Close
About
FAQ
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
/
Do you see what I see? Personality and cognitive factors affecting theory of mind or perspective taking
(USC Thesis Other)
Do you see what I see? Personality and cognitive factors affecting theory of mind or perspective taking
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
DO YOU SEE WHAT I SEE? PERSONALITY AND COGNITIVE FACTORS
AFFECTING THEORY OF MIND OR PERSPECTIVE TAKING
by
Gurveen Singh Chopra
A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(PSYCHOLOGY)
August 2012
Copyright 2012 Gurveen Singh Chopra
ii
Dedication
In the memory of my grandfather – a remarkable man who instilled in me the love
of reading and an unquenchable thirst for knowledge.
iii
Acknowledgements
I would like to thank my mentor, Dr. Stephen J. Read (Ph. D.), for providing
invaluable guidance and support throughout my academic career. It has been an honor
and privilege knowing and working with him. I am also extremely grateful to Dr. Lynn C.
Miller (Ph. D.), Dr. David A. Walsh (Ph. D.), Dr. Margaret L. McLaughlin (Ph. D.), and
Dr. Justin N. Wood (Ph.D.) for their insightful feedback during the progress of this
dissertation.
iv
Table of Contents
Dedication ii
Acknowledgements iii
List of Figures v
Abstract vi
Introduction 1
Egocentrism and ToM 5
Cultural Orientation/Empathy and ToM 9
Cognitive Load and ToM 12
Hypotheses 14
Chapter One: Study 1a 16
Methods 16
Results 26
Table 1: Reliability of personality measures across studies 30
Discussion 33
Chapter Two: Study 1b 36
Methods 36
Results 37
Discussion 42
Chapter Three: Study 2 44
Methods 44
Results 47
Discussion 53
Conclusion 56
Bibliography 61
Appendices
Appendix A: Individualism/Collectivism Scale 65
Appendix B: Interpersonal Reactivity Index 67
Appendix C: Adolescent Egocentrism Scale 69
Appendix D: Hypersensitive Narcissism Scale 70
Appendix E: Manipulation Check Questionnaire (Study 1a/b) 71
Appendix F: Manipulation Check Questionnaire (Study 2) 72
v
List of Figures
Figure 1: Example of a 4x4 Grid display 17
Figure 2: Study 1a: Interaction effect of Condition x Load for
Time_Drag_Drop 29
Figure 3: Study 1a: Interaction effect of Perspective Taking x
Condition for Time_Drag_Drop 31
Figure 4: Study 1a: Interaction effect of Egocentrism x Condition x
Load for #Critical Errors 32
Figure 5: Study 1b: Interaction effect of Empathic Concern x
Condition x Load for #Critical Errors 39
Figure 6: Study 1b: Interaction effect of Empathic Concern x
Condition x Load for Time_Critical 40
Figure 7: Study 1b: Interaction effect of Fantasy x Condition for
#Critical Errors 41
Figure 8: Study 2: Relational trials – Interaction effect of
Empathic Concern x Condition x Load for #Critical errors 49
Figure 9: Study 2: Relational trials – Interaction effect of
Narcissism x Load for #Critical Errors 51
Figure 10: Study 2: Ambiguous Trials (ToM condition) - Interaction
effect of Collectivism x Belief for Time_Drag_Drop 53
vi
Abstract
Prior research on adults’ perspective taking or theory of mind (ToM) ability (i.e.
attributing beliefs, goals, intentions, etc. to others) has indicated an egocentric bias during
decision making tasks (Keysar et al., 2000, 2003). Specifically, while partaking in a task
that requires taking the perspective of another while making decisions, people
occasionally fail in establishing a common ground (common beliefs/knowledge shared
between the parties) and tend to see things from only their point-of-view. Some studies
(Apperly et al., 2010) have suggested that trying to maintain a ToM of others can tax
one’s cognitive resources thereby causing errors in perspective taking tasks. Other
research has indicated the possible effects of personality (such as collectivism, Wu and
Keysar, 2007) on perspective taking ability. In this study, we attempted to study the
effects of low vs. high cognitive load and various personality factors (viz. empathy,
perspective taking, collectivism/individualism, and egocentrism/narcissism) on people’s
ToM ability. Based on the grid task experiments of Keysar et al., (2000, 2003) and
Apperly et al., (2010), we designed and implemented software to assess ToM. Across
various studies, our results indicated that adults do find a ToM task more difficult than a
non-ToM task. Having a high collectivistic orientation leads to faster responses whereas
high individualism and egocentrism leads to more errors. There was also an indication of
the moderating effect of cognitive load on personality. For instance, participants with
high empathy made fewer errors than those with lower empathy but only for the low load
condition. We discuss these and other findings, and also make suggestions for future
studies.
1
Introduction
Human beings are social creatures. Throughout human history we have seen that
the best possible chance of an individual to survive would be to live within a group. In
pre-historic times, early man lived in groups for a variety of different reasons, such as
warding off predators, hunting, gathering food, mating, etc. Like their ancestors, modern
humans have inherited this desire to belong and live with others, be it in terms of family,
work, friendship, etc. In every aspect of our daily lives we are constantly engaging in
interactions with different people. Needless to say, in order to effectively communicate
with others we need to understand a given situation from the other’s perspective and
make sure that both parties are on the same page i.e. know what the other person is
referring to.
Consider a simple example of two people trying to schedule a time and place for
lunch. If these individuals are colleagues/friends who usually have lunch together at a
particular place, then one could ask the other to meet him/her “there” for lunch at noon;
“there” being the place where they usually eat. In this case, the other person should be
able to correctly interpret what the friend is referring to (the restaurant would form a part
of the common ground between the two). On the other hand, if these two people usually
eat at different places, then “there” could refer to any place. In this case, one person is
clearly not thinking from his friend’s perspective. Thus, essentially, we need to have a
“theory of mind” and establish a “common ground” in order to effectively interact and
work with others. In the next few sections we will describe common ground and theory of
mind, and the relationship between the two.
2
According to Herbert Clark (1996), people engage in communication via a set of
joint activities and joint actions. A joint activity (e.g. chess) involves the participation of
at least two people who engage in joint actions (e.g. moving the chess pieces around) by
coordinating with each other. Such joint actions occur when people have to coordinate
with each other to solve coordination problems (Schelling, 1960). These problems are a
result of peoples’ common interests, goals and interdependency of actions. Thus, to
achieve one’s goals, one needs to coordinate one’s own actions into one or more joint
actions. For instance, in a game of chess the two players need to be aware of the position
of the various pieces, time elapsed after each move, turn order, etc. in order to effectively
engage in joint actions and, thus, progress the game. Essentially, the action of one player
is contingent upon the action of the other.
So we see that joint activities and joint actions become part of the common
ground of the people involved (e.g. the dynamically changing state of the chess board
forms the common ground for the two players). We can think of common ground as the
“sum of peoples’ mutual, common, or joint knowledge, beliefs, and suppositions” (Clark,
1996). There can be different types of common ground such as communal common
ground wherein people have mutual beliefs that at least one person engaging in a specific
activity represents a particular community (e.g. doctors, students, Americans, etc.).
Basically, we are categorizing people as being representative of certain cultural groups,
professions, political ideologies, social networks, etc. Based on this categorization we can
then make assumptions about such things as their beliefs and values. Then there is also
3
personal common ground, which is based on joint perceptual experiences (e.g. watching
a movie together) and joint actions (e.g. discussing each other’s opinions of the movie).
As the common ground is continuously evolving, it can be subdivided into:
• Initial common ground: This is composed of the “set of background facts,
assumptions, and beliefs” that each participant engaging in the joint activity
already presupposes or is aware of (Clark, 1996). For example, the rules, etiquette
and strategies of playing chess should be a part of the players’ knowledge base
prior to commencing the game.
• Current state of the joint activity: This is what participants believe/assume to be
the state of the activity at a specific moment in time. For instance, the state of the
chess board after 5 moves.
• Public events so far: These comprise of events that the participants believe have
taken place in public that eventually lead to the current state of the joint activity
(e.g. record of the players’ moves so far).
Common ground between two (or more) individuals can be established by using
theory of mind (ToM) to develop a model of the other person’s mind, which would allow
us to infer various aspects of the other’s mental states. ToM can be defined as one’s
ability to attribute mental states like beliefs, desires, goals, knowledge, etc. to others and
acknowledge that others’ beliefs and intentions may differ from our own (Premack &
Woodruff, 1978). Basically, ToM involves putting yourself in the other person’s shoes
and attempting to think about and view a particular situation from his/her perspective.
Each day we constantly engage in social interactions with others and, either consciously
4
or sub-consciously, we make observations and inferences about others’ mental states. Via
these observations/inferences we try to determine the other’s underlying motivations,
predict possible future actions of the other person, etc. Seeing things from another’s
perspective can shed new light in situations wherein group members have to solve a
specific problem, and can aid problem solving. For instance, if two people negotiating the
terms of a business contract can accurately form mental models of one another, then they
may be more likely to reach a mutually agreeable decision (Thompson & Loewenstein,
2003).
An inability to effectively form a ToM of others can result in people performing
poorly in social interactions as they are unable to gauge and interpret the other’s actions,
facial expressions, emotions, etc. This is specially the case for those suffering from
autism (Langdon, 2003). People with autism are unable to interpret emotions (such as
anger, fear, sadness, etc.) from facial expressions and, thus, may not perceive signs of
potential danger, and may misconstrue what the other person is trying to say (Ozonoff,
Pennington, & Rogers, 1990). Hence, we see that a sound ToM can play a critical role in
facilitating social situations and make one more competitive in achieving one’s goals
(e.g. a salesman with a good understanding of ToM might be more adept in
manipulating/cajoling a consumer and be successful in selling his product).
Most of the work on ToM comes from the developmental psychology literature
focused on the ToM capabilities of children. Research has shown that ToM ability
typically develops in children around the ages of 4-5 years (Wimmer & Perner, 1983).
Some recent studies have also indicated the development of ToM processes in even
5
younger children (Onishi & Baillargeon, 2005). The common view on ToM is that once
formed in childhood, it becomes a relatively automatic process during adulthood.
However, other researchers (e.g., Apperly et al., 2010; Keysar et al., 2000, 2003) have
argued that using ToM is usually not an automatic process and that different factors can
affect its successful implementation, especially in the case of adults.
Some studies in children have indicated that ToM usage is dependent on one’s
cognitive resources or executive function (i.e. processes involving planning,
coordination, etc.) (Perner & Lang, 1999). Researchers like Apperly et al., (2010) have
indicated that this could also be the case in adults. Other researchers have found ToM
being affected by individual differences in personality (Keysar et al., 2000, 2003; Wu and
Keysar, 2007).
In this paper, we shall be discussing some personality factors such as cultural
orientation (individualism vs. collectivism), empathy, and egocentrism/narcissism that
have been known to affect perspective taking or ToM. Moreover, we will also look into
the function of cognitive load as a potential moderator of one’s ToM capability. Even
though some researchers have started looking at the role of personality and cognitive
factors in ToM, there is still a lot of work to be done in this domain. We propose studies
that specifically investigate the effects of personality and cognitive capacity on ToM.
Egocentrism and ToM
Research has shown that children start developing ToM abilities between the ages
of 4-5 years. Much of this work has focused on establishing when children first develop
the ability to represent a false belief; that is, when the child can represent that someone
6
else has a belief about the world that is false, given what the child knows about the world.
For example, a study by Wimmer and Perner (1983) investigated the effect of presenting
young children with situations involving agents/people with false beliefs. The children
were presented with a situation wherein they observe the following. Another child (say,
child_x) places a toy in a box with a lid and leaves the room. The child observer then sees
an adult remove that toy from the box and place it in a cupboard. The researchers were
interested in observing where the children observing the interaction thought that child_x
would look for his/her toy upon returning. Three-year-olds falsely held the belief that
child_x would look for the toy in the cupboard, thus projecting their own knowledge
about the actual location of the toy onto child_x. However, 4-to-5-year-olds correctly
predicted that child_x would follow his own belief and look for the toy in the box where
he had originally left it; they were able to separate their own beliefs from those of
child_x.
