Etwas andere Einblicke in die Wirtschaftswissenschaft, 5. Januar 2009

Hinter Freakonomics verbergen sich verschiedene Geschichten aus der US-Wirtschaftswissenschaft, die aus verschiedenen Daten ausgelesen wurden. Zum Teil wirken diese doch sehr "amerikanisch" - so wird das Wirtschaftssystem einer Gang beleuchtet. Und das Ergebnis ist, dass normale, einfache Straßengangster weniger Geld als ein einfacher Arbeiter verdienen und deswegen noch zu Hause wohnen.Fazit: Mäßig unterhaltsam.

       Statistik mal anders: kurzweilig und spassig!, 10. September 2007

Dieses Buch ist ein Sammelsurium aus verschiedenen kurzen Geschichten, in denen Statistik angewendet wurde, um ungewöhnliche Sachverhalte oder Zusammenhänge aufzudecken, bspw. unter welchen externen Umständen Angestellte sich ehrlicher beim Kauf von Bagels ohne Kontrolle verhalten. Diese sind in einem locker-flockigen Stil geschrieben, der auf Leser ohne tiefer gehende Statistikkenntnisse ausgerichtet ist, und bietet stundenlange kurzweilige Unterhaltung bei der man auch etwas über die konzeptionelle Herangehensweise bei der Suche nach "Wahrheiten" lernen kann.Levitt untersucht verschiedene "natürlich Experimente", also solche Vorkommnisse, die nicht wirklich als Experiment konzipiert sind, jedoch aufgrund der Gegebenheiten genau diesen Charakter aufweisen, beispielsweise ob der besagte Bagelverkäufer seine Waren nun in Büros des Top-Managements oder einfacher Büroangestellter ausstellt. Das interessante daran sind die ungewöhnlichen Beispiele, die hierbei herangezogen werden, beispielsweise unter welchen Umständen Sumo-Ringer Kampfabsprachen treffen oder wie die Einkommensstruktur einer Drogengang aussieht, und die unorthodoxe Denkweise mit welcher Levitt diese untersucht. Natürlich reicht die Lektüre des Buches mitnichten aus, um selbsttätig anspruchsvolle Statistiken zu erstellen, die Neugier daran wird jedoch geweckt und im Endeffekt steht ja auch die Unterhaltung im Mittelpunkt, sonst könnte man ja auch gleich ein "echtes" Statistikbuch lesen. Die Lektüre lohnt schon allein wegen der ungewöhnlichen Geschichten, die von Levitt vorgetragen werden.Insgesamt ein sehr unterhaltsames Buch, das Statistik sexy aussehen lässt!

       Multiple-Regression Statistical Analysis Put on an Easy-to-Understand Pedestal, 22. Mai 2007

Ask most people if they want to understand statistics . . . and they run in the opposite direction. That's too bad because these days anyone who can run a personal computer can perform sophisticated statistical analysis using relatively affordable software like SPSS. Freakonomics may open a few minds by showing that much of what the conventional wisdom is . . . is wrong.Economics has been traditionally focused on writing equations to explain "how things should work" assuming that nothing else changes. That's the rub. Everything else does change . . . and the theories don't work in practice. You've all heard the resulting economist jokes.Steven Levitt does something that academics don't like anyone to do: He looks for interesting, practical questions and devises simple, straightforward solutions.His method is usually pretty simple. He looks for patterns by using regression programs and then thinks about what the regressions might mean. That often leads to a trip to some other data, and eventually the correct cause-and-effect pattern emerges. It's like the invention methods of champion tinkerer Thomas A. Edison. Keep trying until something practical works. Fortunately, with today's computers you don't have to wait very long. The biggest challenges are in finding the right data sets, as this book shows through its example of why drug dealers usually live with their mothers.The book indicts the media and many so-called experts who simply haven't done their homework. As a result, you can spend a lot of time being misinformed by reading the latest Congressional testimony, the latest think-tank study or by watching a talking head debate on television. The lesson: Be skeptical unless you see the data and the analyses, as they are displayed in this book's few examples.In the book, you will find out how statistics can identify some of those who cheat (whether they are teachers or sumo wrestlers) and how economic incentives slant behavior (how real estate brokers sell their own property versus selling yours). You will encounter a novel argument that Roe v. Wade has reduced the violent crime rate. You'll find an even more interesting argument about how to equate the value of reduced crime to the cost of abortions.More favorably, there are case studies on how accurate information trumps bad or misleading information to the benefit of us all.The book ends up on a largely unsatisfying statistical look at nature versus nurture . . . and pretty much dismisses nurture when it comes to child-raising.So it's a grab bag of topics, mixed with lots of hero worship (by co-author Stephen J. Dubner for co-author Steven D. Levitt).Why is this book selling so well? I couldn't figure it out. It doesn't have the elegance and relevance of The Tipping Point. It's about statistics, and hardly anyone wants to read about that.So I asked my wife and younger daughter. They both knew the book was a best seller (obviously it has good media play). They both loved the cover . . . especially the illustration of an apple that when you cut into it reveals an orange. They also liked the title (both finding economics pretty freaky). I nominate whoever came up with that cover concept and title for the best "you can't tell a book by its cover" award for 2005.So what does Freakonomics have to do with apples and oranges? As best I can tell, Freakonomics has very little to do with those fruits in a literal sense. The metaphor seems to be intended to be applied in two ways: First, you have to compare apples and oranges to the right reference to understand what you are examining; and second, sometimes the cause of something comes from an unexpected source when we peel back the skin of surface reality. If you want more, I discuss some applications of the book in my blog posting for today.If you already like and know statistics, you can read Professor Levitt's articles instead of this book. If you like "gee whiz" facts about things you don't know much about, this book is for you.

