Remember that quote made famous by Mark Twain?
“There are three kinds of lies: lies, damned lies and statistics.”
This was a quote used to suggest that the United States’ low PISA scores were being abused to push an education reform agenda because the US has much higher poverty rates than other countries in the PISA study and this falsely skews the US results downward.
Well, I took a second look at the PISA scores and the poverty rates I posted yesterday here; especially the poverty rates. I tried to verify several of the poverty rates from the link I was given and could not. The US Census does not have the poverty rate for the US anywhere near 21.7% even going back to 2009. See here where it is 14.3% in 2009. The only way that one could get to anything close to 21.7% is to use the child poverty rate which can be found here for 2009.
The problem is that I do not believe that the data provided for the other countries is for child poverty alone, but is a national poverty rate; even if the figures could be verified. I found different data outlined below and linked here which is published by the United Nations and compares countries for their “Human Poverty Index”. The HPI is a rate that takes into consideration not only income levels but longevity and education.
The revised table using the HPI data is shown below.
|Country||Human Poverty Index||PISA Score|
*Portugal did not have data in this study. Therefore, I used the data that was in the original table. However, information available from government websites and media suggests that Portugal might have poverty rates as high as 36%. In 2005, it was 18%.
**New Zealand was not in the study either. The poverty rate used in the table is from here.
If this new data is put into a graph, the following results.
This new data still does not seem to show much correlation between poverty and PISA scores. In fact, when you look at where most of the HPIs fall (between 10% and 15%), the PISA scores range from a low of approximately 470 up to a high of approximately 522.
Even if one uses the straight poverty rates (50% of median income) published in the HPI data, the following graph results.
*I used 15% for New Zealand, which is the rate carried through from original data as there is no data for NZ on HPI tables and I used 18% for Portugal as that seems to correlate with what Portugal was reporting for itself around the time the PISA scores in 2009 were released.
Not only is the relation between poverty and PISA scores not “clear”, it is not even evident. In fact, there are some countries that are clearly exceeding expectations if one EXPECTS lower scores with higher poverty rates: Japan, Australia, Canada and New Zealand.
As always, I welcome your comments.