Objection! The pollster is leading the respondent...
Clayton Cramer links to (and The Geek with a .45 is aghast at) this newspaper story reporting on a poll which produced somewhat shocking results about free speech rights: 36% said that the press should be required to obtain government approval before publication.
He's right that it's not a terribly encouraging sign, but I don't think that the end of the world -- or even the end of democracy in America -- is coming as a result. Why aren't I more worried? Simple: the pollster manipulated the results in order to produce a good headline. How? Read the extended entry to find out...
On the other hand... well, it's definitely not encouraging even with the manipulation.
The poll presented the following options (paraphrased, since the exact wording is not available):
This trick is called "creating a false dichotomy", and it works by forcing a complex problem into one of two simple answers, neither one of which is actually correct. Respondents end up choosing the response which is closer to their view, or with which they disagree least, even though they may have reservations. The results of the poll are then reported as if all the respondents agreed with the question as stated, as happened in this case.
The results of such a poll can be further skewed by careful selection of the questions. You have one question designed to be broad in scope -- honestly representating about half of the respondents. Your other question is much narrower, representing perhaps 10% of the respondents in an honest poll. But because it's the only alternative, it receives a significant boost from the remaining 40% whose opinions are not honestly represented. Some of those 40% will choose "No opinion", but the rest will select the option into which they have been funnelled.
So how many people approve of our present free speech restrictions -- specifically, laws which punish libel and slander (eg, publication of knowingly false statements)? Is that "publish freely"? Not really. It's not really government approval before publication, either, but it is government approval of a sort, after the fact. How about the FCC? It's not 'publish freely", but it's not prior government approval either. And so on.
The result is that respondents are funnelled into a more extreme position than they actually meant. The poll could have been fixed by adding one or two additional options to more finely capture the distinction:
See how the additional options change the situation?
Check the groups below and enter your email address to receive updates by email:
The trackback URL for this entry is: http://triggerfinger.org/weblog/servlet/trackback/6190
No trackbacks have been posted so far.
No comments have been posted so far.