THESIS

TYIELIBRAR

Illllllllllllllllllllllllllllllllllllllll lllllollllll

31293 0171200

 

 

 

This is to certify that the

dissertation entitled

THE ROLE OF GOAL ORIENTATION IN
A SELF-REGULATION FRAMEWORK

presented by

Sandra Leigh Fisher

has been accepted towards fulfillment
of the requirements for

Ph: D; degreeinisychology

//Mflc/fl
/

Major professor

 

Date fi/flf

MSU is an Affirmative Action/Equal Opportunity Institution 0-12771

 

 

 

LIBRARY

Mlchlgan State

University

 

 

PLACE IN RETURN BOX
to remove this checkout from your record.
TO AVOID FINES return on or before date due.

 

DATE DUE

DATE DUE

DATE DUE

 

_ . ‘\
(1 ~‘

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1/98 mount-mu

 

 

THE ROLE OF GOAL ORIENTATION IN A
SELF-REGULATION FRAMEWORK
By

Sandra Leigh Fisher

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Psychology

1 998

ABSTRACT

THE ROLE OF GOAL ORIENTATION IN A
SELF-REGULATION FRAMEWORK

By

Sandra Leigh Fisher

This research was designed to examine the role of goal orientation and interest
within a self-regulation framework. It was hypothesized that goal orientation would
affect individuals” self-set goals, as well as reactions to discrepancies between goals
and actual performance. State measures of goal orientation were developed and used
in conjunction with established measures of trait goal orientation to examine the
stability of goal orientation across performance periods.

A longitudinal study was conducted to examine the role of goal orientation in
self-regulation processes over time in a classroom setting. Three waves of data were
collected using two survey administration modes. One-fourth of the participants
completed paper-and-pencil surveys, while three-fourths completed surveys using
electronic mail. Electronic mail resulted in a slightly higher response rate across three
waves, while response mode did not substantially affect scale means and standard
deviations.

Results suggested that goal orientation did impact several processes in the self-
regulation cycle. Both state and trait mastery orientation positively affected self-set
exam goals, although neither performance orientation did. Some evidence was found
for a moderating relationship of state goal orientation between goal-performance
discrepancies and reactions to discrepancies. particularly for planned increases in effort

and feedback credibility.

Results further suggested that state goal orientations had unique effects on the
goal setting process, with state goal orientation affecting self-regulatory outcomes after
accounting for trait goal orientation. State goal orientations tended to have stronger
effects on goal setting and reactions to discrepancies in Waves 2 and 3 of the study.
Mastery orientation was less stable than performance orientation, likely because of its
close relationship with course specific interest. Course specific interest decreased
when students did not meet exam goals. Decreases in course specific interest were

associated with decreased state mastery orientation in the subsequent wave.

ACKNOWLEDGMENTS

I never expected to write this page while sitting in my house in Virginia, over a
year after leaving Michigan State. it’s been a long road, and i want to thank the people
who helped me reach this point. First, I thank my dissertation chair, Rick DeShon.
Thank you encouraging me to complete my proposal in three months, and for
continually reading proposal drafts to allow me to create the proposal in that time
frame. Thanks for helping me fund this research. And finally, thanks for sticking with
me for the months it took to produce a final dissertation. i hope all your future
dissertation students are a little more timely and easier to reach!

I want to thank my committee members, Kevin Ford, Steve Kozlowski, and Neal
Schmitt. Kevin, thanks for pushing me to improve the state goal orientation scale and
to write a thoughtful conclusion section. Steve, thanks for being supportive and
providing so much positive feedback, especially in the proposal defense. Neal, thanks
for pinch-hitting for Ann in the final defense, and offering some great comments as well.
I wouldn’t expect anything else! I also want to thank all four committee members for
allowing me to develop many of the ideas in this dissertation through some highly
relevant comps questions.

Many people made it possible for me to collect all my data. Rick and Neal
assisted by offering extra credit in their statistics classes. Adam Lush was extremely

helpful with data collection and data entry. Mike Wasserman helped me develop a

rather creative e-mail data collection system, and helped stuff envelopes, mail surveys,
or any other tasks when i needed the help.

While i do appreciate the technical assistance from Mike, I appreciate the moral
support even more. We should have been going through our dissertations at the same
time. I hope i can help and support you now with your dissertation as much as you
helped and supported me. Thanks for always being there for me. lcouldn’t have done

it without you. I’m looking fonlvard to starting the post-dissertation phase of our life!

