There are several times in the past few decades when I have focused on a data set that required my thinking, for important insight. One such data set was evident in the early 1990s, when I looked at global population data and had to figure out what it meant for population to explode in part of the world and contract in the rest (the explosion had been caused by expansion of hygiene, which had ended by then).
The same is true for decades-long data on obesity. The data does not present a clear or coherent picture of the results without careful thinking. I finally developed the model of Darwinian selection in the food marketplace to fully explain the obesity data.
Similarly I was able to understand the ‘08 national crash by carefully examining the financial, equity and banking data.
There are a number of other periods where I have developed new ideas by successfully focusing on unusual data sets.
The most important tool for understanding and analyzing data sets is the notion of a systemic problem. Large populations and large pools of data can only be explained when changes occur because of systemic issues.
It is foolish to blame data changes on greed, failed government policy, political ideology or moral failures. Those are not systemic issues. A systemic issue must be a problem within the system that becomes evident when the system is pushed into a new domain.
The data that we get to examine in this blog is the dramatic decline automobile deaths between 2005 and 2009. The number of deaths declined from 43,400 to 33,800. A truly dramatic decline in automobile fatalities.
First, it is important to reject the nonsense we hear all around us all the time. Apparently, the Chicken Littles who think that texting and talking on cell phones is a danger have not looked at the data.
What happened? I can think of three systemic changes that should be considered.
* In the past 5 to 7 years many states have focused on first-time drivers 16 to 21 and made it difficult to get a drivers license without creating a good driving record.
* Many more cars have side airbags.
* Auto stability systems.
The firt two of these systemic changes make logical sense, however neither effects a large enough population to have shrunk the automobile deaths by the magnitude we are seeing (new drivers might be 5% of the population and new side airbag fitted cars, about the same).
Only auto stability systems, which have been increasing for 10 years, is a good explanation.
I will keep thinking about this issue and I count on my wonderful readers to do their part.
Recent Comments