PERSPECTIVES strength and stiffness in tension, high energy carbon fibers used for structural composites 3. H. G. Chae, S. Kumar, J. Appl. Polym. Sci. 100, 791 (2006). absorption, and electrical and thermal con- will thus likely be processed not from poly- 4. M. L. Minus, S. Kumar, JOM 57, 52 (2005). ductivity. The low density of these fibers acrylonitrile alone but from its composites 5. V. B. Gupta, V. K. Kothari, Eds., Manufactured Fibre would provide further weight savings. with carbon nanotubes. Technology (Chapman & Hall, New York, 1997). 6. L. M. Ericson et al., Science 305, 1447 (2004). Carbon nanotubes can also act as a nucle- If processing conditions can be developed 7. P. Miaudet et al., Nano Lett. 5, 2212 (2005). ating agent for polymer crystallization and such that all carbon nanotube ends, catalyst 8. M. Zhang, K. R. Atkinson, R. H. Baughman, Science 306, as a template for polymer orientation (12). particles, voids, and entanglements are elim- 1358 (2004). 9. K. Koziol et al., Science 318, 1892 (2007); published No other nucleating agents are as narrow inated, this would result in a continuous fiber online 15 November 2007 (10.1126/science.1147635). and long as a single-wall carbon nanotube. with perfect structure, low density, and ten- 10. M. Motta, A. Moisala, I. A. Kinloch, A. H. Windle, Adv. The tensile strength of a poly(vinyl alcohol) sile strength close to the theoretical value. Mater. 19, 3721 (2007). film tripled with the addition of 1 wt % Such a carbon nanotube fiber could have 10 11. V. V. Kozey, H. Jiang, V. R. Mehta, S. Kumar, J. Mater. Res. 10, 1044 (1995). of single-wall carbon nanotubes (13). times the specific strength of the strongest 12. M. L. Minus, H. G. Chae, S. Kumar, Polymer 47, 3705 (2006). Similarly, incorporation of 1 wt % of carbon commercial fiber available today. However, 13. Z. Wang, P. Ciselli, T. Peijs, Nanotechnology 18, 455709 nanotubes in polyacrylonitrile increased many challenges have to be overcome to (2007). 14. H. G. Chae, M. L. Minus, A. Rasheed, S. Kumar, Polymer the tensile strength and modulus of the achieve this goal. 48, 3781 (2007). resulting carbon fiber by 64% and 49%, 15. T. Dumitrica, M. Hua, B. I. Yakobson, Proc. Natl. Acad. respectively (14). Polyacrylonitrile/carbon References Sci. U.S.A. 103, 6105 (2006). 1. K. K. Chawla, Fibrous Materials (Cambridge Univ. Press, nanotube composites have good tensile Cambridge, 1998). and compressive properties. Next-generation 2. P. Smith, P. J. Lemstra, J. Mater. Sci. 15, 505 (1980). 10.1126/science.1153911

ECONOMICS Economic models can benefit from incorporating insights from psychology, but behavior in the lab might be a poor

Homo economicus Evolves on November 11, 2008 guide to real-world behavior. Steven D. Levitt and John A. List

he discipline of economics is built on The watershed for behavioral economics research in this area approaches the problem the shoulders of the mythical species came in the 1970s. Kahneman and Tversky from the opposite direction: Observing an THomo economicus. Unlike his uncle, carried out the landmark study on nonstandard unexpected pattern of behavior (e.g., lower Homo sapiens, H. economicus is unswerv- preferences (6), or what has become known stock markets on rainy days in New York ingly rational, completely selfish, and can as “loss aversion” or the City), one looks for a psycho- 4.5 effortlessly solve even the most difficult opti- “endowment effect,” i.e., logical theory consistent with www.sciencemag.org Low price mization problems. This rational paradigm that losses loom larger than 4 that behavior (in this case, has served economics well, providing a coher- gains in decision-making. High price seasonal affective disorder). ent framework for modeling human behavior. In a similar spirit, models 3.5 Given the wide array of psy- However, a small but vocal movement in eco- of social preferences, such 3 chological explanations from nomics has sought to dethrone H. economi- as human reciprocity, in- which to choose, however, a 2.5 cus, replacing him with someone who acts equity aversion, and altru- researcher undertaking such “more human.” This insurgent branch, com- ism (7–9), and modeling of 2 a task has virtually unlim- Downloaded from

