Not Trolling Zaku-Illegal Immigration

[Darius871]

Your experience with sampling bias is hugely biased.

I'm not suggesting my experience is the end-all-be-all here, but it doesn't take a genius to realize that the bulk of cases are solved by documentation of location, employment, associates, etc., all of which are especially lacking amongst the illegal population.

Most simple assault and theft cases in which arrests are made do NOT fall into those categories. Instead, the suspect is caught because the cops find him within a few blocks of where the crime occurred, and he's identified by a victim or witness.

For the life of me I can't think of a google search that will find me that hard statistic, as intuitive as it may sound. Tell me you didn't first hear it on TV.

Even if you're right that it's "most" and not "many," it'd take something like "nearly all" to abate my skepticism about incarceration rates.

[KrazyHorse]

it doesn't take a genius to realize that the bulk of cases are solved by documentation of location, employment, associates, etc.

Now, this is just not true.

Drug arrests, theft, burglary. All solved mostly by the cops actually CATCHING somebody. Not by them carefully investigating and obtaining a ****ing warrant.

[KrazyHorse]

For the life of me I can't think of a google search that will find me that hard statistic, as intuitive as it may sound. Tell me you didn't first hear it on TV.

I'm not arguing anything other than common sense here. Just like you're not. I have no idea why you think that your appeal to non-statistical evidence is any more compelling than mine.

[Darius871]

Now, this is just not true.

Drug arrests, theft, burglary. All solved mostly by the cops actually CATCHING somebody. Not by them carefully investigating and obtaining a ****ing warrant.

...

For the life of me I can't think of a google search that will find me that hard statistic, as intuitive as it may sound. Tell me you didn't first hear it on TV.

Even if you're right that it's "most" and not "many," it'd take something like "nearly all" to abate my skepticism about incarceration rates.

[KrazyHorse]

Even if you're right that it's "most" and not "many," it'd take something like "nearly all" to abate my skepticism about incarceration rates.

Again, if you accept "most" then what it does is place an upper bound on sampling bias. If cops catch half of all criminals on the street and half by careful police work then unless you're claiming that illegals are better at running from the cops than native born Americans, the MOST you can claim is that incarceration rates understate criminality by a factor of 2.

[Darius871]

I have no idea why you think that your appeal to non-statistical evidence is any more compelling than mine.

I don't. But if we're both going to sit here and use "common sense" (i.e. aimlessly speculate), suppose I conceded that, say, 60% of arrested petty crime suspects are arrested at or near the scene. Even then, it could be possible that 99% of illegal immigrant arrested petty crime suspects are arrested at or near the scene (because those who escape scene are virtually impossible to find afterward), compared to only 56% of citizen/resident arrested petty crime suspects (because those who escape the scene can still be found afterward, resulting in a larger total getting caught). In other words, without being able to 1) establish a percentage for petty criminals arrested at or near the scene and 2) separate citizens from illegals within that percentage, its effect on the sampling bias still isn't clear.

Supposing #1 is "many," the upper bound could range from a factor of 3 to a factor of 5 depending entirely on both #1 and #2. Supposing you're right that #1 is "most," then the upper bound could be as high as 2 depending entirely on #2. That factor of 2 could then be raised further by the blatant flaw of conflating legal immigrants (who are unusually law-abiding) with illegal immigrants, and even the effects of deportation, among other flaws.

The bottom line is that there are at least two and perhaps five sources of a potentially huge sampling bias that cannot be quantified with any readily available information, leaving a statistic that's questionable at best. A third problem could be the effect of deportations, though I don't know if your source's definition of "incarceration" controlled for that or not. A fourth problem could be that illegal immigrants are less likely to be caught altogether, whether or not at the scene, because they tend to live in more densely populated areas that make it easier to blend in. A fifth problem could be that illegal immigrants are far more concentrated in some states and localities than others, such that using a nationwide statistic smooths over local discrepancies. Some combination of any of these five could conceivably explain a vast disparity between criminality and incarceration.

On the other hand, biases about "furriners" aside, we have A) the well-established correlation between criminality and poverty combined with B) the strong correlation between poverty and illegal immigrant status. That overlap alone is more compelling to me than the above house of cards.

[KrazyHorse]

I'm not going to read that long-ass exposition. The important part is here:

suppose I conceded that, say, 60% of arrested petty crime suspects are arrested at or near the scene. Even then, it could be possible that 99% of illegal immigrant arrested petty crime suspects are arrested at or near the scene (because those who escape scene are virtually impossible to find afterward), compared to only 56% of citizen arrested petty crime suspects

If we accept a 60% figure (for non-immigrants) and 100% (for illegals) for those captured near the scene then all we have to do is multiply the non-immigrant incarceration rate by 0.6 to get an equivalent criminality rate. And thus the criminality rate for non-immigrants is STILL HIGHER than for illegals.

THAT IS MY POINT.

