Revisiting image theory: decision styles, temptations and image theory’s compatibility test

This thesis project revisits the compatibility test, Image Theory's screening process to form decision choice sets, and considers its elements and mechanisms in the light of three aspects: first, it investigates how the affect heuristic influences the compatibility screening. In this context, the claim of earlier research that only criteria violations are considered during the option screening process is reconsidered; second, a structural model is evaluated establishing links between a decision-maker's decision styles and the variables defining the compatibility test; and third, a neural network is created and tested to predict even irrational choice of decision-makers for a specific screening situation and based on their compatibility test in- and outputs. 741 participants of two populations were administered three online questionnaires to collect required data. 40 questionnaire items have been used to identify the participants decision styles. The participants were tasked to select companies as potential acquisition targets and, thus, performed a compatibility test based on criteria and their importance weights provided by the researcher. Companies met and failed to meet the criteria to differing extent. Two temptation alternatives that outperformed all other companies in the most important criteria multiple times and failed to meet all others were administered to the participants. Based on what companies were selected, the participants rejection threshold and their inconsistent choices were determined. The research provides evidence that the claim of earlier research that Image Theory's compatibility screening process relies only on criteria violations is untenable. Further, a structural equation model was confirmed establishing links between participants' decision styles and the variables defining their compatibility screenings. Eventually, a neural network was generated, trained and tested that correctly predicted with close to 90% reliability a participant's choices, even the objectively irrational ones. It is recommended that future research further develops the idea of neural networks mimicking human decision behaviour.