We can think of this inability to separate one’s own viewpoint from another
person’s (or lacking ToM) in terms of egocentrism – the tendency to prefer one’s own
privileged viewpoint to that of another who does not share the same privileged
information (Royzman, Cassidy, & Baron, 2003). In order to avoid an egocentric point-
of-view one must frequently make a conscious effort to become aware of the common
ground between the two (or more) parties, and try and put aside the knowledge that only
one person has access to. Several studies have demonstrated that adults are prone to
making egocentric errors while attempting to process information that only they are privy
to and not the other party (Keysar, Ginzel, and Bazerman, 1995; Camerer et al., 1989).
7
One experiment (Keysar, 1994), investigating egocentrism in adults, involved
participants reading scenarios about a recommendation for a restaurant. In the scenario,
one person (say John) asked his friend (say Jim) to recommend a restaurant to which he
could take his parents. After John visits the recommended restaurant, he writes a
comment to Jim: “You wanted to know about the restaurant, well, marvelous, just
marvelous” (Keysar, 1994, p.173). Now, some participants were given privileged
information that on the day John visited the restaurant the service and food were terrible,
while other participants were given privileged information that the food and service were
great. This information was not available to Jim (in the scenario). Thus, John’s comment
to Jim could be interpreted as sincere or sarcastic depending on the privileged
information. For the experiment, participants had to ascertain whether Jim would
interpret the comment as sarcastic. The results indicated that participants in the negative
dining experience condition were more likely to take an egocentric approach and judge
that the comment would be interpreted as sarcastic by Jim. Thus, participants were
egocentrically using the privileged information they possessed.
Other research investigating perspective taking/ToM and egocentrism points in
the same direction (i.e. people frequently make egocentric decisions). Various
experiments have been performed to test adults’ ToM abilities in situations wherein two
people are engaged in a joint activity but the perceptual common ground is different for
each person; to effectively perform that joint task one person would have to take the
perspective of another. One such study by Keysar et al. (2000) delved into this issue of
disparate common ground between two participants. In this study, participants were made
8
to sit in front of a table which had a 4x4 grid on top and had to move objects within the
grid. There was a director (supposedly, another participant) sitting on the other side. The
experimenter informed the subject that the director would be giving verbal instructions
asking him to move some of the objects within the grid slots. The participant was
informed that the director would only be able to view certain objects, as some of the grid
slots were occluded from the director’s point-of-view; the participant could view all
objects. For instance, say the grid had three candles – small, medium, and large. Now, the
smallest candle was occluded from the director’s perspective but not from the
participant’s. So, from the director’s perspective, the smallest candle present in the grid
would be the medium one. It was observed that when the director instructed the
participant to move the smallest candle to the slot below, the participant would frequently
fail to take into account the director’s belief about the smallest candle and reach out for
the smallest (incorrect) candle and not the medium (correct) one. This pattern was
replicated across a number of trials with other objects.
Another study by Keysar et al. (2003) provides further evidence that adults
frequently fail at reliably utilizing ToM. This experiment involved a similar 4x4 grid with
certain slots occluded only from the director’s view. In this case, the participant was
asked to hide an object (e.g. a roll of sticking tape) in a paper bag; the director had no
knowledge of the bag’s contents. If participants were to use their ToM, the object in the
bag should never be thought of as the one referred to by the director. The grid had other
objects in its slots including a cassette tape. When the director asked the participant to
move the tape, he meant the cassette tape and not the sticking tape hidden in the bag.
9
However, most participants initially attempted to reach for the tape in the bag
demonstrating their failure at taking the director’s perspective into account and taking an
egocentric approach instead. These participants failed to incorporate the common ground
shared between the director and themselves.
Though the aforementioned studies attempt to give an egocentric explanation for
ToM errors, they have not explicitly looked into individual differences in egocentrism,
i.e. whether some people are more egocentric than others to begin with, and how that
affects ToM. From such experiments it would seem plausible to infer that highly
egocentric individuals would be more likely to make more ToM errors than those who are
low in egocentrism. We will be testing this hypothesis in our study by administering a
ToM task to groups of people who are either high or low in egocentrism. Although there
is a literature on individual differences in egocentrism (e.g., Royzman, Cassidy, & Baron,
2003), this work has not been previously related to ToM and common ground.
Besides egocentrism, we also wish to look at the effect of narcissism on ToM. Previous
research has linked narcissism to egocentrism, inability to empathize with others, and a
general lack of perspective taking (Fiscalini, 1994; Watson et al., 1984). We would like
to see whether people who score high on a measure of narcissism are more prone to
making egocentric decisions even though the correct outcome would require establishing
a common ground between participants and having a ToM of all those involved.
Cultural Orientation/Empathy and ToM
Research suggests that culture can play an important role in perspective taking.
Individualistic (Western) cultures usually place more emphasis on the self and a striving
10
for personal independence/autonomy, whereas collectivistic (Eastern) cultures de-
emphasize the individual and focus more on the other and group interdependence
(Markus & Kitayama, 1991). A grid experiment, similar to the aforementioned ones,
highlighted the difference between perspective taking/ToM across cultures (Wu &
Keysar, 2007). Chinese students, when asked to move an object that could potentially be
occluded from the director’s point-of-view, were more likely to take the director’s
perspective into account than their American counterparts and chose the intended object.
The authors have argued that this was because the Chinese students were more
collectivistic (emphasizing interdependence and focusing on others) and thus, better at
perspective taking than the American students who were more individualistic
(emphasizing independence and focusing on the self).
Perspective taking is also an important aspect of individual differences in
empathy. We can think of empathy as the ability to understand the thoughts and
feelings/emotions of others. Most researchers believe that it comprises of three
components: (1) an affective response to another that might include sharing the other’s
emotional state; (2) a cognitive capacity to take the other’s perspective into account; and
(3) regulatory mechanisms keeping track of the origin of the feelings (self and other)
(Davis, 1996; Ickes, 2003; Decety & Jackson, 2004). Basically, empathy requires that
one has the ability to share/feel others’ emotions (affective component) and an ability to
understand another’s experience (cognitive component). Pro-social behaviors (like
altruism, helping, etc.) are generally dependent on a person’s ability to understand and
11
experience the needs and emotions of another from the other’s perspective (Nichols,
2001; Davis & Stone, 2003).
Studies manipulating perspective taking, wherein subjects are asked to imagine
what the other person feels (via scenario descriptions, viewing confederates experiencing
supposed distress, etc.), elicit greater empathic responses by the participants (Heinke &
Louis, 2009; Batson et al., 1997). In one study (Stotland, 1969) subjects were made to
watch a confederate whose hand was attached to a machine that supposedly induced
extreme heat. When these participants were asked to try and imagine what that person
might be feeling, they experienced physiological changes (viz. palm sweating and vaso-
constriction) implying that participants were capable of placing themselves in the other
person’s shoes and experience the “fake” distress of the confederate. These and other
studies indicate that such perspective taking situational manipulations of empathy (i.e.
asking participants to imagine being in the other’s situation) do elicit an empathic
response from the participants towards the “victim”. However, not much research has
been done linking individual differences in empathy to ToM. In our study we will be
explicitly looking at individual differences in perspective taking and empathy, and their
relation to ToM.
Collectivism has also been linked to empathy in that individuals with high
collectivistic values tend to be more empathic towards others than those with a more
individualistic social orientation (Heinke & Louis, 2009; Triandis, 1995). As
aforementioned, collectivistic cultures are more interdependent and give more importance
to the group over the individual. If people with a collectivistic orientation lacked the
12
ability to empathize with other group members they would probably find it difficult to
accomplish group-oriented goals. In one study wherein participants were asked to read a
scenario depicting someone in distress, it was found that those participants who displayed
more collectivistic values showed greater empathy for the protagonist than those who had
a pre-dominant individualistic orientation (Duan, Wei, & Wang, 2008).
So far we have mentioned different personality factors (viz. cultural orientation,
empathy, and egocentrism) that can affect perspective taking. In our study, we will be
investigating the influence of these factors on individual differences in ToM.
Specifically, we will be administering scales that measure individual differences in
empathy and perspective taking (IRI scale – Davis, 1980), egocentrism (AES scale –
Enright et al., 1980), narcissism (HSNS scale – Hendin & Cheek, 1997), and cultural
orientation (HVIC scale – Triandis & Gelfand, 1998). Additionally, participants will also
have to perform a task assessing their ToM ability.
Cognitive Load and ToM
Besides factors such as empathy, cultural orientation and egocentrism/narcissism,
ToM can also be influenced by cognitive mechanisms. Some researchers have argued that
ToM is dependent on an individual’s executive functioning and that if executive
functioning is taxed then ToM has a greater chance of being overcome by one’s
egocentric bias (Apperly et al., 2008; Samson et al., 2005). Simply put, when people have
to process a lot of information they may resort to relying on their own perspective while
making decisions. Executive function refers to a set of cognitive processes that are
involved in working memory, response inhibition, resistance to interference, and
13
planning (Hughes, 2002; Shallice & Burgess, 1996). It has been proposed that deploying
ToM depends on these abilities and that to the extent that they are strained, ToM will get
degraded (Apperly et al., 2010). However, little work has explicitly examined this
possibility, especially in adults. Thus, more research is needed in order to establish a
significant link between high cognitive load and ToM degradation.
In one study, Apperly et al. (2010) created a software version of the Keysar
(2000, 2003) grid task wherein participants were given pre-recorded audio instructions
(such as, moving an object to another slot in the grid) and had to indicate their selection
by clicking on an object and then clicking on the grid space/slot where that object was
supposed to be placed. The objects themselves were immovable and the grid itself was
static; it recorded mouse-click coordinates and the time it took for participants to carry-
out each instruction. There were two conditions – a ToM condition wherein participants
were informed that certain objects were occluded from the director’s point-of-view and
that they should incorporate that perspective when responding, and a non-ToM condition
wherein the participants were simply asked to avoid placing any objects in the dark-grey
(occluded) grid slots. The results indicated that participants were more prone to making
errors in the ToM condition (by taking egocentric actions on the director’s instructions)
than the non-ToM condition. The researchers believe that this significant difference
between the two conditions is because the ToM task was more cognitively demanding in
that the participants had to actively try and keep the director’s perspective in their minds.
It should be noted that cognitive load was not independently manipulated in this
experiment and that though the researchers believed that cognitive demand in doing the
14
task led to the ToM errors, this possibility was not systematically tested. We propose to
explicitly manipulate cognitive load by having participants perform a task requiring
executive function while simultaneously performing another task assessing ToM.
Hypotheses
In this paper, we wish to test the possible impact of both personality and cognitive
factors on people’s perspective taking/ToM and try to better identify the conditions under
which adults succeed and fail at this task. There has not been much work in terms of
personality and cognitive load explicitly affecting ToM, and we hope that our research
will provide greater insight into the mechanisms underlying ToM/establishing common
ground. The central task we will use for this is a software version of the Keysar (2000,
2003) grid task, similar to the one done by Apperly et al. (2010) but with more
interactivity and enhancements. We will be conducting two preliminary studies to test the
effectiveness of our ToM and cognitive load tasks, and a final study incorporating any
changes (in procedure, design, etc.) resulting from the first two studies.
The following hypotheses will be investigated in all studies:
• Hypothesis 1: Participants in the ToM condition (wherein they have to keep the
director’s perspective in mind) will perform worse (in terms of longer reaction
times and more errors) than those in a non-ToM control condition (avoiding
blacked-out grid slots). These results should corroborate the findings of Keysar et
al. (2000, 2003) and Apperly et al. (2010).
• Hypothesis 2: Participants with a high empathic ability should be better at ToM
tasks than those who have lower empathy. Such individuals should be better at
15
identifying and taking the emotional and cognitive perspective of the other. Thus,
we hypothesize that those who score higher on an individual difference measure
of empathy (Davis, 1980) should also fare better in ToM tasks.
• Hypothesis 3: As collectivism implies group interdependence and placing the
group above the individual, people with this orientation are better able to
empathize and take the perspective of others. Thus, we hypothesize that
participants with a pre-dominant collectivistic social orientation should show
greater empathy and better performance in the ToM task than those with a more
individualistic orientation.
• Hypothesis 4: Keysar’s (2000, 2003) and other studies have indicated that people
often take an egocentric view when interpreting others’ actions, i.e. they fail to
keep track of the common ground between themselves and others. We predict that
participants measuring high on egocentrism and narcissism should have greater
difficulty processing the other’s perspective and, thus, perform more poorly on
ToM tasks.
• Hypothesis 5: Participants experiencing high cognitive load should find it difficult
to keep the other person’s perspective in mind while performing the ToM task.
Specifically, participants when asked to memorize a 7-character (high cognitive
load) string of letters while simultaneously trying to keep the other’s perspective
in focus would have longer reaction times and make more errors than those who
had to memorize a 2-character (low cognitive load) string.
16
Chapter One: Study 1a
This was a pilot study done to test the plausibility of our experimental design.
Specifically, we were interested in learning whether our manipulation was effective,
whether participants had any difficulty following instructions, was the study too easy or
too difficult, etc.
Methods
Participants
We recruited a total of 72 participants (males = 15, females = 57) from the
university subject pool. These participants were all undergraduate students at the
University of Southern California (USC) and were awarded course credit for their
participation. They were all at least 18 years old at the time of their participation and the
average age was 20 years.
Design and Apparatus
A software version of the Keysar (2000, 2003) and Apperley et al. (2010) grid
task was created using Adobe Flash as the development platform and Action Script 3.0 as
the default programming language. The main stimulus was a 4x4 grid with 16 slots (see
figure 1). Five of these grid slots were occluded (as shown in black) to indicate that
certain slots (and objects within those slots) were not visible from the director’s point-of-
view. This pattern of occluded slots remained fixed across 8 experimental grids.
Each grid consisted of six images of various objects (e.g. types of balls, glasses, candles,
etc.) that were placed within different grid slots; these objects were unique to each grid
and were not repeated in other grids. We used ambiguous objects/grids in this study. An
17
example of an ambiguous object/grid would be two balls in the grid – a tennis ball in an
occluded slot that was not visible to the director (but visible to the participant), and a
soccer ball in an unoccluded slot that was visible to both. Thus, if the director asked the
participant to move the ball s/he would be referring to the soccer ball and not the tennis
ball.
Figure 1. Example of a 4x4 Grid display
18
Also, all objects were interactive in that they could be clicked, dragged and
dropped in any slot within a grid. We recorded a wide variety of information from the
mouse movements, such as mouse coordinates when an object was clicked, where it was
dropped, type of object the mouse cursor was hovering over, location of the mouse within
the grid (column1 row1, etc.), and time (in milliseconds) between clicking an object and
dropping it. Essentially, we could track the mouse cursor as it moved within the grid.
Below the grid there was a text chat box that served as the interface for giving written
instructions to the participants. There were four instructions in each grid with an interval
of 8 seconds between each instruction. The first three instructions consisted of one
critical instruction (asking participants to move an object, like a car, one of which was
occluded from the director’s perspective while the other was visible to both; to make a
correct choice the participant would have had to keep in mind what the director could and
could not see), and two neutral instructions that simply asked the participant to move a
neutral object that both could see. The critical instruction appeared randomly in the first,
second, or third position across grids. All of these instructions asked the participant to
move an object either 1 cell up/down/left/right from that object’s current position, or to
the rightmost/leftmost/bottommost/topmost slot within the same row/column of the
object’s location. The fourth –and final– instruction informed participants that in the next
screen they would be asked to recognize whether the letter that appeared there was also
present in a string that they had previously seen. This was part of the cognitive load task.
We manipulated cognitive load by having participants look at a string of letters
before beginning an experimental grid. There were 2 letters for the low load condition
19
and 7 letters for the high load condition. These letters were selected at random without
replacement from the following set: {F, G, K, N, P, Q, R, S, T, X, Y, Z}, and appeared
sequentially in the center of the screen at an interval of 1 second (Wood, 2011).
Following each experimental grid, participants were presented with another screen that
displayed a letter and asked them to click either a “yes” or “no” button to indicate
whether they had seen that letter before.
In the ToM condition, before the load and experimental tasks, participants were
shown a screen with two grid pictures adjacent to one another. One was a picture of the
experimental grid from the participant’s perspective, and the other was a picture of the
same grid but from the director’s perspective. This was done so that the participant would
have a clear idea of what objects were visible to him/her but not visible to the director
and vice versa. The picture appeared on the screen for 10 seconds. For the non-ToM
condition, there were no pictures shown as there was no mention of a director.
Thus, for the ToM condition, a trial consisted of the following sequence:
1. Pictures of the two grids are displayed.
2. Participant is presented with a letter string (either a low or high load).
3. Experimental Grid with instructions
4. Recognition task
And for the non-ToM condition, a trial consisted of the following sequence:
1. Participant is presented with a letter string (either a low or high load).
2. Experimental Grid with instructions
3. Recognition task
20
There were 8 such trials across the ToM/non-ToM condition. Each trial began
immediately after the previous trial ended.
A 3.2 GHz Windows 7 computer was used to run the simulation. All responses
were stored in the form of text files in the computer. The study was conducted in a
laboratory within USC’s Department of Psychology.
Measures
Individualism/Collectivism
Individualistic vs. collectivistic orientation was assessed by using the Horizontal
& Vertical Individualism & Collectivism II scale (see Appendix A) by Triandis and
Gelfand (1998). Responses are made on a 9-point Likert-type scale ranging from 1
(never) to 9 (always). Horizontal orientation emphasizes equality (wherein one is thought
to be similar to others) while vertical orientation (wherein one is different from others)
emphasizes hierarchy. The scale consisted of 27 items with 4 sub-scales:
• Horizontal Individualism (HI – 5 items): This measures the extent to which
individuals strive to be distinct without desiring special status (e.g. “I'd rather
depend on myself than others”).
• Horizontal Collectivism (HC – 8 items): This measures the extent to which
individuals emphasize interdependence but avoid easily submitting to authority
(e.g. “If a coworker gets a prize, I would feel proud”).
• Vertical Individualism (VI – 8 items): This measures the extent to which
individuals strive to be distinct and desire special status (e.g. “It is important that I
do my job better than others”).
21
• Vertical Collectivism (VI – 6 items): This measures the extent to which
individuals emphasize interdependence with in-group members and competition
with out-groups (e.g. “Parents and children must stay together as much as
possible”).
We will be treating HI and VI together as a measure of general Individualism, and
HC and VC together as a measure of general Collectivism.
Empathy/Perspective taking
Empathy and perspective taking were assessed by the Interpersonal Reactivity
Index (IRI) (see Appendix B) (Davis, 1980). Responses are made on a 5-point Likert-type
scale ranging from 1 (does not describe me well) to 5 (describes me very well). The scale
consisted of 28 items with 4 subscales (7 items each):
• Fantasy: This measures the extent to which individuals identify with fictional
characters (e.g. “When I am reading an interesting story or novel, I imagine how I
would feel if the events in the story were happening to me”).
• Perspective taking: This measures the extent to which individuals spontaneously
try to adopt another’s point of view (e.g. “Before criticizing somebody, I try to
imagine how I would feel if I were in their place”).
• Empathic Concern: This measures the extent of one’s feelings of warmth,
compassion, and concern for others (e.g. “When I see someone being taken
advantage of, I feel kind of protective towards them”).
22
• Personal distress: This measures the extent of one’s feelings of anxiety and
discomfort caused by another's negative experience (e.g. “When I see someone
who badly needs help in an emergency, I go to pieces”).
Egocentrism
Egocentrism was assessed with the Adolescent Egocentrism Scale (AES) (see
Appendix C) (Enright et al., 1980). Even though this scale was developed specifically for
adolescents, research has indicated that it can be used to test egocentrism in adults as well
with a reasonable degree of reliability and validity (Frankenberger, 2000). There were a
total of 15 items responded to on a 5-point Likert-type scale ranging from 1 (no
importance) to 5 (very important) and the scale is composed of three sub-scales:
• Personal Fable: This measures one’s belief about the self being special and
unique (e.g. “Accepting the fact that others don't know what it’s like being me”).
• Imaginary Audience: This measures one’s belief about the extent to which others
might be paying attention to them (e.g. “Being able to daydream about great
successes and thinking of other people’s reactions”).
• Self-focus: This measures the extent to which one’s thoughts are directed inwards
toward the self, instead of being directed outwards toward other groups (e.g.
“Becoming real good at being able to think through my own thoughts”).
For our analysis, we will use these sub-scales together as our measure of
egocentrism.
23
Narcissism
Narcissism was measured by the Hypersensitive Narcissism Scale (HSNS) (see
Appendix D) developed by Hendin and Cheek (1997). This scale consists of 10 items and
uses a 5-point Likert-type scale ranging from 1 (strongly disagree) to 5 (strongly agree).
Some of the sample items in this scale are: “I often interpret the remarks of others in a
personal way”, “I dislike sharing the credit of an achievement with others”, etc.
Manipulation Check
At the end of all these aforementioned personality surveys we performed a
manipulation check to ascertain whether participants who were assigned to the ToM
condition actually believed that the instructions were coming from the experimenter in
the adjacent room. This check entailed a series of progressively focused questions (see
appendix E) that asked the participants about their experience with the task. For instance,
the questions started-out with general questions like: “how easy was it to follow my
instructions?”, “how easy was it to move objects around the grid?”, etc. Participants rated
these on a 5-point Likert-type scale ranging from 1 (very easy) to 5 (very difficult).
Subsequent questions then started to ask them about their belief about the source of the
instructions: “were you, at any time, wondering about the source of the instructions?”,
“did you believe that I was giving you the instructions?”, etc. These were again rated on a
5-point Likert-type scale ranging from 1 (definitely yes) to 5 (definitely not). The
subjects could also write open ended comments following each question. This was done
to give them the opportunity to voice their opinion about their belief or disbelief earlier
24
rather than later. We figured that the sooner a subject stated whether they believed (or not
believed) the easier it would be to classify them as believers or non-believers.
There was no manipulation check for the non-ToM condition as participants
assigned to this condition were not asked to follow instructions by explicitly keeping the
experimenter’s perspective in mind.
All personality measures and the manipulation check were administered after the
end of all experimental trials.
Procedure
Upon the participant’s arrival in the lab, the experimenter randomly assigned
him/her to a ToM/non-ToM and Low/High load condition. The participant was asked to
sit in front of a computer and read/sign the informed consent form after which the
experimenter started to explain the procedure.
In the ToM condition, participants were told that they would be participating in a
team performance study and that they would be asked to follow instructions given to
them by the experimenter (serving as the director) who would be in the adjacent room
when the trials began. A picture of a demo grid was then shown and the experimenter
informed the participant that s/he would be asked to move various objects within the grid
with the mouse as per the instructions that would appear in a chat-box below the grid.
The experimenter would be “typing” these instructions from his/her computer and
sending them to the participant. It was explicitly mentioned to the participant that the
director would be looking at a mirror image of the grids that the participant would see;
the black grid slots indicated that not all objects could be viewed from the director’s
25
perspective. This point that the participant had to always keep the director’s point-of-
view in mind while moving objects was repeatedly emphasized during the course of
explaining the procedure. The participant was always shown a picture displaying the
experimental grid from both his/her perspective and from the director’s perspective. This
allowed him/her to compare the two grids and see which objects were not visible to the
director. The experimenter also stated that following each picture-view, a series of letters
would appear and that the participant would be asked to recall one of them later-on
during the experiment.
After this explanation, the participant was asked to go through two practice trials
– each trial consisted of the aforementioned sequence (viz. picture view, letters,
experimental grid, and recognition task). For these practice trials, the experimenter was
sitting next to the participant and giving verbal instructions. This was done so that any
errors made by the participant could be corrected.
At the end of the practice trials, the experimenter started the experimental trials
and went to the other room. The first screen that the participant saw was a “connection”
screen indicating that the computer was communicating with the director’s computer and
establishing a network. We did this to make the experiment seem more realistic to the
participants and to try to get them to believe that the instructions would be coming from
the director. In reality, though the experimenter was in the adjacent room, all instructions
were coming automatically from the computer and the two computers were not connected
in any way. After this connection screen, the experimental trials began. There was no
contact between the participant and director during the course of the experiment.
26
Once the final trial was over, the participant was asked to click a button and
complete the various measures described above. At the end of the last survey, the
participant was asked to contact the experimenter who then proceeded with the
debriefing.
For the non-ToM condition, the procedure was similar except that it was
explicitly mentioned to the participants that they would be taking instructions from a
computer, and that they were to ignore objects in the black slots and to avoid moving any
object onto a black slot. The sequence of trials for the participants consisted of the load
string, the experimental grid, and the recall task. Like before, there were 8 such trials at
the end of which the participant was asked to complete surveys.
The ToM condition took about 16-18 minutes while the non-ToM condition took about
12-14 minutes.
Results
The performance of participants in the experiment was determined in terms of
times and number of errors. We ended-up with the following six dependent variables
(DVs):
1. Time_Neutral: This was the average time (ms) taken by a participant to click
‘any’ object following a neutral (not critical) instruction. There were 2 neutral
instructions per grid. We calculated this “neutral” time by measuring the
difference between the time it took for the participant to respond (i.e. click an
object) after a neutral instruction had appeared.
27
2. Time_Critical: This was the time (ms) taken by a participant to click ‘any’ object
following a critical (ToM) instruction. There was 1 critical instruction per grid.
We calculated this “critical” time by measuring the difference between the time it
took for the participant to respond (i.e. click an object) after a critical instruction
had appeared.
3. Time_Drag_Drop: This was the average time (ms) taken by a participant to drag-
and-drop ‘any’ object after receiving an instruction (neutral or critical). There
were 3 instructions (2 neutral + 1 critical) per grid. We calculated this
“drag_drop” time by measuring the difference between the time when the
participants first clicked on an object and when they dropped it onto a grid slot.
4. #Neutral Errors: These were the errors made when a participant clicked/moved
‘any’ incorrect object following a neutral instruction, or failed to respond to
(missed) a neutral instruction.
5. #Critical Errors: These were the errors made when a participant clicked/moved a
“competitor” object (i.e. an object that was not being referred to by the director)
following a critical instruction.
6. #Cognitive load errors: These were the errors made when a participant failed to
recognize a letter from the string of letters shown previously during the cognitive
load task.
Central Analyses
As the error variables were “count” data and also turned-out to be highly
positively skewed, we decided to log transform these variables. Thus, for all ensuing
28
analyses with the three errors variables we used their log
10
(variable + 1) transformed
values. A ‘1’ was added to the log values to ensure that we got a valid log. An alpha level
of .05 was used for all statistical tests.
We performed a Condition (non-ToM vs. ToM) x Load (Low vs. High) ANOVA
with condition and load as between subjects factors. The distribution of the participants
was as follows: ToM (n=34), non-ToM (n=38), low load (n=40), and high load (n=32).
We found a significant main effect for condition, F(1,68)=5.68, p=.02, in that for
Time_Neutral, the response time for participants was greater in the ToM condition
(M=3706.02, SD=544.59) than the non-ToM condition (M=3320.40, SD=720.54). Thus,
participants were taking more time to respond to neutral instructions in the ToM
condition than the non-ToM condition supporting our hypothesis. There were no
significant main effects of condition on Time_Critical and Time_Drag_Drop.
There were additional main effects for load. Counter to our hypothesis,
participants assigned to the low load condition took a significantly longer time
(M=3689.67, SD=800.14) to respond to a critical instruction (Time_Critical) than those
in the high load condition (M=3313.37, SD=591.25), F(1,68)=5.12, p=.026. However,
in-line with our hypothesis, participants made more cognitive load errors in the high load
condition (M=0.35, SD=0.25) than in the low load condition (M=0.15, SD=0.17),
F(1,68)=15.714, p<.001.
We also found a significant Condition x Load interaction for Time_Drag_Drop
(see figure 2), F(1,68)=5.25, p=.025, indicating that participants in the ToM (low load)
condition took more time to drag and drop an object than those in the ToM (high load)
29
condition. This was counter to our hypothesis in that we would have expected
participants to take more time in the high load condition.
Figure 2. Study 1a: Interaction effect of Condition x Load for Time_Drag_Drop
Including gender as another factor, we found a significant main effect for
Time_Drag_Drop, F(1,64)=4.88, p=.03, wherein males took longer (M=1344.54,
SD=275.85) than females (M=1177.43, SD=227.64) to drag and drop an object. Also,
males tended to make more cognitive load errors (M=0.33, SD=0.25) than females
(M=0.21, SD=0.22), F(1,64)=4.4, p=.04.
Analyses of personality measures
We used a 3-step hierarchical linear regression model for our analyses of all
personality measures. The ratings for each personality scale were averaged and then
centered by subtracting the mean of each scale from each participant’s average score on
that scale. We also used effect coding for condition (non-ToM = -1, ToM = 1) and load
non-ToM ToM
Low load 1125.54 1347.9
High load 1212.87 1178.21
1000
1050
1100
1150
1200
1250
1300
1350
1400
Time (ms)
Condition x Load: Time_Drag_Drop (p=.025)
30
(low load = -1, high load = 1). Interaction effects for condition and load were calculated
as the product of condition/load with the centered scores of the appropriate personality
measure.
In step 1 of our model, we input the centered scores of a personality scale and the
effect coded condition and load variables. In step 2, we added the interactions, viz.
Condition x Load, Personality Measure x Condition, and Personality Measure x Load.
Finally, in step 3 we had the 3-way interaction of Personality Measure x Condition x
Load. We did analyses for all 6 aforementioned DVs. The reliability coefficients for all
personality measures can be seen in table 1.
Table 1. Reliability of personality measures across studies
For the Perspective Taking (PT) facet of the IRI, we got a significant PT x
Condition interaction for Time_Drag_Drop (β=-.44, t(65)=3.76, p<.001). Participants
scoring low in PT and in the ToM condition took longer to drag and drop an object than
those in the non-ToM condition (see figure 3). On the other hand, those who were high
on PT and in the ToM condition responded faster than those in the non-ToM condition.
These results are in line with our hypothesis as we had predicted that people with higher
Study 1a (n=72) Study 1b (n=73) Study 2 (n=243)
Personality Measure Cronbach's Alpha Cronbach's Alpha Cronbach's Alpha
Perspective taking (7 items) 0.81 0.81 0.77
Empathic Concern (7 items) 0.74 0.77 0.78
Fantasy (7 items) 0.84 0.79 0.81
Personal Distress (7 items) 0.77 0.70 0.80
Individualism (13 items) 0.66 0.72 0.75
Collectivism (14 items) 0.80 0.87 0.79
Egocentrism (15 items) 0.82 0.79 0.76
Narcissism (10 items) 0.76 0.73 0.75
31
perspective taking abilities should be able to make quicker decisions while keeping
another’s perspective in mind. Moreover, those with low perspective taking skills would
find it harder (i.e. take more time) to take another’s point-of-view while following
instructions.
Figure 3. Study 1a: Interaction effect of Perspective Taking x Condition for
Time_Drag_Drop
For the fantasy scale (FS), we saw significant main effects for Time_Critical (β=-
.25, t(68)=-2.15, p=.035) and #cognitive load errors (β=-.35, t(68)=-3.36, p<.001). This
indicates that participants with low fantasy scores took longer to respond to a critical
instruction and had more cognitive load errors than those scoring high on fantasy. Thus,
having high fantasy could have aided participants in taking another’s perspective
resulting in quicker responses and fewer errors.
Low PT High PT
non-ToM 1060.15 1315.30
ToM 1369.42 1178.78
0.00
200.00
400.00
600.00
800.00
1000.00
1200.00
1400.00
1600.00
Time (ms)
Perspective Taking x Condition: Time_Drag_Drop (p<.001)
32
Looking at Egocentrism, we get a 3-way interaction (see figure 4) of Egocentrism
x Condition x Load for #critical errors, β=.25, t(64)=2.07, p=.043. In the ToM condition,
participants with high egocentrism made almost twice as many critical errors in the high
load condition than the low load condition. This fits our hypothesis that high cognitive
load, in combination with trying to keep another’s perspective in mind, taxes one’s
executive functioning especially for highly egocentric individuals leading them to make
more critical or ToM errors. As highly egocentric individuals move from high to low load
their critical errors decrease substantially. Thus, load moderates the effect of egocentrism.
The other personality measures, viz. empathic concern, personal distress, individualism,
collectivism, and narcissism yielded no significant main or interaction effects.
Figure 4. Study 1a: Interaction effect of Egocentrism x Condition x Load for #Critical
Errors
Low Egocentrism High Egocentrism
(1) ToM, High Load 0.16 0.34
(2) ToM, Low Load 0.28 0.17
(3) non-ToM, High Load 0.21 0.03
(4) non-ToM, Low Load 0.25 0.25
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
# Errors
Egocentrism x Condition x Load: #Critical Errors (p=.043)
33
Belief Analysis
Using the manipulation check for the ToM, we were able to get an idea of the
number of participants who actually believed that the instructions were coming from the
director and not the computer. We classified these participants into two groups: believers
(n=16) and non-believers (n=15). Thus, about 52% of the participants assigned to the
ToM condition bought the story of the experimenter as the director who was giving
instructions.
Performing a Belief x Load ANOVA resulted in only a significant main effect of
belief for cognitive errors made. Non-believers (M=0.31, SD=0.23) made more cognitive
load errors than believers (M=0.14, SD=0.20), F(1,27)=7.56, p=.011.
Discussion
This was a preliminary study to test our experimental design and to gauge the
effectiveness of the ToM/non-ToM manipulation. Some of the results supported our
hypothesis while others either yielded opposite or non-significant results.
We did find that participants took significantly longer to respond to neutral instructions in
the ToM condition than the non-ToM condition. Moreover, they took more time to
respond to critical instructions, and to drag and drop objects in the ToM condition;
however, these results were not significant but still in the right direction. Thus, the ToM
condition seemed to support our hypothesis that participants would take longer to respond
while simultaneously trying to keep another person’s perspective in mind.
Participants found the high cognitive load task harder (more load errors) than the
low load task which was consistent with our hypothesis. However, they responded more
34
quickly to critical instructions in the high load condition than in the low load condition.
This was exactly opposite to what we had predicted; perhaps, participants were
discounting focusing on the high load task in favor of faster reaction times.
We also saw some main effects and interactions for some of the personality
measures. For instance, those participants with lower perspective taking scores and
assigned to the ToM condition seemed to take longer while dragging and dropping an
object than their counterparts in the non-ToM condition. This could imply that those with
low perspective taking skills were taking additional time to process and keep the
director’s perspective in mind while following instructions. Moreover, there was an
indication that participants with high perspective taking skills tended to be much quicker
while dragging and dropping objects in the ToM condition rather than the non-ToM
condition. This could be because those with higher PT abilities were better suited in
maintaining a mental representation of the director’s perspective while moving objects. In
the non-ToM condition they were not trying to follow another’s perspective. We would
have hoped to see similar results for responses to critical instructions but these were not
significant.
Cognitive load seemed to moderate the effects of egocentrism in that participants
scoring high on our egocentrism measure and assigned to the high load-ToM condition
made more critical errors than those assigned to the low load-ToM condition. This could
mean that higher egocentrism leads one to make more internal/egocentric decisions (i.e.
more critical errors) especially when asked to take another’s perspective, and this is
amplified by high cognitive load.
35
Our manipulation check for the ToM condition indicated that only about half the
participants actually believed that the instructions were coming from the director; the
other half believed the instructions were automated and coming from the computer.
Essentially, these non-believers were treating the ToM condition as a non-ToM
condition. Separate analysis of this group of believers and non-believers only showed that
believers made fewer cognitive load errors than non-believers.
Overall, only some of the results supported our hypotheses. We felt that perhaps
our experimental task was too simple (with only 8 grids) to accurately and effectively
gauge the differences between times/errors across condition/load. Furthermore, our
ToM/non-ToM manipulation could use further improvement given that 48% of the
participants in the ToM condition did not buy it. To reflect these changes, we conducted
study 1b with more grids and a slightly refined manipulation.
36
Chapter Two: Study 1b
This study was similar to the previous one but with improvements to the software
design.
Methods
Participants
We recruited a total of 73 participants (males = 21, females = 52) from the
university subject pool. These participants were all undergraduate students at the
University of Southern California (USC) and were awarded course credit for their
participation. They were all at least 18 years old at the time of their participation and the
average age was 20 years.
Design and Apparatus
The design of this study was virtually identical to the previous one with a few
differences. We increased the number of experimental grids to 16. Again, these grids all
had ambiguous objects. The number of cognitive load tasks also increased to 16. Thus,
we had a total of 16 trials. The sequential order of the trials was the same as before, viz.
ToM (grid pictures, letters for the load task, experimental grid, and load recognition), and
non-ToM (letters for the load task, experimental grid, and load recognition).
Based on participants’ feedback and results we shortened the time interval
between instructions from 8 seconds to 7 seconds. Some of the non-believers from the
previous study had stated that they thought the time interval was too long. By decreasing
this interval we wanted to increase the plausibility of our ToM manipulation. In other
words, we hoped that faster instructions would imply that the director was actually
37
observing the participant’s actions and sending subsequent instruction promptly. We also
induced a few typographical errors into these instructions to, once again, make the
participant think that the director was typing and sending instructions in real time. These
errors were minor (e.g. “moove the ball 1 cell down) and did not change the meaning of
the instruction in any significant way. As before, there were 4 instructions per grid.
We used the same measures of personality and manipulation check.
Procedure
There was no change to the procedure from that of study 1a except that the total
time of the study increased to about 26-28 minutes for the ToM condition and 22-24
minutes for the non-ToM condition.
Results
We used similar types of analyses and coding schemes as in the previous study.
Central Analyses
As before, we performed a Condition x Load ANOVA for the six DVs with
condition (non-ToM, ToM) and load (low, high) as between subjects factors. The
distribution of the participants was as follows: ToM (n=38), non-ToM (n=35), low load
(n=37), and high load (n=36). We got a significant main effect of load for the number of
cognitive load errors, F(1,69)=32.68, p<.001. Similar to the previous study, participants
found the high load task harder and made more errors (M=.54, SD=.25) than in the low
load condition (M=.22, SD=.22).
There was also a marginally significant Condition x Load interaction for #critical
errors which indicated that participants tended to make more errors in the high load
38
(ToM) condition than in the high load (non-ToM) condition. This would imply that high
load was possibly taxing one’s cognitive resources, especially while trying to take
another’s perspective when making choices. Thus, though in line with our hypothesis,
this interaction was not significant, F(1,69)=3.42, p=.069.
We did not get any significant results by including gender as another between
subjects factor.
Analyses of personality measures
We used regression analyses similar to those of the previous study to analyze our
personality measures. The reliability coefficients for all personality measures can be seen
in table 1 above.
For empathic concern (EC), we got a significant 3-way EC x Condition x Load
interaction for #critical errors (see figure 5), β=.31, t(65)=2.49, p=.015. Participants
assigned to the ToM condition and displaying higher empathy made fewer errors than
those with low empathy but only for the low cognitive load condition. For high load,
higher empathy resulted in more critical errors. This could imply that though individuals
with high empathy scores are better at taking another’s perspective (fewer critical errors),
they can only successfully do so when their cognitive resources to process information
are not strained (by high load). Thus, load seems to be moderating the effects of empathy.
39
Figure 5. Study 1b: Interaction effect of Empathic Concern x Condition x Load for
#Critical Errors
There was also a significant 3-way EC x Condition x Load interaction for
Time_Critical (see figure 6), β=.27, t(65)=2.25, p=.028. In the ToM condition, those with
high empathy took less time to respond to a critical instruction but only in the low load
condition. For high load, they took more time. Once again, load appears to be affecting
those with higher empathy.
Low EC High EC
(1) ToM, High Load 0.14 0.25
(2) ToM, Low Load 0.26 0.11
(3) non-ToM, High Load 0.28 0.15
(4) non-ToM, Low Load 0.13 0.39
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.45
# Errors
Empathic Concern x Condition x Load: #Ciritical Errors (p=.015)
40
Figure 6. Study 1b: Interaction effect of Empathic Concern x Condition x Load for
Time_Critical
For the fantasy scale, we got a significant Fantasy x Condition interaction for
#critical errors (see figure 7), β=-.35, t(66)=-2.95, p=.004. Participants with higher scores
on fantasy tended to make significantly fewer critical errors in the ToM condition than in
the non-ToM condition. It would appear that in the ToM condition those with higher
levels of fantasy were better able to follow directions from the director’s perspective and
usually chose only those critical objects that the director was actually referring to. On the
other hand, high fantasy scorers in the non-ToM condition made more critical errors. This
could be because they were not asked to make decisions while following another’s
perspective. In this case, high fantasy could be a potential distractor.
Low EC High EC
(1) ToM, High Load 3528.80 3637.07
(2) ToM, Low Load 3413.90 3166.47
(3) non-ToM, High Load 3093.86 2966.22
(4) non-ToM, Low Load 2880.64 3939.85
0.00
500.00
1000.00
1500.00
2000.00
2500.00
3000.00
3500.00
4000.00
4500.00
Time (ms)
Empathic Concern x Condition x Load: Time_Critical (p=.028)
41
Figure 7. Study 1b: Interaction effect of Fantasy x Condition for #Critical Errors
The scale for collectivism yielded significant main effects for Time_Critical (β=-
.26, t(69)=-2.27, p=.026, and for #cognitive load errors, β=-.25, t(69)=-2.67, p=.01.
Thus, those with high collectivism took less time to respond to a critical instruction and
made fewer errors than those with low collectivism. This supported our hypotheses that
highly collectivistic individuals, in general, are better at making decisions involving
perspective taking.
Belief Analysis
As in study 1a, we used the manipulation check for the ToM condition and got
two groups: believers (n=16), non-believers (n=21). This indicated that more than half
our sample (57%) did not believe that the experimenter was the one giving them
Low Fantasy High Fantasy
non-ToM 0.20 0.29
ToM 0.32 0.04
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
#Errors
Fantasy x Condition: #Critical Errors (p=.004)
42
instructions from the other room. There were no significant results for the Condition x
Load ANOVA.
Discussion
This study was done as a follow-up to study 1a with slight differences in the
design to allow us to tap into the ToM effect with more accuracy and conviction. Our
central analyses only revealed that high load caused participants to make more cognitive
load errors than low load. This result was also seen in study 1a. There was some
indication that high load was increasing the #critical errors for participants, more-so in
the ToM condition than the non-ToM condition.
We also found more effects for the personality measures. Participants who were
assigned to the ToM condition and with higher empathy scores took less time to respond
to a critical instruction and made fewer critical errors but only in the low load condition.
This could mean that those who are highly empathic are better able to see things and take
decisions based on another’s point-of-view but only if their cognitive resources are not
severely taxed as in the high load condition.
Higher levels of fantasy were found to be beneficial for those in the ToM
condition in that they made fewer critical errors than their counterparts in the non-ToM
condition. Perhaps having high fantasy enables one to take another person’s perspective
more effectively. We also found support for our hypothesis that people who are more
collectivistically oriented take less time to respond to critical instructions, thereby
strengthening the link between collectivism and better perspective taking ability.
43
By introducing typographical errors into the instructions we had hoped that more
participants would believe that they were getting instructions from an actual person and
not from the computer. However, our manipulation check revealed that only about 43%
of our sample believed this to be the case. Some participants even commented that they
thought the errors were purposely inputted into the instructions to trick them into thinking
than the director was typing them out.
A key difference between our two studies so far and those conducted by Keysar et
al., (2000, 2003) and Apperly et al. (2010), was that in addition to having grids with
ambiguous objects they also had grids with relational objects. In these relational grids,
the critical instruction referred to the target object by its relative size (compared to similar
objects) and position within the grid. For example, there could be three glasses of
different sizes – small, medium, and large – and the largest glass would be occluded from
the director’s point-of-view. In this case, the critical instruction – “move the large glass 1
cell down” – would refer to the medium glass as that would be visible to both the director
and the participant.
For our main study we decided to follow this design and included both relational
and ambiguous grids. We also made changes to the manipulation and procedure to enable
us to more effectively convince participants in the ToM condition.
44
Chapter Three: Study 2
This was our primary study that included refinements to the software, procedure,
and manipulation.
Methods
Participants
We recruited a total of 243 participants (males = 77, females = 166) from the
university subject pool. These participants were all undergraduate students at the
University of Southern California (USC) and were awarded course credit for their
participation. They were all at least 18 years old at the time of their participation and the
average age was 20 years.
Design and Material
The design of this study was similar to the previous ones with a few key changes.
We decided to incorporate both relational and ambiguous objects for our grids. Thus,
there were 8 relational grids and 8 ambiguous grids for a total of 16 grids. To
accommodate these relational objects, 8 of the grids from study 1b were altered with new
relational objects that differed in relative size and position within a grid.
For instance, an example of a relational grid with objects reflecting varying size
would be three types of candles – thin, thick, and thickest. The thickest candle would be
in an occluded slot (only visible to the participant); the other two candles would be
visible to both the participant and the director. Thus, if the director instructed the
participant to move the thickest candle, s/he would be referring to the medium-thick
candle and not the thickest (occluded) one. On the other hand, an example of a relational
45
grid with objects in different positions would be three cars placed on different rows: one
would be at the topmost row in an occluded slot; the other two would be on rows 2 and 3
visible to both the participant and director. So, if the director asked the participant to
move the topmost car 1 cell down, s/he would be referring to the car in row 2 and not on
row 1 (occluded). The order in which the ambiguous and relational grids appeared was
randomized.
As before, there was a text box for the instructions. The fourth instruction, which
informed participants about the cognitive load task in the subsequent screen, was
removed. We felt that this instruction was repetitive and that once participants
commenced the experimental trials they could proceed more smoothly without having to
look at another instruction. Thus, we only had three instructions per grid – one critical
and two neutral. The time interval between instructions remained at 7 seconds and no
typographical errors were introduced into any of the instructions.
There was no change to the cognitive load task from that of study 1b. We had 16
experimental trials. The sequential order of these trials was the same as before, viz. ToM
condition (grid pictures, letters for the load task, experimental grid, and load recognition),
and non-ToM condition (letters for the load task, experimental grid, and load
recognition).
An important difference in this study was the introduction of an intercom system
that would enable the participant and director to verbally communicate with one another
across rooms. This was done to maintain the aura of authenticity that the director was in
the other room giving instructions.
46
Measures
We used the same personality measures as before. The manipulation check for the
ToM condition used in the previous studies was further refined with additional questions
that asked participants to indicate when (if at all) they were getting suspicious about the
source of the instructions (see Appendix F).
Procedure
The procedure followed a similar pattern as the previous two studies with some
differences.
In the ToM condition, participants were explicitly informed that the
experimenter/director would not be typing the instructions. Instead, in the interest of
time, these instructions had already been pre-typed and entered into the software. The
experimenter would be sending these instructions at a set interval by clicking on a button
on their interface.
For both the ToM and non-ToM conditions, the purpose of the intercoms was
explained. Participants were told that the experimenter would be using this intercom to
communicate with them periodically. The experimenter, after going into the next room,
would contact them and tell them when to begin. About half-way between the trials (after
8 grids), a screen would appear that would instruct the participants to contact the
experimenter who would then tell them to proceed with the trials. Participants were also
asked to contact the experimenter at the end of the surveys using the intercom. Thus, the
first screen that the participants saw in both conditions was a “Wait to begin” screen
wherein the subjects clicked on a continue button after hearing from the experimenter.
47
The “connection” screen appeared next, as in the previous studies and the subsequent
trials began.
This study took about 26-28 minutes for the ToM condition and 22-24 minutes for
the non-ToM condition.
Results
The types of analyses and coding schemes that we used for this study were similar
to those in the previous studies.
Central Analyses
As before, performance of participants was calculated in terms of times and
errors. This gave us the same set of six dependent variables for relational and ambiguous
trials.
We conducted a repeated measures ANOVA with relational and ambiguous
times/errors as within subject variables, and condition and load as between subject
factors. The distribution of the participants was as follows: ToM (n=122), non-ToM
(n=121), low load (n=125), and high load (n=118). The results indicated a significant
between subjects main effect for condition in terms of Time_Critical, F(1,236)=8.72,
p=.003. Participants took longer to respond to a critical instruction for the non-ToM
condition (Relational: M=3984.73, SD=698.51; Ambiguous: M=2913.14, SD=561.49)
than the ToM condition (Relational: M=3802.61, SD=674.62; Ambiguous: M=2690.26,
SD=508.99). This was contrary to our hypothesis as we had predicted that participants
would take longer to respond to a critical instruction when they had to keep another’s
perspective in mind.
48
There was another between subjects main effect for load in terms of #cognitive
load errors, F(1,239)=204.06, p<.001. Participants made more errors in the high load
condition (Relational: M=.45, SD=.24; Ambiguous: M=.39, SD=.23) than in the low
load condition (Relational: M=.09, SD=.15; Ambiguous: M=.15, SD=.21). This was
consistent with our hypothesis and with the results of the previous two studies.
We also found significant within subjects main effects for the three time variables –
Time_Neutral [F(1,239)=69.13, p<.001], Time_Critical [F(1,239)=839.19, p<.001], and
Time_Drag_Drop [F(1,239)=43.23, p<.001]. For each of these variables, participants
took much longer to respond in the relational trials than the ambiguous trials. There was
also a significant main effect for #critical errors in that participants made more errors in
the relational trials (M=.33, SD=.31) than in the ambiguous trials (M=.16, SD=.25),
F(1,239)=120.15, p<.001.
Including gender as another between subjects factor did not yield any significant
results.
Analyses of personality measures
We conducted separate regression analyses for relational and ambiguous trials
using the same 3-step hierarchical linear regression model from the previous studies. The
reliability coefficients for all personality measures can be seen in table 1 above.
For relational trials, we found a significant main effect of empathic concern (EC) for
#critical errors, β=-.19, t(239)=-3.003, p=.003, which indicated that participants scoring
low on the empathy scale made more critical errors than those with high empathy scores.
Thus, those with higher empathy were potentially better perspective takers. There was
49
also a significant 3-way EC x Condition x Load interaction for #critical errors (see figure
8), β=.14, t(235)=2.14, p=.034. Participants assigned to the ToM condition and with
higher empathy made more errors in the high load condition as opposed to the low load
condition. In this case, high load probably increased the strain on one’s cognitive
resources and even individuals with high empathy tended to make more theory of
mind/critical errors. Those with lower empathy, in general, made more errors across load
than those with high empathy; load did not appear to moderate the #errors for those with
low empathy.
Figure 8. Study 2: Relational trials – Interaction effect of Empathic Concern x Condition
x Load for #Critical Errors
Low EC High EC
(1) ToM, High Load 0.43 0.31
(2) ToM, Low Load 0.44 0.23
(3) non-ToM, High Load 0.43 0.26
(4) non-ToM, Low Load 0.25 0.32
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.45
0.50
# Errors
Empathic Concern x Condition x Load: #Critical Errors (p=.034)
50
Looking at Individualism for ambiguous trials, we got a main effect for #critical
errors, β=.14, t(239)=2.17, p=.031, indicating that those with a high individualistic
orientation made more critical errors than those with low individualistic scores.
Moreover, analysis of Egocentrism for ambiguous trials also revealed a main effect for
#critical errors, β=.18, t(239)=2.83, p=.005, showing that highly egocentric people made
more critical errors. This supports our hypothesis that high individualism and high
egocentrism could lead people to take more egocentric decisions while they are trying to
maintain another’s perspective. There was no effect of individualism or egocentrism for
relational trials.
For relational trials, we found a significant Narcissism x Load interaction for
#critical errors (see figure 9), β=.13, t(236)=2.08, p=.039. High narcissism in
combination with high load appears to significantly increase critical errors by
participants. There was also a marginally significant, β=.13, t(236)=1.95, p=.053,
Narcissism x Condition interaction for #neutral errors indicating that participants with
high narcissism scores made more neutral errors in the ToM condition than the non-ToM
condition.
51
Figure 9. Study 2: Relational trials – Interaction effect of Narcissism x Load for #Critical
Errors
Belief Analysis
Using the manipulation check for the ToM condition, we got two groups:
believers (n=62) and non-believers (n=55). Due to missing data, five participants could
not be put into either group. Thus, about 53% of our participants believed that the
director was responsible for the instructions.
Running separate ANOVAs for relational and ambiguous trials with belief (non-
believers, believers) and load as between subjects factors did not yield any significant
results.
We also analyzed the personality measures for this group of believers and non-
believers by using the same 3-step hierarchical regression model from the previous
Low Narcissism High Narcissism
Low Load 0.33 0.28
High Load 0.30 0.42
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.45
# Errors
Narcissism x Load: #Critical Errors (p=.039)
52
studies. So, for this analysis only the believers and non-believers in the ToM condition
were selected. Effect coding was used for belief (non-believers = -1, believers = 1). In
step 1 of the model, we added a personality measure along with the effect coded load and
belief variables. In step 2, the interactions were added, viz. Belief x Load, Personality
Measure x Load, and Personality Measure x Belief. In step 3, the 3-way Belief x Load x
Personality Measure interaction was added.
For empathic concern (EC), we found significant main effects for #critical errors
in both relational (β=-.29, t(113)=-3.31, p<.001), and ambiguous (β=-.19, t(113)=-2.12,
p=.036) trials. This further bolsters our hypothesis that people with high empathy make
fewer critical errors than those with low empathy. In the case of ambiguous trials, there
was also a significant main effect for #cognitive load errors, β=-.21, t(113)=-2.59,
p=.011, indicating that those with higher empathy made fewer load errors than their
counterparts with lower empathy.
The collectivism scale yielded a significant Collectivism x Belief interaction for
Time_Drag_Drop, β=-.19, t(113)=-2.02, p=.046 (see figure 10). This indicated that belief
moderated the effect of collectivism in that for high collectivism, believers took less time
to drag and drop an object than non-believers. Thus, those who were strongly
collectivistic but did not believe that they were getting instructions from another person
performed worse than any of the other groups. This interaction was only found for the
ambiguous trials.
53
Figure 10. Study 2: Ambiguous Trials (ToM condition) - Interaction effect of
Collectivism x Belief for Time_Drag_Drop
Discussion
After having conducted two preliminary studies, we implemented study 2 with a
few design changes that could help us get more effective results. In addition to
ambiguous trials, we added relational trials following the work done on the grid task
experiments by Keysar et al., (2000, 2003) and Apperly et al. (2010). We introduced an
intercom system allowing the participant and director to communicate with each other in
order to strengthen our claim that instructions were coming from the director (in the ToM
condition).
Results indicated that participants took longer to respond to a critical instruction
in the non-ToM condition than the ToM condition. This was opposite to our prediction
that participants would be much faster in the non-ToM condition as they were not being
Low Collectivism High Collectivism
Non-Believers 969.72 1118.54
Believers 992.9 972.7
850.00
900.00
950.00
1000.00
1050.00
1100.00
1150.00
Time (ms)
Collectivism x Belief: Time_Drag_Drop (p=.046)
54
asked to make decisions by keeping another person’s perspective in mind. However,
consistent with results from the first two studies, we found that participants were making
more cognitive load errors in the high load condition than in the low load condition.
Thus, participants, across all studies, found the high load task to be harder than the low
load one.
Looking at the differences between relational and ambiguous trials, we discovered
that participants, in general, took a longer time during relational trials and made more
critical errors. This is consistent with previous research which has shown that relational
trials are usually harder than ambiguous ones.
Analyses of the various personality measures revealed that individuals with high
empathy tended to make fewer critical errors than those with low empathy. This gives
credence to our claim that people who are better able to empathize with others are more
suited to seeing things from another’s point-of-view and are able to make more accurate
decisions (as shown by fewer errors in this case). However, this effect was moderated by
cognitive load in that despite high empathy, high load resulted in more critical errors than
low load. As we discussed previously, this is probably because high load has the potential
to severely tax one’s cognitive resources, negating any effects of personality.
We also found evidence that high individualism and high egocentrism results in
an increase in critical errors. Moreover, there was an indication that high narcissism leads
to more critical and neutral errors. This is probably because highly individualistic and
egocentric/narcissistic people are more prone to taking egocentric interpretations of
55
others’ viewpoints. Or, in other words, they usually tend to see things from only their
perspective.
One of the reasons that we conducted this study with an intercom system was to
maintain and strengthen the illusion that for participants assigned to the ToM condition,
instructions would be coming from the director who would be (supposedly) observing
their actions via another computer in an adjacent room. Unfortunately, the manipulation
check revealed that only about 53% of the participants believed this to be the case. Thus,
throughout the different studies there was a high degree of suspicion among the
participants. This could be one factor that yielded some results counter to our hypotheses.
56
Conclusion
In this paper, we attempted to investigate the possible influences of personality
and cognitive load on people’s theory of mind or perspective taking ability. Our work
was modeled after that of Keysar et al., (2000 & 2003), and Apperly et al. (2010), who
had implemented physical and virtual versions (respectively) of a grid task, designed to
assess whether participants could keep a director’s perspective in mind while moving
objects within the grid, some of which were occluded from the director’s point-of-view.
This research had clearly demonstrated that people, in general, have a harder time trying
to maintain another’s perspective and making accurate judgments based on the common
ground between all those involved.
Other research has indicated that personality can also influence one’s ToM ability.
People with a collectivistic orientation tend to show greater empathy towards others and
are better able to see things from another’s point-of-view than those who are more
individualistic (Wu & Keysar, 2007). Also, those who are highly narcissistic and
egocentric tend to be poor perspective takers in general (Fiscalini, 1994; Keysar, 1994).
For this study, we wished to see how different personalities effect people’s performance
on our version of the grid task.
Besides personality, we also wanted to look at the effect of cognitive load on
ToM. Prior research by Apperly et al. (2010), had hinted that ToM usage has the potential
of taxing one’s cognitive resources leading to egocentric (and often incorrect) decisions.
However, load was not explicitly manipulated in this or other studies. We designed our
57
study so that participants would be either in a low or high load condition while
simultaneously doing the experimental grid task.
To reiterate our hypotheses, we predicted that participants assigned to a ToM
condition (wherein they were asked to explicitly maintain another’s perspective) would
take a longer time to respond to instructions and would also make more errors than their
counterparts in a non-ToM condition (wherein they were not asked to take the
perspective of another). We further hypothesized that higher cognitive load would lead to
an increase in response time and errors. In terms of personality, we theorized that those
with higher empathy, perspective taking skills, and collectivism would be better at the
ToM task (in terms of fewer errors and faster response times) than those who are highly
egocentric, narcissistic, and individualistic.
Across two preliminary studies and one primary study we found some results that
supported our hypotheses and other results that were either not significant or opposite to
our initial predictions. Across all studies, we did find that participants found the high
cognitive load task harder than the low load task. They clearly had trouble remembering
the 7-letter string (high load) than the 2-letter string (low load) as indicated by the
number of cognitive errors in the high load condition.
There was also an indication that participants assigned to a high load task made
more critical errors in the ToM condition than the non-ToM condition. Thus, having to
maintain another’s perspective while simultaneously working on a high load task might
have taxed the participants’ cognitive resources, thereby inducing more errors. However,
though this finding supports our hypothesis, it was found only in one study (1b) and not
58
the others. For future studies, it would be interesting to conduct this experiment by
having a no-load condition as research has indicated that even a low (2-letter) load can be
enough to make a task more difficult (Wood, 2011).
Surprisingly, in our main study (2) we discovered that participants took more time
to respond to critical instructions in the non-ToM condition than in the ToM condition.
This was opposite to what we had predicted and counter to the results of our preliminary
study (1a). We did find that in study 2, participants took a longer time and had more
critical errors for relational trials than ambiguous trials. This supports the results of
Keysar et al., (2000 & 2003), and Apperly et al. (2010). Thus, in general, relational grids
were much harder than ambiguous grids.
In terms of the personality measures, two of the studies (1b and 2) revealed a
moderating effect of load on empathy. Those participants assigned to the ToM condition
and who scored high on the empathy measure made fewer critical errors than those with
lower empathy but only for low load; higher empathy and higher load yielded more
errors. We believe that this is because high load puts more strain on the cognitive
resources of even highly empathic individuals, causing them to make more theory of
mind errors.
We also found some evidence that those who are highly collectivistic take less
time to respond to critical instructions (implying that they are faster at seeing things from
the other’s viewpoint). Moreover, those with high individualism and egocentrism made
more critical errors than those who scored low on these measures. Though these results
59
corroborate our hypotheses, they were primarily main effects; there was no interaction to
indicate that these results were different for the ToM vs. non-ToM conditions.
There was further evidence of the moderating effect of load on egocentrism and
narcissism. High egocentrism/narcissism led to more critical errors in the high load
condition than the low load condition. Thus, once again we see that high load has a
depleting effect on one’s executive functions.
Looking back at our results, we do find an indication that theory of mind can be
affected by personality and cognitive load. However, the results that we did find were not
consistent across all studies. This could be due to several factors. Participants assigned to
the ToM condition, in the three studies, had trouble believing that they were getting
instructions from the director; we only managed about a fifty percent success rate in
convincing them. Thus, if these non-believers treated the ToM condition as a non-ToM
condition, then we cannot be entirely sure whether we are tapping into their theory of
mind or perspective taking ability. Though we did analyses with belief (believers vs. non-
believers) as another factor, this did not seem to make much of an effect. To have a
greater degree of plausibility, perhaps future versions of this task could involve a live
video feed next to the experimental grid showing the director interacting with his/her
computer.
In their experiments with the grid task, Keysar et al,. (2000, 2003) had used eye
tracking equipment to precisely measure whether participants immediately glanced at an
incorrect critical object after receiving a critical instruction. They based their findings on
this immediate physical response of participants to determine whether they resorted to
60
egocentric interpretations of the director’s instructions. Even though our version of the
task was able to accurately record when the participant moved an object, it was not as
fine-grained. Moreover, Keysar et al. (2000, 2003) achieved their results by using actual
physical grids and people without involving deception of any kind. It could be that
participants are better at following verbal instructions (as opposed to text instructions via
our chat interface). In their software version, Apperly et al. (2010) had embedded pre-
recorded audio files that gave instructions and were able to replicate most of the findings
of Keysar et al. (2000, 2003). Additionally, our personality measures were based
primarily on self-report. For future studies, it would be interesting to see the effects of
different personality factors, that have been explicitly manipulated (by priming, etc.), on
theory of mind.
To conclude, we do see some results supporting our predictions, viz. perspective
taking can be quite challenging and that high cognitive load has the potential of making a
ToM task even more difficult for people with certain types of personality (e.g.
egocentrism, narcissism). However, in order to get even more substantial and conclusive
results, a better version of our task would probably not involve any deception and could
use more refined ways to accurately measure theory of mind.
61
Bibliography
Apperly, I. A., Back, E., Samson, D., & France, L. (2008). The cost of thinking about
false beliefs: Evidence from adults’ performance on a non-inferential theory of
mind task. Cognition, 106, 1093-1108.
Apperly, I.A., Carroll, D.J., Samson, D., Humphreys, G.W., Qureshi, A., & Moffatt, G.
(2010). Why are there limits on theory of mind use? Evidence from adults’ ability
to follow instructions from an ignorant speaker. Quarterly Journal of
Experimental Psychology, 63: 6, 1201 – 1217.
Astington, N.J.(2003).Sometimes necessary, never sufficient: False-belief understanding
and social competence. In B. Repacholi & V. Slaughter (Eds.), Individual
differences in theory of mind. Macquarie monographs in cognitive science
(pp.12–38).Hove, E. Sussex: Psychology Press.
Batson, C. D., Early, S., & Salvarani, G. (1997). Perspective taking: Imagining how
another feels versus imagining how you would feel. Personality and Social
Psychology Bulletin. Vol 23(7), Jul 1997, pp. 751-758
Clark, H. H. (1996). Using Language. Cambridge University Press
Camerer, C., Lowenstein, G., & Weber, M. (1989).The curse of knowledge in economic
settings: An experimental analysis. Journal of Political Economy, 97, 1232–1254.
Davis, M. H. (1980). A multidimensional approach to individual differences in empathy.
JSAS Catalog of Selected Documents in Psychology, 1980, 10, 85.
Davis, M.H. (1983). Measuring individual differences in empathy: Evidence for
multidimensional approach. Journal of Personality and Social Psychology, 44,
113-126.
Davis, M.H. (1996). Empathy: a social psychological approach. Boulder, CO:Westview.
Davis, M. H. , & Stone, T. (2003). Synthesis: Psychological understanding and social
skills. In B. Repacholi & V. Slaughter (Eds.), Individual differences in theory of
mind. Macquarie monographs in cognitive science (pp.305–352). Hove, E.
Sussex: Psychology Press.
Decety, J., Jackson, P. L., (2004). The functional architecture of human empathy.
Behavioral and Cognitive Neuroscience Reviews, 3(2), 71-100.
62
Duan, C., Wei, M., & Wang, L. (2008). The Role of Individualism-Collectivism in
Empathy: An Exploratory Study. Asian Journal of Counseling, Vol. 15, No. 1, 57-
81
Enright, R. D., Shukla, D. G.and Lapsley, D. K. (1980). Adolescent egocentrism-
sociocentrism and self-consciousness. Journal of Youth and Adolescence, 9(2),
101-116.
Fiscalini, J. (1994). Narcissism and co-participant inquiry: Explorations in contemporary
interpersonal psychoanalysis. Contemporary Psychoanalysis, 30, 747–776.
Frankenberger K. D. (2000). Adolescent egocentrism: a comparison among adolescents
and adults. Journal of Adolescence, 23, 343±354
Heinke, M. S. & Louis W. R. (2009). Cultural Background and Individualistic–
Collectivistic Values in Relation to Similarity, Perspective Taking, and Empathy.
Journal of Applied Social Psychology, 39, 11, pp.2570–2590.
Hendin, H.M., & Cheek, J.M. (1997). Assessing Hypersensitive Narcissism: A Re-
examination of Murray's Narcissism Scale. Journal of Research in Personality, 31,
588-599.
Hughes, C. (2002) Executive functions and development: Why the interest? Infant and
Child Development, 11, 69-71
Ickes, W. (2003). Everyday mind reading. New York: Prometheus.
Keysar, B. (1994). The illusory transparency of intention: Linguistic perspective-taking
in text. Cognitive Psychology, 26, 165–208.
Keysar, B., Ginzel, L. E.,& Bazerman, M.H.(1995).States of affairs and states of mind:
The effect of knowledge about beliefs. Organizational Behavior and Human
Decision Processes, 64, 283–293.
Keysar, B., Barr, D.J., Balin, J.A., & Brauner, J.S. (2000). Taking perspective in
conversation: the role of mutual knowledge in comprehension. Psychological
Sciences, 11, 32–38.
Keysar, B., Lin, S.H., & Barr, D.J. (2003). Limits on theory of mind use in adults.
Cognition, 89(1). 25-41.
63
Langdon, R.(2003). Theory of mind and social dysfunction. Psychotic solipsism versus
autistic asociality. In B. Repacholi & V.Slaughter (Eds.), Individual differences in
theory of mind. Macquarie monographs in cognitive science (pp.241–269).Hove,
E. Sussex: Psychology Press.
Markus, H. R., & Kitayama, S. (1991). Culture and self: Implications for cognition,
emotion, and motivation. Psychological Review, 98, 224–253.
Mulligan N. W. (1998). The Role of Attention During Encoding in Implicit and Explicit
Memory. Journal of Experimental Psychology: Learning, Memory, and
Cognition, Vol. 24, No. 1,27-47
Nelson, D. W., & Baumgarte, R. (2004). Cross-cultural misunderstandings reduce
empathic responding. Journal of Applied Social Psychology, 34, 391–401.
Nichols, S. (2001). Mind reading and the cognitive architecture underlying altruistic
motivation. Mind and Language, 16, 425–455.
Onishi, K. H., & Baillargeon, R. (2005). Do 15-month-old infants understand false
beliefs? Science, 308(8), 255–258.
Ozonoff, S., Pennington, B. F., & Rogers, S. J. (1990). Are there emotion perception
deficits in young autistic children? Journal of Child Psychology and Psychiatry
and Allied Disciplines, 31, 343–361.
Perner, J., & Lang, B. (1999). Development of theory of mind and executive control.
Trends in Cognitive Sciences, 3(9), 337-344.
Premack, D. G. & Woodruff, G. (1978). Does the chimpanzee have a theory of mind?
Behavioral and Brain Sciences, 1, 515-526.
Royzman, E. B., Cassidy, K. W., & Baron, J. (2003). “I Know, You Know”: Epistemic
Egocentrism in Children and Adults. Review of General Psychology, Vol. 7, No.
1, 38-65
Samson, D., Apperly, I.A., Kathirgamanathan, U. & Humphreys, G.W. (2005). Seeing it
my way: A case of selective deficit in inhibiting self-perspective. Brain, 128,
1102-1111.
Schelling, T. C. (1960). The strategy of conflict. Cambridge MA: Harvard University
Press
64
Shallice, T. and Burgess, P. W. (1996). The domain of supervisory processes and the
temporal organization of behavior. Philosophical Transactions of the Royal
Society of London B, 351, 1405-1412.
Stotland, E. (1969). Exploratory investigations of empathy. In L. Berkowitz (Ed.),
Advances in experimental social psychology (Vol. 4, pp. 271-313). New York:
Academic Press.