[Darius871]

If we accept a 60% figure (for non-immigrants) and 100% (for illegals) for those captured near the scene then all we have to do is multiply the non-immigrant incarceration rate by 0.6 to get an equivalent criminality rate. And thus the criminality rate for non-immigrants is STILL HIGHER than for illegals.

THAT IS MY POINT.

I already admitted that, pretty openly. If you want to ignore the rest of my post pointing out that this is only one of five flaws which all can contribute in different ways to the total discrepancy, so be it. I also did not "accept" the 60% figure or any other to begin with. Bottom line, I'd rather side with the well-established poverty/criminality correlation which doesn't suffer from so many flaws.

[Darius871]

I still can't find any statistic to back up your "common sense" about "most" petty crime arrests, but for what it's worth only 34% of "solved" burglaries in Britain involved arrests at or near the scene. Considering that burglars seem more likely to stay in the same place for several minutes than a purse-snatcher, convenience-store robber, etc. who bolts when he gets the prize, it's tough to see how they would get arrested at or near the scene twice as often. Britain could be an outlier, however.

[Dinner]

Yeah, the whole "being able to feed their family" meme is really passé.

Recently in Missouri one of the Republican legislature members came out against free or reduced cost school lunches for low income children because starving "builds character". A true story.

http://www.stltoday.com/stltoday/news/stories.nsf/editorialcommentary/story/9C70A6D2409BEA60862575D6007F24B9?OpenDocument

[Darius871]

I still can't find any statistic to back up your "common sense" about "most" petty crime arrests, but for what it's worth only 34% of "solved" burglaries in Britain involved arrests at or near the scene. Considering that burglars seem more likely to stay in the same place for several minutes than a purse-snatcher, convenience-store robber, etc. who bolts when he gets the prize, it's tough to see how they would get arrested at or near the scene twice as often. Britain could be an outlier, however.

There's also this, from a legal database:

The most common means of clearing robberies in Boston in 1968 was, according to John Conklin, by an arrest at or near the scene of the crime (38%). Victim identifications through examination of photographs or lineups accounted for another 19% of the clearances. Nearly 40% were “secondary” clearances in which a suspect arrested for one offense was connected with other robberies -19% were connected through photographs or lineups for another offense and 19% through “multiple confessions,” in which the arrestee confessed to other robberies. Police investigation through informants or the careful assembly of various kinds of information accounted for the final 8%. Except for its role in the obtaining of secondary clearances, police interrogation was not identified in the study as a source of clearances.

Floyd Feeney, POLICE CLEARANCES: A POOR WAY TO MEASURE THE IMPACT OF MIRANDA ON THE POLICE, 32 Rutgers L.J. 1 (2000) (citing John E. Conklin, Robbery and the Criminal Justice System 133 (1972))

So not only was the figure less than "most," but the other methods would also tend to be more difficult for illegal immigrants; they're typically less socially prominent and more transient, both of which lower the odds of a positive ID by the present victim or a past victim in a "secondary clearance," they typically speak little to no English, which lowers the odds of successful interrogations for secondary clearances, and they produce little to no documentation, which lowers the odds of "assembly of information," among other things. Maybe we can assume on-site patrols are more effective now than in 1968, but maybe not.

[KrazyHorse]

Errr...did you even read what you just posted?

Police investigation through informants or the careful assembly of various kinds of information accounted for the final 8%.

Victim identifications through examination of photographs or lineups accounted for another 19% of the clearances.

So, to reiterate:

40% of them are caught at or near the scene
20% were caught for another crime and then identified by lineup
20% were caught for another crime and then confessed

which of these has anything to do with being an illegal immigrant again?

I'll give you 27%. You owe me 73%

[Darius871]

which of these has anything to do with being an illegal immigrant again?

Since I did read what I posted, I'd already anticipated and answered this:

So not only was the figure less than "most," but the other methods would also tend to be more difficult for illegal immigrants; they're typically less socially prominent and more transient, both of which lower the odds of a positive ID by the present victim or a past victim in a "secondary clearance," they typically speak little to no English, which lowers the odds of successful interrogations for secondary clearances, and they produce little to no documentation, which lowers the odds of "assembly of information," among other things.

At best, those statistics made us both wrong, not just me. I also said more than once now that the likely means of arrest make up only one of five major flaws with your incarceration rates, not the only one, whereas you've yet to identify even one major flaw with the countervailing poverty/criminality correlation.

[KrazyHorse]

Darius, you really do get too lawyerlike sometimes.

Your hypothesis is weak. Sampling bias is present but is not that important when it comes to the statement "illegals have lower criminality". A quick look at the numbers tells us this. Criminals are usually caught for reasons which have nothing to do with their documentation. Thus the thought that you can finagle a factor of 3 or 4 out of there is ridiculous.

[Darius871]

Whatever. I'm equally tired of dabbling in the unknowable.