Comment by lutusp | Hacker News Reader

Comment by lutusp | Hacker News Reader

Parent Full thread

> The USA poverty line is some 20x world median income.

Check: median --the middle of the range. Not the average value.

http://en.wikipedia.org/wiki/Median

Quote: "The median of a finite list of numbers can be found by arranging all the observations from lowest value to highest value and picking the middle one"

Therefore the USA poverty line is not 20x the world median income, because the world median income is roughly half of Bill Gates' income. (Technically it's (H-L)/2 + L, where H = highest income, L = lowest income.)

> If you're making more than twenty times what the top of half the people on the planet are ...

Wait, that's not the median, that's the average (or mean). They're not the same. In a symmetrical distribution, the mean and median are the same, but most distributions aren't symmetrical. Personal incomes are most certainly not symmetrical, anywhere in the world.

So when specifying median or average, be sure you know the difference.

I know the difference. That's why I said "median".

Median world income is $2/day. USA poverty line is $46/day. That's a 23x difference, 20x if you round in the direction of common qualifiers.

> Median world income is $2/day.

I think you should be using average or mean, in particular because you're comparing a median to a mean (the U.S. poverty level). What's the point of choosing the middle value in a billion income figures when the average produces a more meaningful result? Just look for the point on the distribution that has a first derivative of zero.

Also:

http://www.bbc.co.uk/news/magazine-17512040

Quote: "so the average income is heading towards $10,000 (£6,273) per person per year."

That's an average of $27.37 per person per day. Which means the US poverty line of $46/day (an average value) is 1.7x the world average income level.

Also:

http://www.wisegeek.com/what-is-the-median-income-worldwide....

Quote: "The median income worldwide — the amount that is dead middle between the least and the highest amounts — is $850 US Dollars (USD)."

Which works out to $2.32/day.

Also:

http://money.cnn.com/2012/01/04/news/economy/world_richest/i...

Quote: "In fact, people at the world's true middle -- as defined by median income -- live on just $1,225 a year."

Which works out to $3.35/day.

I'm not sure the median is what you're after.

Interestingly, your other argument does apply here.

It's inappropriate to use the mean for world average income because it's inflated by very high earners. Very high earners (outliers) have a disproportionate effect on the mean that makes it unrepresentative.

Imagine if salaries were ${1,1,1,1,1,1,1,1,1,1000000} Then the mean average would be $200000 - but you could not claim this represents most people. The median is $1, which is more fairly representative.

As with any skewed distribution, the standard way to measure average income is by using the median. There is broad agreement in official figures to use the median.

The BBC article is wrong to use the mean: the journalist appears to be confused in comparing the mean (which she calculates) to the median (standard published figure) and then claiming the difference results from patchy stats.

I'm not sure you want to get my point. I'm sure I do want the median: if half the people on the planet are doing much worse than you, then the term poor doesn't apply ... yet the term "poor" keeps getting revised upwards into historical levels of luxury.

If you're at the USA "poverty line", you are doing better than 87% of people on the planet. That. Is. Not. Poor.

All valid points, but you were comparing a world median with a U.S. average. Surely comparing two medians (or two averages) would be more appropriate. For example, one could establish a median "poverty level" representing the point below which 25% of Americans fall (or another percentage on which people agree). That might be more enlightening.

It would be interesting to know how many Americans fall above and below the average represented by the poverty threshold, i.e. using a straightforward count -- more in keeping with using a median measure.

> yet the term "poor" keeps getting revised upwards into historical levels of luxury.

I agree with that point, entirely. I also think it's bizarre that so many people end up being defined by absence of an arbitrary property -- homeless person, childless couple, unemployed worker -- which to me seems an underhanded way to enforce social conformity.

"you were comparing a world median with a U.S. average."

No, I was comparing to the official legal definition of "poverty line".

My gripe is that in all this talk about "poverty", either nobody defines it or they define it upwards to rediculous. If someone earns more than 87% of everyone on the planet and is still deemed "poor" then the definition of "poor" is absurd.

The median is not found by (H-L)/2+L. Read that Wikipedia entry again.

Disproof: 1, 1, 2, 10, 101. The median of this list is 2, not 51.

> The median is not found by (H-L)/2+L. Read that Wikipedia entry again.

I really did read it, but I managed to take it to mean the midpoint of the range, all evidence to the contrary. It seems I tripped on the words "middle value", which is ambiguous.

Yeah! Modes of centrality can mean lots of things. It's why it can be so frustrating when media folks casually toss out "average" for long-tailed things that require way more specificity.

Reply on news.ycombinator.com