All Terrain Thinking

A Compendium of things I think are Important

Economics: It's not just whats' in your wallet

Overview: Data Analysis
"There are lies, damn lies, and statistics"

How do we make good choices? One of the prerequisites is good information. You will need it when you decide what your major will be and where you will look for a job after graduation; your parents will need it when they are considering refinancing their home or investing in a mutual fund; and American Power Conversion will need it when making the decision on where to locate its next factory or how many power supply units to produce.

But where do you get the information to make the best choices? Increasingly the information needed to make educated choices is presented in quantitative form - statistics on average salaries of college graduates by discipline to be factored into your choice of a major; historic rates of return to guide the choice of mutual funds; and comparative labor cost data for the plant location decision.

The down-side on this move toward summary quantitative representation of information is that as we move to a situation where decisions are based on this quantitative information, we must be aware of the dangers that Darrell Huff pointed out nearly 50 years ago in his little book How to Lie with Statistics. The best way to protect yourself from the 'statistical' lies of others is for you to see first-hand how the lies are created.  If you invest the effort now in understanding the basic questions that need to be asked when creating or interpreting 'numbers', maybe you will not fall victim to those selling you something you really do not want or need. It will also help you get through this course where many of the concepts will be presented in quantitative form.

Where then do we begin? I suggest we begin with some basic questions that should guide anyone interpreting the information supplied by others or creating their own information - a point I first ran across in  Macroeconomics, a book by Michael Rukstad who was attempting to teach macroeconomics with Harvard's case method to MBAs. The five questions that need to be answered are:

WHERE
WHEN
WHICH
WHAT
HOW

You must begin by knowing where to get the data. At the present time you can access data / information in two primary forms, in printed or electronic form. Most of you, as a result of your previous attempts at report writing, are probably familiar with the first of these and have a working knowledge of the library. As for the electronic form of retrieval, information can be accessed electronically, at speeds that will amaze you and make most other methods of search and retrieval seem inefficient.  You are now able to access information via the Internet, and one place to start a search for economic data would be on the Information Sources site for this course.

Once you have tracked down your data, the real 'fun' begins as you attempt to interpret it, or turn it from data into information. It will help if you realize we are truly talking about creating information and in the creation process we must answer four basic questions. We must develop answers to the when question - what time period will we be observing; the which question - which phenomena will be observing; the what question - what aspects of phenomena we will be viewing; and the how question - how will we present our observations.

These are not easy questions to answer and a thorough treatment of them extends well beyond what we will do here.  Although it is an economics course, it is in reality a course designed to prepare you for a world where you will need to create quantitative information and interpret quantitative work done by others.  It is as valuable a skill for a political scientist, a lawyer, a bankers, a buyer, or a chemist. 

In this unit, however, we will focus our attention on the potential value of a good presentation of information, on how quantitative information can be misleading, and on the construction and interpretation of graphs that will be use extensively in the course.

To demonstrate the value of good visuals we will look at some of the work that appears in the work of Edward Tufte.  The first visual is a table which was used in an early trial of John Gotti, the "Teflon don."  In this table, that could be easily produced with any spreadsheet package, the criminal careers of the government's witnesses against Gotti were clearly summarized by the defense attorneys.  By summarizing the criminal track records of the witnesses, the defense was hoping to provide the jury with some reasons to doubt the validity of the witnesses.  Mr.' Polisi's rap sheet was quite impressive, while Mr. Cardinale's credibility would certainly be lowered by his prior conviction on pistol whipping a priest.  As it  turned out, when the trial went to the jury the only piece of evidence from the trial the jury wanted to review was this table.  After their review, they acquitted Mr. Gotti.  The table clearly had an impression on the jurors, which was the goal of the table's creator.

The second visual pertains to data collected during the cholera epidemic that broke out in London on the evening of August 31, 1854.  At that time there was no understanding of the causes of the deaths being recorded and the question was: can the data be presented in such a way as to provide any insight into the causes of the disease? One possibility would be a time-series graph that simply records the dates of the deaths and shows the number of deaths peaking early and then falling continuously.  A second form for presenting the data was produced by John Snow who had suspected contaminated water as the cause of the outbreak.  After a thorough review of the data Snow presented it in a map on which he recorded the residence of those who died and the water pumps.  What he found was the deaths centered around the Broad Street public well, and this was enough evidence to prompt the removal of the handle on the Broad Street pump and the epidemic soon disappeared. 

The final visual pertains to the explosion of the Challenger shuttle on January 28, 1986 that we eventually found to be the result of a failure in an O-ring.  Prior to the Challenger launch there had been a good deal of data collected on the performance of the O-rings in previous launches.   In the first graph we see data ordered chronologically.  Each rocket represents a launch and each black spot represents O-ring damage detected after the rockets were retrieved.  There is no pattern emerging such as the damage happened in the early flights or the later flights.

But what a difference a presentation can make.  In the second graph the launches have been sorted by the temperature at the launch.  What you see here is the damage tends to be concentrated in launches that took place when the temperature was below 70 degrees.  When presented with the data in this form, you would expect the decision to launch the Challenger would have been aborted given the temperature was 29 degrees at launch time. 

As powerful as the visuals can be, they can also be quite misleading, and bad graphics can lead to bad choices.  In this unit we will look at three examples of where a simple graphic has lead many people to derive inappropriate conclusions. The three example are:

• Where would you locate production to satisfy local demand?
• Would you support this person for re-election?
• Has the drug Ansylium reduced Chemical x?

Now its time to work through the simple example of Slippery Slope University to give you a sense of the importance of these questions.  In this example you will be asked to review some information and answer the question:  Does the President deserve our support?  We will end this unit with a review of graphs - how they are constructed and how they can be interpreted.  We will discuss briefly the evolution of graphical representation of data - from the early maps to the time series graphs to the more modern relational graphs.  The emphasis here will be on line graphs which are the most popular tool of the economist.  The best example of a relational line graph used by economists would be the supply and demand graphs which you will see used throughout this course.  Once we have finished this, we will be ready to employ Production Possibility graphs to demonstrate the concept of Opportunity Cost and the Supply & Demand graph to demonstrate the determination of prices in a market system. 

Add to Your Social Bookmarks: Del.icio.us Digg Furl Reddit Simpy Spurl Y! MyWeb - StumbleUpon