Here are a few tricks you can use to “win” a debate with a statistician.
- While rarely saintly and other worldly, statisticians are educated to listen to other points of view and to change their opinion if logic and evidence so dictate. This is their Achilles’ heel. Go for it.
- Present yourself as very self-assured. Use vocabulary, voice tone and body language designed to intimidate. This works against anybody, but will be even more effective when debating statisticians since they have been educated to question their own points of view.
- Statisticians will be expecting a logical case, so use one that makes absolutely no sense. This will confuse them and give you enough time to make another nonsensical point. And then another. Arguments that refute themselves are especially useful since statisticians are trained to avoid making them.
- Statisticians learn to think in terms of conditional probabilities and to consider many variables simultaneously. This takes time, so talk as fast as you can. Don’t give them time to think.
- Statisticians will be expecting data marshaled in support of your case. “New” data your opponent is unaware of will be most effective in any debate and, if you contrive some numbers on the spot, you can be guaranteed they won’t have heard of them.
- Alternatively, confidently cite figures that are irrelevant to your argument. This will confuse statisticians and buy you valuable time.
- When they’ve got you cornered, argue aggressively that statistics is irrelevant to the debate. Cite some mumbo-jumbo from physics or some field your opponent has little background in.
- Cite philosophers or intellectuals no one has thought about since they were in school. It doesn’t matter if your citation is inaccurate.
- Many quant types don’t understand that humans sometimes tell jokes, and interpret all human speech literally. If you have an audience, make a really dumb joke with a really straight face and watch them respond really seriously. They might not realize they’ve been taken but the audience will.
- Claim they’re being theoretical when they’ve punched a big hole in your argument.
- Make it seem like they’re geeky perfectionists when they’ve uncovered fatal flaws in your data or data analysis (if you’ve bothered with either). Sound exasperated.
- Emotional arguments are the best of all. Cite “true” stories that would make even a hardened criminal burst into tears. Make the statistician seem cold and distant by contrast.
- Make it seem in dramatic fashion that anyone who disagrees with you is fighting a worthy cause and, therefore, evil. Ad hominems are the last refuge of the scoundrel and often work brilliantly.
- Be sure to read How To Lie With Numbers and Statistical Mistakes Even Scientists Make!
- Don’t respond to emails or phone calls from the statistician the following day. This almost certainly means they’ve had time to think through your arguments and that you’re in big trouble.
Like many statisticians, I’m a lousy debater and have learned the hard way. Slowly but surely, though, we’re catching on!
Kevin Gray, Marketing Research, Statistics and Data Science Subcontracting and Consulting
Kevin Gray is President of Cannon Gray, a marketing science and analytics consultancy. He has more than 30 years’ experience in marketing research with Nielsen, Kantar, McCann and TIAA-CREF. Kevin also co-hosts the audio podcast series MR Realities.
Originally sourced from Kevin Gray, LinkedIn