[Talk-GB] [Osmf-talk] Live OSM discussion in ~45 minutes (7.30pm UK time)

Simon Poole simon at poole.ch
Thu Jul 27 09:33:52 UTC 2017


The Wikipedia article on the topic:
https://en.wikipedia.org/wiki/Snowball_sampling

It would just seem to be far too easy for advocacy groups to hijack even
a well intentioned selection, which is what I suspect we are seeing
here. Instead of getting a diversity of viewpoints we are simply getting
one, that will then in turn be referenced by the same groups as a result
of "research".

Simon


On 27.07.2017 10:06, Dan S wrote:
> 2017-07-27 8:57 GMT+01:00 Simon Poole <simon at poole.ch>:
>> PS: you do illustrate an interesting point wrt the research we are
>> discussing here, one would expect an unbiased sample of OSM contributors
>> of size 15 to contain zero non-male and zero HOT members obviously the
>> composition of the interviewees was rather different.
> A uniform random sample is not really relevant in small-numbers
> sociological research. This is "snowball sampling" which is a very
> problematic form of sampling, but one thing it definitely does not
> claim is to be a balanced random sample. A social researcher using
> snowball sampling to find a small number of interviewees knows full
> well they aren't gathering a statistically random sample, and one
> hopes that this researcher tried to find a diversity of viewpoints for
> his interviews (I don't know). He could, for example, have
> deliberately designed his sample to have almost-equal numbers of men
> and women.
>
> Best
> Dan




More information about the Talk-GB mailing list