Notwithstanding choices across alternatives, every behavioural choice may be reduced to the decision to perform or not perform a given behaviour. Theory of planned behaviour (TPB) studies are often conducted with a complementarity assumption—the view that cognitions about action and inaction are conceptually mirrored. It is also assumed that the TPB framework operates in the same way when applied to action and inaction. However, a considerable body of literature identifies that information processing is not equivalent in regard to action and inaction. Recent research suggests that anticipated regret (AR) influences intention to perform a behaviour differently depending on whether the behaviour may be categorised as a distal benefit behaviour or an immediate hedonic behaviour. No previous research has established whether AR exerts a differential influence on intention to perform and not perform target behaviours.

This thesis tests the complementarity assumption by evaluating the psychological distinctness of action and inaction within the TPB framework; determines whether the belief-based and extended direct measure TPB models operate differently when applied to action and inaction; and establishes whether AR exerts a differential influence on intentions to perform and not perform target behaviours. The research therefore provides the first in-depth evaluation of the TPB framework when applied to action and inaction.

Three studies were undertaken into blood donation (study 1), sunscreen use (study 2), and high-calorie snack consumption (study 3). All data were collected using cross-sectional surveys and utilised student (study 1) and Amazon’s Mechanical Turk (studies 2 and 3) samples. Each study was comprised of a belief-elicitation study followed by a main TPB study. Main TPB surveys captured the standard TPB constructs in addition to measure(s) of AR. Studies 2 and 3 included measures of actual behavioural control and behaviour. Studies 2 and 3 also captured belief-based measures relating to both action and inaction. Belief-elicitation study data were analysed using content analysis. The belief-based models were compared using SEM, path analysis, and regression analysis; whereas the extended direct measure TPB models were evaluated using SEM and moderation analysis (examining the moderating role of actual capacity and actual autonomy on the intention-behaviour relationships).

This thesis makes several important contributions to knowledge regarding the nature of action and inaction within the TPB framework. Firstly, identifying that action and inaction are not conceptually mirrored concepts highlights that inferences made about the determinants of inaction based on those about action may be invalid and lead to an inaccurate understanding of the motivating factors that influence inaction. Researchers should therefore not draw conclusions about the motivating factors that influence inaction from research into the determinants of action, and vice versa. The research also identifies a dearth in knowledge regarding the determinants of inaction within the TPB literature. Secondly, the thesis demonstrates that the belief-based and direct measure TPB models operate differently when applied to action and inaction. Future research is necessary to establish whether enhancing the accessibility of beliefs in regard to the less-accessible behavioural alternative leads to the models functioning with similar efficacy for action and inaction. Finally, this thesis identifies that AR exerts a differential influence on intention to perform and not perform a target behaviour which may be because action and inaction constitute different types of behaviour. The distal benefit behaviour category identified by Sandberg and Conner (2008) does not adequately account for the way in which anticipated action regret influences intention to not perform distal benefit behaviours. Behaviour type categories of AR should be extended accordingly. Behaviour type categories should be extended to include a type that accounts for the regret people anticipate in regard to errors of omission.