# 05 - Attitude behaviour correlation

# Attitude behaviour correlation

© SPMM Course 
assume that subjective attitudes can be quantified meaningfully. An alternative method of 
measurement is to observe behaviours, but behaviours do not always reflect attitude. 
 Thurstone scale: While constructing a Thurstone scale, hundreds of statements are 
initially produced pertaining to a particular topic. These statements are presented to a 
sample (similar to a panel of judges) who is asked to score the statements on an 11 point 
scale. A set number of statements e.g. 10 each on both extremes (positive and negative 
attitude) are chosen based on the consistency of scores given by the judges. Each of these 
statements will carry a value, which is the average of 100 judgments on the 11 points scale. 
These 20 statements are clubbed together in producing an attitude scale, which is 
administered to the subject. The subject will then indicate what statements he agrees to. It 
is not often used because the method is too tedious. The 11 points (used to rate each 
statement) are assumed to be intervals and averages are used to obtain the value scores. 
This is not entirely accurate as the 11 points scale is, in fact, ordinal. 
 Likert scale includes graded ‘agree’ to ‘disagree’ measures. This is one of the most popular 
and statistically more reliable measures. It is easy to construct, and no assumptions are 
made about the equality of intervals. 
 Sociometry is used to measure interpersonal attitudes in a repertory grid-like fashion i.e. 
who like whom tables. These are called sociograms. 
 Guttman introduced scalograms that include cumulative statements where accepting a 
statement usually means accepting all that comes below a statement, in a step-wise 
fashion. 
 Osgood’s semantic differential scale is used to measure verbally expressed attitudes. It 
allows different attitudes about a particular topic to be measured on the same scale. It 
includes various factors constituting an attitude; e.g. while expressing one’s attitudes 
regarding a politician, one can rate him using an evaluative component (good ---- bad), 
activity component (active ----- inactive) and potency component (powerful ---- weak) etc. 
With these bipolar adjectives being the two extremes, a 7 points scale is designed, and the 
subject is asked to indicate a score for each factor. Osgood’s semantic differential assumes 
that every concept can be represented in a hypothetical semantic space with two extremes. 
 
Attitude behaviour correlation 
Attitudes and behaviours are not correlated in simple linear fashion. Attitudes are only 
predispositions; actual behaviours depend on: 
1. Perceived consequences. 
2. Social desirability.

© SPMM Course 
3. Habitual behaviours. 
4. Situational factors 
 
If attitudes are measured with specified assessment of target, action, context and time element, 
however, then measured attitudes will be closer to actual behaviour e.g. if one wants to measure 
public attitude on the issue of abortion, simply eliciting attitudes on abortion may not be 
appropriate. Instead if we measure attitudes on legal abortion in a married woman after 3 months 
of marriage, it may provide a more accurate measure of the actual behavior of the respondents 
when the issue arises in their personal or family life. Single instances of behaviours are unreliable 
indicators of attitudes. Various attitudes aggregate to result in behaviour; also the strength of an 
attitude is proportional to its influence on behaviour.