Attitude statements form key parts of many surveys. In this post I’m not going to look at how to generate attitude statements for surveys or at the best scales to use (this subject is dealt with here). Rather I am providing a checklist that will help to present your attitude statements to respondents in the best possible way.
1. Attitude statements should be debatable i.e. they should be worded so that respondents can agree or disagree with them.
The statement should not be a widely accepted fact. For example the statement Mobile phones allow us to communicate with other people is undoubtedly true, everyone will agree with it. Re-word or remove any attitude statements from your survey that everyone will almost certainly either agree or disagree with.
2. Avoid statements without feeling
Try to use more extreme words in attitude statements. For example, Makes cars that are OK will tend not to generate a wide variation in response. The vast majority of people will probably agree. If there is no variation in response the data from the attitude statement is not telling us anything about differences in respondents’ attitudes. We may as well have not asked the question.
The statement Makes great cars is a better alternative as the wording is more emotive, the statement becomes something that you can agree or disagree with. Responses for each brand are more likely to show differences.
3. Attitude statements should focus on just one concept.
Many years ago I worked on a tracking survey for a large telecoms company. One of the attitude statements on the survey was Makes a fair and reasonable profit. Results from the statement were difficult to interpret as we were asking respondents to judge the company profit on 2 dimensions, fair and reasonable, which are not necessarily the same thing. Interpretation was also hindered by the fact that the word fair could be interpreted in 2 ways; either as the amount of profit is just, or that the amount of profit is quite good.
4. Avoid superlative words like best
A few days ago I answered a survey about mobile phone manufacturers. A range of attitude statements was presented for each brand I said I had previously heard of. One of the statements was Makes the best mobile phones.
With this statement each brand cannot be considered in isolation as you are forced to make a comparison. One of the brands is the best and the others aren’t no matter how good their phones are – the use of an attitude statement is inappropriate here.
The statement could be turned into a question asking Which brand do you think makes the best mobile phones? Or a better alternative statement would have been Makes great mobile phones. Each brand is therefore considered in isolation and you can agree or disagree with the statement to a varying extent.
5. Keep attitude statements short
Wordy statements tend to be less easy for respondents to understand and more likely to be subject to different interpretations. Always review the wording of your attitude statements with a view to shortening them. A good target is to try to ensure that none of your statements are more than 15 words long.
6. Include a balance of positive and negative statements
Including only positive statements tends to make respondents bored and prone to giving the same answer for each statement. Including a mix of positive and negative statements is more likely to make respondents consider each statement individually avoiding a pattern of answers forming.
A mix of positive and negative statements can also be a good quality control mechanism. If a respondent gives the same answer for each statement then it is likely that they are cheating and just going through the questionnaire as quickly as possible, a practice called “flatlining”. Any respondents identified as flatliners can then be excluded from the final data set.