January 1st, 2009
Somehow I never seem to have much to say about my research project, but that’s going to change over the next 6 months. I need to have the thesis finished by June 2009, which means that my priority for the first half of 2009 is going to be analysing the data and writing the thesis.
I was very pleased with the response to the online survey: I received 206 responses, from all over the world! Most of them look usable, and many more individual free/open source projects were represented than I expected, and one of the things I’ll do here is mention some of the ones that were new to me. The range of responses means that the data is more ‘random’ in some ways than I expected, which is probably a good thing. I’ll start summarising my findings here once I start working with the data.
One comment: the open source software I used to gather the data is from an abandoned project, and hindsight now tells me that I shouldn’t have relied on it, even though our systems administrator told me that other people had used it for big surveys. I used it for a shorter survey in 2007, and I thought that I understood its quirks well enough to use it for this one. Unfortunately, that wasn’t true, and I’ve spend the last two days converting the data from its rather idiosyncratic (and verbose) XML output to something I can work with more easily. The biggest challenge was the fact the it only output data for the questions that people had answered, so I had to identify the missing values before loading it into an anlysis tool. This was complicated by the order of the results, which were not sorted by question number, but were instead presented in the order I had defined them. But I have finally gotten the data in the form I want, through a combination of XSLT, a text editor, and a short php program. I’ve done enough double checking to be confident that everything’s there, and correct (whew!).
If I do something on this scale again, I’ll probably use LimeSurvey, which looks good and seems to have a very supportive community.