Yeah, that was my issue as well. If they don’t give the sample sizes (which they always do), it’s reasonable to assume that both groups are the same size. At no stage did they say they were only two possible options (which seems ridiculous to me). Of course, an experienced epidemiologist would immediately ask the sample size or if those were the only two options.
1 Like
I would have said that the first means the probability is 0.000002 (if I counted the 0s right, that is).
The 1 in 50 seems high, but over a person’s lifetime maybe not… I’ve been bitten by a dog (not a pit bull) so what do I know?
If we assume that 25% of the biting dogs are pit bulls (and ignore the chance of being bitten multiple times by different breeds) that’s a 1 in 200 chance of being bitten by a pit bull.
I would think that 100% more likely means a 2 in 200 chance, no? (1 in 100)
1 Like
I interpreted it as additive, but who knows what it’s supposed to mean.
Though “we”, what ever that group is, does know feelingatically what it means.
There is also the issue of what the time frame should be (which I think should be much shorter than lifetime, but probably doesn’t matter unless the probabilities change over time).
If you have two pieces of pizza and I have 50% more pizza than you… that means I have three pieces of pizza, right?
Your pieces of pizza = x
Pct additional = y
My pieces of pizza = z
z = x * (1 + y)
1 Like
So if you have a 0.5% chance of being bitten by a pit bull (x) and 100% higher chance of falling in love with it (y) then
z = 0.005 * (1 + 1) = 0.01 = 1%
That seems low for the chance of falling in love with a pit bull.
Agreed. Percentages are multiplicative.
Cuz, if you have zero pieces and I have one piece, i do NOT have 100% more pizza than you.
1 Like
This is my theory for lottery tickets. My chances of winning with 1 ticket is infinitely higher than my chances of winning with 0 tickets. But my chances of winning with 2 tickets instead of 1 ticket are 100% more, but that is still vanishingly small. So buying 1 ticket is worth in increase, but buying 2 tickets is not worth the increase over just buying 1.
1 Like
I’m skeptical as well. That website seems to cherry-pick a lot and handwaves a lot of “do your own research,” which is a lot of research (they do have citations of boring studies).
The site also seems to have a lipstick against pitbulls, citing that they bite far more frequently than other breeds in the studies they have (stats they cite involve hospital visits).
Site needs a disclaimer: “The owner and maintainer of this website had a child killed by a pitbull, so read this with that in mind.” ( Maybe not true, but, I’m gonna guess 80% likely, and am taking bets.)
Amend “killed” to “mauled” and you’re probably right with the odds. A lot more people get injured than killed.
If I have a credit card that gives me 2% rewards on groceries and 3% rewards on gas, wouldn’t you say I get 1% more on gas? Perhaps in context you could say 50% more with no problem, because no one would have heard of a 52% rewards card, but 1% more is IMO what would typically be said.
And if I said I had a 1% chance of been bitten by a pitbull, and a 2% higher chance being bitten by a snake, would you expect that I thought the chance a snakebite is 3% or 1.02%?
The pizza example is not really comparable, IMO, because it is talking about 50% more than something that is typically an integer.
In the pitbull example, you are talking about a probability, not explicitly cited, but probabilities are often expressed as a % percent.
Instead of saying 100% more likely it may have been clearer to just say it’s 2 times more likely. This is the type of things that lead to the communication exam for the IFoA.
But if it’s 2 times more likely, isn’t that 3 times as likely?
1 Like
Mmmm, I have always taken 2 times more likely to be z = 2 times y.
Confusing as!
I’m beginning to wonder just who is innumerate.
1 Like
I can understand why you might see things differently from those of us who usually count 1,2,3
1 Like
You could find a winning ticket on the ground somewhere, so buying one isn’t an infinite increase in your chance…
I wouldn’t, no. I’m sure some people would, but I am not among them.
Talking about percentage changes without the context of the actual numbers themselves is confusing. It’s extra confusing talking percentages when the original numbers are percentages too.
It should be the author’s responsibility to clearly indicate if they mean additionally or multiplicative and provide all the context so it’s clear what they mean.
ETA: the whole percentage change thing without numbers has made me pause on the electoral analysis of the reasons why the democrats lost till I came also see the numbers as well.