University of Southern California Dissertations and Theses

The relationship between levels of expertise, task difficulty, perceived self-efficacy, and mental effort investment in task performance

(USC Thesis Other)

The relationship between levels of expertise, task difficulty, perceived self-efficacy, and mental effort investment in task performance

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

THE RELATIONSHIP BETWEEN LEVELS OF EXPERTISE, TASK DIFFICULTY,
PERCEIVED SELF-EFFICACY, AND MENTAL EFFORT INVESTMENT IN TASK
PERFORMANCE

by

Hsin-Ning Ho

A Dissertation Presented to the
FACULTY OF THE USC ROSSIER SCHOOL OF EDUCATION
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree of
DOCTOR OF EDUCATION

May 2010

Copyright 2010 Hsin-Ning Ho
ii

ACKNOWLEDGMENTS

My gratitude and thanks to everyone who had supported me and inspired me during
my academic study. I especially would like to thank my chair and mentor, Dr. Richard
Clark. His guidance and encouragement was the most important support for me to
complete this dissertation. I am very grateful for his patience and understanding
throughout the years. I also would like to thank my committee members Dr. Robert
Rueda and Dr. Robert Keim for their invaluable insights to enrich this study. It is
impossible for me to complete this dissertation without their advice and help.
I would like to express my appreciation to my friend, A-Tsai, who helped me
understand the computer game Warcraft III and made the experimental design possible to
implement.
Moreover, I would like to express my deepest thanks to my dear parents who have
always supported me in every way they could. I am so fortunate to be their daughter. I
also want to thank my husband and my son for their love and care during this journey.
Finally, may the glory and praise to the God our Jesus Christ. He is the one I trust and
always leads me with His unceasing love.

iii

TABLE OF CONTENTS

ACKNOWLEDGMENTS ii

LIST OF TABLES vi

LIST OF FIGURES viii

ABSTRACT xi

CHAPTER I: REVIEW OF THE LITERATURE 1
Introduction 1
Review of the Literature 1
Perceived Self-efficacy 2
Self-efficacy Judgment and levels of expertise 3
Self-efficacy Judgment and levels of task difficulty 5
Summary 6
Self-Efficacy judgmental error – Overconfidence 6
Levels of expertise and overconfidence 7
Perceived Task difficulty and overconfidence 8
Summary 9
The relationship between task familiarity, expertise, perceived self-efficacy,
and mental effort 10
Levels of expertise and mental effort 11
Levels of Task difficulty and mental effort 13
Perceived self-efficacy and mental effort 16
Summary 18
Conclusion 19
The Instrument of the Variables Measurement 21
Mental effort Measures 21
Self-report measure 21
Dual-task measure 22
Perceived Task Difficulty Measure 23
Expertise Measures 23
Perceived Self Efficacy Measures 25
Significance of the problem 26
Purpose of the Study 26
Research Questions 27
Hypotheses 27
Definition of Terms 38

iv

CHAPTER II: METHODOLOGY 40
Apparatus 40
WarCraft III 41
Secondary Task 45
Computers 47
Instrumentations 47
Expertise 47
Perceived self efficacy 48
Mental Effort Investment 48
Performance 49
Subjects 49
Design 50
Data Collection 52
Data Analysis 54

CHAPTER III: RESULTS 56
Introduction 56
Internal Consistency Reliability 56
Descriptive Statistic Analysis of Subjects’ Background and Game Experiences 57
Descriptive Statistic Analysis of Perceived Self Efficacy 61
Descriptive Statistic Analysis of Mental Effort 64
Descriptive Statistic Analysis of Levels of Task Difficulty 65
Hypothesis Number 1a 67
Hypothesis Number 1b 69
Hypothesis Number 2a 70
Hypothesis Number 2b 72
Hypothesis Number 2c 75
Hypothesis Number 3a 76
Hypothesis Number 3b 80
Hypothesis Number 4a 81
Hypothesis Number 4b 84
Hypothesis Number 5a 86
Hypothesis Number 5b 89
Hypothesis Number 6a 91
Hypothesis Number 6b 94
Hypothesis Number 6c 96
Path Analysis 98

CHAPTER IV: DISCUSSION 100
Summary 100
Results of Descriptive Statistics 101
Hypothesis Number1a 102
Hypothesis Number 1b 102
Hypothesis Number 2a 102
v

Hypothesis Number 2b 103
Hypothesis Number 2c 103
Hypothesis Number 3a 104
Hypothesis Number 3b 104
Hypothesis Number 4a 104
Hypothesis Number 4b 105
Hypothesis Number 5a 106
Hypothesis Number 5b 107
Hypothesis Number 6a 108
Hypothesis Number 6b 109
Hypothesis Number 6c 109
Research Question 1 110
Research Question 2 111
Conclusions 113
Recommendations for Future Research 115

REFERENCES 117

APPENDICES: 128
Appendix A: Subject Consent Form 128
Appendix B: Background Questionnaire 132
Appendix C: Game Confidence Level 135
Appendix D: Mental Effort Measure 138
Appendix E: Game Difficulty Level 141

vi

LIST OF TABLES

Table 1: Hypotheses Matrix for Possible IV and DV Combinations 38
Table 2: Reliability of Scales 57
Table 3: Descriptive Statistics of Perceived Self-Efficacy 61
Table 4: Descriptive Statistics of Perceived Self-Efficacy from scale 0 to 100 63
Table 5: Descriptive Statistics of Self-Reported Mental Effort (N =66) 64
Table 6: Descriptive Statistics of Perceived task difficulty among all subjects 66
Table 7: One-Way ANOVA Test (N=66) 67
Table 8: Regression Results (N=66, Predictor: Levels of Task Difficulty,
Dependent Variable: Mental Effort) 68

Table 9: Regression Results (N=66, Predictor: Levels of Task Difficulty,
Dependent Variable: Task Performance) 70

Table 10: Regression Results (N=66, Predictor: Perceived self-efficacy,
Dependent Variable: Mental Effort) 71

Table 11: Regression Results (N=66, Predictor: Perceived self-efficacy,
Dependent Variable: Mental Effort) 73

Table 12: Regression Results (N=66, Predictor: Perceived self-efficacy,
Dependent Variable: Task Performance) 76

Table 13: Regression Results (N=66, Predictor: Pretest1 score, Dependent
Variable: Pretest2 score) 77

Table 14: Percentile Analysis of pretest2 predicted scores (N=64) 78

Table 15: Novice and Expert Groups 79

Table 16: Regression Results (N=21, Predictor: Expertise, Dependent
Variable: Mental Effort) 79

Table 17: Regression Results (N=66, Predictor: Expertise, Dependent
Variable: Task Performance) 81
vii

Table 18: Group Statistics of Mental effort 82

Table 19: Group Statistics of Mental effort 83

Table 20: Regression Results ((N=66, Predictor: Expertise, Levels of task
difficulty, Dependent Variable: Mental Effort) 84

Table 21: Group Statistics of Task Performance 85

Table 22: Regression Results ((N=66, Predictor: Expertise, Levels of task
difficulty, Dependent Variable: Task Performance) 86

Table 23: Group Statistics of Perceived Self-Efficacy 87

Table 24: Group Statistics of Mental effort 88

Table 25: Regression Results ((N=66, Predictor: Perceived self-efficacy,
Levels of task difficulty, Dependent Variable: Mental Effort) 89

Table 26: Group Statistics of Task Performance 90

Table 27: Regression Results ((N=66, Predictor: Perceived self-efficacy,
Levels of task difficulty, Dependent Variable: Task Performance) 91

Table 28: Group Statistics of Perceived Self- Efficacy 92

Table 29: Group Statistics of Mental effort 93

Table 30: Regression Results (Predictor: Perceived self-efficacy, Expertise,
Dependent Variable: Mental Effort) 94

Table 31: Group Statistics of Mental Effort 95

Table 32: Regression Results (N=66, Predictor: Perceived self-efficacy,
Expertise, Dependent Variable: Mental Effort) 95

Table 33: Group Statistics of Task Performance 96

Table 34: Regression Results (N=66, Predictor: Perceived self-efficacy,
Expertise Dependent Variable: Task Performance) 97
viii

LIST OF FIGURES

Figure 1: Suggested inverted U relationship between self-efficacy and
mental effort 18

Figure 2: Suggested Relationship between Expertise, Perceived self-efficacy,
Perceived task difficulty, Mental Effort and Performance 20

Figure 3: The relationship between levels of task difficulty and mental effort 27

Figure 4: The relationship between levels of task difficulty and task
performance 28

Figure 5: The relationship between perceived self-efficacy and self-reported
mental effort 28

Figure 6: The relationship between perceived self-efficacy and dual-task
measured mental effort 29

Figure 7: The relationship between perceived self-efficacy and task
performance 29

Figure 8: The relationship between expertise and mental effort 30

Figure 9: The relationship between expertise and task performance 30

Figure 10: The relationship between expertise and mental effort 31

Figure 11: The relationship between levels of task difficulty and mental effort 31

Figure 12: The relationship between expertise and task performance 32

Figure 13: The relationship between levels of task difficulty and task
performance 32

Figure 14: The relationship between perceived self-efficacy and mental effort 33

Figure 15: The relationship between levels of task difficulty and mental effort 33

Figure 16: The relationship between perceived self-efficacy and task
performance 34

ix

Figure 17: The relationship between levels of task difficulty and task
performance 34

Figure 18: The relationship between perceived self-efficacy and mental effort 35

Figure 19: The relationship between expertise and mental effort 35

Figure 20: The relationship between perceived self-efficacy and dual-task
measured mental effort 36

Figure 21: The relationship between expertise and mental effort 36

Figure 22: The relationship between perceived self-efficacy and task
performance 37

Figure 23: The relationship between expertise and task performance 37

Figure 24: Normal P-P Plot of the regression model of self-reported
mental effort 72

Figure 25: Normal P-P Plot of the regression model of dual task measured
mental effort 74

Figure 26: Scatter Plot for perceived self efficacy and mental effort 74

Figure 27: Performance Model: The Relationship between Expertise,
Levels of Task Difficulty, Perceived Self-Efficacy, Mental Effort,
and Task Performance 98

Figure 28: The Relationship between Levels of Expertise, Levels of
Task Difficulty, and Mental Effort 105

Figure 29: The Relationship between Levels of Expertise, Levels of
Task Difficulty, and Task Performance 106

Figure 30: The Relationship between Perceived Self-Efficacy, Levels of
Task Difficulty, and Mental Effort 107

Figure 31: The Relationship between Perceived Self-Efficacy, Levels of Task
Difficulty, and Task Performance 108

x

Figure 32: The Relationship between Perceived Self-Efficacy, Levels of Expertise,
and Mental Effort 109

Figure 33: The Relationship between Perceived Self-Efficacy, Levels of Expertise,
and Task Performance 110

xi

ABSTRACT

This study examined the impact of different levels of task difficulty and expertise on
self-efficacy judgments. In addition, the study examines how self-efficacy judgments
affect the amount of mental effort investment and task performance under different levels
of task difficulty and expertise. Results from this study are used to build a performance
model that helps illustrate the relationship among these variables.
A quantitative experimental design was used for the study. A strategic computer
game, WarCraft III, was used to examine the subjects’ levels of expertise and task
performances. Three tasks with different difficulty levels - Easy, Moderate, High were
selected for the experiment. A developed program that can randomly present a pop-out
window on the computer screen was used as a secondary task to interrupt subjects’
engagement of the primary task and to examine subjects’ amount of mental effort
investment. In addition, questionnaires were used to examine the subjects’ perceived
self-efficacy and self-reported mental effort. Sixty-six subjects participated in the study.
One-way ANOVA result showed that there was no significant difference between
easy and moderate tasks. As a result, data from these two tasks were combined as
“Normal” task. Data of “Normal” and “High” tasks were then used to test fourteen
hypotheses.
Results from t-test showed that there was a significant and positive relationship
between perceived self-efficacy and expertise. Subjects with high level of expertise have

xii

high self-efficacy and subjects with low level of expertise have low self-efficacy.
However, levels of task difficulty and perceived self-efficacy were negatively correlated.
The higher the levels of task difficulty are, the lower the perceived self-efficacy is found.
Results from the correlation and regression analysis also indicated that the relationship
between perceived self-efficacy and self-reported mental effort was negative. However,
the scatter plot suggested that the relationship between perceived self-efficacy and
dual-task measured mental effort roughly revealed an inverted-U shape – as efficacy
increased, effort also increased until tasks became challenging and then effort appeared to
decrease. Analysis also supported the conclusion that there was a positive relationship
between perceived self-efficacy and task performance. The relationship between levels of
task difficulty and secondary mental effort measures was also positive and the
relationship between levels of task difficulty and task performance was negative.
Moreover, the results also indicated that the relationship between expertise and mental
effort was negative while the relationship between expertise and task performance was
positive. Based on these results, a performance model was established and path analysis
was used to examine the model.

1

CHAPTER I REVIEW OF THE LITERATURE

Introduction

Convincing evidence has shown that perceived self-efficacy is positively related to
task performance in many situations (Bong, 2001; Silver, et al., 1995; Gibbs, 2003).
However, some other research (Vancouver, et. al, 2001, 2002; Pajares, 2002) has also
shown that performance can decline if people have too much self-efficacy. For example,
Hill (2000) found that executives and higher level managers with too much efficacy felt
overconfident and consequently made bad decisions. Therefore, factors affecting people’s
judgment of their own self-efficacy, as it is related to task performance is a topic of
critical importance (Bandura, 1986, 1997; Clark, 1999; Bong, 1997). The following
section will (1) discuss the variables affecting self-efficacy judgment (2) examine the
relationship between perceived self-efficacy and overconfidence, (3) discuss the
relationship between perceived self-efficacy, mental effort, and task performance (4)
review the instruments that have been developed to assess these variables.

Review of the Literature

This literature review is concerned with what role perceived self-efficacy plays in
performance as well as under which conditions the over-estimation of perceived
self-efficacy occurs in affecting task performance. Additionally this review investigates if
judgment of self-efficacy is affected by self-knowledge judgment or the perception of
2

task difficulty and task demand (Bandura, 1997). Research (Clark & Estes, 2002; Schunk,
2003) shows that a sense of too much self-efficacy may affect the amount of mental
effort invested in performing a task and lead to a problem of overconfidence.
Bandura (1997) states that “It is widely believed that misjudgment produces
dysfunction.” Certainly, gross miscalculation of one’s efficacy can get one into trouble.
To act persistently on a belief that one can exercise control over events that are, in fact,
uncontrollable is to tilt at windmills. In activities where the margins of error are narrow
and missteps can produce costly or injurious consequences, personal well-being is best
served by highly accurate efficacy appraisal. Much of the research aimed at devising
strategies for deflating overconfidence stems from concerns over physical risks or
financial losses from faulty judgment” (p.71). Bandura’s statement clearly implies that
the causes of an individuals’ judgmental disparity is a problem worthy of investigation.
The purpose of this literature review is to look at the factors affecting self-efficacy
judgment and its relationship to mental effort investment and task performance. The
research results provide a strong relationship between levels of expertise, perceived task
difficulty or similarity, perceived self-efficacy, mental effort, and task performance.
Perceived Self-efficacy

Bandura (1986, 1977, 1997) has explained that the central focus of self-efficacy
theory is “the dynamic interplay between self-referent thought, action, and affect.”
Self-efficacy relates to one’s belief in one’s capacity to perform a specific task,

3

determines how people feel, think, motivate themselves and behave, and refers to the
confidence in one's ability to behave and to achieve the desired outcome. Pajares and
Miller (1995) have reported that students’ math problem solving performance is affected
primarily by their math self-efficacy among other related factors. Phan and Walker (2000)
also confirmed that self-efficacy plays a direct as well as an indirect role in predicting the
math performance of students from 3
rd
grade to 6
th
grade. Research (Pajares, 1996;
Pajares & Schnk, 2001; Oliver & Shapiro, 1993) also shows that self-efficacy mediates
people’s interpretation of their knowledge, skill, or the results of prior attainments on
performance and is believed to be an important factor to predict achievement. Pajares and
Valiante (in press) found that students’ academic attainments were markedly different
even if their knowledge and skills were of the same level and related this result to the
students’ judgment of their levels of self-confidence. Perceived self-efficacy is
considered a crucial factor in determining people’s choices, persistence, and effort
invested in the task which then determines task performance.
Self-efficacy Judgment and levels of expertise
Individuals with higher self-efficacy will be more likely to report higher levels of
domain-specific knowledge. Conversely, people with lower self-efficacy will be more
likely to demonstrate lower levels of self-knowledge judgment (Gravill, Compeau, &
Marcolin, 2002). The accuracy of self-efficacy judgment on performance prediction is
proved to be domain and situation specific (Bandura, 1986; Hackett & Betz, 1989; Shell,
Colvin, & Bruning, 1995). However, people have limited ability to assess the adequacy
4

of their self-appraisals when confronted with a new task. Given limited familiarity with
the new activity, they tend to make self-efficacy judgments partly from knowledge of
what they have done in similar situations. Individuals use this judgment in predicting
their performance of other similar tasks (Pintrich & Schunk, 1995; Bong, 1997).
However, appearances of similarity or dissimilarity can be misleading and affect the
accuracy of self-efficacy judgment. For example, Farrington (1997) contended that many
children learn early that everything that swims is a “fish”. Koehler, et al. (1996) also
found that people's estimates were at once most inaccurate and made with greatest
certainty when the trait in question was highly similar and familiar to the information
being provided as a basis for judgment. Bandura and Schunk (1981) claimed that when
complex cognitive operations are imbedded in seemingly easy tasks, appearances may be
quite misleading. Research also shows that the classification of judging similarity differs
based on the levels of expertise. Anderson (1995) has also mentioned that the judgment
of task similarity is different between an expert and a novice. Novices chose surface
features to classify a large set of problems into similar categories while experts map the
surface features of a problem onto these deeper principles. De Groot’s (1965) seminal
research with chess players showed that novice chess players were only able to see the
superficial structure of a game board compared to the experts’ deeper recognition of
larger patterns. Adelson (1984) manipulated task instructions with novice and expert
computer programmers, and she found that novices outperformed experts on surface level
tasks, while experts outperformed novices on deep level tasks.
5

Self-efficacy Judgment and levels of task difficulty
In addition, Bandura (1997) claimed that the judgment of personal efficacy also
depends on the levels of task difficulty as perceived by the performer. When the task is
complex and new, people lower their perceived self efficacy and determine their success
rate from others’ performance. On the other hand, people keep their level of efficacy
when confronted with an easy and familiar task. Bandura has cautioned that the accuracy
of perceived self-efficacy prediction should be ”measured in terms of particularized
judgments of capability that may vary across realms of activity, different levels of task
demands within a given activity domain, and under different situational circumstances”
(p. 6).
From another perspective, people will not choose to engage the task if their
perceived self-efficacy is low. People who have high self-efficacy are also more likely to
take challenging task and have confidence that they can achieve the desired results
related to the task (Wood & Bandura, 1989). Bandura (1997) therefore explains and
advises that the most effective learners and performers possess extremely high
domain-general self-efficacy but much lower short-term and specific self-efficacy. For
example, when students have high domain self-efficacy and encounter a novel task within
the domain, their task specific self-efficacy is low and they know that they do not have
immediate access to the knowledge or skills needed to perform the task, however, they
still feel confident in investing the effort needed to carry out the task.

