Testing the bi‐dimensional effects of attitudes on behavioural intentions and subsequent behaviour

Abstract

Attitudes are typically treated as unidimensional predictors of both behavioural intentions and subsequent behaviour. On the basis of previous research showing that attitudes comprise two independent, positive and negative dimensions, we hypothesized that attitudes would be bi-dimensional predictors of both behavioural intentions and subsequent behaviour. We focused on health-risk behaviours. We therefore also hypothesized that the positive dimension of attitude (evaluations of positive behavioural outcomes) would better predict both behavioural intentions and subsequent behaviour than would the negative dimension, consistent with the positivity bias/offset principle. In Study 1 (cross sectional design), N = 109 university students completed questionnaire measures of their intentions to binge-drink and the positive and negative dimensions of attitude. Consistent with the hypotheses, both attitude dimensions independently predicted behavioural intentions and the positive dimension was a significantly better predictor than was the negative dimension. The same pattern of findings emerged in Study 2 (cross sectional design; N = 186 university students) when we predicted intentions to binge-drink, smoke and consume a high-fat diet. Similarly, in Study 3 (prospective design; N = 1,232 speed limit offenders), both the positive and negative dimensions of attitude predicted subsequent (6-month post-baseline) speeding behaviour on two different road types and the positive dimension was the better predictor. The implications for understanding the motivation of behaviour and the development of behaviour-change interventions are discussed.

An application of the prototype willingness model to drivers' speeding behaviour

“…Of the 330 participants who completed Time 1 questionnaire, 60% completed Time 2 questionnaire and were successfully matched, producing a final sample of N = 198 participants. This completion rate compared extremely favourably with previously published research using a similar design (e.g., Elliott, Brewster, Thomson, Malcolm, & Rasmussen, [Study 3]). Overall, the study ran for 9 months, from July 2014 to March 2015.…”

“…2 We acknowledge that the separation of the positive and negative prototype favourability ratings is not consistent with how prototype favourability is often treated in other studies in which researchers have used overall measures of prototype favourability that combine the positive and negative ratings. The factor analytic findings presented in the main text, however, demonstrate that the positive and negative prototype favourability items tapped different (positive and negative) components of prototype favourability and the findings are consistent with research into bi-dimensional attitudes, in which positive and negative valences have been shown to be independent of one another (see Conner et al, 2007;Elliott, Brewster, et al, 2015). Additionally, it should be noted that when the analyses reported in the main text were re-run with an overall measure of prototype favourability (the mean of all the positive and negative items), rather than the separate positive and negative prototype favourability measures, the findings and resulting conclusions were the same.…”

Testing the effects of explicit and implicit bidimensional attitudes on objectively measured speeding behaviour

“…The likelihood of a behaviour being performed is then held to increase with the extent to which it is evaluated positively rather than negatively (e.g., Fishbein, ). More recently, however, attitudes have been conceptualized as bidimensional predictors of behaviour (Elliott, Brewster, Thomson, Malcolm, & Rasmussen, ; McCartan & Elliott, ). This means that attitudes comprise separate unipolar, positive, and negative dimensions (e.g., ‘For me, speeding is not at all positive to extremely positive’ and ‘For me, speeding is not at all negative to extremely negative’).…”

“…Recent research in which attitudes have been treated as bidimensional predictors of behaviour (Elliott, Brewster, et al ., ; McCartan & Elliott, ) has shown that both the positive and negative attitude dimensions independently predict binge‐drinking intentions (Elliott, Brewster, et al ., [studies 1 and 2]), smoking and unhealthy dieting intentions (Elliott, Brewster, et al ., [study 2]), self‐reported speeding behaviour (Elliott, Brewster, et al ., [study 3]), and objectivity measured speeding behaviour (McCartan & Elliott, ). These studies have also shown that the positive dimension of attitude is a significantly stronger predictor of both intentions and subsequent behaviour than is the negative dimension.…”

“…In line with the above discussion, the aim of this research was to test the effects of implicit bidimensional attitudes on behaviour over and above the effects of explicit bidimensional attitudes. In line with previous research (e.g., Elliott, Brewster, et al ., ; McCartan & Elliott, ), hypothesis 1 was that the positive and negative dimensions of explicit attitudes would account for a significant proportion of the variance in behaviour, with the positive dimension being the stronger predictor. Hypothesis 2 was that the positive and negative dimensions of implicit attitudes would account for a significant increment to explained variance in behaviour, with the positive dimension again being the stronger predictor.…”