Development and Evaluation of the Preference for Automation Types Scale

Understanding user preferences for different types of automation is necessary for user-centred designof automation across various sectors. We introduce the Preference for Automation Types Scale (PATS),grounded in the Model of Types and Levels of Automation, providing researchers and practitioners ascale measuring individual preferences for automation vs. human control across four key functions:information acquisition, information analysis, decision selection, and action implementation. Item generation aligned closely with the automation literature and iterated in expert interviews, andpre-testing with users. Three empirical studies assessed dimensionality, reliability, and validity acrossdiverse contexts: generative AI for essay writing (N = 107), an e-commerce decision support system(DSS) for vacation planning (N = 126), and a route-planning DSS used by international seafarers (N =172). Results confirmed theorised dimensionality and demonstrated strong internal consistency andreliability, with the scale effectively differentiating user preferences for automation types acrosscontexts. Construct validity analysis showed that PATS was suitable for evaluating user dispositionstoward automation types but not for detecting shifts in preferences due to system use. Altogether, wedemonstrate that the PATS is a valuable tool for understanding users and work contexts in thedevelopment of automation systems, grounded in a classical human-automation interaction model.