6

Summary
According to the available research, self-efficacy maintains strong predictive and
explanatory power for academics and task performance. Individuals who have strong
self-efficacy usually have confidence and are able to perform the task. The judgment of
self-efficacy depends on people’s self-knowledge and task familiarity or their ability to
judge the difficulty of a task. When individuals perceive the task is familiar, they usually
decide that they have adequate knowledge or skill to accomplish it and possess high
self-efficacy. Conversely, when people perceive the task as novel, their perceived
self-efficacy is low.
However, individuals may misjudge task similarity and make judgment errors about
their self-knowledge (Bandura, 1997; Clark, 1999). They may also misjudge the task
demand, and overestimate their own expertise and capability to solve the task. The
following section will then discuss the relationship between self-efficacy judgment and
overconfidence problem.
Self-Efficacy judgmental error – Overconfidence

Bandura (1997) states that when people’s self-appraisals exceed their performance,
it reflects judgmental disparity and overconfidence. Weinberg (2005) agrees that
overconfidence occurs when people overestimate their ability and knowledge. Yates, et al.
(1998) also defined overconfidence as being that point when probability judgments about
knowledge are higher than the proportion of questions one answers correctly. Dunning, et
al. (2003) have done research on people’s recognition of their own competence and
7

concluded that people evaluate their preconceived beliefs about their skill and use those
beliefs to estimate their performance on any specific test. He contended that people
generally tended to be more overconfident with regarding to solving the task problem
than being underconfident. Clark and Estes (2002) proposed that overconfidence occurs
under the condition that people are fully confident in performing the task but fail the task.
The reason being that they consider the task to be familiar but it is not. These people tend
to use existing strategies or their automated knowledge to solve the problems when what
is actually required is a new approach. When they fail at the task, they blame the task
instead of accounting for the misuse of their problem solving strategy. This research also
contends that overconfident people misjudge their own ability and the novelty of the
tasks they face.
Levels of expertise and overconfidence
Kruger and Dunning (1999; 2002) have argued that inaccurate self-appraisal
judgment mostly comes from unskilled performers because these people’s metacognitive
skills are also poor. Flannelly (2001) found that nursing school students who perform
worse on tests of domain knowledge tend to be more overconfident about their
performance, he then attributed the result to their lack of critical thinking ability. He
therefore recommended that students should be trained to determine their degree of
knowledge accurately while studying, so they can correct their misunderstandings and
choose more effective study strategies. Grimes (2002) also found that students usually
misjudge their prediction of the exam performance and express overconfidence.
8

Some other studies, however, have argued that individuals, with greater self-efficacy,
more active in their day-to-day learning or engaging with activities and increasing their
interaction with their environment, will become more self-aware of their capabilities.
These people are also more likely to demonstrate over-estimation of their abilities.
Stephan & Kiell (2000) found that the judgmental heuristics between expert and novice
made no difference. Gravill, Compeau, & Marcolin (2002) reported that increased
computer self-efficacy relates to increased self-confidence of IT knowledge and
over-estimation of self-knowledge. Low confidence in abilities was found to be related to
under-estimation and lower levels of self-awareness. Hill (2000) also found that business
executives or senior managers have tremendous judgmental disparity about their
performance and often commit errors as a result of being overconfident.
Perceived Task difficulty and overconfidence
Banudra (1986) has mentioned that when performance requirements are ill-defined,
underestimating task demands produces errors in the direction of overassurance;
overestimating task demands will produce errors in the conservative direction. Initially, a
person who perceives a task to be difficult may do so based on low perception of
competence. On the other hand, when they perceive the task as easy, their competence is
perceived to be high. However, studies concerning the relationship between task
difficulty and overconfidence have provided some interesting discussions.
Pulford & Colman (1997) found that hard questions produced significantly higher
levels of overconfidence than medium-difficulty and easy questions, which in turn
9

resulted in underconfidence. Stone (1994) also found that self-efficacy judgment for
cognitively complex tasks tended to be overconfidence. Others (Juslin, Olsson, &
Bjorkman, 1997; Budescu et al., 1997) also found this “hard-easy” effect between task
difficulty and overconfidence. Prinzel (2002) studied that the relationship between
perceived self-efficacy and task difficulty on NASA pilot performance. He found that
low self-efficacy pilots performed better under high workload situations. Although high
self-efficacy pilots performed better under both constant and variable reliability
conditions, the result suggested that they also acted overconfident and limited their use of
other strategies.
On the other hand, Kruger (1999) found that people estimate their performance to be
better than average when the task is easy and estimate themselves to be worse than
average when the task is hard in terms of comparing their personal expertise to those of
others. Salomon (1984) also found that children who possessed high self-efficacy on easy
media (TV) performed worse than people whose self-efficacy is low on hard media
(print-text) when they were learning. The relationship between task difficulty and
efficacy or confidence judgment may need more investigations.
Summary
The relationship between levels of expertise, perceived self-efficacy, levels of
task-difficulty, and overconfidence is complicated. Research has showed that individuals
with lower levels of expertise express more overconfidence and inaccurate self-appraisal
judgment than individuals with higher levels of expertise. Other research has proposed
10

that experts also commit overconfidence and inaccurate self-knowledge judgment as do
novices. Besides, other research has found that the more difficult the task is, the more the
overconfidence occurs. However, some other research has found that perceived
self-efficacy tends to be high when people perceive the task is easy instead of being
difficult. Therefore, the relationship between perceived self-efficacy judgment and
overconfidence is still uncertain when one takes into account both of levels of expertise
and task difficulty.
The relationship between task familiarity, expertise, perceived self-efficacy, and mental
effort

Kahneman (1970, 1973) was the first to define effort as an attentional control
mechanism within an information processing framework. His model rendered effort
synonymous with attention and attentional regulation. Kahneman’s principle of effort
regulation was based on a simple feedback loop; task demand was defined by a standard
amount of effort and failure to invest effort to an appropriate degree resulted in a
degradation of performance. The capacity theory of attention postulates the existence of a
limited pool of processing capacity (Navon & Gopher, 1979; Gopher, 1986, Wickens,
1992). Therefore, later research (Britton, Muth & Glynn, 1986, Cennamo, 1992; Jahns,
1973) has identified the components of mental effort as attention, concentration,
cognitive capacity, and mental workload. Salomon (1984) defined mental effort as “the
number of non-automatic elaborations necessary to solve a problem”. Gimino (2000)
clarified the term “nonautomatic elaborations” and explained it as “the retrieval of
relevant declarative knowledge (i.e., knowledge about related concepts, processes and
11

principals) and the subsequent construction and tuning of action and decision steps in a
procedure to achieve a learning goal.” Salomon’s definition implies that the amount of
mental effort investment relies on individuals’ ability to retrieve knowledge in the long
term memory and ability to process the information. The required mental effort is
depended on novelty or familiarity of the learning task or problem. When the task is
familiar to the person, the relationship between his/her expertise and performance is
direct due to the automated cognitive process or very low mental effort investment.
However, when the task is unfamiliar to the person, the relationship between his/her
expertise and performance is indirect due to the cognitive process is non-automated and
mental effort investment is required.
Sweller (1988, 1994, 1999) proposed cognitive load theory and explained how
individuals process the information or solve the problem. People retrieve schemas
automatically in the long-term memory if the information or task is familiar. However,
when new and unfamiliar information is dealt, people’s working memory is limited in its
capacity for mental processing.
Levels of expertise and mental effort
Experts possess a large number of domain-specific schemas and are able to bypass
working memory capacity limitations because many of their schemas are highly
automated (Kalyuga, Ayres, Chandler, & Sweller, 2003).The difference between an
expert and a novice is that a novice hasn't acquired the relevant schematic knowledge as
an expert’s. The automation is achieved through extended and deliberate practice
12

(Anderson, 1995; Ericsson, Krampe, & Tesch-Romer, 1993). In other words, the change
in performance occurs because as the learner practices extensively and becomes
increasingly familiar with the material, the cognitive characteristics associated with the
material are altered so that it can be handled more efficiently by working memory (Paas,
Renkl & Sweller, 2003). The reduction of the demand of working memory capacity then
leads to the improvement of performance. However, studies also showed that experts
only can perform accurately and efficiently in the specific task domain when the task
patterns are familiar, meaningful, and recognizable (Chase & Simon, 1973). Experts may
commit judgmental biases if the task pattern is unfamiliar even in the same familiar task
domain. Joyce and Biddle (1981) conducted a series of experiments to examine the
probability assessments of experienced auditors. One experiment was about performing a
simple task but one that did not involve the type of decision making that auditors
typically encounter. The results showed the auditors had a strong anchoring bias.
However, when the tasks were more analogous to typical audit judgments, experts
displayed no significant fault judgment.
Weber and Brewer (2003) argued that the experiment of Chase and Simon actually
changed the domain-relevant structure of the stimuli. Therefore, they conducted another
similar experiment but directed experts’ and novices’ attention to different tasks under
the same domain-relevant structure. One group was instructed to judge which hockey
team had advantage in a competition based on the given information and the other group
was instructed to recall each team member’s name and the number of times the name
appears. The first group was believed to be sensitive in domain-relevant structure
13

information and the second group was not. Finally, the participants were required to
answer questions on one response sheet. The result showed that the experts’ recall
performance was better under high degree of domain-relevant structure which is
consistent with the previous studies. Moreover, the experts’ recall performance would
decrease when the experts were instructed not to attend to the information but to the
members’ names. In conclusion, the effect of expert memory advantage should be the
result of a three-way interaction between expertise, domain-relevant structure, and
attention. This implies that experts may not attend to the meaningful information within
the domain when the primary task does not require their attention.
Levels of Task difficulty and mental effort
According to Chandler and Sweller (1996), the nature of the task or material
imposes different mental demands on the learners. For example, some material imposes
very low cognitive load and an example of such a task is the learning of the basic
vocabulary of a second language. Each element or schema is independent from the others
with no interactivity and subsequently the required mental processing is low (Sweller &
Chandler, 1994). Tasks that have low element interactivity can be learnt serially rather
than simultaneously. Tasks with a high degree of element interactivity have a heavy
intrinsic cognitive load and an example of such a task is the learning of the grammar of a
second language as all the words in the phrases need to be considered, that is processed,
at once (Garner, 2000). Working memory load is affected by the inherent nature of the

14

material and by the format in which the material is presented (Sweller, van Merrienboer,
& Paas, 1998).
Two important mental load characteristics from complex cognitive tasks are the
number and nature of the component skills involved and the complexity of the goal
hierarchies of the problems that must be solved in the task domain. Skills which include
many component skills usually show higher mental demands than skills with less
components and increasing complexity of the goal hierarchies of the solutions to the
problem also produce high cognitive load. Paas, Renkl, and Sweller (2003) referred the
mental load imposed by the task as intrinsic cognitive load and considered that it only can
be reduced by constructing additional schemas and automating previously acquired
schemas.
Van Merrienboer (1997) introduced the broader concept of task classes to define
simple-to-complex categories of learning tasks. Learning tasks within a particular task
class are equivalent in the sense that the tasks can be performed on the basis of the same
body of generalized knowledge. A more complex task class requires more knowledge or
more embellished knowledge for effective performance than the preceding and simpler
tasks classes. For a simpler task class, less elements and interactions between elements
need to be processed simultaneously in working memory when performing the tasks in
simpler task classes; for more complex task classes, the number of elements and
interactions between elements increases (Pollock, Chandler, & Sweller, 2002). Therefore,
as the number of items that must be processed in working memory increase, mental effort
must necessarily increase if learning is to be successful (Clark, 1999). The result implies
15

that judgment of task difficulty appears to directly influence the amount of mental effort
invested in a learning task (Salomon, 1984; Clark & Elen, 2006).
The effect of task demand and imposed mental load on subjects’ amount of mental
effort investment can be changed. Pollock, Chandler, and Sweller (2002) suggested that
an isolated elements instructional technique that allows novices to circumvent working
memory limitations by initially presenting complex material as a collection of individual,
isolated elements of information. If element interactivity is artificially reduced in this
way, some partial schemas for the presented information may be developed first,
allowing novice learners to reduce the working memory load during a subsequent attempt
learn from original, high-element interactivity material. Mayer and Chandler (2001) used
scientific instructional materials and found that initially permitting inexperienced learners
to artificially control the speed of an animation and thus allowing the assimilation of
isolated elements within specific animation frames benefited learners far more than
initially presenting learners with a high-element interactivity animation, which they could
not control. Subjects also perceive task difficulty reduced through practicing (Kalyuga,
Ayres, Chandler, & Sweller, 2003). De Crook et al. (1998) studied the effects of
contextual interference on transfer performance and invested mental effort. He found that
subjects in the higher context interference group showed higher performance and lower
invested mental effort on far transfer test problems, relative to subjects in the low
contextual interference group. The reason is that learners practice different task variations
results in multiple and variable encoding processes (Shea & Morgan, 1979; Shea &

16

Zimmy, 1983). These practices are believed to increase learners’ memory and to decrease
the dependency on reinstating the encoding context (Lee & Magill, 1983, 1985).
Perceived self-efficacy and mental effort
Generally, people who perceive that they have the necessary skills and knowledge
will choose to invest more effort in a task. They will withhold the effort if they do not
perceive that they have the adequate knowledge and skill (Pajares & Miller, 1995). Many
studies of the relationship between self-efficacy and effort investment have reported the
relationship between self-efficacy and mental effort is a linear and positive (Covington,
1992; Bong, 1997, 1999; Wigfield and Eccles, 1992; Schunk, 2003). For example,
Brookhart et al. (2002) investigated the relationship between self-efficacy, mental effort
investment, and performance for elementary school students and reported that there is a
positive relationship between mental effort investment and perceived self-efficacy, and
both of them have positive relationship with students’ social studies and science
performance.
However, other researchers (Clark, 1999; Gimino, 2000; Yoshida, 2000) have
claimed that the relationship between self-efficacy and mental effort is an “inverted U”
(see Figure 1 below). Vancouver and Scherbaum (2002) found this phenomenon in his
computational model study. They claimed that individuals with very low self-efficacy do
not choose to accept the goal and allocate no cognitive resources towards it. When the
individual has enough self-efficacy to accept the goal, he or she anticipates that
substantial resources will need to be applied to accomplish it. Finally, the individual only
17

acquires few resources when he or she perceives that the goal is relatively easy for him or
her to achieve.
Clark (1999) suggested that overconfidence could be defined as high perceived
self-efficacy and low mental effort investment on a novel task. He explained that novel or
unanticipated challenges require a great deal of mental effort to succeed and the amount
of invested mental effort is determined by people’s perceived self-efficacy for specific
tasks. People who lack confidence and those who are overconfident tend not to invest
much mental effort in a task. An overconfident person thinks that he knows what he is
doing, so he does not work very hard on a task which leads to poor performance. One
example comes from Salomon’s research. Salomon (1984) conducted a study of 124 sixth
graders to test the relationship between self-efficacy, amount of invested mental effort
(AIME), and learning from either televised film or printed text. Students were assigned to
view a silent film group (TV group) or to read a comparable printed text group (print
group). Students reported significantly higher levels of self-efficacy for learning from
television than from print in the pretest assessment. The posttest result demonstrated that
self-efficacy was positively and significantly related to the amount of mental effort
investment and to achievement for the print group, but negatively related to amount of
mental effort investment and achievement for the TV group. This finding indicated that
subjects invested less effort in television because their self-efficacy indicated that it made
learning easier for them and more effort in print, which they perceived as more difficult.
18

Figure 1 Suggested inverted U relationship between self-efficacy and mental effort
(Gimino, 2000)

High ME

Low ME
Very low SE low-med SE Very high SE

Summary
DeMarie et al. (2004) have suggested that without adequate measurement of prior
knowledge, one cannot determine whether a non-linear or linear model is a better
estimate of the relation between self-efficacy and learning. Pajares and Miller (1995)
supported that people who perceive that they have the necessary skills and knowledge
will choose to invest more effort toward the task. They will withhold the effort if they do
not perceive that they have the adequate knowledge and skill. However, Clark (1999)
asserted that if people believe that they are skilled on a task, their high self efficacy
judgment will result in only low levels of mental effort. Individuals who have strong
self-efficacy usually invest some effort to perform a task and if their efficacy is accurate
they succeed. People withhold their effort from a task when their self-efficacy is low. On
the other hand, individuals with high levels of self-efficacy for a given task may invest
low mental effort if their perceived expertise is high and apply automated knowledge that
requires no mental effort in order to accomplish the task. Therefore, although a
19

considerable amount of research has proposed that the relationship between perceived
self-efficacy and mental effort investment is linear, a number of motivation researchers
suggest that the shape of the relationship between self-efficacy and mental effort is most
often an “inverted U”.
As discussed earlier, experts tend to solve problems they consider familiar by the
automatic processes that characterize expertise (Retamero & Dhami, 2009; Anderson,
1993). Therefore, it may be inferred that people who consider themselves experts but
actually they are not and people who misjudge the difficulty level of the task tend to
invest low mental effort and express overconfidence toward the task.

Conclusion

This literature review discussed the factors affecting the judgment of self-efficacy
and the efficacy judgmental disparity problem of overconfidence, as well as, their
influences on mental effort and task performance. It is proposed that individuals’ levels
of expertise and levels of task difficulty play an important role in determining
self-efficacy and relate to the generation of overconfidence. Studies also show that people
with too much efficacy or confidence reduce their effort on the task that causes decreased
task performance. After reviewing the evidence, a model is suggested below is to briefly
illustrate the relationship among levels of expertise, perceived self-efficacy, levels of task
difficulty, mental effort, and performance.

20

Figure 2 Suggested Relationship between Expertise, Perceived self-efficacy, Perceived
task difficulty, Mental Effort and Performance

Perceived Task Difficulty

Expertise Perceived
Self Efficacy (Self-Confidence)

Mental Effort

Performance
21

The Instrument of the Variables Measurement

Mental effort Measures
Self-report measure
Most measurement instruments for mental effort request self report of the amount of
effort subjects believe they have used to solve problems, learn or transfer knowledge to
new tasks. Salomon (1983) used one direct question “How much effort have you put in
comprehending this story?” and one less direct question “How much did you concentrate
while you were reading this story?” in assessing children’s amount of mental effort
investment (AIME) in the reading comprehension with two different media (TV & Print
text). The AIME measure used a 4 point effort scale from “Not at all” to “A lot of” effort.
Paas (1992) also describes one such measure which involves asking subjects to report the
amount of mental effort they invest in each problem-solving task. His measure used a 9
point effort scale that varies from “very, very high mental effort” to “very, very low
mental effort”.
However, studies have found that the reliability of self-report measure is still in
question. For example, Salomon reported high reliability of his AIME scale at a=. 81
while Cennamo, et al. (1991) only found a reliability of a=.55 for AIME in her study.
Gimino (2000) adapted both self-report measures and a dual task measure. She described
that the subject’s indicated high mental effort investment on the self-report measures
while the dual task measure results were different. The dual task measures of mental
effort are discussed next.
22

Dual-task measure
Some researchers (Brunken et al., 2002, 2003; Chandler & Sweller, 1996) consider a
dual task measure to be a more promising measure of cognitive workload than the
self-report mental effort measure. The dual task paradigm requires that subjects be
interrupted by a tone or a visual image while they are learning and asked to quickly strike
a key. The speed with which they react to the interruption is assumed to represent the
amount of mental effort they are investing in the primary task (Meyer & Kieras, 1997;
Navon & Miller, 2002). When subjects react quickly it is assumed that they are not fully
engaged or investing much effort in the task that was interrupted. The slower their
response, the more effort they are presumed to be investing. This means that when the
primary task demands more mental load or effort from the subjects, the reaction time to
the secondary task from the subjects should take longer. Conversely, subjects response
the secondary task faster if the primary task only requires low amount of effort from the
subjects.
One of the dual task measurement processes is that a baseline measure (B.M.) of the
subjects’ reaction time to the secondary task is recorded first, after that, the subjects are
asked to begin the primary task and then strike a key when manipulating the secondary
task. The difference between their reaction time is viewed as the amount of mental effort
or cognitive workload. For example, Kee and Davis (1988) used a secondary task that
asked subjects to tap a key as quick as possible with their finger when they were learning.
They recorded subjects’ finger taping performance first as their baseline scores. After

23

asking subjects to perform the primary taskwhen they were manipulated, they found the
response time increased when the primary learning task became more novel.
Another dual task measure is developed to record the stimuli-response time of the
subjects (Gimino, 2000). Brunken, et al. (2002) also confirmed that the stimuli- response
performance of dual task measure is significant enough that subjects delayed their
response time to the randomly presented computer stimuli when they are engaged in the
primary task.

Perceived Task Difficulty Measure

In his AIME study, Salomon (1983) developed the self-reported perceived task
difficulty scale questions for measuring students’ perception of task difficulty level about
two different media (TV vs. Print). Gimino (2000) reported moderate to high reliability (a
= .7907 for set 1 to a = .9192 for set 5) of Salmon’s perceived task difficulty scale in his
study. The reported levels of task difficulty also had a significant relationship to the
amount of mental effort investment.

Expertise Measures

Traditional methods used to evaluate learners’ expertise depends on tests involving
the solution of a series of problems representing a given task domain. However, experts
are often not able to state precisely what they know and what they do to become experts
(Villachica et al. 2001). Self reporting for individuals to estimate their levels of expertise
24

by asking them relevant domain knowledge is usually not reliable and time consuming.
To avoid the problem, the method of cognitive task analysis, such as “think-aloud”
approach (Rowe, et al., 1996) and a concept map (Schvaneveldt, 1990), has been as a
way of complementing the self report for measuring the expertise (Chipman, Schraagen,
& Shalin, 2000). Another method used to assess levels of expertise is to measure
performance accuracy and time consumed in performing the task (McCloskey, 1983).
This method is based on actual performance that can be used to reflect the expertise
knowledge and the underlying performance. However, time and accuracy measures can
only provide indirect evidence of measuring expertise knowledge without being able to
demonstrate people’s direct knowledge of their expertise (Staggers & Norcio, 1993).
Kalyuga and Sweller (2004) pointed out the problem with traditional methods for
measuring domain-specific knowledge. They argued that traditional expertise measure
may emphasize observable the elements of the problem state instead of on the appropriate
steps experts use to solve problems (Chase & Simon, 1973; Ericsson & Kintsch, 1995).
This means that automated procedural knowledge is difficult and time consuming to
measure. A more appropriate schema-based measure of long term working memory
content should be considered. In their experiments, subjects were asked to indicate an
immediate next step towards solution rather than asking for the complete solution steps
for the task. The results were promising and validated the expertise measure. However,
the study was limited to well-defined mathematical tasks with predictable sequences of
solutions steps. It raised the generality problem when applied to more complex tasks and

25

suggested that the measure should be tested to verify its reliability and validity extent on
different problem areas.

Perceived Self Efficacy Measures

In his study of self efficacy measurement scales, Bandura (1997) recommended that
presenting subjects with items portraying different levels of task demands, and ask them
to rate the strength of their belief in their ability to execute the requisite activities. The
items are phrased in terms of can do rather than will do. Bandura cautioned that
researchers need to draw on conceptual analysis and expert knowledge of what it takes to
succeed in a given pursuit before designing the scales. Pajares (1996) were indicated that
domain-specific assessments, such as asking students to provide their confidence to learn
mathematics or writing, are more explanatory and predictive than omnibus measures and
preferable to general academic judgments, but they are inferior to task-specific judgments
because the sub-domains differ markedly in the skills required.
The self-efficacy scale developed by Bandura records the strength of people’s belief
on a 100-point scale, ranging in 10-unit intervals from 0 (“Can not do”); through
intermediate degrees of assurance, 50 (“Moderately certain can do”); to completely
assurance, 100 (“Certainly can do”). The efficacy scores will be calculated by summing
the total scores and divided by the total number of items to indicate the strength of
perceived self efficacy for the activity domain.

26

Significance of the problem

Studies have shown that students overestimate their knowledge of an area of study
(Flannelly, 2001; Pressley & Ghatala, 1990). The overconfidence problem seriously
affects students’ perception about their expertise and need to learn.
The judgment of self-efficacy is highly correlated with task difficulty judgment
(Bandura, 1997, 1986; Pulford & Colman, 1997, 1996). However, rare studies on
overconfident problem examined the influence of self-efficacy and expertise at the same
time (Beach, et al., 1986). Therefore, the current study intends to study the problem of
overconfidence by including both variables - levels of expertise and levels of task
difficulty, and consider how these two variables affect perceived self-efficacy judgment
and their relationship with mental effort investment and task performance.

Purpose of the Study

The purpose of this study is to address on the following issues:
1. Determine the relationship between expertise, levels of task difficulty, and
self-efficacy judgments
2. Determine the relationship between levels of task difficulty, perceived
self-efficacy, expertise, mental effort, and task performance.

27

Research Questions

1. What is the relationship between expertise, levels of task difficulty, and
self-efficacy judgment?
2. What is the relationship between levels of task difficulty, perceived self-efficacy,
expertise, mental effort, and task performance?

Hypotheses

H1a: There is a positive relationship between levels of task difficulty and mental effort.
The relationship is illustrated in Figure 3.

