Framing and Conflict

8/14/2019 Framing and Conflict

1/18

Journal of Experimental Psychology:Learn ing , Memory , and Cogni t ion1992, Vol. 18, No. 5, 1040-1057Copyright 1 992 by the Am erican Psychological Association Inc0278-7393/92/3.OO

Framing and Conflict: Aspiration Level Contingency,the Status Quo, and Current Theories of Risky ChoiceSandra L. Schneider

University of South FloridaThe effect of positive versus negative frames on risky choice was examined for a variety ofscenarios and risks. Preferences in the positive domain were strong and mainly risk averse, withnotable exceptions. Preferences in the negative domain, however, were marked by their incon-sistency, shown both by an overwhelming lack of significant majority preferences and a surpris-ingly strong tendency of individual subjects to vacillate in their negatively framed choices acrosspresentations. This finding is accounted for by a proposed aspiration level contingency in whichaspiration levels are systematically set to be more difficult to achieve in the face of a perceivedloss than a gain. The implication s of the results, and the aspiration level contingency, are exploredwith respect to current theories of risky choice, including Kahneman and Tversky's (1979)prospect theory and Lopes's (1987, 1990) security-potential/aspiration theory.

When making choices, people are sensitive to the way inwhich the options are presented. The framing effect, initiallydescribed by Tversky and Kahneman (1981), exemplifies thissensitivity. They found that people's preferences among op-tions were dependent on how those options were described.Different representations frequently elicited different prefer-ences even though the objective outcomes remained the samefrom one representation to the next.The most well-known example of the framing effect in-volves the "Asian disease problem," in which the outcomesof potential courses of action are described or framed posi-tively in terms of lives saved or negatively in terms of liveslost, whereas the actual potential outcomes of the problemremain objectively the same in both cases. Tversky andKahneman (1981) found that when the outcomes of thescenario were framed positively, a majority of their 152subjects (72%) had risk-averse preferences, selecting a certainalternative over a risky alternative of equal expected value. Incontrast, when the same outcomes were framed negatively, amajority of an independent group of 155 subjects (78%) hadrisk-seeking preferences, choosing the risky alternative overthe certain one.Tversky and Kahneman (1981, 1986; Kahneman & Tver-sky, 1984), as well as others, have dem onstrated several waysin which decisions may be influenced by the framing ofoptions. Framing effects have been demonstrated in the con-text of domina ted alternatives, one- versus two-stage gambles,

This research was supported in part by Grant RCS 32-89F fromthe University of South Florida Research Council and Division ofSponsored Research, Research and Creative Scholarship Grant Pro-gram. Many thanks to Marianne Martin, Diana Kares, and TonyJohnson for their assistance on the project. I would also like to thankJerome Busemeyer, Deborah Frisch, Lola Lopes, and an anonym ousreviewer for their insightful comments and suggestions.Correspondence concerning this article should be addressed toSandra L. Schneider, Department of Psychology, University of SouthFlorida, BEH 339, Tampa , Florida 33620-82 00. Electronic mail maybe sent to Sandra @ ultrix.csc.usf.edu.

"paying" versus "losing" effects, and effects due to changes inthe "mental accounts" being considered (see also Bierman,1989; Metzger & Krass, 1988; Slovic, Fischhoff, & Lichten-stein, 1982; Thaler, 1980, 1985; Thaler & Johnson, 1990).Perhaps the most interesting example of framing effects in-volves a valence manipulation in which the same outcomesare described in positive or in negative terms. The manipu-lation has much in common with the proverbial descriptionof the same glass of water as half full or half empty. Whatseems remarkable is the fact that this seemingly superficialchange in perspective can have a substantial impact on deci-sion making.

Prospect Theory and Framing EffectsKahneman and Tversky (1979, 1984; Tversky & Kahne-man, 1981, 1986) offer prospect theory as a potential expla-nation of framing effects. Prospect theory was designed as analternative to more traditional expected utility theories ofdecision making under risk. According to prospect theory,risky decision making involves (a) an editing phase in whichavailable options or prospects are coded to yield simplifiedmental representations, and (b) an evaluation phase in whichsubjective outcome values and probability weights are psy-chophysically determine d for each prospect and are integratedinto a single prospect value. The integration equ ation for eachprospect is the weighted sum of each of the prospect's subjec-

tive outcome values. The final prospect values are comparedso that the prospect with the highest value may be chosen.Of the editing operations, the most impo rtant for explainingframing effects involving valence is the coding operation. Thisoperation is responsible for describing each possible outcomerelative to a psychologically neutral reference poin t. O utcomesabove the reference point are perceived as gains, and outcomesbelow the reference p oint as losses.Within the evaluation phase, the value function plays acritical role in explaining differences in behavior as a functionof positive and negative frames. The psychophysical relation-ship between objective and subjective outcome values is char-

1040


2/18

FRAMING AND CONFLICT IN RISKY CHOICE 1041acterized as an S-shaped function.' In general, the shape ofthis value function (in conjunction with the integration equa-tion) implies that people will prefer a sure thing to a gambleof equal expected value (risk aversion) for gains but will prefera gamble to an equivalent sure thing (risk seeking) for losses.Hence, when outcomes are framed positively, they will becoded as gains and, as a result, risk-averse choices are pre-dicted. When outcomes are framed negatively, however, theywill be coded as losses, resulting in risk-seeking choices. Thevalue function is also steeper in the loss domain than in thegain domain. In consequence, "losses loom larger than gains,"or a loss is perceived to be more aversive than an equivalentgain is pleasant.The appeal of prospect theory as an explanation of framingeffects has been widespread. The S-shaped value function hasbeen implicated in a multitude of studies investigating theeffects of positive and negative frames, not only in basicresearch but also in applied contexts, including negotiation(e.g., Bazerman, Magliozzi, & Neale, 1985; Neale, Huber, &Northcraft, 1987), industry (e.g., Quails & Puto, 1989), na-tional security (e.g., Kramer, 1989), social dilemmas (e.g.,Brewer & Kramer, 1986; Fleishman, 1988), and health care(e.g., McNeil, Pauker, Sox, & Tversky, 1982; Meyerowitz &Chaiken, 1987). Moreover, practitioners in many areas arebeing warned of the potential impact of framing on their ownor their clients' judgm ents (e.g., Bazerma n, 1983; Burton &Babin, 1989; Travis, Phillippi, & Tonn, 1989).

Risk Preferences and FramingSeveral studies of framing effects amply demonstrate thatchanges in the representation of information can have potenteffects on decision behavior. However, they are not as apt to

demonstrate that these effects are due exclusively to changesfrom risk-averse preferences in the positive frame to risk-seeking preferences in the negative frame. In fact, in m any ofthe applied studies, it is difficult to determine exactly howrisk andrisk-takingpropensity should be defined or measured.Even in those studies in which the characterization of risk isrelatively straightforward, results tend to differ substantiallyfrom one investigation to the next.Framing studies in which the characterization of risk isclear-cut tend to use scenarios that are similar to, or havebeen modeled after, Tversky and Kahneman's (1981) Asiandisease problem. Despite the structural similarity of the tasksacross studies, however, results range from no differencebetween positive and negative frame choices to sizable differ-ences in choice that vacillate not only in response to changesin frame but also to changes in other variables.Fagley and Kruger (1986), for example, found no framingeffects when they asked a large group of school psychologiststo respond to a hypothetical scenario involving programs toprevent school dropo uts. For half of the subjects, the programoutcomes were framed in terms of the number of studentsstaying in school, and for the other half in terms of the nu mb erof students dropping out of school. In both frame conditions,the majority of subjects (70% in the positive frame and 67%in the negative frame) made risk-averse choices, preferringthe certain option to the risky option. The lack of a framing

effect is attributable to the fact that, contrary to predictions,subjects did not predominantly prefer the riskier option inthe negative frame condition.In subsequent work using a problem involving lives threat-ened by cancer, Fagley and Miller (1987) again found that amajority of subjects preferred the certain option in the positiveframe in each of two studies. However, preferences for thecertain option were also relatively strong in the negativeframe, with a majority of subjects (master's of business ad-ministration students) in thefirststudy and half of the subjects(college of education graduate students) in the second studypreferring the certain option over the risk. Hence, positiveframe results were generally consistent with predictions, butnegative frame results were not supportive of predictions.Responses to risky choices in the negative frame were notpredominantly risk seeking in either study.Sometimes, framing effects will appear and disappear withvarious changes in scenarios. Levin (1990) found relativelystrong framing effects for a problem involving lives threatenedby acquired immune deficiency syndrome (AIDS) if the vic-tims were hemophiliacs. The results were exactly those ex-pected. Eighty percent of the subjects made the risk-ave rsechoice in the positive frame, and 80% of another group ofsubjects made the risk-seeking choice in the negative frame.When the scenario was again presented but this time with thevictims changed to homosexuals or drug users, no framingeffect was found. Only half of the positive frame subjectschose the sure thing, and just less than half (40%) of thenegative frame subjects chose the risk. In this version, aconsensus in preference was not clearly evident for eitherframe. Levin concluded that factors such as the perceiveddesirability of particular social groups can m ediate the im pactof framing.

