Featured Video Play Icon

KEYNOTE: Consequences, Norms, and Generalized Inaction in Moral Dilemmas: The CNI Model of Moral Decision-Making

Bertram Gawronski
Professor, David Wechsler Regents Chair in Psychology
University of Texas at Austin


Abstract. Research on moral dilemma judgments has been fundamentally shaped by the distinction between utilitarianism and deontology. According to the principle of utilitarianism, the moral status of behavioral options depends on their consequences; the principle of deontology states that the moral status of behavioral options depends on their consistency with moral norms. To identify the processes underlying utilitarian and deontological judgments, researchers have investigated responses to moral dilemmas that pit one principle against the other (e.g., trolley problem). However, the conceptual meaning of responses in this paradigm is ambiguous, because the central aspects of utilitarianism and deontology—consequences and norms—are not manipulated. We illustrate how this shortcoming undermines theoretical interpretations of empirical findings and describe an alternative approach that resolves the ambiguities of the traditional paradigm. Expanding on this approach, we present a multinomial model that allows researchers to quantify sensitivity to consequences (C), sensitivity to moral norms (N), and general preference for inaction versus action irrespective of consequences and norms (I) in responses to moral dilemmas. We present 8 studies that used this model to investigate the effects of gender, cognitive load, question framing, and psychopathy on moral dilemma judgments. The findings obtained with the proposed CNI model offer more nuanced insights into the determinants of moral dilemma judgments, calling for a reassessment of dominant theoretical assumptions.

The Paper


2 thoughts on “KEYNOTE: Consequences, Norms, and Generalized Inaction in Moral Dilemmas: The CNI Model of Moral Decision-Making”

  1. Hi Bertram! Very cool and helpful paper! (Also, I’ve been appreciating your comments on other papers in the conference, so thanks for that too.)

    I find myself wanting to hear a bit more about the psychopathy findings, about the omission bias, and about the action-vs-inaction parameter (I). Specifically,

    1. I want to make sure I understand a background point. As I read the paper, your approach shows how prior claims about psychopathy being associated only with utilitarian processing are not quite right. After all, when we zoom in on various factors involved in psychopath’s moral processing (i.e., sensitivity to consequences, sensitivity to norms, and reluctance to act) we find that psychopaths are not disposed toward only the so-called utilitarian response. Rather, they are — depending on the factor — disposed toward both utilitarian and deontological responses. So, in short, psychopath’s net disposition might be closer to the so-called utilitarian response, but it does not follow from that that psychopaths are therefore disposed toward utilitarian responses and not deontological responses. (Please correct me as needed).
    2. I wonder if you measured any metric of analytic/reflective disposition (e.g., need for cognition, the cognitive reflection test, etc.) in studies 2a or 2b. I ask because I am curious to know whether people who score higher on the NFC scale or the CRT would be less likely to end up cognitively loaded and then less likely to show omission bias. After all, it’d be interesting if more reflective people would be more likely to overcome a bias in their moral reasoning.
    3. I want to make sure I understand the third parameter about inaction (I). I understand that the (I) parameter shouldn’t be capturing anything about consequences (C) or norms (N) because those parameters are calculated out before calculating the (I) parameter. But I wonder why we should consider the (I) parameter to be about action vs. inaction in particular rather than, say, everything that is left over once you factor out sensitivity to consequences and norms — which might include more than just preferences for action and inaction. So could you rehearse the reasons for thinking that the third parameter (I) is capturing preferences for action vs. inaction in particular (as opposed to something more general)?

    Thanks again Bertram!

  2. Hi Nick,

    Thanks for your questions. I’m glad you liked the paper. To answer your questions…

    1) I would disagree with the conclusion that psychopaths are disposed to make utilitarian judgments. Our data actually show the opposite: higher levels of psychopathy were associated with a weaker sensitivity to consequences in the CNI analyses. Hence, it makes no sense to say that psychopaths are disposed to make utilitarian judgments (which is the common conclusion from findings with the traditional approach). That being said, you are absolutely right that that the difference between psychopaths and non-psychopaths involves more than that. In our analyses with the CNI model, psychopaths were also less sensitive to moral norms and they showed a greater willingness to act regardless of consequences and moral norms. But to say that psychopaths have a tendency to make utilitarian judgments is misleading if one considers that psychopaths were less (not more) sensitive to consequences in a utilitarian sense.

    2) Although the CNI model has several advantages over the traditional approach (as well as Conway & Gawronski’s, 2013, process dissociation model), it has one major limitation: it is not amenable to individual difference analyses. That’s because the number of dilemmas is too small to fit the model at the individual level. With our validated set of 24 dilemmas, the model is limited to analyses at the group level (e.g., for tests of between-group differences in experimental designs or known-group differences such as gender). Your question is very interesting, but it’s unfortunately not possible to test correlations between the three parameters and individual differences variables such as need for cognition.

    3) The short answer is: the I parameter captures general inaction versus action regardless of consequences and norms by virtue of how it is defined the response patterns in Figure 1. As shown in the table on the right side of the figure, “I” leads to inaction across all four types of dilemmas and “1 minus I” leads to action across all four types of dilemmas. That being said, we cannot rule out that our model is incomplete in the sense that there are other determinants of moral dilemma responses that are not captured by the three parameters. The fact that our model fits the data suggests that it does a good job in describing responses to the four types of dilemmas. But that doesn’t rule out the possibility that there are other determinants that influence moral dilemma responses over and above the three determinants of our model. You are absolutely right that there might be more determinants of moral dilemma responses that are different from what is captured by the C, the N, and the I parameter. I can’t think of anything in particular, but if you have any thoughts on that, I’d be curious to hear them.


Comments are closed.