Several articles criticising the methodology behind the Washington Post Racism vs. Economic Freedom story have appeared online. The original piece by Max Fisher appeared to show HK to be the most racist state in the world.
“26. For people with these background, which do you find unacceptable to be your neighbour? [show card]”
a – with criminal record
b – different race
c – heavy drinker
d – emotionally unstable
Answer: (1) acceptable (2) unacceptable
The source data for the article presents participant’s answers differently. It summarises responses as if participants were presented with ‘tick boxes’ whereby each person checks off each of the various groups of people they would be uncomfortable living next to. The data for HK is presented as a plain: ‘participant mentioned‘ or ‘participant not mentioned‘, as opposed to ‘acceptable‘ or ‘unacceptable‘ for each one. This suggests the ‘mentioned’ numbers actually related to the ‘acceptable’ column, meaning the figures are inverted. As ‘cf18’ notes…
“…it looks like 78.1% of people in Hong Kong dislike other races… But don’t you find it more unbelievable that the next two questions say 80% people don’t mind having drunks or emotionally unstable peoples as neighbours Or even more crazy stuff on the next page, which say 90% people think it’s fine to have drug addicts as neighbours … The correct result should be: In Hong Kong 78.1% found it acceptable to have different race neighbours, while less than 20% found it acceptable to have drunks, emotionally unstable and drug addicts as neighbours.”
Apple Daily has since reported on the problems with the methodology.
There was also some scathing commentary by Siddhartha Mitter, who also noticed bugs in the data points and found that surveys were conducted years apart in different territories. Some figures were just a few years old, some dated back to the 1990s.
Meanwhile, scholar Steve Saideman showed how different people may hear the question differently. To relate this to Hong Kong – participants may be thinking mostly of neighbours from India, Bangladesh or Pakistan when asked. The ‘neighbour’ question is also far from the best measure of tolerance anyway. Sideman presents an Institute on Public Policy study from 2003 as a better measure of tolerance – it asks a number of questions…
Fisher blocked Siddhartha Mitter from his Twitter account where he has been defending his article…
Coincidentally, readers from red countries are much more likely to say they doubt the methodology behind this study wapo.st/13mbhcm
— Max Fisher (@Max_Fisher) May 15, 2013
…Cool Fisher isn’t being a dick about it. Though he has, however, conceded on several points in a follow-up post here.