Thompson, L., & Loewenstein, J. (2003). Mental models of negotiation: Descriptive,
prescriptive, and paradigmatic implications. In M.A. Hogg & J. Cooper (Eds.),
Sage handbook of social psychology. London: Sage.
Triandis, H. C. (1995). Individualism and collectivism. Boulder, CO: Westview.
Triandis, H. C. & Gelfand, M. J. (1998). Converging Measurement of Horizontal and
Vertical Individualism and Collectivism. Journal of Personality and Social
Psychology, Vol. 74. No. 1, 118-128
Watson, P. J., Grisham, S. O., Trotter, M. V., & Biderman, M. D. (1984). Narcissism and
empathy: Validity evidence for the narcissistic personality inventory. Journal of
Personality Assessment, 45, 159-162.
Williams, H. M., Parker, S. K., & Turner, N. (2007). Perceived dissimilarity and
perspective taking within work teams. Group and Organizational Management,
32, 569–597.
Wimmer, H., & Perner,J. (1983).Beliefs about beliefs: representation and constraining
function of wrong beliefs in young children’s understanding of deception.
Cognition, 13,103–128.
Wood, J. N. (2011). When do spatial and visual working memory interact? Attention,
Perception, & Psychophysics. Vol. 73, pp. 420-439
Wu, S. & Keysar, B. (2007). Cultural effects on perspective taking. Psychological
Science, 18, 600-606.
65
Appendix A: Individualism/Collectivism Scale
For each of the following statements, please indicate the extent to which you agree with
the statement. Please provide a rating from 1 to 9 using the following scale:
1 2 3 4 5 6 7 8 9
not at all very much
1. I'd rather depend on myself than others.
2. I rely on myself most of the time; I rarely rely on others.
3. I often do "my own thing."
4. My personal identity, independent of others, is very important to me.
5. I am a unique individual.
6. It is important that I do my job better than others.
7. Winning is everything.
8. Competition is the law of nature.
9. When another person does better than I do, I get tense and aroused.
10. Without competition it is not possible to have a good society
11. It annoys me when other people perform better than I do
12. I enjoy working in situations involving competition with others
13. Some people emphasize winning; I’m not one of them
14. If a coworker gets a prize, I would feel proud.
15. The well-being of my coworkers is important to me.
16. To me, pleasure is spending time with others.
17. I feel good when I cooperate with others.
18. If a relative were in financial difficulty, I would help within my means
66
19. It is important to maintain harmony within my group
20. I feel like sharing little things with my neighbors
21. My happiness depends very much on the happiness of those around me
22. Parents and children must stay together as much as possible.
23. It is my duty to take care of my family, even when 1 have to sacrifice what I want.
24. Family members should stick together, no matter what sacrifices are required.
25. It is important to me that I respect the decisions made by my groups.
26. I would sacrifice an activity that I enjoy very much if my family did not approve
of it
27. Children should be taught to place duty before pleasure
67
Appendix B: Interpersonal Reactivity Index
The following statements inquire about your thoughts and feelings in a variety of
situations. For each item, indicate how well it describes you. Please provide a rating
from 1 to 5, using the following scale:
1 2 3 4 5
does not describe describes me
me well very well
1. I daydream and fantasize, with some regularity, about things that might happen to
me.
2. I often have tender, concerned feelings for people less fortunate than me.
3. I sometimes find it difficult to see things from the "other guy's" point of view.
4. Sometimes I don't feel very sorry for other people when they are having
problems.
5. I really get involved with the feelings of the characters in a novel.
6. In emergency situations, I feel apprehensive and ill-at-ease.
7. I am usually objective when I watch a movie or play, and I don't often get
completely caught up in it.
8. I try to look at everybody's side of a disagreement before I make a decision.
9. When I see someone being taken advantage of, I feel kind of protective towards
them.
10. I sometimes feel helpless when I am in the middle of a very emotional situation.
11. I sometimes try to understand my friends better by imagining how things look
from their perspective.
12. Becoming extremely involved in a good book or movie is somewhat rare for me.
13. When I see someone get hurt, I tend to remain calm.
14. Other people's misfortunes do not usually disturb me a great deal.
68
15. If I'm sure I'm right about something, I don't waste much time listening to other
people's arguments.
16. After seeing a play or movie, I have felt as though I were one of the characters.
17. Being in a tense emotional situation scares me.
18. When I see someone being treated unfairly, I sometimes don't feel very much pity
for them.
19. I am usually pretty effective in dealing with emergencies.
20. I am often quite touched by things that I see happen.
21. I believe that there are two sides to every question and try to look at them both.
22. I would describe myself as a pretty soft-hearted person.
23. When I watch a good movie, I can very easily put myself in the place of a leading
character.
24. I tend to lose control during emergencies.
25. When I'm upset at someone, I usually try to "put myself in his shoes" for a while.
26. When I am reading an interesting story or novel, I imagine how I would feel if the
events in the story were happening to me.
27. When I see someone who badly needs help in an emergency, I go to pieces.
28. Before criticizing somebody, I try to imagine how I would feel if I were in their
place.
69
Appendix C: Adolescent Egocentrism Scale
For each of the following statements, please indicate how important they are to you.
Please provide a rating from 1 to 5, using the following scale:
1 2 3 4 5
not important very important
1. Accepting the fact that others don’t know what it’s like being me.
2. Getting other people to better understand why I do things the way I do.
3. Explaining my unique feelings and viewpoints to others so they can get some idea
about what I am like.
4. Trying to get other people to know what it is like being me.
5. Coming to accept that no one will ever really understand me.
6. When walking in late to a group meeting, trying not to distract everyone's
attention.
7. Trying to figure out how other people will react to my accomplishments and
failures.
8. Being able to daydream about great successes and thinking of other people's
reactions.
9. Being able to think about having a lot of money someday and how people will
admire that.
10. Trying and being able to figure out if two people are talking about me when they
are looking my way.
11. Becoming real good at being able to think through my own thoughts.
12. Thinking about my own feelings
13. Being real good at knowing what others are thinking of me.
14. Knowing my own thoughts and feelings.
15. Thinking about myself.
70
Appendix D: Hypersensitive Narcissism Scale
Please answer the following questions by deciding to what extent each item is
characteristic of your feelings and behavior. Please provide a rating from 1 to 5, using the
following scale:
1 2 3 4 5
strongly disagree strongly agree
1. I can become entirely absorbed in thinking about my personal affairs, my health,
my cares or my relations to others.
2. My feelings are easily hurt by ridicule or the slighting remarks of others.
3. When I enter a room I often become self-conscious and feel that the eyes of others
are upon me.
4. I dislike sharing the credit of an achievement with others.
5. I feel that I have enough on my hands without worrying about other people's
troubles.
6. I feel that I am temperamentally different from most people.
7. I often interpret the remarks of others in a personal way.
8. I easily become wrapped up in my own interests and forget the existence of
others.
9. I dislike being with a group unless I know that I am appreciated by at least one of
those present.
10. I am secretly "put out" or annoyed when other people come to me with their
troubles, asking me for my time and sympathy.
71
Appendix E: Manipulation Check Questionnaire (Study 1a/b)
The following questions ask about your opinion of the interface that you interacted with.
1. How easy was it to follow my instructions?
o Very Easy
o Easy
o Neutral
o Difficult
o Very Difficult
2. Any comments?
3. How easy was it to move objects around within the grids?
o Very Easy
o Easy
o Neutral
o Difficult
o Very Difficult
4. Any comments?
5. What did you think about the experiment? Do you have any questions or
observations?
6. Were you, at any time, wondering about the source of the instructions?
o Definitely yes
o Probably yes
o Maybe
o Probably not
o Definitely not
7. What did you think?
8. Who did you think was giving the instructions?
9. Did you believe that I was in the other room giving you instructions?
o Definitely yes
o Probably yes
o Maybe
o Probably not
o Definitely not
10. Any comments?
72
Appendix F: Manipulation Check Questionnaire (Study 2)
The following questions ask about your opinion of the interface that you interacted with.
1. How easy was it to follow my instructions?
o Very Easy
o Easy
o Neutral
o Difficult
o Very Difficult
2. Any comments?
3. How easy was it to move objects around within the grids?
o Very Easy
o Easy
o Neutral
o Difficult
o Very Difficult
4. Any comments?
5. What did you think about the experiment? Do you have any questions or
observations?
6. Were you, at any time, wondering about the source of the instructions?
o Definitely yes
o Probably yes
o Maybe
o Probably not
o Definitely not
7. Any comments?
8. While you were moving the objects, did you believe that I was the one who was
choosing and sending you the pre-typed instructions?
o Definitely yes
o Probably yes
o Maybe
o Probably not
o Definitely not
9. Any comments?
73
10. While you were moving the objects, did you believe that it was a computer
program that was automatically sending you the pre-typed instructions?
o Definitely yes
o Probably yes
o Maybe
o Probably not
o Definitely not
11. Any comments?
12. If at any time you started to get suspicious about the source of the instructions,
when was that?
o I was never suspicious
o I became suspicious while moving the objects
o I became suspicious when the instructions started appearing
o I became suspicious when you started asking me these questions
13. Any comments?
Abstract (if available)
Abstract
Prior research on adults’ perspective taking or theory of mind (ToM) ability (i.e. attributing beliefs, goals, intentions, etc. to others) has indicated an egocentric bias during decision making tasks (Keysar et al., 2000, 2003). Specifically, while partaking in a task that requires taking the perspective of another while making decisions, people occasionally fail in establishing a common ground (common beliefs/knowledge shared between the parties) and tend to see things from only their point-of-view. Some studies (Apperly et al., 2010) have suggested that trying to maintain a ToM of others can tax one’s cognitive resources thereby causing errors in perspective taking tasks. Other research has indicated the possible effects of personality (such as collectivism, Wu and Keysar, 2007) on perspective taking ability. In this study, we attempted to study the effects of low vs. high cognitive load and various personality factors (viz. empathy, perspective taking, collectivism/individualism, and egocentrism/narcissism) on people’s ToM ability. Based on the grid task experiments of Keysar et al., (2000, 2003) and Apperly et al., (2010), we designed and implemented software to assess ToM. Across various studies, our results indicated that adults do find a ToM task more difficult than a non-ToM task. Having a high collectivistic orientation leads to faster responses whereas high individualism and egocentrism leads to more errors. There was also an indication of the moderating effect of cognitive load on personality. For instance, participants with high empathy made fewer errors than those with lower empathy but only for the low load condition. We discuss these and other findings, and also make suggestions for future studies.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Investigating the effects of personality factors on goal shielding
Asset Metadata
Creator
Chopra, Gurveen Singh
(author)
Core Title
Do you see what I see? Personality and cognitive factors affecting theory of mind or perspective taking
School
College of Letters, Arts and Sciences
Degree
Doctor of Philosophy
Degree Program
Psychology
Publication Date
07/09/2014
Defense Date
06/11/2012
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
cognitive load,collectivism,egocentrism,Empathy,executive function,individualism,narcissism,OAI-PMH Harvest,Personality,perspective taking,theory of mind,ToM
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Read, Stephen J. (
committee chair
), McLaughlin, Margaret L. (
committee member
), Miller, Lynn C. (
committee member
), Walsh, David A. (
committee member
)
Creator Email
gchopra@usc.edu,gurveen_c@yahoo.com
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c3-54375
Unique identifier
UC11289207
Identifier
usctheses-c3-54375 (legacy record id)
Legacy Identifier
etd-ChopraGurv-931.pdf
Dmrecord
54375
Document Type
Dissertation
Rights
Chopra, Gurveen Singh
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 a...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA
Tags
cognitive load
collectivism
egocentrism
executive function
individualism
narcissism
perspective taking
theory of mind
ToM