This whole survey thing really exploded, huh?
I've caught up on a lot of the posts about it between classes and such, and have found it hilarious and infuriating by turns. The "slash is like dudes liking transsexuals" comment
was bewildering; the responses to it warmed the cockles of my heart. I think that what throws me the most is the staggering number of hugely broad assumptions they are working from already
- they are treating things like givens when they are not actually givens, and that anybody who is supposedly some kind of qualified researcher would not know better than that blows my mind.
I have seen it mentioned that the people behind the survey have implied that, because they were not intending any minors to be involved, they did not need IRB approval. (Although I did not actually follow any specific link to a comment containing that implication, so I have no first-hand knowledge.) If that is the case, then I would like to weep for all the time I wasted filling out paperwork just to get an expedited
review of my thesis project by UVM's IRB when no minors were going to be involved.
I'm not sure exactly what I was thinking when I first saw the banners for the survey going around; I've seen that kind of thing before for people's smaller statistics projects and such, and I guess I never really realized that this was at least trying to be an actual official research project. That being the case, the problematic survey items take on a whole new meaning. We spent ages
in Measurement of Communication Processes learning about just how careful you have to be about everything
when you're making a good survey, including the wording of the items and the various options for the response setup. We had to run our project surveys on a pilot sample and test the validity of the items and chuck the really awful ones and rephrase the only mildly terrible ones and argue for, like, a day and a half about whether we had really picked the right response setup. And that was just for a class project
, not for something we were planning on turning into an actual research tool.
Other people have talked quite a bit about the myriad other problems, so: linkspam roundup
. I would be so lost by these blowups if it weren't for these roundups.
... And now I have to go to Human Cultures. Sooner or later I will put up a riveting account of my Tuesday, though.
ETA: Obviously I should have checked my e-mail before I left, or I would have noticed I had a comment (on LJ, I mean) before this posted, but somehow my last post about the survey actually ended up in the linkspam
. It's like I have some kind of actual presence on the internet or something. o.O
ETA 2: Also, something so basic I forgot to even mention it: every statistics class I have ever taken has hammered home the point that convenience samples are almost totally worthless. Obviously this is a situation where you are aiming to get a particular population, and so a convenience sample, as long as it's taken from that target population, is perhaps slightly
less useless than usual, but.