monly referred to as behavioral economics, temporal decision-making Quality delivered 1.5 ited freedom to explain any argues that actual human behavior deviates (10) have substantially observed behavior ex post facto. from the rational model in predictable ways. influenced economic re- 1 Perhaps the greatest chal- Incorporating these features into economic search. These theoretical 0.5 lenge facing behavioral eco- models, proponents argue, should improve approaches are buttressed nomics is demonstrating its 0 our ability to explain observed behavior. by an entire body of empir- Lab treatment Field treatment applicability in the real world. The roots of behavioral economics date ical evidence drawn from Selfish subjects. Quality as a func- In nearly every instance, the back to Adam Smith, who viewed decisions as laboratory experiments that tion of price offered for sports trading strongest empirical evidence in a struggle between “passions” and an “impar- lends strong support to cards of different quality (13). (Left) favor of behavioral anomalies tial spectator”—a “moral hector who, looking their critical modeling as- Results of offering high ($65) and low emerges from the lab. Yet, there over the shoulder of the economic man, scru- sumptions and findings. ($20) prices in a laboratory setting. are many reasons to suspect tinizes every move he makes” (1). Simon’s (2) Most of this research (Right) Results in a natural environ- that these laboratory findings pioneering analysis of bounded rationality eschews a narrow concep- ment with the sellers unaware they might fail to generalize to real represents an early attempt to incorporate cog- tion of rationality, while were participating in an experiment. markets. We have recently dis- Dealers in the lab provide much nitive limitations into economic models, as continuing to embrace pre- cussed (11) several factors, higher quality on average for the same did later work on bias and altruism (3–5). cisely stated assumptions offer, and respond to generous offers ranging from the properties of that produce a constrained in the lab by increasing the quality of the situation—such as the optimization problem. A nature and extent of scrutiny— The authors are in the Department of Economics, University the goods. This suggests that dealers of Chicago, Chicago, IL 60637, USA. E-mail: jlist@ less “scientific,” and in our behave far more selfishly in natural to individual expectations and uchicago.edu view less productive line of settings than in the lab. the type of actor involved. For

www.sciencemag.org SCIENCE VOL 319 15 FEBRUARY 2008 909 Published by AAAS PERSPECTIVES

example, the competitive nature of markets Other field data yield similar conclusions. choice such as information provision or encourages individualistic behavior and selects For example, farm worker behavior is consis- financial education. for participants with those tendencies. Com- tent with a model of social preferences when Behavioral economics stands today at a pared to lab behavior, therefore, the combina- workers can be monitored (14). Yet, this dis- crossroads. On the modeling side, researchers tion of market forces and experience might appears when workers cannot monitor each should integrate the existing behavioral models lessen the importance of these qualities in other, which rules out pure altruism as the and empirical results into a unified theory everyday markets. underlying cause. Being monitored proves to rather than a collection of interesting insights, Recognizing the limits of laboratory be the critical factor. Similar data patterns are allowing the enterprise to fulfill its enormous experiments, researchers have turned to observed in work that explores the endow- potential. To be empirically relevant, the anom- “field experiments” to test behavioral mod- ment effect (15). alies that arise so frequently and powerfully in els (12). Field experiments maintain true Stigler (16) wrote that economic theo- the laboratory must also manifest themselves in randomization, but are carried out in natural ries should be judged by three criteria: gen- naturally occurring settings of interest. Further environments, typically without any knowl- erality, congruence with reality, and trac- exploring how markets and market experience edge on the part of the participant that their tability. We view the most recent surge in influence behavior represents an important line behavior is being scrutinized. Consequently, behavioral economics as adding fruitful of future inquiry. field experiments avoid many of the impor- insights—it makes sense to pay attention to tant obstacles to generalizability faced by good psychology. At the very least, psycho- References lab experiments. logical insights induce new ways to con- 1. N. Ashraf, C. F. Camerer, G. Lowenstein, J. Econ. Perspect. 19, 131 (2005). Some evidence thus far suggests that ceptualize problems and provide interest- 2. H. A. Simon, Q. J. Econ. 69, 99 (1955). behavioral anomalies are less pronounced ing avenues of research. In their finest 3. R. H. Strotz, Rev. Econ. Stud. 23, 165 (1956). than was previously observed in the lab (13) form, such insights provide a deeper means 4. G. S. Becker, The Economics of Discrimination (Univ. of Chicago Press, Chicago, 1957). (see the figure). For example, sports card to describe and even shape behaviors. One 5. G. S. Becker, J. Polit. Econ. 82, 1063 (1974). dealers in a laboratory setting are driven important practical example involves sav- 6. D. Kahneman, A. Tversky, Econometrica 47, 263 (1979). strongly by positive reciprocity, i.e., the seller ings decisions, where it has been shown 7. M. Rabin, Am. Econ. Rev. 83, 1281 (1993). provides a higher quality of good than is nec- that decision-makers have a strong ten- 8. E. Fehr, K. Schmidt, Q. J. Econ. 114, 817 (1999).