In another set of studies, Wagenaar, Keren, and Lichten-stein (1988) demonstrated that differences in the surfacestructure of a problem can have a profound impact on riskpreferences. Although in the three studies they conductedonly negatively framed scenarios were used, the degree of riskseeking changed drastically as various aspects of the scenariochanged. The basic scenario, involving islanders exposed to alife-threatening disease, was adopted from a study by Ham-merton, Jones-Lee, and Abbott (1982) in which it was re-ported tha t only 17% of subjects m ade risk-seek ing choiceseven though outco mes were framed in terms of probability ofdying. In their first study, Wagenaar et al., found that ex-panding the story, providing more explicit outcome probabil-ity information, and clarifying the decision maker's role (is-lander o r med ical officer) led to differences in risk preferences,

1 Prospect theory does not necessitate that all individuals in allsituations have an S-shaped value function; functions of this shapeare merely hypothesized to be among the most common. Neverthe-less, it is the S-shape of the value function that is critical for theprediction of framing effects characterized by risk-averse preferencesin the positive frame and risk-seeking preferences in the negativeframe. Because Tversky and Kahneman (e.g., 1981, 1986) themselvesregularly rely on the presence of an S-shaped value function to predictand explain framing effects, it seems reasonable to make the samegeneral assumption in this context.


3/18

1042 SANDRA L. SCHNEIDERwith the percentage of subjects choosing the risky optionranging from 39% to 63%.In the second study, they found that preferences wereaffected by whether the uncertain outcomes were associatedwith action or inaction on the part of the decision maker. Ingeneral, risky outcomes tended to be m ore popular when theyrequired taking action; however, the effect was mediated bothby the decision maker's role and by the subject population(Dutch vs. American students). In all cases, agreement amongsubjects was weak. Between 42% and 68% of subjects chosethe risky option across each of the conditions, indicating thatsubjects were relatively evenly split between risk-averse andrisk-seeking preferences.In the third and final study, Wagenaar et al. (1988) at-tempted to generalize their findings from the islanders prob-lem to a problem involving children held hostage in a kinder-garten. The change in context resulted in a dramatic changein preferences. Regardless of decision-maker role or the asso-ciation of risky outcom es with action versus inaction, the vastmajority of subjects (85% to 92%) preferred the risky optionover the certain death of a child. As the authors point out:It is abundan tly clear that cha nge of cover story affected the deepstructure of the problem as it was analyzed by subjects. Even ifwe had proposed a theory that would account for all resultsobtained with the islander cover story, it is not clear how such atheory would predict these dramatically different results for thehostages problem, (p. 186)They conclude that the prospect theory explanation of fram-ing cannot account for these results because all proposedediting and evaluation operations are carried out on proba-bilities and outcomes, essentially independent of scenariocontext.Besides the inconsistency in the presence and strength offraming effects across studies, an additional concern is thegenerality of the framing effect w ith respect to the individ ual.In the original presentation of the Asian disease problem, forexample, different groups of subjects responded to the positiveand negative frame versions. Although majority preferenceswere opposite across frames, it is not clear to what ex tent eachindividual may have succumbed to a framing effect. More-over, within each frame, we can be assured that 28% of thepositive frame subjects and 22% of the negative frame subjectswere behaving in a fashion inconsistent with the predictionof risk aversion for gains and risk seeking for losses.In one portion of a larger study, Frisch (in press) askedeach of her subjects to respond to both the positively framedand negatively framed versions of the Asian disease problem.Although 59% of the 136 subjects made risk-averse choicesin the positive frame and 62% made risk-seeking choices inthe negative frame, the percentage of subjects who showedthe expected pattern of preferences across both domains wasonly 29%. Hence, almost three fourths of the subjects showedan overall preference pattern that is inconsistent with theframing prediction based on prospect theory. Although thelack of framing effects at the level of the individual may bedue in part to carryover effects, these were minimized byinterspersing other problems between frame pairs. Regardlessof potential carryover effects, these results suggest that groupdata m ay not acc urately reflect the susceptibility of individuals

to the effects of frame and that framing effects for individualsmay not be as common as is generally believed.Risk Aversion and Risk Seeking

Although studies of the framing effect routinely invok e theprospect theory prediction of risk-averse preferences in thepositive frame and risk-seeking preferences in the negativeframe, there is a noticeable absence of systematic assessmentsof the degree to which this prediction actually captures deci-sion-making behavior. How ever, in studies in which m onetarygambles replace the decision scenario and actual gains andlosses replace positive and negative frames, there is evidencethat preferences often do not conform to the prediction ofrisk aversion for gains and risk seeking for losses. Althoughthere are cases in which the predicted pattern is observed,there are as many cases, if not more, in which the pattern isnot observed.For example, Hershey and Schoemaker (1980) examinedpreferences between two-outcome gambles and their certaintyequivalents. They found that majority preferences could notbe accurately predicted for either domain and that preferencereversals from the gain to th e loss domain were the exceptionrather than the rule. Although the option pairs varied widelyin both amounts and probabilities, there were many cases inwhich no significant majority preferences were observed, es-pecially amo ng o ption pairs involving losses. Of the 25 op tionpairs presented in the first two experim ents, there were signif-icant majority preferences for only 15 of the 25 gain pairs. Ofthose 15, 12 represented a pre dom inance of risk-ave rsechoices and the other 3 represented a predominance of risk-seeking choices. For losses, only 10 of the 25 option pairsshowed significant majority preferences, with 9 representinga risk-seeking majority and 1 a risk-averse majority. Overall,evidence of predominantly risk-averse preferences for gainsand risk-seeking preferences for losses was weak and incon-sistent.

In a study of multi-outcome lottery preferences for bothgains and losses, Schneider and Lopes (1986) found thatpreferences varied across subjects and across lottery types. Fo rthose subjects who had been preselected for their risk-aversebehavior in a task involving two-outcome gambles represent-ing gains, multi-outcome lottery preferences were generallyrisk-ave rse for gains but were not consistently risk seeking forlosses. In contrast, for those subjects preselected for their risk-seeking behavior in the same two-outcome task, multi-out-come lottery preferences were not consistent across the do-main of gains but were generally risk-seeking for losses. Forboth groups of subjects, the combined prediction of risk-averse preferences for gains accom panied byrisk-seekingpref-erences for losses was only supported for a small minority ofpairs.

Understanding Framing Effects for RisksWhat all of this evidence seems to suggest is that riskpreferences frequently do no t exhibit the pattern of risk-ave rsepreferences in the positive domain coupled with risk-seekingpreferences in the negative dom ain. Although risk-aversepref-


4/18

FRAMING AND CONFLICT IN RISKY CHOICE 1043Table 1Decision Scenario Key

Code/issue Source DescriptionAD : Asian diseaseCA: CancerPO : PoisoningPE: PesticidePR: PreservationDO : DropoutsPC: Plant closingWO: WorkersEQ: Equipment

Adapted from Tversky&Kahneman, 1981Adapted from Fagley &Miller, 1987Adapted from Fischhoff,1983Original

OriginalAdapted from Fagley &Kruger, 1986Adapted from Bazer-man, 1982OriginalOriginal

Threat to 600 human livesfrom diseaseThreat to 1,200 humanlives from cancerThreat to 3,000 humanlives from poisoningThreat to 1,200 endangeredanimal livesThreat to last 1,200 lives ofone animal speciesThreat of 900 studentsdropping out of highschoolThreat to 600 jobs fromplant closingThreat to 600 potentialjobs in a new factoryThreat to recovery of$30,000 from forced sale

erences are relatively common for positive options, risk-seeking preferences do not seem to be reliably associated w ithnegative options. Given that the framing effect described byTversky and Kahneman (e.g., 1981, 1986) has been explainedon the basis of this relationship in risk preferences acrossdomains, it becomes important to establish the prevalence ofthis particular pattern of preferences. Moreover, for thoseinstances in which the predicted pattern is not found butframing effects are still evident, it becomes necessary to searchfor other contributing factors that may explain the observedchange in preferences.The purpo se of this investigation is to exam ine the strength,generality, and form of framing effects across a variety ofchoices involving risk. To obtain a broad sampling of prefer-ences, several scenarios are presented and options are widelyvaried both in terms of probability of occurrence and size ofpotential outcomes. The scenarios tap several topics includingthreats to human lives from disease or disaster, threats toendangered animal lives, threats to jobs, threats to students,and threats to financial investments. The options vary notonly in the extremeness of probabilities and outcomes butalso in the presence or absence of riskless components (i.e.,risks in which all compon ent outcomes are nonzero).The study is designed to compare the size and direction offraming effects across different scenarios and options, focusing

on the extent to which positive frame preferences are riskaverse and negative frame preferences are risk seeking. Togain a more complete understanding of the prevalence ofparticular preference patterns, comparisons are conductedboth at th e level of group m ajority preferences and at the levelof individuals' preferences. Strength of preference w ithin eachframe is examined in terms of the degree of consensus acrosssubjects as a group and in terms of the level of consistency ofeach individual's choices over time. Throughout, preferencepatterns are identified and assessed in an effort to isolate thosefactors that may contribute to choice, either in concert or inopposition, as one m oves from a positive to a negative frame.

