Rating:  -
I am completely disappointed by this book. It does not fullfil what it promises on its cover. [An exploration in] the hidden side of everything?? I think it's more like A speculation into a handful of unrelated topics.
The book is very thin in its content. Even with only 200 pages in the version that I have, it feels stretched. The last chapter is particularly bemusing with a bunch of names listed. I would have gotten 99% of what the book has to say by just reading the introduction. While I do not doubt Prof. Levitt scientific approach to his work, this book seems more like a bunch of speculation. The rigorous analysis was just not apparent in the book. This is acceptable for a popular writer from whom I would be expecting something fun to read. But for an author, who hold himself out as a Chicago economist, I expect more.
There are few useful take-away from the book. The authors presented their "great" findings of real estate agents working to maximize their benefits instead of the home owner. This is hardly suprising. They are agents and I think it is well-documented that the interest of the agent does not always coincide with that of the principal. Home owner still engage them not because they are ignorant of this fact but the benefits of having these agents in the process is probably higher than the costs (the agency fee plus not selling at the maximum price). An home owner do not have the knowledge, time or may be more importantly, access to the pool of clients. Teachers cheating. Is this a news? Afterall, they are human and they respond to incentive...we often watch in the news that even highly religious people will commit crime too. How surprising is that a teacher will cheat?
I said the book is poorly written because it's not even entertaining. there are people out there, who are better story-teller. The narrative of the book is too plain, the lead to the main story is too long (& sometimes not entirely relevant).
The only up side of the book is probably the way how Prof. Levitt uses ingenious way to measure things that are very hard to measure. But, the robustness and limitations of those meausres were not discussed. so, overall, it's very hard to judge whether the "observed" patterns are real or are they just results of data mining.
A hugely disappointing book.
Rating:  -
I almost wonder if the book is an inside joke, because the authors commit virtually every sin they accuse other experts of committing.
That aside, they go through a massive study to prove such immediately apparent things like: lower class people glom onto names for their children that are popular among the rich, and once that happens, rich people find new names to make popular. Oh, and naming your kid with a name popular among the rich doesn't mean your kid will be rich.
Thanks for that blinding glimpse of the obvious.
Then there is this bit of chicanery: They conclude that it isn't what parents DO for their kids that matter, but who they are. As an example of this, they cite statistics that show that kids who grow up in a house filled with books do well in school -- but reading to your kids is not actually a factor in later academic achievement.
Hmmm. Without a Ph.D., I can figure that perhaps they just aren't asking the right questions here. What is it about that house filled with books that promotes academic success? The authors acknowledge that people who have a lot of books are probably educated and read a lot. But they leave it at that saying "it is who the parents are and not what they do that matters".
Perhaps the lesson to be learned here is that a parent must practice what he preaches and lead by example. But no, the authors say that "by the time a parent picks up a parenting book, it is too late to do anything about the type of parent they will be -- they have already formed their own personality and that is the determining factor."
Way less than impressed.
Rating:  -
Surprisingly superfical. Read the whole book in two days. Could have easily done so in a single sitting, except that Real Life(tm) intruded. Is entertaining in spots here and there. The writers obviously have a gift in distilling data into understandable concepts, applicable to seemingly intractable problems that modern societies confront. The book does not shy from controversy, dabbling in topics such as abortion and race. The chapter delving into names was the weak point of the book, getting bogged down in the mundane data. Overall, an average read.
Rating: -
Overall, I found this worth reading, but it irritated me how the authors think that to make this a mass-market product they need to portray its upperclass heroes as underdogs. Levitt is presented as a brilliant dude with a six-figure income, "fancy" degrees, and wide recognition but at the same time he's a "rogue economist" (interestingly according to my dictionary "rogue" as a noun means "a dishonest, knavish person; scoundrel"); a source Levitt uses calls himself a "rogue sociologist"; and the CEO of the Chicago public school system in 2002 was "an unlikely candidate to hold such a powerful job" who had never before had "a job important enough to have his own secretary". Also when he responded to the research Levitt presents on teacher cheating by readministering tests, this CEO only had the resources to retest 120 classrooms. Poor baby.
The teacher cheating results aren't as sensational as the ideas about legal abortions linked to a drop in crime, but the teacher cheating allegations resulted in people losing their jobs so they're worth a closer look. First we get the motive - newly introduced laws meant that if a class wasn't up to a certain score on their standardised tests, that could lead to teachers being laid off. Then Levitt introduces an algorithm used to detect teacher cheating, which looks for identical strings of answers across tests in the same class. He also expects these identical strings to appear towards the end of the test where he assumes the harder questions would reside. Therefore a teacher using her limited time to correct test answers could memorize a short string of answers and have an impact on the grades of her class. He provides results from one normal class that doesn't show such patterns and one suspect class that does. To top it off the suspect class hasn't performed as well in previous or subsequent years.
This evidence definitely puts teacher cheating on the table, but I can think of another explanation. Questions that appear one after another in an exam are often related. A string of questions often deal with the same topic, and the result of one question may even be required to answer the next. Freakonomics said there were a lot of tests from that class with strings of blanks elsewhere that contained the suspect string of answers, but it's at least possible that the teacher of that class managed to convey one topic clearly and others poorly, or maybe that class studied that topic in the final class before the exam so it was fresh in everyone's memory. Students may have been left with gaps in their knowledge elsewhere, and yes, that may be a sign they didn't have the best teacher so it would make sense if those students didn't show improvement in the long run. It would also make sense if after a year or more - and the research compares the suspect results to those of the same students in their next year so I assume it was a long time till the retest - the students had forgotten about that topic and couldn't answer the same questions. I'd really like to see what those questions were before passing judgement.
Finally, if as Levitt says, teacher cheating is rarely looked for, what's with the assumption that teachers would need to memorize answers? Teachers carry papers - one more piece of paper wouldn't be noticed and probably not considered incriminating even if someone did notice that it had all the test answers written on it. Then assuming answers are weighted equally, there wouldn't be much incentive for a teacher to fill out a relatively long sequence of correct answers one after the other - even a mediocre teacher would know that's how cheating gets picked up. How much of a time saver would that be anyway versus randomly correcting wrong - or even better, some of the many blank - answers on as many tests as you could, assuming you had all the right answers written down?
So I would say the evidence as presented in Freakonomics isn't enough by itself to justify firing people, but as you can tell, it did get me thinking. Although the authors repeatedly say there is no unifying theme in their book, there's a unifying formula - question some common sense, preferably about a hot topic, think about the "incentives" - financial and otherwise - of the people involved, and then use/skew some statistics using techniques and ideas from economics to support a conclusion that is hopefully shocking or at least counter intuitive.
I do think it's still worth a look and I agree with one premise: question everything.
Including Levitt. At least till he gives me a bunch of winning horses to bet on, like he says he can.
Rating:  -
Freakonomics is everything and more than I imagined. It is genius and entertaining!! It's a must read!
|