The Plural of Anecdote is not Data
Or: Why Winnipeg is a uniquely evil city
For all that we consider Canada to be a pluralistic nation, filled with nice people that apologize too much, the patterns seems to break in Winnipeg. I couldn’t tell you why, but Winnipeg shows just how bad Canada can be. Don’t believe me? Let me show you:
- A man charged with multiple aggravated assaults and firearm offences shot a 50 year old man (source)
- Two men charged with first-degree murder and arson by the RCMP (source)
- A man was fatally stabbed in the parking lot of a Tim Horton’s (source)
- A cyclist died in a hit and run (source)
- A late night bar fight left one man dead, lying on the street (source)
I could keep going, but I think this proves my point. Something about Winnipeg drives people to commit violent, destructive, and homicidal acts on a regular basis. This isn’t my opinion — you can follow the links for yourself and read more.
When people claim Canada is a “great nation”, remind yourself that Winnipeg exists.
Was that convincing?
I hope not.
There’s a particular style of debate I find frustrating, that I’ll call proof by quantity of examples. I was inspired to write this after listening to a debate hosted by the Intelligence Squared podcast. The resolution was: “Nationalism is a force for good”. While there were interesting arguments on both sides, one of the debaters used this tactic to great effect.
To prove her point, one of the individuals arguing against the resolution decided to list a series of bad events she attributed to nationalism: Britain leaving the EU, Donald Trump’s election, and more. She had a very big list detailing all the ways in which nationalism has caused bad events to occur.
I’ve just shown why this is a bad rhetorical tactic with the case of Winnipeg. As far as I know, Winnipeg is not uniquely evil as a city. Winnipeg has a population of about 750k, so if even a tiny percentage of people were violently inclined, we’d still see hundreds of examples of Winnipeg residents committing heinous acts.
We see this debate tactic a lot when discussing contentious issues. What this does is pull the conversation away from data-driven arguments into a more shady territory. When the IQ2 debaters were pointing out examples of bad nationalist outcomes, it put the pro-nationalist debaters in an awkward spot. Of course they condemn specific instances of nationalist overreach — in fact, everyone in the debate confirmed they were strongly against authoritarian states — but making the conversation revolve around anecdotes means there can be no discussion of net harm or benefit.
Imagine if, in a hypothetical debate about Winnipeg’s dark character, I told the opposition about a particularly awful murder in the city, and asked them to “tell me why we shouldn’t condemn Winnipeg when instances of cruelty like this occur regularly”. It would be remarkably tone deaf of my interlocutors to push that example aside as irrelevant and focus on the “statistics”. I could even accuse them of not caring about the people whose lived experiences we’re discussing. And yet for all my arguments, I still have not proven my point — only made it socially awkward for the opposition to object.
Why this works
I’ve written before about confirmation bias, and how we seem to reason about arguments we agree or disagree with. Simply put:
- When we agree with the conclusion of an argument, we ask ourselves “can I believe it?”
- When we disagree with the conclusion of an argument, we ask ourselves “must I believe it?”
In the case of the nationalism debate, listeners who already agreed that nationalism was bad, probably didn’t care that the arguments supporting their prior position were weak. It got to the right conclusion, so they ask themselves “can I believe [x number of] examples are convincing?”
A particularly strident anti-Winnipeg activist may also agree with the series of examples I put forward. But they would only do so in the context of already agreeing with me. I doubt it would convince many Winnipeg residents, for example.
The Plural of Anecdote is not Data
No matter your position on nationalism, or Winnipeg’s moral character, this proof by quantity of example rhetorical style is not a coherent argument. What we instead need is data-driven, critical thinking about the resolutions placed before us. If nationalism is bad, go get data about the relative success of nationalist and not-nationalist countries. If you truly believe you’re right, you should have no problem proving your point the right way.
Now that I’ve pointed this out, you’ll find yourself seeing this rhetorical style pop up everywhere in your life. People routinely fall back on anecdotes — and especially quantity of anecdotes — to prove their points instead of reaching for the data.