MethodStimuli

Decision scenarios. The stimuli included nine different decisionscenarios.2 The scenario types are listed in Table 1 and includevariations of several previously published problem s as well as severaloriginal problems. The topics of the scenarios included threats tohuman lives from new and existing diseases and from poisoning,threats to endangered animal lives, threats of students dropping outof high school, threats to new and existing jobs, and threats ofmonetary losses. Regardless of topic (or source), each scenario wasstructured using a systematic approach that ensured that each scenariooutlined (a) the type, cause, and extent of threat; (b) the course ofaction used to generate potential solutions; (c) the information thatit is only possible to implem ent a single option; and (d) the provisionof an explicit role in choosing an option along with a clearly impliedgoal.Each scenario was presented to subjects both visually and orally.Scenarios were visible on an overhead projector screen while theexperimenter read the scenario aloud. The scenario remained visiblethroughout all relevant trials.Options. Four basic options were created and edited for eachframe and each scenario. All options had an expected value of onehalf of the total number threatened. The general pre-edited forms ofthese options are presented in Table 2 for the positive (and negative)frames. Although fractions are listed in Table 2, the actual optionslisted the specific number of lives, jobs, and so on, threatened. Withthe exception of the sure thing (ST), each option is coded accordingto the probability of the best outcome followed by the probability ofthe worst outcome regardless of frame.

2 In actuality, the stimulus set contained a tenth scenario involvinga threat to the recovery of $600 from the purchase of an unused planeticket. Because of the belated discovery of a typographical errorpresent in one of the negative options associated with this scenario,all data regarding this stimulus were later omitted from the analysis.For ease of exposition, reference to the plane ticket scenario has beendeleted throughout.


5/18


6/18

FRAMING AND CONFLICT IN RISKY CHOICE 1045ST vs 75/25100

80O UJui M 6 0

u. *O COI- E4 02 0

0 A D C A F O P E P R r j D F C W O B Q

ST vs 50/50

O POSITIVE FRAME NEGATIVE FRAME

A D C A F O P E P R D O F C W O B Q

1 0 08 06 04 02 0

0

ST VS 25/75

O POSITIVE FRAME NEGATIVE FRAME

A D C A F O P E P R D O P C W O B Q

25/75 vs 75/25is


7/18

1046 SANDRA L. SCHNEIDERof pair and of frame. Although more subtle, there also appearto be systematic differences in preferences across scenarios.Focusing on the more traditional ST-risk pairs in the toprow, one cannot help but notice the striking differences inpreference as the risk changes. Only when the ST is pittedagainst the 75 /25 risk are there large framing effects across allscenarios. When the ST is paired with the 50/50 risk, sizableframing effects are only evident for pairs involving human oranimal lives. When the ST is paired with the 25/75 risk, thereis virtually no evidence of a framing effect and majoritypreferences in both domains tend to be risk seeking.There seems to be a rough parallel in preferences for therisk-risk pairs presented in the bottom row. The results forthe 25/75 versus 75/25 pair are similar to, but less pronouncedthan, the results for ST versus 75/25. The same kind ofsimilarity is apparent between the top and bottom centralgraphs involving the 50/50 option, although scenarios involv-ing lives are less differentiated. In the 75/25 versus 50/50graph, there are few discernible differences in preference fromthe positive to the negative frame; a small majority in bothframes prefers the 75/25 option.Relationship of frame to ch oice. Table 3 statistically con-firms the observed impressions of choice relationships acrossframe. For each option and scenario, Table 3 presents a Fisherexact test of association and a lam bda coefficient. A significantFisher exact test indicates that the proportion of subjectsmaking the risk-averse choice in the positive frame conditiondiffers reliably from the prop ortion making the risk-ave rsechoice in the negative frame condition (although not neces-sarily that preferences switch from risk aversion to risk seek-ing). The lambda coefficient indicates the proportional reduc-tion in the probability of error in predicting choice due to thea priori specification of frame. Lambda coefficients varybetween .00 (indicated by a dash in Table 3) and 1.00; thelarger the coefficient, the greater the increase in the ability topredict choice, given information about whether the scenariois framed positively or negatively.

For the ST versus 75/25 option pair, there is a significantrelationship between choice and frame across all of the sce-narios. Except for the PR scenario, this relationship substan-

tially increases the ability to predict choices, given frame. Forthe ST versus 50/50 option pair, relationships are only evidentin about half of the scenarios, including those involvinghuman lives and finances. Here again, predictability is en-hanced by frame information. For the pairs involving animalextinction, students, and jobs, however, there is no discernibleinterdependence between choices and frame. For the 25/75versus 75/25 pair, just over half of the Fisher exact tests reachsignificance; however, there is little, if any, increase in theability to predict choice when frame information is available(i.e., regardless of the frame, a prediction of risk aversion isas good as or better tha n a prediction of risk seeking). Acrossthe remaining option pairs, there is little evidence of system-atic relationships between choices and frame. Overall, thetable highlights the substantial differences in the presence offraming effects across pairs and scenarios. Framing effectsoccur consistently across all scenarios for only a single o ptionpair and occur in a majority of scenarios for only two of theremaining five pairs.

It may seem tempting to conclude that framing effectsgenerally decline as the difference in riskiness between thetwo available options decreases. However, this hypothesis isonly partially supported. Recall that the options increase inriskiness from ST to 25/75, 75/25 , and then 5 0/50. Althoughthe data for the ST versus 25/75 option and the 75/25 versus50/50 option appear to confirm that framing effects are leastlikely to occur for pairs adjacent in riskiness, he 25/75 versus75/25 data seem to disconfirm this possibility. In addition,given that 50/50 is the riskiest of the options, one wouldexpect framing effects to be larger and more consistent forthe pairs involving 50/50 than for comparable pairs involving75/25. However the data are clearly in contradiction to thishypothesis. Hence, differences in riskiness alone cannot ade-quately account for the differences in framing effects acrosspairs.

It is even more difficult to summarize the differences inframing effects across scenarios. Nevertheless, it does seemclear that the PR scenario was least likely to yield framingeffects. This scenario involves a threat to the last 1,200 ani-mals of a particular species. Regardless of frame, subjects wereTable 3Fisher Exact Tests and Lambda Coefficients for C hoices, Given Frame

Test/Option pairFisherST vs. 75/25

ST vs. 50/50ST vs. 25/7525/75 vs. 75/2525/75 vs. 50/5075/25 vs. 50/50LambdaST vs. 75/25ST vs. 50/50ST vs. 25/7525/75 vs. 75/2525/75 vs. 50/5075/25 vs. 50/50

A D.000***.007**1.000.002**.176.117.500.381

CA.000***.005**.767.010**.157.554.556.333.133.091

PO.038*.005**.755.107.241.107.286.222

PE.001***.031*1.000.008**.301.760.421

ScenarioPR

.002**1.000.769.014*1.000.745

D O.000***.083.239.006**.008**.191.400.100

PC.001***.049*1.000.015*.041*1.000.267

W O.012*.157.038*.216.191.352.235.286

EQ.002**.010**.377.140.010**.745.353.133.095.191.133

Note. ST = sure thing; AD = Asian disease; CA = cancer; PO =closing; WO = workers; EQ = equipme nt. Dashes indicate A = 0.* p < .05. **p< .01. ***/>


8/18

FRAMING AND CONFLICT IN RISKY CHOICE 1047predominantly risk averse in their choices. It is easy to spec-ulate that this particularly strong pattern of preferences resultsfrom the fact that total extinction is a particularly aversiveoutcome that is seen as unacceptable regardless of frame.Hence, avoiding the total depletion of a valued and nonre-newable resource may act as a highly salient goal that isrelatively unaffected by positive and negative descriptiveframes.Elsewhere, especially in the negative frame, the threat ofloss of human or animal lives seems to be responded todifferently than threats involving jobs or school drop outs an d,to a lesser extent, financial losses. It may be that framingeffects are more likely to occur for more substantial threats(such as lives vs. school dropouts). Unfortunately, the differ-ences observed here seem less than systematic and do noteasily lend themselves to the prediction of scenario-relevantinfluences on framing effects.In fact, the analysis of subject inconsistency levels acrosschoices, which is presented in the upcoming final stage of theanalysis, cautions against a premature attempt to preciselyspecify the influence of the nine scenarios on preferences.Some of the app arent differences in preference across scenar-ios seem more likely to reflect inconsistencies in choice thanreliable changes in preference with changes in scenario.Hence, a realistic assessment of the systematic influence ofscenario on choice must await not only a more exactingattempt to control and isolate the relevant components ofscenarios but also a serious attem pt to differentiate weak andvacillating preferences from strongly held preferences.

Risk aversion and risk seeking. To what extent are grouprisk preferences consistent with the prediction of risk aversionin the positive frame and risk seeking in the negative frame?Looking back to Figure 1, an examination of the shadedregions within each graph reveals that risk aversion in thepositive domain is generally quite strong. For four of the pairs,a significant majority have risk-averse preferences for at leasteight of the nine scenarios. For 75/25 versus 50/50, a signifi-cant majority have risk-averse preferences for only three ofthe nine scenarios, and for ST versus 25/75, majority prefer-ences actually tend to be risk seeking, although a significantrisk-seek ing majority was only dem onstrated for a singlescenario.The latter tendency toward risk seeking is not entirely asurprise, given that the positive option 25/7 5 h as a risklesscomponent (i.e., its worst outcome is better than zero). Otherstudies (e.g., Kahneman & Tversky, 1979; Schneider & Lopes,1986) have also documented majority preferences for thistype of risk over a sure thing. Nevertheless, with this onenotable exception, majority preferences in the positive do-main do tend to be risk averse as predicted.Preferences within the negative domain, on the other hand,oscillate in a less than systematic fashion between risk-averseand risk-seeking majorities, and only rarely do majority pref-erences represent a significant consensus. In fact, of the 6cases in 54 in which significant majority preferences can beidentified, 5 involve a predominance of risk-averse choices.The only significant majority preference for the riskier optionin a pair occurs for the ST versus 75 /25 choice w ithin the CAscenario. Hence, clear majority preferences in the negativedom ain are virtually never risk seeking, nor are they generally

risk averse. It seems, instead, that strong majority preferencesin the negative domain frequently do not exist.It is true that where the framing effect does occur in thepresent data, it virtually always involves a decrease in riskaversion as one m oves from the positive to the negative frame.Yet, rarely, if ever, does the shift involve a strong tendencytoward risk seeking in the negative domain. Moreover, thevariability in framing effects across pairs and scenarios seemsto have less to do with positive frame preferences (which, inat least four of the six pairs, can be counted on to be stronglyrisk averse) and more to do with vacillations in majoritynegative frame preferences.Changes in Preference with Changes in Frame

Given that the previous analysis included only the datafrom subjects' first exposure to a single frame condition , onlybetween-subjects framing effects could be exam ined. Th e nextanalysis considers all of the da ta, aggregated over pairs, in anattempt to identify how changes in frame from one session tothe next, along with changes in other variables, influence thepreferences of the same individuals. The data for this analysisare the total numb er of times each option was chosen in eachscenario and each frame, with a maximum of six (3 pairs x 2presentations) and a minimum of zero. An analysis of vari-ance (ANOVA) was conducted with the between-subjectsvariable of group (pos-neg vs. neg-pos) and the repeatedmeasures of frame, option, and scenario.The main effect for option was significant, F(3, 129) =21.11, MS, = 35.35, p < .001. Overall, the most popularoption was 25/75, followed by ST, 75/25, and 50/50, respec-tively. However, these preferences were significantly affectedby the frame, the scenario, and the group. These effects arecomplex, appearing in first-order and higher order interac-tions. Figure 2 presents a com prehensive view of these results.The graphs on the left illustrate the results for the pos-neggroup, with the positive frame results from their first day oftesting at the top of the figure and the negative frame resultsfrom their second session at the bottom. The graphs on theright show the results for the neg -pos group, with the negativeresults for Day 1 at the top and the positive results for Day 2at the bottom of the figure.

There is a significant Frame x Option interaction, ,F(3,129) = 12.85, MS, = 13.65, p < .001. As can be seen in Figure2, the preference pattern for the positive frame (top left andbottom right) and negative frame (bottom left and top right)are ordinally similar, yet the pattern for both groups is muchmore distinct in the positive frame than in the negative frame.These results corroborate the findings from the previousanalysis, in which majority preferences could be clearly iden-tified in the positive domain but not in the negative domain.The Scenario x Option interaction was also significant,F(24, 1032) = 3.55, MSe = 1.19, p < .001. Rather thanindicating substantial differences in the preference orderingof the options, however, the interaction primarily shows thatthe strength of the overall preference pattern changed fromone scenario to the next, that is, the preference orderingremained virtually identical across the nine scenarios, but themagnitude of the difference in popularity of one optionrelative to another varied across the scenarios. For all nine


9/18

1048POS-NEG GROUP, DAY 1

SANDRA L. SCHNEIDERNEG-POS GROUP, DA Y 1

im


10/18

FRAMING AND CONFLICT IN RISKY CHOICE 1049described positively, but those same subjects have weak andvacillating preferences when the options are framed nega-tively.The data for the neg -pos group show a similar pattern. OnDay 1, when optio ns were framed negatively, preferences wereweak and variable across scenarios. However, 1 week later,when the options were presented in a positive frame, thesesubjects' preferences became much clearer and more consist-ent across scenario. Although the differential effect of positiveand negative frames can be clearly identified in both groups,either minor preexisting group differences or relatively smallbut noticeable carryover effects also exerted an influence onchoice.Framing Effects an d Conflict

The final analysis focuses on the characteristics of therelationship between positive and negative frame preferencesfor each of the subjects individually. The analysis serves toidentify not only clear-cut changes in preference from oneframe to the other but also changes in the levels of choiceagreement and consistency across frames. Both disagreementacross different subjects and inconsistency on the part ofindividual subjects are examined as evidence of potentialconflict in choice.Recall that for each scenario and pair within a given frame,subjects were asked to report on their preferences twice. Ineach case, subjects could choose the less risky option on bothpresentations, they could choose the riskier option on bothpresentations, or they could be inconsistent in their choices,selecting each option once. Hence, for both the positive andnegative frame, it is possible to ascertain whether subjectswere consistently risk averse or risk seeking in their prefer-ences for each pair or whether they were inconsistent in theirchoices. Moreover, it is then possible to compare preferencesin one frame with preferences in the other frame. Just such acomparison is presented in Table 4.

Preference patterns across frames. Table 4 shows the per-centage of subjects in the pos-neg group (left column) andneg-pos group (right column) with each of the nine (3 x 3)possible positive frame versus negative frame preference pat-terns, presented separately for each option pair. Because theproportions were generally similar across scenarios, each cellvalue actually represents the mean percentage of subjectshaving that preference pattern across the nine scenarios. Al-though not shown in Table 4, the standard deviations for eachcell tend to be small, ranging from 0% to 12.1 % for the pos-neg group and from 2.5% to 11.9% for the neg-pos group,with an average cell standard deviation of approximately 5%for both groups.The cell at the upper left of each 3 x 3 table indicates thepercentage of subjects who had risk-averse preferences in thepositive frame condition coupled with risk-seeking prefer-ences in the negative frame condition for each pair. If an S-shaped value function were prevalent among subjects, thepercentage within this cell would be expected to be quite high.Notice, however, that less than 20% of the subjects fall intothis category regardless of the group or pair being considered.Moreover, in no case is this the cell with the largest percentage

