
the
behind-the-scenes work with surveys can let me know (private or
this
list), and I'll set you up with access to manage surveys
yourselves.
This will also allow you to see the results of surveys in various
ways.
I have built a template for our surveys that comes with a standard
introduction and a data consent form at the very end; you should
be
able
to see that in the survey demo I linked to. I had the wording in
the
template refined by our resident "editor-in-chief", Andrew
Wetmore,
but
things can always be improved upon further, so do not hesitate to
suggest changes.
There isn't a whole lot more to say, take the survey for a spin
and
see
if this survey tool fits the brief.
With regards,
Daniel.
[1] Access to survey administration is restricted to members of
the
D&I
committee for data privacy reasons; this is explained in more
detail
in
the survey itself.
