Lee Jussim is the chairman of the psychology department at Rutgers and perhaps the leading expert on stereotypes and bias. He has a new academic book out:
contests the received wisdom in the field of social psychology that suggests that social perception and judgment are generally flawed, biased, and powerfully self-fulfilling. Jussim reviews a wealth of real world, survey, and experimental data collected over the last century to show that in fact, social psychological research consistently demonstrates that biases and self-fulfilling prophecies are generally weak, fragile, and fleeting. Furthermore, research in the social sciences has shown stereotypes to be accurate.
Jussim overturns the received wisdom concerning social perception in several ways. He critically reviews studies that are highly cited darlings of the bias conclusion and shows how these studies demonstrate far more accuracy than bias, or are not replicable in subsequent research. Studies of equal or higher quality, which have been replicated consistently, are shown to demonstrate high accuracy, low bias, or both. The book is peppered with discussions suggesting that theoretical and political blinders have led to an odd state of affairs in which the flawed or misinterpreted bias studies receive a great deal of attention, while stronger and more replicable accuracy studies receive relatively little attention. In addition, the author presents both personal and real world examples (such as stock market prices, sporting events, and political elections) that routinely undermine heavy-handed emphases on error and bias, but are generally indicative of high levels of rationality and accuracy. He fully embraces scientific data, even when that data yields unpopular conclusions or contests prevailing conventions or the received wisdom in psychology, in other social sciences, and in broader society.
The funny thing is that this successful academic has a rather non-academic style. Jussim has summaries of each chapter up online here.
Chapter 17. Pervasive Stereotype Accuracy
Abstract
This chapter reviews every high quality study of stereotype accuracy that I could find. It presents the evidence with respect to personal and consensual accuracy, using both correlations and discrepancy scores (see Chapter 16 for an explanation of what these are). It includes sections reviewing the empirical research on racial, gender, and other stereotypes. When the original studies addressed conditions under which accuracy was higher or lower, that, too, is included here. Furthermore, each study is critically evaluated, highlighting both its strengths and its limitations. Overall, this review indicates that the high quality, scientific research consistently shows that stereotype accuracy is one of the largest effects in all of social psychology.
EXCERPTS:
WARNING: TURN BACK NOW, BEFORE IT IS TOO LATE
This chapter contains contents that may be deeply upsetting to anyone committed to the view of stereotypes as inherently or generally inaccurate and irrational. If you have read this book continuously, you undoubtedly do not need these warnings and know what to expect. However, these warnings are necessary for anyone reading this chapter without reading the rest of the book.
Warning I: DO NOT READ THIS CHAPTER without having first read Chapters 10-12, 15 and 16. You will need those chapters to understand what I mean by accuracy generally, and when I describe the results of the studies reviewed below as showing that people’s beliefs were “accurate,” “near misses” or “inaccurate” in this chapter.
Warning II: DO NOT READ THIS CHAPTER, unless you are willing to consider the possibility that stereotypes are often accurate. DO NOT READ THIS CHAPTER, if you think that merely considering the possibility that many of people’s beliefs about groups (stereotypes) have a great deal of accuracy makes someone a racist, sexist, etc. DO NOT READ THIS CHAPTER if you believe that stereotypes are inherently inaccurate, flawed, irrational, rigid, etc., and this belief cannot be or should not be revised if empirical scientific data fail to fully support it.
