Tuesday, February 25, 2014

Juxtaposition X: In Which I'm Wrong Several Times in a Row

I wanted to find that Alexander's data wasn't all buggered, but unfortunately I kept finding new reasons for cynicism. It doesn't help that I'm bad at counting.

This kind of iterative analysis is normal for me. It's unfortunate that it seems I work best by confidently asserting a position, (it naturally causes me to question myself). Seems like it would be misleading to those I confidently state it at. On the plus side I never forget that I'm wrong all the time, due to proving myself wrong more often than right.

On issues not internal comparison; this is not an argument. I'm simply reporting my observations. I'm not saying you should trust them, but saying I shouldn't trust my own observations isn't going to fly.

First, my notes, then stats for amusements, then my final attempt at a useful summary.

Suicide: straw man - suicide vs. suicide attempts / buggered

Debt: buggered / manipulated / point in favour

Homicide: straw man - homicide vs. homicide attempts / buggered

American homicide: dubious. If true, part of worldwide, nonprogressive trend.

Worldwide anti-violence: nonprogressive, point in favour of nrx. True.

War: manipulated. Probably true?

Hours worked: true but contaminated with welfare data, so buggered for our purposes.

Literacy: worldwide nonprogressive trend, but true.

Poverty: poverty of data, you mean. Also see previous.

Hunger: point in favour of nrx, probably buggered

World per capita: true, irrelevant to nrx position, but true.

US GDP: point in favor of nrx, misinterpreted

Harvard/leftist opinion table: favour of nrx, probably buggered

Blob graph: data is buggered, values aren't 2-D.

Arrow graph: buggered, point in favour of nrx.

Political involvement: point in favour of arguments about Singapore, HDI probably buggered.

World GDP: true but irrelevant

Pop density trifecta: true, misused. Buggered by confounders.

Divorce: buggered, but even if true in favour of nrx. Premarital sex by the lady increases your risk of divorce by 15% absolute, about 1 in 6.

Divorce 2.0: not exactly news to nrx. probably true. Mostly point in favour.

US Pop growth: Straw man - nrx talks native pop growth, SSC includes immigration. Nrx doesn't know what's going on either, though. Best is here; plausible, but I can hardly blame proggies for dismissing it.

Discrimination: good news, probably true, but straw man. See Richwine etc. What the masses think doesn't matter anymore; we live in a democracy only descriptively, not in any sense prescriptively.

4: straw men.
2: graphs cut off at awkwardly convenient moments.
8: data actually favours nrx predictions.
16: data interpreted overgenerously for progressivism.
10: data most likely true.
12: data is untrustworthy.

Total: 22
By descending seriousness of error:
12: probably untrue or self-contradictory.
2: straw men.
1: time range manipulated or chosen for convenience.
4: easily interpreted to favour nrx predictions.
3: irrelevant.

Unfortunately, that's all of them. None of the graphs are evidence neoreactionaries need to be worried about. While it is statistically unlikely that Alexander made this many errors by chance, it is at least as unlikely that neoreactionaries are in fact correct on all these points. It shouldn't be necessary to pick such flawed data to construct a telling criticism.

No comments: