attitude questions, those that have an agree-disagree scale. And they are quite popular but they're also quite troublesome, so we thought it might be a good idea to introduce you to some issues related to these agree-disagree scales. Here's an example, "To what extent do you agree or disagree with the following statements?" "Abortion should be legal" would be the first statement. And then, "I am certain of my views on abortion", would be the second statement. Each of these statements can be answered with a five point scale. The endpoints are labeled "strongly agree" and "strongly disagree" and in between fully labeled "agree," "neither," "disagree." Very common scale labels, very common type of question. What is the problem here? The advantage for this format is that they are pretty easy to administer. People are familiar with that format, they have seen it before. In the paper by Krosnick from 2012, he states that respondents are about two thirds the time faster to this format than other formats, that we have few I don't know answers for this kind of format, and that respondents actually like these kind of scales. However, there are a few disadvantages that we should mention here, and they have been pointed out in earlier literature, but again summarized in the Krosnick paper referenced here. The first one is, they are subject to acquiescence bias. Respondents want to answer polite and in agreeable way. In other context, acquiescence bias means respondents are more likely to answer yes to a question. Here, respondent is on the agreeable side. But also, respondent might defer to higher status and adopt satisficing strategies. So, it's all, you know, leading to more agreeableness or "Yes, sure," "There's a statement. I probably should agree to that." Another problem is that they often are double barreled, in the way they administer it. So, here's an example. "America is getting so far away from the true American way of life that force may be necessary to restore it." That's just, you know, a statement that you can find in a survey and it's pretty easy to ask whether you agree or disagree with this statement. But here you have to agree or disagree to two different things, one has the America getting further away from the true America And the second piece of it, people might have a different opinion in terms of agreeing or disagreeing. Often the disagree end can be ambiguous, so consider this item here, "I'm seldom depressed." So, what does disagree mean, really, for this item? You are never depressed or are you depressed all the time, you know? That can be on two different scale points here and it's not clear what is, what does it mean to disagree with any of these items. So, instead, when, you know, in particular Krosnick argues that you can think of, other ways to ask the same question. So, on the left hand side in this table you see agree-disagree formats for a particular issue, which probably should be avoided. And on the right hand side you see construct-specific response choices that are recommended. So, on the abortion item, "Abortion should be legal" you can see the five point scale again. On the right hand side the same question is reformulated to say, "Some people think that abortion should be legal but others think that abortion should not be legal. Which one is closest to your opinion? Legal or illegal?" That would be one way to ask this question to avoid some of these problems we just talked about. "I'm certain of my views on abortion" is paired up here in a reformulation "How certain are you of your views on abortion? Very certain, somewhat certain, not very certain, not certain at all." So, that too would be an alternative formulation for the same issue. If you read through Jon Krosnick's papers and, in particular, the one we cite here, his verdict is never, ever, ever, under any circumstances ask a question with agree/disagree or strongly agree to strongly disagree scale. Let's try a few more examples listed below here. "The staff worked together as a team to coordinate my care." That could be a question to some patients in the hospital. "Electrons are smaller than atoms." "Research results from industry scientists are less reliable than those from university scientists." To which of those would you be able to have an agree/disagree scale in your opinion and which one would you prefer an alternative formulation? We've provided a sheet here for you so that you can enter your alternative formulations for those questions where you would prefer that. Okay, and on the next segment we'll talk about don't know options.