Sequential human redundancy: Can social loafing diminish the safety of double checks?

Abstract

It is often assumed that if two people work on a failure-detection task one after the other, they will observe more failures than when only one person undertakes the task (4-eyes principle). However, human beings have also been found to exert less effort on tasks that they share responsibility for, a phenomenon called social loafing. In the current research, we assessed the effectiveness of sequential human redundancy in light of possible social loafing. In two laboratory experiments, teams of two participants performed a qualitycontrol task in a blinded and in a nonblinded condition, operationally defined by whether or not evaluations of the first checker were forwarded to the second one. In the blinded condition, no social loafing was found, and a near-perfect overall team performance was observed. In contrast, nonblinded redundancy led to a substantial effort reduction of the second checker. However, despite this social-loafing effect at the second position, even nonblinded redundancy led to an overall safety advantage over a single-checker condition. Our research suggests that social loafing in sequential-human-redundancy work settings can occur but does not necessarily reduce the gains in overall reliability. Blinded processes, however, seem to provoke less social loafing than nonblinded processes. Public Significance StatementDouble checks are generally implemented to enhance safety. This research found that specific implementations of double-check processes may lead to reduced individual effort and may therefore limit the positive impact double checks theoretically might have.

Lean back or lean in? Exploring social loafing in human–robot teams

“…Therefore, an experiment was conducted in which participants worked either alone or in a team with a robot on a realistic quality inspection task. Our assumption was that the amount of effort people put into the quality inspection, i.e., the area of the board they searched and/or the amount of time they spent searching, would be lower when working together with the robot than when working alone on the quality inspection, similar to findings of redundant quality control in human teams (Cymek and Manzey, 2022). We also assumed that the individuals working in a team with the robot would reduce their effort over time more than the individuals working alone.…”

“…Specifically, social loafing is higher when the evaluation potential is low, when the task has low perceived value, when a coworker performs well on the task, and when task inputs of the group members are redundant. Social loafing in human teams occurs across different task types and group sizes-even in small teams consisting of only two people (Cymek, 2018;Cymek and Manzey, 2022). For example, in a study by Cymek and Manzey (2022), social loafing was found when two people double-checked the quality of chemical products one after the other.…”

“…Social loafing in human teams occurs across different task types and group sizes-even in small teams consisting of only two people (Cymek, 2018;Cymek and Manzey, 2022). For example, in a study by Cymek and Manzey (2022), social loafing was found when two people double-checked the quality of chemical products one after the other. When individuals in the second position in the quality check experienced that the first person was working almost error-free, they checked the quality less often over time and therefore missed more undetected defects than individuals who did the quality check alone.…”

“…In the current study, we aimed to further investigate the question of the occurrence of social-loafing effects in human-robot teams. While social loafing in redundant quality control has already been demonstrated in humans (Cymek and Manzey, 2022), we wanted to know whether we would also find social-loafing effects in a qualityinspection task performed by a human-robot team, similar to the one described above for electronic components. If social loafing occurs in such a setting, the expected improvement in outcomes due to the redundant quality inspection may not materialize.…”