Maximizers versus satisficers Decision-making styles, competence, and outcomes

《Maximizers versus satisficers Decision-making styles, competence, and outcomes》由会员分享，可在线阅读，更多相关《Maximizers versus satisficers Decision-making styles, competence, and outcomes（15页珍藏版）》请在装配图网上搜索。

1、Judgment and Decision Making, vol. 2, no. 6, December 2007, pp. 342-350.Maximizers versus satisficers: Decision-making styles, competence, and outcomesAndrew M. Parker1* , Wndi Bruine de Bruin2, and Baruch Fischhoff2,31 RAND Corporation, Pittsburgh PA2 Department of Social and Decision Sciences, Car

2、negie Mellon University3 Department of Engineering and Public Policy, Carnegie Mellon UniversityOur previous research suggests that people reporting a stronger desire to maximize obtain worse life outcomes (Bruine de Bruin et al., 2007). Here, we examine whether this finding may be explained by the

3、decision-making styles of self-reported maximizers. Expanding on Schwartz et al. (2002), we find that self-reported maximizers are more likely to show problematic decision-making styles, as evidenced by self-reports of less behavioral coping, greater dependence on others when making decisions, more

4、avoidance of decision making, and greater tendency to experience regret. Contrary to predictions, self-reported maximizers were more likely to report spontaneous decision making. However, the relationship between self-reported maximizing and worse life outcomes is largely unaffected by controls for

5、measures of other decision-making styles, decision-making competence, and demographic variables.Keywords: maximizing, satisficing, decision making, competence, decision style.1 IntroductionBehavioral decision research (Edwards, 1961; Hastie & Dawes, 2001; Yates, 1990) characterizes behavior in terms

6、 of its consistency with the axioms of utility maximization (Bernoulli, 1738/1954; von Neumann & Morgenstern, 1953). A half-century of research has revealed both consistency with and departures from that norm (e.g., Baron, 2000; Plous, 1993). The latter include “satisficing,” choosing an alternative

7、 that is “good enough,” rather than “maximizing,” selecting the option with the highest expected utility (Simon, 1978). Such strategies can be beneficial if they save enough cognitive effort to justify any loss in expected payoff (Simon, 1955, 1956, 1957).Historically, decision-making research has f

8、ocused on general processes underlying deviations from normative theory, such as satisficing instead of maximizing (Lopes, 1987). More recently, attention has turned to individual differences in decision making (e.g., Stanovich & West, 1998, 2000), asking whether, through preference or ability, indi

9、viduals make decisions in consistent ways, across tasks and situations (Bromiley & Curley, 1992). Individual differences that have been examined include risk aversion and risk judgments (Slovic, 1962; Weber, Blais, & Betz, 2002); preference for rational, intuitive, dependent, avoidant, or spontaneou

10、s decision-making styles (Scott & Bruce, 1985); and decision-making competence (Bruine de Bruin et al., 2007; Finucane et al., 2002, 2005; Parker & Fischhoff, 2005).Building on Simons work, Schwartz et al. (2002) developed a scale measuring the degree to which individuals report trying to maximize,

11、rather than satisfice. It includes items such as “When I watch TV, I channel surf, often scanning through the options even while attempting to watch one program.” The other items capture ways in which one might explore as much information as possible when making a choice. Given the many options ofte

12、n available in modern life (e.g., TV channels, cars, jobs, prospective mates), maximizing is no small feat (Iyengar & Lepper, 1999; Schwartz, 2004a; 2004b; Tversky & Shafir, 1992).Perhaps because of the challenges of successfully implementing a maximizing strategy, people who attempt to do so fare l

13、ess well in life, in the sense of experiencing less happiness, optimism, self-esteem, and life satisfaction, while incurring more depression, perfectionism, and regret (Schwartz et al., 2002). Moreover, individuals who try to maximize may have less constructive decision-making styles (Schwartz et al

14、., 2002). For example, aspiring maximizers make more upward social comparisons, thereby inducing regret and counterfactual thinking about what might have been. They rely more on external information sources (Iyengar, Wells, & Schwartz, 2006), which might lead them to further question their choices.