9. J. Andreoni, J. Miller, Econometrica 70, 737 (2002). on November 11, 2008 essary, especially when the buyer offers to dency to adhere to whatever plan is pre- 10. D. Laibson, Q. J. Econ. 112, 443 (1997). pay a generous price. This is true even though sented to them as the default option, 11. S. D. Levitt, J. A. List, J. Econ. Perspect. 21, 153 (2007). the buyer has no recourse when the seller regardless of its characteristics. With this 12. G. W. Harrison, J. A. List, J. Econ. Lit. 42, 1009 (2004). 13. J. A. List, J. Polit. Econ. 114, 1 (2006). delivers low quality in the lab experiment. tendency in mind, Madrian and Shea (17) 14. O. Bandiera, R. Iwan Rasul, I. Barankay, Q. J. Econ. 120, Yet, this same set of sports card traders in a worked with a Fortune 500 company to 917 (2005). natural field experiment behaves far more change the default option for the firm’s 15. J. A. List, Q. J. Econ. 118, 41 (2003). 16. G. Stigler, Essays in the History of Economics (Univ. of selfishly. They provide far lower quality on retirement plans, dramatically influencing Chicago Press, Chicago, 1965). average when faced with the same buyer asset allocations. The changes in behavior 17. B. C. Madrian, D. F. Shea, Q. J. Econ. 116, 1149 (2001).

offers and increase quality little in response to induced by changing default rules dwarf www.sciencemag.org a generous offer from the buyer. more “rational” approaches to influence 10.1126/science.1153640

MEDICINE Host and bacterial proteins essential for bacterial survival offer new avenues for developing nonantibiotic-based treatments

Combating Impervious Bugs Downloaded from for infections. Richard P. Novick

taphylococcus aureus has always been United Sates, MRSA infections now terium requires for growth and for detoxify- a serious human pathogen, and during account for more deaths each year than ing ROS. Liu et al. report that certain choles- Srecent decades it has become more AIDS (4). But two reports, by Corbin et al. terol-lowering drugs have an entirely unex- serious owing to its acquisition of antibiotic on page 962 in this issue (5), and by Liu et pected activity against S. aureus—blocking resistance. In the past few years, a new al. in Science Express (6), are cause for synthesis of staphyloxanthin, the pigment strain of methicillin-resistant S. aureus some cautious optimism about new thera- that imparts the organism’s characteristic (MRSA), known as USA300, and its close peutic approaches to treat such infections. color (aureus means “golden” in Latin) and relatives, have emerged that are not only Both studies describe possible strategies also chemically detoxifies ROS. resistant to antibiotics but are more virulent for interfering with the ability of S. aureus to Staphylococcal infection is an especially and highly contagious. MRSA is presently thwart attacks that are mounted by the serious health threat in individuals with weak- spreading throughout the world, in hospitals immune system during infection. Bacteria ened immune systems, impaired circulation and also in community settings where peo- defend against lethal reactive oxygen species (as with diabetics), and surgical wounds. In ple are in close contact (1–3). Indeed, in the (ROS) produced by neutrophils, immune deep-tissue sites, staphylococci can be life- cells that are mobilized to sites of infection. threatening, even in otherwise healthy individ- Corbin et al. show that calprotectin, a well- uals. Staphylococcal abscesses form when the The author is at the Kimmel Center for Biology and Medicine, New York University School of Medicine, New known mammalian calcium-binding protein, host immune system recognizes certain bacte- York, NY 10016, USA. E-mail: [email protected] chelates manganese (Mn2+), which the bac- rial products, including cell wall components

910 15 FEBRUARY 2008 VOL 319 SCIENCE www.sciencemag.org Published by AAAS