Carl Westin

26 Strategic Conformance: A Literature Survey either lead to a safe state or a loss of separation. This allowed controllers to freely decide and implement their own preferred solution to traffic conflicts. In a follow-up trial two weeks later, the same controllers interacted with an iden- tical system, only now supported by a decision aid that would recommend solutions by plotting them in the constraint-based interface. This was done deliberately to make controllers believe they could perceive the decision-making criteria of the au- tomation, while also opening up the opportunity to veto automation and implement another preferred solution. In two separate simulations consisting of four scenarios repeated four times, each controller was subjected to a total of 32 advisories (one per scenario). In half of the cases the controller’s own previous solutions were presented (conformal advisories). In the other half, a colleague’s different, but equally work- able and safe solution were presented (nonconformal advisories). Results showed that conformal advisories were accepted 33% more often than nonconformal advi- sories (76.2% and 56.6% respectively). Standardized agreement ratings, measured on a 1-100 rating scale, supported acceptance data, with conformal advisories re- ceiving higher agreement ratings. Controllers also responded on average one second faster to conformal advisories (4.9 s and 5.9 s respectively). This study illustrates that individually matched solutions can indeed improve automation acceptance. To summarize, empirical evidence has shown that acceptance issues of decision aiding automation can be attributed to a mismatch between human and automaton decision-making strategies. Larger mismatch gaps, as exemplified by technology- centered approaches (e.g., URET), result in larger resistance to automation. Reduc- ing the gap, by heuristic forms of automation (e.g., CORA), can benefit acceptance. The greatest benefits, however, may be achieved by tuning decision aiding automa- tion to the individual’s problem-solving style. However, we do not only have to rely on empirical studies to make such a claim, as it appears that other research fields outside of CE explicitly acknowledge the importance of compatibility between hu- man and technology for acceptance. 2-3 Toward a new perspective Acceptance research can, in general, be divided into three main streams depending on the field in which it has been studied. In addition to CE research, addressed in the previous section, the other main streams are innovation-diffusion theories (IDT) originating from the broad fields of sociology and psychology, and technology ac- ceptance theories studied within the information systems community. Surprisingly, both interaction and knowledge transfer across these three communities have been sparse. 103 As such, it is valuable to synthesize and perhaps unite the different views into a single concept that addresses acceptance issues and how to resolve them.