Thurstone scale:
One of the best-known attitude scales is the "equal appearing interval" scale developed by Thurstone. To construct the scale, the researcher selects a number of statements - usually about 20 - which cover all the possible attitudes towards the topic of the survey. A panel of "experts" then rate them according to their "favourableness" towards the topic and place them along a scale, ranging from "most unfavourable" through to "most favourable". Each statement should be equally spaced along the scale. The statements are then incorporated at random in a questionnaire and respondents are asked to indicate their degree of favourableness towards the attitudes expressed.
It is assumed that respondents will only agree with statements which are clustered round their own position on the attitude scale, and will disagree with the more extreme statements on either side - it is, therefore, essential to have both highly positive and highly negative statements in the set. An average score is then calculated for each respondent. An extract from a questionnaire using a Thurstone scale is shown below.
Here is a list of statements about attitudes people have towards banks. Please indicate how much you agree or disagree with each statement by placing a tick in the appropriate box.
A questionnaire using a Thurstone scale avoids the problem of respondents not indicating the strength of their views accurately, because judgements about distances between statements have already been made by the panel. However, the construction process is laborious and is dependent upon the views of the "experts" being representative of those of the target population. Furthermore, although a Thurstone scale attempts to achieve an ordinal scale of measurement for attitudes, and hence permit statistical manipulation, there is doubt among statisticians as to whether it can actually do so. The third technique we will examine - the Likert scale - attempts to overcome some of these problems.