Pest Patrol: Dealing with bias in pest management research trials

#PestPatrol with Mike Cowbrough, OMAFRA

I have biases.

Perhaps that’s a surprising statement coming from someone who is supposed to be impartial, but experiences lead to biases. It’s human nature.

The bigger question is how does one account for and neutralize those biases when evaluating something such as a weed control program? That’s the topic that I wanted to tackle with the Ontario Agriculture College’s (OAC) Soil and Crop Club this past fall.

Let’s begin with the premiss that if one truly had no biases, then when evaluating control of weeds with herbicides, the value given for “per cent visible control” would be identical whether or not the treatment name was known at the time of evaluation. But what if you know the name of the treatment that you are evaluating? Would past experiences influence how you would “score” or rate the treatment?

Related Articles

I think it most certainly would, and to demonstrate how bias and prejudice can influence evaluations, I showed students three brands of kettle chips: Miss Vickie’s, President’s Choice and Cade Code (GMO free). The students were asked to do two things. First, identify which of the three brands they like the best. Second, using a blind taste test, identify the best-tasting sample, which in theory should be identical to the brand identified in the first question. The results of this exercise are shown in Table 1 below.

Although all students chose Miss Vickie’s as the superior-tasting chip when asked in advance, only half of them identified it as their favourite during the blind tasting. This exercise with the OAC Soils and Crops Club certainly demonstrates that preconception can influence evaluation. So how do you overcome that?

When I’m involved in weed control trials, I ask someone else to randomize the list of treatments so that I don’t know what plot number corresponds with a particular herbicide. I then only evaluate the numbered plot and it is not until I’m done evaluating the entire trial that I will look up the name of the treatment. If I was more disciplined, I would wait until all evaluations were done throughout the season, but human nature creeps in and I get excited when I see things work really well and I just want to know what it is.

Although it is an asset to have experience, we must recognize that it can lead to preconceptions which can affect the fairness of an evaluation. This is understandable, but failing to recognize this and take the steps necessary to neutralize the bias is unacceptable.

Have a question you want answered? Hashtag #PestPatrol on Twitter to @cowbrough or email Mike at [email protected].

About the author



Stories from our other publications