In reply to David Hardman's comment on my thoughts about the heuristics and biases paradigm and the current approach to the whole of psychology, I went back and read Dhami's (2003) write up of the bail decision study, trying to keep a more objective stance. It seems that I had made some unfounded assumptions at the first place, which warrant further attention to my own reflective processes (more on this later). To begin with I failed to see Dhami's findings in their particular context. That is, the paper discussed bail decision and not jury or juror decision making processes. There is certainly a qualitative difference between a bail or jail decision by a magistrate, pending later trial, and later stage legal decision making. The pescriptive validity of fast and frugal heuristic (at least in the context described in the study discussed here) is evident. Nonetheless, what is the importance of judgemental accuracy of such heuristics if people don't actually use them? There is nothing to suggest that we can be certain that people actually use such heuristics in general or exclusively, and even if they do under certain circumstances (i.e. time constraints) it seemes dubious to generalise.
To stay within the context of decision making in courts, consider the widely popularised R v Maxwell case. In January 1996, Maxwell bros were cleared of all charges to defraud pension funds of £400 million, against the weight of evidence and public opinion. This case is a prime example of a 'perverse acquittal', i.e. a jury decision that goes against public opinion and what the evidence suggests. Surely, jury decision making is a group process and as such different than individual decision making discussed here, but leaving group dynamics or conformity etc, aside, it would be reasonable to suggest that individual jurors 'world knowledge', i.e. their beliefs about 'how things are' may have their role in such cases and some form of cognitive architectural equation would be unlikely to predict the outcome in such cases. Thus, jurors may as well employ heuristics but a judgement may also be influenced by individual differences and by the use of one's own world knowledge. Enquiring into the phenomena of consciousness, by asking people what they think while judging, or, from a more quantitative point of view, measuring attitudes and biases by using scales and questionnaires would also elucidate aspects of judgement other than heuristics.
Therefore one's world knowledge may affect one's decision/behaviour/judgement, and particularly on my own world knowledge, this is exactly what happened when I wrote the previous post. That is, the way I see the world, and thus the way I see psychology, cognitive psychology and research methodology caused me to adopt a somewhat close-minded attitude, when reading about the role of heuristics in human judgement. On second thought, heuristics seem to be part of the picture and as such give a better insight to phenomena of decision making (even though I still don't see how a mathematical equation could address the complexity of judgement and not miss exceptional cases and more subjective aspects). It is important not to see things as mutually exclusive and to return to my previous post, both quantitative and qualitative methodologies seem to have their own strengths and limitations but could perhaps be used to describe different aspects of reasoning.
Dhami, M. K. (2003). Psychological models of professional decision-making. Psychological Science, 14, 175-180.
Tuesday, 17 November 2009
Monday, 9 November 2009
On Judges, Judging...or the quest for 'essence' in psychology
In an applied Cognitive Psychology module concerned with Judgement processes such as the one for which this blog is devised and undertaken, it seems almost natural to talk about those who judge by profession, i.e. court judges. Drawing upon a research article by Dahmi (2003) that explored the processes of decisions taken by professional judges in courts, I have had to participate in devising a small presentation on the role of heuristics in such contexts.
