![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
The Problem With Anecdotes
Dec. 24th, 2011 12:19 amI'll often get into debates with people on assorted hot topics. This happened today after archery, which is why it's on the top of my head.
Invariably, when there is a disagreement, someone will come up with the "well I heard this from person X" argument. Person X isn't present, but their experience is being held up as proof of that person's point of view. The problem being that I have literally no reason to believe the hearsay of a person not present. However it's impolite to point out why. Here's some reasons:
However, as Gregory House says, "Everyone lies."
I remember once where an acquaintance told me that they had a friend who got pregnant from petting (not sex) in a hot tub. They assured me that the person had no reason to lie about this. I told them the more likely scenario was that they were simply lying about it - they had sex, got pregnant, and were embarrassed by it, so they came up with a barely plausible scenario to save face. However, the story is barely plausible so maybe I'm wrong. But if I have to make a bet based on "odds that a person is lying" vs. "odds that a person can become pregnant in a hot tub from a handjob", I know where I'm putting my money.
Anyway, I got into an argument with some folks about flat tax rates vs. progressive tax rates. I won't hash over the arguments here (you can check out the pros and cons at Wikipedia) - suffice to say I find the arguments in favour much more compelling than the arguments against. More to the point, I don't think we live in a world where the poor need a bigger burden and the rich need less of one.
However one fellow did make the argument that progressive taxes give a disincentive to work because of the apocryphal story of people who make less money because they've moved to a higher tax bracket. He used his Father moving up a pay grade in the military as an example of just such a thing happening. Dad got a raise, take-home pay went down.
I didn't believe him, but as I mentioned, Saying someone's Dad was wrong is not diplomatic.
Now I can imagine progressive tax systems where this can happen. Imagine a situation where the tax rate goes up a percent for every $10K - you pay 1% up to $10K, 2% if you make up to $20K, 3% if you make up to $30K, etc. In this scenario there are edge cases where you will lose money. Would you rather take home 99% of $10000 ($9900) or 98% of $10001 ($9801)? Seems obvious, aside from the fact that these are small edge cases.
However, that's not how Canada implements it's taxation. From Revenue Canada's website:
The edge cases described above never occur - no amount of increase in your income will result in your take-home pay going backwards.
So what about that anecdote? Maybe Canada changed methods (though I doubt it. The rates may have changed, but the method for calculating them probably hasn't - it's a simple solution to an obvious argument against it). Maybe the people who handled his payroll were idiots that kept too much back (in which case it was returned as a refund at the end of the year). Maybe his Father overstated the case (I got a raise, but it doesn't look like much when taxes are applied and it's divided over every paycheque of the year). Maybe my debating opponent assumed it all, based on flawed "common knowledge". The current numbers don't support the argument though.
And that's why I hate anecdotes. Lots of reasons to reject them, but people get pissy if you do. But I have to put up with them all the time because I'd like to not alienate everyone I know. Still, it bothers me. You know what I do like? I like data. The plural of "anecdote" is not "data". A thousand people can tell me how much good Chiropractors are, but that doesn't prove it's effective - it only proves a thousand people believe it was worth it (I know people who think being tied up and beaten with whips is worth it - that doesn't make it effective medicine). Multiple, properly peer-reviewed, studies showing a statistically significant effect greater than placebo? Sure, I'll buy that. Your Uncle Bob having back pain for years until he went to the bone breaker regularly - not so much.
Do I use anecdotes? Sure, but not as evidence, and I take pains to label them as hearsay.
I might get into it here about the progressive taxation thing, but not right now.
Invariably, when there is a disagreement, someone will come up with the "well I heard this from person X" argument. Person X isn't present, but their experience is being held up as proof of that person's point of view. The problem being that I have literally no reason to believe the hearsay of a person not present. However it's impolite to point out why. Here's some reasons:
- The person doesn't actually exists (i.e. debating opponent is lying).
- The person exists, but the debating opponent is deliberately misstating the person's experience (lying again).
- The person exists, but the debating opponent has misunderstood the person's experience (the debating opponent is mistaken, but honest in it).
- The person exists, but lied or was mistaken themselves.
However, as Gregory House says, "Everyone lies."
I remember once where an acquaintance told me that they had a friend who got pregnant from petting (not sex) in a hot tub. They assured me that the person had no reason to lie about this. I told them the more likely scenario was that they were simply lying about it - they had sex, got pregnant, and were embarrassed by it, so they came up with a barely plausible scenario to save face. However, the story is barely plausible so maybe I'm wrong. But if I have to make a bet based on "odds that a person is lying" vs. "odds that a person can become pregnant in a hot tub from a handjob", I know where I'm putting my money.
Anyway, I got into an argument with some folks about flat tax rates vs. progressive tax rates. I won't hash over the arguments here (you can check out the pros and cons at Wikipedia) - suffice to say I find the arguments in favour much more compelling than the arguments against. More to the point, I don't think we live in a world where the poor need a bigger burden and the rich need less of one.
However one fellow did make the argument that progressive taxes give a disincentive to work because of the apocryphal story of people who make less money because they've moved to a higher tax bracket. He used his Father moving up a pay grade in the military as an example of just such a thing happening. Dad got a raise, take-home pay went down.
I didn't believe him, but as I mentioned, Saying someone's Dad was wrong is not diplomatic.
Now I can imagine progressive tax systems where this can happen. Imagine a situation where the tax rate goes up a percent for every $10K - you pay 1% up to $10K, 2% if you make up to $20K, 3% if you make up to $30K, etc. In this scenario there are edge cases where you will lose money. Would you rather take home 99% of $10000 ($9900) or 98% of $10001 ($9801)? Seems obvious, aside from the fact that these are small edge cases.
However, that's not how Canada implements it's taxation. From Revenue Canada's website:
- 15% on the first $41,544 of taxable income, +
- 22% on the next $41,544 of taxable income (on the portion of taxable income between $41,544 and $83,088), +
- 26% on the next $45,712 of taxable income (on the portion of taxable income between $83,088 and $128,800), +
- 29% of taxable income over $128,800.
The edge cases described above never occur - no amount of increase in your income will result in your take-home pay going backwards.
So what about that anecdote? Maybe Canada changed methods (though I doubt it. The rates may have changed, but the method for calculating them probably hasn't - it's a simple solution to an obvious argument against it). Maybe the people who handled his payroll were idiots that kept too much back (in which case it was returned as a refund at the end of the year). Maybe his Father overstated the case (I got a raise, but it doesn't look like much when taxes are applied and it's divided over every paycheque of the year). Maybe my debating opponent assumed it all, based on flawed "common knowledge". The current numbers don't support the argument though.
And that's why I hate anecdotes. Lots of reasons to reject them, but people get pissy if you do. But I have to put up with them all the time because I'd like to not alienate everyone I know. Still, it bothers me. You know what I do like? I like data. The plural of "anecdote" is not "data". A thousand people can tell me how much good Chiropractors are, but that doesn't prove it's effective - it only proves a thousand people believe it was worth it (I know people who think being tied up and beaten with whips is worth it - that doesn't make it effective medicine). Multiple, properly peer-reviewed, studies showing a statistically significant effect greater than placebo? Sure, I'll buy that. Your Uncle Bob having back pain for years until he went to the bone breaker regularly - not so much.
Do I use anecdotes? Sure, but not as evidence, and I take pains to label them as hearsay.
I might get into it here about the progressive taxation thing, but not right now.
