Wednesday, April 16, 2008

Best Places to Live in USA
Another example of poor use of statistics to present information.
Soon to be found in Vic's new book ... (2009 Publication date) "STUPIDOMICS (Stupid in = Stupid out)"

This site is selling a book to people who are looking for a better place to live. The site ranks NJ as Number FIVE !!! It ranks Louisiana as number 50. I have been in both and have lived in NJ and I am confident that parts of LA are far more livable than many parts of NJ. Note in the ranking method they AVERAGE a series of criteria about each state to arrive at a conclusion. I have not read the book, but I can imagine the list includes things like "mean household income" and "percent of population under the US poverty level" - both of which are HIGHLY relative to the geography. A better score on these would be "percent of income spent on housing" and "percent of people spending more than 40 percent of income on food". Also - depending on your own preferences you would apply greater or less weight to the many "livability factors". So, a retiree would rank "highly available geriatric care" with more weight than would a young family. A young family would rank "percent of high school graduates that attend college" with more weight than "percent of population in service sector jobs". Statistics do not "LIE" - they just provide data slices that, while real, are (more often than not) biased to the publisher of the data.
Want some interesting data on states, counties and towns? See THIS LINK from census data.
Post a Comment