What's dumber than a mollusk?

Now, anyone who doesn't flat out admire "Captain Capitalism" Aaron Clarey has to be won over by this post. Here, a guy who has repeatedly demonstrated a complete lack of understanding of how to do statistics, calling people "mollusks" for - wait for it - not doing statistics well! That, friends, is cojones. It is truly impressive. Let's go through the post a bit, shall we? This is a bit outside of the usual remit of this blog but fortunately there's some offensive speech, not just the usual brainless patter.

Now, a lot of people - even, say, buisness oriented folk - have found a lot of bad news in the Census report on poverty. But here he's picking on a left-leaning group's comments on it.

He begins by making fun of a graph showing how many more people are living in poverty in America over the past few years.: there has been almost a 20% increase -- not beanbag. He then says they had to "mainpuilate the scale on the Y-axis." Uh, no. In fact, real stats packages real economists use - and for that matter even the crappy Excel graphs Aaron uses - will do this by default. The natural Y-axis ranges basically over the same range as your data, no? More on axes later. A graph is designed to help you present your data, kiddies, and that's all - "the right way" to set it up is so what you want seen can be seen.

He says, well, of course there's more people who're poor, there's more people, isn't the rate the relevant number? And here, Deputy Bullshit Spotters, is the point to listen to your inner voice that says, "Is this moronic toad really trying to suggest the population of the United States went up by something like 20% between 2000 and 2005?" Because it didn't -- it went up, by the Census Bureau's estimates, about 5%. That means, a plot of the rates is going to look, well, pretty much like the plot of the absolute numbers, with pretty much the same trends - NOT flat. How does he make it look flat? He screws with the Y-scale. Excellent.

And here's something for all you stats wannabes out there. You want to compare to pieces of data graphically? Make sure you show them with the same scaling, whatever scaling you use. Anyone who doesn't, is hiding something, or perhaps stupid, or both.

Now, Pathetic Ignorant Little Man (PILM) Aaron goes on to worry about "context" - you see, it's not fair to look just at a short time scale. Of course, PILM himself often makes judgements about trends from even shorter time spans - like, for example, here. But never mind. So he shows a plot of the poverty rates beginning with 1959. Basically it shows a gentle falloff from the '50s to 1970 -- from which one might infer that things like the New Deal and Great Society worked OK. Aaron thinks that's the relevant context. Um, why, oh Deputy Bullshit Spotters, is it relevant to compare poverty rates now, to a time when the US was segregated, and half the anti-poverty measures in action now, weren't in place? This is the challenge of all data analysis: define the meaningful "context." But for a group like CAP looking at immediate policy effects and solutions it's hard to defend Aaron's concept of "context."

"Thus, official economists would come to the professional conclusion," (says the PILM, who is not an economist, contra just about every economist involved in policy anywhere in the political spectrum) "that 'People should shut the hell up and stop their whining. Poverty is not a problem in America.'"

But, hey, there's never been a problem, right? After all, Aaron Clarey is the one who said that

"I'm sick and tired of your greedy scum bucket parasites known as 'the poor' who think you have it so rough"

so, how can there ever have been a problem, eh? I mean, screw these people!

And that's in a way as offensive as anything Aaron has ever written, and goes back to the point below about guys sitting in their rooms playing Super Mario who have no idea how actual Americans think. My personal guess on that thinking being, that while we recognize that how you help poor people may be a complicated issue, there is a real and terrible poverty problem in America if just a few hundred thousand people - let alone 40 million - are in need, and we desperately want to help, and anyone who denies that vision is a moral cretin, more fit to be called a "retard" or "mollusk," simply for lacking that understanding, than anyone at CAP. And I don't think it's just "liberal" or "conservative" Americans who'd agree with that.

Moving on - he again takes a shot at assessing the administration's policy effects just by looking at the past 4-5 years instead of longer (uh...OK, you don't need me to point out the problem). But then he says, hey, incomes fell for a while because we're coming out of a recession. Which is funny, because Aaron has heaped scorn upon the idea that we're coming out of a recession! And because, OK, we're coming out of a recession, incomes are recovering - but poverty numbers AREN'T, which seems to be CAP's point, no? Note that on the graph the Census Bureau doesn't identify the "recession" the same way Aaron (now) does. It's interesting to note, by the way, that while median incomes are stangant or declining, mean (per capita) incomes are going up, so Aaron's graphs are actually not addressing the "somebody's getting screwed" question.

Lastly he hits on the idea that illegal immigrants might be causing the jump - and, well, I think you know how to handle that question, it's going to look a lot like how you clever folk handled the population question. You know, "Is the PILM really saying....?" That way.

And check out the apparatchiks in the comments. It is tragic that after all the good work on decent people like Philip Mangano and others showing how new ideas and approaches can help the poor and the homeless, there are still fools who will spit out invented statistics about the percentage of homeless that are mentally ill or the like. These are not, of course, "fact-oriented" people.