In contrast to legal theory, which in a somewhat commonsensical way, assumes that judges evaluate all information that is available to them in order to arrive to a decision, cognitive psychological research such as the paper mentioned above has claimed that people use heuristic strategies that depart from weighing and combining all available information appropriately. Such a claim was based on Dahmi’s (2003) study that tested the ability of two decision making models to predict bail decisions in courts. On one hand, heuristic strategies search only a part of all available cues and even base a decision on one single cue. For instance the matching heuristic (Dhami & Ayton, 2001) was tested in the aforementioned research as a predictor of judges decisions are taken in a non-compensatory way (i.e. without ‘weighing’ or judging the importance of all pieces of information available). It attracts attention to information that is pre-consciously selected as relevant (Evans, 1998) as follows: If the first cue that is encountered gives a reason for a punitive decision then such a decision will be predicted, while if the first cue doesn’t give a reason for being punitive the second one will be examined. If the second cue gives a reason for a punitive decision then such a decision can be predicted, if not a non-punitive decision will be reached (it should be noted that a subset can encompass more than two cues). For instance, in the court study prosecution requests and previous convictions where two cues that were used in such a way. In contrast, historically it has been proposed that linear regression models are used to reach judgements. That is, all available pieces of information are searched, differentially weighed and combined to reach a decision (Dahmi, 2003). Such a process is described by Franklin’s rule (Dahmi, 2003). When Benjamin Franklin, one of the founding fathers of the USA, had an important decision to make, he would examine the situation and make a list of all the factors favouring each of his possible options (i.e. pros and cons). Then based on all the information on his list, he would make his decision. In a computational model, such a principle would mean that a weighting function would be applied to all cues available and therefore a decision would be reached in a due process, slowly and statically. Psychological research such as the study described here, has indicated that judges bail or non-bail decision processes are ‘fast and frugal’ (Dahmi, 2003). Judgements are better predicted by a simple matching heuristic, than a process integrating all available information, and might even be based on a single cue!
At first, such findings contradict legal theory and perhaps public perception of justice, That is, people tend to think that a judge would examine all available information carefully to arrive to a decision that might have a dramatic impact on someone’s life. Yet what is demonstrated by cognitive psychology here is that out perceptions of justice may be unrealistic. At some level there is something horrific about this. The almost un-thoughtful rapidity and perhaps ‘close-minded’ equations of different (but similar) cases make me wonder how many court rulings have been as ‘fast and frugal’ as implied here as opposed to the analytical, thoughtful and carefully weighed decisions put forward by the legal discourse.
It was suggested in class that such findings highlight the need for cognitive training of judges in order to help them consciously use more of the information available. Suggestions even went as far as to suggest that a computational equation derived by simple heuristics would actually be well suited to ‘decide and judge’ rather than people doing so. Nevertheless I am still grappling with the idea that decision making is in fact ‘fast and frugal’, especially in courts or similar contexts. Firstly, it seems a bit dubious to generalise such findings and prescribe simple heuristics as ‘explanations’ of human judgement. If anything, such heuristics simply describe processes but don’t explain much about them. Still, just because a descriptive model is a better predictor of final decisions, doesn’t necessarily mean that decisions were actually reached the way the model prescribes. Moreover, assuming that shorter decision times mean less information examined ignores emotional influences and individual differences of those who decide and different characteristicsbetween cases to decide upon..
Being the liberal that I am, I find the idea of reducing human judgement to a computational equation sinister. Such an approach seems valid as a description (if we accept the inherent assumptions that it entails) but has little or no prescriptive validity. The idea that a computational heuristic can simulate human judgement seems constructed and arbitrary as it ignores the phenomenological characteristics of human judgement. That is, it attempts to study the content of though than the actual experience of thinking and deciding and by doing so it ignores consciousness. Why wouldn’t a cognitive psychologist enquire into the phenomena of consciousness by asking people (judges in particular) how they arrived at a particular judgement? How can one assume that this is irrelevant? How can one assume that all cases are treated the same way and that qualitative characteristics of those cases are dismissed or that the emotional influences that such cases or particular characteristics of those cases induce don’t come to play? Why is there a problem with each case being treated as an exception?
It is ironic that cognitive psychology has left experiential and subjective aspects of experience behind and it has attempted to reduce thinking and deciding to an information processing mechanism, leaving the role of feelings, affective and non-affective or sensory, aside (Clore, 1992). From the spreading activation model of memory (Collins & Loftus, 1975) to the ‘revolutionary’ insights to perception, that might actually be more relevant to ophthalmology and not psychology (Klein, 1989) subjective experience, emotion (as such), and individuality have been overlooked.
