Here is the original article from Washington Post
Apparently they have messed up the numbers for Bangladesh and Hong Kong (They've reversed the Yes and No in the questionnaires)
See the bottom of the article where author Max Fisher provided an explanation for the correction
"Correction: This post originally indicated that, according to the World Values Survey, 71.7 percent of Bangladeshis and 71.8 percent of Hong Kongers had said that they would not want a neighbor of a different race. In fact, those numbers appear to be substantially lower, 28.3 percent and 26.8 percent, respectively. In both cases, World Values appears to have erroneously posted the incorrect data on its Web site. Ashirul Amin, posting at the Tufts University Fletcher School’s emerging markets blog, looked into the data for Bangladesh and discovered the mistake. My thanks to Amin, who is Bangladeshi and was able to read the original questionnaire, for pointing this out. His analysis is worth reading in full, but here’s his conclusion:
The short answer is, yes, someone did fat finger this big time. “Yes” and “No” got swapped in the second round of the survey, which means that 28.3% of Bangladeshis said they wouldn’t want neighbors of a different race – not 71.7%.
26K Facebook likers and 2.5K Tweeters, take note.
Amin adds, “Bangladeshis are a tolerant bunch — it’s ok to come visit.” The error in the Hong Kong data, first discovered by Chinese-speaking users on Reddit, was flagged by Engadget Chinese editor Richard Lai. Ng Chun Hung, a University of Hong Kong professor who was the principal investigator on World Values’ survey there, confirmed via e-mail that the data had been transposed on the survey company’s Web site. He added that he has written the World Values Survey team to alert it to this and ask it to remove the faulty data. My thanks to him, as well as to Lai and the Reddit users who dug through original Chinese-language survey forms to demonstrate the error."