Close
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Countering the effects of selective exposure
(USC Thesis Other)
Countering the effects of selective exposure
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
COUNTERING THE EFFECTS OF SELECTIVE EXPOSURE
by
Chadwick J. Snow
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(PSYCHOLOGY)
December 2005
Copyright 2005 Chadwick J. Snow
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UMI Number: 3220155
INFORMATION TO USERS
The quality of this reproduction is dependent upon the quality of the copy
submitted. Broken or indistinct print, colored or poor quality illustrations and
photographs, print bleed-through, substandard margins, and improper
alignment can adversely affect reproduction.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if unauthorized
copyright material had to be removed, a note will indicate the deletion.
®
UMI
UMI Microform 3220155
Copyright 2006 by ProQuest Information and Learning Company.
All rights reserved. This microform edition is protected against
unauthorized copying under Title 17, United States Code.
ProQuest Information and Learning Company
300 North Zeeb Road
P.O. Box 1346
Ann Arbor, Ml 48106-1346
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ii
TABLE OF CONTENTS
LIST OF TABLES iii
LIST OF FIGURES iv-vi
ABSTRACT vii
CHAPTER I: INTRODUCTION 1
The Value of Inconsistency 1
The Paradigms of Cognitive Dissonance 5
Countering the Effects of Cognitive Dissonance 6
Key Components of the Present Research 9
The Present Research 13
CHAPTER II: STUDY 1 15
Method 16
Results 20
Discussion 27
CHAPTER III: STUDY 2 31
Method 31
Results 33
Discussion 38
CHAPTER IV: STUDY 3 41
Method 41
Results 43
Discussion 56
CHAPTER V: STUDY 4 59
Method 62
Results 65
Discussion 79
CHAPTER VI: GENERAL DISCUSSION 82
REFERENCES 92
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table
Table
Table
Table
LIST OF TABLES
3.1. Reliability analyses for both pre-mood and post
mood measures of emotion.
4.1. Reliability analyses for personality measures.
4.2. Reliability analyses for both pre-mood and post
mood measures of emotion.
4.3. Main effects of frames on information preference
ratings.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
iv
Figure 1.1.
Figure 1.2.
Figure 1.3a.
Figure 1,3b.
Figure 2.1.
Figure 2.2.
Figure 2.3a.
Figure 2.3b.
Figure 3.1.
Figure 3.2.
Figure 3.3a.
LIST OF FIGURES
Overall distribution for choice of information based
on frames. 22
Main effects of framing manipulations on non
chosen alternative information choice. 24
Information choice distributions for framing of the
chosen alternative information conditions versus
the no framing control condition. 25
Information choice distributions for framing of the
non-chosen alternative information conditions
versus the no framing control condition. 26
Overall distribution for choice of information based
on frames. 34
Main effects of framing manipulations on non
chosen alternative information choice. 36
Information choice distributions for framing of the
chosen alternative information conditions versus
the no framing control condition. 37
Information choice distributions for framing of the
non-chosen alternative information conditions
versus the no framing control condition. 38
Overall distribution for choice of information based
on frames. 45
Main effects of framing manipulations on non-
chosen alternative information choice. 46
Information choice distributions for framing of the
chosen alternative information conditions versus
the no framing control condition. 47
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
V
Figure 3.3b.
Figure 3.4.
Figure 3.5.
Figure 3.6.
Figure 4.1.
Figure 4.2a.
Figure 4.2b.
Figure 4.3.
Figure 4.4a.
Figure 4.4b.
Figure 4.5.
Information choice distributions for framing of the
non-chosen alternative information conditions
versus the no framing control condition. 48
Interaction between framing manipulations for
chosen alternative information choosers on levels
of post-information choice happiness. 50
Interaction between information choosers for
framing of the non-chosen alternative information
on levels of post-information choice happiness. 53
Interaction between framing manipulations for
chosen alternative information choosers on levels
of post-information choice calmness. 55
Overall distribution for choice of information based
on frames. 66
Overall distribution for preference rating of
information based on frames for consistency and
inconsistency preferences only. 68
Overall distribution for preference rating of
information based on frames. 69
Main effects of framing manipulations on non-
chosen alternative information choice. 71
Information choice distributions for framing of the
chosen alternative information conditions versus
the no framing control condition. 72
Information choice distributions for framing of the
non-chosen alternative information conditions
versus the no framing control condition. 73
Interaction between need for closure personality
type and framing of the non-chosen alternative
information for magnitude of preference for the
chosen alternative information. 75
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 4.6. Number of participants choosing the non-chosen
alternative information based on high versus low
need for closure and positive versus negative
framing of the non-chosen alternative information.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ABSTRACT
In social psychology, research on the phenomenon of cognitive dissonance
has shown that the experience of inconsistency is perceived negatively and is
avoided. However, other research has demonstrated the value of experiencing
inconsistency. The specific concern addressed in the present research was how to
increase the likelihood of choosing potentially valuable information that is
inconsistent with a prior decision given the general tendency to choose consistent
information due to cognitive dissonance. All 4 studies found that the inclusion of
framing manipulations successfully manipulated information choice. Positively
framing information about a non-chosen alternative (inconsistent information) or
negatively framing information about a chosen alternative (consistent information)
increased non-chosen alternative information choice relative to oppositely framed
conditions and a no frame control condition. Study 3 indicated a relationship
between the framing manipulations and changes in affect. Study 4 determined that
people who are low in need for closure are more receptive to framing
manipulations of non-chosen alternative information than people who are high in
need for closure. The possible motivational processes activated by the framing
manipulations are briefly considered along with suggestions for future research on
this topic.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1
CHAPTER I: INTRODUCTION
Anyone relatively familiar with social psychological literature has likely
heard of the theory of cognitive dissonance. Festinger explained that cognitive
dissonance occurs when an individual encounters an inconsistent relationship
among two or more cognitions or actions (Aronson, 1968). An abundance of
research in the field has discovered the human tendency to employ a variety of
cognitive or behavioral strategies with the intent of avoiding or reducing
dissonance. Research also shows that dissonance causes people to experience
negative affect due to the fact that it is aversive (Harmon-Jones & Mills, 1999).
The goal of the present research was to determine if under certain circumstances
people will prefer information that is inconsistent with a prior choice and not
experience the negative affect typically associated with the experience of cognitive
dissonance as a result of having chosen that information.
The Value of Inconsistency
The notion that the experience of inconsistency can be something attractive
and experienced positively is not new to many areas of psychological research.
Evidence for the human tendency to seek out novel experiences can be found as
early as infancy. The habituation literature has demonstrated a preference for
novelty among infants only 3 months old. Infants quickly lose interest in the
repetitious display of the same image (consistency) (Jeffrey & Cohen, 1971;
Wetherford & Cohen, 1973). However, they will immediately reorient themselves
to the presentation of a novel image (inconsistency) (Pancratz & Cohen, 1970).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2
What’s interesting about this research is that it not only demonstrates an attraction
to inconsistency as evidenced by the infant tendency to reorient themselves visually
to a novel presentation, but additionally there is a clear disinterest in consistency as
evidenced by the infant tendency to look away from a repetitious display.
These results support the “two-stage process” of infant attention presented
by Hunt (1963). When infants initially display an interest in objects, they are first
captivated by the familiar (consistency) as they experiment with their ability to
recognize aspects of their environment (8 weeks). Once this ability has developed,
infants become interested in objects that are novel (inconsistency) (10 and 12
weeks).
Research has also shown that children will prefer to play with toys that
they have never seen before than toys that are familiar. In fact, if given the option
of playing with 2 novel toys that are identical to each other or 2 familiar but non
identical toys, they will choose the former. This demonstrates the powerful
attraction children have to inconsistency. While both choices of toys possess
inconsistent properties (the identical toys being new and thus inconsistent from
familiar experiences and the non-identical toys being inconsistent as variation
within the familiar), the option of interacting with something unfamiliar is the most
appealing (Wentworth and Witryol, 1983).
Personality psychology has repeatedly demonstrated that some people have
a natural tendency towards highly creative thought and behavior (Amabile, 1983;
Stein, 1974; MacKinnon, 1962). However, social psychology has shown that all
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3
humans are capable of and experience moments of creativity (Amabile, 1983;
Wallach, 1970; Torrance, 1965). The fact that it is possible for everyone to be
creative, yet not all the time, suggests that there must be certain requirements that
promote creativity. Within the creativity literature we find compelling evidence
that the ability to create is dependent upon encounters with inconsistency. One
must be both willing and able to interact with inconsistency in order to ultimately
produce something that is new and therefore inconsistent with the norm. A study
by Maddi, Charlens, Maddi, and Smith (1962) found that when people were
required to attend to monotonous stimulation, they were significantly less able to
subsequently produce creative stories based upon pictures of people engaged in
various activities than were people who were not required to attend to monotonous
stimulation. This research demonstrates the detrimental impact unyielding
consistent stimulation can have on human creativity. When left to our own devices
or exposed to highly novel stimulation, we are far better at engaging the creative
process.
Perhaps one of the most effective demonstrations of the role of novelty and
inconsistency in creativity is in the examination of art. Drawing from Berlyne’s
(1960) and Zuckerman’s (1994) explanations for the appeal of art, whether it be
visual, musical, or humorous, we find that the presence of inconsistency is an
essential ingredient. Put succinctly, a form of art must contain a level of
inconsistency appropriate in magnitude to promote optimal arousal within its
perceiver. However, the inconsistency must also be resolvable allowing for a
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4
coherent impression of the art form. Without resolution, the inconsistency becomes
aversive, thereby making the art form unappealing.
When considering visual art, such as a painting, Berlyne (1960) notes that
interest in an art form is dependent upon visual conflict or inconsistency among its
elements. He also comments that the inconsistency must be resolved and
synthesized into a coherent whole in order to be appealing.
Berlyne (1960) also explains that there exists in music the need for both
inconsistency and relief from inconsistency. At the most fundamental level, the
notes in music are inconsistent from each other. However, the composer is
required to arrange them in such a way that a certain level of coherence can be
perceived as the various notes are played. From a more macro level of analysis,
repetition between sections of a composition keep the listener interested, but only
in so far as there are also elements suggesting upcoming and uncertain changes.
Once experienced, those changes must be inconsistent enough from what was being
played beforehand to be perceived as appealing, but limited to the extent to which
the listener is capable of integrating the changes within the rest of the composition.
Zuckerman (1994) describes the role of inconsistency in humor. He
explains that there exists inconsistency between the elements of the joke, but that
inconsistency is relieved with an unexpected twist causing the inconsistent
elements to gel in a congruous way. Without inconsistency, there is no joke.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5
The Paradigms of Cognitive Dissonance
As described above, there is a substantial body of research supporting the
positive aspects of inconsistency (infant learning, child behavior, creativity, art,
music, humor, etc.). Therefore, within social psychology, it is reasonable to
speculate that the experience of inconsistency could sometimes be potentially
rewarding in situations that concern cognitive dissonance. However, several
paradigms have flourished within the dissonance literature demonstrating the
aversiveness of inconsistency. In the “free-choice” paradigm, people must choose
between 2 closely ranked alternatives and then re-evaluate the alternatives after
choosing 1 of them. People reduce dissonance by increasing their rated preferences
for the chosen alternative and decreasing their rated preferences for the non-
chosen-altemative (Harmon-Jones & Mills, 1999). In the “belief-disconfirmation”
paradigm, people are given information that is inconsistent with their beliefs.
People reduce dissonance by changing their beliefs to be consistent with the new
information or by rejecting the new information (Harmon-Jones & Mills, 1999). In
the “effort-justification” paradigm, people are asked to participate in an unpleasant
activity in order to achieve some desired goal. By bolstering the perceived quality
of the goal, one increases justification for engaging in the unpleasant activity, thus
reducing dissonance (Harmon-Jones & Mills, 1999). In the “induced-compliance”
paradigm, people are given some incentive to act in a way that contradicts their
attitudes and beliefs, such as being offered some reward or being threatened with
punishment, and therefore they feel justified in their aberrant behavior and
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6
experience little dissonance (Harmon-Jones & Mills, 1999). In the “selective
exposure” paradigm, people reduce dissonance by choosing information that is
supportive (consistent) with a prior choice as opposed to information that does not
support (inconsistent) a prior choice (Mills, 1968).
One commonality shared among these paradigms is that the consequences
of not reducing dissonance in these particular situations are threatening to the self-
concept (Greenwald & Ronis, 1978; Abelson, 1983). Under these paradigms, not
reducing or avoiding dissonance would lead to feelings of stupidity (free-choice
paradigm), loss of belief system (belief-disconfirmation paradigm), wasted effort
(effort-justification paradigm), the perception of oneself as an untrustworthy
individual (induced-compliance paradigm), or the belief that one is a poor decision
maker (selective exposure paradigm). Whether intentional or not, it is clear from
any review of the literature that dissonance studies in social psychology have put
people in uncomfortable situations in which an inconsistent thought or action
would be threatening or without value (Harmon-Jones & Mills, 1999).
Countering the Effects of Cognitive Dissonance
Clearly, the effects of cognitive dissonance are quite powerful. As alluded
to earlier, the concern that arises from this is that there are inevitably times in
which adopting a thought or taking an action that is inconsistent with a prior
thought or action may in some way be beneficial. Due to the effects of cognitive
dissonance, people in these situations would be prevented from acting in their best
interests.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
7
For example, suppose someone just bought a computer and 1 month later is
given the opportunity to read information that supports her choice of computer or
information that suggests that an alternative choice would have been better.
According to Festinger’s 2n d hypothesis of dissonance: “When dissonance is
present, in addition to trying to reduce it, the person will actively avoid situations
and information which would likely increase the dissonance.” (Festinger, 1957, p.
3). Therefore, this person would choose to read supportive information about the
computer they purchased because such information is consistent with her purchase.
Reading supportive information about the other computer would increase
dissonance because it would suggest that this person may have not made a wise
decision. This example of the “selective exposure” paradigm is similar to the study
by Ehrlich, Guttman, Schonbach, and Mills (1957), which found that new car
owners preferred to read advertisements about their newly purchased cars than ads
for other cars. Regarding the prevalence of this finding in the “selective exposure”
literature, Mills (1968) states, “To summarize the current status of the hypothesis
that people prefer supporting information to discrepant information, there are a
number of studies that corroborate the hypothesis and there are none which provide
substantial evidence against the hypothesis.” (p. 744).
However, if reading the information about the other computer model would
have informed this person that she would have been happier with this other model,
then such information could potentially benefit her decisions regarding future
computer purchases. With this understanding, this person might choose to read
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
8
information about the other computer model and be happy with that choice of
information.
One can imagine several other examples of times when information that is
inconsistent with a choice could be quite valuable. Therefore, if inconsistent
information is presented as potentially having some value, such as serving as an
opportunity to learn or to facilitate becoming a better decision maker, an individual
should be more likely to choose that information than had that information not been
presented as potentially valuable. As a consequence, the individual should not find
that information to be aversive and instead experience positive affect from knowing
that he or she has chosen valuable information even if that information is
inconsistent with a prior choice. However, given the strength of the human desire
to prefer information that is consistent, as opposed to inconsistent, with a prior
choice, as demonstrated in countless cognitive dissonance studies looking at
selective exposure to information (Mills, 1968), is it possible to alter the strong
tendency to prefer consistent information? The primary purpose of the present
research was to design a classic cognitive dissonance study within the “free-choice”
and “selective exposure” paradigms and manipulate the potential value of choosing
information that is inconsistent versus consistent with a prior choice so as to
increase the likelihood of choosing information that is inconsistent with a prior
choice.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
9
Key Components of the Present Research
The Free-Choice Paradigm. An important component of the present
research is the “free-choice” paradigm in cognitive dissonance. Under this
paradigm, when a person has made an irreversible decision, any negative
characteristics about the choice made and any positive characteristics of any
alternative choices conflict with the decision. Such characteristics are inconsistent
with the belief that one has decided correctly and so they are experienced as
aversive. In contrast, both the positive elements of the choice made and the
negative elements of any alternative choices are in support of the decision, thus
making these pieces of information attractive. The “free-choice” paradigm in the
cognitive dissonance literature has repeatedly demonstrated the various methods
people implement to reduce the dissonance associated with having made a difficult
decision between alternatives. People can reduce dissonance by (1) adding positive
characteristics and/or eliminating negative characteristics of the choice made or (2)
adding negative characteristics and/or eliminating positive characteristics of any
existing alternative choices. This method of dissonance reduction has been termed
the “spreading of alternatives” (Harmon-Jones & Mills, 1999).
The original study that examined the free-choice paradigm asked female
participants to rate their preferences for different household products and to then
decide between two of the products. After making a decision when the two
products were both highly desired (i.e. difficult decision), participants reevaluated
the chosen product to be even more preferred and the alternative product to be
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
10
significantly less preferred, thus illustrating the spreading of alternatives (Brehm,
1956).
The Selective Exposure Paradigm. Another major component of the present
research is the “selective exposure” paradigm. As described earlier, under this
paradigm, people prefer information that is consistent with a prior choice as
opposed to information that is inconsistent with a prior choice (Mills, 1968). A
specific line of research in “selective exposure” that is of particular interest to the
present research is the work of Dieter Frey and his colleagues (Frey, 1982; Frey &
Rosch, 1984, Frey, 1986). Their research has shown that when inconsistent
information has the potential of improving a participant’s immediate situation, such
information is actively sought. In one study, Frey (1982) had participants play a
card game. Each participant decided whether he or she wanted to be in position A
or B. The experimenter assumed the position not taken by the participant. For
each trial in the game, the participant had to decide how much money they wanted
to bet (5-20 cents). After 12 trials, the participant was asked if he or she would like
to switch positions. Before deciding, the participant was given the opportunity to
read 3 essays supporting either position A or B from a selection of 12 possible
essays (6 favoring their chosen position and 6 favoring the alternative position).
Participants that didn’t win or lose much money typically chose consistent
information (essays supporting their decision). With little gain or loss, Frey argues
that there was little evidence to suggest that their current position was any worse
than the alternative position. Inconsistent information (essays opposing their
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
11
decision) was clearly preferred when participants lost a substantial amount of
money. Frey suggests that if participants lost a significant amount of money
compared to the experimenter, they doubted the quality of their chosen position.
Participants therefore sought information inconsistent from their original position
to support their desire to change positions. When participants won a great deal of
money, they had no preference for consistent or inconsistent information. Frey
points out that if participants won a lot of money relative to the experimenter, then
their winnings supported their chosen position making both additional consistent
and inconsistent information equally irrelevant (Frey, 1982).
Much of Frey’s work has demonstrated this preference for inconsistent
information when participants’ decisions are made reversible and when there is
good evidence that the original decision was a poor one. It is sensible to speculate
that classic cognitive dissonance researchers may not consider these studies a tme
test of dissonance since the theories driving traditional dissonance research have
typically, if not exclusively, implemented irreversible decisions and situations.
Therefore, a major goal of the present research was to make inconsistent
information seem potentially valuable to people despite not being able to reverse
their prior choice. One way to do this (and the way it was done in the present
research) would be to frame inconsistent information such that the potential value
of choosing inconsistent information would be beneficial to people beyond their
participation in the current research.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
12
The Effectiveness of Frames. Several researchers in psychology have
studied the effectiveness of frames in making something appealing. For example,
Tversky and Kahneman (1981) have demonstrated that people’s preferences for
outcomes can shift when one outcome is framed positively while the other is
framed negatively. In particular, shifts will occur when outcomes are framed as
gains or losses, relative to a neutral reference outcome. As another example,
Higgins, Shah, and Friedman’s (1997) research on promotion focus versus
prevention posits that when an individual evaluates a stimulus and any interactions
with that stimulus as potentially positive (i.e. What positive things can I gain or
lose from this?) he is considered to be promotion focused. Stimuli with which
people are promotion focused tend to elicit “approach” oriented behavior. In
contrast, when an individual evaluates a stimulus and any interactions with that
stimulus as potentially negative (i.e. What negative things will I experience or
avoid from this?) she is considered to be prevention focused. Stimuli with which
people are prevention focused tend to elicit “avoidant” oriented behavior (Higgins,
Shah, & Friedman, 1997; Higgins, 1997).
Considering the effectiveness of frames, the logical result of framing an
encounter or potential encounter with information that is inconsistent with a prior
choice as potentially valuable should lead to preferences for that information. This
should be especially so when, in addition to framing inconsistent information
positively, the option of choosing consistent information is framed negatively.
Similarly, framing inconsistent information negatively should promote an
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
13
avoidance of information inconsistent with a prior choice. This should be
especially so when, in addition to framing inconsistent information negatively, the
option of choosing consistent information is framed positively.
According to Higgins, Shah, & Friedman (1997), an individual who
successfully acts in accordance with how something is framed (promotion focus
versus prevention focus) experiences positive affect such as “happiness”,
“satisfaction”, “calmness”, and “relaxation”. Therefore, people who are made
aware of the potential value of choosing information that is inconsistent with a
prior choice (positive framing) and choose inconsistent information should be
happier and more relaxed about that choice than people who choose inconsistent
information, but receive a negative framing of that information. Additionally,
people who receive a positive framing of information that is inconsistent with a
prior choice and choose to read that information should be just as happy and
relaxed about their decision as people who receive a positive framing of
information that is consistent with a prior choice and choose consistent
information.
The Present Research
The main hypothesis was that people who are led to believe that there is
value in choosing information that is inconsistent with a prior irreversible choice,
or a lack of value in choosing information that is consistent with a prior irreversible
choice, will choose to read inconsistent information more often than people who
are led to believe the opposite or who are not led to believe that either piece of
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
14
information is more useful than the other, such as in a classic cognitive dissonance
situation. Furthermore, it was hypothesized that through these manipulations,
people who choose inconsistent information when it is presented as potentially
valuable will experience more positive affect than people who choose inconsistent
information when it is presented as not potentially valuable.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
15
CHAPTER II: STUDY 1
Study 1 was the first in a series of 4 studies designed to test the hypothesis
that people may be more receptive to information that is inconsistent with a chosen
alternative if the value in choosing to read that information is framed positively,
rather than negatively. Specifically, the prediction was that people who received a
positive framing of the information that supports the non-chosen alternative
(inconsistent information) would be more likely to choose to read that information,
than people who received a negative framing of that information. Additionally, it
was predicted that if people received a negative framing of the information that
supports the chosen alternative (consistent information), then they too would be
more likely to choose to read inconsistent information than people who received a
positive framing of the consistent information.
It was also predicted that people who received a positive framing of the
non-chosen alternative information and/or a negative framing of the chosen
alternative information, would be more likely to choose to read the non-chosen
alternative information than people who received no framing manipulations at all.
The condition in which people did not receive framing manipulations was designed
to capture the dissonance effect typical to the “free-choice” and “selective
exposure” paradigmatic approaches to examining people’s reactions to cognitive
dissonance.
Finally, it was predicted that people’s emotional reactions to choosing to
read the chosen alternative information or the non-chosen alternative information
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
16
would be affected by the framing manipulations. For example, people who chose
to read the non-chosen alternative information were expected to be happier about
that choice if they received a positive framing of the non-chosen alternative
information than if they had received a negative framing of the non-chosen
alternative information. Similarly, people who chose to read the chosen alternative
information were expected to be less happy about that choice if they received a
positive framing of the non-chosen alternative information than if they had received
a negative framing of the non-chosen alternative information.
Method
Participants. This study consisted of 346 participants (73 male, 273 female)
ranging in age from 18 to 75 with a mean age of 36. Participants were recruited
online through either an advertisement placed on the internet or email.
Participation occurred online.
Materials and Procedure. After consenting to take part in the study,
participants filled out a pretest emotion measure to assess baseline emotions. The
list consisted of 15 randomly ordered items, with 3 items measuring each of 5
different emotions (happiness, calmness, sadness, fear, anger). The scale ranged
from “Very Slightly or Not At All” to “Extremely”. This list was constructed from
the PANAS-X Affect schedule (Watson & Clark, 1994). Another list capturing the
same 5 emotions, but using different terms from the PANAS-X, was constructed
and administered at the end of the study. Subsets of the PANAS-X were used for
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
17
both lists to reduce the amount of time participants needed to devote to filling out
these measures.
On the next webpage, participants were told that they would be taking part
in a role-playing game in which they had to make believe that they would be
venturing out into a lush forest all alone. They were told that the goal was to
survive for 3 days and ultimately reach an undisclosed location. It was further
explained that the only resources available to them would be 2 items chosen from a
list of 8 possible items and that they must decide which items they feel are the 2
most essential items to have, given the type of environment that they would be
entering. Participants then rated on a scale from l(not at all) to 11 (very much) just
how much they would like to have each of the 8 possible items available for their
survival trek (pocket knife, portable radio, nine nutrition bars, flashlight canteen
with water, matches, insect repellent, sleeping bag). The purpose of these initial
ratings was simply to get the participants thinking about how they felt about the 8
items, so that they would subsequently be better able to choose their 2 most highly
preferred items.
Following their assessment of the 8 items, the next webpage asked
participants to choose their 2 most highly preferred items. On the webpage that
followed, they were then asked to rate once again on the same l(not at all) to
11 (very much) scale how much they would like to have their 2 most highly
preferred items (these ratings served as the pre-choice spreading of alternatives
measure). The next webpage brought participants to “The Survival Shop” where
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
18
they were told that this was the shop that sold the 2 items that they wished to take
with them on their adventure. It was further explained that unfortunately the
shopkeeper could only spare 1 of their 2 preferred items, but was willing to give
them whichever one they wanted the most. Participants were then asked to choose
which of the 2 items they wanted to take with them on their adventure with the
understanding that once they had made their decision, it could not later be changed.
After choosing which item to take on their adventure, participants were
asked to rate once again on the same l(not at all) to 11 (very much) scale how much
they would like to have their 2 most highly preferred items (these ratings served as
the post-choice spreading of alternatives measure). Participants were also asked to
rate on the same l(not at all) to 1 l(very much) scale how confident they were that
they had chosen the best possible item to take with them on their adventure.
On the next webpage, participants were told that they would now have the
opportunity to read information that describes the value of either the item they
chose to take with them on their adventure or the item that they did not choose to
take with them on their adventure. Participants were randomly assigned to 1 of 5
possible framing conditions. Thus, a 2(the chosen alternative information: positive,
negative) X 2(the non-chosen alternative information: positive, negative) between
subjects design was implemented, with the fifth condition being the no framing
control condition.
For the chosen alternative information, the positive frame was, “by
choosing to read information about the item you chose, you will likely feel more
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
19
confident about your choice”. This is an example of promotion focus framing in
that it instigates “approach” oriented behavior (Higgins, 1997) because people
would want to choose this information to increase their level of confidence
regarding having chosen the best item. The negative frame for the chosen
alternative information was, “by choosing to read information about the item you
chose, you are not likely to learn anything valuable about the item other than what
you already know”. This is an example of prevention focus framing in that it
instigates “avoidant” oriented behavior (Higgins, 1997) because people would want
to avoid this information since choosing it would serve no value.
For the non-chosen alternative information, the positive frame was, “by
choosing to read information about the item you did not choose, you will likely
learn more about that item and possibly become a better decision maker”. This is
an example of promotion focus framing in that it instigates “approach” oriented
behavior (Higgins, 1997) because people would want to choose this information to
learn more about that item and possibly become a better decision maker. The
negative frame for the non-chosen alternative information was, “by choosing to
read information about the item you did not choose, you will likely feel less
confident about your choice”. This is an example of prevention focus framing in
that it instigates “avoidant” oriented behavior (Higgins, 1997) because people
would want to avoid this information since choosing it would decrease their level
of confidence regarding having chosen the best item. These frames were designed
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
20
to be effective, but to ultimately be relatively simple and mild in terms of
magnitude of impact.
After choosing which information they wanted to read, participants were
taken to a webpage that described the value of that item. After reading the
information, participants clicked through to a webpage that asked them once again
to rate on the same l(not at all) to 11 (very much) scale how confident they were
that they had chosen the best possible item to take with them on their adventure.
Following this rating participants were asked to fill out a few other questions
designed for exploratory analysis purposes (i.e. “Using the provided scale, please
indicate how much you liked the information you read.”, “Using the provided scale,
please indicate how much you believe that the information you read is useful.”,
etc.).
On the final webpage of the study, participants completed the second
measure of emotions that was ultimately excluded from analysis. Lastly,
participants clicked through to a webpage assessing demographical information and
were then debriefed.
Results
Overall Distributions of Information Choosers based on Framing
Conditions. The number of participants choosing to read the non-chosen
alternative information was highest (51%) for participants who received a negative
framing of the chosen alternative information and a positive framing of the non-
chosen alternative information. Participants who received a positive framing of the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
21
chosen alternative information and a negative framing of the non-chosen alternative
information or who were in the no framing control condition, were least likely to
choose to read the non-chosen alternative information, 13% and 6% respectively.
The distribution of information choice in the control condition, 94% of participants
choosing to read the chosen alternative information and 6% of participants
choosing to read the non-chosen alternative information, supports the classic
dissonance effect. For participants that received positive frames for both pieces of
information or negative frames for both pieces of information, 31% in both
conditions chose to read the non-chosen alternative information (see Figure 1.1).
Spreading of Alternatives. An important element of the “free-choice”
paradigm in cognitive dissonance is the “spreading of alternatives”. The paradigm
predicts that the differences in rated likings for 2 alternatives should be greater after
choosing 1 alternative over the other than prior to that choice. The rationale behind
this prediction is that people want their opinions to be consistent with their choice.
So, people’s ratings for the chosen item should increase while ratings for the non-
chosen item should decrease. To measure this phenomenon, for each of the two
preferred items, we first subtracted the item’s pre-choice value from its post-choice
value. Results from these difference scores indicated that the spreading of
alternatives occurred between pre-decisional to post-decisional chosen item
difference scores and pre-decisional to post-decisional non-chosen item difference
scores, t(311) = 8.188, p < .001 (Ms = 0.01 vs. -0.36, respectively). The magnitude
of this difference was then calculated by subtracting the latter from the former.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
22
This was done so that additional analyses could be conducted to determine if the
amount of spreading might predict choice of information. A logistic regression
analysis indicated that the amount of spreading did not predict information choice,
(/3= .10, p = .517).
Figure 1.1. Overall distribution for choice of information based on frames.
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
pos CA Info pos CA Info neg CA Info neg CA Info no frame
/ pos NCA I neg NCA / pos NCA / neg NCA control
Info Info Info Info
Note. CA Info = framing of the chosen alternative information
NCA Info = framing of the non-chosen alternative information
It was predicted that the amount of spreading could possibly interact with
the framing manipulations, thus affecting information choice. For example, people
who have a large amount of post-decisional spreading, may be less affected by the
frames because their clear preference for one item over the other might negate any
■ Chosen Alternative
□ Non-Chosen Alternative
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
23
interest in considering information about the alternative item. However, after
centering the spreading of alternatives scores, logistic regression analyses
determined that the magnitude of the spreading of alternatives modeled with the
framing manipulations also did not predict information choice, (/3 = .07, p = .686),
nor were there any interactions. Even when examining only those participants in
the no framing control condition, the magnitude of the spreading of alternatives did
not predict information choice, (/3 = -.06, p = .944).
Effects of Frames on Information Choice. Analyses were then conducted to
see if just the framing manipulations alone would predict information choice.
Logistic regression analyses revealed main effects for both the framing of the
chosen alternative information, (/3 = .93, p < .001), and the framing of the non-
chosen alternative information, (/3 = -.91 >E< .001), on choice of information. The
number of participants choosing to read the non-chosen alternative information was
highest when they received a negative framing, as opposed to a positive framing, of
the chosen alternative information or when they received a positive framing, as
opposed to a negative framing, of the non-chosen alternative information (see
Figure 1.2). No interactions were detected.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
24
Figure 1.2. Main effects of framing manipulations on non-chosen alternative
information choice.
50%
40%
30%
Positive
Negative
20%
10%
CA Info NCA Info
Note. CA Info = framing of the chosen alternative information
NCA Info = framing of the non-chosen alternative information
To better understand the effects of the framing manipulations relative to the
classic cognitive dissonance phenomenon, as represented in the no framing control
condition, a series of chi-square tests were performed on information choice (see
Figures 1.3a & 1.3b). Overall, the presence of framing manipulations led to more
the non-chosen alternative information choice than in the control condition:
positive framing of the chosen alternative information, x2(l, n = 202) = 8.84, p <
.01, negative framing of the chosen alternative information, *2(1, n = 214) = 28.19,
P < .001, negative framing of the non-chosen alternative information, % 2(1, n =
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
25
208) = 9.31, 2 < .01, positive framing of the non-chosen alternative information,
*2(1, n = 214) = 28.38, p < .001.
Figure 1.3a. Information choice distributions for framing of the chosen
alternative information conditions versus the no framing control condition.
1 00 ^ ~
I Chosen Alternative
Non-Chosen Alternative
No Frame CA Info Frame CA Info Frame
Control Positive Negative
Note. CA Info = the chosen alternative information
Confidence Levels for Information Choice. An analysis of variance
revealed an effect of reading information on levels of confidence regarding having
chosen the best possible item, F(l, 343) = 100.60, p < .001, such that participants
who chose to read the chosen alternative information were significantly more
confident that they chose the best item to take with them into the wilderness after
having read that information than were participants who chose to read the non
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
26
chosen alternative information (Ms = 0.97 vs. -1.34, respectively). It was decided
that nothing conclusive could be determined regarding change in confidence levels
because there’s no way to tell if any change is due to the act of choosing to read the
non-chosen alternative information versus the chosen alternative information or the
effects of reading the information itself. It is the case, however, that when looking
at consistency information choosers and inconsistency information choosers
separately, no significant differences were observed across the different framing
conditions for changes in confidence levels, (Fs < 1).
Figure 1.3b. Information choice distributions for framing of the non-chosen
alternative information conditions versus the no framing control condition.
100% n
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
No Frame NCA Info NCA Info
Control Frame Positive Frame
Negative
Note. NCA Info = the non-chosen alternative information
■ Chosen Alternative
a Non-Chosen Alternative
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
27
Effects of Framing Manipulations on Emotions. No analyses were
conducted on any of the emotion measures because the time 2 emotion measure
was placed at the very end of the study making it impossible to determine if any
changes in emotion observed in time 2 were due to the act of choosing the chosen
alternative information vs. the non-chosen alternative information based on the
framing manipulations or due to the experience of having actually read the
information. Both of these issues were remedied and emotion states were measured
again in Study 3.
Discussion
Study 1 yielded some rather interesting findings. First and foremost, the
distribution of participants choosing the chosen alternative information over the
non-chosen alternative information in the control condition, suggests that the goal
of designing a study that effectively parallels a classic cognitive dissonance study
in the “selective exposure” paradigm was achieved. The inclusion of the framing
manipulations successfully altered people’s preferences regarding the chosen
alternative information and the non-chosen alternative information as compared to
the control condition. The main effects for the framing manipulations account for
the distribution of information choice across the 4 framing conditions. Specifically,
the main effects of framing the pieces of information positively versus negatively,
demonstrate that by providing people with the means to consider the possible value
of choosing to read information that is about a non-chosen alternative or to consider
the possible lack of value in choosing to read information that is about a chosen
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
28
alternative the likelihood is increased that they will choose to information that is
about a non-chosen alternative. So, among the overall distributions across framing
conditions, the highest level of preference for inconsistency should have occurred
in the condition in which people received a positive framing of the non-chosen
alternative information and a negative framing of the chosen alternative
information. This is exactly what happened. Furthermore, among all the framing
conditions the lowest level of preference for inconsistency should have occurred in
the condition in which people received a negative framing of the non-chosen
alternative information and a positive framing of the chosen alternative
information. This too occurred.
One might intuitively think that the distribution of information choice
among people who received a positive framing of both pieces of information or a
negative framing of both pieces of information would closely resemble the
distribution of information choice in the control condition due to a canceling out
effect for the frames. However, the main effects for the frames would suggest that
instead the distributions in these 2 conditions should be more in the middle range.
Specifically, if you think of the positive and negative framing manipulations as
opposing forces, then mathematically, receiving equal frames should produce
distributions that are averages of the distributions for conditions in which people
receive opposite frames.
It was not surprising to see that people who received a negative framing of
the chosen alternative information or a positive framing of the non-chosen
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
29
alternative information were more likely to choose the non-chosen alternative
information than were people in the control condition. This finding demonstrates
that the frames not only affected information choice relative to each other, but also
relative to the control condition. This means that framing information can shift
people away from the chosen alternative information and towards the non-chosen
alternative information when compared to what happens in the classic dissonance
situation in which no outside influences (i.e. framing manipulations) are present to
affect people’s actions in response to dissonance.
Only those participants who specifically received a positive framing of the
chosen alternative information and a negative framing of the non-chosen alternative
information should have not significantly differed from the control condition
regarding information choice, which is exactly what happened x2(l, n = 137) =
2.38, p = .12 (87% versus 94% choosing the chosen alternative information,
respectively). This finding makes sense because people in this condition received
no suggestions to do anything other than what they would have done without the
presence of frames.
The effects found for the spreading of alternatives raised some concern for
part of the study’s design. While the spreading of alternatives was significant and
thereby supports the contention that the study was successfully designed as a “free-
choice” paradigm study of cognitive dissonance, it was a relatively small spread
between pre-decisional to post-decisional chosen item difference scores and pre-
decisional to post-decisional non-chosen item difference scores (Ms = 0.01 vs. -
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
30
0.36, respectively). It is clear from looking at these means that people’s ratings for
the chosen item did not significantly increase, rather it was the decrease in people’s
ratings for the non-chosen item that resulted in the significant spreading of
alternatives. This sparked the idea that perhaps by having people choose between
their 2 most highly preferred items, the available range for spreading near the top of
the scale was restricted. Therefore, it would be better to ask participants to decide
between 2 items that are highly preferred, but that are not the most highly preferred
items, thereby improving the available range and amount of available spreading.
Greater spreading would potentially enhance the effects of cognitive dissonance.
This was the foundation for Study 2.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
31
CHAPTER III: STUDY 2
Study 2 was designed to increase the magnitude of the spreading of
alternatives, thereby depicting a more accurate model of cognitive dissonance and
possibly enabling a more accurate understanding of how people respond to the
framing manipulations and the no framing control condition in terms of information
choice (chosen alternative vs. non-chosen alternative). Rather than asking people
to decide between their 2 most highly preferred items and risk a ceiling effect for
ratings of liking as was the case in Study 1, people in this study were asked to
choose between their second most preferred pair of items. The hypothesis was that
when choosing between 2 items that are highly desirable, but not the 2 most highly
desirable, the reduced level of interest among the former pair compared to the latter
pair would allow for greater spreading from pre-choice to post-choice rated liking
between the 2 items and thus establish a stronger “free-choice” paradigm study of
cognitive dissonance. Additionally, it was hypothesized that greater magnitude of
spreading of alternatives might influence information choice.
Method
Participants. This study consisted of 1219 participants (348 male, 860
female) ranging in age from 18 to 86 with a mean age of 36. Participants were
recruited online through either an advertisement placed on the internet or email.
Participation occurred online. The high number of participants was due to an
unexpectedly heightened influx of participants over a period of time that was
predicted to attract far fewer participants.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
32
Materials and Procedure. The materials and procedure were identical to
Study 1 with the exception of the items that were used to establish the spreading of
alternatives. Specifically, the initial instructions differed from Study 1 in that they
explained to participants that the only resources available to them on their
adventure in the wilderness would be 4 items, as opposed to 2 items as was the case
in Study 1, chosen from a list of 8 possible items. Furthermore, they were told that
they must decide which items they feel are the 2 most essential items to have, given
the type of environment that they would be entering and that additionally, they
would be choosing another 2 “backup” items in case they lost one or both of their 2
most preferred items. After providing their initial ratings of the 8 items and then
choosing their 2 most highly preferred items, participants were then asked to
choose their 2 “backup” items to take with them on their adventure.
Participants then rated how much they would like to have their 2 “backup”
items and then clicked through to “The Survival Shop” where they were told that
this was the shop that sold the 4 items that they wished to take with them on their
adventure. It was further explained that unfortunately the shopkeeper was sold out
of their 2 most preferred items and could only spare 1 of their 2 “backup” items,
but was willing to give them whichever one they wanted the most. Participants
then chose which of the 2 “backup” items they wanted to take with them on their
adventure and then the study proceeded from there on identically to Study 1.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
33
Results
Overall Distributions of Information Choosers based on Framing
Conditions. The overall distribution for choice of information for all 4 framing
conditions and the no framing control condition adhered to roughly the same
pattern found in study 1 (see Figure 2.1). The number of participants choosing to
read the non-chosen alternative information was highest (53%) for participants who
received a negative framing of the chosen alternative information and a positive
framing of the non-chosen alternative information. Participants who received a
positive framing of the chosen alternative information and a negative framing of the
non-chosen alternative information or who were in the no framing control
condition, were least likely to choose to read the non-chosen alternative
information, 13% and 21% respectively. As in Study 1, the distribution of
information choice in the control condition, with 79% of participants choosing to
read the chosen alternative information and 21% of participants choosing to read
the non-chosen alternative information, supports the classic dissonance effect.
However, in the control condition of this study, preference for the chosen
alternative information dropped 15% from Study l ’s distribution (79% vs. 94%,
respectively). Participants receiving positive frames for both pieces of information
or negative frames for both pieces of information were nearly identical in the
likelihood of choosing to read the non-chosen alternative information, 24% and
25% respectively.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
34
Figure 2.1. Overall distribution for choice of information based on frames.
■ Chosen Alternative
■ Non-Chosen Alternative
pos CA Info pos CA Info neg CA Info neg CA Info no frame
/ pos NCA / neg NCA / pos NCA / neg NCA control
Info Info Info Info
Note. CA Info = framing of the chosen alternative information
NCA Info = framing of the non-chosen alternative information
Spreading of Alternatives. All calculations regarding the spreading of
alternatives were performed identically to those done in Study 1. As observed in
Study 1, the spreading of alternatives occurred between pre-decisional to post-
decisional chosen item difference scores and pre-decisional to post-decisional non-
chosen item difference scores, t(965) = 17.783, p < .001 (Ms = 0.31 vs. -0.59,
respectively). The increased magnitude of spreading suggests that by having
participants choose between 2 highly preferred items that were not their most
highly preferred items, the available range for item ratings was successfully
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
35
increased. A logistic regression analysis indicated that the amount of spreading did
not predict information choice, (/3 = -.09, p = .098). After centering the spreading
of alternatives scores, logistic regression analyses determined that the magnitude of
the spreading of alternatives modeled with the framing manipulations also did not
predict information choice, (/3 = -.07, p = .191), nor were there any interactions.
Even when examining only those participants in the no framing control condition,
the magnitude of the spreading of alternatives did not predict information choice, (/?
= -.14, p = .416).
Effects of Frames on Information Choice. As in Study 1, analyses were
conducted to determine if just the framing manipulations alone would predict
information choice. Logistic regression analyses revealed main effects for both the
framing of the chosen alternative information, (/3 = 1.08, p < .001), and the framing
of the non-chosen alternative information, (/3 = -1.03, p < .001), on choice of
information. The number of participants choosing to read the non-chosen
alternative information was highest when they received a negative framing, as
opposed to a positive framing, of the chosen alternative information or when they
received a positive framing, as opposed to a negative framing, of the non-chosen
alternative information (see Figure 2.2). No interactions were detected.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
36
Figure 2.2. Main effects of framing manipulations on the non-chosen
alternative information choice.
50%
40%
30%
Positive
Negative
20%
10%
CA Info NCA Info
Note. CA Info = the chosen alternative information
NCA Info = the non-chosen alternative information
To better understand the effects of the framing manipulations relative to the
classic cognitive dissonance phenomenon, as represented in the no framing control
condition, a series of chi-square tests were performed on information choice (see
Figures 2.3a & 2.3b).
Participants that received a positive framing of the chosen alternative
information versus those that received no frames were not significantly different in
their choice of information, with both groups mostly choosing to read the chosen
alternative information, % 2(1, n = 749) = .40, p = .53. By contrast, significantly
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
37
more participants who received a negative framing of the chosen alternative
information chose to read the non-chosen alternative information, ^2(1, n = 684) =
19.64, p < .001, than did participants in the no framing control condition.
Figure 2.3a. Information choice distributions for framing of the chosen
alternative information conditions versus the no framing control condition.
No Frame CA Info Frame CA Info Frame
Control Positive Negative
l Chosen Alternative
l Non-Chosen Alternative
Note. CA Info = the chosen alternative information
Participants who received a negative framing of the non-chosen alternative
information versus those that received no frames were not significantly different in
their choice of information, with both groups mostly choosing to read the chosen
alternative information, y2(l, n = 690) = .52, p = .47. However, significantly more
participants who received a positive framing of the non-chosen alternative
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
38
information chose to read the non-chosen alternative information, x2(l, n = 743) =
16.42,E < .001, than did participants in the no framing control condition.
Figure 2.3b. Information choice distributions for framing of the non-chosen
alternative information conditions versus the no framing control condition.
100% - I
No Frame NCA Info NCA Info
Control Frame Positive Frame
Negative
■ Chosen Alternative
■ Non-Chosen Alternative
Note. NCA Info = the non-chosen alternative information
Discussion
The purpose of this study was to attempt to increase the magnitude of the
spreading of alternatives from what occurred in Study 1. By having participants
rate their pre-choice and post-choice liking for 2 highly preferred items that weren’t
their 2 most preferred items, the magnitude of the spreading of alternatives
occurred and was successfully increased. This increase in spreading was observed
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
39
for the chosen item (Ms: Study 1 = .01, Study 2 = 0.31) and for the non-chosen
item (Ms: Study 1 = -0.36, Study 2 = -0.59). The effects of this change in the
design of the study were also observed in the distribution of information choosers
in the no framing control condition. Instead of 94% of participants choosing the
chosen alternative information, as was the case in Study 1, 79% of participants
chose the chosen alternative information. These results suggest that having
participants choose between 2 highly preferred items as opposed to their 2 most
preferred items, allows for the spreading of alternatives to be greater for all
participants and the likelihood of choosing the chosen alternative information to go
down somewhat among participants in the no framing control condition.
However, as in Study 1, it is still the case that the magnitude of the
spreading of alternatives did not predict information choice among participants,
whether looking at all the participants, just the no framing control condition, or the
framing manipulation conditions. This suggests that the spreading of alternatives
inevitably occurs as a result of dissonance, but that its magnitude does not affect
whether people will choose the chosen alternative information or the non-chosen
alternative information nor will it interact with any framing manipulations to affect
information choice.
The main effects for the framing manipulations were replicated from Study
1 suggesting good reliability in their effects across studies. This time, however,
only for the instances in which participants received either a negative framing of
the chosen alternative information or a positive framing of the non-chosen
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
alternative information was the number of participants choosing inconsistency
greater than the number of participants choosing inconsistency in the no framing
control condition. Unlike participants in Study 1, participants in this study who
received a positive framing of the chosen alternative information or a negative
framing of the non-chosen alternative information resembled the distribution of
information choosers found in the no framing control condition. This change does
not seem to be due to any significant change in the effectiveness of the framing
manipulations, but is rather due to the decrease in the number of participants in the
no framing control condition choosing the chosen alternative information.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
41
CHAPTER IV: STUDY 3
With the effectiveness of the framing manipulations appearing stable and
the magnitude of the spreading of alternatives increased, the next task at hand was
to more accurately and sensitively assess emotions and run a study designed to
detect differences in emotional experiences among consistency and inconsistency
choosers based on the framing manipulations. The main hypotheses for this study
were that people who chose the chosen alternative information would be less happy
and more anxious if they had received a positive framing, rather than a negative
framing, of the non-chosen alternative information or if they had received a
negative framing, rather than a positive framing, of the chosen alternative
information. Similarly, it was predicted that people who chose the non-chosen
alternative information would be more happy and less anxious if they had received
a positive framing, rather than a negative framing, of the non-chosen alternative
information or if they had received a negative framing, rather than a positive
framing, of the chosen alternative information. It was also predicted that sadness
and calmness would demonstrate effects due to being on the opposite end of the
concepts of happiness and anxiety, respectively, but it was uncertain if any
significant differences would show up in terms of anger, given the pleasant nature
of the experiment (i.e. a game).
Method
Participants. This study consisted of 360 participants (111 male, 247
female) ranging in age from 18 to 63 with a mean age of 33. Participants were
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
42
recruited online through either an advertisement placed on the internet or email.
Participation occurred online.
Materials and Procedure. The materials and procedure were identical to
Study 2 with the exception of the items that were used to assess changes in
emotion. Participants filled out a pretest emotion measure to assess baseline
emotions. The list consisted of 15 randomly ordered items, with 3 items measuring
each of 5 different emotions (happiness, calmness, sadness, anxiety, anger). The
scale ranged from “Very Slightly or Not At All” to “Extremely”. This list was a
subset constructed from the PANAS-X Affect schedule (Watson & Clark, 1994).
A subset was used to reduce the amount of time participants needed to devote to
filling out these measures. The same list was then given a new random order and
then presented to participants a second time immediately after they chose which
information they wanted to read. After completing the list for the second time,
participants read the information that they chose to read and the procedure
continued from there identically to Study 2.
Factor analyses of both the pre-mood emotions and the post-mood emotions
generated a five factor structure (all loadings > .4). Based on the 5 component
matrix, reliability analyses were then performed on each of the components. All 5
components formed highly reliable composites for both the pre-mood emotions and
the post-mood emotions; a testament to the validity of the PANAS-X as a measure
of emotion (see Table 3.1). The composites capturing these 5 emotions were used
for subsequent analyses.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
43
Table 3.1. Reliability analyses for both pre-mood and post-mood measures of
emotion.
Pre-Mood Post-Mood
Happiness Alpha = .83 Alpha = .90
Cheerful
Happy
Enthusiastic
Calmness Alpha = .85 Alpha = .90
At Ease
Calm
Relaxed
Sadness Alpha =.91 Alpha =.90
Blue
Downhearted
Sad
Anxiety Alpha = .78 Alpha = .78
Nervous
Afraid
Jittery
Anger Alpha = .81 Alpha = .83
Hostile
Angry
Irritable
Results
Overall Distributions of Information Choosers based on Framing
Conditions. Across all 4 framing conditions and the no framing control condition,
the overall breakdown for choice of information followed the same pattern as was
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
44
observed in studies 1 and 2 (see Figure 3.1). Specifically, the number of
participants choosing to read the non-chosen alternative information was highest
(54%) among participants who received a negative framing of the chosen
alternative information and a positive framing of the non-chosen alternative
information. Participants who received a positive framing of the chosen alternative
information and a negative framing of the non-chosen alternative information or
who were in the no framing control condition, were least likely to choose to read
the non-chosen alternative information, 16% and 22% respectively. As in the 2
prior studies, the distribution of information choice in the control condition, with
78% of participants choosing to read the chosen alternative information and 22% of
participants choosing to read the non-chosen alternative information, supports the
classic dissonance effect. Participants receiving positive frames for both pieces of
information or negative frames for both pieces of information were highly similar
in likelihood for choosing to read the non-chosen alternative information, 31% and
25% respectively.
Spreading of Alternatives. As was the case in studies 1 and 2, the spreading
of alternatives was observed between pre-decisional to post-decisional chosen item
difference scores and pre-decisional to post-decisional non-chosen item difference
scores, t(293) = 10.066, p < .001 (Ms = 0.27 vs. -0.71, respectively). A logistic
regression analysis indicated that the amount of spreading did not predict
information choice, (/3 = .02, p = .867). After centering the spreading of
alternatives scores, logistic regression analyses determined that the magnitude of
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
45
the spreading of alternatives modeled with the framing manipulations also did not
predict information choice, ((3 = -.03, p = .734), nor were there any interactions.
Even when examining only those participants in the no framing control condition,
the magnitude of the spreading of alternatives did not predict information choice, (j3
= -.19, p =.472).
Figure 3.1. Overall distribution for choice of information based on frames.
■ Chosen Alternative
m Non-Chosen Alternative
pos CA Info pos CA Info neg CA Info neg CA Info no frame
/ pos NCA / neg NCA / pos NCA / neg NCA control
Info Info Info Info
Note. CA Info = framing of the chosen alternative information
NCA Info = framing of the non-chosen alternative information
Effects of Frames on Information Choice. As in studies 1 and 2, logistic
regression analyses revealed main effects for both the framing of the chosen
alternative information, (/3 = .764, p < .01), and the framing of the non-chosen
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
46
alternative information, (y 3 = -1.131 > E < .001), on choice of information. The
number of participants choosing to read the non-chosen alternative information was
highest when they received a negative framing, as opposed to a positive framing, of
the chosen alternative information or when they received a positive framing, as
opposed to a negative framing, of the non-chosen alternative information (see
Figure 3.2). No interactions were detected.
Figure 3.2. Main effects of framing manipulations on the non-chosen
alternative information choice.
50%
40%
30%
20%
10%
I Positive
Negative
CA Info NCA Info
Note. CA Info = the chosen alternative information
NCA Info = the non-chosen alternative information
To better understand the effects of the framing manipulations relative to the
classic cognitive dissonance phenomenon, as represented in the no framing control
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
47
condition, a series of chi-square tests were performed on information choice (see
Figures 3.3a & 3.3b).
Figure 3.3a. Information choice distributions for framing of the chosen
alternative information conditions versus the no framing control condition.
No Frame CA Info Frame CA Info Frame
Control Positive Negative
I Chosen Alternative
I Non-Chosen Alternative
Note. CA Info = the chosen alternative information
Participants that received a positive framing of the chosen alternative
information versus those that received no frames were not significantly different in
their choice of information, with both groups mostly choosing to read the chosen
alternative information, x2(l, n = 182) = .01, p = .94. By contrast, significantly
more participants that received a negative framing of the chosen alternative
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
48
information chose to read the non-chosen alternative information, %2(1, n = 250) =
5.13, £ < .03, than did participants in the no framing control condition.
Figure 3.3b. Information choice distributions for framing of the non-chosen
alternative information conditions versus the no framing control condition.
No Frame
Control
NCA Info
Frame Positive
NCA Info
Frame
Negative
Chosen Alternative
i Non-Chosen Alternative
Note. NCA Info = the non-chosen alternative information
Participants that received a negative framing of the non-chosen alternative
information versus those that received no frames were not significantly different in
their choice of information, with both groups mostly choosing to read the chosen
alternative information, x2(l, n = 235) = .02, p = .90. However, significantly more
participants that received a positive framing of the non-chosen alternative
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
49
information chose to read the non-chosen alternative information, x2(l, n = 197) =
10.04, g < .01, than did participants in the no framing control condition.
Effects of Framing Manipulations on Emotions for Chosen Alternative and
Non-Chosen Alternative Information Choosers. The first analyses of interest
regarding the effects of the framing manipulations on emotions were to determine
whether or not the frames were effective for the chosen alternative information and
the non-chosen alternative information choosers separately. For example, is it the
case that people who choose the non-chosen alternative information can be made to
feel significantly happier if they believed that there was potential value in that
choice than if they had been told there was little to no value in that choice?
Regarding calmness, is it the case that people who choose the chosen alternative
information can be made to feel less calm if they believed that there was potential
value in choosing the non-chosen alternative information than if they had been told
there was little to no value in choosing the non-chosen alternative information?
The following analyses were designed to answer these questions.
A 2(the chosen alternative information frame: positive, negative) X 2(the
non-chosen alternative information frame: positive, negative) analysis of
covariance was conducted on the chosen alternative information choosers and their
levels of post-information choice happiness, controlling for pre-information choice
happiness. This analysis revealed an interaction between the chosen alternative
information and the non-chosen alternative information frames, F(3, 189) = 5.64, p
< .02 (see Figure 3.4). Specifically, if the chosen alternative information choosers
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
50
received a negative framing of the chosen alternative information, then they were
equally happy regardless of whether they received a positive or negative framing of
the non-chosen alternative information (Ms = 3.02 vs. 2.91, respectively).
Figure 3,4. Interaction between framing manipulations for chosen alternative
information choosers on levels of post-information choice happiness._________
3.5
3.4
3.3
3.2
3.1
Means 3
2.9
2.8
2.7
2.6
2.5
Positive CA Info Negative CA Info
Framing Framing
— Positive NCA
Info Framing
” 'S 8 ~ " Negative NCA
Info Framing
Note. CA Info = the chosen alternative information
NCA Info = the non-chosen alternative information
However, if the chosen alternative information choosers received a positive
framing of the chosen alternative information, then those participants who also
received a positive framing of the non-chosen alternative information were
significantly less happy than participants who also received a negative framing of
the non-chosen alternative information (Ms = 2.67 vs. 3.11, respectively). No
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
51
effects on happiness levels were observed among consistency choosers when
comparing participants who received framing manipulations versus participants in
the no framing control condition, (Fs < 1).
A 2(the chosen alternative information frame: positive, negative) X 2(the
non-chosen alternative information frame: positive, negative) analysis of
covariance was conducted on the non-chosen alternative information choosers and
their levels of post-information choice happiness, controlling for pre-information
choice happiness. This analysis uncovered a significant main effect of the non-
chosen alternative information frames on happiness levels, F(3, 85) = 5.30, p < .03.
Specifically, the non-chosen alternative information choosers were significantly
happier if they received a positive framing of the non-chosen alternative
information than if they had received a negative framing of the non-chosen
alternative information (Ms = 2.96 vs. 2.46, respectively). No effects on happiness
levels were observed among inconsistency choosers when comparing participants
who received framing manipulations versus participants in the no framing control
condition, (Fs < 1), although the means were moving in the right direction (Ms:
positive framing the non-chosen alternative information = 2.96 vs. no framing
control condition = 2.79 vs. negative framing the non-chosen alternative
information = 2.46).
Next, analyses of covariance were conducted to determine whether or not
the chosen alternative information and the non-chosen alternative information
choosers varied in levels of post-information choice happiness depending on
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
52
whether they received a positive or negative framing of the chosen alternative
information or the non-chosen alternative information. The 2(the chosen
alternative information frame: positive, negative) X 2(information choice: chosen
alternative, non-chosen alternative) analysis of covariance on levels of post
information choice happiness (controlling for pre-information choice happiness)
revealed no effects.
However, the 2(the non-chosen alternative information frame: positive,
negative) X 2(information choice: chosen alternative, non-chosen alternative)
analysis of covariance on levels of post-information choice happiness (controlling
for pre-information choice happiness) revealed a significant interaction, F (3 ,189) =
4.33, p < .04 (see Figure 3.5). Specifically, if participants received a negative
framing of the non-chosen alternative information, then they were less happy if
they chose the non-chosen alternative information than if they chose the chosen
alternative information. However, if participants received a positive framing of the
non-chosen alternative information, then the participants who chose the non-chosen
alternative information were just as happy as the participants who chose the chosen
alternative information.
Interestingly, participants who received a positive framing of the non-
chosen alternative information and chose to read the non-chosen alternative
information were just as happy as participants who received a positive framing of
the chosen alternative information and chose to read the chosen alternative
information (Fs < 1).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
53
Figure 3.5. Interaction between information choosers for framing of the non-
chosen alternative information on levels of post-information choice happiness.
4
3.8
3.6
3.4
3.2
Means 3
2.8
2.6
2.4
2.2
2
CA Info Choosers NCA Info Choosers
-♦ -P o s itiv e NCA
Info Framing
'’S — Negative NCA
Info Framing
Note. CA Info = the chosen alternative information
NCA Info = the non-chosen alternative information
A 2(the chosen alternative information frame: positive, negative) X 2(the
non-chosen alternative information frame: positive, negative) analysis of
covariance was conducted on the chosen alternative information choosers and their
levels of post-information choice calmness, controlling for pre-information choice
calmness. This analysis uncovered a marginally significant main effect of the non-
chosen alternative information frames on calmness levels, F(3, 189) = 5.30, p =
.066. Specifically, the chosen alternative information choosers were less calm if
they received a positive framing of the non-chosen alternative information than if
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
54
they had received a negative framing of the non-chosen alternative information (Ms
= 2.99 vs. 3.32, respectively).
This analysis also revealed an interaction between the chosen alternative
information and the non-chosen alternative information frames that provided more
insight into the main effect above, F(3, 189) = 4.71, p < .04 (see Figure 3.6).
Specifically, if the chosen alternative information choosers received a negative
framing of the chosen alternative information, then they were equally calm
regardless of whether they received a positive or negative framing of the non-
chosen alternative information (Ms = 3.10 vs. 3.18, respectively). However, if the
chosen alternative information choosers received a positive framing of the chosen
alternative information, then those participants who also received a positive
framing of the non-chosen alternative information were significantly less calm than
participants who also received a negative framing of the non-chosen alternative
information (Ms = 2.87 vs. 3.53, respectively). No effects on calmness levels were
observed among consistency choosers when comparing participants who received
framing manipulations versus participants in the no framing control condition, (Fs
< !)•
Interestingly, participants who received a positive framing of the non-
chosen alternative information and chose to read the non-chosen alternative
information were just as calm as participants who received a positive framing of the
chosen alternative information and chose to read the chosen alternative information
(Fs < 1).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
55
Figure 3.6. Interaction between framing manipulations for chosen alternative
information choosers on levels of post-information choice calmness.
4
3.9
3.8
3.7
3.6
3.5
3.4
3.3
3.2
3.1
Means 3
2.9
2.8
2.7
2.6
2.5
Positive CA Info Frame Negative CA Info Frame
- Positive NCA
Info Frame
Negative NCA
Info Frame
Note. CA Info = the chosen alternative information
NCA Info = the non-chosen alternative information
A 2(the chosen alternative information frame: positive, negative) X 2(the
non-chosen alternative information frame: positive, negative) analysis of
covariance was conducted on the non-chosen alternative information choosers and
their levels of post-information choice calmness, controlling for pre-information
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
56
choice calmness, but no significant differences were observed among the
conditions (Fs < 1). No other emotion effects (sadness, fear, anger, negative
arousal) were detected in this study.
Discussion
The main effects for the framing manipulations replicated from Study 1 and
Study 2 once again suggesting good reliability in their effects across studies.
Additionally, the distribution in the no framing control condition regarding
information choice and the magnitude of the spreading of alternatives measures
were all nearly identical to Study 2 suggesting good reliability for the effects of
having people choose between 2 highly preferred items, as opposed to the 2 most
highly preferred items.
Some interesting effects were discovered in terms of the emotions people
experienced based on their information choice and the framing manipulations.
Specifically, people who chose the chosen alternative information after having had
that information framed positively, were significantly less happy if the non-chosen
alternative information had been framed positively instead of negatively. This
finding suggests that by mildly reframing information, you can actually get people
to feel less happy about choosing information about a chosen alternative!
Furthermore, due to the fact that this study demonstrated that people who chose the
non-chosen alternative information were significantly happier if they received a
positive framing, instead of a negative framing, of the non-chosen alternative
information, indicates that the reduced happiness predicted by cognitive dissonance
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
57
due to choosing the non-chosen alternative information can be reduced by simply
framing the non-chosen alternative information positively.
Also, the finding that the non-chosen alternative information choosers were
just as happy as the chosen alternative information choosers when both groups
received a positive framing of the non-chosen alternative information demonstrates
the effectiveness of positively framing the non-chosen alternative information. By
doing so, the non-chosen alternative information choosers do not experience a
reduction in happiness as they do when they receive a negative framing of the non-
chosen alternative information or no frames at all, although the latter effect was not
significant. This idea is supported by the nearly identical means for happiness
among people who received a positive framing of the non-chosen alternative
information and then chose the non-chosen alternative information and among
people who received a positive framing of the chosen alternative information and
then chose the chosen alternative information (Ms = 2.96 vs. 2.94, respectively).
Another interesting finding was that the chosen alternative information
choosers were less calm if they received a positive framing of the non-chosen
alternative information than if they had received a negative framing of the non-
chosen alternative information. This finding indicates that by mildly reframing
information, you can actually get people to feel more anxious about choosing
information about a chosen alternative! Also interesting was that participants who
received a positive framing of the non-chosen alternative information and chose to
read the non-chosen alternative information were just as calm as participants who
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
58
received a positive framing of the chosen alternative information and chose to read
the chosen alternative information. This demonstrates that by getting people to
consider the possible value in choosing the non-chosen alternative information, it
not only increases their likelihood of choosing the non-chosen alternative
information, but it also reduces the anxiety typically associated with dissonance to
a level equal to that of people who choose the chosen alternative information.
Across 3 studies, a reliable trend was observed for information choice based
on framing manipulations. Perhaps indicative of the frames serving only as mild
manipulations of information choice, is the overall trend for people to choose
information about a chosen alternative over information that is about a non-chosen
alternative. Even in the condition in which people receive a positive framing of the
non-chosen alternative information and a negative framing of the chosen alternative
information, the distribution of information choice is roughly split down the
middle. Is it possible that certain individual differences might help to explain why
half of these people choose the non-chosen alternative information as suggested by
the frames, while the other half are unaffected by the frames and choose the chosen
alternative information? Study 4 was designed to address this question.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
59
CHAPTER V: STUDY 4
Study 4 was designed to determine if certain individual differences were at
least partially responsible for the distribution of information choice in Studies 1, 2,
and 3. An examination of how different people respond to the “selective exposure”
paradigm in cognitive dissonance in terms of their willingness to expose
themselves to information about a non-chosen alternative seemed warranted. It was
hypothesized that some people may be more open to information that does not
support a prior decision, if they perceive that such information may improve their
decision making abilities. Three different scales were implemented to capture
individual differences relevant to the decision making processes in the present
studies.
The Need for Closure Scale offers the ability to discriminate between
people who simply want an answer quickly (close-minded) and people who are
willing to investigate further to increase the likelihood of being correct (open-
minded). People who are high on need for closure prefer predictability and are
uncomfortable in ambiguous situations. These people do not like to re-evaluate
past decisions. On the other hand, people who are low on need for closure are very
open to reconsidering past decisions and alternative opinions (Neuberg, Judice, &
West, 1997). Therefore, if information is available that can clarify what is the best
decision, low need for closure people should be more open to reading such
information than people who are high on need for closure even if that information
is about a non-chosen alternative.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
60
It was predicted that people who were high on need for closure would be
less affected by the framing manipulations than people low on need for closure
because the former group would want to choose information that is about a chosen
alternative so that they don’t have to reconsider their decisions. Low need for
closure people should be more receptive to the frames because they are concerned
with making the best possible decisions. Therefore, if prompted to think about the
value of choosing information that is about a non-chosen alternative, they would be
willing to take advantage of that opportunity.
The Fear of Invalidity Scale is used to differentiate between people who
show little concern for the possibility of being incorrect and therefore make
decisions rather quickly (low fear of invalidity) and people who worry about
making good decisions and are consequently indecisive and more likely to second-
guess themselves (high fear of invalidity) (Neuberg, Judice, & West, 1997). So, if
information is available that can potentially help people alleviate their concerns that
they made an incorrect decision, people who are high on fear of invalidity should
be more interested in that information than people who are low on fear of invalidity
because the latter group is less bothered by the possibility of being wrong.
It was predicted that people who are high, as opposed to low, on fear of
invalidity would be more affected by the frames and therefore more likely to
consider information about a non-chosen alternative in the hopes of determining
whether or not they chose the correct item to take with them on their adventure.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
61
Low fear of invalidity people should be less concerned with being incorrect and
therefore less receptive to the frames.
The Preference for Consistency Scale is designed to discriminate between
people who have a strong desire to be consistent among their thoughts and actions
(high preference for consistency) versus people who are relatively unconcerned
with whether or not they think and behave in a consistent manner (low preference
for consistency) (Cialdini, Trost, & Newsom, 1995). Therefore, people who are
high on preference for consistency should be resistant to considering information
about a non-chosen alternative because they want their decisions to be consistent
with each other. Contrarily, people who are low on preference for consistency
should be more receptive to information about a non-chosen alternative because
they are less concerned with maintaining consistency among their decisions.
It was predicted that people who were high on preference for consistency,
would be less affected by the framing manipulations than people low on preference
for consistency because the former group would want to choose information that is
about a chosen alternative so as to be consistent among their item and information
choices. Low preference for consistency people should be more receptive to the
frames because they are less bothered by being inconsistent among their thoughts
and actions. Therefore, if given the opportunity to read potentially valuable
information that is about a non-chosen alternative, they would not be bothered by
accepting that opportunity.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
62
Method
Participants. This study consisted of 421 participants (74 male, 344 female)
ranging in age from 18 to 78 with a mean age of 36. Participants were recruited
online through either an advertisement placed on the internet or email.
Participation occurred online.
Materials and Procedure. The materials and procedure were identical to
Study 3 with the exception of the inclusion of personality measures added to the
end of the study. Additionally, a new continuous measure of information
preferences was included to see how accurately people’s preferences for
information rated on a continuous scale would reflect people’s actual information
choices. This continuous measure of information preferences was an 11 point scale
ranging from -5 (strongly prefer to read information about the OTHER item) to +5
(strongly prefer to read information about the CHOSEN item) and was positioned
directly before the question asking people to make their actual choice of
information.
The 3 personality measures used to assess how different types of people
respond to dissonance, were items selected from the Need For Closure Scale
(Neuberg, et al, 1997), Fear of Invalidity (Neuberg, et al, 1997), and Preference for
Consistency Scales (Cialdini, et al, 1995). Due to concerns regarding fatigue
effects, only a selection of questions was used from each scale to reduce the total
number of items that participants would be required to complete. Additionally,
many of the items from these scales were of less interest to the current research.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
63
For example, many of the preference for consistency items measure people’s
interest in appearing consistent in their opinions and actions, rather than actually
being consistent among their opinions and actions. The current research is
primarily interested in how people actually are and not how they try to appear.
Additionally, other items were constructed as variants of the items from the 3
scales to bolster the ability to measure the individual differences of interest.
Factor analyses of the personality measures indicated three main factors that
corresponded to the three intended measures, although not all of the items from
each measure loaded. Some of the additional items that were constructed for this
study also mapped onto those factors (all loadings > .4). Based on the loadings
across the 3 components in the factor analysis, reliability analyses were then
performed on each of the components. All 3 components formed highly reliable
composites (see Table 4.1). Additionally, none of the scales were significantly
correlated with each other suggesting that they all measured separate individual
differences (all r’s < .3). The composites for these 3 personality measures were
used for subsequent analyses. Factor analyses of both the pre-mood and post-mood
emotions generated essentially the same loadings as in Study 3 (all loadings > .4).
Based on the 5 component matrix, reliability analyses were then performed on each
of the components. As in Study 3, all 5 components formed highly reliable
composites for both the pre-mood and post-mood emotions, once again
demonstrating the reliability of the PANAS-X as a measure of emotion (see Table
4.2). The composites for these 5 emotions were used for subsequent analyses.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 4.1. Reliability analyses for personality measures.
64
Need For Closure Alpha = .7745
ADD8.1 find it important to understand the details of a different opinion, so
that I can have a better understanding of my own opinion.
NC4. When thinking about a problem, I consider as many different opinions on
the issue as possible.
ADD7.1 find value in exploring alternatives, even if those alternatives may not
benefit me immediately.
NC3. When considering most conflict situations, I can usually see how both
sides could be right.
NCI. Even after I’ve made up my mind about something, I am eager to consider
a different opinion.
ADD2.1 try to leam from my mistakes, even if thinking about where I went
wrong is unpleasant.
Fear of Invalidity Alpha = .7715
FI4.1 wish I didn’t worry so much about making errors.
FI1. 1 can be reluctant to commit myself to something because of the possibility
that I might be wrong.
FI2. Even after making an important decision, I continue to think about the pros
and cons to make sure that I am not wrong.
NC2. When considering most conflict situations, I can usually see how both
sides could be right.
FI3.1 wish I didn’t worry so much about making errors.
ADD4.1 do not like to question myself by considering information that might
suggest I made a poor choice.
Preference For Consistency Alpha = .6160
PC3. I’m uncomfortable holding two beliefs that are inconsistent with each
other.
PCI. It is important to me that my actions are consistent with my beliefs.
PC4. It doesn’t bother me much if my actions are inconsistent.
ADD3. It’s important to me to remain positive regarding a choice I made.
PC2.1 get uncomfortable when my behavior contradicts my beliefs.
ADD5. Once I’ve made a decision, I stick with it.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
65
Table 4.2. Reliability analyses for both pre-mood and post-mood measures of
emotion.
Pre-Mood Post-Mood
Happiness Alpha = .87 Alpha = .89
Cheerful
Happy
Enthusiastic
Calmness Alpha = .92 Alpha = .88
At Ease
Calm
Relaxed
Sadness Alpha = .91 Alpha = .80
Blue
Downhearted
Sad
Anxiety Alpha = .81 Alpha =.80
Nervous
Afraid
Jittery
Anger Alpha = .76 Alpha = .89
Hostile
Angry
Irritable
Results
Overall Distributions of Information Preferences and Choice based on
Framing Conditions. Across all 4 framing conditions and the no framing control
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
66
condition, the overall breakdown for choice of information followed the same
pattern as was observed in studies 1, 2, and 3 (see Figure 4.1).
Figure 4.1. Overall distribution for choice of information based on frames.
■ Chosen Alternative
■ Non-Chosen Alternative
pos CA Info pos CA Info neg CA Info neg CA Info no frame
I pos NCA / neg NCA / pos NCA I neg NCA control
Info Info Info Info
Note. CA Info = framing of the chosen alternative information
NCA Info = framing of the non-chosen alternative information
Specifically, the number of participants choosing to read the non-chosen
alternative information was highest (49%) among participants who received a
negative framing of the chosen alternative information and a positive framing of the
non-chosen alternative information. Participants who received a positive framing
of the chosen alternative information and a negative framing of the non-chosen
alternative information or who were in the no framing control condition, were least
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
67
likely to choose to read the non-chosen alternative information, 15% and 11%
respectively. As in the 3 prior studies, the distribution of information choice in the
control condition, with 89% of participants choosing to read the chosen alternative
information and 11% of participants choosing to read the non-chosen alternative
information, supports the classic dissonance effect. About 1 out of 4 participants
receiving positive frames for both pieces of information or negative frames for both
pieces of information chose to read the non-chosen alternative information, 26%
and 23% respectively.
Based on the framing manipulations, ratings of information preference
followed the same pattern as was observed in the breakdown for choice of
information (see Figure 4.2a). Among the 4 framing conditions and the no framing
control, a similar number of participants, ranging from 13% to 21% across
conditions, were indecisive regarding their information preferences, as indicated by
their use of the 0 point (see Figure 4.2b).
A logistic regression analysis found that ratings of information preference
predicted actual information choice, (/3 = -.50, p < .001), such that the more
someone preferred to read a particular type of information, the more likely they
were to choose to read that information. Similarly, an analysis of variance
determined that people’s choice of information was consistent with their rated
preference for reading information F(l, 400) = 239.45, p < .001 (Ms: consistency
information choosers = 3.16 vs. inconsistency information choosers = -1.62).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
68
Figure 4.2a. Overall distribution for preference rating of information based
on frames for consistency and inconsistency preferences only.
■ Chosen Alternative
□ Non-Chosen Alternative
pos CA Info pos CA Info neg CA Info neg CA Info no frame
/ pos NCA / neg NCA / pos NCA / neg NCA control
Info Info Info Info
Note. CA Info = framing of the chosen alternative information
NCA Info = framing of the non-chosen alternative information
Spreading of Alternatives. As was the case in studies 1, 2, and 3, the
spreading of alternatives was observed between pre-decisional to post-decisional
chosen item difference scores and pre-decisional to post-decisional non-chosen
item difference scores, t(368) = 10.977, p < .001 (Ms = 0.26 vs. -0.58,
respectively). A linear regression analysis demonstrated that the amount of
spreading did not predict rated preferences for information, (/3 = .01, g = .956), and
a logistic regression analysis confirmed that the amount of spreading did not
predict information choice, (/3 = -.08, g = .366).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
69
Figure 4.2b. Overall distribution for preference rating of information based
on frames.
■ Chosen Alternative
a Neutral
1 Non-Chosen Alternative
pos CA Info pos CA Info neg CA Info neg CA Info no frame
/ pos NCA / neg NCA / pos NCA / neg NCA control
Info Info Info Info
Note. CA Info = framing of the chosen alternative information
NCA Info = framing of the non-chosen alternative information
After centering the spreading of alternatives scores, linear and logistic
regression analyses determined that the magnitude of the spreading of alternatives
modeled with the framing manipulations also did not predict ratings of information
preferences, (/? = .08, p = .583), or information choice, (/3 = -.10, p = .316), nor
were there any interactions. Even when examining only those participants in the no
framing control condition, the magnitude of the spreading of alternatives did not
predict ratings of information preferences, (/3 = -.23, p = .349), or information
choice, (/? = .01, p = .989).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
70
Effects of Frames on Rated Information Preferences and Information
Choice. A 2(the chosen alternative information frame: positive, negative) X 2(the
non-chosen alternative information frame: positive, negative) analysis of variance
was conducted on information preference ratings. This analysis revealed main
effects for both the framing of the chosen alternative information, F(3, 313) =
18.58, p < .001, and the framing of the non-chosen alternative information, F(3,
313) = 12.71, p < .001, on ratings of information preferences. Specifically,
participants were significantly less likely to prefer to read the chosen alternative
information when they received a negative framing, rather than a positive framing,
of the chosen alternative information. Participants were also significantly less
likely to prefer to read the chosen alternative information when they received a
positive framing, rather than a negative framing, of the non-chosen alternative
information (see Table 4.3). There were no interactions between the framing
manipulations (Fs < 1).
Table 4.3. Main effects of frames on information preference ratings.
Direction
Framing (Chosen
Alternative Information)
Framing (Non-Chosen
Alternative Information)
No Framing
Control
Positive 2.63 1.06
2.86
Negative 0.96 2.49
Note. Higher mean values represent greater preference for the chosen
alternative information.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
71
Logistic regression analyses revealed main effects for both the framing of
the chosen alternative information, (j3 = -.81,p < -01), and the framing of the non-
chosen alternative information, (/3 = .98,2 < .001), on choice of information. The
number of participants choosing to read the non-chosen alternative information was
highest when they received a negative framing, as opposed to a positive framing, of
the chosen alternative information or when they received a positive framing, as
opposed to a negative framing, of the information about the non-chosen
altemative(see Figure 4.3). No interactions were detected.
Figure 4.3. Main effects of framing manipulations on the non-chosen
alternative information choice.
50% n
40%
30%
20%
10%
0%
Note. CA Info = the chosen alternative information
NCA Info = the non-chosen alternative information
■ Positive
s Negative
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
72
To better understand the effects of the framing manipulations relative to the
classic cognitive dissonance phenomenon, as represented in the no framing control
condition, a series of one-way ANOVAs were performed on ratings of information
preferences (see Table 4.3) and a series of chi-square tests were performed on
information choice (see Figures 4.4a & 4.4b).
Figure 4.4a. Information choice distributions for framing of the chosen
alternative information conditions versus the no framing control condition.
100% n
■ Chosen Alternative
■ Non-Chosen Alternative
No Frame CA Info Frame CA Info Frame
Control Positive Negative
Note. CA Info = the chosen alternative information
Participants who received a positive framing of the chosen alternative
information versus those that received no frames were not significantly different in
their ratings of information preferences (Fs < 1), nor in their choice of information,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
73
with both groups mostly choosing to read the chosen alternative information, *2(1,
n = 251) = 2.99, p = .08. By contrast, significantly more participants who received
a negative framing of the chosen alternative information ended up preferring to
read the non-chosen alternative information, F(l, 245) = 18.17, p < .001, and
choosing to read the non-chosen alternative information, y2(l, n = 259) = 17.83, p
< .001, than did participants in the no framing control condition.
Figure 4.4b. Information choice distributions for framing of the non-chosen
alternative information conditions versus the no framing control condition.
100% n
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
No Frame NCA Info NCA Info
Control Frame Positive Frame
Negative
Note. NCA Info = the non-chosen alternative information
Participants who received a negative framing of the non-chosen alternative
information versus those who received no frames were not significantly different in
■ Chosen Alternative
H Non-Chosen Alternative
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
74
their ratings of information preferences (Fs < 1), nor in their choice of information,
with both groups mostly choosing to read the chosen alternative information, ^2(1,
n = 261) = 2.36, p = .13. However, significantly more participants who received a
positive framing of the non-chosen alternative information preferred to read the
non-chosen alternative information, F(l, 240) = 14.68, p < .001, and chose to read
the non-chosen alternative information, y2(l, n = 249) = 20.29, p < .001, than did
participants in the no framing control condition.
Effects of Personality Measures on Rated Information Preferences and
Information Choice. To test for any possible differences between participants who
were low on need for closure versus people who were high on need for closure, a
variable was constructed that divided participants into low and high need for
closure categories with “low” representing the third of the sample with the lowest
need for closure scores and “high” representing the third of the sample with the
highest need for closure scores.
A 2(need for closure: low, high) X 2(the chosen alternative information
frame: positive, negative) X 2(the non-chosen alternative information frame:
positive, negative) analysis of variance was conducted on information preference
ratings. A main effect was revealed for need for closure personality type, F(7, 192)
= 6.07, p < .02, indicating that among people who received framing manipulations,
low need for closure participants preferred to read the non-chosen alternative
information significantly more than did high need for closure participants (Ms =
1.27 vs. 2.28, respectively). Interestingly, an interaction occurred between need for
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
75
closure personality type and the framing of the non-chosen alternative information,
F (7,192) = 5.13, p < .03 (see Figure 4.5).
Figure 4.5. Interaction between need for closure personality type and framing
of the non-chosen alternative information for magnitude of preference for the
chosen alternative information.
3
2.5
2
1.5
Means
1
0.5
0
Low Need for Closure High Need for Closure
Positive NCA
Info Framing
Negative NCA
Info Framing
Need for Closure Personality Type
Note. NCA Info = the non-chosen alternative information
Specifically, high need for closure participants were not affected by the
framing of the non-chosen alternative information. These participants’ preferences
for the chosen alternative information was the same regardless of how the non-
chosen alternative information was framed. Flowever, the low need for closure
participants were affected by the framing of the non-chosen alternative information.
These participants preferred to read the chosen alternative information significantly
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
76
less if they received a positive framing, as opposed to a negative framing, of the
non-chosen alternative information. Low and high need for closure participants in
the no framing control condition did not significantly differ in their rated
preferences for information (Fs < 1), with both personality types strongly preferring
to read the chosen alternative information (Ms = 3.25 vs. 2.47, respectively).
A logistic regression analysis revealed that need for closure personality type
modeled with the framing manipulations predicted information choice, (/3 = 1.71, p
< .01), such that low need for closure participants were significantly more likely to
choose to read the non-chosen alternative information than were high need for
closure participants (35% versus 23%, respectively). Unlike the analysis looking at
rated information preferences, actual information choice yielded no interaction
between need for closure personality type and framing of the non-chosen
alternative information, (ft = -1.27, p = .07). Although not significant, looking at
percentage breakdowns of information choice, it is clear that low need for closure
participants are moving in the direction of being more affected by the framing of
the non-chosen alternative information than high need for closure participants (see
Figure 4.6). Specifically, low need for closure participants leaned more towards
choosing to read the non-chosen alternative information if they received a positive
framing, as opposed to a negative framing, of the non-chosen alternative
information ^2(1, n = 99) = 6.31, p < .02. Low and high need for closure
participants in the no framing control condition were equally as strong (86%) in
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
77
their likelihood for choosing to read the chosen alternative information, rather than
the non-chosen alternative information, x2(l, n = 64) = 0.00,2 = .96.
Figure 4.6. Number of participants choosing the non-chosen alternative
information based on high versus low need for closure and positive versus
negative framing of the non-chosen alternative information.
50%
40%
30%
20%
10%
Low Need for Closure High Need for Closure
Need for Closure Personality Type
Positive
NCA Info
Framing
i Negative
NCA Info
Framing
Note. NCA Info = the non-chosen alternative information
When looking at need for closure scores as a continuous variable, linear
regression analyses determined that centered need for closure scores modeled with
the framing manipulations predicted ratings of information preferences, (/? = .27, p
< .05), such that as need for closure decreases, preference for reading the non-
chosen alternative information increases. Similarly, logistic regression analyses
determined that centered need for closure scores modeled with the framing
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
78
manipulations predicted information choice, ((3 = -.20, p < .04), such that as need
for closure decreases, the number of participants choosing to read the non-chosen
alternative information increases. No significant interactions were observed when
looking at need for closure as a continuous variable.
No significant differences were observed between low and high fear of
invalidity participants on rated information preferences (Fs < 1) or information
choice ((3 = .04, p = .621). Additionally, no significant differences were observed
between low and high preference for consistency participants on rated information
preferences (Fs < 1) or information choice (/S = .10, p = .413).
When looking at fear of invalidity scores and preference for consistency
scores as continuous variables, linear regression analyses determined that centered
fear of invalidity scores and centered preference for consistency scores each
modeled with the framing manipulations did not predict ratings of information
preferences, (|3 = .01, p = .972; ft = .18, p = .247, respectively). Not surprisingly,
logistic regression analyses determined that centered fear of invalidity scores and
centered preference for consistency scores each modeled with the framing
manipulations did not predict information choice, (/3 = .07, p = .242; (3 = .02, p =
.889, respectively).
Effects of Framing Manipulations on Emotions for Chosen Alternative and
Non-Chosen Alternative Information Choosers. The effects of the framing
manipulations on emotions observed in Study 3 did not replicate in this study, nor
were any other emotion differences detected in this study.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
79
Discussion
The main effects for the framing manipulations replicated from Studies 1, 2,
and 3, strongly affirming excellent reliability in their effects across studies.
Additionally, the distribution in the no framing control condition regarding
information choice and the magnitude of the spreading of alternatives measures
were all nearly identical to Studies 2 and 3, once again demonstrating good
reliability for the effects of having people choose between 2 highly preferred items,
as opposed to the 2 most highly preferred items.
The continuous measure of rated information preference performed nearly
identically to the actual information choice measure, suggesting that they are both
effective means of assessing information preferences.
The most interesting results in Study 4 are those found among the
personality measures. Unfortunately, no differences were detected for fear of
invalidity or preference for inconsistency. However, an interesting interaction
showed up when looking at the ANOVA for high versus low need for closure
personality type on people's rated preferences for information. Specifically, people
who are high in need for closure are not affected by the different framings of the
non-chosen alternative information in terms of their rated preferences of
information. However, people who are low in need for closure are affected by the
framing of the non-chosen alternative information. Specifically, these people are
significantly less likely to prefer the chosen alternative information if they receive a
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
80
positive framing of the non-chosen alternative information than if they had received
a negative framing of the non-chosen alternative information.
This finding suggests that the effectiveness of positively framing
information about a non-chosen alternative is largely dependent upon people’s
orientation on need for closure. Given that people’s orientation on need for closure
did not affect people’s choice of information in the control condition, it would
appear that low need for closure people require an appeal to think about the
potential value of information about a non-chosen alternative in order to actually
consider that information. With such an appeal, the preferences for and the
likelihood of choosing the non-chosen alternative information dramatically increase
for low need for closure people. Without the presence of framing manipulations,
people’s preferences and choices of information will follow the pattern predicted
by cognitive dissonance, regardless of need for closure orientation. Specifically,
people will predominantly prefer and choose information about a chosen
alternative.
The apparent lack of significant differences for change in emotion from pre
information choice to post-information choice was surprising given the differences
observed in Study 3. One possibility is that by giving people an opportunity to
specifically rate how much they prefer the chosen alternative information versus
the non-chosen alternative information, people feel that they have clearly stated
their position and are somewhat emotionally pacified as a consequence regardless
of which information they actually choose to read based on the framing
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
81
manipulations. Implementing the complete PANAS-X Affect schedule (Watson &
Clark, 1994) may yield more reliable significant differences in future studies.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
82
CHAPTER VI: GENERAL DISCUSSION
Cognitive dissonance is a powerful social psychological phenomenon. A
vast body of research across several paradigms has convincingly shown that people
do not like to experience cognitive dissonance and that the presence of dissonance
motivates people to engage in a variety of activities in the pursuit of reducing
dissonance (Harmon-Jones & Mills, 1999). A key ingredient of dissonance is
inconsistency. If two or more cognitions or actions are inconsistent with each
other, an individual will experience cognitive dissonance (Aronson, 1968). This
phenomenon has been found to be especially strong when the self-concept is
threatened (Greenwald & Ronis, 1978; Abelson, 1983).
However, other research in psychology has demonstrated preference for
inconsistency over consistency. Developmental psychology has discovered the
human attraction to novelty (things new and different from prior experience)
(Jeffrey & Cohen, 1971; Wetherford & Cohen, 1973; Pancratz & Cohen, 1970).
Research on human creativity shows us that it is exposure to novelty, thus
inconsistency, that sparks the creative impulse (Maddi, Charlens, Maddi, & Smith,
1962).
With the understanding that the experience of inconsistency is often a
positive experience with valuable consequences, one must consider the
implications this has for the effects observed in studies examining cognitive
dissonance. In the present research, the finding that people will predominantly
choose information about a chosen alternative was investigated. The concern
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
83
raised in the present research was that if information about a non-chosen alternative
offered a potential benefit to an individual, such an individual would not be able to
experience that benefit since the human tendency is to choose information about a
chosen alternative.
In all 4 studies, it was determined that it is possible in a classic cognitive
dissonance situation, in this case the “free-choice” and “selective exposure”
paradigms, to change people’s information preferences by positively or negatively
framing both information about a chosen alternative and information about a non-
chosen alternative. Given the effectiveness of relatively mild frames to alter
information preferences either by positively framing the non-chosen alternative
information or by negatively framing the chosen alternative information, the
concern raised in the present research is reasonably surmountable. Specifically,
when people could possibly benefit from information about a non-chosen
alternative, people’s likelihood of choosing that information can be increased by
either pointing out the potential value of that information or the lack of value of
information about a chosen alternative.
Furthermore, not only did the framing manipulations serve as a method for
increasing the likelihood of choosing information about a non-chosen alternative,
they also affected people’s levels of happiness and calmness regarding information
choice. These findings regarding affect must be considered cautiously, however,
since they did not replicate in Study 4.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
84
In Study 3, the main findings for regarding happiness were that people who
chose the non-chosen alternative information were significantly happier if they
received a positive framing, as opposed to a negative framing, of the non-chosen
alternative information. Additionally, the non-chosen alternative information
choosers who received a positive framing of the non-chosen alternative information
were just as happy as the chosen alternative information choosers who received
either piece of information framed positively. These findings suggest that should
people choose information about a non-chosen alternative, pointing out the value of
the non-chosen alternative information can buffer any reduction in positive affect
that would have been experienced had that positive framing not been there or had a
negative framing of that information been present. In sum, the reduction in positive
affect that is a predicted outcome of cognitive dissonance when choosing
information about a non-chosen alternative can be remedied through the use of
framing manipulations.
In terms of calmness, people who chose the chosen alternative information
were less calm about that choice if both pieces of information were framed
positively than if the chosen alternative information had been framed positively and
the non-chosen alternative information had been framed negatively. These findings
suggest that by pointing out the value of information about a non-chosen
alternative, the level of calmness typically associated with choosing information
that is about a chosen alternative with a choice is reduced. This could either be
because when people receive positive framings of both pieces of information, they
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
85
are uncertain as to which piece of information to choose or because the positive
framing of the non-chosen alternative information brings into question the relative
value of the chosen alternative information.
Looking at individual differences, people who are low on need for closure
appear to be very receptive to the notion that the non-chosen alternative
information can be highly valuable. These people are open-minded and concerned
about being good decision-makers (Neuberg, Judice, & West, 1997). Therefore, if
information is available that can possibly help low need for closure people to
improve their decision-making abilities, they are more likely to take advantage of
that opportunity than high need for closure people. These findings demonstrate that
mildly persuasive framings of information about a non-chosen alternative can affect
information choice among people low on need for closure. Perhaps stronger
framings of the non-chosen alternative information might affect information choice
among high need for closure people. An increase in strength might override high
need for closure people’s tendency to simply want an answer and stick with it.
Does the relationship of need for closure to other personality variables
provide greater insight into the nature of need for closure? When considering the
relevance of need for closure orientation to other personality characteristics,
Webster and Kruglanski (1994) point out several related constructs. They note that
people who are high in need for closure share traits with the authoritarian
personality in that these people have a greater tendency towards rigidity of thinking
and an intolerance of other people who defy conventional norms. Need for closure
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
86
orientation is also conceptually related to intolerance of ambiguity in that people
who are high in need for closure are also more likely to be intolerant of ambiguous
situations in which decisions are questionable. Webster and Kruglanski (1994) also
found that people who are high in need for closure are also more likely to be
impulsive because their desire for a quick answer precludes making decisions
carefully.
Also, need for closure orientation maps onto some of the traits associated
with dogmatism. Rokeach (1960) used this term to describe differences between
the open and closed mind. A high need for closure may promote closed
mindedness because openness to conflicting information could potentially impede
closure. Related to this is the concept of cognitive complexity, which refers to the
ability to consider several factors when making decisions. As cognitive complexity
increases, the need for closure should decrease because these people are oriented
towards thinking things through before committing to a decision (Webster &
Kruglanski, 1994).
In sum, the tendency to choose information about a chosen alternative as
observed in classic cognitive dissonance research can be altered through the use of
making appeals to the value of the non-chosen alternative information or the lack of
value of the chosen alternative information. Such manipulations can also improve
the way people feel about choosing information about a non-chosen alternative.
Finally, people who are low on need for closure are more likely to choose
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
87
information about a non-chosen alternative should such information be potentially
valuable than would high need for closure people.
To better our understanding of the findings observed in the preceding series
of studies, future research should attempt to clarify the motivational processes that
are affected by the framing of information in the “selective exposure” paradigm of
cognitive dissonance. Three possibilities seem most plausible.
One possibility is that the framing of information creates demand
characteristics, such that people want to choose information that they are being told
is the information that they should choose. The problem with this explanation is
that people overall have a tendency to still choose the chosen alternative
information over the non-chosen alternative information. Therefore, if the frames
are creating demand characteristics, they are not overwhelmingly demanding.
Additionally, one would think that high need for closure people would be as
susceptible to demand characteristics as low need for closure people. However,
high need for closure people did not differ in their information preferences based
on the framing of the non-chosen alternative information, while low need for
closure people did differ.
A second possibility is that the framing of information is serving as another
manipulation of cognitive dissonance. Specifically, in addition to the potential
dissonance experienced from choosing information about a non-chosen alternative,
is the potential dissonance of choosing information that is inconsistent with the
framing manipulations. Some people may let the potential dissonance between the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
88
chosen item and chosen information take the backseat to the potential dissonance
between the frames and the chosen information. Still, the differences observed for
need for closure orientation suggests that this is not the motivational process behind
the framing manipulations. Specifically, need for closure does not correlate with
preference for consistency and instead measures a separate phenomenon.
Therefore, in terms of information preferences, the fact that only low need for
closure people are affected by the framing of the non-chosen alternative
information, suggests that the motivational process fueled by the framing
manipulations is something other than a second manipulation of dissonance. If it
were a second manipulation of dissonance, the frames would likely affect high need
for closure people in the same way that they affect low need for closure people
because in the no frame control condition where dissonance is the only
motivational force at work, both high and low need for closure people were equal
in their preferences for information type.
This leads us to our third possibility which is that the motivational process
sparked by the presence of frames is one of utility. Specifically, the frames serve
as a manipulation of the potential value of the two pieces of information. To quote
Tversky and Kahneman (1981), “When faced with a choice, a rational-decision-
maker will prefer the prospect that offers the highest expected utility” (p. 453).
Therefore, low need for closure people should want to capitalize on that value and
try to improve their abilities to make good decisions. On the other hand, high need
for closure people would prefer to simply be settled with their earlier decision
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
89
regarding the item they chose and not question it by choosing information that does
not support that item.
While the third possibility seems the most plausible explanation for the
motivational processes at work due to the framing manipulations, additional
research could possibly further clarify this issue. One possibility would be to
implement other personality measures or to include the complete versions of the
measures used in the present research. To address the concern that the full versions
of these measures would significantly increase the number of questions to be
complete by participants, a new version of the preceding studies could exclude the
emotion measures to offset the higher number of personality measure items.
The research presented here is a beginning to uncovering the positive
effects of framing manipulations on information choice and the resultant changes in
affect within dissonance research. Additionally, the differences between high and
low need for closure people help to clarify the role of framing manipulations on
information choice within dissonance research. It is relieving to find that the
powerful effects of cognitive dissonance are not absolute in determining
information preferences. When information that does not support a prior choice is
presented as potentially valuable, some people (low need for closure) are receptive
to that information. Their decisions regarding what information they will consider
is not controlled by the forces of cognitive dissonance. This is not to say that there
is little hope for people high on need for closure, rather it may be that they need a
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
90
more persuasive argument or greater incentive to consider non-chosen alternative
information.
It would be interesting to see how the current findings regarding need for
closure could be applied to the decision making processes of liberal and
conservative political officials. As mentioned earlier, Webster and Kruglanski
(1994) point out how several constructs share a relationship with need for closure.
Given the steadfast position that the Republican Party has maintained concerning
the war in Iraq and their unwillingness to consider the possibility that their decision
to invade Iraq was wrong, one might suspect that they are high on need for closure
and therefore unlikely to be affected by appeals to consider information that does
not support their decision. On the other hand, given the repeated attempts of the
Democratic Party to encourage the Bush administration to consider the evidence
against their actions in Iraq, one might suspect that democrats are low on need for
closure and therefore very interested in considering information that suggests a
different course of action should have been taken regarding Iraq. Our results
suggest that the Democrats might be more responsive to appeals that point out the
value of information about the non-chosen alternative.
The framing manipulations in the present research were designed to be just
strong enough to cause a shifting of information preferences. Clearly, stronger
framing manipulations could be implemented and undoubtedly such manipulations
occur in the daily lives of high need for closure people, let alone all people for that
matter. Future research on this topic should design frames to further increase the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
91
number of people who are willing to consider information about a non-chosen
alternative, especially when considering the resiliency of high need for closure
people. Additionally, more sensitive measures of affect, such as the complete
PANAS-X Affect schedule (Watson & Clark, 1994) should be used to understand
the effects of frames on emotional changes that occur when choosing information
that is about a chosen alternative or about a non-chosen alternative.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
92
REFERENCES
Amabile, T. M. (1983). The social psychology of creativity. New York: Springer-
Verlag.
Abelson, R. P. (1983). Whatever became of consistency theory? Personality and
Social Psychology Bulletin. 9, 37-54.
Aronson, E. (1968). Dissonance theory: Progress and problems. In R.P. Abelson, E.
Aronson, W.J. McGuire, T.M. Newcomb, M.J. Rosenberg, & P.H.
Tannenbaum (Eds.), Theories of cognitive consistency: A sourcebook (pp.
5-27). Chicago: Rand McNally & Company.
Berlyne, D. E. (1960). Conflict, arousal and curiosity. New York: McGraw-Hill.
Brehm, J. W. (1956). Post-decision changes in the desirability of alternatives.
Journal of Abnormal and Social Psychology, 52, 384-389.
Cialdini, R. B., Trost, M. R., & Newsom, J.T. (1995). Preference for consistency:
The development of a valid measure and the discovery of surprising
behavioral implications. Journal of Personality & Social Psychology, 69,
318-328.
Ehrlich, D., Guttman, I., Schonbach, P., & Mills, J. (1957). Post-decision exposure
to relevant information. Journal of Abnormal and Social Psychology, 54,
98-102.
Festinger, L. (1957). A Theory of Cognitive Dissonance. Stanford, CA: Stanford
University Press.
Frey, D. (1982). Different levels of cognitive dissonance, information seeking, and
information avoidance. Journal of Personality and Social Psychology, 43,
1175-1183.
Frey, D. (1986). Recent research on selective exposure to information. In L.
Berkowitz (Ed.), Advances in Experimental Social Psychology (Vol. 19, pp.
41-80). New York: Academic Press.
Frey, D., & Rosch, M. (1984). Information seeking after decisions: The roles of
Novelty of information and decision reversibility. Personality and Social
Psychology Bulletin, 10, 91-98.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
93
Greenwald, A. G., & Ronis, D. L. (1978). Twenty years of cognitive dissonance:
Case study of the evolution of a theory. Psychological Review, 85, 53-57.
Harmon-Jones, E., & Mills, J., (1999). An introduction to cognitive dissonance
theory and an overview of current perspectives on the theory. In E.
Harmon-Jones, & J. Mills (Eds.), Cognitive dissonance: Progress on a
pivotal theory in social psychology (pp. 3-21). Washington, DC: American
Psychological Association.
Higgins, E. T. (1997). Beyond pleasure and pain. American Psychologist, 5 2 ,1280-
1300.
Higgins, E. T., Shah, J., & Friedman, R. (1997). Emotional responses to goal
attainment: Strength of regulatory focus as moderator. Journal of
Personality and Social Psychology, 72, 515-525.
Hunt, J. M. (1963). Piaget’s observations as a source of hypotheses concerning
motivation. Merrill-Palmer Quarterly, 9, 263-275.
Jeffrey, W. E., & Cohen, L. B. (1971). Habituation in the human infant. In H.
Reese (Ed.), Advances in child development and behavior (Vol. 6, pp. 63-
97). New York: Academic Press.
MacKinnon, D. W. (1962). The nature and nurture of creative talent. American
Psychologist, 17, 484-495.
Maddi, S. R., Charlens, A. M., Maddi, D. A., & Smith, A. J. (1962). Effects of
monotony and novelty on imaginative productions. Journal of Personality,
30,513-527.
Mills, J. (1968). Dissonance theory: Progress and problems. In R.P. Abelson, E.
Aronson, W.J. McGuire, T.M. Newcomb, M.J. Rosenberg, & P.H.
Tannenbaum (Eds.), Theories of cognitive consistency: A sourcebook (pp.
771-776). Chicago: Rand McNally & Company.
Neuberg, S. L., Judice, T. N., & West, S. G. (1997). What the Need for Closure
Scale measures and what it does not: Toward differentiating among related
epistemic motives. Journal of Personality & Social Psychology, 72, 1396-
1412.
Pancratz, C. N., & Cohen, L. B. (1970). Recovery of habituation in infants. Journal
of Experimental Child Psychology, 9, 208-216.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
94
Rokeach, M. (1960). The open and closed mind. Oxford, England: Basic Books.
Stein, M. I. (1974). Stimulating creativity (Vol. 1). New York: Academic Press.
Torrance, E. P. (1965). Rewarding creative behavior: Experiments in classroom
creativity. Englewood Cliffs, New Jersey: Prentice-Hall.
Tversky, A., & Kahneman, D. (1981). The framing of decisions and the psychology
of choice. Science, 211, 453-458.
Wallach, M. A. (1970). Creativity. In P. H. Mussen (Ed.), Carmichael’s Manual of
Child Psychology (Vol. 1). New York: Wiley.
Watson, D., & Clark, L. A. (1994). The PANAS-X: Manual for the Positive and
Negative Affect Schedule - Expanded Form. University of Iowa.
Webster, D. M., & Kruglanski, A. W. (1994). Individual differences in need for
cognitive closure. Journal of Personality & Social Psychology, 67, 1049-
1062.
Wentworth, N., & Witryol, S. L. (1983). Is variety the better part of novelty? The
Journal of Genetic Psychology, 142, 3-15.
Zuckerman, M. (1994). Behavioral expressions and biosocial bases of sensation
seeking. New York: Cambridge University Press.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Abstract (if available)
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
The role of quantitative novelty in self-focused attention: A figure/ground relationship
PDF
Children's and parents' exposure to community violence: Links with neighborhood disadvantage and parenting
PDF
Early adolescent drug use among multiethnic males: A prospective examination of the influences of psychological distress, relationship with family and school, law abidance, guilt and peer drug use
PDF
Displaced aggression as a function of target power
PDF
Attachment, accommodation, and the outcomes of romantic relationships
PDF
Adolescent exposure to alcohol advertising: A prospective extension of Strickland's model
PDF
Individual motivation loss in group settings: An exploratory study of the social -loafing phenomenon
PDF
Cognitive load, trigger salience, and the facilitation of an aggressive response in triggered displaced aggression
PDF
Collective efficacy, anxiety, creativity/innovation and work performance at the team level
PDF
Ageism and looksism among gay men
PDF
The effects of coherence -based reasoning on betting decisions
PDF
Integrating acceptance with imaginal exposure to core schema: Process and outcome in generalized anxiety disorder
PDF
Examining the differences in family functioning and meaning in life between Reform, Conservative, and Orthodox Jewish families
PDF
A daily diary approach to compare the accuracy of depressed and nondepressed participants' estimation of positive and negative mood: A test of the depressive realism hypothesis
PDF
Gender differences in symptom presentation of depression in primary care settings
PDF
Considering the impact of culture on the therapeutic relationship: A look at Latino school-age children in a special education setting
PDF
Factors influencing the interpretation of novel words as adjectives in 4-year-old Spanish -speaking children
PDF
Biological and genetic factors in childhood psychopathy: A twin study
PDF
Atypicality: Benefit and bane. Provocation, trigger, typicality, and the expression of aggression and generalization
PDF
Data-driven derivation of skills for autonomous humanoid agents
Asset Metadata
Creator
Snow, Chadwick J.
(author)
Core Title
Countering the effects of selective exposure
School
Graduate School
Degree
Doctor of Philosophy
Degree Program
Psychology
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
OAI-PMH Harvest,psychology, social
Language
English
Contributor
Digitized by ProQuest
(provenance)
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c16-616281
Unique identifier
UC11341309
Identifier
3220155.pdf (filename),usctheses-c16-616281 (legacy record id)
Legacy Identifier
3220155.pdf
Dmrecord
616281
Document Type
Dissertation
Rights
Snow, Chadwick J.
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA
Tags
psychology, social