Figure 3 The relationship between levels of task difficulty and mental effort

Levels of task difficulty
High

H1b: There is a negative relationship between levels of task difficulty and task
performance. The relationship is illustrated in Figure 4.
Mental Effort
Low
Low High
28

Figure 4 The relationship between levels of task difficulty and task performance
Levels of task difficulty
High
Low
Task Performance
Low High

H2a: There is a negative relationship between mental effort and perceived self-efficacy
from self-reported measure results. The relationship is illustrated in Figure 5.

Figure 5 The relationship between perceived self-efficacy and self-reported mental
effort

Self-Efficacy
High
Low
Mental effort
Low High

29
2b: There is an inverted U relationship between perceived self-efficacy and mental
Figure 6 The relationship between perceived self-efficacy and dual-task measured mental
effort
low SE low-med SE Very high SE

2c: There is a positive relationship between perceived self-efficacy and task
igure 7 The relationship between perceived self-efficacy and task performance

H
effort from dual-task measure results. The relationship is illustrated in Figure 6.

High ME

Low ME
Very
H
performance. The relationship is illustrated in Figure 7.

F
Task Performance
Self-efficacy
Low
High
Low High
30

3b: There is a positive relationship between expertise and task performance. The
trated in Figure 9.
Figure 9 The relationship between expertise and task performance

H3a: There is a negative relationship between expertise and mental effort. The
relationship is illustrated in Figure 8.
Figure 8 The relationship between expertise and mental effort

H
relationship is illus
Task performance
Expertise
Low
High
Mental effort
Expertise
Low High
Low
High
Low High
31

H4a: There is a significant difference between levels of task difficulty, expertise, and
mental effort. There is a positive relationship between levels of task difficulty and
mental effort while there is a negative relationship between expertise and mental
effort. Both levels of task difficulty and levels of expertise contribute the prediction
to mental effort. The relationship is illustrated in Figure 10 and Figure 11.
Figure 10 The relationship between exp fort.

Figure 11 The relationship between levels of task difficulty and mental effort

ertise and mental ef
Expertise
Low
High
Mental effort
Low High
Mental Effort
Levels of task difficulty
Low
High
Low High
32

here is a negative relationship between levels of task difficulty and
task performance while there is positive relationship between expertise and task
performance. Both expertise and levels of task difficulty contribute the prediction to
task pe ce. The relationship is illustrated in Figure 12 and Figure 13.
Figure 12 The relationship between expertise and task performance

y and task performance

H4b: There is a significant difference between levels of task difficulty, expertise, and task
performance. T
rforman
Figure 13 The relationship between levels of task difficult
Task performance
Expertise
Low
High
Low High
Task Performance
Levels of task difficulty

Low
High
Low High
33

H5a: There is a significant difference between levels of task difficulty, perceived
self-efficacy, and mental effort. There is a negative relationship between levels of
task difficulty and perceived self-efficacy. There is a positive relationship between
levels of task difficulty and mental effort while there is a negative relationship
between perceived self-efficacy and mental effort. Both levels of task difficulty and
perceived self-efficacy contribute the predictio he relationship is
illustrated in Figure 14 and Figure 15.

Figure 15 The relationship between levels of task difficulty and mental effort
n to mental effort. T
Figure 14 The relationship between perceived self-efficacy and mental effort

Mental Effort
Levels of task difficulty
Low
High
Low High
Mental effort
Perceived self-efficacy
Low High
Low
High
34

H5b:
ere is a negative relationship between levels
ween
mance

Figure 17 The relationship between levels of task difficulty and task performance

There is a significant difference between levels of task difficulty, perceived
self-efficacy, and task performance. Th
of task difficulty and task performance while there is a positive relationship bet
perceived self-efficacy and task performance. Both levels of task difficulty and
perceived self-efficacy contribute the prediction to task performance. The
relationship is illustrated in Figure 16 and Figure 17.
Figure 16 The relationship between perceived self-efficacy and task perfor
Task Performance
Levels of task difficulty
Low
High
Task Performance
Perceived self-efficacy
Low
High
Low High
Low High
35

H6a: There is no significant difference between expertise, perceived self-efficacy, and
mental effort from self-reported measure results. However, there is a negative
relationship between expertise and self-reported mental effort as well as there is a
ceived self-efficacy and mental effort. The
relationship is illustrated in Figure 18 and Figure 19.
Figure 18 The relationship between perceived self-efficacy and mental effort

Figure

H6b: There is a significant difference between expertise, perceived self-efficacy, and
mental effort from dual-task measure results. There is a positive relationship
negative relationship between per

19 The relationship between expertise and mental effort
Mental Effort
Expertise
Low
High
Mental effort
Perceived self-efficacy
Low
High
Low High
Low High
36

re is an inverted U
expertise and perceived self-efficacy contribute the prediction to dual-task measured
mental effort. The relationship is illustrated in Figure 20 and Figure 21.
Figure relationship between perceived self-efficacy and dual-task measured
mental effort
Figure

H6c: and task

between expertise and perceived self-efficacy while the
relationship between self-efficacy and dual-task measured mental effort. Both
20 The

21 The relationship between expertise and mental effort
There is a significant difference between expertise, perceived self-efficacy,
performance. There is a positive relationship between expertise and perceived
Mental Effort
Expertise
Low
High
Mental effort
Pe elf-efficacy rceived s
Low High
Low
High
Low High
37

n to task performance.
icacy and task performance
Figure 23

self-efficacy. There is a positive relationship between expertise and task
performance as well as between perceived self-efficacy and task performance. Both
expertise and perceived self-efficacy contribute the predictio
The relationship is illustrated in Figure 22 and Figure 23.
Figure 22 The relationship between perceived self-eff

The relationship between expertise and task performance
Task Performance
Expertise
Low
High
Low High
Task Performance
Perceived self-efficacy
Low
High
Low High

38

Table ns 1: Hypotheses Matrix for Possible IV and DV Combinatio
Levels of task difficulty
DV self-efficacy Easy Moderate High
IV Expertise Perceived

D D D

D D

Mental effort
D

D D D

D D D D

D D D

D D

Performance
D
D D D

D D D D

Definition of Terms

Self-efficacy: Beliefs about one’s capabilities and confidence to learn or perform
behaviors at designated levels (Bandura, 1986, 1977).
Overconfidence: Overconfidence refers to the failure to know the limits of one’s
knowledge and task demand so that they believe in a capability to perform a task may be
overestimated (Bandura, 1997; Russo & Schoemaker, 1992).
Mental Effort: Tthe number of non-automatic elaborations necessary to solve a
problem (Salomon, 1984).

39

ive knowledge (i.e.,
t cepts, pr princip
uction and tuning of action and s in a ure ing
ino, 2000).
Self-knowledge jud nt: Subjects’ se rception-ba d knowle ge of the
eed at task articular dom in (Gravill ., 2002
Task-specific Self-efficacy: Beliefs about one’s capac perfo speci ask
Non-automatic elaborations: The retrieval of relevant declarat
knowledge abou related con ocesses and
decision step
les) and the subsequent
constr proced to achieve a learn
goal (Gim
gme lf-pe se d ir
ability to succ s in the p a et al ).
ity to rm a fic t
(Bandura, 1997).

40

CHAPTER II METHODOLOGY
ficacy judgments. Another purpose is to determine how the
, sampling, instrumentation, issues of validity and reliability, and
procedures for data collection and analysis.

Apparatus

In order to measure subjects’ mental effort investment, a dual-task measure was
used and subjects received a primary task and were interrupted of a secondary task. To
measure subjects’ levels of expertise and performance objectively, the measures of
primary task performance was assessed to reflect subjects’ performing accuracy and
speed (Ericsson & Charness, 1994). A strategic computer game called WarCraft III was
used as the primary task. A developed program that can randomly present a pop-out
window on the computer screen was employed as a secondary task to interrupt subjects’
engagement of the primary task and to measure subjects’ reaction time which is a
measure of mental effort investment.

The purpose of this study is to determine how different levels of task difficulty and
expertise affect self-ef
self-efficacy judgments affect the amount of mental effort investment and task
performance under different levels of task difficulty.
This chapter includes a description of the research methodology, experimental
design procedures
41

WarCraft III
WarCraft III was developed by Blizzard Entertainment Corporation and is classified
uses.
e
“Human” race. The
“Hum
ere
gain war advantages and strengths with
limited gold mine (resources) and food. The abilities to utilizing the resources and
personnel, upgrading the heroes’ levels taining and using the treasures
effic
of

task by task in order to teach the players
verything about the race. The task usually requested the player is to find one enemy hero
r save another captured Human hero from one location to another. During the search

as a strategic computer game. There are four types of races for players to select and each
of the races differs from the others in terms of heroes, units, weapons, and strategy
Players can play the game in two ways. First, players can choose to play the gam
by following its script. The script describes a story starting from
an” race encountered an attack from “Undead” race. Therefore, the leader (Hero) of
the “Human” race needs to protect his race and to beat the enemy. In order to do so, th
are many tasks the hero needs to complete to
quickly, ob
iently are the key factors to win the opponent.
At the beginning, the Human hero needs to create his base, slaves, and soldiers for
the war. Therefore, the players need to know how to do so and to learn the functions
different weapons, buildings, and soldiers. A dialog box displaying on the screen
instructs how to start the base first. In addition, there are some graphic icons representing
treasures or weapons the player can use to fight the enemy. A text box also appears
beside each graphic icon for the explanation when the player moves the cursor on it.
Players therefore learn the functions of the icons and how to use them. The story
continues to ask the player to complete
e
o
42

journey, the Human hero usually encounters attacks from enemy or other wild monsters.
succ
e

of the
in the search journey, the player
need

its

to the
started
The Human hero then needs to use the weapons and trained soldiers to kill the enemy and
monsters. However, the task difficulty increases depending on the requested tasks in
sequence. Most of the time, new weapons and soldiers are provided that the player can
essfully complete the new task. As a result, when starting a new task, the dialog box
appears again on the screen to instruct what new resources, weapons, and strategies th
player should use in order to complete the task. For example, the Human hero needs to
fly across a mountain in order to destroy the enemy base in one of the sequential tasks.
The dialog box instruction would show the player how the flying ship can be obtained
and use. The player therefore can learn all of the strategies, weapons, and resources
Human race after completing all of the tasks.
If unfortunately, the Human hero lost the battle
s to recreate the hero and soldiers to redo the task until the task is completed.
However, it is possible that the players’ base is destroyed by the enemy first. The reason
is when player is doing the task; the opponent (computer) is also trying to increase
strength and to stop the player to complete the task. The player will not allow continuing
to the next task if he or she cannot complete the prior task. When the Human hero
successfully completes the mission, his expertise level is upgraded to a higher level.
The Human race story ends when the hero found a treasure to transform him
most powerful “Undead” leader. He then killed the original “Undead” leader and
invade the other races. Because of this story arrangement, the player starts learning all of
the strategies, weapons, and resources of “Undead” race in a similar way the player
43

player has a faster speed in searching the
enem
e
p
selec )

learned in Human race story. The final story ends when the player learned all of the
strategies, weapons, and resources of the four races (Human, Undead, Orc, Night Elf). In
a word, the script option provides step-by-step learning procedures of the game.
After learning the game from the script option, players can choose to play a
“customized game” in WarCraft III. As what happens in the script option, players are also
required to beat the opponents to win the game. The difference is that players need to use
all of knowledge and skills he or she learnt from script option without help. For
example, players need to know how to build up their supply bases, to train their soldiers
and slaves, to investigate the environment, to trace who their enemies are, and to decide
which strategies they should use to beat their opponents. They should have learnt all of
these skills in the script option. An experienced
y, advancing the weapons, training the soldiers, upgrading heroes’ levels, and
utilizing the resources. He or she also is able to judge accurately in the number of
soldiers’ creation, weapon usages, and the priority of the resource usages.
It is possible that the player only has experience in playing one or two race(s) in th
script option. In this case, the player can only select the race(s) he or she learnt in the
script option and plays the “customized game” with the race(s).
The “customized game” also provides different task difficulty choices, various ma
tions, and race options. As a result, players can assign their favorite or skilled race(s
to themselves or to their opponent (computer), choose the maps, and decide which
difficulty level they want to play. In addition, the choice of the “customized game” also
allows multiple players to play with computers at the same time. In this research, the
44
ing
the g
ses
ed by
On the other
hand
’
e computer. Besides,
the o
e

researcher used two different maps in the “customized game” and designed the
experiment only for single player.
Although the game does not restrict the playing time, players are actually play
ame with a time limit. The reason is when players are generating strategies,
resources, and units; the opponent (computer) is also generating its strategies, resources,
and units to attack players. If the opponent (computer) is well-prepared, it will directly
attack the player and destroy the player’s base. The game ends when one party’s ba
are completely destroyed.
One of the important features for the choice of the “customized game” is to allow
players to control the health point (hp). The hp point determines the strength of attack
during the battles and the speed of the soldiers’ recovery. The manipulation of this
function allows the researcher to create the most difficult task that is impossible for
subjects to win and named this task level “High”. The “High” task level was design
assigning subjects’ race hp point 50% from regular assigned 100% hp point.
, the hp point of computer’s race still stayed in default condition (computer: 100%
hp). This design slows the player’s speed in resource creation, unites creation, soldiers
and heroes’ recovery that makes the player very difficult to win th
riginal customized game also provides different levels of task difficulty choice; the
researcher used this feature to include two tasks from “Easy” to “Moderate” difficulty
levels.
There are a number of reasons for the researcher to have chosen this game. First,
WarCraft III provides four types of races and each of the races can be considered as on
45
the game
layers can also be experts in more than one race,
depe
ame

lows the researcher to manipulate different levels of task
peed.

The use of the secondary task is to measure subjects’ amounts of mental effort
investment. The secondary task was designed to make subjects be interrupted by a tone or
a visual image and the subjects should response the interruption by pressing a key as
quickly as possible. The speed with which the subjects react to the interruption is
assumed to represent the amount of mental effort they are investing in the primary task
(Meyer & Kieras, 1997; Navon & Miller, 2002). This means that when the primary task
demands more mental load or effort from the subjects, the reaction time to the secondary

domain. Players would act as a novice in one race if they never or rarely played
with this race before. Conversely, p
nding on the frequency with which the players have practiced. In common cases,
players can be the experts only for one race (domain). As a result, the use of this g
can help the researcher to distinguish the players’ actual levels of expertise based on the
assigned race (domain) by the researcher. Second, this game is a strategic game that
requires players to use their cognitive resources and effort to generate strategy and solve
the problem. Third, the game al
difficulty as discussed before. Finally, WarCraft III is very detailed in providing
performance information and the length of players’ elapsed time to complete the task.
These features allow the researcher to evaluate players’ performance accuracy and s
Secondary Task

46
d
e
d,
rimary
et
available free cognitive capacity. To meet these criteria, the measure of reaction time is
proved to be used successfully in previ emory research (Verwey &
Veltman, 1996; Wickens, 1984). The learner has to react to a specific signal during a very
simple continuous monitoring task as soon as possible. This design minimizes the
interference between the tasks and maximizes the exhaustion of the free capacity. Thus,
reaction time in a secondary monitoring task is a valid measure of cognitive and mental
workload.
Based on these principles, a pop-out window program designed by a student from
the Computer Science Department at one of the universities in Southern California area
will be used as secondary task for this experimental design. The components of the
program include:

task from the subjects should take longer. Conversely, subjects response the secondary
task faster if the primary task only requires low amount of effort from the subjects.
The design of the secondary task needs to meet some criteria of methodological an
technical challenges (Brunken et al., 2003). First, for the secondary task to be a sensitiv
measure it has to require the same cognitive resources as the primary task; otherwise,
secondary task performance will be independent of primary task performance. Secon
the performance measure for the secondary task has to be reliable and valid. Third, the
secondary task has to be so simple that it does not suppress simultaneous learning
processes; otherwise, the secondary task may affect the learning outcome of the p
task by requiring cognitive resources that are no longer available for learning (Marcus
al., 1996). Finally, the secondary task has to be able to consume flexibility of all of the
ous working m
47
2.
Instrumentations
n,
total scores were summed from three categories: unit scores, resources scores,
and h

1. A pop out window with “tone” sounds that can randomly appear in different
locations each time.
A function for recording the time when the window popped out as well as the time it
took the subjects to respond by pressing the computer space-bar.
3. The window can automatically disappear after the subjects press the space-bar to
respond.

Computers

The researcher used two laptop computers with mouse in the experiment. One of
them had WarCraft III Frozen Throne version installed and the researcher installed the
pop out windo.w program onto the other computer.

Expertise
Subjects’ level of expertise was measured based on their performance accuracy and
speed on two pre-tests with “Moderate” task level (Ericsson & Charness, 1994; Anderso
1995). The
ero scores. The total scores were used to interpret subjects’ expertise levels.

48
ts with several items were developed by using Bandura’s (1997)
by rating themselves
ndix C). Subjects were asked to respond to a 100 point scale that ranges from
) cannot win/beat at all, (50) Moderately certain can win/beat, to (100) certainly can
win/beat. Subjects’ scores across items problem set were summed to
dete
asured by using both the self report
and dual task measures. The self report measure used Paas’s (1992) and Salomon’s (1983)
mental effort scale. Paas’s (1992) ra sted and ranged from (1) No effort,
(4) Moderate effort, to (7) Extreme effort (see Appendix D). The researcher also adapted
Salomon’s (1983) effort measurement questions and restated the questions as: (1) How
much effort, or work, did you put into trying to play the game? (2) How much did you
concentrate while playing the game? The rating range for the first question is from (1) No
effort to (4) A huge amount of effort, and for the second question is from (1) Not at all to
(4) Extremely Concentrating. Although self-ratings may appear questionable, it has been
emonstrated that people are quite capable of giving a numerical indication of their
erceived mental burden (Gopher & Braune, 1984). The study also used and modified the

Perceived self efficacy
Two problem se
guide to construct self efficacy scales to measure subjects’ confidence on the game
performance. Subjects needed to answer how confident they were
on the items such as the estimated playing time and predictions of performance scores
(see Appe
(0
within each
rmine a total self efficacy scores.
Mental Effort Investment
The variable of mental effort investment was me
ting scale was adju
d
p
49
proposed by Salomon (1983) to measure how much difficulty the
subj
sk
n of
process. The time elapsed between the time when
the pop-out window appeared and the subjects’ response time was recorded to reflect the
required cognitive task demand of the subjects.

The study included 66 volunteers from colleges in Los Angeles, Orange, and San
Bernardino counties. Subjects were selected based on their experiences in playing the
computer game WarCraft III. A background questionnaire was also distributed to identify

7 point liker-type scales
ects felt regarding the completion of the tasks (see Appendix E ). After completing
five trials, subjects were required to answer six questions about their perception of ta
difficulty, such as “How easy was the “Moderate” level task?” and “How difficult was
the “Moderate” level task?”
In the mean time, the researcher adapted the dual task measure to avoid possible bias
of self report. Subjects were requested to respond to the secondary task by pressing the
keyboard ten times whenever the pop-out window randomly appeared on the scree
the other computer during the playing
Performance
Subjects’ performances were determined based on their performances on the tasks
(win or fail the task) and performance scores of three tasks (Easy, Moderate, High) in the
second phase of the experiment.
Subjects

50

Design
This research used quantitative approach for the study. The study had two phases
consisting of five trials. There were two trials in the first phase and three trials in the
second phase. It took ten minutes to complete each trial. In the first phase, two pre-tests
were given to determine subjects’ levels of expertise. The task difficulty level was set at
“Moderate” in these two pre-tests. After taking a ten minutes break, subjects then started
n the second phase and were required to complete the given task with
diffe
only had information about what their race was and did
not have any idea about what the computer’s race was prior to each trial. In the first phase,
a map named “Ogre Mound“ was used and a map named “Booty Bay” was used in the
second phase.
First Phase

the volunteers’ ages, gender, race, and experiences in playing WarCraftIII Frozent
Throne version (see Appendix B). Questions like “please specify your degree of the
familiarity with the four races in order” and “How many hours do you play with the
race(s) per week?” were used to identify the subjects’ game playing experiences.

the third trial i
rent levels of task difficulty (Easy, Moderate, High). During the two phases, The
“Human” race was set as subjects’ default race and computer’s race was randomly
selected. This means that subjects
Subjects were asked to fill out background and perceived self-efficacy
questionnaires before the first trial. There were two laptop computers in front of the
subjects. One laptop has the strategic computer game WarcraftIII Frozen Throne installed
51
our
ow
total
scores by clicking the “save replay” icon on the computer screen when the researcher
asked them to stop the trial. Subjects then played the game again in the second trial and
needed to respond to a randomly appearing pop-out window by pressing a spacebar
button as quickly as they can on the second computer. Like the first trial, the subjects
were given ten minutes to play the game and required to save their second trail procedure
and total scores in “save replay” when the researcher asked them to stop the trial.
Second Phase

and the other has a pop-out window program installed. Subjects were asked to “do y
best” to win the game during the trial. In the first trial, subjects should only play
WarcraftIII Frozen Throne on the first laptop without responding to the pop-out wind
task on the second computer. Subjects were asked to save the procedure and their
From the third to the fifth trial, subjects were asked to play
WarcraftIII Frozen Throne on the first laptop and needed to respond to a randomly
appearing pop-out window by pressing a spacebar button as quickly as they can on the
second computer. They were also asked to save the procedure and their total scores by
clicking the “save replay” icon on the computer screen when the researcher asked them to
stop each of the three trials. Subjects then were asked to rate their mental effort
investment scores after each of the three trials. Finally, the subjects were asked to
compare the difficulty level of the three trials by rating the perceived task difficulty
questionnaire.
The entire study will take approximately 90-110 minutes: (a) approximately 10
minutes for introduction and procedure demonstration, (b) approximately 5 minutes for
research consent form and background questionnaire, (c) approximately 5 minutes for
52
time (g)
.
Data Collection
Subjects were first asked to read and signed the Informed Consent Form for the
researcher and each of them was seated in front of two laptop computers. The researcher
then explained the purpose of the research and what procedures subjects would go
through during the experimental process. At the same time, the researcher explained what
the subjects needed to do with the two laptop computers, as well as, demonstrated the
secondary task to the subjects. The researcher then distributed the background
information sheet to the subjects and informed them that one assigned ID number on the
top of the sheet was used for their identity in this research. All of the information they
provided was confidential and only for the research purposes. Once the subjects
understood the researcher’s instruction, they were requested to fill out their background
information and perceived self efficacy questionnaires before the first trial. The subjects
were also told that they could consult the researcher if they did not understand the
questions. The researcher then collected their answer sheets and informed the subjects to

perceived self-efficacy questionnaire, (d) approximately 10 minutes for the first trial (e)
approximately 10 minutes for the second trial (f) approximately 10 minutes break
approximately 10 minutes for the third trial, (e) approximately 3 minutes for invested
mental effort questionnaire (h) approximately 10 minutes for the fourth trial (i)
approximately 3 minutes for invested mental effort questionnaire (j) approximately 10
minutes for the fifth trial (k) approximately 3 minutes for invested mental effort
questionnaire, (l) approximately 5 minutes for perceived task difficulty questionnaire
53

uter. The
005”
Similar data collection procedures were used in the second phase. However, the
subjects were asked to save their trial tests according to the difficulty level of the task.
For example, subject A’s file’s name for the third trial was “005Easy” if subject A’s
assigned number was “005” and the difficulty level of the task for his or her third trial
was “Easy”. Then, the subjects were asked to answer AIME questionnaire. The
researcher then collected their answer sheet and asked them to prepare for the fourth trial.
The subjects repeated the same data collection procedure for their fourth trial.
Before the fifth trial began, the subjects were reminded to do the same thing as what
they did in the fourth trial. The researcher followed the same procedure to collect data
and ended the experiment after the subjects finished the mental effort questionnaire and
perceived task difficulty questionnaire.
The performance scores, scores of mental effort investment, perceived self-efficacy
scores, and standard deviations under different levels of task difficulty were examined;
the relationship among perceived self efficacy, levels of task difficulty, expertise, and
mental effort investment on task performance was also analyzed.

