Economy, too broad factor

Regression model is one powerful tool of statistics, and is widely applied in many sciences and analyses. People applied it to solve relations between things in order to explain what factors drive an interested variable; for example, suppose that you are McDonalds fans, and you are interested in what factors are relevant to Mc’s burger sales. You can make hypotheses, collect data, and form regression models. Although this tool has some limitation such as if you are interested in some non-quantitative factors such as people taste, health concern, or a collection of Mc’s toys, you must convert them to quantitative scales, its concept is powerful and useful.

In Where have all the criminals gone? chapter, Levitt & Dubner say that a strong economy is definitely not related to crime rate reduction in U.S. in 1990s. (2006, pp. 121 – 122) They give three reasons in order to support their conclusion; that is, a strong economy is related to only non-violent crimes such as burglary and robbery, and not related to violent crimes; however, violent crimes were the major reduction in that period. In addition, they strengthen their conclusion by providing evidence that in spite of high growth economy in U.S. in 1960s, crimes increased significantly.

I like their analyses because they can convey a qualitative and controversial issue to quantitative analyses based on fact and data, and I am practicing the way as they did because this is a good and effective method to make low-error and low-bias results of group meetings. However, this analysis is seemed as it followed the concept of regression, and I only concern that they regress a specific issue on a broad factor; thus, its result generally returns a weak relation or no relation; moreover, I concern that we might ignore some missing links between a strong economy and crime rate. For example, you regress Mc’s burger sales on an economy. You will find that its relation is low or weak (or you fail to reject the claim that its beta value is zero). However, an economy relates some degree with Mc’s burger sales; that is, when Mc’s do its marketing campaign by launching toys’ collection, it will be succeeded if an economy is strong, and some people who are not usually eat Mc’s will spend their wealth consume Mc’s in order to take its toys. Otherwise, this campaign will be failed if an economy is slow down.

Their analyses are convinced me that a strong economy is not a factor that drove crime rate down in U.S. in 1990s, and my concern is a little point for their conclusion. However, I have to raise this concern in order to remind me that their analyses and explanations are the study of specific events, and I cannot throw any factors into a bin if I think they maybe relate with my study when I do my analyses for my tasks by referring their conclusion.

__________

Reference

Levitt, S. & Dubner, S. (2006). Where have all the criminals gone? In Freakonomics: A Rogue Economist Explores the Hidden Side of Everything (p. 121 – 122). London: Penguin Books.