A consensus analysis

310 views

I saw this in the context of a team of people calling out another team saying they outlined a consensus analysis and I don’t understand what it is.

In: Biology

It’s basically when you have a bunch of different answers to the same question (because the question is hard, because people have different opinions, because some of the data might be contradictory, because it requires a lot of judgement, etc.) and so you present the “average” answer as *the* answer.

To make a really trivial example…suppose I flip a coin. Half the observers say it came up heads, half say it came up tails. Obviously, it either came up heads or tails and half of the observers are wrong. The consensus analysis would be “the coin came up half heads and half tails”. Which illustrates the potential problem with consensus analysis…you can get an answer that makes no sense if you’re not careful. At least “heads” or “tails” *could* be right…the consensus analysis *can’t* be right.

There is a place for consensus analysis…when you’re doing things with no rigorous process that don’t have verifiable answers at the time, some kind of “average” (in quotes because it doesn’t have to be the literal average) of the answers from several people who ought to know what they’re doing can be better than any one answer because it smooths out outliers. Consensus analysis is routinely used in the business world for forecasting future demand.