Table 4Average Percentage ofPos-Neg and Neg-Pos Subjects WithEach Positive Frame Versus Negative Frame PreferencePattern for Each Option Pair

frameRSInRATotalRSInRATotalRSInRATotalRSInRATotalRSInRATotalRSInRATotal

RA

18224282

19114979

641323

15244382

16204682

14123359

PositivePos-neg subjects

In

55616

43613

125926

43411

257

14

57921

RS Total %ST vs. 75/252

002

252748100

ST vs. 50/505218

281656100

ST vs. 25/752614105025/75321625/75122575/25

106420

44233299

frameNeg-pos subjects

RA

17142455

11183362

6121230

vs. 75/2522294899

11203566

vs. 50/50192755

101

13164069

vs. 50/50292546

10 0

7133353

In971

17

38314

78823

35513

456

15

471223

R S

224228

1465252911848

163221

574

16

107623

Total %482527

100

283241101

423128101

302842100

222850

100

21275199

Note. ST = sure thing; RA = consistently risk averse; In = incon-sistent; RS = consistently risk seeking.of subjects. For five of the six pairs, the most frequent pref-erence pattern involves risk-averse preferences in both thepositive and negative frame conditions. For the ST versus 25/75 pair, the most com mon preference pattern involves risk-seeking preferences in both domains. Hence, it appears thatsubjects are more likely to maintain their preferences fromone frame to the other than they are to change their prefer-ences in accord with the proposed prospect theory valuefunction.The totals for each 3 x 3 table provide information aboutthe predominance of preferences within each frame. In thepositive frame, a majority of subjects show clear risk-aversepreferences in five of the six pairs. T his tendency is especially


11/18

1050 SANDRA L. SCHNEIDERstrong for the pos-neg group (as seen previously in Figure 1).In both groups, however, roughly half of the subjects areclearly risk seeking in their preference of 25/75 over the ST.Overall, subjects show high levels of agreement in their posi-tive frame preferences.As seen earlier, preference patterns in the negative domainare not as strong as their positive frame counterparts. Forfour of the six pairs, approximately half of the subjects inboth groups displayrisk-aversepreferences, with the rest beingabout equally divided between risk-seeking and inconsistentpreferences. The same pattern holds for ST versus 75/25 forthe pos-neg group, however, the pattern is reversed for theneg-pos group such that half of the subjects are risk seeking.As in the positive frame, preferences for the ST versus 25/75pair are most apt to be risk seeking for both groups. Overall,neither group displayed a strong m ajority of risk-seek ingpreferences for any of the negative frame pairs.

Framing and inconsistency. Table 5 summ arizes the pref-erence patterns and types of inconsistency found for eachgroup and each option pair. The most striking finding is thelack of clear-cut framing effects. Regardless of the group orthe option pair, less than 20% of the subjects showed a reliablechange in preference with a change in frame (as shown in theSwitch column). For most pairs across both groups, subjectswere most likely to maintain the same preferences from oneframe to the other. Nevertheless, there remains a large pro-portion of subjects for each pair whose preferences wereinconsistent in one or both do mains.The three columns on the right in Table 5 represent abreakdown of the types of inconsistency displayed by subjects.

Notice tha t in alm ost all cases, the percentage of subjects whoare inconsistent in the negative domain far exceeds the num-ber who are inconsistent in the positive frame or in bothframes. Indeed, inconsistency in the negative domain is gen-erally twice as likely as inconsistency in the positive domain.In fact, subjects frequently have a greater likelihood of show-ing inconsistent preferences in the negative domain than risk-seeking preferences.The generality of the preference patterns and inconsistencylevels across option pairs and groups is nothing short ofremarkable. To document the statistical significance of muchof the information in Table 5, an ANOVA was conducted onthe relevant consistency data. For each frame and pair, thedata consisted of the nu mb er of times across the nine scenariosthat each subject was consistent from the first to the secondpresentation.As expected, the main effect for frame was highly signifi-cant, F(\, 43) = 26.62, MSe = 3.46, p < .001 . Subjects weresubstantially more consistent in their choices in the positiveframe than in the negative frame. The subject's group, how-ever, was not related to th e observed consistency levels, eitheroverall or in com bination with frame (both Fs < 1, with MSss= 11.72 and 3.46, respectively). This suggests that previousexposure to options framed from an opposing perspective hasno effect on the level of consistency observed. Regardless,subjects are less consistent in negative frame choices.Although there were no overall differences in consistencyacross pairs, there was a significant Frame x Pair interaction,F(5, 215) = 2.28, MS, = 2.16, p < .05. A simple-effectsanalysis of pair at each of the two frames revealed that

Table 5Summary of Positive Versus Negative Frame Preference Patterns by Group an dOption Pair

Option pairST vs. 75/25ST vs. 50/50ST vs. 25/7525/75 vs. 75/2525/75 vs. 50/5075/25 vs. 50/50

MeanST vs. 75/25ST vs. 50/50ST vs. 25/7525/75 vs. 75/2525/75 vs. 50/5075/25 vs. 50/50

Mean

Preference iSwitch

18201616181818

19161413171315

Same44543946474346

46474151454346

jatternIncons.

Pos-neg group38264437363937

Neg-pos group35384636384339

Total %10010099991011001011001009999101100101

Type of inconsistencyPos111021891412

106158101611

Neg2213182622182018242323232022

Both53535757885577

Note. Preference patterns show the mean percentage of subjects whose preferences switched from oneframe to the other (Switch), were the same in both frames (Same), or were inconsistent in at least oneframe (Incons.). Inconsistency values indicate the mean percentage of subjects who were inconsistentin only the positive frame (Pos), only the negative frame (Neg), or in both frames (Both). ST = surething.


12/18


13/18

1052 SANDRA L. SCHNEIDERThe lack of consistency in choice in the negative framerelative to the positive frame is troublesome from this per-spective. Inconsistent preferences are only to be anticipatedwhen the perceived values of the available options are equiv-alent, implying that it would not matte r to the perceiver whichoption is selected. Using an example from this study, thiswould suggest that subjects feel a strong commitment to aparticular option when they are trying to save people's livesbut that they do not care which option is selected when theyare trying to keep those same people from dying. Surely, theydo care which option is selected, but they experience greaterdifficulty in determining which of the two options is prefera-ble. A psychological theory that cannot distinguish the expe-rience of indifference from that of conflict seems equallyunable to ad equately cap ture the processing of difficult trade-offs inherent in risky choice.The inability to explain (or even acknowledge) conflict inchoice is not the only problem raised by the results of thisexperiment. The very notion of risk attitudes seems suspect.Again, the evidence presented here clearly suggests that no

strong majority preferenceslet alone strong risk-seek ingpreferencesexist in the negative dom ain. Of the 54 negativeframe choices ma de, only one indicated a significant majoritytendency toward risk-seeking preferences (with five choicesrevealing a significant co nsensus toward risk aversion). Mo re-over, the examination of individual preference patterns dem-onstrated that only about 30% of the subjects could be ex-pected to have risk-seeking preferences for any given pair ofnegative frame options.On the o ther han d, m ajority preferences for a full two thirdsof the choices in the positive domain favor the hypothesis ofrisk aversion. For five of the six positive frame pairs, anaverage of 70% of the subjects maintained risk-averse prefer-ences. Although this appears to provide some support for therole of risk attitudes in the case of gains, it is less compellingwhen one recognizes that there is also evidence that riskseeking may predominate for some choices involving gains.These include, but may not be limited to, choices for whichthe risky option has a riskless component. This occurs whenboth potential outcomes of the risky option are positive,eliminating any chance of obtaining zero. It may be temptingto dismiss this preference for the riskier option as uninterest-ing, given that the option entails some sure positive outcome.However, to do so is to dismiss the fact that in some circum-stances involving potential gains, many people would rathernot be certain of the outcome ahead of time but wouldactually prefer an option for which the outcomes will bedetermined probabilistically.Overall, then, in neither do main is there evidence for strongand stable risk attitudes. It should not be surprising to find,in addition, a general lack of any systematic relationshipbetween risk preferences from the domain of gains to thedomain of losses. Risky decision making in the negativedomain is neither the antithesis nor the equal of decision

these factors would affect c onsistency except inasm uch as they mightinfluence the closeness of prospects' perceived overall values and, asa consequence, the possibility of indifference.