be ready for the two pre-tests. The subjects were instructed about time limits to complete
the two pre-tests and were reminded to response to the secondary task as quickly as they
could during the second pre-test. After they were asked to stop the game, the subjects
were instructed to save their trial procedure and performance scores in the comp
file’s name was their assigned ID plus the number of the test. For example, subject A’s
file’s name for pre-test1 was “005pretest1” if subject A’s assigned number was “
and the subject A’s file name for the second trial (pre-test2) would be “005pretest2”.
54
diffe
so
aining the
relat
ral relationships (Kerlinger, 1986; Pedhazur,
1997
m

Data Analysis
A test of internal consistency reliability (Cronbach’s alpha) was conducted for each
survey administration. The statistical software SPSS 17.0 was used to analyze data.
Descriptive statistics for subjects’ playing experiences, perceived self efficacy, mental
effort, and perceived task difficulty were analyzed first.
An independent-samples t-tests were conducted to determine if there were any
differences between the low level and high level of expertise groups as well as the
rent task difficulty groups. One-way ANOVA analysis was also used to check if
there is a significant difference between perceived self efficacy and mental effort.
Pearson’s correlation was used to check hypotheses 1a, 1b, 2a, 2b, 2c, 3a, and 3b.
Regression analysis was used to test all fourteen hypotheses. Effect size method was al
used for power analysis.
Finally, path analysis was used to test each path in the performance model showed
in Figure 2 and examined as to whether the model is reliable and valid for expl
ionship between independent variables and dependent variables. Path analysis
involves path diagrams, graphically representing the a priori structures, and if
assumptions are met, ordinary least squares estimates of regression coefficients can be
used to estimate the strengths of the structu
). It is possible to test the patterns of intercorrelations between the variables in the
equation. In general, path analysis techniques allow for the testing of bi-directionality
between the variables. However, Pajares et al. (1999) also cautioned that inferences fro

55
findings cannot explain the causal inference of the variables; the
expl

path analytical
anations must be made carefully and modesty.

56

CHAPTER III RESULTS
Introduction

This study was designed to determine how different levels of task difficulty and
expertise affect self-efficacy judgments. Another purpose is to determine how
self-efficacy judgments affect the amount of mental effort invested and the resulting task
performance under different levels of task di this study could be
used to build a performance model to illustrate the relationship among these variables.
Data was collected after the trials. Five trials were used in this study. The first two
trials were used to determine subjects’ levels f expertise and the other three trials were
used to determine the subjects’ task performa ce under different levels of task difficulty.
Sixty-six subjects participated in the study. However, two subjects’ pretest2 scores were
missing because an unexpected technology error occurred. These two subjects’ levels of
expertise therefore were determined by their pretest1 scores. All available data were used.
This chapter presents the results of data Statistical analyses included the
test of internal consistency reliability, descriptive statistics, multiple regression, T-test,
One-way ANOVA, effect size, and Two-way ANOVA.
Internal Consistency Reliability

Perceived self-efficacy and mental effort surveys were tested for reliability in all
five trials. A Cronbach’s alpha coefficient w tained for each subscale and the overall
ale. As seen in Table 2, the alpha values for mental effort, perceived self-efficacy, and

fficulty. Results from
o
n
analysis.
as ob
sc
57

the overall scale were 0.89, 0.88, and 0.87, respectively. The high alpha values ( α > 0.8)
dicated good internal consistency of the survey items.

Tab
in
le 2 Reliability of Scales
N=66 Reliability ( α) Items
Mental Effort 0.89 9
ME-Normal 0.70 3
ME-Insane 0.78 3
ME-Advanced 0.84 3
Perceived Self Efficacy

0.88 20
SE 1_1, 2, 3 0.82 3

SE 2_2_1-16 0.89 17

Overall Scale

0.87 29

s, thirty-five Chinese, two
Vietnamese, four Taiwanese, ten Asian Indians, and one Korean. Fifty-four subjects
ported to have experience in playing other strategy games within the past one year and
twelve subjects reported “No experiences

Descriptive Statistic Analysis of Subjects’ Background and Game Experiences

Subjects included eleven Caucasians, three Hispanic
re
in playing other strategy games” before.

58
Question number 1: Which races do you play with most in Warcraft III?

WarCraft III provides four types of races and each of the races can be considered as
one domain. Players would act as a novice in one race if they never played the game with
this race before. Conversely, players can also be experts in more than one race,
depending on the frequency with which the
p domain). Around 42% of the subjects
f rcraft III and e
s o
p
Q g ne

players have practiced. In common cases,
layers can be the experts only for one race (
requently play with the race “Human” in Wa this result implies that thes
ubjects should have more kn wledge and expertise in playing the game than the other
articipants.

uestion number 2: How lon have you played Warcraft III Frozen Thro version in
general?
The responses of this question indicated that three su d “Never”, two
onth”, ten subjects answered “one to three months” ,
r to six months”, three subjects answered “six months to one
year”

e
bjects answere
subjects answered “less than one m
three subjects answered “fou
, and the remaining fourty-five answered “More than one year.”
Question number 3: How many hours do you play Warcraft III with computer in on
week in average?
Fifty-six subjects played the Frozen Throne version with computer less than seven
ered “eight to fourteen hours”, and four hours in one week in average, six people answ

59
eople answered “fifteen to twenty-one hours”. Most of the subjects mentioned that they
now play the upgrade version of Warcraft III instead of the Frozen Throne.

p
Question number 4: How many hours do you play Warcraft III Frozen Throne version
online with people in one week in average?
Question number 4 asked that “How many hours do you play Warcraft III Frozen
Throne version online with people in one week in average?” Thirty-eight subjects played
the Frozen Throne version online with people less than 7 hours in one week in average,
sixteen people answered “eight to fourteen hours”, seven people answered “fifteen to
twenty-one hours”, two people answered “twenty-two to twenty-eight hours”, one person
answered “twenty-nine to thirty-five hours”, and two people answered “more than
thirty-five hours”. The reason that most of the subjects played the Frozen Throne version
only for few hours a week is that they now play the upgrade version.

Question number 5: How long have you played Warcraft III Frozen Throne version with

the race Human?
When asking “How long have you played Warcraft III Frozen Throne version with
the race Human?”, six subjects answered “Never”, fifteen answered “less than one
month”, eight answered “one to three months”, two answered “four to six months”, four
answered “six months to one year”, and the other thirty-one answered “More than one
year.”

60

Question number 6: How many hours do you play Warcraft III Frozen Throne version
with computer with the race Human in one week in average?

When asking “How many hours do you play Warcraft III (with the race Human)
ours”, one answered “eight to fourteen hours”, and one answered “fifteen to twenty-one
hours”.
Question number 7: How many hours do you play Warcraft III Frozen Throne version
with computer in one week in average?”, sixty-four subjects answered “less than seven
h

online with people with the race Human in one week in average?

Finally, fifty-six subjects played the Frozen Throne version (with the race Human)
online with people less than seven hours in one week in average. Seven subjects pla
“eight to fourteen hours”, one subject played “fifteen to twenty-one hours”, one subject
played “twenty-nine to thirty-five hours”, and one subject played “more
yed
than thirty-five
ours” with the race Human with people online in one week in average.
f Caucasians and Hispanics. At least 81% of the subjects had
expe

However, most of them currently spend less than seven hours or eight to

h
In conclusion, the study contained a much higher percentage of Asians and a much
lower percentage o
riences in playing other computer strategic games and more than 40% of the subjects
often used the race Human when playing the Warcraft III.
Most of the subjects had experiences in playing the Frozen Throne version more than
one year. They also had experience in playing the game with the race Human more than
one year.
61

in certain time limits, question of “Within the following given time-limits, I am
onfident that I can win the game” was given. There were six answers for the subjects to

that the subjects were confident to win the gam
respectively).

fourteen hours in playing the Frozen Throne version either with computer or online with
people in one week in average.
Descriptive Statistic Analysis of Perceived Self Efficacy
To analyze the subjects’ confidence level and self-efficacy beliefs in playing the
game with
c
choose under three task difficulty levels: (1) Less than 10 min. (2) 10-15 min. (3) 16-20
in. (4) 21-30min. (5) 31-35 min. (6) More than 35 min. The results in Table 3 suggested m
e in a longer time limit if the level of task
difficulty was high and were confident to win the game in a shorter time limit if the level
of task difficulty was low (M=3.45, 4.67, 5.59 with level of normal, insane, and advanced,

Table 3 Descriptive Statistics of Perceived Self-Efficacy
N=66

Within the following given
time-limits, I am confident
that I can win the game
when the level of task
difficulty is Easy
Within the following given
time-limits, I am confident
that I can win the game
when the level of task
difficulty is Moderate
Within the following
I am
confident that I can win
of task difficulty is High
given time-limits,
the game when the level
Mean
3.45 4.67. 5.59
Standard
Deviation
0.182 0.189 0.112
62

were given. The scale was between 0 and 100:
0=Cannot win/beat at All, 50=Moderately certain can win/ beat, 100=Certainly can
win/beat. Questions of “I can win the game when my race setting is Human” and “I can
in by recording in each of the blank spaces a number when my race setting is Human”
were asked and the results were also reported in Table 4.
The average self-report score for the first question was between “Moderately certain
can win” and “Certainly can win” (M=68.50). For the second question, the average item
scores were ranged from 52 to 69. The results showed that the subjects were less
confident in the “Lumber Harvested” (M=52.95) and were more confident in the “Unites
Killed” (M=68.41). In general, the subjects had “moderate to high” confidence in
winning the computer.

To understand the subjects’ confidence level in winning the game when the race was
set “Human”, two scale questions
w
63

Table 4 Descriptive Statistics of Perceived Self-Efficacy from scale 0 to 100
N = 66
Mean Standard Deviation
I can win the game when my race
setting is Human
68.50 2.676
I can win by recording in each of the blank spaces a number when my race setting is
Human
units produced
56.97 3.138
units killed
68.41 2.717
building produced
59.47 2.493
units grazed
66.29 2.945
largest army
61.59 3.09
hero killed
61.94 3.269
items obtained
56.97 2.729
experience gained
65.38 2.351
gold mined
54.50 3.398
lumber harvested
52.95 3.242
technology percent
59.39 2.393
gold lost to upkeep
54.47 2.905
64

Questions of “How much effort, or work, e?”,
“How much did you concentrate while playing the game?” (Salomon, 1983) and “How
much mental effort did you invest in order to win the game?” (Paas, 1992) were used to
measure the subjects’ mental effort investment.
The first two questions used the scales of 1= No effort, 2= Not much effort, 3=A lot of
effort, 4= A huge amount of effort and 1= Not at all, 2=Not much, 3=A lot 4=Extremely.
As seen from Table 5 5, when level of task difficulty was “Easy”, the average scores for
Descriptive Statistic Analysis of Mental Effort
did you put into trying to play the gam
Table 5 Descriptive Statistics of Self-Re Mental Effo ported rt (N =66)
Questions of Mental Effort Measurement
Mean Standard Deviation
How much effort, or work, did you put into
trying to play the game? (Easy)
2.89 .747
How much did you concentrate while playing the
game? (Easy)
3.18 .579
How much mental effort did you invest in order
to win the game? (Easy)
5.17 1.131
How much effort, or work, did you put into
trying to play the game? (Moderate)
3.12 .755
How much did you concentrate while playing the
game? (Moderate)
3.32 .683
How much mental effort did you invest in order
to win the game? (Moderate)
5.62 .973
How much effort, or work, did you put into
trying to play the game? (High)
3.39 .653
How much did you concentrate while playing the
game? (High)
3.42 .634
How much mental effort did you invest in order
to win the game? (High)
5.89 .994
65
the first two question effort” (M1=2.89,
M2 scores for the
firs 2, M2 Fi
task difficulty was “High”, the average scores for the firs ons w A lot of
effo
No Effort, 4=Moderate
Effo e mean scores of the three difficulty levels (Easy, Moderate,
High) were 5.17, 5.62, and 5.89, respectively. The results suggested that most subjects
rep era rts in orde in the
gam
s of Task Difficulty

sy, Moderate, High) were
pro the subjects, a perceived task difficulty survey was used.
The

s were between “Not much effort” and “A lot of
=3.18). When level of task difficulty was set “Moderate”, the average
t two questions were “A lot of effort” (M1=3.1 =3.32). nally, When level of
t two questi ere “
rt” (M1=3.39, M2=3.42)
The third question used Paas (1992)’s scale from 1 to 7 (1=
rt, 7=Extreme Effort). Th
orted to invest a lot of efforts or more than the mod te effo r to w
e.

Descriptive Statistic Analysis of Level
To determine if the three task difficulty levels (Ea
gressively more difficult to
results showed that subjects felt the “High level” the most difficult and less easy task
among the three tasks. (M=4.97, M=3.02). However, their perceptions of the difficulty
level for easy and moderate tasks were inconsistent (M=3.92, 4.43; M=3.97, 4.31).

66

Table 6 Descriptive Statistics of Perceived task difficulty among all subjects
N= 66
Mean Standard Deviation
How easy was the “Easy” level task? 3.92 1.964
How difficult was the “Easy” level task? 4.43 1.767
How easy was the “Moderate” level task? 3.97 1.709
How difficult was the “Moderate” level task? 4.31 1.767
How easy was the “High” level task? 3.02 1.493
How difficult was the “High” level task? 4.97 1.561

One-way analysis of variance (ANOVA) and post hoc test were then tested based on
the subjects’ game performance scores. The results demonstrated that there was no
significant difference between easy task performance and moderate task performance
(p<.05). However, there may be a significant difference between “Easy” level and “High”
level because a p-value (p=0.097) was found close to the significance level of p=0.05. As a
result, both “Easy he data of the
“Norm task
d

” and “Moderate” tasks were combined to “Normal” task. T
al” task and “High” task then were used to represent two different levels of
ifficulty.
67

Table 7 One-Way ANOVA Test (N=66)

df F p

Between Groups 2 2.651 .073
Within Groups 195

Total 197

LSD Post Hoc Test for task difficulty (Easy, M te, High) odera
Levels of
task
difficulty
Mean Difference
Stdandard
Error p
Moderate -516.07576 949.11605 .587 Easy
High 1580.93939 949.11605 .097
Easy 516.07576 949.11605 .587 Moderate
High 2097.01515
*
949.11605 .028
Easy -1580.93939 949.11605 .097 High
Moderate -2097.01515
*
949.11605 .028
( * p<.05)

Hypothesis Number 1a

There is a positive relationship between levels of task difficulty and mental effort.
Findings supported the hypothesis and indicated that there is a significant and
positive relationship between levels of task difficulty and mental effort.

This question was addressed by the mental effort self-reported measurement of
aas’s (1992) and Salomon’s (1983) mental effort scales. Paas’s (1992) rating scale ranged
om 1= No effort, 4= Moderate effort, to 7= Extreme effort.

P
fr
68
The rating = No effort to 4= A huge amount of
effort for the first question, and for the second question is from 1= Not at all to
4=Extreme wo task difficu lev rmal and High) were also used to
test the relationship. Pearson’s correlation sis result was used to explain the
relationshi
index of m t investment was 0.120 and there was a significant relationship
between le k ifficulty an .01).
Reg n ana as used to tween thes
variables ficulty as om the regression
analyses indicated that levels of task difficulty significantly predicted mental effort
(F=8.686,
explained by levels of task difficulty (R square=0.014). The regression equation obtained
in this analysis was: Y’ (mental effort) = .354*(levels of task difficulty) + 3.53. The
ANOVA result also showed that the model was able to explain any variable in the
dependent variable (p=0.003). The Partial Eta Squared value suggested that the effect size
Table 8

of Salomon’s (1983) scale ranges from 1
ly Concentrate. T lty els (No
analy
p between levels of task difficulty and mental effort investment. The correlation
ental effor
vels of tas d d mental effort (p<
ressio lysis then w test the linear model be e two
. Using levels of task dif the predictor, results fr
p=.003). However, only 1.4% of the variance in mental effort could be
for levels of task difficulty was small (partial eta squared=0.017).
Regression Results (N=66, Predictor: Levels of Task Difficulty,
Dependent Variable: Mental Effort)
Predictor B SE B β
Level of Task
Dif
0.354 0.120 0.120
ficulty
Notes: R =0.014 (p<0.01)
2

69
t
of task
Hypothesis Number 1b
T
gative
mance.

of

le
w
ta
v
an erformances
(F=5.0 . However, only 2.5% of the variance in task performances could be
explained by levels of task difficulty (R square=0.025). The regression equation obtained in
thi ormances) = -1838.977*(levels of task difficulty)+22856.65.
The ANOVA result also showed that the model was able to explain

In summary, when levels of task difficulty increased, the investment of metal effor
also increased. There was a significant and positive relationship between levels
difficulty and mental effort from self-reported measurement.
here is a negative relationship between levels of task difficulty and task performance.
Findings supported the hypothesis and indicated that there is a significant and ne
relationship between levels of task difficulty and task perfor

This question was addressed by the game performance scores under the two levels
task difficulty (Normal and High).
Pearson’s correlation analysis result was used to explain the relationship between
vels of task difficulty and task performance. The correlation index of task performance
as -0.158 and there was a significant relationship between levels of task difficulty and
sk performances (p=0.013).
Regression analysis then was used to test the linear model between these two
ariables. Using levels of task difficulty as the predictor, results from the regression
alyses indicated that levels of task difficulty significantly predicted task p
24, p=.026)
s analysis was: Y’ (task perf
70

any d
Table 9 Regression Results (N=66, Predictor: Levels of Task Difficulty, Dependent
Variable: Task Performance)
variable in the dependent variable (p=0.026). The Partial Eta Squared value suggeste
that the effect size for levels of task difficulty was small (partial eta squared=0.018).

Predictor B SE B β
Level of Task -1838.977 820.48 -0.158
Difficulty
Notes: R
2
=0.025 (p<0.05)

Overall, the scores of game performances decreased when the levels of task difficulty
increased. There was a significant and negative relationship between levels of task
difficulty and task performance.

ship between mental effort and perceived self-efficacy from
self-reported measure results.
Findings partially supported the hypothesis and indicated that there is no
significant but negative relationship between perceived self-efficacy and self-reported
mental effort.

. Perceived self-efficacy is people’s beliefs in their capabilities in performing the
tasks. As a result, subjects’ scores in the self-efficacy question of “I can win the game

Hypothesis Number 2a
There is a negative relation

71

wh
se
Pearson’s correlation analysis results were used to explain the relationship between
perce
-0.005. The result suggested that m al effort decreased when perceived self efficacy
incre there was nt differenc en self-report al effort
and p 47).
Regression analysis then was used to test the linear model between these two variables.
Using perceived self efficacy as the predictor, results from the regression analysis indicated
that perceived self efficacy could not significantly predict mental effort (F=0.017, p=
0.895). The ANOVA result also showed that the regression model was not able to explain
any variable in the dependent variable (p=0.895).

able 10 Regression Results (N=66, Predictor: Perceived self-efficacy, Dependent
V
en the race setting is Human” were used for the analysis. In addition, data from
lf-reported mental effort were also used to test the hypothesis.

ived self efficacy and mental effort. The correlation index for the relationship was
ent
ased. However, no significa e betwe ed ment
erceived self efficacy (p=0.4
T
ariable: Mental Effort)
Predictor β B SE B
Perc
self-efficacy
eived 0 -0.005 -0.005
Notes: adjusted R
2
=-0.002 (p=0.895)

The normal probability plot (also called a P-P Plot) is an alternative method plotting
observed cumulative probabilities of occurrence of the standardized residuals on the Y axis
and of expected normal probabilities of occurrence on the X axis. If and the ideal
distribution line in regression analysis is not the 45 degree straight line (the black line in
Figure 24), it is probably that the data have broken the assumption of linearity.

72

e pattern
analysis. As seen from
Figure 24 Normal P-P Plot of the regression model of self-reported mental effort
As a result, to check whether the data of self-reported mental effort matched th
of the linearity assumption, a normal probability plot was used for
Figure 24, the data demonstrated the pattern of linearity imperfectly. It then may be
concluded that the assumption of the regression model was not violated.

Hypothesis Number 2b
rt.

There is an inverted U relationship between perceived self-efficacy and mental effort
from dual-task measure results.

Findings did not support either the inverted U relationship hypothesis or the linear
relationship between perceived self-efficacy and dual-task measured mental effo
However, the scatter plot showed the relationship roughly demonstrated the shape of
inverted U. The result of this hypothesis remains open for further research.
73

easured mental
effort and perceived self efficacy (p=0.409).
A linear regression analysis was used to analyze if perceived self-efficacy can
predict mental effort.
Using perceived self efficacy as the predictor, results from the regression analysis
indicated that perceived self efficacy could not significantly predict mental effort (F=0.053,
p= 0.818). Meanwhile, the value of R square also showed that no variance of the mental
effort could be explained by the model. The ANOVA result also showed that the regression
model was not able to explain any variable in the dependent variable (p=0.818).
Table 11 Regression Results (N=66, Predictor: Perceived self-efficacy, Dependent
Variable: Mental Effort)
Pearson’s correlation analysis results were used to explain the relationship between
perceived self efficacy and mental effort. The correlation index for the relationship was
0.005. The result suggested that mental effort increased when perceived self efficacy
increased. However, there was no significant difference between dual task m
Predictor B SE B β
Perceived
self-efficacy
0.005 0.02 0.005
Notes: R
2
=0.005 (p=0.818)

The Normal P-P Plot was used again to test whether there was a violation of the data
from dual-task measured me the data of dual-task
the
ntal effort. Figure 25 showed that
measured mental effort (the blue line) was a non-linear relationship between the outcome
and the predictor. The path line was roughly like an inverted U shape and illustrated
trend of the data.

74

Figure 25 Normal P-P Plot of the regression model of dual task measured mental effort

The scatter plot in Figure 26 also demonstrated that there was a linear relationship
(followed the trend of the equation line) between perceived self efficacy and mental effort
when the perceived self efficacy scores were between 20 and 85. However, the curve
changed and showed a minor drop after reaching the efficacy score of 85.

Figure 26 Scatter Plot for perceived self efficacy and mental effort

75
assumption of the regression model but also the scatter plot showed that the relationship
between perceived self-efficacy and dual-task measured mental effort closed to the shape
of inverted U.

Hypothesis Number 2c

There is a positive relationship between perceived self-efficacy and task performance.
Findings supported the hypothesis and indicated that there is a significant and positive
relationship between perceived self-efficacy and task performance.

se nce scores
fr levels of tasks (Normal, High) were also used to measure the subjects’ task
p
Pearson’s correlation analysis results were used to explain the relationship between
perceived self-efficacy and mental effort. The correlation index for the relationship was
0.197. The result suggested that task performance increased when perceived self-efficacy
increased. There was a significant difference between task performance and perceived
self-efficacy (p=0.003).

In conclusion, not only the curve linear shape revealed in the P-P Plot violated the
Subjects’ scores in the self-efficacy question of “I can win the game when the race
tting is Human” were used again for the analysis. In addition, the performa
om the two
erformances.

76
05). The ANOVA result also showed that the regression model was able
explain any variable in the dependent variable (p=0.005). 3.9 % of the variance of task
performance could be explained by the regression equation (R square= 0.039). The
regression equation obtained in this analysis was: Y’ (task performance) =
the
Table 12 Regression Results (N=66, Predictor: Perceived self-efficacy,
Dependent Variable: Task Performance)

Regression analysis then was used to test the linear model between these two
variables. Using perceived self-efficacy as the predictor, results from the regression
analyses indicated that perceived self-efficacy significantly predicted task performance
(F=7.93, p= 0.0
to
50.131*(perceived self-efficacy)+16970.688. The Partial Eta Squared value suggested that
effect size for perceived self-efficacy was large (partial eta squared=0.212).

Predictor B SE B β
Perceived 50.131 17.802 0.197
self-efficacy
Notes: R
2
=0.039 (p=0.005)

In summary, perceived self-efficacy was one of the predictors to the task performance.
When perceived self-efficacy increased, task performance also increased. There was a
significant and positive relationship between perceived self-efficacy and task performance.
Hypothesis number 3a

s a negative relationship between expertise and mental effort.

There i
77

Findings supported the hypothesis for novice and expert subjects. The results
indicated that there is a significant and negative relationship between expertise and
mental effort.
Expertise Measurement
The measurement of expertise was addressed by the performance scores of pretest1
and pretest2. A regression analysis was used to find the pretest2 predicted score. If
subjects’ pretest2 scores were higher than the predicted score, they were co
level of expertise. Conversely, they were classified to low level of expertise.
Using pretest1 as predictor, results from the regression analyses indicated tha
nsidered high
t
retest1 significantly predicted pretest2 performances (F=82.626, p<.01). The ANOVA
Ta was able to explain any variable in the
dependent variable (p<.01). 57.1% of the variance of the pretest2 performance could be
explained by the regression equation (R square= 0.571). The regression equation obtained
in this analysis was: Y’ (pretest2 scores) = 0.786*(pretest1 scores)+6775.706.
Variable: Pretest2 score)
p
ble also showed that the regression model

Table 13 Regression Results (N=66, Predictor: Pretest1 score, Dependent
Predictor B SE B β
Pretest1 score 0.786 0.087 0.756
Notes: R
2
=0.571(p<0.01)

expertise. If subjects’ pretest2 scores were higher than the average predicted value
The average predicted value was used as the index to determine subjects’ levels of
78

(M=
nd 32
igh level of expertise.
t and subjects’ levels of expertise were used to test the
as a
and
w
6938.42 were considered novices. The scores between 30196.58 and 36968.97 were
con
4 Percentile Analysis of pretest2 predicted scores (N=64)
23424.19), they were considered high level of expertise. Conversely, they were
classified to low level of expertise. Finally, 34 subjects were low level of expertise a
subjects were h
Self-reported mental effor
hypothesis. Pearson’s correlation analysis was used. The correlation index for the
relationship between these two variables was .047. The result suggested that there w
positive relationship between expertise and mental effort. There is no significant
difference between subjects’ level of expertise and mental effort (p=0.124).
To better understand whether the correlation results were different from novice
expert groups, subjects’ pretest2 scores were used to find novice and expert subjects. The
percentile analysis was used for the selection.
As seen from Table 14, the minimum predicted value was 10452.65 and the
maximum predicted value was 36968.9; the 25% percentiles score was 16938.42 and the
75% percentiles score was 30196.58. It was decided that the pretest2 scores belo
1
sidered experts.
Table 1
Percentile Pretes dicted Score t2 Pre
0 10452.65
25 16938.42
50 23424.19
75 30196.58
100 36968.97

79

d 8

Finally, the result showed that 13 subjects were included in the novice group an
subjects were included in the expert group.
Table 15 Novice and Expert Groups

N Subject ID Number
Novice 13 1, 4, 6, 10, 11, 24, 38, 39, 40, 55, 56,62, 66
Expert 8 29, 34, 47, 51, 52, 57, 64, 65

Pearson’s correlation analysis was used to test if the hypothesis was true betw
novice and expert subjects. The correlation index for the relationship between these tw
een
o
varia
ear model between these two
v alysis
indicated that subjects’ levels of expert
p= lain
any variable in the dependent variable (
could be explained by the regression equation (R square= 0.022). The regression equation
obtained in this analysis wa
(expertise)+4.552.
Table 16 Regression Results (N=21, Predictor: Expertise, Dependent Variable:
Mental Effort)
bles was -.149. The result suggested that there was a negative and significant
relationship between expertise and mental effort (p=0.021).
Regression analysis then was used to test the lin
ariables. Using subjects’ levels of expertise as predictor, results from regression an
ise significantly predicted mental effort (F=4.234,
0.041). The ANOVA result also showed that the regression model was able to exp
p=0.041). 2.2% of the variance of the mental effort
s: Y’ (mental effort) = -0.47*
Predictor B SE β B
Expertise -0.47 0.229 -0.149
Notes: R
2
=0.022(p<0.05)

80

Th
the mental effort
dec
There was a significant and ne ive relationship bet ental effort.
The ta Squared value gested e was small (partial
ta squared=0.022).
d task performance.

This question was addressed by the two levels of expertise (Low, High) and
performance scores under the two levels of task difficulty (Normal and High).
Pearson’s correlation analysis was used again. The correlation index for the
relationship between these two variables was 0.609. The result suggested that there was a
positive and significant relationship between expertise and task performance (p<.01).
Regression analysis then was used to test the linear model between these two
varia ects’ level of expertise as predictor, results from the regression
analyses indicated that expertise significantly predicted task performance (F=115.812,
p<0.01). The ANOVA result also showed that the regression model was able to explain

e results revealed that the subjects’ level of expertise was one of the predictors to
the mental effort. When subjects’ level of expertise increased,
reased.
gat ween expertise and m
Partial E sug that the effect size for expertis
e

Hypothesis Number 3b

There is a positive relationship between expertise an
Findings supported the hypothesis and indicated that there is a significant and
positive relationship between expertise and task performance.

bles. Using subj
81
he
ation obtained in this analysis was: Y’ (task performance) =
t
on Results (N=66, Predictor: Expertise, Dependent Variable:
Task Performance)

any variable in the dependent variable (p<0.01). 37.1% of the variance of the task
performance could be explained by the regression equation (R square= 0.371). T
regression equ
6687.222*(expertise)+10475.17. The Partial Eta Squared value suggested that the effec
size for expertise was large (partial eta squared=0.406).
Table 17 Regressi
Predictor B SE B β
Expertise 6687.222 621.396 0.609
Notes: R
2
=0.371(p<0.01)

p erformance also increased.
There was a significant and positive relationship between expertise and task performance.

Hypothesis Number 4a
t
nly levels of task difficulty contributes the prediction to mental effort.
In summary, subjects’ level of expertise was one of the predictors to the task
erformance. When their levels of expertise increased, the task p

H4a: There is a significant difference between levels of task difficulty, expertise, and
mental effort. There is a positive relationship between levels of task difficulty and mental
effort while there is a negative relationship between expertise and mental effort. Both
levels of task difficulty and expertise contribute the prediction to mental effort.

Findings partially supported hypothesis and indicated that there is a significan
difference between levels of task difficulty, expertise, and mental effort. There is a
positive relationship between levels of task difficulty and mental effort while there is a
negative relationship between expertise and mental effort. Regression results showed that
o
82

nt

The independent-samples t-test was used first to check if there was a significa
difference between mental effort, two levels of expertise groups (Low, High), and two
levels of task difficulty groups (Normal, High).
Table 18 presented the descriptive statistics of mental effort for the two levels of
expertise groups and two levels of task difficulty groups.
Table 18 Group Statistics of Mental effort
N=66
Mean Std. Deviation
low 3.94 1.43 Expertise
high 4.07 1.34
normal 3.88 1.36 Levels of task
difficulty
high 4.24 1.41

The mental effort mean for the low level of expertise and high level of expertise
were 3.94 and 4.07, respectively; while the corresponding figures for the normal task and
high task were 3.89 and 4.24, respectively. The Levene’s test for equality variance
suggested that the variances of the two levels of expertise groups and two levels of task
difficulty groups were not signifi 8, p = 0.178; F = 0.278, p =
of
expertise (t= -1.155, p = .248). However, there was a significant difference in mental
effort between the groups of normal task and high task (t=2.947, p=0.003).
The independent-samples t-test was also used to check the results of novice and
expert subjects.

cantly different (F = 1.81
0.599). The independent-samples t-test results indicated that there was no significant
difference in mental effort between the groups of low level of expertise and high level
83
T ert

able 19 presented the descriptive statistics of mental effort for the novice and exp
groups, as well as the normal and high levels of task difficulty groups.
Table 19 Group Statistics of Mental effort
N=21
Mean Std. Deviation
novice 4.08 1.41 Expertise
expert 3.61 1.45
normal 3.89 1.36 Levels of task
difficulty
high 4.24 1.41

resp the c ding figure he norma and high task were 3.89
and 4.24, respectively. T vene’s test for ariance suggested that the
41).
ults
as
ately 1.7%
as a significant difference between levels of task difficulty, expertise,
The mental effort mean for novice and expert groups were 4.08 and 3.61,
ectively; while orrespon s for t l task
he Le equality v
variances of novice and expert groups were not significantly different (F = 0.153, p =
0.696). The independent-samples t-test results indicated that there was a significant
difference in mental effort between the novice and expert groups (t= -2.058, p = 0.0
The regression analysis was used to check if there was a significant difference
between the levels of task difficulty, expertise, and mental effort.
Using levels of expertise and levels of task difficulty as predictors, regression res
indicated that only levels of task difficulty was the significant predictor of mental effort
(t=2.948, p = .003). However, the ANOVA result showed that the regression model w
able to explain any variable in the dependent variable (p=0.007). Approxim
of the variance in mental effort could be explained by expertise and levels of task
difficulty. There w

84
an
T or: Expertise, Levels of task difficulty,
Dependent Variable: Mental Effort)

d mental effort. The Partial Eta Squared value suggested that the effect size for levels of
task difficulty and expertise was small (partial eta squared=0.001).
able 20 Regression Results ((N=66, Predict
Predictor B SE β B
Expertise 0.132 0.113 0.047
L
di
.354 0.12 0.120 evels of task
fficulty
0
Notes: R
2
=0.017(p<0.01)

Hypothesis Number 4b

There is a significant difference between levels of task difficulty, expertise, and tas
performance. There is a negative relationship between levels of task difficulty and task
performance while there is positive relationship between expertise and task performan
Both expertise and levels of task difficulty contribute the prediction to task performa

Findings supported the hypothesis and indicated that there is a significant difference
k

ce.
nce.
e
hile there is a
xpertise groups and two levels of task difficulty groups.
between levels of task difficulty, expertise, and task performance. There is a negativ
relationship between levels of task difficulty and task performance w
positive relationship between expertise and task performance. Both levels of task
difficulty and expertise contribute the prediction to task performance.

The independent-samples t-test was used to check if there was a significant
difference between task performance, two levels of expertise groups (Low, High), and
two levels of task difficulty groups (Normal, High).
Table 21 presented the descriptive statistics of task performance for the two levels of
e
85

Table 21 Group Statistics of Task Performance

Mean Std. Deviation
low 17162.39 4280.04 Expertise
high 23849.61 4463.44
normal 21017.67 5565.08 Levels of task
difficulty
high 19178.70 5186.50

The task performance mean for low level of expertise and high level of expertise
were 17162.39 and 23849.61, respectively; while the corresponding figures for the
ormal task and high task were 21017.67 and 19178.70, respectively. The Levene’s test
for equality variance suggested that the vari rtise groups and two
vels of tasks were not significantly different (F = 0.086, p = 0.77; F = 0.004, p = 0.949).

= .00
on results
indicated that levels of task difficulty and expertise were the significant predictors of task
performance (t=-2.841, p = .005, t=10.954, p=0.000, respectively). The ANOVA result
also showed that the regression model was able to explain any variable in the dependent
variable (p<0.01). There was a significant difference between expertise, levels of task
difficulty, and task performance. Approximately 39.6% of the variance in task
n
ances of the two expe
le
The independent-samples t-test results indicated that there was a significant difference in
task performance between low level of expertise and high level of expertise (t=10.762, p
0). There was also a significant difference in task performance between the groups
of normal task and high task (t=-2.241, p=0.026).
The regression analysis was used to check if the levels of task difficulty and
expertise have significant difference with task performance.
Using expertise and levels of task difficulty as predictors, regressi
86
perform
square=0.396). The eq n obtained was erfor -1838.977 * (levels
of task difficulty) + 668 * (expertise) + . The ta Squared value
sug the effect or levels of tas and e as small (partial eta
squared=0.002).
Table 22

ance could be explained by expertise and levels of task difficulty (R
uatio : Y’ (task p mance) =
7.222 12927.139 Partial E
gested that size f k difficulty xpertise w
Regression Results ((N=66, Predictor: Expertise, Levels of task difficulty,
Dependent Variable: Task Performance)
Predictor B SE B β
Expertise 6687.222 610.478 0.609
Levels of task
difficulty
-1838.977 647.212 -0.158
Notes: R =0.396 (p<0.01)
2

Hypothesis Number 5a

and mental effort. There is a negative relationship
There is a significant difference between levels of task difficulty, perceived self-efficacy,
between levels of task difficulty and
perceived self-efficacy. There is a positive relationship between levels of task difficulty
and m cacy
and mental effort. Both levels of task difficulty and perceived self-efficacy contribute the
ficant
ental effort while there is a negative relationship between perceived self-effi
prediction to mental effort.

Findings partially supported the hypothesis and indicated that there is a signi
difference between levels of task difficulty, perceived self-efficacy, and mental effort.
There is a negative relationship between levels of task difficulty and perceived
self-efficacy. However, regression results showed that only levels of task difficulty
contributes the prediction to mental effort.

87

iven
ented the descriptive statistics of perceived self-efficacy for two levels
of
Subjects’ scores in the self-efficacy question of “Within the following g
time-limits, I am confident that I can win the game when the task level is Easy, Moderate,
and High” were used to check if there was a significant difference between perceived
self-efficacy and two levels of task difficulty groups.
Table 23 pres
task difficulty groups.
Tab p Statistics of ceived Self-Effic le 23 Grou Per acy

Mean Std. Deviation
normal 4.06 1.62 Levels of task
difficulty
high 5.59 0.91

The perceived self-efficacy means for the normal task and high task were 4.06 and
5.59. The Levene’s test for equality variance suggested that the variances of the normal
sk and high task were significantly different (F = 48.863, p = 0.000). As a result, data
p=0.007).
ental
ental effort and perceived self-efficacy. The

ta
referred to Equal variances not assumed were used to explain the results. The
independent-samples t-test results indicated that there was a significant difference in
perceived self-efficacy between the groups of normal task and high task (t=-2.735,
Subjects’ scores in the self-efficacy question of “I can win the game when the race
setting is Human”, two levels of task difficulty (Normal, High), and self-reported m
effort were used for mental effort analysis. The one-way ANOVA was used to check if
there was a significant difference between m
88

independent-samples t-test was used to check if there was a significant difference
between mental effort and two levels of task difficulty groups.
Table 24 presented the descriptive statistics of mental effort for perceived self
efficacy and levels of task difficulty.
Table 24 Group Statistics of Mental effort

Mean Std. Deviation
Perceived self
efficacy

4.00 1.39
normal 3.89 1.36 Levels of task
difficulty
high 4.24 1.41

The mental effort mean and standard deviation for the perceived self-efficacy were
4.00 and 1.39. The mental effort means for the normal task and high task were 3.89 and
4.24. The one-way ANOVA results showed that there was no significant difference
between mental effort and perceived self-efficacy (F=0.844. p=0.614). The
independent-samples t-test results indicated that there was a significant difference in
mental effort between the groups of normal task and high task (t=2.947, p=0.003).
The regression analysis was used to check if levels of task difficulty and perceived
self-
ble to explain any variable in the dependent variable (p=0.013). There was a significant
difference between perceived self-efficacy, levels of task difficulty, and mental effort.
efficacy had the significant difference with mental effort.
Using perceived self-efficacy and levels of task difficulty as predictors, regression results
indicated that only levels of task difficulty was the significant predictor of mental effort
(t=2.945, p = .003). However, the ANOVA result showed that the regression model was
a
89
ed
artial Eta Squared value
sugg was

Table 25 Regression esults ((N=66, Pre Perceived self-efficacy, Levels of
task difficulty, Dependent Variable: Mental Effort)

Approximately 1.4% of the variance in mental effort could be explained by perceiv
self-efficacy and levels of task difficulty (R square=0.014). The P
ested that the effect size for levels of task difficulty and perceived self-efficacy
small (partial eta squared=0.006).
R dictor:
Predictor B SE B β
P
self-efficacy
0 0.003 -0.005 erceived
Levels of task
difficulty
0.354 0.12 0.12
Notes: R
2
=0.014 (p<0.05)

y,
and task performance. There is a negative relationship between levels of task difficulty
self-efficacy and task performance. Both levels of task difficulty and perceived
self-e

is a significant

Hypothesis Number 5b
There is a significant difference between levels of task difficulty, perceived self-efficac
and task performance while there is a positive relationship between perceived
fficacy contribute the prediction to task performance.
Findings supported the hypothesis and indicated that there
difference between levels of task difficulty, perceived self-efficacy, and task performance.
There is a positive relationship between perceived self-efficacy and task performance as
well as levels of task difficulty and task performance. Both perceived self-efficacy and
levels of task difficulty contribute the prediction to task performance.

90

e

rformance and perceived self-efficacy. The
dependent-samples t-test was used to check if there was a significant difference
be
Table 26 presented the descriptive statistics of task performance for perceived
se nd two levels of task difficulty groups.

Subjects’ scores in the self-efficacy question of “I can win the game when the rac
setting is Human”, two levels of task difficulty (Normal, High), and task performance
scores were used for the analysis. The one-way ANOVA was used to check if there was a
significant difference between task pe
in
tween task performance and two levels of task difficulty groups.
lf-efficacy a
Table 26 Group Statistics o Performance f Task

Mean Std. Deviation
Perceived self
efficacy

20404.68 5497.75
normal 21017.67 5565.08 Levels of task
difficulty
high 19178.70 5186.50

The task performance mean and standard deviation for the perceived self- efficacy
was 20404.68 and 5497.75. The task performance mean for the normal task and high ta
were 21017.67 and 19178.70, respectively. The one-way ANOVA results show
sk
ed that
there fficacy

iculty and
perceived self-efficacy had the significant difference with task performance.
was a significant difference between task performance and perceived self-e
(F=2.441. p=0.004). The independent-samples t-test results also indicated that there was
a significant difference in task performance between the normal task and high task groups
(t=-2.241, p=0.026).
The regression analysis was used to check if the levels of task diff
91
to
nd task performance.
Appr ived
064). The equation obtained was:
Y’ ( acy) + -1838.977 (levels of task
difficulty) +19422.658. The Partial Eta Squared value
leve difficulty and perceived self- as small (partial eta squared=0.036).
T ion R N=66, P erceived s y, Levels of
ta y, Depen ariable: Tas nce)

Using perceived self-efficacy and levels of task difficulty as predictors, regression
results indicated that perceived self-efficacy and levels of task difficulty were the
significant predictors of task performance (t=2.846, p = 0.05, t=-2.282 p=0.024,
respectively). The ANOVA result also showed that the regression model was able
explain any variable in the dependent variable (p=0.002). There was a significant
difference between levels of task difficulty, perceived self efficacy, a
oximately 6.4% of the variance in task performance could be explained by perce
self-efficacy and levels of task difficulty (R square=0.
task performance) = 50.131 (perceived self-effic
suggested that the effect size for
ls of task efficacy w
able 27 Regress
sk difficult
esults (( redictor: P elf-efficac
dent V k Performa
Predictor B SE B β
Pe
self-efficacy
rceived 50.131 17.614 0.197
Levels of task -1838.977 806.01 -0.158
difficulty
Notes: R
2
=0.064 (p<0.01)

Hypothesis Number 6a

There is no significant difference between expertise, perceived self-efficacy, and mental
effort from self-reported measure results.

92
e
ffort

Findings supported the hypothesis and indicated that there is no significant
difference between expertise, perceived self-efficacy and mental effort. There is a
positive relationship between expertise and perceived self-efficacy.
Subjects’ scores in the self-efficacy question of “I can win the game when the rac
setting is Human”, two levels of expertise (Low, High), and self-reported mental e
were used for the analysis.
The independent-samples t-test was used to check if there was a significant
difference between the perceived self-efficacy and two levels of expertise groups(Low,
High).
Table 28 presented the descriptive statistics of mental effort for perceived self-
efficacy and expertise.
Table 28 Group Statistics of Perceived Self- Efficacy

Mean Std. Deviation
low 60.00 21.78 Expertise
high 77.53 17.96

ceived self-efficacy means for the low level of expertise and high level of
xpertise were 60.00 and 77.53. The Levene’s test for equality variance suggested that
the variances of the two expertise groups were not significantly different (F = 0.379, p =
.54). The independent-samples t-test results also indicated that there was a significant
en low level of expertise and high level of
t= 3.556, p = .001).

The per
e
0
difference in perceived self-efficacy betwe
expertise (
93

T and able 29 also presented the descriptive statistics of mental effort for expertise
perceived self-efficacy.
Table 29 Group Statistics of Mental effort

Mean Std. Deviation
Perceived self
efficacy

4.00 1.39
low 3.94 1.43 Levels of expertise
high 4.07 1.34

.39. The mental effort mean for the low level of expertise and high level of
t
etween mental effort and perceived self-efficacy
(F= s also indicated that there was
no significant difference in mental effort between the low level of expertise and high
level of expertise (t= -1. = .248). How ere wa ificant difference in

23, p = 0.187, t=-0.658, p=0.511, respectively). The
VA result also showed that the regression model was not able to explain any

The mental effort mean and standard deviation for the perceived self-efficacy were
4.00 and 1
expertise were 3.94 and 4.07, respectively. The one-way ANOVA results showed tha
there was no significant difference b
0.844. p=0.614). The independent-samples t-test result
155, p ever, th s a sign
mental effort between novice and expert subjects (t= -2.058, p = 0.041).
The regression analysis was used to check if expertise and perceived self- efficacy
had the significant difference with mental effort.
Using perceived self-efficacy and expertise as predictors, regression results
indicated that expertise and perceived self-efficacy were not the significant predictors of
self-reported mental effort (t=1.3
ANO
94
va
ived self-efficacy, and self-reported mental effort.
T ceived self-efficacy, Expertise,
Dependent Variable:

riable in the dependent variable (p=0.414). There was no significant difference
between expertise, perce
able 30 Regression Results (Predictor: Per
Mental Effort)
Predictor B SE β B
P
self-efficacy
(
-0.002 0.003 -0.03 erceived
N=66)
Expertise(N=66) 0.165 0.125 0.059
Expertise (N=21) -0.337 0.24 -0.107

Hypothesis Number 6b
There is a significant relationship between expertise, perceived self-efficacy, and men
effort from dual-task measure results. There is a positive relationship between expertise
and perceived self-efficacy while there is an inverted U relationship between self-efficacy
and dual-task measured mental effort. Both expertise and perceived self-efficacy
contribute the prediction to dual-task measured mental effort.

Findings did not support the hypothesis and indicated that there is no significant
difference between expertise, perceived self-efficacy, and mental effort. Th

tal

ere is also no
of “I can win the game when the race
roups.
inverted U relationship between perceived self-efficacy and mental effort.
Subjects’ scores in the self-efficacy question
setting is Human”, two levels of expertise (Low, High), and results from dual-task
measured mental effort were used for the analysis.
The independent-samples t-test was used to check if there was a significant
difference between the dual-task measured mental effort and two levels of expertise
g
95

Table 31 presented the descriptive statistics of mental effort for two levels of
expertise.
Table 31 Group Statistics of Mental Effort

Mean Std. Deviation
low 3.77 22.50 Expertise
high 2.96 14.52

rt means level of expertise and high level of expertise were
3.77 and 2.96. The Levene’s test for equality variance suggested that the variances of the
o levels of expertise were not significantly different (F = 2.745, p = 0.098).
s used to check if expertise and perceived self- efficacy
had the significant difference with dual-task measured mental effort.
tors of
ual-task measured mental effort (t=-1.128, p = 0.259, t=0.669, p=0.504, respectively).
The ANOVA result also showed that the regression model was not able to explain any
variable in the dependent variable (p=0.515). There was no significant difference
between expertise, perceived self-efficacy, and mental effort from dual-task measured
mental effort.
Table 32 Regression Results (N=66, Predictor: Perceived self-efficacy,
The mental effo for low
tw
The regression analysis wa
Using perceived self-efficacy and expertise as predictors, regression results
indicated that expertise and perceived self-efficacy were not the significant predic
d
Expertise, Dependent Variable: Mental Effort)
Predictor B SE B β
Perceived 0.015 0.022 0.017
self-efficacy
Expertise -1.063 0.942 -0.028
96

significant difference between expertise, perceived self-efficacy, and task
performance. There is a positive relationship between expertise and perceived
self tween expertise and task performance and
between erceived self-e ficacy and task per ce. nd perceived
self-effic te e prediction to tas rma
Findings partially supported the hypoth d indica t there is a significant
ifference between expertise, perceived self-efficacy, and task performance. There is a
positive relationship between expertise and task performance as well as there is a positive
relationship between perceived self-efficacy and task performance. Regression results
showed that only expertise contributes to the prediction of task performance.

game when the
e scores
ived
Hypothesis Number 6c
There is a
-efficacy. There is a positive relationship be
p
acy contribu
f
th
forman
k perfo
Both expertise a
nce.

esis an ted tha
d
Subjects’ scores in the self-efficacy question of “I can win the
race setting is Human”, two levels of expertise (Low, High), and task performanc
were used for the analysis.
The independent-samples t-test was used to check if there was a significant
difference between task performance and two levels of expertise.
Table 33 presented the descriptive statistics of task performance for perce
self-efficacy and expertise.
Table 33 Group Statistics of Task Performance

Mean Std. Deviation
Perceived self
efficacy

20404.68 5497.75
low 17162.39 4280.04 Expertise
high 23849.61 4463.44

97

The task performance mea for the perceived self-efficacy
and
n task performance and
perc
f

Using perceived self efficacy and expert e as predictors, regression results indicated
that expertise was the significant predictor of task performance (t=10.228, p<0.01). The
ANOVA result also showed that the regression model was able to explain any variable in
the dependent variable (p<0.01). Approximately 37.4% of the variance in task
performance could be explained by perceived self-efficacy and expertise (R
square=0.374). The Partial Eta Squared value suggested that the effect size for expertise
and perceived self-efficacy was large (partial eta squared=0.14).
Table 34 Regression Results (N=66, Predictor: Perceived self-efficacy, Expertise
n and standard deviation
was 20404.68 and 5497.75. The task performance mean for the low level of expertise
high level of expertise were 17162.39 and 23849.61, respectively. The one-way ANOVA
results showed that there was a significant difference betwee
eived self-efficacy (F=2.441, p=0.004). The independent-samples t-test results also
indicated that there was a significant difference in task performance between low level o
expertise and high level of expertise (t=10.762, p=0.000).
The regression analysis was used to check if expertise and perceived self- efficacy
had the significant difference with task performance.
is
Dependent Variable: Task Performance)
Predictor B SE B β
Perceived
self-efficacy
-15.332 15.757 -0.06
Expertise 6955.838 680.106 0.634
Notes: R
2
=0.374 (p<0.01)

98

Figu
Difficulty, Perceived Self-Efficacy, Mental Effort, and Task Performance

ce
t from
fort (F=4.234, p=0.041)
for 03),
levels of task difficulty to performa (F=5.024, p=0.026), and perceived self-efficacy
to p ( =7.93, p=0.00 is also an in fect from expertise,
perc efficacy to task p (r=0.689) and from levels of task difficulty,
perceived self-efficacy to task performance (r=-0.196). To find out if there is an indirect
ship between expertise, levels of task difficulty, perceived self-efficacy and task

re 27 Performance Model: The Relationship between Expertise, Levels of Task

Levels of Task Difficulty

Perceived Self Efficacy

β =-0.005
β =0.197
Expertise Mental Effort

Performan

Path Analysis

Figure 27 demonstrates the relationships and paths among all of the variables, as
revised from Figure 2.
In summary, the data analysis results suggest that there is a direct effec
expertise to perceived self-efficacy (F=12.644, p=0.001), levels of task difficulty to
perceived self-efficacy (F=50.811, p<0.01), expertise to mental ef
novice and expert groups, levels of task difficulty to mental effort (F=8.686, p=0.0
nce
erformance F 5). There direct ef
eived self- erformance
relation
β=0.406
β =-0.149
β =-0.054
β =-0.192
β =-0.158
β =0.609
β =0.12
99
onducted. The results suggest that there is a significant relationship between levels of
ance
=99.120, p<0.01). Expertise and levels of task edictors to task
rfo a 8.2 0 ever, mental effort
may
ance rough

performance by going through the mental effort, a multiple regression analysis was
c
task difficulty, expertise, perceived self-efficacy, mental effort, and task perform
(F difficulty are the pr
pe rm =1 5 <0.01; t=-4.724, p<0.01, respectively). How nce (t , p
is not a significant predictor, although the p-value is close to 0.05 (t=-1.685, p=0.093). As
a result, the present study can only infer that expertise and levels of task difficulty
have an indire k ct relationship with tas by going th mental effort. perform

100

CHAPTER IV DISCUSSION
Summary

the
self-efficacy judgments affect the amount of mental effort investment and task
erformance under different levels of task difficulties. Results from this study could be
sed to build a performance model and to improve task performance by understanding
e relationship between these variables. The study used a strategic computer game
arcraft III as the primary task to examine the subjects’ levels of expertise and task
erformances. A developed program that can randomly present a pop-out window on the
omputer screen was also used as a secondary task to interrupt subjects’ engagement of
e primary task and to examine subjects’ amount of mental effort investment during
dual-tasks.
Sixty-six subjects volunteered to participate in the study. The subjects played a
computer game Warcraft III: The Frozen Throne. Designed questionnaire data from five
trials and other quantitative data were collected during the experiment of the study. The
designed questionnaires measured subjects’ perceived self-efficacy, perceived task
difficulty, and mental effort. In addition, quantitative data were collected from the

Results of the data analysis in Chapter III are summarized and discussed in this
section.

This study was designed to determine how different levels of both task difficulty and
expertise affect self-efficacy judgments. Moreover, the study aims to determine how
p
u
th
W
p
c
th
101

subjects’ game performance scores and secondary task response records. All data were
used to test the proposed hypotheses and to address the research questions. Statistical
analy ts,
NOVA test, effect size, and two-way ANOVA test.

Results of e Statistics

owever,

han
one
an
In addition, participants generally had “Moderately certain can win” and
“Ce ing
lot of
ses included the test of reliability, descriptive statistics, multiple regression, t-tes
one-way A
Descriptiv
Most of the study subjects had more than one year of experience playing both
Warcraft III: The Frozen Throne and the race setting of Human in the game. H
on average per week, they spent less than seven hours or eight to fourteen hours playing
the Frozen Throne either on the computer alone or online with other people. These
subjects explained that they played the upgraded version of Warcraft III more often t
the Frozen Throne version. In general, subjects were confident to win the Frozen Thr
game in a longer time limit if the level of task difficulty was high and were confident to
win the game in a shorter time limit if the level of task difficulty was low. This is not
surprising given that people are usually less confident in performing a difficult task th
an easy task.
rtainly can win” confidence assessments in winning the game when the race sett
was Human. Finally, the mean scores of self-reported mental effort in three tasks with
different difficulty levels also showed that more effort was invested when the level of
task difficulty increased. On average, at higher difficulty levels, subjects invested a
effort or more than moderate effort in order to win the game.
102

Hypothesis Number 1a
nvestment of metal
effort also increased. However, the effect size signifies that the difference has small
practical significance.

t
is
There is a positive relationship between levels of task difficulty and mental effort.
Regression results suggest that levels of task difficulty significantly predict mental
effort. As expected, when levels of task difficulty increased, the i
Hypothesis Number 1b
There is a negative relationship between levels of task difficulty and task performance.

Regression results suggest that levels of task difficulty significantly predict task
performances. However, the effect size signifies that the difference has small practical
significance. Overall, the scores of game performances decreased when the levels of task
difficulty increased. There is a significant and negative relationship between levels of
task difficulty and task performance.

Hypothesis Number 2a
There is a negative relationship between mental effort and perceived self-efficacy from
self-reported measure results

Regression results indicate that perceived self-efficacy does not significantly predic
self-reported mental effort, but the relationship between self-efficacy and mental effort
negative.

103

T t
support that people’s effort investment
gradually increases when their efficacy increases. However, people who have very high
self-efficacy judgment tend to inv cause the nonconscious and
a
the resu
9;
Hypothesis Number 2c
There is a positive relationship between perceived self-efficacy and task performance.

Regression results indicate th significantly predicts task
p
sign
increases, task performance also increases.
Hypothesis 2b
here is an inverted U relationship between perceived self-efficacy and mental effor
from dual-task measure results.

The results of dual-task measures found that there is no significant difference between
perceived self-efficacy and dual-task measured mental effort. However the scatter plot
shows that the relationship between perceived self-efficacy and mental effort roughly
reveals the shape of inverted U. The findings
est low mental effort be
utomated cognitive processes are applied to perform the task. The difference between
lts of self-report and dual-task measures may be that people who automatically
defaulted to invest less effort actually did not know this change consciously (Clark, 199
Gimino, 2000).

at perceived self-efficacy
erformance. The effect size also signifies that the difference has large practical
ificance. As many studies suggest (Bandura, 1997; Pintrich & Schunk, 2002),
perceived self-efficacy is one of the predictors of task performance. When perceived
self-efficacy
104

Hypothesis Number 3a
There is a negative relations ertise and mental effort.
Regression results sugge edictor of self-reported
men
re
Hypothesis Number 3b

Hypothesis Number 4a
There is a significant difference between levels of task difficulty, expertise, and mental

mental effor
lts also
tise,

hip between exp
st that expertise is a significant pr
tal effort between novice and expert subjects. However, the effect size signifies that
the difference has small practical significance. Findings suggest that people who are mo
familiar with a task invest less effort than people who are less familiar with a task.

There is a positive relationship between expertise and task performance.
Regression results suggest that expertise is a significant predictor of task
performance. The effect size also signifies that the difference has large practical
significance. As expected, when expertise increases, so does the task performance.
effort.
The independent-samples t-test results indicate that there is a significant difference in
t between novice and expert subjects. In addition, there is a significant
difference in mental effort between normal task and high task. The regression resu
support that there is a significant difference between levels of task difficulty, exper
and mental effort. However, the effect size signifies that the difference has small practical
significance. The relationship is illustrated in Figure 28.
105

Figure 28 The Relationship bet , Levels of Task Difficulty, and
Mental Ef

There is a significant difference between levels of task difficulty, expertise, and task
performance.
The independent-samples t-te ere is a significant difference in
t
lso
e,

ween Levels of Expertise
fort
Hypothesis Number 4b

st results indicate that th
ask performance between the low level of expertise and high level of expertise. There is
a a significant difference in task performance between the normal task and high task
The regression results also find that there is a significant difference between expertis
levels of task difficulty, and task performance. However, the effect size signifies that the
difference has small practical significance. Figure 29 illustrates the relationship.

Expertise
Level of Task
Difficulty
Mental Effort
106

nce
Hypothesis Number 5a
There is a significant difference fficulty, perceived self-efficacy,
The independent-samples t-test results indicate that there is a significant difference in
perceived self-efficacy between normal task and high task. The results also show that
there is a significant difference in mental effort between normal task and high task. In
addition, one-way ANOVA results state that there is no significant difference between
mental effort and perceived self-efficacy.
Finally, the regression results support that there is a significant difference between
signifies that the difference has small practical significance.
Figure 30 illustrates the relationship.

Figure 29 The Relationship between Levels of Expertise, Levels of Task Difficulty, and
Task Performa
Expertise
Level of Task
Dif
Performance
ficulty

between levels of task di
and mental effort.

perceived self-efficacy, levels of task difficulty, and mental effort. However, the effect size

107
re 30 The Relationship between Perceived Self-Efficacy, Levels of Task Difficulty,
nd Mental Effort

There is a significant difference be difficulty, perceived self-efficacy,
and task performance.

nd
hat there is a significant difference between
ificance. Figure 31
illus rates the relationship.

Figu
a

Perceived Self-Efficacy
Lev
Difficulty
Mental Effort
el of Task

Hypothesis Number 5b
tween levels of task
One-way ANOVA results show that there is a significant difference between task
performance and perceived self-efficacy. The independent-samples t-test results also
indicate that there is a significant difference in task performance between normal task a
high task.
Finally, the regression results suggest t
levels of task difficulty, perceived self-efficacy, and task performance. However, the
effect size signifies that the difference has small practical sign
t

108
igure 31 The Relationship between Perceived Self-Efficacy, Levels of Task Difficulty,

There is no significant difference between expertise, perceived self-efficacy, and mental

e in
ficant relationship between expertise and mental effort for novice and expert
sub
rted

F
and Task Performance
Perceived Self-Efficacy
Level of Task
Difficulty
Performance

Hypothesis Number 6a
effort from self-reported measure results.
The independent-samples t-test results indicate that there is a significant differenc
perceived self-efficacy between low level of expertise and high level of expertise. There is
also a signi
jects. However, as expected, the regression results show that there is no significant
difference between expertise, perceived self-efficacy, and mental effort from self-repo
mental effort. Figure 32 illustrates the relationship.

109

Hypothesis 6b
There is a significant relationship between expertise, perceived self-efficacy, and mental
effort from dual-task measure results.
sign
xpertise, perceived self-efficacy, and
e.

s described in Hypothesis 5b, one-way ANOVA results show that there is a
significant difference between task performance and perceived self-efficacy. The

Figure 32 The Relationship between Perceived Self-Efficacy, Levels of Expertise, and
Mental Effort

Perceived Self-Efficacy
Expertise
Mental Effort

The regression results show that expertise and perceived self-efficacy were not the
ificant predictors from dual-task measured mental effort results. There is no
significant difference between expertise, perceived self-efficacy, and mental effort from
dual-task measure results.

Hypothesis Number 6c
There is a significant difference between levels of e
task performanc
A
110
ifference between
nce. The effect size also signifies that
ce. Figure 33 illustrates the relationship.

, Levels of Expertise, and
T
Research Question 1

What is the relationship betw task difficulty, and self-efficacy
997;
Kruger, 1999) that people have high perceived self-efficacy when they believe the task to

independent-samples t-test results also indicate that there is a significant difference in
task performance between low level of expertise and high level of expertise.
The regression results also show that there is a significant d
perceived self-efficacy, expertise, and task performa
the difference has large practical significan
Figure 33 The Relationship between Perceived Self-Efficacy
ask Performance
Performance
Perceived Self-Efficacy

Level of Expertise
een expertise, levels of
judgments?

This study found that higher levels of expertise are associated with higher
self-efficacy judgments. This result is consistent with previous findings (Bandura, 1
111
win d that
r the
lf-efficacy, the lower was the mental effort invested. As discussed in the Literature
overconfident person thinks that he knows what he is doing, so he
nce. Therefore, the result
s when they

elf-efficacy,
e, mental effort, and task performance?

the amount of mental effort investment
decreases when level of expertise increases. This result confirms that experts have their
schemas highly automated when s alyuga, Ayres, Chandler, &
Sweller, 2003). The result applied m
Secondly, the current study also showed onship between self-reported
mental effort and perceived self-efficacy is a linear shape while the relationship between
dual-task measured mental effort and perceived self-efficacy roughly showed the shape
of inverted U. The results align with Clark’s (Clark, 1999) automated cognitive default

be familiar. The results also confirmed that people’s self-efficacy judgments decrease
when tasks become difficult. In this study, subjects believed it took more time for them to
the high task than moderate and easy tasks. In addition, the present study foun
there is a significant difference between expertise and perceived self-efficacy. The higher
the level of expertise, the higher was the subject’s confidence. Conversely, the highe
se
Review section, an
does not work very hard on a task which leads to poor performa
provides the inference that experts may commit overconfidence problem
perceive a task familiar but it is not (Clark & Estes, 2002; Hill, 2000).
Research Question 2

What is the relationship between levels of task difficulty, perceived s
expertis
This study, like previous studies, found that
olving familiar tasks (K
ost strongly to the novice and expert subjects.
that the relati
112

Besides, the present study results generalize previous findings that the levels of task

een expertise,
perceived self-efficacy, and task performance, as well as, between levels of task difficulty,
perceived self-efficacy, and task performance. The result is consistent with Bandura’s
assertion that people’s self-efficacy judgment is influenced by their ability in applying the
domain skill and know difficulty. The
stu
notion that once subjects reach an efficacy threshold, they tend to invest fewer efforts to
the task.
difficulty affect the performance and effort investment. The more difficult the task, the
worse is the performance and the more efforts are needed (Flad, 2002; Gimino, 2000).
Results also show that there is a direct relationship between expertise and task
performance. People with higher levels of expertise perform better than people with
lower levels of expertise. The result is consistent with the expert theory that expert’s
performance is superior to non-expert in a specific task domain (Ericsson & Charness,
1994; Ericsson & Smith, 1991). In addition, the study also found that perceived
self-efficacy is a strong predictor to task performance as described in many other studies
(Pajares, 1996; Pajares & Schnk, 2001; Oliver & Shapiro, 1993).
Path analysis results show that there is an indirect relationship betw
ledge as well as their perception of levels of task
dy also shows that perceived self-efficacy and expertise are two important factors in
predicting task performance not only in the statistic setting but also in the practical
situation.

113
ture study to prove the significance.
dy failed to find that efficacy belief was a significant predictor to self-reported
men
study to
Conclusions
The results of this study conclude that the relationships between expertise, levels
of task difficulty, perceived self-efficacy, mental effort, and task performance are as
follows:
1. There is a positive relationship between levels of task difficulty and
mental effort.
2. There is a negative relationship between levels of task difficulty and
task performance.
3. There is a negative relationship between self-reported mental effort and
perceived self-efficacy.

The present study failed to find that expertise and levels of task difficulty have an
indirect relationship with task performance through mental effort. It is necessary for
fu
Finally, unlike the results of previous studies (Gan, 2005; Bandura, 1986, 1997), the
present stu
tal effort although the correlation index showed the relationship is negative. The
study also failed to find a significant inverted U relationship between perceived
self-efficacy and dual-task measured mental effort although the scatter plot roughly
reveals the inverted U shape. It is possible that more subjects are needed for the
find the significance.

114
There is a positive relationship between perceived self-efficacy and task
performance.
5. There is a negative relationship between expertise and mental effort for
novice and expert groups.
ce.

y.
ulty
se, and task performance. Both levels of task difficulty and
expertise contribute the prediction to task performance.
11. There is a significant difference between levels of task difficulty,
perceived self-efficacy, and mental effort. However, only levels of task
difficulty contribute the prediction to mental effort.
12. There is a significant difference between levels of task difficulty,
perceived self-efficacy, and task performance. Both levels of task
difficulty and perceived self-efficacy contribute the prediction to task
performance.

4.
6. There is a positive relationship between expertise and task performan
7. There is a negative relationship between levels of task difficulty and
perceived self-efficac
8. There is a positive relationship between expertise and perceived
self-efficacy.
9. There is a significant difference between levels of task difficulty,
expertise, and mental effort. However, only levels of task diffic
contribute the prediction to mental effort.
10. There is a significant difference between levels of task difficulty,
experti
115
dations for Future Research

The jor lim
The Frozen Throne had been a popular strategy game from about year 2003 to 2006. This
ersion of the game was then upgraded, and more players began playing the new version.
The up ed ver
researcher’s knowledge, no other strategy games are as popular as the Warcraft series. As
a result the new
problem of small o the result of unsuccessfully predicting the
relation p betw
failure inding
self-efficacy and al effort. Moreover, the effect size and
R-squa alues i lso
weak because of ple subjects. Future studies would need to address this
limitati
Another limit
study, the researc at the average time required
to win l the ers
might be able to w ers to
complete the gam h a ten minute
trial tim se
would be able to
performance scor ulty levels.

Recommen
ma itation of this study is the problem of small sample size. Warcraft III:
v
grad sion is not appropriate for use in this study. In addition, from the
of version, it was extremely difficult to find volunteers for the study. The
sample size may have led t
shi een mental effort and perceived self-efficacy. It may also have led to the
of f a more expressive inverted U-shape relationship between perceived
dual-task measured ment
re v n explaining the significance results of regression analyses were a
the lack of sam
on.
ation of the study is the length of the trial time. Before designing the
her worked with a few players and found th
or fai game was around twenty to twenty-five minutes. While expert play
in the game in ten to fifteen minutes, it would be difficult for play
e in ten minutes. As a result, the study was designed wit
e to control the testing length and pace. Players with higher levels of experti
complete most of the games and have higher scores. Moreover, the
es of the different tasks would also reflect their diffic
116

However, it was
avoid attacks and
This may be a key reason why the results of perceived task difficulty were not consistent.
The determination of trial time, therefore, may be required for future study to be
cautiously considered.
Finally, the study uses a strategic computer game as the cognitive task to measure
subjects’ expertise and task performance. Future research may consider using different
cognitive tasks in other fields to replicate the study for generalization.

found that many subjects chose to stay in the base much of the time to
did not encounter the enemies until the last few minutes of the game.

117

e
with expertise. Journal of Experimental Psychology, 10, 483-495.
Allport, D. A., Antonis, B., & Reynolds, P. (1972). On the division of attention: A
ology,
24, 225-235.
Anderson, John. (1995). Cognitive Psychology and its Implications. 4
th
Ed. W. H.

theory.
Englewood Cliffs, NJ: Prentice Hall.
Bandura, A. (1997) Self efficacy: The exercise of control. New York: W. H. Freeman.

Bandura, A. & Schunk, D. H. (1981). Cultivating competence, self-efficacy, and
Psychology, 41, 586-598.
Beach, L. R., Barnes, V. E., & Christensen-Szalanski, J. J. (1986). Beyond heuristics and
biases: A contingency model of judgmental forecasting. Journal of Forecasting, 5,
143-157.

Beatty, J. (1982). Task-evoked papillary responses, processing load and the structure of
processing resources. Psychological Bulletin, 91(2), 176-292.

Berger, H. (1969). Hans Berger on the electr ncephalogram of man (P. Gloor, Trans.),
Electroeneceph. Clinical Neurophysiol., Suppl. 28.

Bong, M. (1997). Generality of academic self-efficacy judgments: Evidence of
hierarchical relations. Journal of Educational Psychology. 89(4). 696-709.

Bong, Mimi (1999). Personal Factors Affecting the Generality of Academic Self-efficacy
Judgments: Gender, Ethnicity, and Relative Expertise. The Journal of Experimental
Education, 67(4), 315-331.

REFERENCES
Adelson, B. (1984). When novices surpass experts: The difficulty of a task may increas

disproof of the single channel hypothesis. Quarterly Journal of Experimental Psych

Freeman and Company. New York.
Anderson, John (1993). Rules of the mind. Hillsdale, NJ: Erlbaum.
Bandura, A. (1977) Social learning theory. Englewood Cliffs, NJ: Prentice Hall.
Bandura, A. (1986). Social foundations of thought and action: A social cognitive

instrinsic interest through proximal self-motivation. Journal of Personality and Social

oe
118

Bong Mimi (2001). Between and within dom in relations of academic motivation among
middle and high school students: self-efficacy, task value, and achievement goals.
Journal of Educational Psychology, 93(1), 23-34.

Britton,B.K.,Muth,K.D.,&Glynn,S.M.(1986). Effects of text organization on memory:
Test of a cognitive effort hypothesis with lim ed exposure time. Discourse Processes, 9,
475-487.

Brookhart, M. S., Walsh, M. J., Zientarski, A. W. (2002). Motivation and Effort in
Middle School Science and Social Studies Tests and Performance Assessments. Paper
presented at the annual meeting of the National Council on Measurement in Education,
April, New Orleans, LA.

Budescu, D. V., Erev, I., & Wallsten, T. S. (1997). On the importance of random error in
in
runken, R., Steinbacher, S., Schnotz, W., & Leutner, D. (2002). Assessment of
rt and video-based
arning: The relationship of preconceptions and the effects of interactive and covert
y, 10, 1-20.

5-81.

en, S. F. Chipman, & V. L. Shalin (Eds.), Cognitive Task
nalysis (pp. 3-23). Mahwah, NJ: Lawrence Erlbaum Associates.
lark, R. E. (1999a) The CANE model of motivation to learn and to work: A two-stage
a
it
the study of probability judgment: Part I: New theoretical developments. Journal of
Behavioral Decision Making, 10, 157-171.

Brunken, R., Plass, J. & Leutner, D. (2003). Direct Measurement of Cognitive Load
Multimedia Learning. Educational Psychologist, 38(1), 53-61.

B
cognitive load in multimedia learning using dual-task methodology. Experimental
Psychology, 49, 109-119.

Cennamo, K. S., Savenye, W. C. & Smith, P. L. (1991). Mental effo
le
practice. Educational Technology Research and Development, 39(1), 5-16.

Chandler, P. & Sweller, J. (1996). Cognitive load while learning to use a computer
program. Applied Cognitive Psycholog
Chase, W. G. & Simon, H. A. (1973). Perception in chess. Cognitive Psychology, 4,
5

Chipman, S. F., Schraagen, J. M., & Shalin, V. L. (2000). Introduction to cognitive task
analysis. In J. M. Schraag
A

Clark R.E. & Fred Estes (2000). Turning Research Into Results. CEP Press.

C
process of goal commitment and effort. Lowyck, J. (Ed.) Trends in Corporate Training.
Leuven Belgium, University of Leuven Press.

119
ational Processes Operating in
ultimedia Learning Environments. In van Merrienböer, J. (Ed.) Cognition and
lark, R.E & Elen, Jan (2006), When Less is More:Research and Theory Insights about
ovington, M.V. (1992). Making the grade: A self-worth perspective on motivation and
chool reform. Cambridge: Cambridge University Press.
s on
mputers in Human Behavior, 14(2),
49-267.
eMarie, D., Aloise-Young, A. P., Prideaux, L. C. & Gerda, H. J. (2004). College
on
,
unning, David, Johnson, Kerri, Ehrlinger, Joyce, & Kruger, Justin (2003). Why People
,
ricsson, A. K. & Kintsch, W. (1995). Long-Term Working Memory. Psychological
ricsson, K. A., & Charness, N. (1994). Expert performance: Its structure and acquisition.
. T., & Tesch-Romer, C. (1993). The role of deliberate
ractice in the acquisition of expert performance. Psychological Review, 100, 363-406.

gical

Clark, R. E. (1999b), Yin and Yang Cognitive Motiv
M
Multimedia Design. Herleen, Netherlands: Open University Press.

C
Instruction for Complex Learning. Handling Complexity in Learning Environments:
Theory and Research. Elsevier Ltd.

C
s

de Croock, M. B., van Merrienboer, J .G., & Paas, F. G. (1998). High versus low
contextual interference in simulation-based training of troubleshooting skills: effect
transfer performance and invested mental effort. Co
2

DeGroot, A. (1965). Thought and Choice in Chess. The Hague, Netherlands: Mouton.

D
Students’ Memory for Vocabulary in Their Majors: Evidence for a Nonlinear Relati
Between Knowledge and Memory. Canadian Journal of Experimental Psychology, 58:3
181-195.

D
Fail to Recognize Their Own Incompetence. Current Directions in Psychological Science
12, 83-87.

E
Review, 102(2), 211-245.

E
American Psychologist, 49(8), 725-747.

Ericsson, K. A., Krampe, R
p

Ericsson K. A., Smith, J. (1991). Towards a general theory of expertise: Prospects and
limits. New York: Cambridge University Press.

Fairclough, H. S. & Houston, k. (2004). A metabolic measure of mental effort. Biolo
Psychology, 66, 177-190.

120

ty
t and
ual-task measures of mental effort during problem solving. Unpublished doctoral
ce students’ judgment bias on test
uestions. Journal of Nursing Education, 40(1), 10-16.
ical,
f Psychophysiology,
1, 129-145.
nce
rse. Unpublished doctoral dissertation, University of
outhern California.
em Solving domains.

rol of
. Amy (2000). Factors that influence students’ investment of mental effort in
cademic tasks: A validation and exploratory study. Unpublished doctoral dissertation,

Braune R (1984) On the psychophysics of workload: Why bother with
bjective measures? Human Factors, 26, 519-532
f attention. In G.R.J Hockey, A.W.K. Gaillard, & M.G.H. Coles (Eds.),
Energetics and human information processing. Dordrecht: Martinus Nijhoff Publishers.
Deborah R., & Marcolin, Barbara L. (2002). Meta Cognition
nd IT: The Influence of Self-efficacy and Self-awareness. Paper Presented in the
Farrington, Jeanne Ruth (1997). The impact of structural features and context on
improving transfer in concept acquisition. Unpublished doctoral dissertation, Universi
of Southern California.

Flad, Julie Ann (2002). The effects of increasing cognitive load on self-repor
d
dissertation, University of Southern California.

Flannelly, T. L. (2001). Using feedback to redu
q

Fournier, R. Lisa, Wilson, F. Glenn, & Swain, R. Carolyne (1999). Electrophysiolog
behavioral, and subjective indexes of workload when performing multiple tasks:
manipulations of task difficulty and training. International Journal o
3

Gan, Sandy (2005). An evaluation of perceived task value, self-efficacy, and performa
in a geography blended distance cou
S

Garner, Stuart (2000). Cognitive Load Reduction in Probl
Gibbs, Colin (2003). Explaining effective teaching: self-efficacy and thought cont
action. Journal of Educational Enquiry, 4(2), 1-14.

Gimino, E
a
University of Southern California.

Glass, A. (1966). Comparison of the effect of hard and easy mental arithmetic upon
blocking of the occipital alpha rhythm. Q. J. Experimental Psychology. 8, 142-154.

Gopher D,
su

Gopher, D. (1986). In defense of resources: On structures, energies, pools, and the
allocation o

Gravill, Jane I. Compeau,
a
meeting of Eighth Americas Conference on Information Systems.

121
xamination of a metacognitive skill. Journal of Economic Education, 33(1), 15-30.
al changes in the ongoing EEG related
the difficulty of mental tasks. Brain Topogr, 5, 17-25.
1989). An exploration of the mathematics self-efficacy
athematics performance correspondence. Journal of Research in Mathematics
ducation, 20, 261-273.
ill Jr., Jay James, (2000). Are executives overconfident? Ed.D. Dissertation: University
oncept of Operator Workload in Manual Vehicle Operations.
orschungsinstitut Anthropotechnik, Meckenheim, West Germany (Technical report No.
ias in probabilistic assessment: On the interpretation of stochastic components of
yce E. & Biddle, G. (1981a). Anchoring and adjustment in probabilities inference in
Research, 19, 323-349.
ahneman, D.(1973). Attention and effort. Englewood Cliffs NJ: Prentice-Hall.
Kalyuga, S. & Sweller, J. (2004). Measuring Knowledge to Optimize Cognitive Load
yuga, S. Chandler, P., & Sweller, J. (1999). Managing split-attention and redundancy
P., & Sweller, J. (2003). The Expertise Reversal
ffect. Educational psychologist, 38(1), 23-31.
ental
nalysis. Contemporary Educational Psychology, 13, 221-228.

Grimes, W. P. (2002). The overconfident principles of economics student: an
e

Gundel, A. & Wilson, G. F. (1992). Topographic
to

Hackett, G. & Betz, N .E. (
m
E

H
of Southern California.

Jahns, D. W. (1973). A C
F
14).

Juslin, P., Olsson, H., & Bjorkman, M. (19976). Brunswikian and Thurstonian origins of
b
judgment. Journal of Behavioral Decision Making, 10, 189-209.

Jo
auditing. The Accounting Review, LVIII(1), 78-97

Joyce E. & Biddle, G. (1981b). Are auditors’ judgments sufficiently regressive? Journal
of Accounting

Kahneman, D.(1970). Remarks on attentional control. In A.F. Sanders(Ed.), Attention
and performance, 3, 118-131.

K

Factors During Instruction. Journal of Educational Psychology, 96(3), 558-568.

Kal
in multimedia instruction. Applied Cognitive Psychology, 13, 351-371.

Kalyuga. S., Ayres. P., Chandler.
E

Kee, W. Daniel & Davies, Leslie (1988). Mental effort and Elaboration: A developm
A
122
ch

Kerlinger (1986). Foundations of behavioural resear (3 Ed.) Fort Worth Holt,
inehart & Winston.
nd
ial
ruger, J., & Dunning, D. (1999). Unskilled and unaware of it: How difficulties in
cognizing one’s own incompetence lead to self-inflated assessments. Journal of
ruger, J., & Dunning, D. (2002). Unskilled and unaware-but why? A reply to Krueger &
ee, T. D. & Magill, R. A. (1983). The locus of contextual interference in motor skill
n, 9,
ee, T. D. & Magill, R. A. (1985). Can forgetting facilitate skill acquisition? In D.
ntrol, Amesterdam, The Netherlands, Elsevier Science
usblishers.
ational Psychology, 88, 49-63.
ional

use in applied environments. Biology
sychology, 34, 205-236.
997). A computational theory of executive cognitive
rocesses and multiple-task performance: Part 1. Basic mechanisms, Psychological
rd
R

Koehler, D. J., Brenner, L. A., Liberman,V., & Tversky, Amos (1996). Confidence a
Accuracy in Trait Inference: Judgment by Similarity. Acta Psychology, 92 , 33-57.

Kruger, J. (1999). Lake Wobegon be gone! The "below-average effect" and the
egocentric nature of comparative ability judgments. Journal of Personality and Soc
Psychology, 77, 221-232.

K
re
Personality and Social Psychology, 77, 1121-1134

K
Mueller. Journal of Personality and Social Psychology, 82, 189-192.

L
acquisition. Journal of Experimental Psychology: Learning, Memory, and Cognitio
730-746.

L
Goodman, R. B. Wilberg, & I. M. Franks (Eds.) Differing perspectives in motor
learning, memory and co
P

Marcus, N., Cooper, M., & Sweller, J. (1996). Understanding instructions. Journal of
Educ

Mayer, R., & Chandler, P. (2001). When learning is just a click away: Does simple user
interaction foster deeper understanding of multimedia messages? Journal of Educat
Psychology, 93, 390-397.

McCloskey, M. (1983). Naïve theories of motion. In D. Gentner & A. L. Stevens (Eds),
Mental Models, 299-323. Hillsdale, NJ: Erlbaum.

Mecklinger, A., Kramer, A. F., & Strayer, D.I. (1992). Measurement and analysis
methods of heart rate and respiration for
P

Meyer, D. E., & Kieras, D. E. (1
p
Review, 104, 3-65.

123
ocessing
stem. Psychological Review, 86, 214-255.
f the
ngle-bottleneck notion. Cognitive Psychology, 44, 193-251.
y and computers. Journal of
omputer-Based Interactions, 20, 81-85
g skills in statistics: A
ognitive load approach. Journal of Educational Psychology. 84, 429-434.
ecent Developments. Educational Psychologist,38(1), 1-4.
001)
erception (p.239-266). London: Ablex Publishing.
y
school students, Journal of Educational Research. Review of
ducational Research, 53, 445-459.
cs
ournal of Educational Psychology, 86,
93-203.
d mathematical
roblem solving: Implications of using varying forms of assessment. Paper presented at
ank (1996). Self-efficacy Beliefs and Mathematical Problem Solving of Gifted
tudents. Contemporary Educational Psychology, 21, 325-344.

ajares, M. & Graham, L. (1999) Self-efficacy, Motivation Constructs, and Maths
:

Navon, D., & Gopher, D. (1979). On the economy of the human information pr
sy

Navon, D., Miller, J. (2002). Queuing or sharing? A critical evaluation o
si

Oliver, T. A., & Shapiro, F. (1993). Self-efficac
C

Paas, F. (1992). Training strategies for attaining problems solvin
c

Paas, F., Renkl, Alexander, & Sweller, John (2003). Cognitive Load Theory and
Instructional Design: R

Pajares, F. & Schunk, H. D. (2001). Self-beliefs and School Success: Self-efficacy,
Self-concept, and School Achievement. Chapter in R. Riding & S. Rayner (Eds.), (2
P

Pajares, F. & Valiante, G. (in press). Predictive and mediational roles of the self-efficac
beliefs of upper elementary
E

Pajares, Frank & Miller, M. D. (1995a). Mathematics Self-efficacy and Mathemati
outcomes: The need for specificity of assessment. J
1

Pajares, Frank & Miller, M. D. (1995b). Mathematics Self-efficacy an
p
the meeting of the American Psychological Associations, New York.

Pajares, Fr
S

Pajares, Frank (2002). Overview of Social Cognitive Theory and Self-efficacy. from
http://www.emory.edu/EDUCATION/mfp/eff.html.

P
Performance of Entering Middle School Students. Contemporary Educational
Psychology, 23, 124-139.

Pedhazur, E. J. (1997). Multiple Regression in Behavioral Research (3rd Ed). Fort Worth
Hartcourt Brace College Publishers.

124
Presented in the meeting of Australia
ssociation for Research in Education. Sydney, Australia.
arch, and
liffs, NJ: Prentice Hall.
ollock, E. Chandler, P., & Sweller, J. (2002). Assimilation complex information.
earning and Instruction, 12, 61-86.
rconfidence, base rates and outcome
ositivity/negativity of predicted events. The British Journal of Psychology, 87, 431-445.
back and item difficulty
ffects. Personality and Individual Differences, 23, 125-133.
onship of Self-Efficacy and Complacency in
ilot-Automation Interaction. NASA STI Report # NASA/TM-2002-211925.
t-novice
ecision strategies for residential burglary. Psychonomic Bulletin & Review, 16(1),
owe, L. A., Hall, P. E., Cooke, J. N., & Halgren, L. T. (1996). Toward an On-Line
ing
ugg. M .D, & Dicnkens, A.M.J. (1982). Dissociation of alpha and theta activity as a
usso, J. E., and Schoemaker, P. J. (1992). Managing overconfidence. Sloan
ources. Educational Psychologist. 18(1), 42-50.
ent

Phan, Huy & Walker, Richard (2000). The Predicting and Mediational Role of
Mathematics Self-efficacy: A Path Analysis. Paper
A

Pintrich, P. R. & Schunk, D. H. (1995). Motivation in Education: Theory, Rese
Applications. Englewood C

Pintrich, P. R. & Schunk, D. H. (2002). Motivation in education: Theory, research
and application (2
nd
ed.). Upper Saddle River, NJ: Merrill Prentice Hall.

P
L

Pulford, B. D. & Colman, A. M. (1996). Ove
p

Pulford, B. D. & Colman, A. M. (1997). Overconfidence: Feed
e

Prinzel III, Lawrence J (2002). The Relati
P

Rocio Garcia-Retamero, Mandeep K Dhami. (2009). Take-the-best in exper
d
163-170.

R
Knowledge Assessment Methodology: Building on the Relationship Between Know
and Doing. Journal of Experimental Psychology, 2(1), 31-47.

R
function of verbal and visuospatial tasks. Electroenceph. Clinical Neurophysiol, 53,
201-207.

R
Management Review, 23, 7-17.
Salomon, G. (1983). The differential investment of mental effort in learning from
different s

Salomon, G. (1984). Television is "easy" and print is "tough": The differential investm
of mental effort in learning as a function of perceptions and attributions. Journal of
Educational Psychology. 76(4), 647-658.

