Religion and computer language use survey results.
February 16, 2008
I’ve put up a first pass at visualizing the data from the survey I put up last month about computer language use and religion. This was an attempt to see if there are any correlations between languages people *prefer* to use and religious identity. The results are available to be viewed at http://www.kimsal.com/reldevsurvey/results.php.
There’s a few things to note:
- I didn’t deal with the country data. I’m not sure how to properly visualize or account for it.
- The options changed during the first few hours, and eventually I had an ‘add your own’ option. I should have put the ‘added’ options in the available options for future people, as that would have reduced repeats with slightly different misspellings.
- The raw data is available to download at the link provided above. If anyone cares to take this data and visualize it better, give me your results and I’ll link to them or post them.
I don’t (yet?) have just raw results by religion, but out of the 3814 results, something like 1800+ identified as “Atheist”, and another 700 for ‘Agnostic’. Clearly this is a big bunch, and is higher than I anticipated. 1000 of those Atheist and Agnostics claimed “US” as their country, which again surprised me a bit.
Any feedback/comments, feel free to discuss here or email me privately - mgkimsal@gmail.com.
Did you like this post? Buy me a hot chocolate!
Posted in 
February 16th, 2008 at 4:34 pm
The link is broken — the “v” “survey” is a second “y”.
February 16th, 2008 at 4:37 pm
Fixed - thanks!
February 16th, 2008 at 5:23 pm
I think the most interesting trend your survey results show is the number of developers who say they are athiest or agnostic. Software developers are typically pretty intelligent rational people so perhaps we can draw a correlation between being intelligent and not believing in a silly religion.
February 16th, 2008 at 9:23 pm
Shouldn’t the numbers be based on percent of respondents, and not raw number?
February 16th, 2008 at 9:34 pm
As opposed to believing in a non-silly religion?
February 16th, 2008 at 9:59 pm
Alan:
They probably should. This was just a ‘first pass’ at getting something out there. As I wrote before, there’s nothing in here that is specifically dealing with filtering/reporting by country either. I will likely get around to redoing the reporting to include those sorts of things (including the percentages). In the meantime, the small graphs now do relative percentages for the top results (up to the top 10), so those pie chart sizes are relative to each other based on percents. The data is there if you’d care to pull it down and run some charts via OpenOffice or Excel or whatever (or give it to someone who loves number crunching eight ways from Sunday).
Cheers.
February 17th, 2008 at 8:38 am
Would it be possible to show a graph of *all* languages vs. religions chosen? A quick glance at the “big” languages makes me think it’d be ~45% atheist, but I’m curious all the same.
February 17th, 2008 at 8:43 am
It’s possible, but I’m not sure how to do it. You can see just all the religions chosen and all the religions chosen by looking at the pulldowns. Getting a number from them is probably something easiest done in a spreadsheet. The data’s there to plug away at, but I’m not sure I can have anything proper in the near future. I plan to look at this in more detail in the next couple weeks, time permitting.
February 17th, 2008 at 9:57 am
This was interesting. Thanks for making this data available.
I agree with chris - the only trend that I could immediately see was that Athiests / Agnostics made up well over 50% in every language that I checked.
I assume that you’re going to spend some time looking for further correlations (or complete lack thereof).
Finally, I would be interested to see the results minus languages / religions where there were only 1 respondant. A lot of the religions, and some of the languages, were obviously not serious - Jedi Knight, Pastafarian, occultist, etc. (although it’s not always easy to tell with religion). And in the cases where there were only 1 respondent, removing them would cut out a lot of noise without dramatically affecting any potential correlations.
Matt
February 17th, 2008 at 10:21 am
I just found this post linking here with a touch more analysis on the top languages and beliefs: http://blog.mobocracy.net/2008/02/beliefs-and-programming.html Thanks Blake!
February 17th, 2008 at 10:13 pm
Good point Ted. Since we’re talking religion I didn’t really need to throw silly in there. That’s a given.
February 22nd, 2008 at 1:19 pm
Where was this survey taken? If it was taken from subscribers to your site it could be that a large majority of your subscribers are Atheist, tainting the results.
If it wasn’t a random sample, and rather whoever you could get to come take the survey in the time frame allowed, then I’m not sure you can even draw conclusions based on it.
February 22nd, 2008 at 1:26 pm
Hi Tim:
The link to the survey was posted to people following me on Twitter (about 80 people or so), to my blog which gets fed in to some other aggregators, and to reddit.com. I’m not sure where else it may have been reposted. I can say that the large majority of interest came from reddit.com readers.
Wouldn’t a ‘random sample’ also happen to be people would I could get to take the survey in a time frame too?
I consider it to be about as random as one could reasonably expect to get without spending money on organizing something more controlled.
You can draw whatever conclusions you want to from it. I wasn’t able to find any major parallels between languages and religion like I thought I might, but having the majority of respondents identify as atheist or agnostic was interesting. It doesn’t indicate is atheism causes one to become a developer, or vice versa, but was still interesting to me and others.
February 22nd, 2008 at 1:27 pm
@Tim - thanks for the input, by the way