Konradin Maier

A semi-automated approach for structuring multi criteria decision problems

This article seeks to enhance multi criteria decision making by providing a scientific approach for decomposing and structuring decision problems. We propose a process, based on concept mapping, which integrates group creativity techniques, card sorting procedures, quantitative data analysis and algorithmic automatization to construct meaningful and complete hierarchies of criteria. The algorithmic aspect is covered by a newly proposed recursive cluster algorithm, which automatically generates hierarchies from card sorting data. Based on comparison with another basic algorithm and empirical engineered and real-case test data, we validate that our process efficiently produces reasonable hierarchies of descriptive elements like goal- or problem-criteria.

A Multidimensional Model of Trust in Recommender Systems

As e-commerce providers increasingly utilize recommender systems (RS) to support their customers, trust is emerging as a key factor for the design of such technologies. Although a considerable number of researchers addressed the issue of trust towards RS, there is still no common understanding of how trust relates to the acceptance of RS and what factors influence the perception of trustworthiness in RS. After a discussion on the peculiarities of RS, we build on a literature review of trust in order to analytically distinguish basic concepts of trustworthiness. We propose an integrated model of trustworthiness that accounts for the multiple dimensions and perspectives on trustworthiness in RS. Additionally, we will point out several implications for practice and conclude with suggestions for further research in this area.

Incomplete Information within Relative Pairwise Comparisons as Utilized by the AHP

Motivated by a rising demand for feasible and practical decision aids this article relaxes the need for complete pairwise comparisons in Analytic Hierarchy Process (AHP) based approaches regularly applied in multiple criteria decision making practice. We propose a method based on the k-walk procedure from Graph Theory to extract a weighted ranking profile from an incomplete pairwise comparison matrix offering increased flexibility and practicability to the user. Our simulation tests show relative advantages over Harker’s method. The connection into Graph Theory offers an understandable interpretation of weight vectors derived from either complete, incomplete, consistent or inconsistent pairwise comparison data.

An Effort Feedback Perspective on Persuasive Decision Aids for Multi-Attribute Decision-Making

Decision strategies and the level of cognitive effort humans devote to decision-making are highly sensitive. This study investigates the role of feedback interventions in decision aids (DAs) to direct the user’s attention and consequently increase the level of effort spent on the thinking in multi-attribute selection problems. Guided by four research hypotheses, we conducted an experiment with two groups, one with feedback enabled, the other one with it disabled, and provide post hoc click data analysis. The self-developed persuasive DA used in the experiment featured a continuous feedback mechanism based on the users investment of time. This DA led the users through a smartphone decision scenario with altering levels of complexity. Results show that normative effort feedback increases the decision maker’s willingness to spend more effort. We provide new evidence supporting the view that DAs should pay more attention to soft persuasion by guiding the decision maker towards working harder rather than only confronting the user with final recommendations.

Success of Multi Criteria Decision Support Systems: The Relevance of Trust

In recent years the concept of trust has found its way into Information Systems research, especially in the field of technology acceptance. Trust is considered an important driver for the success of technologies. In this context, trust is particularly relevant for Multi Criteria Decision Support Systems (MCDSS), due to the inherent complexity of the methods and the often high impact of the underlying decision. In the course of this paper we present a consolidated view on different dimensions of trust and use a multidimensional model to discuss the specific characteristics and dynamics of trust in MCDSS. Using a quantitative empirical study we were able to confirm our theoretical findings and validate the proposed model empirically. After a discussion of the results we conclude our paper and suggest some areas for further research.

The Effects of Effort-Feedback on Time Spent for Information Processing in Multi-criteria Decision Making

When making decisions, many people tend to rely on their instinctive choices instead of trusting and adopting those suggested by decision support systems. Feedback interventions can be used to subliminally direct and assist decision makers’ attention to increase the accuracy of the decision. This paper seeks to further develop our understanding of such persuasive decision aids for a preferential choice problem. For this purpose, we applied an experiment design with either enabled or disabled feedback interventions in a self-developed computerized decision aid with varying decision complexities. We applied continuous feedback mechanisms using normative rules to evaluate the participants’ time investments during their information processing stage. Our findings demonstrate that normative effort feedback impacts both time spent on single items and time spent for overall decision processing. We conclude that effort feedback can be a viable feedback option to implement in decision aids.