re: #252 Obdicut

It’s a ridiculous quiz on many levels.

For example, the question:

A ‘rock-ribbed’ conservative could honestly answer “Completely agree” to this, because in their (sadly deluded) mind, the government could do more by deregulating the insurance industry, by imposing at the federal level that insurance companies could sell to other states, etc. However, ‘conservatives’ taking this test know that the ‘completely agree’ position would label them as a Democrat, so, even if they think it’s true that the government could take action that would result in more affordable health care, they’ll answer ‘Completely disagree’.

Likewise, “I never doubt the existence of God” somehow means my Catholic parents and all the other observant, religious Democrats aren’t sincere in their religion or something. A real question would be “We should teach creationism in schools”, not about simply having faith.

There are a number of ways of designing surveys to extract attitudes. Back when I went for my Psych degree, we were taught to design surveys where questions were asked repeatedly, generally differently each time. The use of intentional slanting along with more neutral forms allows for removal of false bias such as in your example.

Almost all surveys are designed with the type of false bias you’re talking about in mind, and unless the survey designers intentionally want a false result, and I have seen several like that recently, what may look like poorly formed, or bias laden, questions may not be so poorly chosen.