While doing a literature search for another story I encountered a classic pitfall in scientific research that one can all too often find. I wanted to know how many venomous snakebites occur each year, and of those, how many are fatal.
I dug into the literature to see what was there. My first stop was Wikipedia. With experience, I have learned that for science issues, Wikipedia is often very good. The wiki format mirrors the scientific publication process where contributors double check facts and each other, and on subjects that I know something about I find it to be a good general source.
I found a statement that there are approximately 7,000 – 8,000 people bitten by venomous snakes each year, and of those about 5 die. Even better, this statistic was given a reference (Henkel ?), as well it should. So, being curious about the original research, I looked up the citation. And here is where the pitfall begins.
In order to conduct science and add to human knowledge we must necessarily build on the work of others. One of the characteristics of science is that results must be repeatable—that is, in theory I could redo the work of any other scientist and get similar results. It is impractical for me to begin every research project by completely re-doing the work of all those researchers before me, so we cite their work and trust in it. (This also explains why scientists can be vicious anytime someone is caught fudging results—we have to be able to trust each other). However, sometimes you encounter what I will call the “chain of citations” pitfall.
This pitfall is in the sloppy application of the scientific process, not with the process itself, and happens I guess out of laziness. It is where an author cites a point of fact from a research paper which may reference the fact (secondary source), but was not the paper where the research was presented in the first place (primary source).
I looked up Henkel and it turns out to be an interesting blurb from an appendix on safety, and he cites Gold et al. (2002), an article with the promising title “Bites of venomous snakes.” Ah, I think, here is the original research, so I look it up.
Those authors say “The true incidence of bites by venomous snakes in the United States is probably 7,000 to 8,000 per year, of which 5 or 6 result in death,” (pg 347) but they cite other authors for this fact, Langley and Morrow (1997).
This is getting silly, I think. How deep does this rabbit hole go? So I look up Langley and Morrow (1997).
The Langley and Morrow paper is a summary of deaths caused by animals of all kinds between 1979 through 1990. They compiled data from the US Department of Health and Human Services as published in Vital Statistics of the United States. They found that an average of 157 deaths occur each year as a result of injuries from animals, and about 60 of those came from venomous animals of all kinds, the majority being bees and wasps. They found an average of 5.5 deaths from venomous snakebites (Langley and Morrow 1997, table 2).
So, here at least we do have some original research on the number of fatalities each year. But how many bites? Langley and Morrow say “Approximately 45,000 snakebites occur each year, of which 8,000 are inflicted by venomous snakes,” (pg 12) and they cite Gold and Wingert 1994. Ok, let’s see what they say.
Gold and Wingert (1994) say “Approximately 45,000 snakebites occur in the United States each year. Poisonous snakes account for an estimated 8,000 of these bites, which result in approximately 9 to 15 fatalities,” (pg 579) and they cite Parrish (1966).
Finally, with Parrish (1966) we got to a paper that tries to determine not just the number of deaths from snakebites, but how many snakebites there are. Parrish (1966) conducted a survey of hospitals and physicians in both 1958 and 1959 to determine the number of snakebite victims treated. He sent questionnaires to a representative sample of both, and based upon the results, extrapolated to the total number of bites.
He wrote “On the basis of all these various reports, I estimate that approximately 6,680 persons in the United States (excluding Alaska and Hawaii [they don’t have venomous snakes]) were treated for poisonous snakebites during 1959,” (pg 272).
There are two numbers I cannot find in Parrish: 45,000 total snakebites and 8,000 venomous snakebites. I have no idea where these numbers originated since all the authors cite it, and all references point back to Parrish. Maybe the 45,000 number is a wild guess, and 8,000 is some adjustment to Parrish’s 6,680 number based on population growth? There is no way for me to know.
But this clearly shows the danger of relying on what the last guy reported and citing him as the authority when that is not what his paper was about. I tell my students to find the original publications and this is why. The number of deaths per year (about 5) is reasonably documented with the work of Langley and Morrow (1997), and in a more recent follow up study (Langley 2005). But, how many venomous snakebites occur in the United States in a year? Based on this evidence, we do not have a clue.
So here is a research tip, free of charge, to all you young budding scientists: go find the answer—just send me a copy of the results. Scour the literature to see if I missed something. Then, determine a good way to address the question and see if you can improve our understanding. The most recent data are over 40 years old, and inquiring minds want to know.
This post is in the dangerous animals series. Check out that post for more information.
Gold, B. S., and W. A. Wingert. 1994. Snake venom poisoning in the United States: A review of theapeutic practice. Southern Medical Journal 87(6):579-589.
Gold, B. S., R. C. Dart, and R. A. Barish. 2002. Bites of venomous snakes. New England Journal of Medicine 347(5):347-356.
Henkel, J.? For goodness snakes! Treating and preventing venomous bites. U.S.D.A. emergency response.
Langley, R. L. 2005. Animal-related fatalities in the United States–an update. Wilderness & Environmental Medicine 16:67-74.
Langley, R. L., and W. E. Morrow. 1997. Deaths resulting from animal attacks in the United States. Wilderness & Environmental Medicine 8(1):8-16.
Parrish, H. M. 1966. Incidence of treated snakebites in the United States. Public Health Report 81(3):269-276.Share This