Yet I feel it is important here to position myself in relation to the concepts discussed here and my own feelings about these concepts described above. As a very liberal individual, I often think of science as socially constructed. I have often questioned the so called scientific methodology and the essentialist assumptions that it entails and I have advocated qualitative research in psychology as an alternative to the prevailing quantitative methods employed. Cognitive psychology has always seemed dubious to me, throughout this degree course as it seems to ignore the very notion that gave birth to it: cognitivism. Dhami’s (2003) study is, to me, another example of how the discipline has been reduced to the study of information processing and I can’t help but wonder: Do we need more ‘quality’ in cognitive psychology? Do we need to revisit the concept of ‘cognitivism’ that once gave rise to this discipline as alternative?
References
Clore, G. (1992). Cognitive Phenomenology: Feelings and the Construction of Judgement. In L. L. Martin & A. Tesser (Eds.). The construction of Social Judgement (pp. 133-163). Hillsdale, NJ: Erlbaum.
Collins, Allan M.; Loftus, Elizabeth F., "A spreading-activation theory of semantic processing", Psychological Review, 82, 407-428.
Dhami, M. K. (2003). Psychological models of professional decision-making. Psychological Science, 14, 175-180.
Dhami, M. K., & Ayton, P. (2001). Bailing and jailing the fast and frugal way. Journal of Behavioral Decision Making, 14,141-168.
Evans, J. (1998). Matching Bias in Conditional reasoning. Do we understand it after 25 years? Thinking and Reasoning, 4, 45 – 110.
Kein, P. (1989). Psychology Exposed or The Emperor’s New Clothes. New York: Routledge.
In contrast to legal theory, which in a somewhat commonsensical way, assumes that judges evaluate all information that is available to them in order to arrive to a decision, cognitive psychological research such as the paper mentioned above has claimed that people use heuristic strategies that depart from weighing and combining all available information appropriately. Such a claim was based on Dahmi’s (2003) study that tested the ability of two decision making models to predict bail decisions in courts. On one hand, heuristic strategies search only a part of all available cues and even base a decision on one single cue. For instance the matching heuristic (Dhami & Ayton, 2001) was tested in the aforementioned research as a predictor of judges decisions are taken in a non-compensatory way (i.e. without ‘weighing’ or judging the importance of all pieces of information available). It attracts attention to information that is pre-consciously selected as relevant (Evans, 1998) as follows: If the first cue that is encountered gives a reason for a punitive decision then such a decision will be predicted, while if the first cue doesn’t give a reason for being punitive the second one will be examined. If the second cue gives a reason for a punitive decision then such a decision can be predicted, if not a non-punitive decision will be reached (it should be noted that a subset can encompass more than two cues). For instance, in the court study prosecution requests and previous convictions where two cues that were used in such a way. In contrast, historically it has been proposed that linear regression models are used to reach judgements. That is, all available pieces of information are searched, differentially weighed and combined to reach a decision (Dahmi, 2003). Such a process is described by Franklin’s rule (Dahmi, 2003). When Benjamin Franklin, one of the founding fathers of the USA, had an important decision to make, he would examine the situation and make a list of all the factors favouring each of his possible options (i.e. pros and cons). Then based on all the information on his list, he would make his decision. In a computational model, such a principle would mean that a weighting function would be applied to all cues available and therefore a decision would be reached in a due process, slowly and statically. Psychological research such as the study described here, has indicated that judges bail or non-bail decision processes are ‘fast and frugal’ (Dahmi, 2003). Judgements are better predicted by a simple matching heuristic, than a process integrating all available information, and might even be based on a single cue!
At first, such findings contradict legal theory and perhaps public perception of justice, That is, people tend to think that a judge would examine all available information carefully to arrive to a decision that might have a dramatic impact on someone’s life. Yet what is demonstrated by cognitive psychology here is that out perceptions of justice may be unrealistic. At some level there is something horrific about this. The almost un-thoughtful rapidity and perhaps ‘close-minded’ equations of different (but similar) cases make me wonder how many court rulings have been as ‘fast and frugal’ as implied here as opposed to the analytical, thoughtful and carefully weighed decisions put forward by the legal discourse.
