Are false alarms not as bad as supposed after all? A study investigating operators’ responses to imperfect alarms
Are false alarms not as bad as supposed after all? A study investigating operators’ responses to imperfect alarms
复制标题
误报真的没有想象中的那么糟糕吗？一项研究调查了操作员对不完美警报的反应。
DOI：
--
复制
发表时间：
2010
期刊：
影响因子：
--
通讯作者：
N. Gérard
中科院分区：
文献类型：
--
作者：
N. Gérard
在二进制报警系统的领域中，当前的研究已经特别解决了假警报的问题，假警报被假定为导致所谓的“哭狼效应”，即操作员对警报的忽视或至少对警报的延迟响应。然而，问题是，这些发现是否可能是由于实验设置迫使操作员在立即对警报做出反应或完全忽略警报之间做出选择。警报系统的第二个问题是失误的行为影响，这被认为是导致警报系统因缺乏对警报功能的依赖而被停用。本实验室实验探讨了误报、漏报和操作员行为之间的关系。实验范式代表了一个双任务环境，其中两个并发任务之一涉及响应二进制报警系统。与早期的研究相比，该范式为参与者提供了通过检查可用的原始数据来主动验证给定警报的可能性。五个不同级别的报警可靠性，操作定义为错过和误报的数量进行了调查。研究结果没有显示出任何狼来了效应的证据。即使在出现大量假警报的情况下，参与者也继续通过对照原始数据核对警报的有效性来对警报作出反应。在警报保持安静的试验中，即使警报系统的细微变化也会强烈影响交叉检查行为。这些结果提供了另一个经验证据，区分两种行为倾向，在响应警报，已被称为依赖和遵守。
In the domain of binary alarm systems, current research has especially addressed the problem of false alarms which are assumed to lead to a so-called “cry-wolf effect”, i.e. a disregard of alarms by operators or at least a delayed response to alarms. However, the question arises if these findings might be due to experimental settings that forced the operators to choose between reacting to the alarm immediately or to ignore it completely. A second issue of alarm systems regards the behavioural effects of misses which are assumed to lead to disuse of alarm systems in terms of a lack of reliance on its alerting function. The present laboratory experiment explored the relationship between false alarms, misses and operator behaviour. The experimental paradigm represented a dual-task environment where one of the two concurrent tasks involved responses to a binary alarm system. In contrast to earlier research, the paradigm offered participants the possibility to actively validate a given alarm by checking available raw data. Five different levels of alarm reliability, operationally defined as number of misses and false alarms were investigated. The results did not reveal any evidence for a cry-wolf effect. Even in case of a large number of false alarms, the participants continued to respond to the alarms by cross-checking the validity of alarms towards the raw data. In trials where the alarm stayed quiet, crosschecking behaviour was strongly influenced by even subtle changes of misses of the alarm system. These results provide another empirical evidence for the distinction between two behavioural tendencies in response to alarms that have been referred to as reliance and compliance.