[Osmf-talk] OSM science

[john whelan]

Trust is the most important thing in surveys.  To get accurate surveys you
need to select the target group with care to get a random sample and then
you need to think about the response rate.  Running the same survey in
Canada and the US the response rate was above 90% when Stats Canada ran it
in Canada.  In the US it was done by a private company and the response
rate was around 60%.

The problem is those missing 30% are probably the most interesting ones and
it takes away from the authority of the results.

So if you're an academic do you want to base your PhD on a survey with low
response rates or would you prefer to run your own?

The software choice is basically irrelevant.

The first step has to be to talk to the academics and see how they would
like to see a random sample generated and what safeguards they can come up
with to get the response rate to a reasonable standard.

Have you thought about languages by the way or that in many parts of the
world governments are not trusted?

I'm not sure if the above can be addressed but it might be worth thinking
about in the planning stages.

Cheerio John

On 10 Jan 2018 12:40 pm, "Clifford Snow" <clifford at snowandsnow.us> wrote:

> Joost,
> I like your suggestion. I've use Lime Survey in the past. It is a solid
> package. It's easy to create and modify questions in different formats. The
> web interface is easy for users to understand.
>
> Setting it up on a linux machine isn't too hard. I've done it once on an
> AWS ubuntu instance, although I'd also be happy to defer to someone with
> more experience.
>
> Let me know how I can help.
>
> Clifford
>
> On Wed, Jan 10, 2018 at 7:21 AM, joost schouppe <joost.schouppe at gmail.com>
> wrote:
>
>> Hi,
>>
>> There has been some talk on surveys and the like in the OSM ecosystem,
>> and it was mentioned that it might not be a problem to set up a survey
>> engine ourselves. This would make it easier to survey different kinds of
>> mappers, as well as have some sort of common ground between researchers and
>> OSM, and make it possible to pool questions about similar subjects so as to
>> avoid over-surveying people.
>>
>> So I would like to formally propose we do that. I'm a big fan of the open
>> source Limesurvey package. I've only used it as a user, but if I can get an
>> OSMF machine, I'd be willing to see if I can set it up. But I would be more
>> than happy to leave it to more experienced people.
>>
>> To discuss all the stuff related to this, I just asked for the creation
>> af a science mailing list, as described on the mialing lists wiki page.
>>
>> Last, maybe we need a separate working group to handle the basic
>> infrastructure. Some of the data on that server will be privacy-sensitive.
>> It is perfectly possible to have different roles for different people, so
>> it would only need to be one or two people who need access to the sensitive
>> stuff. A similar non-disclosure agreement to the one used in the MWG might
>> be enough.
>>
>> --
>> Joost Schouppe
>> OpenStreetMap <http://www.openstreetmap.org/user/joost%20schouppe/> |
>> Twitter <https://twitter.com/joostjakob> | LinkedIn
>> <https://www.linkedin.com/pub/joost-schouppe/48/939/603> | Meetup
>> <http://www.meetup.com/OpenStreetMap-Belgium/members/97979802/>
>>
>> _______________________________________________
>> osmf-talk mailing list
>> osmf-talk at openstreetmap.org
>> https://lists.openstreetmap.org/listinfo/osmf-talk
>>
>>
>
>
> --
> @osm_seattle
> osm_seattle.snowandsnow.us
> OpenStreetMap: Maps with a human touch
>
> _______________________________________________
> osmf-talk mailing list
> osmf-talk at openstreetmap.org
> https://lists.openstreetmap.org/listinfo/osmf-talk
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/osmf-talk/attachments/20180110/2843caa0/attachment.html>