Deriving accurate measurements of the strength of attitudes (or opinions or beliefs) from quantitative data is not straightforward, as each individual respondent's views will be different. Techniques of attitude measurement have, therefore, been devised to quantify responses to questions of opinion.
Underlying the various approaches to attitude measurement is the notion that there are scales along which individual attitudes can all be positioned. Respondents can be asked a number of questions about aspects of a particular topic and their responses may then combined to find an average score which reflects their overall attitude on the scale. There are three aspects to this.
• Devising attitude statements
In order to examine fully a particular attitude, you need to have a number of statements which relate to it. These can then be presented as a set within the questionnaire for the respondent to select from or rate in some way.
The sources of such statements may be derived from your research questions, the literature review or focus group discussions with members of the proposed target population. Whichever source is used, you need to give them careful consideration. Attitude statements should be meaningful and interesting to the respondents, easy to understand and free from jargon.
• Selecting statements to be included in the final scale
You will need to decide which statements differentiate most effectively between individuals with favourable and unfavourable attitudes towards the topic under consideration. You should discard any statements which do not seem to relate well to the attitude dimensions you are interested in.
When making your selection of statements for the final scale, take care to ensure that they span the full range of attitudes likely to be found in the target population and give careful consideration to the order in which they appear, to avoid introducing bias.
• Checking the reliability and validity of the scale
Once the scale has been finalised, it is essential to check that it measures the variable it is supposed to measure, and not others, and produces consistent results when applied more than once to the same person under similar conditions. This will usually involve a pilot survey, which will also provide you with the opportunity to identify any problems in the administration of the questionnaire before you conduct your full study.
There are three main approaches to attitude scaling - rating scales, Thurstone scales and Likert scales. We shall now examine each one in turn and consider their advantages and disadvantages.