User:DavidJWeiss/sandbox

David J. Weiss (born February 20, 1944) is an American psychologist and professor. He received his undergraduate degree from the University of Pennsylvania, and Ph. D. from the University of California, San Diego. He spent his teaching career at California State University, Los Angeles, where he retired as Professor Emeritus. He was a Fulbright Scholar to India.

Research
Weiss has contributed more than 65 works to the professional literature, including a statistics text and an edited book featuring papers written by Ward Edwards. He has received grant support from the National Science Foundation, the Federal Aviation Administration, the Air Force Office of Scientific Research, and the Department of Homeland Security.

In his early work, Weiss carried out studies on psychophysical judgment, which had been the corner of functional measurement research assigned to him by his Ph.D. mentor Norman H. Anderson. He exhibited a contrarian streak by demonstrating that magnitude estimation was "biased and invalid" and that the "power law" could not be a general principle no matter what empirical results obtained because the units of the stimuli were arbitrary. After publishing that argument, Weiss decided to move away from perception research.

Theoretical innovations
Weiss is a co-developer of the CWS index of expert performance. He has published several papers illustrating the use of the index, many in collaboration with James Shanteau. CWS serves as a partial theory of expertise because judgment underlies most of what experts do. The index is particularly useful in evaluating performance when no valid external standard is available.

Together with Jie W. Weiss and Ward Edwards, he developed the hierarchical descriptive multi-attribute utility model. This conceptualization attempts to account for everyday decisions by attaching a momentary salience parameter to each anticipated consequence of a decision option, along with the classical subjective value and subjective probability parameters used in subjective expected utility |Subjective expected utility| theory. Consequences not considered at the moment of decision receive zero salience, which mitigates Herbert Simon's objection that utility models cannot be descriptive because the cognitive burden of integrating a large number of potential consequences would be too great. Carefully considered policies are a core element of the model. People set them to simplify the daily repetitive decisions governed by the applicable ones.

Factorial Forecasting was developed to predict how a proposed political action will be received by those whose behavior is required to change if the innovation is to be successful. For example, if a community is considering building a subway system or a bike path, will they be used by the public? Individuals are asked to predict only how they would act if the new measure were in place.

Statistical ideas
Weiss has extended the analysis of variance methodology by considering what to do when participants drop out of a study, suggesting implanting the researcher's best guess as to what the missing score would be when there is an obvious reason for its absence. In another paper, he proposed a test for whether attrition could be considered random.

The type of data suitable for analysis was a concern as well. Weiss showed that, despite classical strictures, ordinal data could be handled well by analysis of variance, with no loss of information for ranks and surprisingly little loss for categorized data. Even nominal data could be handled using an analysis of variance framework. Which data should be analyzed was also a concern. Weiss suggested that in studies of health behavior, the analysis should focus on the behavior rather than on physiological outcomes. For example, a study whose purpose is to increase exercise or reduce caloric intake should examine physical activity or eating rather than weight change.

Weiss showed how one might extract individual contributions to a team's performance, where the analytic challenge is that members have different roles. He illustrated the method using performance data from professional basketball.

A paper written with Jie W. Weiss and Ward Edwards argued that assessing clinical (as opposed to statistical) significance was an important aspect of applied research that could not be handled formulaically, but rather was a decision that required analysis of costs and benefits. With Edwards, he wrote a paper that similarly called for judgment in deciding how to summarize data and whether to apply a transformation.