Photo: Survey Crew by Rusty Clark via flickr.com, CC BY 2.0
Last week I asked you how much time you spent writing. In my writing book I suggest that most academics spend a large fraction of their time writing. I’d been surprised to hear a few people argue the opposite – and to find this estimate of 2% to apparently back that up. Hence last week’s post, in which I gave a few reasons for doubting that 2% number; but had to admit I didn’t have any real backup for my claim that most of us spend a lot of time writing. So, I included a completely, 100% unscientific poll, and I promised I’d come back with the results.
So here they are. There’s some variance, but a substantial majority of you believe you spend at least 25% of your work time writing; and a decent fraction believe it’s over 50%. I’m not surprised by this, although I don’t think we can quite take the numbers at face value either.
For one thing, and rather obviously, it’s an internet poll. I didn’t recruit respondents randomly, or stratify them by likely correlates like career stage or field. Much worse: I primed responses by stating my own opinion first, and arguing against the contrasting 2% estimate. Now, I didn’t claim I was doing great science, but there’s an important point here: survey data are hard. There are a million ways to trip up, and I didn’t try to avoid them because the stakes were low. When the stakes are higher (because we actually want to interpret survey data), we should acknowledge that as scientists we tend not to be very savvy about surveys, and we should consult those who do have that expertise*.
It also seems pretty likely that my survey is measuring something different from the 2% survey. Remember that I asked respondents to define writing fairly broadly, including not just papers but grants, reports, and the like, and including revision, figure drafting, as well as the production of new text. The definition of “writing” as an activity in the 2% survey isn’t entirely clear to me, but it’s certainly narrower (referring, for instance, only to “papers”). The populations surveyed are different too. I could go on about all this, but we’d both get bored.
So: I don’t claim to have measured how much time scientists spend writing, but I do think we have (weak) evidence that many scientists perceive that it’s a lot. This makes me feel better about the claim in my book, and about my own perception of my writing time (I’d be in the 25-50% bar). It also suggests that a really large, well-designed activity-budget study would be worthwhile. Anyone need a PhD thesis idea in science studies?
As for me: I have some (other) writing I should be doing.
© Stephen Heard (email@example.com) June 30, 2016
*^Typically, this expertise is available from colleagues in the social sciences, for whom survey data are extremely important. I’m involved in a project using survey data to understand public attitudes to forest pest insects and their suppression, and it’s clear that it has to be driven (as it is) by social scientists. My “hard”-science colleagues and I simply do not know how to run a survey or interpret its results.