Introducing and Testing the Creepiness of Situation Scale (CRoSS)

Abstract

When people interact with novel technologies (e.g., robots, novel technological tools), the word “creepy” regularly pops up. We define creepy situations as eliciting uneasy feelings and involving ambiguity (e.g., on how the behave or how to judge the situation). A common metric for creepiness would help evaluating creepiness of situations and developing adequate interventions against creepiness. Following psychometrical guidelines, we developed the Creepiness of Situation Scale (CRoSS) across four studies with a total of N = 882 American and German participants. In Studies 1–3, participants watched a video of a creepy situation involving technology. Study 1 used exploratory factor analysis in an American sample and showed that creepiness consists of emotional creepiness and creepy ambiguity. In a German sample, Study 2 confirmed these subdimensions. Study 3 supported validity of the CRoSS as creepiness correlated positively with privacy concerns and computer anxiety, but negatively with controllability and transparency. Study 4 used the scale in a 2 (male vs. female experimenter) × 2 (male vs. female participant) × 2 (day vs. night) field study to demonstrate its usefulness for non-technological settings and its sensitivity to theory-based predictions. Results indicate that participants contacted by an experimenter at night-time reported higher feelings of creepiness. Overall, these studies suggest that the CRoSS is a psychometrically sound measure for research and practice.

Highly automated job interviews: Acceptance under the influence of stakes

“…On the one hand, this may indicate that Potosky's () aspect of transparency appears to be relatively low for automated high‐stake interviews. On the other hand, this is in line with former research that suggested that feelings of uncertainty and ambiguity are especially prevalent in novel situations where people do not know what to do, what to feel, or how to judge the situations (Langer & König, ; Tene & Polonetsky, ). This might also account for the result that the low‐stake interview was perceived as being less ambiguous, as it might be more familiar to use automation for low‐stake scenarios such as training or work scheduling (see also Lee, ; Zyda, ).…”

“…One reason for this might be that people only observed the situation; thus they might not have experienced enough emotional immersion into the situation. This finding might also indicate that perceptions of the virtual character in the current study were not affected by potential negative emotional effects caused by the uncanny valley (Kätsyri et al, ; Langer & König, ; Mori, ; Mori et al, ).…”

“…Creepiness was measured with 10 items from Langer and König (). The creepiness scale consists of two facets, namely emotional creepiness and creepy ambiguity, both measured with five items.…”

“…Furthermore, another potentially negative influence of using virtual characters comes from the uncanny valley hypothesis (Mori, ; Mori et al, ). This hypothesis suggests that acceptance of virtual character rises as they become more human‐like up to a certain point where acceptance rapidly drops—this drop in acceptance is associated with negative feelings that constitute the uncanny valley (Langer & König, ). Consequently, having a virtual interviewer might lead to negative feelings elicited by an uncanny valley effect.…”

“…Specifically, we examine social presence (the feeling of present interpersonal warmth and empathy during an interaction; Walter, Ortbach, & Niehaves, ) interpersonal treatment (feelings of being treated with respect and dignity; Bauer et al, ), perceived behavioral control (the belief that one can influence a specific situation [perceived controllability], the confidence of being able to execute a specific behavior and to control a situation [perceived self‐efficacy]; Ajzen, ), consistency (in the sense of equal treatment; Bauer et al, ), and fairness. Furthermore, we examine acceptance variables stemming from research on novel technologies and virtual characters: privacy concerns (concerns about what happens to ones’ private data; Malhotra, Kim, & Agarwal, ), and creepiness (an ambiguous negative affective reaction toward a specific situation, paired with feelings of not knowing how to judge and handle this situation; Langer & König, ). Higher social bandwidth and interactivity in videoconference interviews should condense in the acceptance variables’ social presence, interpersonal treatment, perceived behavioral control, and consistency.…”