My first thought was “I wonder if there’s a relation to urban versus rural respondents”. Article didn’t seem to say, but did have this to say about their sample:

"Each country’s sample consists of ca. 2000 individuals in Japan and the United States, ca. 1000 individuals in Australia, Brazil, Canada, China (mainland), France, Germany, Great Britain, Italy, and Spain, and ca. 500 individuals in Argentina, Belgium, Chile, Colombia, Hungary, India, Ireland, Malaysia, Mexico, the Netherlands, Peru, Poland, Russia, Saudi Arabia, South Africa, South Korea, Sweden, Switzerland, and Turkey.

The samples in Argentina, Australia, Belgium, Canada, France, Germany, Great Britain, Hungary, Italy, Japan, the Netherlands, Poland, South Korea, Spain, Sweden, Switzerland, and the United States can be taken as representative of these countries’ general adult population under the age of 75.

The samples in Brazil, Chile, China (mainland), Colombia, India, Malaysia, Mexico, Peru, Russia, Saudi Arabia, Singapore, South Africa, and Turkey are more urban, more educated, and/or more affluent than the general population. The survey results for these markets should be viewed as reflecting the views of the more “connected” segment of their population."

So while it’s mostly representative, there’s room for improvement it seems.