Saturday, April 30, 2011

Yemen uprising. Who represent the people?



The ability of the Yemeni president to launch massive demonstrations in his support struck many people. It made the simple minded wonders; if so many people want him to stay then what is the evidence that the majority of people want him to go?
In politics the majority of people are silent. A minority speaks.  One of the ways in which they speak is demonstration. The size of the demonstration gives an idea about the proportion of people the demonstrators represent. But is it an accurate representation??
The question reminds me of the wisdom of random sampling in statistics! For a sample (the demonstrators) to be representative of the whole (the people) it should be a random sample. So in an ideal theoretical situation we should ask expert statisticians to choose a proportion of the community using random tables or whatever method of proper random sampling they suggest and we then ask the opinion of the individuals in the sample about the matter at stake. Their opinions will then represent the opinions of the nation. Of course this is all imagination but it helps to clarify the fact that demonstrators are not accurate representatives of the people. They are, in statistical terms, selected (not random) samples. The man in power has got a clear advantage in recruiting people to his demonstration. He can intimidate, threatens and bribes. His followers can demonstrate without fear of being jailed or killed. On the opposition side, demonstrators risk being jailed or killed and those who recruit them have not got the power to defend them and cannot offer them material advantages. So although the majority of people who want the Yemeni president to stay and the majority of those who want him to depart are silent, the proportions who speak are quite different in the two sides. It is much larger on the side who is in power. A thousand on the opposition side may represent a larger section of the community than ten thousand on the regime side.
I wish random sampling of statisticians can be applied in politics. That is of course a day dream or, if you like, a mental exercise!

No comments:

Post a Comment