Monday, September 19, 2011

Lies, damned lies, and statistics

  "Lies, damned lies, and statistics" is a popular phrase that is often used to disparage statistics that do not properly support a particular position. Humans have a fondness for numbers, but unfortunately those numbers are sometimes misused or even abused.

   Mark Twain popularized the saying in "Chapters from My Autobiography," published in the North American Review in 1906. "Figures often beguile me," he wrote, "particularly when I have the arranging of them myself; in which case the remark attributed to Disraeli would often apply with justice and force: 'There are three kinds of lies: lies, damned lies, and statistics.'"

  This phrase cannot be found in any of Disraeli's writings, however; moreover, it did not appear until after his death. It has been attributed to many people before Twain since then. The actual authorship does not really matter, but the phrase does. It is memorable, although misleading, since it suggests that statistics are the worst kind of lies. 
   Statistics are part of everyday life. One cannot avoid them. Over breakfast, you read the sports pages to find out how your favorite team fared, or you turn to the business pages to check how your stocks (assuming you have any) did yesterday. You can also find these and other statistics online during the course of the day.
   Statistics are not necessarily true or false, but like everything else in creation they can be misused, willfully or not, and even blatantly abused. Thus we ought to be very careful in how we interpret the statistics that are fed to us. The media are sometimes guilty of contributing to this misuse. We must learn to take statistics with a large grain of salt (even though salt is dangerous for our health, as statistics show).

   Just a few examples of serious abuse of statistics. Extrapolation is a very common but potentially fallacious one. Pity the poor woman who is subjected to the following argument:
  My graduate students often confuse correlation and causation. The latter is very difficult to prove, as I constantly had to remind them. The chart below illustrates this point well.
   There are other popular statistical devices that are commonly used by writers. I admit that I too have used the following device occasionally in my postings, although I used it in order to illustrate a problem and to motivate a change in attitude rather than to convince people to give up something they enjoy (like ice cream) for an abstraction (such as sending children in Africa to school). Laying a guilt trip on people is not an appropriate use of statistics, but it is easy to do.

   Yet there is an appropriate use of statistics that do not lie, especially those that relate to the natural world.  For example, Russia is the largest country in the world at 17,075,400 sq km and Canada is in second place with 9,970,671 sq km. Such statistics are not debatable, give or take a few sq km. 
   Other statistics are more debatable, however. To give just one example, in Nigeria the results of the 2006 census revealed that the population of Kano in the largely Muslim north was bigger than that of Lagos. Not surprisingly, southern Nigerians dispute these statistics. 
   In fact, they dispute many of the 2006 census results, including the fact that there are at least 3.5 million more males than females of all ages in Nigeria. I was present during this census. It was a mess. Even the dead were counted in some cases in order to inflate population figures.
   Every census in Nigeria has been highly disputed. The reason for the dispute is not really religious, but it is political, since seats in the legislature and federal funds are all apportioned on the basis of population.
   Religion was excluded from the census on purpose, since it is such a divisive issue. Muslims claim that they constitute the majority of the Nigerian population. Christians, not surprisingly, disagree.
   The CIA World Factbook supports the Muslim claim, insisting that 50% of the country is Muslim and only 40% Christian. Regarding the population of major cities, this source reveals that Lagos is more than three times larger than Kano, although it does support the sex ratio result of the 2006 census.
  Operation World, on the other hand, insists that currently the population of Nigeria is 51% Christian and 45% Muslim. These two sources cannot both be right at the same time. However, both are biased, although in different directions. 
  There is one reliable way to discover the religion of people without resorting to a census. Every child in Nigeria is required to have a health card, and mothers must dutifully list the religion of their children. I suggested this as a useful piece of research to some of my students, but no one has done so yet.
  When it comes to religion, Operation World is very reliable. I use it often when researching the faiths of various countries. As with all statistics, it is wise to compare it with other sources. 
   Religion is too important in life to let people misuse statistics to bolster their own faith and to attack other faiths. Yet religion must be talked about openly, and statistics are one way important way to do that.
   Neither the CIA World Factbook nor Operation World can afford to abuse their figures. Yet both have a readership that they are trying to reach using their own statistics. To be aware of this is to be forearmed.
    Another source that I find very helpful is Pocket World in Figures, published by the Economist. In my opinion, this is the best news magazine in the world, and this pocket-sized work is very useful. Its major failing is that it excludes religion, which is not surprising considering the economistic bias of the magazine. If you want to know what country has the highest GDP per head (Luxembourg), here is the source. Bermuda is number two, with a per capita GDP that is more than twice that of the US (which is in 20th place).

   I hope this has been helpful. I am not an expert on statistics. I only want to show how easy it is to misuse them. Statistics are not lies per se, but they can be used to support a lie. As I learned a long time ago, even the greatest lie has to contain a grain of truth if people are going to believe it.
   Let the reader beware!


1 comment: