Organizational Behavior and Human Decision Processes 95 (2004) 198–207

[Link]/locate/obhdp

Effects of attribute framing on cognitive processing and evaluation

Bård Kuvaasa,*, Marcus Selartb
a
Department of Leadership and Organization Management, Norwegian School of Management, Box 580, N-1302 Sandvika, Norway
b
Department of Strategy and Management, Norwegian School of Economics and Business Administration, N-5045 Bergen, Norway

Received 13 June 2003

Available online 13 October 2004

Abstract

Whereas there is extensive documentation that attribute framing influences the content of peopleÕs thought, we generally know
less about how it affects the processes assumed to precede those thoughts. While existing explanations for attribute framing effects
rely completely on valence-based associative processing, the results obtained in the present study are also consistent with the notion
that negative framing stimulates more effortful and thorough information processing than positive framing. Specifically, results from
a simulated business decision-making experiment showed that decision makers receiving negatively framed information had signif-
icantly better recall than those receiving positively framed information. Furthermore, decision makers in the negative framing con-
dition were less confident than decision makers in the positively framed condition. Finally, compared to a no-framing condition,
decision makers receiving positive framing deviated significantly more in evaluation than decision makers receiving negative framing
did.
Ó 2004 Elsevier Inc. All rights reserved.

Keywords: Attribute framing; Cognitive processing; Susceptibility to framing

There is extensive documentation that the formula- ers take. Even though framing effects have been
tion of a decision stimulus or an event influences how documented across several decision making situations
decision makers think and decide. According to Tversky (see Kühberger (1998) and Levin et al. (1998) for reviews
and Kahneman (1981), a decision frame may be defined of the framing literature), the search for a deeper under-
as referring to the decision makerÕs conception of acts, standing of the cognitive processes that underlie valence
outcomes, and contingencies associated with a particu- framing has been limited (Levin et al., 1998). Indeed,
lar choice. Hereby, the frame that a decision maker Kühberger (1998) refers to the cognitive processes that
adopts depends heavily on how a decision problem is de- are responsible for framing effects as a ‘‘stepchild’’ of
scribed, although norms, habits, and personal character- framing research.
istics also play a part (Kahneman & Miller, 1986; In an attempt to better explain how positively and
Kühberger, 1998; Levin, Schneider, & Gaeth, 1998). negatively valenced information affect judgments and
Since Kahneman and TverskyÕs (1979) seminal work, a decisions, Levin et al. (1998) developed a typology to
large number of studies have demonstrated that framing distinguish between three distinct types of valence fram-
normatively equivalent information in positive versus ing effects with different underlying mechanisms and
negative ways, so called valence-based framing, may sys- consequences. Risky choice framing effects occur when
tematically affect the decisions or actions decision mak- willingness to take a risk depends on whether the poten-
tial outcomes are framed positively or negatively. Goal
*
Corresponding author. Fax: +47 6755 7678.
framing effects occur when a persuasive message has dif-
E-mail addresses: [Link]@[Link] (B. Kuvaas), [Link]@ ferent appeal depending on whether it stresses the posi-
[Link] (M. Selart). tive consequences of performing an act to achieve a

0749-5978/$ - see front matter Ó 2004 Elsevier Inc. All rights reserved.
doi:10.1016/[Link].2004.08.001
 B. Kuvaas, M. Selart / Organizational Behavior and Human Decision Processes 95 (2004) 198–207 199

particular goal or the negative consequences of not per- ing effects in evaluations. Improving our understanding
forming the act. Finally, attribute framing effects occur of the processes that underlie attribute framing will not
when evaluations of an object or event are more favor- only provide more satisfactory accounts of attribute
able if a key attribute within any given context is framed framing effects, but may also have implications for man-
in positive rather than negative terms. agerial decision-making and the emerging research on
Due to its simplicity (i.e., positive framing supports how to attenuate or eliminate framing biases among ‘‘re-
more favorable evaluations and negative framing sup- al world’’ decision makers (e.g., Hodgkinson, Bown,
ports less favorable evaluations), attribute framing is Maule, Glaister, & Pearman, 1999; Wright & Goodwin,
particularly suitable for providing insights into the nat- 2002). In order to detect a potential asymmetry in sus-
ure of the information processing differences resulting ceptibility to framing we compare whether positive
from positive and negative framing (Levin et al., and negative framing decision-makers deviate more or
1998). Levin and Gaeth (1988), Levin, Johnson, Russo, less in framing effects compared to decision makers
and Deldin (1985), and Levin et al. (1998) argue that receiving no framing information. To explore whether
attribute framing effects occur because information is framing may act as a catalyst for different modes of cog-
encoded relative to its descriptive valence. Positive label- nitive processing we include three different conceptual-
ing of an attribute will thus lead to an encoding of the izations of cognitive processing; self-reported
information that tends to evoke favorable associations analytical thinking reflecting the level of analytic versus
in memory, whereas negative labeling of the same attri- intuitive processing (Mantel & Kardes, 1999), and deci-
bute will cause an encoding that evokes unfavorable sion makersÕ amount of recall of decision event informa-
associations. According to such an associative model tion and their level of confidence in estimating the
explanation, material that is associatively linked to the accuracy of their own recall (Kuvaas & Kaufmann,
framing manipulation is more likely to be used in vari- 2004). Dunegan (1993) pointed to memory as an impor-
ous constructive cognitive tasks, leading to framing con- tant distinguishing characteristic of different cognitive
gruency in attention, learning, memory, associations, processing modes. After decision makers have made a
and eventually to positivity and negativity biases in eval- decision, they tend to recall much more information
uations and judgments. about the decision event when controlled or more effort-
The associative model may be valid in explaining why ful processing is used (Kernan & Lord, 1989; Langer,
attribute framing affects the nature of evaluations or the 1989a; Maheswaran & Chaiken, 1991). Regarding level
content of cognition (i.e., what people think). It may be of confidence, dozens of laboratory studies have docu-
less useful, however, in explaining more subtle effects mented the prevalence of too much confidence in judg-
attribute framing may have on the process of cognition ment, i.e., decision makersÕ certainty that their
(i.e., how people think). First, the associative model does predictions are correct most often exceeds the accuracy
not take into account the possibility that negative fram- of those predictions (Simon & Houghton, 2003). The
ing may stimulate more effortful and thorough cognitive tendency to be highly confident is found to be greater
processing than positive framing. Although only indi- when decision makers act as ‘‘cognitive misers’’ (Maha-
rectly evidenced in the context of attribute framing jan, 1992). High levels of confidence is thus more likely
(Dunegan, 1993), it is frequently observed that people to occur when decision makers engage in less effortful
encode and respond to positively and negatively va- processing, as indicated by limited information search
lenced affective and informational stimuli in systemati- (Cooper, Folta, & Woo, 1995; Harvey, 1994), truncated
cally different ways (e.g., Bless, 2002; Chatterjee, mental search where potentially critical information is
Heath, Milberg, & France, 2000; Dunegan, 1994; For- ignored (Mahajan, 1992) or used incorrectly (Au, Chan,
gas, 2002a; Peeters & Czapinski, 1990; Taylor, 1991). Wang, & Vertinsky, 2003).
Second, if the valence of attribute framing has system-
atic processing consequences, decision makers in posi-
tive and negative framing conditions may also differ in Theory and hypotheses
how affected they are by the framing information. Then,
however, the explanation of results showing that posi- Most contemporary models of individual informa-
tively valenced framing creates a positivity bias and neg- tion processing view discrepancy between a desired or
atively valenced framing a negativity bias solely by way expected state and an experienced state as an activator
of valence-based associative processes, may represent an of more rigorous and systematic information processing
oversimplified picture of how attribute framing affects (e.g., Beach & Mitchell, 1990; Lord & Maher, 1990;
cognitive processing. Louis & Sutton, 1991). In accordance with evolutionary
This being the case, the current study was undertaken and adaptive arguments (e.g., Peeters & Czapinski,
to explore potential asymmetrical influences of posi- 1990; Taylor, 1991), it is argued that because negative
tively versus negatively valenced framing on modes of or unfavorable incoming information may signal incon-
cognitive processing and susceptibility to attribute fram- sistency between expected or experienced conditions, or
200 B. Kuvaas, M. Selart / Organizational Behavior and Human Decision Processes 95 (2004) 198–207

adverse or threatening events, there is a tendency for flicted about which option to choose, rather than risk
decision makers to engage in more deliberate and careful seeking, Schneider (1992) suggested an aspiration level
analysis (e.g., Dunegan, 1994; Wofford & Goodwin, contingency where negative framing leads to higher
1990). Conversely, when information is positive, cogni- aspiration levels and higher awareness than positive
tive processing tends to be less thorough and systematic framing, consistent with a negative framing-systematic
(e.g., Dunegan, 1994; Klein, 1989). Consistent with the processing view. In the context of attribute framing,
conventional wisdom that ‘‘if it ainÕt broke donÕt fix Dunegan (1993) reported two studies that indirectly sup-
it’’, decision makers may feel justified in using a more port the view that attribute framing systematically influ-
cognitively economic response (Dunegan, 1994). ences modes of cognitive processing. Specifically, he
Most affect-cognition models make a similar proposi- found significant relationships between how subjects
tion (e.g., Bless, 2002; Forgas, 2002a, 2002b). According interpreted a situation and how they responded only in
to the feelings as information model, for instance, nega- the negative-frame condition and that current and tra-
tive affect informs the individual that current conditions jectory images were perceived to be less compatible
are problematic and await solution (Clore, Schwarz, & when feedback information was framed negatively.
Conway, 1994; Schwarz, 1990). To solve problems, Thus, we hypothesize:
information processing is characterized by adherence
Hypothesis 1. Decision makers receiving negatively
to established rules and procedures as well as attention
framed information will report a higher degree of
to detail and systematic information processing. Positive
analytical thinking, recall more information, and be less
affect, in contrast, informs the individual that current
confident than decision makers receiving positively
conditions are benign and signals that systematic infor-
framed information.
mation processing is unnecessary because no problem
awaits solution. Accordingly, in studies where partici- So far, we have argued that there may be systematic
pants are induced to feel positive or negative affect processing consequences associated with the valence of
and then are given a task in which their type of process- attribute framing. In this section, we discuss contrasting
ing can be inferred, those in negative emotional states or views on how potential processing differences may affect
moods are usually found to engage in more bottom-up susceptibility to attribute framing effects in evaluation.
and systematic processing, whereas those in positive The conventional explanation for cognitive biases is that
emotional states or moods are found to engage in more people in general tend to rely on heuristic information
top-down and more heuristic processing (e.g., Forgas, processing or judgmental short cuts to form inferences
2002a; Tiedens & Linton, 2001). Although there are dif- (Tversky & Kahneman, 1974). Accordingly, if negative
ferent explanations for this effect (Bless, 2002), most af- framing decision makers engage in more systematic
fect-cognition models stipulate that the information and relatively less heuristic processing than positive
processing consequences arise from the valence of the framing decision makers, they should also deviate less
affective state (Tiedens & Linton, 2001). It should be from rational norms of decision making and be less sus-
noted, however, that the finding of positive–negative ceptible to attribute framing effects. Indirect support for
information processing asymmetry is qualified by excep- this prediction is provided by research evidence showing
tions (e.g., Ashby, Isen, & Turken, 1999; Isen, 1993; that induced elaborated thinking eliminates or reduces
Staw & Barsade, 1993), contextual contingencies (e.g., risky choice framing effects, such as when decision mak-
Kuvaas & Kaufmann, 2004; Martin, 2001; Martin & ers are asked to provide justification for their choice or
Stoner, 1996), and evidence that other affective dimen- spend more time on a decision making task (e.g., Sieck
sions than valence can influence processing (e.g., Tie- & Yates, 1997; Takemura, 1982, 1994). Another piece
dens & Linton, 2001). of indirect evidence stems from studies revealing that
While activating more systematic or effortful cogni- subjects with higher need-for-cognition scores1 are less
tive processing when experiencing an unfavorable or un- susceptible to risky choice framing effects (Chatterjee
happy rather than a favorable or happy situation may et al., 2000; Smith & Levin, 1996). Similarly, Stanovich
seem quite logical and functional, the question ad- and West (1998) found that students with higher aca-
dressed here is whether a similar response will emerge demic aptitude, as measured by SAT scores, were not
when equivalent information is merely framed in a posi- as affected by risky choice framing as students with low-
tive or negative way. Whereas Levin et al. (1998) did not er academic aptitude. Furthermore, although Levin,
acknowledge positive–negative information processing Gaeth, Schreiber, and Lauriola (2002) found no effect
asymmetry as a potential theoretical account for attri- of need-for-cognition on attribute framing, they found
bute framing effects, they relied on this mechanisms to that of those subjects showing large difference scores2
explain goal framing effects, and Kühberger (1997) dis-
cussed this possibility in a risky choice framing context. 1
A measure of the tendency to engage in effortful, analytical
Furthermore, based on several experiments where par- thinking (Cacioppo & Petty, 1982).
ticipants in the negative risky choice frame were con- 2
Equal to or greater than 1.0 in magnitude on a 7-point scale.
 B. Kuvaas, M. Selart / Organizational Behavior and Human Decision Processes 95 (2004) 198–207 201

between the positive and negative condition, 39 were in likely to invest the cognitive effort required to search
the positive direction while only 3 were in the negative for, find, and remember subtle themes that run through
direction. the information to which they are exposed. Similarly,
In affect-cognition research, the kind of vigilant, sys- Rothman and Salovey (1997) argue that cognitive pro-
tematic attention to stimulus details caused by negative cessing mediate the influence of framing and remind
affect has been used to explain enhanced stereotype and us that in order for framing to have an effect, a mini-
priming effects under positive mood and greater sensitiv- mum level of attention or systematic processing is re-
ity of negative mood subjects to strong and weak argu- quired. Thus, one may also argue that more effortful
ments in persuasion (Fiedler, 2000). Moreover, Forgas or systematic processing will result in higher selective
(1998) has found processing consequences of negative attention to framing-congruent information and to the
mood to reduce and even eliminate such common judg- forming of more framing-congruent associations than
mental biases as the fundamental attribution error. He less effortful or systematic processing. For instance, in
has also reported evidence that positive affect tends to explaining stronger impact of negative than of positive
increase, and negative effect to reduce, the likelihood goal (message) framing, Levin et al. (1998) refer to the
of other kinds of cognitive mistakes in social thinking negativity bias that people pay greater attention to
(Forgas, 2000). Even though there is much evidence that and are influenced more by negative than by compara-
positive affect may also produce distinct processing ble positive information. Affect-cognition research
advantages (e.g., increased flexibility and more creative points to a similar effect, arguing that affect infusion
thinking3), the findings reported above suggest that po- is greater when people engage in more extensive and
sitive attribute framing should create greater framing ef- elaborate thinking (e.g., Forgas, 2001). Furthermore,
fects than negative attribute framing: research reviewed by Cacioppo, Petty, Feinstein, and
Jarvis (1996) supports the view that individuals high
Hypothesis 2. Compared to decision makers receiving
in need-for-cognition are more susceptible to the influ-
no framing information, decision makers receiving
ence of affective states on cognition than are individuals
positively framed information will be more influenced
low in need-for-cognition. Thus, if negative framing
by framing information in evaluation than will decision
facilitates more effortful or systematic processing than
makers receiving negatively framed information.
positive framing, we may expect that negative attribute
The position that cognitive effort should reduce the framing will create greater framing effects than positive
effects of framing has received modest empirical support attribute framing:
(LeBoeuf & Shafir, 2003) and is challenged by the sug-
Hypothesis 2 0 . Compared to decision makers receiving
gestion that effortful processing may sometimes increase
no framing information, decision makers receiving
framing effects. First, studies indicating that need-for-
negatively framed information will be more influenced
cognition and SAT scores reduce framing effects have
by framing information in evaluation than will decision
used risky-choice framing tasks. In such tasks (e.g.,
makers receiving positively framed information.
the ‘‘Asian Disease Problem’’), where the ability to
understand and use rule-based decision aids in a consis-
tent manner would be appropriate, decision makers
engaging in systematic processing may have an advan- Methodology
tage. This may not necessarily be the case for attribute
framing, however. Second, Crawford and Skowronski Subjects and procedure
(1998) found that participants with higher need-for-cog-
nition scores exhibited greater sensitivity to biasing ef- Seventy-three undergraduate students participated in
fects than participants with lower need-for-cognition the study. They were all enrolled in an organization the-
scores when the decision task involved recalling stereo- ory course at a business school in Norway. As part of a
type-consistent information, and research on persuasion voluntary classroom exercise, participants were asked to
suggests that high need-for-cognition subjects may take part in a decision-making study where they should
sometimes be more affected by framing than their low act as project managers of a student organization in a
need-for-cognition counterparts (Wegener, Petty, & simulated business situation. The task was to decide
Klein, 1994). Consistent with the associative model on how much the student organization should invest
explanation of attribute framing, as suggested by Craw- in a particular project. Participants were given a booklet
ford and Skowronski (1998); Levin et al. (1998) argue of materials that included a business scenario, a manip-
that higher need-for-cognition individuals are more ulation check measure, an instruction to make an invest-
ment decision, and a measure of self-reported analytical
thinking. At the end of the task, the responses were col-
3
See for instance the special issue of Psychological Inquiry on the lected and a second instrument designed to measure re-
role of affect in social thinking and behavior (2002, Vol. 13, No. 1). call and level of confidence was presented to the
202 B. Kuvaas, M. Selart / Organizational Behavior and Human Decision Processes 95 (2004) 198–207

subjects. At the end of both tasks, respondents were de- answer (‘‘what are the names of the student organiza-
briefed about the study. tionÕs three main sponsors’’ and ‘‘in addition to yourself
being the project leader, what are the titles of the two
Variables other members of the management team’’). These were
scored 5 if one out of two answers were correct, and
Framing 3.33 if one out of three were correct. An observed recall
The brief business scenario (slightly less than one score was created for each subject by averaging over the
page) in which an investment decision had to be made 10 questions. Since recall was assumed to be a formative
about funding allocations for a project team was used as opposed to reflective measure (i.e., the items/observed
to manipulate framing (see the Appendix). All partici- variables are assumed to cause the latent variable/con-
pants received the same scenario, with one slight modi- struct, rather than that the items/observed variables
fication: the positive framing group (n = 25) read, ‘‘of are caused by the latent variable/construct (e.g., Bollen
the projects undertaken by this team, seven of the last & Lennox, 1991)). Under these conditions, internal con-
10 have been successful.’’ The corresponding scenario sistency between items was of minimal importance be-
provided to the negative framing group (n = 23)4 read, cause items that might even be negatively related can
‘‘of the projects undertaken by this team, three of the serve as meaningful indicators of a construct (Nunnally
last 10 have been unsuccessful,’’ while the no-framing & Bernstein, 1994). Accordingly, we did not compute
group (n = 25) was not informed about prior perfor- CronbachÕs alpha for the index.
mance. With the exception of this sentence, all informa-
tion was the same. The framing effect was measured by a Level of confidence
three-item 9-point semantic-differential scale (Mittal & Level of confidence was measured by first asking
Ross, 1998) reflecting the degree of negative–positive respondents to estimate the level of certainty for correct
evaluation of the scenario (a threat/opportunity for the answers on the recall-questions. Immediately after each
organization, a potential for losing/making money, of the 10 recall question they were asked to indicate, on
and a positive–negative situation). The CronbachÕs al- a 0–100%-scale, how sure they were that each of their
pha for the scale was .82 and a composite score was cre- answers were correct. We then divided these responses
ated for each subject by averaging the items. by ten, and used the difference between certainty ratings
and accuracy or actual recall to calculate 10 confidence
Self-reported analytical thinking scores ranging from
10 to 10, where a positive score
Self-reported analytical thinking was assessed using implied overconfidence, a negative score implied under-
six previously used items (Mantel & Kardes, 1999) that confidence, and a score of 0 implied a perfectly cali-
were modified to fit the current research problem. The brated response. For instance, if the certainty score for
items were presented to the participants immediately a correct answer (scored 10) were 100%, the confidence
after they had made the investment decision. Examples score of this item would be perfectly calibrated (10–
of items are: ‘‘The answer just came to me’’ (reverse 10 = 0). If the certainty score for a correct answer were
coded); ‘‘My decision was based on facts rather than 90%, the confidence score would constitute an under-
on general impressions and feelings’’; ‘‘My decision confident item response (9–10 =
1). Conversely, if the
was based on careful thinking and reasoning.’’ Agree- certainty score for a wrong answer (scored 0) were
ment was expressed on a 1 = ‘‘strongly disagree’’ to 90%, this would constitute a highly overconfident item
7 = ‘‘strongly agree’’ scale. CronbachÕs alpha for the response (9–0 = 9). For the recall items that were asking
scale was .78 and a composite score was created for each for two titles and three names (of main sponsors),
subject by averaging the items. respondents were asked to provide confidence estimates
that both titles and all three names were correct.
Recall Respondents were instructed not to provide certainty
The business scenario used to manipulate framing estimates of unanswered recall items (resulting in a per-
contained several pieces of factual information such as fectly calibrated confidence score of 0, that is, neither
names and numbers that were used to define 10 recall- overconfident nor underconfident). A composite index
able questions (a sample questions is ‘‘how many stu- was created for each subject by averaging the 10 confi-
dents are members of the student organization’’). Each dence scores. Being based on the recall questions, confi-
question was scored from 0 to 10 dependent on the cor- dence was also seen as a formative index and no
rectness of the answer. With the aim of presenting par- CronbachÕs alpha was computed.
ticipants with items varying in degree of difficulty, we
developed two questions that requested more than one Manipulation check

The three-item 9-point semantic-differential scale

4
Two responses suffered from missing data. measuring the degree of negative–positive evaluation
 B. Kuvaas, M. Selart / Organizational Behavior and Human Decision Processes 95 (2004) 198–207 203

of the scenario was used to check the framing manipula- the framing effects scores given by the negative and
tion. Scores were from 1 to 9, where a higher score im- positive framing subjects. ANOVA revealed that the
plies a more positive evaluation. A one-way analysis of positive framing subjects deviated significantly more
variance (ANOVA) with framing manipulation (nega- from the mean of the no-framing group than the
tive, no-framing, or positive) as factor showed that the negative framing subjects (F1,45 = 18.30, p < .001;
group means for framing effects were different Mnegative = 1.06, SD = 1.22; Mpositive = 2.28, SD = .70).
(F2,70 = 68.68 p < .001), and TukeyÕs honestly significant A disadvantage with this method is that it does not take
difference test revealed that all three means were signif- into account the standard deviation of the no-framing
icantly different from each other (Mnegative = 4.42, condition. We hence computed the standardized differ-
SD = 1.22; Mno-framing = 5.48, SD = 1.06; Mpositive = ences (CohenÕs d) for the no-framing group and each
7.76, SD = .70). experimental group (Cohen, 1988). In these computa-
tions, pooled standard deviations were applied in each
Analysis and results case (Rosnow & Rosenthal, 1996). The standardized dif-
ference between the no-framing group and the positive
Multivariate analysis of covariance (MANCOVA) framing group (d c-pf = 2.53, with 100% non-overlap in
was used prior to testing of Hypothesis 1 because recall confidence intervals) and the standardized difference be-
and confidence were significantly correlated (r =
.60, tween the no-framing group and the negative framing
p < 0.001). MANCOVA can test whether there is a sig- group (d c-nf = .93, with 52% non-overlap in confidence
nificant multivariate effect of the independent variable intervals) did, however, not indicate that the standard
on the dependent variables after controlling the correla- deviation of the no-framing group in any way would dis-
tion between the latter. Using WilksÕs lambda (k), this confirm the results obtained from the ANOVA. Thus,
analysis revealed that framing (k = .70, F3,43 = 6.22, the hypothesis stating that when compared to decision
p < .01) was significantly related to the set of dependent makers receiving no framing information, decision mak-
variables. These results suggest that separate hypotheses ers receiving positively framed information will be more
testing for each dependent variable could be performed influenced by framing information in evaluation than
without a high risk for an inflated Type I error. will decision makers receiving negatively framed infor-
The impact of framing on each of the dependent vari- mation (Hypothesis 2), was supported, while the com-
ables was tested using between-subjects analyses. These peting hypothesis (Hypothesis 2 0 ) was not.
analyses revealed that the two groups did not differ sig-
nificantly from each other in self-reported analytical
thinking (F1,45 = .35, n.s.; Mnegative = 4.10, SD = .94; Discussion
Mpositive = 3.93, SD = 1.01), but that the negative fram-
ing group had significantly better recall (F1,45 = 14.56, Consistent with prior theorizing (Levin et al., 1998),
p < .001; Mnegative = 7.93, SD = 1.15; Mpositive = 6.43, valence-based associative processing is probably a valid
SD = 1.53) and lower confidence (F1,45 = 9.93, p < explanation of how attribute framing affects the content
.001; Mnegative =
.37, SD = .96; Mpositive = .66, SD = of peopleÕs thought, and thus why positive attribute
1.26) than the positive framing group.5 Accordingly, framing led to more positive evaluations of the business
the hypothesis (Hypothesis 1) that decision makers scenario than negative attribute framing in our study.
receiving negatively framed information will report a The obtained differences in recall, confidence, and sus-
higher degree of analytical thinking, recall more infor- ceptibility to framing effects, however, suggest that
mation, and be less confident than decision makers asymmetrical information processing may also play a
receiving positively framed information was supported role in explaining attribute framing effects. Accordingly,
for recall and confidence (the positive framing subjects this study contributes to the framing literature by
were slightly overconfident and the negative framing extending previous accounts for attribute framing ef-
subjects slightly underconfident), but not for self-re- fects. It should be noted, however, that there is some dis-
ported analytical thinking. agreement on how to characterize negative–positive
To test the competing hypotheses on susceptibility to information processing asymmetry. In the framing liter-
framing we first calculated the differences between the ature, Schneider (1992), for instance, refers to aspiration
mean value framing effect of the no-framing group and levels that are more difficult to achieve and higher
awareness in the face of negative framing, while Levin
et al. (1998) refer to the negativity bias and that the im-
5
Scales: analytical thinking scores were from a 1 to 7 Likert scale, pact of negative information is systematically stronger
where a higher score implies a higher level of analytical thinking; recall than the impact of objectively equivalent positive
scores were from 0 to 10, where a higher score implies better recall;
confidence scores were from
10 to 10, where a positive score implies
information in goal framing. Furthermore, while prior
overconfidence, a negative score implies underconfidence, and a score affect-cognition research suggested that people experi-
of 0 implies a perfectly calibrated response. encing positive affect tend to employ more superficial
204 B. Kuvaas, M. Selart / Organizational Behavior and Human Decision Processes 95 (2004) 198–207

information processing strategies, more recent research effects, we need more research that investigates process
argues that the processing consequences are better and output separately.
understood in terms of use of general knowledge struc- This study has limitations that should be considered
tures and attention to the details of a situation (e.g., when interpreting the results. First, this study used an
Bless, 2002; Fiedler & Bless, 2000; Forgas, 2002a, 2002b). experimental design and, thus, suffers from the limita-
Prior attribute framing research has usually investi- tions of generalization to more complex natural situa-
gated evaluation effects of labeling a key attribute in po- tions. Moreover, the use of a student sample further
sitive versus negative terms without questioning limits the generalizability of the results. Second, to be
differences in susceptibility to framing effects. By includ- able to test potential effects of framing on recall and con-
ing a no-framing condition, we were able to examine the fidence, we had to present the framing manipulation in
magnitude of framing effects and found that participants the beginning of the business scenario (before the factual
receiving positive framing were more affected by attri- information used to measure recall and confidence). This
bute framing than those receiving negative framing. may have resulted in weaker manipulation effects than if
Also this finding is consistent with asymmetrical infor- the manipulation information had been presented in the
mation processing effects of negative versus positive last sentence, as is usually the case. Third, even though
framing and the position that more effortful and less participants were randomly assigned to the three groups,
heuristic processing may reduce susceptibility to cogni- we can not completely rule out the possibility that other
tive biases (e.g., Sieck & Yates, 1997; Takemura, 1994; factors than framing may have influenced the results, for
Tversky & Kahneman, 1974). This is not to say, how- instance individual differences in expectations and norms
ever, that effortful processing should always reduce (Kahneman & Miller, 1986) regarding the prior perfor-
framing effects. The business scenario used in our study mance information used to manipulate framing. Thus,
contained more information and was considerably more it is not impossible that the stronger framing effect that
complex and ambiguous than is usually the case in attri- was obtained in the positive frame could be interpreted
bute framing research. In such a research context, deci- as the deviation from some previously established norm.
sion makers using relatively more effortful processing The fact that we were stressing the potential successes of
may have been better able to counterbalance the fram- the funding in favor of failures in this framing condition,
ing information with other and more relevant informa- may thus by the audience have seemed like a ‘‘pitch’’
tion, than those using comparably less effortful constructed to enhance the likelihood that more funding
processing. Thus, for some tasks, probably most com- would be forthcoming. Hence, the positive frame may
plex or analytical challenging tasks, a facilitating effect have been regarded by some as an attempt to persuade
of more detailed processing should be more likely than rather than an invitation to evaluate the information.
an inhibiting or biasing effect (e.g., Au et al., 2003; This criticism is well in line with the views of constructiv-
Cacioppo et al., 1996; Huber & Seiser, 2001; Langer, ist theorists who argue that frames can be distorted by
1989b; LeBoeuf & Shafir, 2003; Levin, Huneke, & Jas- communicators acting strategically to secure their own
per, 2000; Louis & Sutton, 1991; Reger & Palmer, instrumental interests.
1996). However, when the task used to investigate attri- Finally, our measures of recall, confidence and self-
bute framing effects contains little information other reported analytical thinking represent indirect opera-
than the framing manipulation, which is often the case tionalizations of cognitive processing. This limitation is
(for instance whether beef is 80% lean versus 20% fat), evident in the non-significant differences for self-reported
more effortful processing should lead to more framing analytical thinking, which may in part be explained by a
congruent cognitive associations and probably result in social desirability bias, leading participants to report
greater framing effects than less effortful processing. having thought more carefully about their decisions than
Furthermore, particularly within the affect-cognition lit- they actually did. It is also possible that it is difficult for
erature, there is growing consensus that processing con- people to give accurate descriptions of their level of
sequences of both negative and positive stimuli have information processing (e.g., Svenson, 1989). We know,
distinct advantages (e.g., Forgas, 2002b; Isen, 2002). Po- for instance, that people often think of themselves as ac-
sitive mood has for instance been found to promote tive and purposeful decision makers, while research in
more flexible and creative thinking styles (Forgas, managerial cognition suggest that they may be predis-
2002a; Isen, 2002). It is probably safe to conclude then, posed toward using relatively effortless information pro-
that the impact of processing differences on evaluation, cessing most of the time (e.g., Louis & Sutton, 1991;
judgment, choice and behavior depends on context, in Mitchell & Beach, 1990; Reger & Palmer, 1996; Walsh,
particular the nature of the task. This, in turn, makes 1995). In support of these arguments, self-reported ana-
it very difficult to explain how people think (i.e., cogni- lytical thinking was not significantly correlated with re-
tive and affective process) based on studies of what peo- call or confidence. However, social desirability should
ple think (e.g., evaluation and decisions). Accordingly, not be a problem for our incidental recall test and the
to better understand the processes that underlie framing confidence measure. Moreover, using the same instru-
 B. Kuvaas, M. Selart / Organizational Behavior and Human Decision Processes 95 (2004) 198–207 205

ment in a similar research context, Kuvaas and Kauf- consists of 14 different project teams that are engaged
mann (2004) found that need for cognition was positively in planning and implementation of different student pro-
related to recall and negatively related to confidence. jects. Of the projects undertaken by the largest project
Accordingly, these two measures should be adequate indi- team, ‘‘The Student Week,’’ seven of the last 10 have
cators of cognitive processing, and clearly better than been successful (three of the last ten have been unsuc-
speculating about cognitive processes on the bases of out- cessful). The student organization is commercially run
put measures only (e.g., evaluation or decisions). Still, and needs more profitable than unprofitable projects.
alternative operationalizations of level of processing such The operations are organized by a management team
as response latencies (e.g., Svenson & Benson, 1993) or consisting of a marketing manager, a project coordina-
verbal protocols (e.g., Maule, 1994) are needed before tor and yourself being the leader of the group. The bud-
any firm conclusion can be drawn. get for 2000 is approximately 1 million NOK. The most
Despite these limitations, the results of this study may important source of income is membership fee from
have one important practical implication. Many in the around 1500 students. In addition, the student organiza-
popular press celebrate the benefits of a positive outlook tion has three main sponsors; Lillebrand Inc., Storeborg
(Dunegan, 1993) and strategic cognition researchers ar- Inc. and the Burseth Group.
gue that the potential power of viewing environmental In late March, the Student Week project team asks
events, trends and developments in a positive way is so you for 50,000 NOK more than originally budgeted in
strong that organizations should actively construct order to finish a project called ODIN – a project that
opportunities, and suppress threats (e.g., Dutton, 1993). should have been completed five weeks ago. ODIN
However, even though positive framing and opportunity has been running for six months and has already been
interpretation has the capacity to enhance motivation funded with 200,000 NOK. Despite the delay, the pro-
and flexible processing, and very high levels of threat per- ject group has great faith in the project. ODIN is a com-
ceptions may result in fixed or rigid processing (Staw, bined social and educational arrangement. Those
Sandelands, & Dutton, 1981), negative framing may have responsible for the project are students with different
the advantage of fostering more systematic processing. levels of experience in arranging student projects. The
Furthermore, managers are fare from immune to positive turnover of members is relatively high because students
illusions (Bazerman & Hoffman, 1999; Hayward & Ham- that get their degrees leave the project and continuity is
brick, 1997; Kahneman & Lovallo, 1993; Kuvaas, 2002) provided by written project reports. The members of the
and they often act as cognitive misers when making deci- project are not paid, but are motivated by the ‘‘work
sions, even important ones (Beach & Mitchell, 1990). experience’’ and networking associated with working
Thus, managers may benefit from knowing that both po- in student projects.
sitive and negative framing have distinct advantages and For the coming year, you have 500,000 NOK at your
disadvantages. This line of reasoning also implies that an disposal for funding new projects and you have to use
interesting avenue for future research is to explore poten- this budget if you want to spend the additional 50,000
tial benefits from purposely framing key information in NOK on ODIN. If ODIN fails, you will also have to
both positive and negative terms in ‘‘real world’’ deci- provide coverage of deficit.
sion-making or strategic conversations.
Framing research is only beginning to understand the
cognitive processes that underlie valence-based framing References
(Kühberger, 1997, 1998; Levin et al., 1998). The present
study was undertaken to increase our knowledge of Ashby, F. G., Isen, A. M., & Turken, A. U. (1999). A neuropsycho-
attribute framing effects and obtained support for posi- logical theory of positive affect and its influence on cognition.
Psychological Review, 106, 529–550.
tive–negative information processing asymmetry using
Au, K., Chan, F., Wang, D., & Vertinsky, I. (2003). Mood in foreign
both process (recall and confidence) and output (evalu- exchange trading: Cognitive processes and performance. Organi-
ation) indicators. Accordingly, even though Levin et zational Behavior and Human Decision Processes, 91, 322–338.
al.Õs (1998) typology of valence framing has made a very Bazerman, M. H., & Hoffman, A. J. (1999). Sources of environmen-
important contribution to the framing literature by tally destructive behavior: Individual, organizational, and institu-
tional perspectives. In R. I. Sutton & B. M. Staw (Eds.), Research
organizing and explaining risky choice, goal, and attri-
in Organizational Behavior (Vol. 21, pp. 39–79).
bute framing, the explanation of attribute framing may Beach, L. R., & Mitchell, T. R. (1990). Image theory: A behavioral
be further developed. theory of decision making in organizations. In B. Staw & L. L.
Cummings (Eds.). Research in organizational behavior (Vol. 12,
pp. 1–41). Greenwich, CT: JAI Press.
Bless, H. (2002). Where has the feeling gone? The signal function of
Appendix. Story used to induce positive(negative) framing
affetive states. Psychological Inquiry, 13, 29–31.
Bollen, K., & Lennox, R. (1991). Conventional wisdom on measure-
Imagine you are the project manager of a student ment: A structural equation perspective. Psychological Bulletin,
organization at a university. The student organization 110, 305–314.
206 B. Kuvaas, M. Selart / Organizational Behavior and Human Decision Processes 95 (2004) 198–207

Cacioppo, J. T., & Petty, R. E. (1982). The need for cognition. Journal cognition and decision making under uncertainty. Strategic Man-
of Personality and Social Psychology, 42, 116–131. agement Journal, 20, 977–985.
Cacioppo, J. T., Petty, R. E., Feinstein, J. A., & Jarvis, W. B. G. Huber, O., & Seiser, G. (2001). Accounting and convincing: The effects
(1996). Dispositional differences in cognitive motivation: The life of two types of justification on the decision process. Journal of
and times of individuals varying in need for cognition. Psycholog- Behavioral Decision Making, 14, 69–85.
ical Bulletin, 119, 197–253. Isen, A. M. (1993). Positive affect and decision making. In M. Lewis &
Chatterjee, S., Heath, T. B., Milberg, S. J., & France, K. R. (2000). J. Haviland (Eds.), The handbook of cognition and emotion
The differential processing of price in gains and losses: The effects (pp. 521–539). New York: Wiley.
of frame and need for cognition. Journal of Behavioral Decision Isen, A. M. (2002). Missing in action in the AIM: Positive affectÕs
Making, 13, 61–75. facilitation of cognitive flexibility, innovation, and problem solv-
Clore, G. L., Schwarz, N., & Conway, M. (1994). Affective causes and ing. Psychological Inquiry, 13, 57–65.
consequences of social information processing. In R. S. Wyer & T. Kahneman, D., & Lovallo, D. (1993). Timid choices and bold
K. Srull (Eds.), Handbook of social cognition (2nd ed.) forecasts: A cognitive perspective on risk taking. Management
(pp. 323–417). Hillsdale, NJ: Lawrence Erlbaum. Science, 39, 17–31.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences Kahneman, D., & Miller, D. T. (1986). Norm theory: Comparing
(2nd ed.). Hillsdale, NJ: Lawrence Erlbaum. reality to its alternatives. Pscyhological Review, 93, 136–153.
Cooper, A. C., Folta, T. B., & Woo, C. (1995). Entrepreneurial Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of
information search. Journal of Business Venturing, 10, 107–120. decisions under risk. Econometrica, 47, 263–291.
Crawford, M. T., & Skowronski, J. J. (1998). When motivated thought Kernan, M. C., & Lord, R. G. (1989). The effects of explicit goals and
leads to hightened bias: High need for cognition can enhance the specific feedback on escalation processes. Journal of Applied Social
impact of stereotypes on memory. Personality and Social Psychol- Psychology, 19, 1125–1143.
ogy Bulletin, 24, 1075–1088. Klein, H. J. (1989). An integrated control theory model of work
Dunegan, K. J. (1993). Framing, cognitive modes, and image theory: motivation. Academy of Management Review, 14, 150–172.
Toward an understanding of a glass half full. Journal of Applied Kuvaas, B. (2002). An exploration of two competing perspectives on
Psychology, 78, 491–503. informational contexts in top management strategic issue interpre-
Dunegan, K. J. (1994). Feedback sign and mindful vs. mindless tation. Journal of Management Studies, 39, 979–1003.
information processing. In C. Stubbart, J. R. Meindl, & J. F. Porac Kuvaas, B., & Kaufmann, G. (2004). Impact of mood, framing, and
(Eds.). Advances in managerial cognition and organizational infor- need for cognition on decision makersÕ recall and confidence.
mation processing (Vol. 5, pp. 315–336). Greenwich, Connecticut: Journal of Behavioral Decision Making, 17, 59–74.
Jai Press Inc. Kühberger, A. (1997). Theoretical conceptions of framing effects in
Dutton, J. E. (1993). The making of organizational opportunities: an risky decisions. In W. R. Crozier & O. Svenson (Eds.),
interpretive pathway to organizational change. In L. L. Cummings Decision making: Cognitive models and explanations. London:
& B. M. Staw (Eds.). Research in Organizational Behavior (Vol. 15, Routledge.
pp. 195–226). Greenwich, CT: JAI Press. Kühberger, A. (1998). The influence of framing on risky decisions: A
Fiedler, K. (2000). Toward an integrative account of affect and meta-analysis. Organizational Behavior and Human Decision Pro-
cognition phenomena using the BIAS computer algorithm. In J. cesses, 75, 23–55.
P. Forgas (Ed.), Feeling and thinking: The role of affect in social Langer, E. (1989a). Mindfulness. New York: Addison-Wesley.
cognition (pp. 223–252). Cambridge: Cambridge University Langer, E. (1989b). Minding matters: The consequences of mindless-
Press. ness-mindfulness. Advances in Experimental Social Psychology, 22,
Fiedler, K., & Bless, H. (2000). The formation of beliefs at the interface 137–173.
of affective and cognitive processes. In N. H. Frijda, A. S. R. LeBoeuf, R. A., & Shafir, E. (2003). Deep thoughts and shallow
Manstead, & S. Bem (Eds.), Emotions and belief: How feelings frames: On the susceptibility to framing effects. Journal of
influence thoughts (pp. 144–170). New York: Cambridge University Behavioral Decision Making, 16, 77–92.
Press. Levin, I. P., & Gaeth, G. J. (1988). Framing of attribute informationh
Forgas, J. P. (1998). Happy and mistaken? Mood effects on the before and after consuming the product. Journal of Consumer
fundamental attribution error. Journal of Personality and Social Resarch, 15, 374–378.
Psychology, 75, 318–331. Levin, I. P., Johnson, R. D., Russo, C. P., & Deldin, P. J. (1985).
Forgas, J. P. (2000). Affect and information processing strategies: An Framing effects in judgment tasks with varying amounts of
interactive relationship. In J. P. Forgas (Ed.), Feeling and thinking: information. Organizational Behavior and Human Decision Pro-
The role of affect in social cognition (pp. 253–280). Cambridge: cesses, 36, 362–377.
Cambridge University Press. Levin, I. P., Schneider, S. L., & Gaeth, G. J. (1998). All frames are not
Forgas, J. P. (2001). Affective intelligence: The role of affect in social created equal: A typology and critical analysis of framing effects.
thinking and behavior. In J. Ciarrochi, J. P. Forgas, & J. D. Mayer Organizational Behavior and Human Decision Processes, 76,
(Eds.), Emotional intelligence in everyday life: A scientific inquiry 149–188.
(pp. 45–79). Philadelphia, PA: Psychology Press. Levin, I. P., Huneke, M. E., & Jasper, J. D. (2000). Information
Forgas, J. P. (2002a). Feeling and doing: Affective influences on processing at successive stages of decision making: Need for
interpersonal behavior. Psychological Inquiry, 13, 1–28. cognition and inclusion-exclusion effects. Organizational Behavior
Forgas, J. P. (2002b). Toward understanding the role of affect in social and Human Decision Processes, 82, 171–193.
thinking and behavior. Psychological Inquiry, 13, 90–102. Levin, I. P., Gaeth, G. J., Schreiber, J., & Lauriola, M. (2002). A new
Harvey, N. (1994). Relations between confidence and skilled perfor- look at framing effects: Distribution of effect sizes, individual
mance. In G. Wright & R. Ayton (Eds.), Subjective probability differences, and independence of types of effects. Organizational
(pp. 321–351). New York: Wiley. Behavior and Human Decision Processes, 88, 411–429.
Hayward, M. L. A., & Hambrick, D. C. (1997). Explaining the Lord, R. G., & Maher, K. J. (1990). Alternative information-
premiums paid for large acquisitions: Evidence of CEO hubris. processing models and their implications for theory, research and
Administrative Science Quarterly, 42, 103–127. practice. Academy of Management Review, 15, 9–28.
Hodgkinson, G. P., Bown, N. J., Maule, A. J., Glaister, K. W., & Louis, M. R., & Sutton, R. I. (1991). Switching cognitive gears: From
Pearman, A. D. (1999). Breaking the frame: An analysis of strategic habits of mind to active thinking. Human Relations, 44, 55–76.
 B. Kuvaas, M. Selart / Organizational Behavior and Human Decision Processes 95 (2004) 198–207 207

Mahajan, J. (1992). The overconfidence effect in market management M. Sorrention (Eds.), Handbook of motivation and cognition (2nd
predictions. Journal of Marketing Research, 29, 261–265. ed.) (pp. 527–561). New York: Guilford Press.
Maheswaran, D., & Chaiken, S. (1991). Promoting systematic Sieck, W., & Yates, F. (1997). Exposition effects on decision making:
processing in low-motivation settings: Effect of incongruent infor- Choice and confidence in choice. Organizational Behavior and
mation on processing and judgment. Journal of Personality and Human Decision Processes, 70, 207–219.
Social Psychology, 61, 13–25. Simon, M., & Houghton, S. M. (2003). The relationship between
Mantel, S. P., & Kardes, F. R. (1999). The role of direction of overconfidence and the introduction of risky products: Evidence
comparison, attribute-based processing, and attitude-based pro- from a field study. Academy of Management Journal, 46, 139–149.
cessing in consumer preference. Journal of Consumer Research, 25, Smith, S. M., & Levin, I. P. (1996). Need for cognition and choice
335–352. framing effects. Journal of Behavioral Decision Making, 9, 283–290.
Martin, L. L. (2001). Mood as input: A configural view of mood Stanovich, K. E., & West, R. F. (1998). Evaluating principles of
effects. In L. L. Martin & G. L. Clore (Eds.), Theories of mood and rational indifference: Individual differences in framing and con-
cognition: A userÕs guidebook (pp. 135–157). Mahwah, NJ: Law- junction effects. Thinking and Reasoning, 4, 289–317.
rence Erlbaum Associates. Staw, B., Sandelands, L., & Dutton, J. (1981). Threat-rigidity effects in
Martin, L. L., & Stoner, P. (1996). Mood as input: What people think organizational behavior: A multilevel analysis. Administrative
about what they feel moods determines how they think. In L. L. Science Quarterly, 26, 501–524.
Martin & A. Tesser (Eds.), Striving and feeling: Interactions Staw, B. M., & Barsade, S. G. (1993). Affect and managerial
between goals, affect, and self-regulation (pp. 279–301). Hillsdale, performance: A test of the sadder-but wiser versus happier-and-
NJ: Erlbaum. smarter hypotheses. Administrative Science Quarterly, 38, 304–331.
Maule, A. J. (1994). Framing elaborations and their effects on choice Svenson, O. (1989). Eliciting and analysing verbal protocols in process
behavior: A comparison across problem isomorphs and subjects studies of judgement and decision making. In H. Montgomery &
with different levels of expertise. In J.-P. Caverni, M. Bar-Hillel, H. O. Svenson (Eds.), Process and structure in human decison making
Barron, & H. Jungermann (Eds.), Contributions to decision making- (pp. 65–83). Chichester: Wiley.
I (pp. 281–300). Amsterdam: North-Holland. Svenson, O., & Benson, L. I. (1993). Framing and time pressure in
Mitchell, T. R., & Beach, L. R. (1990). ...Do I love thee. Let me decision making. In O. Svenson & A. J. Maule (Eds.), Time
count... Toward an understanding of intuitive and automatic pressure and stress in human judgment and decision making
decision making. Organizational Behavior and Human Decision (pp. 133–144). New York: Plenum Press.
Processes, 47, 1–20. Takemura, K. (1982). Effect of decision time on framing of decision: A
Mittal, V., & Ross, W. T. J. (1998). The impact of positive and case of risky choice behavior. Psychologia, 35, 180–185.
negative affect and issue framing on issue interpretation and risk Takemura, K. (1994). Influence of elaboration on the framing of
taking. Organizational Behavior and Human Decision Processes, 76, decision. Journal of Psychology, 128, 33–39.
298–324. Taylor, S. E. (1991). Asymmetrical effects of positive and negative
Nunnally, J. C., & Bernstein, I. H. (1994). Psychometric theory (3rd events: The mobilization-minimalization hypothesis. Psychological
ed.). New York: McGraw-Hill. Bulletin, 110, 67–85.
Peeters, G., & Czapinski, J. (1990). Positive–negative asymmetry in Tiedens, L. Z., & Linton, S. (2001). Judgment under emotional
evaluations: The distinction between affective and informational certainty and uncertainty: The effects of specific emotions on
negativity effects. European Review of Social Psychology, 1, information processing. Journal of Personality and Social Psychol-
33–60. ogy, 81, 973–988.
Reger, R. K., & Palmer, T. B. (1996). Managerial categorization of Tversky, A., & Kahneman, D. (1974). Judgment under uncertainty:
competitors: Using old maps to navigate new environments. Heuristics and biases. Science, 185, 1124–1131.
Organization Science, 7, 22–39. Tversky, A., & Kahneman, D. (1981). The framing of decisions and the
Rosnow, R. L., & Rosenthal, R. (1996). Computing contrasts, effect psychology of choice. Science, 211, 453–458.
sizes, and counternulls on other peopleÕs published data: General Walsh, J. P. (1995). Managerial and organizational cognition: Notes
procedures for research consumers. Psychological Methods, 1, from a trip down memory lane. Organization Science, 6, 280–321.
331–340. Wegener, D. T., Petty, R. E., & Klein, D. J. (1994). Effects of mood on
Rothman, A. J., & Salovey, P. (1997). Shaping perceptions to motivate high elaboration attitude change: The mediating role of likelihood
healthy behavior: The role of message framing. Psychological judgments. European Journal of Social Psychology, 24, 25–43.
Bulletin, 121, 3–19. Wofford, J. C., & Goodwin, V. L. (1990). Effects of feedback on
Schneider, S. L. (1992). Framing and conflict: Aspiration level cognitive processing and choice of decision style. Journal of Applied
contingency, the status quo, and current theories of risky choice. Psychology, 75, 603–612.
Journal of Experimental Psychology: Learning, Memory, and Wright, G., & Goodwin, P. (2002). Eliminating a framing bias by using
Cognition, 18, 1040–1057. simple instructions to think harder and respondents with manage-
Schwarz, N. (1990). Feelings as information: Informational and rial experience: comment on breaking the frame. Strategic Man-
motivational functions of affective states. In E. T. Higgins & R. agement Journal, 23, 1059–1067.