Unintentional Sampling Bias

I am reading Thomas D. Seeley’s book Honeybee Democracy, which explains the collective wisdom and effective decision-making of the hardest working insect known to man: honeybees. I could write an extremely long post about how amazing honeybees are (don’t get me started!), but I wanted to highlight a great example of unintentional sampling bias detailed by Seeley himself.

In 1975, Seeley worked as a research assistant at the Dyce Laboratory for Honey Bee Studies at Cornell in Ithaca, NY. Under the guidance of Professor Roger A. Morse, Seeley focused his research on determining the ideal honeybee hive, based on observation of existing wild beehives found in the local Ithaca area. Seeley put an ad in the local newspaper, The Ithaca Journal, asking community members to contact him if they saw a tree housing a live colony of honeybees. Within one week, he had secured the rights to 18 accessible bee trees in the woods around Ithaca. He ultimately collected and sampled 21 bee tree nests by killing the hive with calcium cyanide powder, cutting down the tree, and carefully dissecting the dead hive. He also located another 18 nests in trees to gather information about hive openings.

So what does this all have to do with research methods? Well, Seeley observed trends towards small size, floor level, southern orientaton indicating elements that were favorable to honeybees. But what perplexed him was the preponderance of nest entrances that were just a few feet from the ground. Being that close to the ground makes honeybee nests vulnerable to detection by predators, such as bears, whose attacks can be fatal to the hive. But, as Seeley later determined, honeybees actually have a strong preference for nesting cavities¬† with entrances located high above the ground. Seely’s initial findings were affected by unintentional sampling bias. The nests that he sampled were noticed inadvertently by a person walking past a bee tree, people are more likely to notice honeybees trafficking from a ground-level nest entrance than a tree-top one. Seeley determined this when he learned to “line” bees (locating bee trees by baiting foragers from flowers and observing their flights back to their nests). He found that every hunt ended with him straining to spy the bees zipping in and out of a nest entrance high in a tree. He has since determined that the average height of a honey bee nest entrance is 6.5 meters (21 feet). Seeley reports in his book: “Needless to say, I’m now alert to the hidden danger of unintentional sampling bias” (p. 52).

This is an excellent illustration of sampling bias, and why we need to be aware of its effects in our own research.

Leave a Reply

Blog authors are solely responsible for the content of the blogs listed in the directory. Neither the content of these blogs, nor the links to other web sites, are screened, approved, reviewed or endorsed by McGill University. The text and other material on these blogs are the opinion of the specific author and are not statements of advice, opinion, or information of McGill.