Freakonomics

[ctenidae]

Have you read it? If not, you should. Pretty amazing stuff. I'm about 3/4 through it now.
Basically, it's two economists looking at everyday kinds of things, but with no emotional filters- just the numbers, pure and simple.
For instance, they look at teachers and mandated tests, and how the incentives encourage teachers to cheat for their students; they look at the crack trade, and figure out why crack dealers still live with their moms; they look at crime, what causes it and why we didn't get the "bloodbath" that was predicted in the mid-90's (the answer is quite surprising, and goes back to an important 1973 Supreme Court ruling- again, no emotions, just looking at the data); which is more dangerous- guns or swimming pools (swimming pools, as it turns out). That sort of thing. Pretty crazy, and easy to read (not filled with economics mumbo-jumbo).

 

[bvibert]

Looks interesting. I may have to check it out...

http://www.freakonomics.com/

 

[hammer]

Somehow the following quote comes to mind...

"There are three kinds of lies: lies, damned lies and statistics." - Mark Twain

 

[roark]

It's a good, quick, and easy read.

Another quote that's good to have in mind: "Correlation is not causation."

 

[riverc0il]

i took a look at that book and passed. roark and hammer's quotes are spot on the mark. looks like an over simplification of things. great stuff to put a book onto the best sellers list, quick and easy reading for sure. but if you really want an education on all of life's problems, you can't just look at the numbers and you have to dig deeper. if you want to read about the stuff in that book, find a reputable specialist's title covering the specific topic. i am sure it is an interesting read, but i doubt this type of foot hold could ever carry much validity in academia.

 

[salida]

i took a look at that book and passed. roark and hammer's quotes are spot on the mark. looks like an over simplification of things. great stuff to put a book onto the best sellers list, quick and easy reading for sure. but if you really want an education on all of life's problems, you can't just look at the numbers and you have to dig deeper. if you want to read about the stuff in that book, find a reputable specialist's title covering the specific topic. i am sure it is an interesting read, but i doubt this type of foot hold could ever carry much validity in academia.

Steve's right, it wont carry much clout in academia, however, I disagree that its a simple book of statistical over simplification. Having read the book, and being a baby economist myself (still in school). I think it holds some valid points and stories, while backing them up with real numbers. For instance the substantial drop in crime in the mid-90's was not due to federal and local crack downs on crime, but an aging criminal population due to the legalization of abortion in the late 70's. Thus poor and less well off people, whom were more likely to have unwanted children, we're now allowed to have abortions, were doing so, thus not raising criminal children.

*Disclaimer: Not that I agree with all of the above points, IE poor people turning into criminals. It is an interesting read, and I don't feel that it was an oversimplification of numbers and statistics. More good theories backed up with numbers as proof.

-Porter

 

[ctenidae]

salida's right- it's absolutely not an oversimplification. These guys applied some pretty rigorous statistics to things in an attempt to figure out if conventional wisdom is right. Turns out, it's not. Their point isn't to convince people that abortion is good because all the aborted kids would have grown up to be criminals. Instead, they're saying that "experts" aren't always right, and are often serving their own needs.

On the crime statistic, for instance, first they set out to prove that more cops equals less crime. Turns out the data just doesn't show it. So then they thougth maybe innovative police methods worked. Not so much. Incarceration, maybe? Some effect, but not enough to explain the drop in crime rates. So, they looked at abortion- turns out that most people who have an abortion do so to keep from having a child in bad circumstances (logical). They also found that kids born into bad circumstances have a greater tendancy to become criminals (again, logical). It follows, then, that fewer kids in bad circumstances = less crime. To prove it, they looked at crime statistics in the 5 states that legalized abortion before Roe v Wade. Turns out their crime rates fell sooner, by about the same amount of time they led Roe v Wade.

They take great pains to say that they are not supporting abortion as a method of crime control or anything like that. They leave the ethics to someone else. What they are saying is that spending gobs of cash putting tons of police on the streets really isn't all that effective, despite the "experts" claim that it is.

Interestingly, academically they hold great weight, since they have been very careful in their methodology. Where they don't carry much weight is with the "experts", because what they say runs counter to the experts source of bread and butter.

