Record honestly, interpret later

It’s hard to be a good machine. Raw material comes in, wheels turn, and you have to pop out an identical processed product, over and over again. For all of our amazing strengths as people, we’re not good at all doing exactly the same thing in exactly the same way over and over again. We think too much, we get distracted, and different people inevitably interpret information in different ways. These can be problems when counting the wild.

The solution is to minimize the amount on interpretation done during a count. Don’t think, just count. This can be done through both choosing a well-designed counting method, and in the way you use the method. In essence, you want to always record honestly and interpret later. Count exactly what you see and hear, and push all interpretation of what it means until after the counts have been made.

Things get messy when you are asked to record your interpretation of what you see or hear, rather than what you actually see or hear. For example, one common method in NZ requires that you estimate an exact count of how many individual birds of a species you saw in five minutes. That takes a lot of thought and interpretation of what you’re seeing and hearing.

In another method, you need to submit an exact list of all the bird species you saw in an hour, with no way of entering the identifications you weren’t certain of. It requires that you make your best guess for things you’re not sure about.

At best, such interpretations create noise in the count data and make it more difficult to reveal patterns and trends. At worst, spurious patterns and trends are created when the interpretations people make differ in different communities, different habitats, or at different times.

I’ve written before about the importance of not guessing, and of counting in minimum-maximum ranges rather than absolutes. If you think a bird you glimpsed flying over looked like a starling, the correct thing to record is starling?,and definitely not starling. If you heard what you were sure were starlings, and thought there were three or perhaps four birds calling, then the correct thing to enter is 3–4 or ~3, not 3 and not 4.

Too often I see the efforts made to simplify scientific methods for citizen science projects resulting in more interpretation being pushed onto the counter. It needs to be the opposite of this. Less interpretation and more honest counting can lead to more complicated instructions and data entry forms. However, I find that the counts themselves then become much easier to do. Without the interpretation, you can become a counting machine recording the moment. That makes for better data, and data that are ripe for interpretation later.