Yes, I saw that. The text I quoted was the footnote to the first sentence in that paragraph. I thought it pretty clearly indicated that survey respondents were perfect strangers, i.e. not pre-selected.
It's worryingly vague on the precise details of who was asked and who wasn't. 'Random' here doesn't exclude 'picked from pre-approved lists.'
I think you are speculating too much. You do see people doing surveys on the streets here, at least in Shanghai and Hangzhou. I could totally see this survey being done in that way. How would you pre-approve pedestrians? Or do you think the Chinese authorities went through the trouble of generating a list of phone numbers whose owners would answer in a politically acceptable manner?
Also, the statement that "most Chinese say they approve of internet control and management" is a global summary of the survey responses. No question was phrased in exactly that way. Based on what the actual questions were, I think the statement is rather exaggerated and even misleading.
After all, how do Internet users in Western countries feel about Internet control on the following?
87% of internet users would control or manage pornography; 86% violent content; 83% spam or junk mail; 66% advertisements; 64% slander against individuals.
You're assuming it was a face to face survey. The methodology in the full paper says that in fact it was a phone survey.
It does say they were randomly selected. But the sample size of 2000 seems on the small side if you're trying to represent five cities across all age groups and use patterns. Especially when around a third of the sample weren't Internet users.
And there were no questions about censorship, only about 'controlling' specific kinds of traffic.
I can accept this wasn't a pre-approved list, although obviously pre-approved lists do exist, which limits the usefulness of surveys which use them.
I'm not so convinced that the study proves anything much, because if you want make a claim about censorship, it doesn't seem a stretch to require a survey to ask explicit questions about censorship.
The fact that that wasn't possible here is - of course - a kind of censorship itself.