making in the positive domain; they are independent inoutcome and perhaps in process. This lack of relationship isnot only evident throughout this study but has also beennoted in several other studies that have documented thegeneral independence of choices in the gain and loss domains(e.g., Cohen, Jaffray, & Said, 1987; Hershey & Schoemaker,1980; Schneider & Lop es, 1986).The evidence suggests that risky choice is subject to a riche rsource of variation than just the psychophysical principlesthat relate objective quantities to subjective experience. Ac-cepting the empirical fact that commodities such as moneyand other tangibles have decreasing marginal utility may bewell-advised. Nevertheless, knowing the relationsh ip betweenthe subjective values of given potential outcomes may not besufficient for predicting or understanding how the introduc-tion of various amounts of risk will mediate choices aboutthose entities.Direct support for this position has been presented by K eller(1985) who demonstrated, first, that subjective value (i.e.,utility) functions for riskless judgm ents were different fromthe subjective value functions for comparable risky judg-men ts. Second, she showed that the risk attitudes of individualsubjects varied qualitatively as the attribute to be judged (e.g.,time, money, or grades) varied. Moreover, studies investigat-ing the concept of risk have consistently demonstrated thatpreference judgments are distinct from risk judgments (e.g.,Lopes, 1984; Luce & Weber, 1986; Payne, 1975). This impliesthat risky choice cann ot be reduced to preferences determ inedpredominantly by risk attitudes.Indirect evidence of the shortcomings of weighted valuetheories comes from the apparent need and subsequent at-tem pts to develop subjective probability o r decision-weightingfunctions to describe how uncertainty influences choice (e.g.,Edwards, 1962; Kahneman & Tversky, 1979; Karmarkar,1978; Savage, 1954). Unfortunately, this type of function hasnot been shown to systematically increase the ability to predictchoice (e.g., Hershey & Schoemaker, 1980; Schneider &Lopes, 1986). These essentially psychophysical solu tions alsofunctionally deny the affective and stress-inducing impact ofrisks, ignoring the difficulty of trading off between uncertaintyand outcome.Framing Effects from a Larger Perspective

Framing effects are more than just the by-product of riskychoice. In fact, framing effects are quite common in judg-ments and choices that do not even involve risk. For instance,Levin and Gaeth (1988) have demonstrated that consumersrespond more favorably to ground beef when its characteris-tics are framed positively (75% lean) rather than negatively(25% fat). Maheswaran and Meyers-Levy (1990) have docu-mented framing effects involving the persuasiveness of posi-tively and negatively framed health-related messages. Huber,Neale, and Northcraft (1987) found that personnel selectiondecisions were influenced by whether the participa nts framedtheir task positively (applicants to accept for an interview) ornegatively (applicants to reject for an interview). Finally, anin-depth series of negotiation studies have demonstrated thata positive frame, as opposed to a negative frame, generally


14/18

FRAMING AND CONFLICT IN RISKY CHOICE 1053results in more successful performance, more concessionarybehavior, and the completion of more transactions both foramateurs and for experts (Bazerman et al., 1985; Neale &Bazerm an, 1985; Neale, Hube r, & Northcraft, 1987; Neale &Northcraft, 1986).In a related vein, the effect of actual changes in the valenceof outcomes has been explored in numerous studies designedto test long-standing approach-avoidance theories of conflict(Lewin, 1938, 1951; Miller, 1944, 1959). It is interesting thatthe vast majority of these studies corroborate my finding thatchoices between two negative or undesirable options (avoid-ance-avoidance) create greater conflict than choices betweentwo positive or desirable options (approach-app roach).As early as 1942, Barker demonstrated that subjects (boys,age 9-11) took longer to decide which of two aversive liquids(e.g., salt water vs. vinegar) they would rather drink than theydid to choose between two desirable liquids (e.g., pineapplevs. orange juice). In what must have been the first study offraming, Barker (1946) later demonstrated that choices be-tween negatively framed personal characteristics were moreoften uncertain than choices between virtually the same pos-itively framed characteristics.In a series of studies that followed Barker's lead, severalinvestigators (e.g., Arkoff, 1957; Minor, Miller, & Ditrichs,1968; Murray, 1975; Schill, 1966) demonstrated that makingchoices between obtaining mutually exclusive unfavorablepersonality characteristics took significantly longer and wasjudged as significantly more difficult than making choicesbetween obtaining one of two favorable personality charac-teristics. Moreover, when an undecided response category wasintroduced into the task (Murray, 1975), subjects were muchmore likely to use it in negative avoida nce-av oidanc e conflicts(42%) than in positive approach-approach conflicts (6%).Using a pencil-and-paper maze task modeled after behav-ioral studies in rats, Smith and Epstein demonstrated thatcollege students' speed of response was slower, with moreerrors and more response variability, when the choice in-volved an avoidance-avoidance conflict rather than an ap-proach-approach conflict (Epstein & Smith, 1967; Smith &Epstein, 1967; Smith & Gehl, 1974). In another study, skinconductance measures suggested that arousal was higher justbefore a possible loss than before a possible gain (Losco &Epstein, 1977).Oddly enough, studies within the a pproach-avoidanc e par-adigm bring us full-circle back to an investigation of riskychoice with A tthowe's (1960) psychophysical study of conflictin selecting gambles. He demonstrated not only that it tooksubjects longer to make their choices when gamble pairs hadnegative rather than positive expected values but also thatmost subjects tended to minimize the worst outcome and, inconsequence, made predom inantlyrisk-aversechoices in boththe positive and negative domains.In combination, this evidence presents a remarkably strongcase for the presence of exaggerated levels of conflict innegative as opposed to positive choice situations, whether ornot risk is involved. Just as the Lewinian (1938) explanationof this difference in experienced conflict relies on assum ptionsconcerning the interaction of unfulfilled needs and environ-mental forces, an understanding of the effects of framing on

risky choice requires an explicit recognition of both the mo-tivational and situational factors that have a primary impacton the decision-making process.An Alternative Approach to Risky C hoice

Although em pirical investigations of conflict defined withinthe approach-avoidance tradition have become increasinglyrare in recent years, the importance of motivational as wellas situational factors in risky choice has not gone completelyunnoticed. In her SP/A (security-potential/aspiration) theory,Lopes (1984, 1987, 1990) proposed that risky choice is pri-marily the consequence of motivational predispositions en-acted within situational constraints. Rather than relying onrisk attitudes, the theory posits that individuals can be dispo-sitionally located on a motivational con tinuum between needsfor security and needs for potential. Those who have relativelystrong security needs (security seekers) are predominantlymotivated to avoid relatively bad outcomes, whereas thosewho have relatively strong potential needs (potential seekers)are predominantly motivated to obtain relatively good out-comes. Both of these dispositions lead to the overweightingof some potential outcomes relative to others. Security seek-ers, who are hypothesized to represent the m ajority of people,tend to focus on and overweight the worst outcomes inchoices, whereas potential seekers tend to focus on and over-weight the best outcomes.

Behaviorally, security seekers will frequently act in a risk-averse fashion. Rather than reflecting an attitude toward riskper se, however, this behavior may reflect a general tendencyto experience a stronger desire to avoid the worst outcome ofthe risk than to approach the best outcome of the risk. Hence,a sure thing of equal expected value may seem more appeal-ing. Nevertheless, security seekers need not always behave ina risk-averse fashion. When the difference in worst outcomesacross available options is small, security seekers may take arisk if it seems to offer a substantial advantage in terms ofpotential. (See Lopes, 1987, and Schneider & Lopes, 1986,for more detailed, empirically based accounts.)The security-potential continu um is just one of two majorfactors proposed to influence risky choice. SP/A theory alsoholds that risky choice is affected by situationally co nstrainedaspiration levels or goals. An aspiration level represents thesmallest outcome that would be deemed satisfactory by thedecision maker, given the current choice situation. Althoughthe setting of aspiration levels may not be completely inde-penden t of the disposition of the individual, it is quite sensitive

to situational constraints and immediate needs. Once set, theaspiration level leads to the relative overweighting of out-comes that would achieve or exceed that level.In explaining the results of their study, Schneider and Lopes(1986) suggested that typical subjects (i.e., security seekers)had strong risk-averse preferences over a set of gain lotteriesbecause security needs and aspiration levels favored the sameselections. For the mirror-reflected loss lotteries, however,subjects sometimes chose the riskier option and sometimeschose the lessriskyoption. Here, it was suggested that securityneeds and aspiration levels favored opposing selections andconsequently led to conflict in choice. Risk-averse preferences


15/18

1054 SANDRA L. SCHNEIDERpredom inated when the worst possible outcom es of the riskieroption were particularly large, but risk-seeking preferencespredominated when there was virtually no possibility ofachieving the aspiration level with the less risky op tion.Although the SP/A theory account seems to map quitereadily onto the major results of this study, there is oneproblematic consideration. In this experiment, the actualoptions did not differ from the positive to the negative con-ditions, as they did in Schneider and Lopes's experiment.Because options framed positively or negatively represent thesame objective state of affairs, security (and potential) needswould not be expected to differ across frames because theworst (and best) outcomes remain the same. Hence, unlessaspiration levels change systematically from one frame to theother, it is not clear how SP/A theory could accommodatethe observed results.The Aspiration Level Contingency