Look at their piece on school choice. They found that whether a kid went to one school or another made little difference (except in the case of trade and vocational schools- interesting, that). Instead, the determinant of a kids school performance lay in the fact their parents were interested in what school they went to (ie, interested in their child's education). It's much easier to raise a rabble clamouring for school choice than it is to convince parents that their child's success is up to them. Easier to blame the school than yourself.

The conventional wisdom and the power of experts is what they're attacking. So, hammer and roark's quotes are very appropo- Experts use statistics to generate their lies, and lots of people are big fans of correlation equaling causation.

I have a tiger repelling rock. Do you see any tigers around here?
Works, doesn't it?

 

[riverc0il]

Steve's right, it wont carry much clout in academia, however, I disagree that its a simple book of statistical over simplification. Having read the book, and being a baby economist myself (still in school). I think it holds some valid points and stories, while backing them up with real numbers. For instance the substantial drop in crime in the mid-90's was not due to federal and local crack downs on crime, but an aging criminal population due to the legalization of abortion in the late 70's. Thus poor and less well off people, whom were more likely to have unwanted children, we're now allowed to have abortions, were doing so, thus not raising criminal children.

over reaching. you can not explain criminology using economic numbers. criminologists are the folks to for answers to crime patterns. as a former CJ major, i would not even attempt to put an over simplified answer like that on the table, much less sell it in a book. you will also notice trends that correlate crime with the economy... the mid-90s were the best of times! tech boom in full effect, everyone has jobs, hey presto, low crime rates. also, changing styles of law enforcement should also be taken into effect. while i completely agree, most changes in crime patterns don't have much to do with law enforcement, there are a multitude of factors and that example merely correlates an event with a pattern, that is NOT causation. the fact is, with or without the ability to have an abortion, unwanted children are born all the time and equating unwanted with criminal is a HUGE step not to be undertaken lightly. i am not claiming to be much of an expert hear, but i would read that book would a lot of scepticism, especially when a correlation of event with a statistic is passed off as causation.

 

[riverc0il]

ctenidae, i agree with all your points. if anything, i always say my education in the CJS caused me to be one of the most outspoken critics of the system. i couldn't even bring myself to appraoch criminal justice as a career, even trying to work from the inside to change issues that just aren't right. it is disgusting and you are right, people on the inside are manupulating statistics to prove points that support their position and employment, a huge bias. i think it is great that they are attacking the experts. but are they not just switching the playing field? just taking the opposite side? using numbers to back up opposite claims to standard accepted knowledge? how are their statistics any better, especially when they are correlating events. the thing is, i have seen enough statistics to have come to the conclusion that unless a correlation is extremely high, it is essentially meaningless. and even when the correlation is high, it could be because of confounding circumstances. take that crime pattern for example, did they control ALL other variables including socio-economic factors?

i am just saying reader beware. judging from what you guys have wrote, i think this is probably a GREAT book for most people to read because of the calling into question of generally accepted knowledge, expert opinion, and use of manipulating statistics to prove a point. but approach with skepticism, verify their methodology, question whether all other variables were considered (there are an ennormous amount to hold constant), and what not.

 

[ctenidae]

Steve- this book would be preaching to the choir for you.
The part of the methodology that gives me some faith is that Levitt wasn't looking for contrarian views and then finding the data to prove it. Instead, he took existing conventional wisdom, and tried to match it to the data. Levitt doesn't give the actual statistics, so there's no way of knowing exactly what the correlations are, but even taken with a heaping teaspoon of salt, he does a pretty good job of calling the experts into question.

I agree that criminologists are closer to the issue than economists are. Levitt tries to apply some of the tools of economics to crime, among many other things. On teh abortion issue, probably the most contentious in the book, he looked at the things that are popularly cited as driving the drop in crime rate, and found little or no correlation (no correlation does equal no causation) for most of them. The answer, then, must be something else, and legalized abortion is the thing he found with the highest correlation. In the end, he finds that Sherlock Holmes was right- when you eliminate the impossible, whatever remains, however improbable, must be the truth.

 

[riverc0il]

word, sounds good. though often when you eliminate the impossible, there are a lot of things that remain as possibilities. i like that the authors started with an assumption and then disproved it before asserting statistics for other possible causes. however, not showing your stats and backup work is suspect in my book, especially when dealing with hot botton issues like that. it doesn't take much additional effort to cite and show your research and data. only problem is that it lets other people verify your work.