15、Thus, these decision-making styles may undermine the very satisfaction that attempted maximizers try so hard to achieve (Schwartz et al., 2002).Even when maximizing pays off with better outcomes, satisfaction with those outcomes still may not follow. Iyengar, Wells & Schwartz (2006) found that recen

16、t college graduates who described themselves as maximizers secured jobs with 20% higher starting salaries, but felt less satisfied during the job search and afterward. One possible explanation is that attempting to maximize encourages focusing on one easily compared feature (salary), while neglectin

17、g other features important to job satisfaction. Were that the case, then those who attempt to maximize may make poorer decisions, despite strongly desiring the opposite. Conversely, in decisions that lack easily compared criteria, would-be maximizers may face cognitively intractable situations, like

18、 those that led Simon to propose the advisability of satisficing.Consistent with these hypotheses, Bruine de Bruin et al. (2007) found that people with higher self-ratings on Schwartz et al.s (2002) maximizing scale had lower scores on a measure of Decision-Making Competence (DMC), which is describe

19、d below (r= .19, p.001). In addition, self-identified maximizers also reported worse outcomes on the Decision Outcome Inventory (DOI), which includes 41 negative life events that might reflect poor decision making. These events range broadly in their impacts and frequency, and include ruining clothe

20、s in the laundry, having a check bounce, having a mortgage or loan foreclosed, being in jail overnight, and having been diagnosed with type 2 diabetes (which is more likely among people who have made poor lifestyle choices).The analyses in Bruine de Bruin et al. (2007) focused on developing and vali

21、dating the DMC and DOI measures. In that context, Schwartz et al.s (2002) self-reported maximization scale was one of several comparison measures. As a result, the paper reported zero-order correlations of maximizing with DMC and DOI, but not with other decision-making styles or demographic characte

22、ristics. In particular, the analyses did not examine the extent to which lower DMC scores and problematic decision-making styles account for the correlation between self-reported maximizing and poorer life outcomes. Here, we examine this question, using Bruine de Bruin et al.s (2007) diverse communi

23、ty sample and rich dataset.We begin by asking whether self-reported maximizers tend to report several decision-making styles. One such measure is behavioral coping, or taking action to resolve difficult tasks, rather than working around them (Epstein & Meier, 1989; Katz & Epstein, 1991). Because sel

24、f-reported maximizers may set unattainable goals, they should report less of such coping. Five other measures come from Scott and Bruces (1985) suite of decision-making style scales. Self-reported maximizers should report engaging in (1) more rational decision making, reflecting their perception of

25、systematic deliberation about their choices; (2) less intuitive decision making, attempting to avoid relying on feelings and instincts (e.g., Slovic, Finucane, Peters, & McGregor, 2004); (3) more dependence on others, reflecting interpersonal comparisons and the quest for information; (4) more avoid

26、ant decision making, postponing decisions to search for more information and ponder the possibilities; and (5) less spontaneous decision making, in the sense of taking more time to carefully decide. Finally, we expect self-reported maximizers to report greater regret about their past decisions, repl

27、icating Schwartz et al.s (2002) finding in a diverse community sample.Subsequently, we take advantage of the diversity of Bruine de Bruin et al.s sample to examine how self-reported maximizing varies with socio-demographic variables. Finally, we examine whether the correlations between self-reported

28、 maximizing and the two performance measures, A-DMC and DOI, are reduced after controlling for the other styles and demographics.2 Method2.1 SampleWe recruited 360 people from the Pittsburgh area through social service organizations (46.1%) and other community groups. The social service organization

29、s were located in poorer sections of the city and served disadvantaged populations. Other community groups were located in relatively more affluent locations and did not address the needs of disadvantaged populations. Among participants responding to demographic questions, ages ranged from 18 to 88

30、(M=47.7, SD=17.0); 73.8% were women, 65.5% white, 28.2% African-American, and 6.3% other racial minorities. Highest level of education was 2.8% no degree, 44.6% a high school degree, 13.0% an associates degree, 29.1% a bachelors degree, 9.5% a masters degree, and 0.9% a Ph.D. Except for the proporti

31、on of women, the sample resembles U.S. Census figures for the Pittsburgh area.2.2 MeasuresSelf-reported maximizing. We used Schwartz et al.s (2002) 13-item measure of tending to maximize, rather than satisfice, which uses a scale anchored at 1 (=completely disagree) and 5 (=completely agree).Other d

32、ecision-making styles. We used the 15-item behavioral coping module of the Constructive Thinking Inventory (e.g., “When I realize I have made a mistake, I usually take immediate action to correct it;” Epstein & Meier, 1989; Katz & Epstein, 1991), with a scale anchored at 1 (=definitely false) and 5

