Research Cluster 5

Basic principles of effective warning strategies in complex systems

The focus of Research Cluster 5 is to work on basic and applied studies which examine different aspects of validity in alarm systems as well as to examine the behavioral consequences of the operator when handling these systems. Experimental paradigms will be developed for investigating different levels of confidence or trust in the alarm system.

This Research Cluster is supervised by Prof. Dr. phil. Dietrich. Manzey.

Completed doctoral theses:

Currently supervised colleagues/scholarship holders and projects:

 

Decision making in the alarm domain: An experimental assessment of irrational behavior and its specifity to the alarm context

Dipl.-Psych. Torsten Günzler

1. Betreuer: Prof. Dr. Dietrich Manzey  
2. Betreuer: Prof. Dr. Harmut Wandke

Alarm systems are usually designed with the assumption that the user will behave in a rational manner. However, the reaction to an alarm can often be described as a decision under uncertainty - a constellation which constitutes a problem frequently analyzed in decision research for good reason. Thus not too surprisingly, there are reports of irrational behavior also in alarm research, found in rather simple alarm tasks in laboratory experiments. In order to understand the fundamental mechanisms of such behavior it therefore appears sensible and essential to specifically investigate alarm related decision making also within the scope of basic science.

One of the objectives of the research project is the analysis of the specific influence the alarm context might have on decision making. Do people react differently to cues that are perceived as alarms and what does influence their perception of the cues? In order to answer these questions and to improve the general understanding of decision making in the alarm domain experimental Framing studies are conducted. Furthermore it is intended to eventually asses the transferability of general decision research findings.

 

Performance relevant impacts of automatic warning functions of surgical navigation systems

Dipl.-Psych. Maria Luz

1st supervisor: Prof. Dr. Dietrich Manzey 
2nd supervisor: Prof. Dr. Marc Kraft

Surgical navigation systems support the surgeon in spatial orientation in a complicated and confusing site. Doing this they raise patient safety. Such systems are used in ENT surgery, in skull base surgery, and in spine surgery. The so called Navigation Control (NC) system is a recent development. It presents the position of a surgical trephine on the navigation screen, and automatically switches off the trephine, if it reaches the border of a preoperatively defined work space. Thus the NC system actively protects critical anatomical structures from unintended damages.

Previous studies have shown that, using the NC system, apart from the advantage regarding patient safety a lot of switch-offs occur due to technical reasons (loss of line of sight). The frequent interruptions of the surgeons’ workflow result in prolonged surgery duration and increased frustration. It was concluded that, to reduce this negative outcome, the switch-off mechanism should be replaced by an alert. The research project studies, whether the alert can reduce the negative consequences and whether advantages concerning patient safety persist.

 

Behavioral Effects of Likelyhood Alarm Systems

Magali Balaud, M. Sc.

1st supervisor: Prof. Dr. Dietrich Manzey

Classical binary alarm systems can lead to unexpected effects like the cry-wolf effect, due to high rates of false alarms. In this case, a loss of trust in the alarm system appears, which eventually causes low or non-compliance. Likelihood Alarm Systems (LAS) provide information about the Positive Predictive Value of events and therefore constitute an interesting alternative to binary alarm systems. LAS allow for a more appropriate compliance and a better allocation of attention. Most of the research to date has been focused on the potential benefits of LAS in terms of user’s performance compared to binary alarm systems. Another focus of research has been the moderation of user’s performance by LAS characteristics or by situational factors. However, the processes affecting the performance have still not been clarified. Is it only a phenomenon of probability matching? Do LAS allow a better allocation of attentional resources due to their greater informativeness? Or does a trust effect occur because of the configuration of the LAS itself?

The goal of this PhD thesis is to identify the processes involved in the use of LAS and to investigate if they are identical or different from binary alarm systems. We will empirically test our hypotheses by running laboratory experiments.

Current theses

Finished theses

"Human Factors Aspekte von Cockpit-Alarmsystemen: Übersicht und Konsequenzen" [Human Factors aspects of cockpit alarm systems: overview and consequences] (Bachelor Thesis)
Martina Becker

"Eine experimentelle Untersuchung zum Einfluss von Risiko auf den Umgang mit Alarmsystemen im Kontext der Mensch-Maschine-Interaktion" [An experimental study investigating the impact of risk on the use of alarm systems in the context of human-machine interaction] (Master Thesis)
Maria Schinkmann

"Umgang mit Alarmen unter Bedingungen hoher Arbeitsbeanspruchung: Eine experimentelle Untersuchung" (Master Thesis)
Ulrike Schmuntzsch

"Verhaltenseffektivität von Alarmsystemen in Abhängigkeit vom Prüfaufwand: Eine experimentelle Untersuchung" (Diploma Thesis)
Juliane Zorn