125
fluence of Modeling,
oal Setting, and Self-Evaluation. Reading & Writing Quarterly, 19, 159-172.
nd schemata. In R. W.
chvaneveldt (Ed.). Pathfinder associative networks: Studies in knowledge organization.
. (1979). Contextual interference effects on the acquisition,
tention, and transfer of a motor skill. Journal of Experimental Psychology, Human
ntext effects in memory and learning movement
formation. In R. A. Magill (Ed.), Memory and control of action. New York: Elsevier.
outcome
hology, 87, 386-398.
self-efficacy on the relationship
etween performance and attributions. Organizational behavior and Human Decision
taggers, N. & Norcio, A. F. (1993). Mental models: Concepts for human-computer
tephan, Ekkehard & Kiell, Guido. Decision Processes in Professional Investors: Does
rence

cision
rocesses, 59(3), 452-424.

nd Instruction, 4, 295-312.
98). Cognitive architecture
nd instructional design. Educational psychology review, 10, 251-296.

Schunk, D. H. (2003). Self-efficacy For Reading and Writing: In
G

Schvaneveldt, R. W. (1990). Proximities, networks, a
S
Norwood, NJ:Albex.

Shea, J. B. & Morgan, R. L
re
Learning and Memory, 5, 179-187.
Shea, J. B. & Zimny, S. T. (1983). Co
in

Shell, D. F., Colvin, C. & Brunning, R. H. (1995) Self-efficacy, attributions, and
expectancy mechanisms in reading and writing achievement: Grade-level and
achievement level differences. Journal of Educational Psyc

Silver, S. W., Terence R. M., & Gist, E. M. (1995). Responses to successful and
unsuccessful performance: the moderating effect of
b
Processes, 62(3), 286-299.

S
interaction research. International Journal of Man-Machine Studies, 38, 587-605.

S
Expertise Moderate Judgmental Biases? Paper presented at the 29th EMAC Confe
in Rotterdam, May 2000.

Sweller, J. (1988) Cognitive load during problem solving: effects on learning. Cognitive
Science, 12, 257-285.

Stone, D. N. (1994). Overconfidence in initial self-efficacy judgments: Effects on
decision processes and performance. Organizational Behavior & Human De
P

Sweller, J. (1994) Cognitive load theory, learning difficulty and instructional design.
Learning a

Sweller, J. (1999). Instructional design. Melbourne: ACER.

Sweller, J., van Merrienboer, J. K. G., & Paas, F. G. W. C. (19
a
126
nt
al design model for technical training. Englewood Cliffs, NJ: Educational
echnology Publications
M., Tischner, E. C. & Putka, D. J. (2002). Two studies
lied
ancouver, J. B., & Scherbaum, K. C. (2002). Self-Efficacy and Resource Allocation
lanning: A Discontinuous Model. Poster presented at the annual meeting of the Society
erwey, W. B. & Veltman, H. A. (1996). Detecting short periods of elevated workload:
tal
illachica, W. S., Lohr, L. L., Summers, L., Lowell, N., Roberts, S., Javeri, M., Hunt, E.,
:
f
munications and Technology, Atlanta, GA.
tructure, Attention, and Expertise. Applied Cognitive Psychology, 17, 295-308.
. R.
ineering psychology and human performance, 2nd Ed. New
ork, NY: Harper Collins.
eoretical analysis. Developmental Psychology, 27, 552-565.
Management Review, 14, 361-384.

Van Merrienboer, J. J. G. (1997). Training complex cognitive skills: A four-compone
instruction
T

Vancouver, J. B., Thompson, C.
examining the negative effect of self-efficacy on performance. Journal of Applied
Psychology, 87, 506-516.

Vancouver, J. B., Thompson, C. M., Williams, A. A. (2001). The changing signs in the
relationships between self-efficacy, personal goals and performance. Journal of App
Psychology, 86, 605-620.

V
P
for Industrial and Organizational Psychology, Toronto, Canada.

V
A comparison of nine workload assessment techniques. Journal of Experimen
Psychology: Applied, 2, 270-285.

V
Mahoney, C. & Conn, C. (2001). A Cognitive Map of Human Performance Technology
A Study of Domain Expertise. Practice Papers Presented at the National Convention o
the Association for Educational Com

Weber, Nathan & Brewer, Neil (2003). Expert Memory: The Interaction of Stimulus
S

Weinberg, A. B. (2005). A model of overconfidence. from
http://econ.ohio-state.edu/weinberg.

Wickens, C. D. (1984). Processing resources in attention. In R. Parasuraman & D
Davies (Eds.), Varieties of attention. London: Academic.

Wickens, C.D. (1992). Eng
Y

Wigfield, A., Eccles, J. (1992). The development of achievement task values: A
th

Wood, R. E. & Bandura, A. (1989a). Social cognitive theory of organizational
management. Academy of

127
., Patalano, A. L., & Sieck, W. R. (1998).
ross-cultural variations in probability judgment accuracy: Beyond general knowledge
eh, Y. Y., Wickens, C. D. (1988). Dissociation of performance and subjective measures
ip between Self-efficacy and Amount of Mental
ffort Invested in Mathematics Problem Solving By Adults. Unpublished doctoral

Yates, J. F., Lee, J. W., Shinotsuka, H
C
overconfidence? Organizational Behavior and Human Decision Processes, 74(2),
89-117.

Y
of workload. Human Factors, 30, 111-120.
Yoshida, M. C. (2000). The Relationsh
E
dissertation, University of Southern California.

128
APPENDICES

University of Southern California

****************
CONSENT TO PARTICIPATE IN RESEARCH

mental
asked to participate in a research study conducted by Hsin-Ning Ho, the student
f Doctor of Education program, and Richard E, Clark, Professor, from the
d are eligible to participate in this research
re
ctoral
lected from the College campuses in Los
Angeles, Orange, and San Bernardino counties to participate. Your participation is
voluntary. You should read the information below, and ask questions about anything you
o not understand, before deciding whether or not to participate

PURPOSE OF THE STUDY
he purpose of this study is to determine how different levels of task difficulty and levels
f knowledge about a computer game affect people's confidence judgments related to
eir game playing expertise. Another purpose is to determine how the confidence
dgments affect people's decisions to invest mental effort that may influence task
erformance.
he researcher will use a strategic computer game Warcraft III as the "task" to measure
our levels of knowledge. Since this computer game allows you to set three different
ifficulty levels (Easy, Normal, Insane-very difficult), the researcher will also be able to
anipulate the levels of task difficulty encountered by you. Survey questions will be
sed to measure your confidence levels under different task difficulty levels (easy,
ormal, and insane). At the same time, the researcher will attempt to estimate your
vested mental effort by asking you to respond (by pressing a keyboard button quickly)

Appendix A
Subject Consent Form
Department of Educational Psychology
Los Angeles, CA 90089

**************************************************

Will Levels of Expertise and task difficulty affect Self-efficacy judgments on
effort and task performance?

You are
o
Educational Psychology Department at the University of Southern California because you
responded to an email or recruitment flyer an
study, and have knowledge in playing a strategic computer game WarcraftIII and a
aged 18 years or older. The results will be contributed to Ms. Hsin-Ning Ho’s do
dissertation. A total up to 181 subjects will be se
d
T
o
th
ju
p

T
y
d
m
u
n
in
129

Appendix A (Continued)
a randomly pop-out window on the computer screen. Finally, the scores of the
omputer game will be used to measure your task performance.
ROCEDURES
he study contains five trials. There are two trials in the first phase and three trials in the
second phase. It will take 10 minutes to complete each trial. You will be asked to fill out
background and perceived self-efficacy questionnaires before the first trial,and then again
arcraftIII installed and the other has a pop-out window program installed. You will
eed to “do your best” to win the game during the trial. In the first trial, you should only
lay WarcraftIII on the first laptop without responding to the pop-out window task on the
second computer. From the second to the fifth trial, you are asked to play WarcraftIII on
the first laptop and need to respond to a randomly appearing pop-out window by pressing
a keyboard button as quickly as you can on the second computer.
The entire study will take approxima tes: (a) approximately 10 minutes
for introduction and procedure demons proximately 5 minutes for research
consent form and background question roximately 5 minutes for perceived
self-efficacy questionnaire, (d) approximately 10 minutes for the first trial (e)
approximately 10 minutes for the mately 10 minutes break time (g)
approximately 10 minutes for the third trial, (e) approximately 3 minutes for invested
mental effort questionnaire e fourth trial (i)
approximately 3 minutes ately 10
minutes for the fifth trial (k) app nvested mental effort
questionnaire, (l) approximately 5 minutes for perceived task difficulty questionnaire.
here will be no direct or potential risk for you to participate in this study. There is no
foreseeing conditio pation. However,
you are able to stop the game at any point you want or under any condition when you feel

OTENTIAL BENEFITS TO SUBJECTS AND/OR TO SOCIETY

to
c

P
T
after each of the remaining trials. There are two laptop computers in front of you when
you are participating in these five trials. One laptop has the strategic computer game
W
n
p
tely 90-110 minu
tration, (b) ap
naire, (c) app
second trial (f) approxi
(h) approximately 10 minutes for th
for invested mental effort questionnaire (j) approxim
roximately 3 minutes for i

POTENTIAL RISKS AND DISCOMFORTS
T
n to cause you feeling discomfort during your partici
any discomfort.
P
You may not receive any benefit from your participation in this research study.
However, it is hoped that the researching on the problems caused by the judgments of
self-efficacy and invested mental effort can help college students understand their
accurate knowledge level and motivate them to progress in their academic achievement.

PAYMENT/COMPENSATION FOR PARTICIPATION
You will be paid $15.00 for the participation. However, your participation is voluntary
and you are able to stop the participation at any point.

130
Appendix A (Continued)
main confidential and will be disclosed only with your permission or as
quired by law.

PARTICIPATION AND WIT
ou can choose whether to be in this study or not. If you volunteer to be in this study,
main in the study. The
vestigator may withdraw you from this research if circumstances arise which warrant
stigator will ask you to terminate the study only in the condition that

Ms.

CONFIDENTIALITY
Any information that is obtained in connection with this study and that can be identified
with you will re
re
The data will be placed in a secure, locked office location, and only Dr. Richard E. Clark
and Ms. Hsin-Ning Ho will have access. A random identification number will be
assigned to you in the beginning of the study. Only assigned number, instead of your
names, will be used for analysis. There will be no identifiable information collected. The
data will be stored for three years after the study has been completed. When the results of
the research are published or discussed in conferences, no information will be included
that would reveal your identity.
HDRAWAL
Y
you may withdraw at any time without consequences of any kind. You may also refuse
to answer any questions you don’t want to answer and still re
in
doing so. The inve
you do not know how to play the strategic computer game Warcraft III or your age is
under 18.

IDENTIFICATION OF INVESTIGATORS
If you have any questions or concerns about the research, please feel free to contact
Hsin-Ning Ho at 626-617-7976 or via email: hsinho@usc.edu any day from 9:00am
9:00pm and at the address of USC, Rossier School of Education, Los Angeles, CA
90089.

RIGHTS OF RESEARCH SUBJECTS
You may withdraw your consent at any time and discontinue
to

participation without
r
a
r
695, (213)
penalty. You are not waiving any legal claims, rights or remedies because of you
participation in this research study. If you have questions regarding your rights as
research subject, contact the University Park IRB, Office of the Vice Provost fo
Research, Grace Ford Salvatori Hall, Room 306, Los Angeles, CA 90089-1
821-5272 or upirb@usc.edu.

131

understand the procedures described above, and I understand fully the rights of a
y questions have
been

Appendix A (Continued)

I
potential subject in a research study involving people as subjects. M
been answered to my satisfaction, and I agree to participate in this study. I have
given a copy of this form.

Name of Subject

ignature of Subject
Date

S

SIGNATURE OF INVESTIGATOR

I have explained the research to the subject, and answered all of his/her questions. I
believe that he/she understands the information described in this document and freely
consents to participate.

Name of Investigator

ignature of Investigator S Date

132
endix B
____________________ Gender (male, female)
#:________________________________________ Age (in years):________
. Please indicate how you would categorize yourself? (Please circle one)
) Caucasian (2) Hispanic (3) African American/Blac,
) Chinese (5) Japanese (6) Vietnamese (7) Taiwanese
) Pacific Islander (9) Native American (10) Asian Indian
1) Other:___________

2. Have you ever played any other strategy g e(s) within the past one year? (Yes/No)
If Yes, please specify the name of the gam __________________
2. Orc
n
. Please indicate what percentage your Warcraft 3 play is with each of the following
an _________
2. Orc _________
3. Undead _________
4. Night Elf _________
100%

. Please answer the following questions based on the game IN GENERAL:
-

App

Background Questionnaire

Date:_________________

Name:___________________

ID

1
(1
(4
(8
(1
am
e
(include the WarCraft I and II)

3. Which race(s) do you play with most in WarCraft III ? (check all that apply)
1. Human

3. U dead
4. Night Elf

4
races:
1. Hum

5
1. How long have you played WarCraft III “Frozen Throne” version?
1. Never
2. Less than 1month
3. 1-3 months
4. 4 6 months
r 5. 6 months – 1 yea
6. More than 1 year

133
Appendix B (Continued)

2. How many hours do you play WarCraft III “Frozen Throne” version alone on the

computer in one week on the average?
1. Less than 7 hour
2. 8 – 14 hours
3.15 – 21 hours
4. 22 – 28 hours
5. 29 – 35 hours
6. More than 35 hours

3. How many hours do you play WarCraft III on-line with people in one week on the
average?
1. Less than 7 hours
2. 8 -14 hours
3. 15 – 21 hours
4. 22 - 28 hours
5. 29 – 35 hours
6. More than 35 hours

6. Please answer the following questions based on your playing experiences with the
“HUMAN” race:
1. How long have you played WarCraft III “Frozen Throne” version?
1. Never
2. Less than 1month
3. 1-3 months
4. 4-6 months
5. 6 months – 1 year
6. More than 1 year

2. How many hours do you play WarCraft III “Frozen Throne” version alone on the
computer in one week on the average?
1. Less than 7 hour
2. 8 – 14 hours
3.15 – 21 hours
4. 22 – 28 hours
5. 29 – 35 hours
6. More than 35 hours

134

Appendix B (Continued)

3. How many hours do you play WarCraft III on-line with people in one week on the
1. Less than 7 hours
3. 15 – 21 hours
5. 29 – 35 hours
average?

2. 8 -14 hours

4. 22 - 28 hours

6. More than 35 hours
135

Appendix C
nfidence Level
e settings are described below that may make your game playing time
timate how certain you are that you can

Game Co

1. A number of gam
different. Please es win the game within the given
elow when your race setting is “Human” time-limit options b .
en time-limits, I am confident that I can win the game:

ess than 10 min. (2) 10 – 15 min. (3) 16 – 20 min. (4) 21 – 30 min.
6) More than 35 min.
ng of task level is “MODERATE”, I can win in:
10 min. (2) 10 – 15 min. (3) 16 – 20 min. (4) 21 – 30 min.
in. (6) More than 35 min.
3. when task level is setting “HIGH” while your race hp (health point) is 50% and
min. (2) 10 – 15 min. (3) 16 – 20 min. (4) 21 – 30 min.

Within the following giv
1. when the setting of task level is “EASY”, I can win in:
(1) L
(5) 31 – 35 min. (

2. when the setti
(1) Less than
(5) 31 – 35 m

computer’s is 100%, I can win in
(1) Less than 10
(5) 31 – 35 min. (6) More than 35 min.

136

. When the race setting is “Human”:

Appendix C (Continued)
2
rate your degree of confidence that you can win by recording in each of the
ber from 10 to 100 using the scale given below.
30 40 50 60 70 80 90 100
Moderately Totally certain
certain can win can win
Confidence
(0 – 100)

1. I can win the game when my race setting is “Human” ______

Please
blank spaces a num

0 10 20
Cannot
win at all

137
Appendix C (Continued)

2. When the race setting is “H

uman”:

0 10 20 30 40 50 60 70 80 90 100
beat at all certain can beat can beat

(0 – 100)
When playing alone, what is your confidence that you will be able to beat the
co
1
2 ______
3. Buildings produced ______
______
6
7 ______
8. Experience gained ______
9. Gold mined ______
10. Lumber harvested ______
11. Technology percentage ______
12. Gold lost to upkeep ______
13. Unit score ______
14. Heroes score ______
15. Resource score ______
16. Total score ______
Please rate your degree of confidence that you can win by recording in each of the
blank spaces a number from 10 to 100 using the scale given below.

Cannot Moderately Totally certain
Confidence

mputer on the following items?
. Units produced ______
. Unites killed
4. Buildings razed ______
5. Largest army
. Heroes killed ______
. Items obtained

138

Appendix D

Mental Effort Measure (Easy)

Ple no right answers.
lease circle the answer that best fits how you feel.
. How much did you concentrate while playing the game?
1. Not at all
2. Not much
3. A lot
4. Extremely

. How much mental effort did you invest in order to win the game?
1 2 3 4 5 6 7
o Moderate Extreme
ffort Effort Effort
ase answer the following questions as truthfully as possible, there are
P

1. How much effort, or work, did you put into trying to play the game?
1. No effort
2. Not much effort
3. A lot of effort
4. A huge amount of effort

2
3

N
E

139
Appendix D (Continued)

Mental Effort Measure (Moderate)

Ple no right answers.
lease circle the answer that best fits how you feel.
1. No effort
2. H
3. H
1
No
Ef

ase answer the following questions as truthfully as possible, there are
P

1. How much effort, or work, did you put into trying to play the game?
2. Not much effort
3. A lot of effort
4. A huge amount of effort

ow much did you concentrate while playing the game?
1. Not at all
2. Not much
3. A lot
4. Extremely

ow much mental effort did you invest in order to win the game?
2 3 4 5 6 7
Moderate Extreme
fort Effort Effort

140

Appendix D (Continued)
Mental Effort Measure (High)
s possible, there are no right answers.
lease circle the answer that best fits how you feel.

1. Ho m rk, did you put into trying to play the game?
3. A lot of effort

2. Ho m u concentrate while playing the game?
t all
. A lot

o Moderate Extreme
ffort Effort Effort

Please answer the following questions as truthfully a
P

w uch effort, or wo
1. No effort
2. Not much effort
4. A huge amount of effort
w uch did yo
1. Not a
2. Not much
3
4. Extremely

3. How much mental effort did you invest in order to win the game?
1 2 3 4 5 6 7
N
E

141

Appendix E
7
ot at all Average Extremely
asy Easy
2. ?
1 4 5 6 7
Not Average Extremely
Diffi Difficult
1 3 4 5 6 7
Not Average Extremely
Easy Easy

5. How easy was the “HIGH” level task?
1 2 3 4 5 6 7
Not at all Average Extremely
Easy Easy

6. How difficult was the “HIGH” level task?
1 2 3 4 5 6 7
Not at all Average Extremely
Difficult Difficult

Game Difficulty Level

1. How easy was the “EASY” level task?
1 2 3 4 5 6
N
E

How difficult was the “EASY” level task
2 3
at all
cult

4. How easy was the “MODERATE” level task?
2
at all

4. How difficult was the “MODERATE” level task?
1 2 3 4 5 6 7
Not at all Average Extremely
Difficult Difficult

Abstract (if available)

Abstract This study examined the impact of different levels of task difficulty and expertise on self-efficacy judgments. In addition, the study examines how self-efficacy judgments affect the amount of mental effort investment and task performance under different levels of task difficulty and expertise. Results from this study are used to build a performance model that helps illustrate the relationship among these variables.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Employing cognitive task analysis supported instruction to increase medical student and surgical resident performance and self-efficacy

PDF

The relationship of students' self-regulation and self-efficacy in an online learning environment

PDF

The relationship between state self-efficacy and state worry and sales performance

PDF

The relationship between motivational factors and engagement in an urban high school setting

PDF

The relationship between students’ computer self-efficacy, self-regulation, and engagement in distance learning

PDF

The effect of reading self-efficacy, expectancy-value, and metacognitive self-regulation on the achievement and persistence of community college students enrolled in basic skills reading courses

PDF

What is the relationship between self-efficacy of community college mathematics faculty and effective instructional practice?

PDF

The relationship between parenting styles, career decision self-efficacy, and career maturity of Asian American college students

PDF

Influence of feedback, resources and interaction with superiors on work self-efficacy levels and employee engagement in informal learning activities in the workplace

PDF

STEM + design thinking training: investigation of perceived changes in self‐efficacy, pedagogy, and conceptual development at the K-5 level

PDF

The tracking effect: tracking and the impact on self-efficacy in middle school students

PDF

What are the relationships among program delivery, classroom experience, content knowledge, and demographics on pre-service teachers' self-efficacy?

PDF

The relationship of gratitude and subjective well-being to self-efficacy and control of learning beliefs among college students

PDF

Self-determination among adult Chinese English language learners: the relationship among perceived autonomy support, intrinsic motivation, and engagement

PDF

What is the relationship between early childhood teachers' training on the development of their teaching self-efficacy?

PDF

The effect of traditional method of training on learning transfer, motivation, self-efficacy, and performance orientation in comparison to evidence-based training in Brazilian Jiu-Jitsu

PDF

The effect of cognitive task analysis based instruction on surgical skills expertise and performance

PDF

Musical self-efficacy of graduate students in South Korea and the United States

PDF

The relationship among gender, race/ethnicity, sense of validation, science identity, science self-efficacy, persistence, and academic performance of biomedical undergraduates

PDF

Self-efficacy beliefs and intentions to persist of Native Hawaiian and non-Hawaiian science, technology, engineering, and mathematics majors

Asset Metadata

Creator Ho, Hsin-Ning (author)

Core Title The relationship between levels of expertise, task difficulty, perceived self-efficacy, and mental effort investment in task performance

School Rossier School of Education

Degree Doctor of Education

Degree Program Education

Publication Date 05/07/2010

Defense Date 03/09/2010

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag expertise,levels of task difficulty,mental effort,OAI-PMH Harvest,self-efficacy,task performance

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Clark, Richard E. (committee chair), Keim, Robert G. (committee member), Rueda, Robert S. (committee member)

Creator Email ho_jessie@yahoo.com,hojessie@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m3050

Unique identifier UC1184411

Identifier etd-Ho-3555 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-340678 (legacy record id),usctheses-m3050 (legacy record id)

Legacy Identifier etd-Ho-3555.pdf

Dmrecord 340678

Document Type Dissertation

Rights Ho, Hsin-Ning

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Repository Name Libraries, University of Southern California

Repository Location Los Angeles, California

Repository Email cisadmin@lib.usc.edu