Syndicate content

Attitudes, Opinions, and Why Dinner Matters

Anne-Katrin Arnold's picture

In the general slander of public opinion and public opinion polls ("leaders who pander to public opinion lose respect", see John Kay in the Financial Times), people often mistake attitudes for opinion. It's a technical detail, but from a governance reform view it makes all the difference. Attitudes are predispositions. Opinions are expressions, speech acts. Opinions precede and determine behavior. And that, after all, is where we aim in working toward governance reform.

This technical detail is relevant when we want governments to heed public opinion, and if governments then turn to opinion polls to discern said opinion. Ouch! says the serious student of public opinion. Here's what you get in opinion polls: Someone calls you at a very inconvenient time (dinner's on the stove!) and asks whether you are in favor of the latest policy of the month, somewhat in favor, a little in favor, or not in favor at all. You answer off the top of your head because at this very moment you have neither the time nor the inclination to actually think about this seriously. You just came home after a long work day and care more about dinner than about the policy of the month. So you answer according to your general feeling toward this policy, but you don't consider implications, alternatives etc. If you do, dinner will be cold by the time you've made up your mind, and the pollster will hang up anyway because she gets paid per completed interview. What you get in this kind of snapshot opinion polls are attitudes at best, random impulses at worst.

Attitudes are broad predispositions that are not necessarily consistent with what is eventually expressed and acted upon. According to public opinion scholar Vincent Price, an attitude is “a global enduring orientation" toward issues, whereas, as research psychologist G. D. Wiebe states, opinions are “calculated for purposes of social interaction, for use in public”. Attitudes are stable but barely structured, subjective, and intuitive neural patterns that consist of many different experiences. Attitudes are assumed to be more affective than cognitive, whereas opinions are more cognitive than affective.

If you think your opinion actually matters, you will take dinner off the stove for the time being and consider the question for a moment. Then you will probably start discussing with the pollster about how it all depends. Opinions precede and determine behavior. They are well considered. They are not expressed on the spur of the moment. An opinion does not necessarily represent an individual’s attitude because attitudes are “states of mind unencumbered by the restrictions of deciding what one will actually do about the issue” (Wiebe, 1953, p. 340).

So, if politicians turn to opinion polls as "here's what the people want" they base their policies on my state of mind right before dinner. Researching attitudes is very valuable for psychology, but for students of public opinion and governance, opinions seem more relevant. Unfortunately, these "real" opinions are hard to measure. It takes thorough research and innovative approaches such as Deliberative Polling - which of course limits the scope of what you can find out about public opinion. And in any event, it's not going to happen before I had my dinner.

Picture credit: Flickr user 27147


Submitted by Ximena HC on
Happy to see that someone agrees with me about the value and veracity of opinion polls. I have actually asked one caller to call me back after a while so I could think about his question and answer more accurately. Thanks for a good article.

Submitted by Kay Engelhardt on
Thanks for highlighting the often-overlooked difference between opinions and attitudes! Despite sounding easy at first, opinion polls are actually quite complex when it comes to the details that you outlined. Besides the difference between attitudes and opinions I would like to put two more issues into the equasion: - quality of the questionnaire - at times questions in "opinion surveys" are quite technical and not suited too well for the general public they are aimed at. This has a quite negative impact on the overall quality of the data, as people might just answer "something" (for the sake of not letting the dinner burn on the stove, so to say). - 500 interviews might not be "500 representative" interviews. When a certain number of people just do not have time when the interviewer calls and such people are replaced by people, who have the time, the results might not be representative at all. For example, it is bad practice to call during the national news in certain countries, where one or two news program dominates the media during a fixed time of the day (1900-1930hrs or 2000-2015 hrs): the ones, who want to see the news (and thus can be assumed to be better informed) will not answer, because they want to see the news after all... on the other side, the ones, who have the time (and not watch) can be assumed to be not as well informed as the former.

Add new comment