It may be that goals or aspiration levels are contingent onone's perspective of the situation as positive or negative.Indeed, this possibility is supported by the popular notionthat losses loom larger than gains, or that losses are moreaversive than equivalent gains are attractive (e.g., Kahneman& Tversky, 1979: Kogan & Wallach, 1967; Slovic & Lichten-stein, 1968). This virtual truism implies that people have astronger desire to minimize losses than to maximize gains.Hence, in setting goals, people may be more insistent onachieving small losses than they will be on achieving largegains. All in all, people may be willing to settle for less in thedomain of gains than in the domain of losses.However, even if losses do loom larger than gains andpeople do set more ambitious aspiration levels when negativeoutcomes are involved, the question remains as to why thishappens. Although Kahneman and Tversky (1979) have sug-gested that losses are perceptually of a larger magnitude thanobjectively equivalent gains, there is also a more cognitiveexplanation that is worth considering.The crux of this explanation is that (a) setting the aspirationlevel involves an anchoring and adjustment process, and (b)the status quo serves as the fundamental reference point, oranchor, in that process.5 In particular, the claim is that aspi-ration levels are set relative to gaining nothing or losingnothing, depending on the frame. That is, no change is thebenchmark against which aspiration levels are determined.To dem onstrate how this can explain differences in aspirationlevels across domains, consider the following example basedon this study's variation of the Asian disease problem involv-ing a threat to 600 lives.If one assumes only that setting aspiration levels involvesan anchoring and adjustment process wherein the status quoserves as the ancho r, then one m ight expect that the aspirationlevel for both gains and losses involves an adjustment ofapproximately the same amount. Hence, the aspiration levelof a hypothetical subject in the positive frame might be tosave at least 200 people, and in the negative frame might beto avoid losing any more than 200 people. At first glance, thetwo aspiration levels have an intuitively appealing symmetryand may even appear to be equivalent. On second glance, it

becomes clear that saving at least 200 people is actuallyequivalent to losing no more than 400 people. In the sameway, to avoid losing more than 200 people is to save at least400 people.Notice that the aspiration level in the positive frame is lessthan the expected value of 300 lives saved or lost, whereas theaspiration level in the negative frame is more than the ex-pected value. Thus, the use of a straightforward, intuitivelyappealing, and superficially reasonable strategy for settingaspiration levels results in easy-to-achieve goals in the positivedomain and hard-to-achieve goals in the negative domain.The consequence is an aspiration level contingency on do-main.SP/A Theory and the Aspiration Level Contingency

The aspiration level contingency in conjunction with SP/Atheory's need for security (being typically stronger than theneed for potential) suggests one explanation for why prefer-ences tend to be risk averse in the positive domain butconflicted in the negative domain. As a general case, aspira-tion levels less than the expected value favor risk-averseresponses (i.e., obtaining the sure thing is satisfactory), andaspiration levels greater than the expected value favor risk-seeking responses (i.e., obtaining the sure thing is not goodenough). As noted earlier, the need for security (i.e., foravoiding the worst outcome) favors risk-averse responding.Hence , with the low aspiration levels postulated in the positivedomain, both the need for security and the modest aspirationlevel can be satisfied by selecting a sure thing. With highaspiration levels in the negative domain, however, conflictensues as security needs favor a risk-averse choice but thestringent aspiration level favors arisk-seekingchoice.The results of this study can be used once again to illustratehow these two factors might interact in risky choice. In thepositive frame, the aspiration level will typically be less thanthe expected value; again, using the Asian disease problem asan example, it might be set at saving at least 200 lives. Boththe ST and 25/75 in the positive frame provide the means toachieve this modest aspiration level with certainty and arestrongly preferred by subjects over 75/25 and 50/50.6 In thenegative domain, however, where the aspiration level is typi-

5 Although the status quo has been implicated as a typical referencepoint for risky choice within prospect theory, it has not been deemedessential, or of fundamental impo rtance, to the decision process or tocognitive representations of the decision task.6 This account is also consistent with the observed tendency toprefer the 25/75 risk over the sure thing (ST) in the positive domain.Although both of these options satisfy the proposed aspiration levelof 200 lives saved, the 25/75 option may seem preferable due to itsoutcom e ch aracteristics. There is little difference in the security (worstoutcomes) of 25/75 and the ST (200 vs. 300 lives saved), whereas the25/75 option enjoys a substantial advantage over the ST in terms ofpotential. Specifically, the 25/75 option allows for the possibility thatall of the people will be saved. Thu s, whereas extrem e security seekersmight maintain a preference for the ST on the basis of the smallimprovement in worst outcome, moderate security seekers would belikely to prefer the 25/75 option.


16/18

FRAMING AND CONFLICT IN RISKY CHOICE 1055cally greater than the expected value, the goal might be set atlosing fewer than 200 lives (i.e., saving at least 400 lives). Inthis example, the more severe aspiration level cannot be metfor certain by any of the options. As a result, conflict isexpected, and observed, for virtually all pairs.At first, the option that offers at least a chance of reachingthe aspiration level might seem preferable, but then the real-ization that one could do substantially worse than in the otheroption may cause the riskier option to look less appealing,especially for individuals who are strongly security-oriented.The consequence is likely to be vacillation back and forth ina frustrating attempt to isolate "the lesser of two evils." Risk-seeking choices would be expected to occur whenever subjectsdecide they must maintain some chance of attaining theaspiration level, and risk-averse choices whenever subjectsdecide that their first priority is to avoid the maximum loss. 7Docum enting the Aspiration Level Contingency

Although there is, as yet, no direct evidence regarding theplausibility of the aspiration level contingency, there is at leastone study that provides interesting preliminary support. Puto(1987) asked over 300 professional buyers to participate in ahypothetical decision task involving a choice between a riskyversus a sure contract price from two competing suppliers.Along with many other results, Puto found that among buyerswho initially set an easy-to-attain aspiration level, those whowere then given a negatively framed sales message (i.e., "avoidlosing an important cost advantage to your major competi-tors") shifted to a m ore difficult-to-attain final aspiration levelthan those who were given a different sales message. In thesame way, buyers who initially set a difficult-to-attain aspi-ration level and were then given a positively framed salesmessage (i.e., "gain an important cost advantage over yourmajor competitors") later shifted to a relatively easy-to-attainfinal aspiration level.

In addition to this preliminary support, the aspiration levelcontingency is appealing because it may contribute to theunderstanding of framing and conflict in choice, not onlywhere risk is present but also where choices do not involverisk. For instance, in the Levin and Gaeth (1988) study ofconsumer choice, subjects' expectations regarding the pres-ence of positive features in ground beef (% lean) may berelatively less demanding than their expectations for the lackof negative features (% fat). In a similar manner, in negotia-tions, where excess concessionary behavior is commonly at-tributed to low aspiration levels (Pruitt, 1981), it seems likelythat the high level of concessionary behavior observed underpositive framing (relative to negative framing) can also beattributed to relatively low aspiration levels (e.g., Neale &Bazerman, 1985).It is obvious that much work needs to be done in theinterest of empirically specifying individu als' aspiration levelswithin particular situations as well as documenting aspirationlevel contingencies on domain. Moreover, the cognitive proc-esses responsible for the aspiration level contingency must beexplored. The general process proposed here, staying close tothe status quo through an anchoring and adjustment processfor setting aspiration levels, suggests a natural and potentially

adaptive m eans of safeguarding against radical changes of life-style, especially in the negative direction.The Status Qu o as Frame

Everything is relative, not the least of which are ou r cogni-tive representations of the events and decisions we encounterthroughout our lives. It has long been known that cognitiverepresentations tend to be categorical in nature and that theyare organized in principled ways into structures generallyreferred to as schemata (see, e.g., Anderson, 1990; Fiske &Taylor, 1991; Rosch & Lloyd, 1978; Smith & Medin, 1981).The form of these schemata depends critically on our under-standing and interpretation of the kinds of experiences wehave had previously. Needless to say, the goals and aspirationswe set are developed with reference to the organized andcategorical structure of our knowledge.Kahneman and Tversky (1984) have suggested that gainsand losses are categorically distinct. This is obviously consist-ent with current theories of the nature of concepts. It seemsreasonable to expect that the status quo is also categoricallydistinct. In fact, the concept of the status quo may be ofspecial importance, not only because it is a prerequisite forestablishing gain and loss categories but also because it mayserve as the foundation for the construction of cognitiverepresentations of choice and, specifically, of decision frame.Consider, for example, what is implied by terms such assave and lose in the context of framing problems such as theAsian disease problem. If people need to be saved, it must bethat the perceived given state of affairs presupposes theirdeath. It is only for a status quo that accepts the deaths asimminent that the construct of saving has meaning. In muchthe same way, the construct of losing only has meaning whenassociated with a status quo that implies their unthreatenedexistence.The framing of a decision problem is tantamount to amanipulation of the perceived status quo. A frame determinesthat which the decision mak er assumes to be true or inevitableat the current point in time. The cognitive representation ofthe decision task is constructed or modified relative to thestatus quo that is adopted by, or imposed on, the decisionmaker. The frame influences the perception and representa-tion of reality by implying what should be accepted as thecurrent state of affairs.A focus on the status quo as the reference point for inter-preting and cognitively representing decision tasks em phasizesthe role of perspectives, expectations, and assumptions in

7 The primary purpose of this article is to bring the role of conflictto the attention of those interested in how frames may influencechoice and to provide a possible explanation for why different framesmay produce different levels of conflict. Nevertheless, it is also criticalto consider the process by which conflicts in choice are resolved.Busemeyerand Townsend (1991; Townsend & Busemeyer, 1989) arecurrently refining an elegant quantitative model of the conflict reso-lution process inspired by Lewin's (1951) field theory. Meanwhile,Coombs and Avrunin (1988) have explored not only the quantitativecharacteristics of conflict situations in general but also have recom-mended a structure for evaluating methods of resolving conflict.


