Combining factors in predicting behavioural clusters: a connectionist approach

Authors

  • G. Nudelman
  • S. Shiloh

Abstract

Background: Health behaviours differ in their characteristics, suggesting that the predictive value of predictors might differ as a function of the targeted behaviour. Health behaviours can also be clustered in categories according to their similarity. Therefore, consistent with a connectionist approach, our goal was to identify patterns of cognitive constructs that predict expectations to engage in clusters of health behaviours. Method: A sample of lay people (N=1709) judged health behaviours representing clusters from the Health Behaviour Taxonomy on 14 constructs drawn from the literature, and rated their expectations to perform these behaviours. Analysis was conducted using Stepwise Multiple Regression. Findings: Expectations to engage in all behavioural clusters were positively associated to varying degrees with ‘frequency of performance’, ‘perceived behavioural control’, and ‘anticipated regret’, and negatively associated with ‘effort’. Moreover, expectations to engage in each behavioural cluster were primarily predicted by a unique combination of constructs: Nutrition behaviours (e.g., fruits and vegetables consumption) by ‘effort’ and ‘positive affect’; Health Maintenance behaviours (e.g., teeth brushing) by ‘impact on health’ and ‘effort’; and Risk Avoidance behaviours (e.g., seat belt use) by ‘perceived behavioural control’ and ‘anticipated regret’. Discussion: The findings demonstrate the utility of a connectionist approach to the field of health behaviours and support the structure representing their cognitive schema. This approach can help in developing an optimal solution for predicting different behavioural expectations. The identification of unique combinations of predictors for different behavioural clusters suggests that distinct factors need to be emphasized in interventions, according to the targeted behaviours.

Published

2017-12-31

Issue

Section

Oral presentations