This belongs into every statistics class: * Use guesstimated data from a time when the basis for the data was not even established ("money" in 1820, when lots

This belongs into every statistics class: * Use guesstimated data from a time when the basis for the data was not even established ("money" in 1820, when lots of people where still exchanging goods or labor). * Just go back far enough that you can claim the "overall trend is positive". Ten years increased poverty levels in the US? Stop whining, it is still better than 56 years ago! * Expand the scope of data collection: Just look how well China and India did. Never mind that their definition of "poverty level" is not even comparable to most other countries. * Correlation equals causation: In a time when, according to some metric, global poverty decreased, so did infant mortality and infectious diseases (using data going back to 1900, again). School enrollment increased (now back to 1820 again!) * Bury a controversial point at the very end: "The World Bank cautions that these gains have been unequal across countries, and the OECD reports that in many cases, the growth that allowed for reduced poverty was accompanied by increasing income inequality. Still, the World Bank has set a goal of eliminating extreme poverty, currently defined as living on $1.90 a day, by 2030. That kind of goal would have been unthinkable in 1950, to say nothing of a century earlier." The author should try to live on that $1.90 a day in every country for a month and then make this same "all is good" statement again, with a straight face.