>I would like to hear more about this random selection process. > I grabbed my list of countries from the 1999 Human Development Report, pasted them into a column of an Excel spreadsheet, used Excel's random number function to assign each one a random number, sorted the spreadsheet in increasing order by the random numbers assigned, and picked the first five. Random samples aren't necessarily representative. Botswana is unrepresentative of Africa; Cuba is unrepresentative of Latin America. But I daresay a larger sample would give very similar overall results: development in GDP per Capita pretty much everywhere except sub-Saharan Africa, development in human indicators--life expectancy, infant mortality, education--pretty much everywhere except in the former Soviet Union. Brad DeLong