It was suggested in class that such findings highlight the need for cognitive training of judges in order to help them consciously use more of the information available. Suggestions even went as far as to suggest that a computational equation derived by simple heuristics would actually be well suited to ‘decide and judge’ rather than people doing so. Nevertheless I am still grappling with the idea that decision making is in fact ‘fast and frugal’, especially in courts or similar contexts. Firstly, it seems a bit dubious to generalise such findings and prescribe simple heuristics as ‘explanations’ of human judgement. If anything, such heuristics simply describe processes but don’t explain much about them. Still, just because a descriptive model is a better predictor of final decisions, doesn’t necessarily mean that decisions were actually reached the way the model prescribes. Moreover, assuming that shorter decision times mean less information examined ignores emotional influences and individual differences of those who decide and different characteristicsbetween cases to decide upon..
Being the liberal that I am, I find the idea of reducing human judgement to a computational equation sinister. Such an approach seems valid as a description (if we accept the inherent assumptions that it entails) but has little or no prescriptive validity. The idea that a computational heuristic can simulate human judgement seems constructed and arbitrary as it ignores the phenomenological characteristics of human judgement. That is, it attempts to study the content of though than the actual experience of thinking and deciding and by doing so it ignores consciousness. Why wouldn’t a cognitive psychologist enquire into the phenomena of consciousness by asking people (judges in particular) how they arrived at a particular judgement? How can one assume that this is irrelevant? How can one assume that all cases are treated the same way and that qualitative characteristics of those cases are dismissed or that the emotional influences that such cases or particular characteristics of those cases induce don’t come to play? Why is there a problem with each case being treated as an exception?
It is ironic that cognitive psychology has left experiential and subjective aspects of experience behind and it has attempted to reduce thinking and deciding to an information processing mechanism, leaving the role of feelings, affective and non-affective or sensory, aside (Clore, 1992). From the spreading activation model of memory (Collins & Loftus, 1975) to the ‘revolutionary’ insights to perception, that might actually be more relevant to ophthalmology and not psychology (Klein, 1989) subjective experience, emotion (as such), and individuality have been overlooked.
Yet I feel it is important here to position myself in relation to the concepts discussed here and my own feelings about these concepts described above. As a very liberal individual, I often think of science as socially constructed. I have often questioned the so called scientific methodology and the essentialist assumptions that it entails and I have advocated qualitative research in psychology as an alternative to the prevailing quantitative methods employed. Cognitive psychology has always seemed dubious to me, throughout this degree course as it seems to ignore the very notion that gave birth to it: cognitivism. Dhami’s (2003) study is, to me, another example of how the discipline has been reduced to the study of information processing and I can’t help but wonder: Do we need more ‘quality’ in cognitive psychology? Do we need to revisit the concept of ‘cognitivism’ that once gave rise to this discipline as alternative?
References
Clore, G. (1992). Cognitive Phenomenology: Feelings and the Construction of Judgement. In L. L. Martin & A. Tesser (Eds.). The construction of Social Judgement (pp. 133-163). Hillsdale, NJ: Erlbaum.
Collins, Allan M.; Loftus, Elizabeth F., "A spreading-activation theory of semantic processing", Psychological Review, 82, 407-428.
Dhami, M. K. (2003). Psychological models of professional decision-making. Psychological Science, 14, 175-180.
Dhami, M. K., & Ayton, P. (2001). Bailing and jailing the fast and frugal way. Journal of Behavioral Decision Making, 14,141-168.
Evans, J. (1998). Matching Bias in Conditional reasoning. Do we understand it after 25 years? Thinking and Reasoning, 4, 45 – 110.
Kein, P. (1989). Psychology Exposed or The Emperor’s New Clothes. New York: Routledge.
Subscribe to:
Posts (Atom)