« Changing Your Metaethics | Main | Setting Up Metaethics »

July 27, 2008


This type of bias seems like it results from a sort of generalized availability heuristic. For doctors (well hospital docs anyway) the cost of guns is obvious and thus easy to bring to mind, while the benefits are more difficult to bring to mind.

I suspect a generalized availability heuristic is responsible for a lot of "selfish biases." For example, in industrial settings, engineers tend to evaluate proposals in terms of how it affects their own "area" because the costs and benefits are very easy to see there since they think about their area all the time, while the costs and benefits to other "areas" are tougher to bring to mind.

Oh poor Bastiat. One wishes he would get more attention.

Maybe even my pet idea of distribution of his collected works to lawmakers? Call the program, "The Law for Lawmakers". Is that too utopian?

I'm not sure it's a good idea to refer to "economist John Lott" with a straight face like that. I think that real economists don't make up sock puppet characters to sing their praises, or quote survey numbers when on being queried they can't produce a shred of evidence that the survey even happened.

This might be along the lines of Paul's post. What does the best empirical research have to say about gun control vs. gun proliferation and overall public safety? This might help me understand if the Lott example you're citing is cherry-picked or not (and if your use of it serves an agenda of overcoming bias or promoting a particular ideological bias).

A policy bias against the unseen is plausible, but after all these years do we have any studies showing it in action? And I fear again that people tend to most point out biases that could plausibly afflict their opponents when their own positions are weakest.

Dear Paul Crowley:

Detailed evidence on statements from two people who took the survey, the replication of the survey results, and additional statements from others is available here.

Dear Robin:

I am very happy to accept the argument that people as a whole don't make systematic mistakes. But it is interesting to see the different views between public health academics versus criminologists and economists. If you have an alternative explanation, I would be happy to hear it. Even if people get it right on average, it seems possible that there will be some groups that make errors in each direction, both over and underestimating the net costs. There are two issues here. 1) The relationships that exist between guns and crime or suicide. 2) Why certain groups in the population believe what they do.

Is it also possible that individuals advocating policy changes (such as the abolition of the FDA) are biased because they don't regularly see people dying from contaminated foods or medicines?

Just a thought...

John, the key issue is how we determine which side is more biased in a dispute. Merely being able to list possible biases each side might suffer doesnt' get us very far.

Robin - I don't get it. Are you proposing that it is not very useful to discuss bias unless it can be quantified? How do you propose to quantify the bias - except by comparing the biased judgment to the truth? But for this you need the truth. But knowing the bias was supposed to help you get at the truth (hence the point of this blog). But you just said that just knowing the bias is not very useful without quantifying it. But for this you need.... I'm caught in a loop. I'm probably missing something.

Speaking of the FDA, I questioned Lott about intellectual property here. He thought the case in favor was so strong as to make discussion of it a waste of time. I suspect there are unseen innovations that don't come to market for fear of "patent trolls" and the like

Paul Crowley, Lott has an economics phD and I believe has taught economics courses. He may be a bad economist or a bad person, but he's still an economist.

HA, I believe the general consensus is that gun laws make no detectable difference for good or ill, with the possible exception of suicide being easier (and even that is questionable).

To clarify, I am not accusing Lott of badness, I just want to leave as much room for those with a low opinion of Lott as possible while still making them concede he is a real economist.

TGGP, is that demonstrated between countries too? (Because I anticipate the obvious criticism if that finding is only between states in the USA).

Also proliferation isn't the same as gun laws of course.

Usually there is more criticism for between country studies because of all the difference that must be controlled for! Why would they be better than comparison between states (or counties)? Ease of moving guns from lax to tight contiguous areas? I mostly hear about studies on American gun laws (especially after Heller). I remember Mark Kleiman (one of the most intelligent anti-drug-legalization folks I've come across) saying that our homicide rates would be lower if our ownership rates were like Japan's, but that gun control and right-to-carry laws in the U.S still don't seem to have any effect one way or the other. Here's a paper on gun control in England, but it's from an American libertarian think tank. Googling for studies on UK gun control I found that David Sharp had written one in 2006 with the sub-title "Still a Matter for Debate", but it's behind a subscription-wall.

There's little evidence, as I understand it, that guns contribute to crimes not committed in proportion comparable to the damage done by the crimes they exacerbate. But that's not my point.

My point is that non-MDs DON'T see all this damage, and so may underestimate its prevalence.