17/18

1056 SANDRA L. SCHNEIDERdetermin ing specific goals and beh avior. It also highlights thedanger of making unwarranted assumptions or of adoptingoverly rigid outlooks on events. Finally, it paves the way forunderstanding when people are likely to set easy or difficultgoals, under what circumstances people will be more apt toagree or disagree with one ano ther, and when people are likelyto experience conflict.

ReferencesAnderson, J. R. (1990). Cognitive psychology and its implications(3rd ed.). San Francisco: Freeman.ArkofT, A. (1957). Resolution of approach-approach and avoidance-avoidance conflicts. Journal of Abnormal and Social Psychology,55, 402-404.Atthowe, J. M., Jr. (1960). Types of conflict and their resolution: Areinterpretation. Journal of Experimental Psychology, 59 , 1-9.Barker, R. G. (1942). An experimental study of the resolution ofconflict by children. In Q. McN ema r & M. A. Merrill (Eds.), Studiesin personality (pp. 13-34). New York: McGraw-Hill.Barker, R. G. (1946). An experimental study of the relationship

between certainty of choice and the relative valence of the alter-natives. Journal of Personality, 15 , 41-52.Bazerman, M. H. (1982). The framing of organizational behavior.Paper presented at the meeting of the Academy of Management,New York. NY.Bazerman, M. H. (1983). Negotiator judgment: A critical look at therationality assumption. American Behavioral Scientist, 27, 2 1 1 -228.Bazerman. M. H.. Magliozzi, T., & Neale, M. A. (1985). Integrativebargaining in a competitive market. Organizational Behavior an dHuman Decision Processes, 35, 294-313.Bierman. H., Jr. (1989). The Allais paradox: A framing perspective.Behavioral Science, 34. 46-52.Brewer. M. B., & Kramer, R. M. (1986). Choice behavior in socialdilemm as: Effects of social identity, group size, and decision fram-ing. Journal of Personality an d Social Psychology, 50 , 543-549.Burton, S., & Babin, L. A. (1989). Decision-framing helps make thesale. Th e Journal oj Consumer Marketing, 6, 15-24.Busemeyer. J. R., & Townsend, J. T. (1991). Decision field theory: Adynamic-cognitive approach to decision-making in an uncertainenvironment. Manuscript submitted for publication.Cohen, M.. Jaffray. J. Y.. &Said, T. (1987). Experimental comparisonof individual behavior under risk and under uncertainty for gainsand for losses. Organizational Behavior and Human Decision Proc-esses, 39, 1-22.Coombs. C. H.. & Avrunin. G. S. (1988). The structure of conflict.Hillsdale, NJ: Erlbaum.Edwards, W. (1962). Subjective probabilities inferred from decisions.Psychological Review, 69 , 109-135.Epstein, S., & Smith, B. (1967). Modes and adequacy of resolutionof three basic types of cognitive-motor conflict. Journal of Experi-mental Psychology, 74 , 264-271.Fagley, N. S., & Kruger, L. (1986, August). The effect of problemframing on the program choices of school psychologists. Paperpresented at the 94th Annual Convention of the American Psycho-logical Association, Washington, DC.Fagley, N. S., & Miller, P . M. (1987). T he effects of de cision framingon choice of risky vs certain options. Organizational Behavior an dHuman Decision Processes, 39, 264-277.

Fischhoff. B. (1983). Predicting frames. Journal of ExperimentalPsychology: Learning, Memory, and Cognition, 9, 103-116.Fiske. S. T.. & Taylor. S. E. (1991). Social cognition (2nd ed.). NewYork: McGraw-Hill.

Fleishman, J. A. (1988). The effects of decision framing and others'behavior on cooperation in a social dilemma. Journal of ConflictResolution, 32, 162-180.Frisch, D. (in press). Reasons for framing effects. OrganizationalBehavior and Human Decision Processes.Ham merton, M ., Jones-Lee, M. W., & Abbott, V. (1982). Equity andpublic risk: Some empirical results. Operations Research, 30, 203-207.Hershey, J. C , & Schoemaker, P. J. H. (1980). Prospect theory'sreflection hypothesis: A critical examination. Organizational Be -havior and Human Performance, 25 , 395-418.Huber, V. L., Neale, M. A., & N orthcraft, G . B. (1987). Decision biasand personnel selection strategies. Organizational Behavior andHuman Decision Processes, 40, 136-147.Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysisof decision under risk. Econometrica, 47 , 263-291.Kahnem an, D., & Tversky, A. (1984). Choices, values, and frames.American Psychologist, 39 , 341-350.Karmarkar, U. S. (1978). Subjectively weighted utility: A descriptiveextension of the expected utility model. Organizational Behaviorand Human Performance, 21 , 61-82.Keller, L. R. (1985). An empirical investigation of relative riskaversion. IEEE Transactions on Systems, Man, and Cybernetics,15, 475-482.Kogan, N., & Wallach, M. A. (1967). Risk taking as a function of thesituation, the person, and the group. In G. Mandler, P. Mussen, N.Kogan, & M. A. Wallach, New directions in psychology (Vol. 3,pp . 111-278). New York: Holt, R inehart & W inston.Kramer, R. M. (1989). Windows of vulnerability or cognitive illu-sions? Cognitive processes and the nuclear arms race. Journal ofExperimental Social Psychology, 25 , 79-100.Levin, I. P. (1990, May). Moderating influences on the effects ofinformation framing. Paper presented at the Third InternationalConference on Judgm ent and Decision M aking, Nags Head, NC.Levin, I. P., & Gaeth, G. J. (1988). How consumers are affected bythe framing of attribute information before and after consumingthe product. Journal of Consumer Research, 15, 374-378.Lewin, K. (1938). Th e conceptual representation and the measure-ment of psychological forces. Durha m, N C: Duke University Press.Lewin, K. (1951). Field theory in social science. Westport, CT:Greenwood Press.Lopes, L. L. (1984). Risk and distributional inequality. Journal ofExperimental Psychology: Human Perception an d Performance, 10,465-485.Lopes, L. L. (1987). Between hope and fear: The psychology of risk.In L. Berkowitz (Ed.), Advances in experimental social psychology(Vol. 20, pp. 2 55-295). New York: Academic Press.Lopes, L. L. (1990). Re-modeling risk aversion: A comparison ofBernoullian and rank dependent value approaches. In G. M. vonFurstenberg (Ed.), Acting under uncertainty: Multidisciplinary con-ceptions (pp. 267-298). Boston: Kluwer Academic.Losco, J., & Epstein, S. (1977). Relative steepness of approach andavoidance gradients as a function of magnitude and valence ofincentive. Journal of Abnormal Psychology, 86 , 360-368.Luce, R. D., & Weber, E. U. (1986). An axiomatic theory of conjoint,expected risk. Journal of Mathematical Psychology, 30 , 188-205.Maheswaran, D., & Meyers-Levy, J. (1990). The influence of messageframing and issue involvement. Journal of Marketing Research,27 , 361-367.McNeil, B., Pauker, S. G., Sox, H. C , & Tversky, A. (1982). On theelicitation of preferences for alternative therapies. New EnglandJournal of Medicine, 306, 1259-1262.Metzger, M. A., & Krass, K. (1988). Replication of judgmentalheuristics with simplified items and familiar content. PsychologicalReports, 63, 671-677.


18/18

Framing and Conflict

Documents

Transcript of Framing and Conflict