Today as I read the Daily Nation (A Kenyan newspaper), the front page headline news was about the MP salaries. It was a piece that tried to explain to Kenyans the severity of the situation in terms of how much of a burden the MP's salaries will be to the treasury, as well as comparing average incomes to the MP's salaries. According to the article, currently MP's earn 79 times the average income in the country and this will rise to 113 times the average income in the country. This is shocking I guess, but I feel that the article has missed an important point. The Daily Nation as well as the Central Bureau of statistics forget that it is statistically incorrect to use average incomes in decision making.

I hope you don't mind a short lesson in statistics. I want you to think back to your high school or even elementary maths where you would, given a sample of numbers, be asked to calculate the average, median and mode of that sample of numbers. The average would include adding the numbers up and dividing them by the number of numbers, getting the mode would be found by getting the figure that appears the most times and getting the median would be found by arranging the numbers in ascending or descending order and finding the figure that appears in the middle. Income statistics are meant to be analysed through the median rather than the mean, the reason is simple. The outlier effect, imagine 20 earners, the bottom 10 all earn 50,000 per month, the following 5 earn 75,000, the following four earn 90,000 while the top earns 600,000. The average income from this sample of earners is 91,750 per month. Is this really representative of the sample? no it isn't it is highly skewed to the person who earns 600,000. The best thing to do if one is to analyse the sample of earners is to take their median income which would be 62,500 per month. This number is more representative of the sample even though it heavily discounts the huge earner.

The Daily Nation article should therefore focus on median salaries instead of average incomes especially due to the severity of Kenya's income inequality that would lead to massive outlier effects. However, I do empathise with them as median income statistics are almost impossible to find in Kenya. I lay the blame squarely on the Central Bureau of Statistics. However before I digress, my point is that the discrepancies would be much more alarming had the article used median incomes rather than average incomes. This extends to house prices to income statistics and other ratio's that measure standards and costs of living. The issue is important as decision making is compromised if people use the wrong statistics as they either get a false sense of security or alarm.