# What is this inverse statistical terminology?

What I mean is, what is the term for describing a statistic in terms of how much it doesn’t happen rather than in terms of how much it does.

Example (totally made up): my boss says I only worked 1 hour last week, and I say “but this week I doubled how much I worked, isn’t that great?” But of course I was supposed to work 40 hours, so my slacking went from 39/40 (97.5%) to 38/40 (95%)–a mere 2.5% improvement.

What’s that called (statistically speaking)?

I have no clue what you are asking. In general, statistics don’t care a whit about what “didn’t happen.” You seem to be getting more into analysis, or something similar.

Both would be frequencies. You were working 1/40th of the time you were not working 39/40th of the time. I suppose you could call them complementary frequencies, but “complementary” is usually applied only to “probabilities” and “events” in my experience and not frequencies. And even if you called them complementary, each is complementary to the other. There’s no special name I can think of for the “not happening event”.

Having thought about it more, I think the term is “data.”

There’s no standard statistical term for it.

Y’all are being harsh. I think what the OP is asking is “Lieing With Statistics”. In most cases you can choose how to present something to make it the most favorable for your case. If I buy something for \$1 and sell it for \$2 that can be called either a 50% margin or a 100% markup. The market was twice as high 6 months ago vs it fell 50% in 6 months, etc.

This is perhaps distantly related to Hempel’s Raven Paradox.

I’ve seen this argument used in relation to buying lottery tickets, e.g. Person A says that buying two tickets means they are twice as likely to win the jackpot than buying one ticket, then Person B says, no, your chance of not winning has only gone down from 13,983,815/13,983,816 to 13,983,814/13,983,816, or from 99.999992848% to 99.999985697%.

I think that’s disingenuous. Of course your chance of winning has doubled.

It’s the difference between proportions and absolute differences. One is additive in function and one is multiplicative (and their inverses, subtraction and division). Both provide true data. Both must be ‘parsed’ in context and knowing which function is being used so that the data isn’t misunderstood. And then the value of the data must be evaluated as applied to a particular situation.