A Model of Trust

Round 1

Reviewer 1 Report

Referee report on Games manuscript “A Model of Trust”

 

Summary

 

This paper presents an interesting model of trust – the vulnerability model, where the trustor decides the level of trust by weighing the positive utility of the expected return from the trustee and negative utility of the vulnerability from the trusting behavior. Variation on the individual-specific parameter in the vulnerability model could represent different trust attitude.  Through simulation, the author shows that predictions made by the vulnerability model is consistent with behavioral results.

 

Major comments

 

1. It would be helpful if the author could show how the predictions of the vulnerability model differ from the classic selfish model, where players play Nash but has different expectations of the trustee’s return. In the vulnerability model, trustful players are described using higher value of t. They show higher sharing proportions and trust levels. How would the predictions for Nash players with high expectation of return from trustee differ from the current model? Such difference may provide insights on the positive correlation between trusting level and expectation of trustee’s return found in experimental evidence, especially the part of which that could not be fully explained by the rational model. In the repeated trust game analysis, the author may also explore the comparison of the rational model with rational belief updating.
2. As a model trying to capture the behavior of people in the trust game, more experimental evidence is needed to support the parameterization of the model and its predictions. For example, in the discussion of breach of trust and repair of trust, it would be helpful to see existing (or new) experimental evidence on the change of trust level and whether it conforms or not with the model predictions. It would also shed some light on an estimate of the t.
3. It would be helpful to provide more discussion (theoretically or experimentally) on why the parameter tis built into the utility function in this fashion. For example, the psychological foundation for the direct link between the weights on vulnerability and expectation.

 

Minor comments

 

1. The structure of the paper is not very clear. The section titles could be numbered for a better organization.
2. Some descriptions of the model parameters need to be more accurate, for example, on page 6, the trust game generally needs m>1.
3. There are many empty parentheses, for example on page 7 and 11.
4. The paper could also benefit from some proofreading.

Author Response

Please see attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

See attached file

Comments for author File: Comments.docx

Author Response

Please see attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

In "A Model of Trust" the author computationally formalizes verbal models of trust in a simple model that combines current and prospective action values with beliefs and expectancies about a partner's behavior. The resulting vulnerability model generates behaviors that can be interpreted as different trust attitudes, depending on an individual's loss aversion and expectations of the partner's behavior. The author also shows how the proposed vulnerability model can be extended towards new investment game paradigms to account for inferences on different psychological traits of a partner that have been previously described as determinants of trustworthiness impressions central to trust, and argues it can be used as is or as a utility function within more complex Bayesian frameworks to fit participants' behavior. 

How and why trust emerges among humans and how best to describe it theoretically is an intensely investigated subject with obvious practical ramifications. Methods of agent-based simulations and network science have been applied successfully and with much effect recently to shed light on the problem from many different perspectives, and also to outline many different ways on how solutions in terms of better understanding trust could be obtained. In this sense, the study addresses a relevant setup, and it also delivers results that will surely be of interest to the readership of Games.

I have very much enjoyed reading this paper. I find it comprehensive and clearly written, and introducing new, timely, and important results that will surely also inspire future research along these lines. The introduction is also quite comprehensive and informative, although overlooking recent research where trust has been studied in terms of its evolution and promotion. For these reasons, I am in favor of publication subject only to a minor revision.

1) As noted, the introduction should mention recent related work studying the evolution of trust, in The evolution of trust and trustworthiness, J. R. Soc. Interface 17, 20200491 (2020), where such models could be used to further the subject. An up-to-date review of these models is Mathematical foundations of moral preferences, J. R. Soc. Interface 18, 20200880 (2021), where likewise synergies with the newly proposed model could easily emerge.

2) The presentation of the results in terms of figures often turns out quite poor in clarity and quality. I can not see axis labels in many figures, and the space allocated to particular panels is so small that the color maps are not really eligible. This should be improved for better clarity and adherence to standards in scientific publishing.

Apart from the above, I am happy congratulate the author to excellent research.

Author Response

Please see attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

The manuscript has been sufficiently improved to warrant publication in Games.