       astonishing answers to unusual questions, 11. Februar 2007

The Introduction of this book is called: "The Hidden Side of Everything". And this is the programme of the book. Steven D. Levitt, a famous economist, gives unusual answers to seemingly bizarre questions: "What Do Schoolteachers and Sumo Wrestlers Have in Common?" "How is the Ku Klux Klan Like a Group of Real-Estate Agents?" Or: "Why Do Drug Dealers Still Live with Their Moms" (in spite of their apparently high income)? Together with the journalist Stephen J. Dubner whom he got to know through his columns for the New York Times Magazine, he gathers all these strange answers in a book. In fact it is a highly interesting book that might interest every reader who is not satisfied with the first solution for a problem. That means a reader who is keen to find "The Hidden Side of Everything". So it is not a typical book for economists.For example, the author pretends that the enormous drop in the American crime rate is not a result of modern police strategies but of the legalisation of abortion in 1973 by the U.S. Supreme Court. As a consequence, unwanted children with a higher risk to become a criminal, were not born anymore. A thesis hard to accept but convincingly proved by the author.Furthermore, the reader learns how the Crack Mafia in the States is organised (as a highly profitable company) or how he can guess the mother's income and the colour of her skin by knowing her child's name.But why does the author pretend, although all scientific studies say the opposite, that long prison terms and the capital punishment have got a deterrent effect on future criminals? Even if he does not affirm that a great decline in crimes is a result of capital punishment, he does not point out its negative aspects resolutely enough.All in all, you will have a very interesting book in your hands that incites you either to laugh or to reflect. Very often you will only be perplexed: Is it really possible that ...?

       Multiple-Regression Statistical Analysis Put on an Easy-to-Understand Pedestal, 30. Januar 2007

Ask most people if they want to understand statistics . . . and they run in the opposite direction. That's too bad because these days anyone who can run a personal computer can perform sophisticated statistical analysis using relatively affordable software like SPSS. Freakonomics may open a few minds by showing that much of what the conventional wisdom is . . . is wrong.Economics has been traditionally focused on writing equations to explain "how things should work" assuming that nothing else changes. That's the rub. Everything else does change . . . and the theories don't work in practice. You've all heard the resulting economist jokes.Steven Levitt does something that academics don't like anyone to do: He looks for interesting, practical questions and devises simple, straightforward solutions.His method is usually pretty simple. He looks for patterns by using regression programs and then thinks about what the regressions might mean. That often leads to a trip to some other data, and eventually the correct cause-and-effect pattern emerges. It's like the invention methods of champion tinkerer Thomas A. Edison. Keep trying until something practical works. Fortunately, with today's computers you don't have to wait very long. The biggest challenges are in finding the right data sets, as this book shows through its example of why drug dealers usually live with their mothers.The book indicts the media and many so-called experts who simply haven't done their homework. As a result, you can spend a lot of time being misinformed by reading the latest Congressional testimony, the latest think-tank study or by watching a talking head debate on television. The lesson: Be skeptical unless you see the data and the analyses, as they are displayed in this book's few examples.In the book, you will find out how statistics can identify some of those who cheat (whether they are teachers or sumo wrestlers) and how economic incentives slant behavior (how real estate brokers sell their own property versus selling yours). You will encounter a novel argument that Roe v. Wade has reduced the violent crime rate. You'll find an even more interesting argument about how to equate the value of reduced crime to the cost of abortions.More favorably, there are case studies on how accurate information trumps bad or misleading information to the benefit of us all.The book ends up on a largely unsatisfying statistical look at nature versus nurture . . . and pretty much dismisses nurture when it comes to child-raising.So it's a grab bag of topics, mixed with lots of hero worship (by co-author Stephen J. Dubner for co-author Steven D. Levitt).Why is this book selling so well? I couldn't figure it out. It doesn't have the elegance and relevance of The Tipping Point. It's about statistics, and hardly anyone wants to read about that.So I asked my wife and younger daughter. They both knew the book was a best seller (obviously it has good media play). They both loved the cover . . . especially the illustration of an apple that when you cut into it reveals an orange. They also liked the title (both finding economics pretty freaky). I nominate whoever came up with that cover concept and title for the best "you can't tell a book by its cover" award for 2005.So what does Freakonomics have to do with apples and oranges? As best I can tell, Freakonomics has very little to do with those fruits in a literal sense. The metaphor seems to be intended to be applied in two ways: First, you have to compare apples and oranges to the right reference to understand what you are examining; and second, sometimes the cause of something comes from an unexpected source when we peel back the skin of surface reality. If you want more, I discuss some applications of the book in my blog posting for today.If you already like and know statistics, you can read Professor Levitt's articles instead of this book. If you like "gee whiz" facts about things you don't know much about, this book is for you.