J. Frank Yates - US grants

There are large, reliable differences in probability judgments made by individuals in Western countries like the United States and those in many Asian countries. Most differences pertain to calibration. (Good calibration exists, for example, if about 40% of the events that are assigned probability judgments of 40% actually happen.) However, differences in discrimination have been documented, too. (Discrimination is the ability of judgments to sharply distinguish instances when a target event will occur from those when it will not.) Several potential explanations of the observed differences will be tested with subjects in the United States and several Asian locations. The major research hypothesis is that the variations rest on cognitive processing habits rooted in different cultural traditions, including ones pertaining to social interactions. Another proposition is that the differences are an artifact of previous researchers' sampling of judgment tasks. According to a third explanation, the distinctions are specific to particular judgment topics and to student subjects. A final explanation says that the differences reflect cross-national variations in the relative importance of various aspects of judgment quality. This research promises insights into universal judgment mechanisms as well as ones peculiar to given cultures. It should also provide guidance about problems in cross-national communication and collaboration and for improving decision technologies and their transferability.

9911301

Yates, J. Frank

The Role of Memory Retrieval in Probability Judgment

People are often overconfident about their answers to general knowledge
questions (e.g. whether chimpanzees have longer or shorter gestation
periods than humans). And there is ample evidence that "real world"
judgments, such as physicians' diagnoses of myocardial infarction, are
often less than ideal in a similar way. The proposed research is intended
to further develop and test a recently proposed model which asserts that
features of how people bring to mind ("recruit") arguments pertinent to a
given judgment situation lead to deficient judgment. This research is
further anticipated to illuminate procedures that yield improved
judgments. The "argument recruitment model" proposes that people rarely
base their judgments on more than a few facts or arguments. And when they
do bring to mind multiple arguments from memory, the ones retrieved later
in the process tend only to support those presented earlier. The model
predicts that judgment should be improved if the amounts and "balance" of
retrieval of facts from memory are changed. Thus, the primary
manipulation to be explored in this research is that of requesting study
participants to recall all of the relevant facts (e.g., similar previous
cases) they can prior to rendering their judgments. Support for the model
will be indicated to the extent that such demands tend to reduce the
overconfidence evident in participants' judgments.

In recent years, decision behavior scholarship has achieved numerous exciting breakthroughs. For instance, we now know where particular chronic decision-making deficits are localized in the brain. We know that certain aspects of people's mechanisms for anticipating future events are characteristically different in some parts of the world than in others. And we also know that people's risk-taking behavior sometimes results from a complex-but orderly-interaction of cognition and emotion. Such advances promise more than a deeper scientific understanding of natural decision processes. They also offer hope for guidance in preventing at least some of the devastating breakdowns of our society's organizations and institutions that are so familiar, from failed corporate governance practices and rules (e.g., Enron) to space shuttle disasters to everyday medical errors that kill thousands each year. Unfortunately, however, it is doubtful that this great promise will come close to being realized. A major reason is that good, solid courses in decision behavior are available to only a remarkably small fraction of the nation's undergraduates and graduate students. Thus, the great majority of the people we could expect to exploit the decision behavior knowledge that is being created are destined to remain woefully ignorant of it. Perhaps even more troubling, this fact necessarily means that future research in the field cannot possibly progress as it should because there will be few trained scientists capable of doing the required work. Moreover, even current research suffers from the absence of the eager, inquisitive, and demanding students who serve as essential catalysts in every vibrant scientific field. The proposed effort is intended to redress this problem.

The core of the project is a small, intensive, "working" conference that is followed by print and electronic publications as well as symposia at annual meetings of major professional societies. Conference participants will include decision scholars who have demonstrated their excellence not just as researchers, but also as master teachers, communicators, and innovators. The invitees will also include experts in the development of instructional techniques and technologies as well as dissemination. The conference will be preceded by initial surveys intended to help inform conferees of the full magnitude and dimensions of the decision behavior instruction problem and also its foundations, e.g., in faculty and student incentives. The sessions of the conference will include ones devoted to the topics that arguably should have priority in decision behavior courses, to reasons for the current dearth of decision courses, to techniques and technologies that have established their value elsewhere, and to means for developing new methods and tools. Before they disperse, conferees will design and set in motion concrete action plans. These plans will guide efforts to complete and disseminate the conference's publications. Just as importantly, these plans will also result in specific teaching and development efforts at institutions in this country and elsewhere.