I was pretty scathing about CASA’s Survey Mapper in my “Just because you can put something on a map…” talk at W3G.
When I met up with Andy Hudson-Smith after the talk in Stratford we had a very good conversation about Survey Mapper, he made a good case for putting the tools into peoples’ hands and seeing what they could do with them and I argued that self selecting polls on topics that may not have a significant spatial distribution or controlled sample might be misleading or pointless. I think we both ended up agreeing with each other.
It turned out that the poll I had been most critical about in my talk, Public Perceptions of Crime, was not a real survey, they just fed in some test data to check that the system was working and generating maps etc! So not surprisingly the perceptions bear no comparison with real crime statistics, unfortunately there was no indication of this on the survey page. I suppose my criticism of that survey was valid, the results portrayed were misleading (and they did get quoted around twitter a bit, so they could have got picked up by a journo and gained undue publicity) but to be fair to Andy and Survey Mapper the survey was a beta test and not intended to be an example of citizen science.
So I thought I would try out Survey Mapper and set up a survey to see whether it is possible to get any useful results and understanding from the tool. I have launched a survey about the budget cuts that will be announced in the Comprehensive Spending Review on October 20th. The survey is pretty simple because I was constrained by the toolset. I wanted to ask a series of questions but I was limited to only one and then I could only use three multiple choice options as answers. I tried to list a series of statements that might resonate with peoples’ opinions but when I published the survey some of the lengthy text strings were cut short, unfortunately once published I couldn’t find a way to edit the survey (a bug or two there that no doubt they will soon fix).
My survey is called CutsMap, for good or bad it’s out here now. Please try it out, please retweet it, Facebook it or socialise in any other way you can. If you happen to work for the Guardian, BBC, Telegraph or any other paper (even though it pains me the Daily Mail) and you feel like blogging it that would be great.
I guess we need at least 25,000 responses to have anything approaching a meaningful sample (any geodemographic expert want to comment on that?) I hope that I can convince someone to do some analysis of the responses (assuming we get enough) to try to correlate against income, age etc. Offers of help will be gratefully accepted.
Thanks to Andy Hudson-Smith for charming me into rethinking my cynicism and giving Survey Mapper a try.
More thoughts as we see whether CutsMap gets any significant response and whether there are any meaningful patterns.
6 thoughts on “CutsMap – another useless survey?”
Steven – neat idea. If you need a geodem profile of the postcodes, more than happy to help out using ACORN. Only thing to be wary of – you’ll really need the full postcode of the respondent to do anything meaningful with the data. You might ask the guys at Survey Mapper to amend this part of their app? Postcode district (which is all it asks respondent to provide) will provide extremely broad-brush analysis.
For good, robust, surveys you really need a minimum of 1,200 people to get anything meaningful out. So you’ve still got a little way to go (!) – have blogged the link (www.mappingpeople.blogspot.com), so you may get one or two more…
Paul – Thanks for the offer of help on the ACORN analysis. Although the postcode field only requires the district it will accept and store the full postcode, it would be better of course if it stated that.
As you say I have a long way to go to get a meaningful sample, which is one of the problems with a Citizen Survey unless the BBC or someone with reach publicises. However I hope that on Oct 21st there may be an opportunity to republicise and get a flurry of responses after the CSR has been announced.
25,000 responses! , you need to use the standards of the cosmetics industry, 80% of the 40 people interviewed…
Nope it’s not chrome, it’s my fault. I entered too long text strings and didn’t foresee the way they would display. There isn’t a way to edit the survey once published that I can find.
This is all part of the learning process I guess. Hopefully you can get the sense of the different opinions and choose which one best matches your own opinion.
Hmm, when I just looked at the survey now, the ends of the questions had been cut off! I don’t know if that’s a subliminal thing (or just because I’m using google chrome…)
The nice guys at Survey Mapper are helping out and should be able to correct some of the glitches in CutsMap. Come back soon to try again.