Dan Bolnick (on Eco-Evo-Evo-Eco) and Meg Duffy on Dynamic Ecology have both posted stories of how they were confronted with, and subsequently addressed, the need to retract or correct their published papers. This fall into the “scientists are humans; humans make mistakes; therefore scientists make mistakes” logical tenet, and they both addressed it wonderfully. Sadly, that’s not always the case.
So to further demonstrate that scientists, as all humans, make mistakes, here is my tale of finding a fairly significant error in my first ever paper.
In my undergrad, I spent a spring at the Point Lepreau Bird Observatory in southern New Brunswick. Yes, past the nuclear exclusion zone and next to a 19th-century lighthouse was a little hut with electricity, a portable heater, radio, and view of the majestic Bay of Fundy. As a one of the more southerly points in the area, it was also a hotspot for migratory birds, mostly seaducks, on their way north to breed. My job: figuring out how many scoters (3 species of mostly-dark seaducks: the Black, Surf, and White-winged Scoters) passed the site in April and early May. I was also generously allowed to analyse the previous 8 years’ data (and have since heard through the grapevine that a student may be updating this work soon!).
I was, at the time, terrible at data analysis and statistics. I had more pivot tables than I knew what to do with, graphs were made in Excel, and I think I used JMP for the various ANOVAs.
But I, and my supervisors, were able to churn out some basic stats on the timing of migration, the peaks, and come up with a crude estimate of how many birds passed the point each year. I would almost certainly analyse the data completely differently today (and I hope the aforementioned student does!). After some fairly minor revisions, it was published in 2007 in Waterbirds, and I was elated – my first publication!
One of the challenges was that the counts were done in 15-minute stints (15 on, 15 off), so in essence I had to double all the counts with the assumption that the number and composition of birds was identical in the counted and uncounted periods.
Except I forgot to do that.
I got an email from a member of one of the naturalist club’s members (the observatory was run by the Saint John Naturalists Club at the time) in 2009 pointing out that he thought my numbers were too low. I dug into the terribly formatted awkward files, and realised what we had done (or rather, not done).
I was devastated.
I immediately wrote my supervisors, contrite, and apologetic. We quickly prepared a correction (which essentially doubled the population, so not that insignificant), and emailed the editor who agreed a correction was in order, which we subsequently published.
Unlike Dan or Meg’s stories, this wasn’t a high profile paper, but it was my first one, and one of the very few for which I have a printed issue of the journal on my shelf. But everyone understood it was an honest mistake, and we did what we could to fix it.
I’ve opined before about why there are so few retractions or corrections in conservation biology/ecology, and I don’t see this changing, or being any different. But in the meantime, if anyone finds an error in any of my other published papers (I’m sure there are some floating around), I will happily try to set the scientific record straight.