It's certainly the case that people not trained in atmospheric physics are for the most part MUCH less concerned about CO2 accumulation than are those of us who do "see" the extent of the change in the energy balance as compared with natural forcings. This sort of blindness doesn't especially have a political skew, as one might gather from your examples.

Dear TGGP:

The comment about Japan doesn't make a lot of sense because Japan had a very low murder rate prior to the institution of their gun control laws. Indeed, the murder rate prior to the ban was at least as low as it was after the ban. Similarly, for the UK, their murder rates prior the first gun control laws in 1920 were extremely low. In London in 1900, a city of millions of people, there were a total of two gun murders and five armed robberies.

As to the claim that Right-to-carry laws had no effect on crime rates, here is some of the published academic research in refereed journals that find a statistically significant drop. The large majority of the refereed research finds that RTC laws reduce murder and violent crime rates. No refereed academic journal article has found a bad effects from these laws.

Your comment about Japan is consistent with what I said: gun laws don't really seem to have an effect one way or another. Remember the distinction between gun laws and gun ownership. Wasn't it the case that Japan didn't have much ownership of guns before the ban?

It may also be the case that no refereed article has found bad effects and some have found good effects while the preponderance of evidence favors no effect. I am of course not an expert on the issue, though I should lay out my biases in that I am a gun owning hunter radically in favor of the second amendment (even for bazookas, as suggested by Reason magazine).

So TGGP and Lott, you're saying that there are two situations which result in lower gun-related or other violent death in a country than is the status quo in the areas of the USA without right to carry laws or low levels of gun proliferation (low defined as Japan or UK levels): (1) right to carry laws, and (2) lower levels of gun proliferation. What does the best evidence have to say (if anything) on which results in the lowest gun-related or other violent deaths: (1) right-to-carry laws in areas with high (high means current USA) levels of gun proliferation, (2) right-to-carry laws in areas with low (Japan/UK) levels of gun proliferation, or (3) no right-to-carry laws in areas with low levels of gun proliferation. Also, please just compare (1) and (3) if there isn't enough information on (2).
I'm especially interested in Lott's take on this since he seems to have done more research on it than the rest of us.

If the best evidence suggests I'd be safest, for example, in a society with right-to-carry laws and low levels of gun proliferation, than that's the type of society I'd want to live in. But intuitively, I admit I think I'd be safer in a society where guns are generally rare, than if they're common, even if that society had right-to-carry laws. Still, the best science trumps unarticulated intuition, in my opinion.

I think the homicide rate in Japan is much lower than the average right-to-carry area in the U.S. The causality is hard to figure out though.

I should add some more on gun proliferation. Switzerland and Israel both have lots of automatic weapons available to citizens because of conscription. They also have lower-than-average homicide rates. Finland has lots of guns and a higher-than-average (for Europe) homicide rate, but those homicides tend to be with stuff like knives. It's gun homicide rate is pretty low. I think the overriding factor is that Japan is full of Japanese people, who don't seem to have high crime rates anywhere, Switzerland full of the Swiss and Finland full of Finns. From what I've read at the Inductivist, Finnish Americans also have high gun-ownership rates but low crime rates.

I'm especially interested in Lott's take on this since he seems to have done more research on it than the rest of us.

The key problem, as always, is to identify experts.
For politically sensitive issues like this, though, there's the separate question of whether to trust experts. In particular, Lott is a well-documented liar. Even for an objective fact like the RTC literature, I value his claim 1 bel less than that of a generic economist.

Luckily, the direct consequences of politically sensitive issues are not relevant to most people. If you're merely choosing where to live and don't have the opportunity to change the laws, the instrumental value of gun laws in affecting crime rates is irrelevant, since you can measure crime rates.

I happen to be reading "Priceless: On Knowing the Price of Everything and the Value of Nothing", which takes a critical look at cost-benefit analysis and how a dollar value is assigned to intangible risks and rewards. It's easy to ask the questions. This blog post didn't do much on the constructive side.

The point made in this post probably has merit. However, John Lott's intellectual misconduct over the years should have long ago gotten him banned from particpating in intellectual discourse. He makes the pie smaller, not bigger.

The comments to this entry are closed.

Less Wrong (sister site)

May 2009

Sun Mon Tue Wed Thu Fri Sat
          1 2
3 4 5 6 7 8 9
10 11 12 13 14 15 16
17 18 19 20 21 22 23
24 25 26 27 28 29 30