33、(=definitely true). Scott and Bruce (1985) provided scales for self-reported attempts to (a) make decisions rationally (4 items; e.g., “I make decisions in a logical and systematic way”), (b) base decisions on intuitions (5 items; e.g., “I generally make decisions that feel right to me”), (c) depend

34、 on others (5 items; e.g., “I often need the assistance of other people when making important decisions”), (d) avoid making decisions (5 items; e.g., “I postpone decision making whenever possible”), and (e) make decisions spontaneously (5 items; e.g., “I make quick decisions”). The response scale wa

35、s anchored at 1 (=completely disagree) and 5 (=completely agree). Finally, we used Schwartz et al.s (2002) 5-item measure of the tendency to feel regret (e.g., “When I think about how Im doing in life, I often assess opportunities I have passed up”), using the same response scale as Scott & Bruce (1

36、985).Adult-Decision-Making Competence. A-DMC has six component tasks, selected to cover the skills central to normative theories of decision making (Bruine de Bruin et al., 2007; Parker & Fischhoff, 2005). The A-DMC tasks were based on ones studied by behavioral decision researchers, drawing on the

37、understanding derived from multiple rounds of experimental research. Resistance to Framing uses valence-framing problems (Levin, Schneider & Gaeth, 1998) to measure whether choices are affected by formally irrelevant variations in how options are described. For example, one pair of items asks for qu

38、ality judgments of ground beef described as either (a) “20% fat” or (b) “80% lean.” Fourteen such pairs are presented in two sets, with one containing the positively-valenced member of each pair, and the other containing the corresponding negatively-valenced items. Recognizing Social Norms asks “out

39、 of 100 people your age, how many would say it is sometimes OK” to engage in each of 16 undesirable behaviors (e.g., “steal under certain circumstances”). These estimates are compared to the percent of respondents from this study who had reported earlier that “it is sometimes OK” to engage in each b

40、ehavior, with each persons score being the within-subject correlation between judged norms and observed norms. Under/Overconfidence uses a 34-item true/false knowledge test, with each answer accompanied by a probability judgment (on a scale ranging from 50%=just guessing to 100%=absolutely sure) tha

41、t it is correct. Each persons score is the absolute value of the difference between their mean probability judgment and the actual percentage correct. The questions were representatively drawn from 17 Complete Idiots guides advising on a wide variety of decisions. Applying Decision Rules assesses th

42、e ability to apply specified decision rules (e.g., elimination by aspects) to ten hypothetical choices, with each option characterized on several attributes in a table. Consistency in Risk Perception assesses the ability to make risk judgments that are internally consistent (e.g., giving a lower pro

43、bability to dying in a terrorist attack than to dying from any cause), with 20 paired judgments. Resistance to Sunk Costs uses ten single-choice sunk cost problems to assess the ability to ignore irrecoverable prior investments, and consider only future consequences when making decisions. The aggreg

44、ate Adult Decision-Making Competence (A-DMC) measure is the unweighted average of standardized task scores. It represents the extent to which individuals can make decisions normatively as a potential correlate of their self-reported attempts to maximize.Table 1: Relationships among decision-making s

45、tyles.Decision-making StyleMaximize Cope Rat.Int.Dep.Avoid.Spon.To cope behaviorally.20* To decide rationally.06* .58* To decide intuitively.05* .42* .44* To depend on others.29* .04* * .23*.24* To avoid decisions.37* .43* .11+ * .00*.40* To decide spontaneously.31* .26* .21*.19*.10+ * .44* To feel

46、regret.47* .31* .01*.01*.34*.41* .18*+ p.10; *p .05; * p .01; * p .001 two-sided.Decision Outcomes Inventory (DOI). The Decision Outcome Inventory elicits self-reports of having experienced each of 41 negative events, varying widely in domain and severity (e.g., threw out food or groceries you had b

47、ought, locked yourself out, got divorced, had an unplanned pregnancy). For 35 of these events, respondents only received credit for avoiding a negative outcome if they indicated having made a decision crucial to experiencing it (e.g., only those who reported having a drivers license received credit

48、for not having lost it). As a proxy for severity, each outcome was weighted by the proportion of participants who reported not experiencing it (among those who had the opportunity), assuming that more severe outcomes tend to be less frequent (as is the case with spending a night in jail, versus forg

49、etting a birthday). Severity was computed specifically using responses from this sample. Weighted outcomes were averaged and subtracted from 0, so that higher scores reflect better outcomes. Thus, the DOI score reflects the number of negative outcomes respondents had avoided, out of those they had t

50、he opportunity to experience, weighted by severity. The A-DMC and DOI measures are available from the authors or online at http:/sds.hss.cmu.edu/risk/ADMC.htm, as well as in this issue of Judgment and Decision Making (http:/journal.sjdm.org/vol2.6.htm).Demographics. Education is assessed from answer

51、s to, “What is currently your highest level of education?” with the options of no degree, high school, associate, bachelor, masters, and Ph.D. A dummy variable for being recruited from a social service organization targeting residents with lower levels of socio-economic status was used as a proxy fo

52、r socio-economic status (SES; 0 = lower, 1 = higher).Table 2: Relationships between decision-making styles and demographics. Gender Age SESEducationTo maximize *.11* .01* .16*.16*To cope behaviorally .01* .20* .24*.20*To decide rationally .04* .19* .16*.16*To decide intuitively .10* .20* .05*.07*To

53、depend on others .07* .11* .05*.05*To avoid decisions .09* .07* .20*.12*To decide spontaneously .01* .08* .25*.27*To feel regret .04* .08* .11*.09*+ p.10; *p .05; * p .01; * p .001 two-sidedNotes: Gender is coded as 0 if male and 1 if female. SES is 1 (higher) if respondent was interviewed at a soci

54、al service organization (0 otherwise). Education is coded as 1 if no degree, 2 if high school, 3 if associates degree, 4 if bachelors degree, 5 if masters degree, and 6 if Ph.D.2.3 ProcedureRespondents were run in group survey sessions held in their communities. On a cover letter, respondents were t

55、old that the study was about decision styles, and that they would “be given several decision problems, items from an intelligence test, as well as questions about decision styles, decision outcomes, and demographic information.” A-DMC tasks were self-paced, in the order: (a) positive versions of the

56、 Resistance to Framing items, (b) Recognizing Social Norms questions asking if “it is sometimes OK” to engage in different behaviors, (c) Under/Overconfidence, (d) Applying Decision Rules, (e) Consistency in Risk Perception, (f) Resistance to Sunk Costs, (g) negative versions of the Resistance to Fr

57、aming items, and (h) Recognizing Social Norms questions asking about their peers reported behaviors. This order maximized the distance between paired tasks (Resistance to Framing, Recognizing Social Norms). Subsequently, participants completed the decision-making styles measures, in the order: regre

58、t, maximizing, behavioral coping, rational, intuitive, dependent, avoidant and spontaneous decision making. Finally, they completed the DOI and the demographic questions. Participants received two envelopes containing $17.50 each, with the option to donate to the organization that recruited them. As

59、 stated in the recruitment materials: “We will give $35 for your time and effort. You will be given a choice between (a) taking home $35; (b) giving $35 to the organization through which you were recruited or (c) taking home $17.50 and giving the organization $17.50.”2.4 Analysis strategyFirst, we e

60、xamine Pearson correlations between self-reported maximizing and the other decision-making styles, as well as the demographic variables. Then, we report hierarchical multiple linear regressions predicting A-DMC, entering maximizing on the first step, the other the decision-making styles on the secon

61、d step, and demographics on the third. Finally, we report hierarchical regressions predicting DOI, entering maximizing and the other decision-making styles on the first two steps, demographics on the third, and A-DMC on the fourth.3 Results3.1 Scale propertiesAll eight decision-making style variable

62、s had ranges of 1 to 5, with higher scores indicating greater endorsement of the style. The mean self-reported maximizing score was 2.9 (Cronbach = .76). Other means were 3.0 for regret (Cronbach = .65), 3.8 for behavioral coping (Cronbach =.86), 3.8 for rational ( = .85), 3.6 for intuitive ( = .87)

63、, 3.4 for dependent ( = .83), 2.6 for avoidant ( = .89), and 2.6 for spontaneous ( = .87) decision making. As reported in Bruine de Bruin et al. (2007), A-DMC has an of .85 and test-retest reliability of .73. DOI has an of .88.3.2 Self-reported maximizing and other decision-making stylesTable 1 pres

64、ents correlations between self-reported maximizing and the other decision-making styles. Consistent with Schwartz et al.s account, self-reported maximizing is related to less behavioral coping, more depending on others, and a stronger tendency to avoid decisions. However, we did not find the expected corrections with self-reports of deciding rationally and deciding intuitively and found the opposite of the expected correlation with deciding spontan