TABLE OF CONTENTS

LIST OF TABLES ................................................................................................ vi
LIST OF FIGURES ............................................................................................. vii
iNTRODUCTION ................................................................................................. 1
REVIEW OF THE LITERATURE ......................................................................... 9
Self-Regulation ........................................................................................ 9
Self-set Goals ............................................................................ 12
Reactions to Goal-Performance Discrepancies ......................... 14
Change effort ................................................................. 15
Change goal ................................................................... 15
Reject feedback ............................................................. 16
Summary of Self-Regulation ...................................................... 16
Goal Orientation and Self-Regulation .................................................... 18
Goal Orientation and Self-set Goals .......................................... 18
Goal Orientation and Goal-Perfonnance Discrepancies ............ 19
Change Effort ................................................................. 21
Goal Change .................................................................. 21
Reject Feedback ............................................................ 22
Summary of Goal Orientation in Self-Regulation ....................... 23
Operationalization of Goal Orientation .................................................. 24
Trait Theory ............................................................................... 27
Conceptual issues .......................................................... 28
lnteractionist perspective ................................................ 32
Measurement issues ...................................................... 33
Example: State and Trait Anxiety .................................. 34
Summary of State and Trait Characteristics ................... 36
Goal Orientation Revisited ......................................................... 36
Conceptual Issues .......................................................... 36
Goal orientation in the lnteractionist Perspective ........... 39
Measurement Issues ...................................................... 40
Measurement of Goal Orientation in the lnteractionist
Perspective .................................................................... 44
Summary ........................................................................ 45
Antecedents of State Goal Orientation .................................................. 46
Within Episode Antecedents ...................................................... 46
Trait Goal Orientation ..................................................... 46

vi

Interest ........................................................................... 47

Between Episode Antecedents .................................................. 49

Effects of Perfonnance—Goal Discrepancies .................. 49

Effects of Changing Interest ........................................... 50

Summary of State Goal Orientation Antecedents ...................... 51

A Model of Goal Orientation in the Goal Setting Process ...................... 52

Wrthin Episode Linkages ........................................................... 52

Antecedents of state goal orientation ........................................ 52

Antecedents of self-set goals ......................................... 53

Antecedents of discrepancy reactions ............................ 56

Between Episode Linkages ....................................................... 58

Research Setting ............................................................................ 59

METHOD ................................................................................................... 62

Sample Characteristics .......................................................................... 62

Measures ............................................................................................... 63

Procedure .............................................................................................. 67

RESULTS ................................................................................................... 72

Data Analysis Strategy .......................................................................... 72

Factor Analyses ..................................................................................... 72

Correlations ........................................................................................... 77

Hypothesis Testing: Wrthin Wave ........................................................ 81

Hypothesis 1 .............................................................................. 82

Hypothesis 2 .............................................................................. 83

Hypothesis 3 .............................................................................. 84

Hypothesis 4 .............................................................................. 85

Hypothesis 5 .............................................................................. 86

Hypothesis 6 .............................................................................. 87

Hypothesis 7 .............................................................................. 87

Hypothesis Testing: Between Wave ..................................................... 90

Hypothesis 8 .............................................................................. 90

Hypothesis 9 .............................................................................. 90

Hypothesis 10 ............................................................................ 91

DISCUSSION ................................................................................................... 92

Role of Goal Orientation and Interest in Self-Regulation ....................... 92

Effects of goal orientation and interest on goals ........................ 94

Effects on goal orientation on reactions ..................................... 98

State and Trail Goal Orientation .......................................................... 102

Measurement issues ................................................................ 104

Constmct issues ...................................................................... 104

Survey Administration Mode ................................................................ 109

Study Limitations ................................................................................. 109

Implications ......................................................................................... 1 10

Summary ............................................................................................. 1 12
APPENDICES

APPENDIX A: Survey 1 .................................................................................. 114

vii

APPENDIX B: Survey 2 ................................................................................. 118

APPENDIX C: Course-specific Interest Scale ................................................ 121
APPENDIX D: General and Academic Self-efficacy Scales ........................... 122
APPENDIX E: Trait Goal Orientation Scales .................................................. 123
APPENDIX F: State Goal Orientation Scale .................................................. 124
APPENDIX G: Self-set Goal Scales ............................................................... 125
APPENDIX H: Discrepancy Reaction Scales ................................................. 126
APPENDIX I: Tables ..................................................................................... 127
APPENDIX J: Figures .................................................................................... 148
LIST OF REFERENCES ................................................................................. 156

viii

LIST OF TABLES

Table 1 - Factor Loadings for Confirrnatory Factor Model of Trait and

State Goal Orientation: Four-Factor Model: Wave 1 Data .............. 127
Table 2 - Wave 1 Intercorrelations .................................................................. 128
Table 3 - Wave 2 Intercorrelations .................................................................. 129
Table 4 - Wave 3 Intercorrelations .................................................................. 130
Table 5 - Test-retest reliability estimates for goal orientation scales ............... 131
Table 6 - Between-wave effort and exam score correlations .......................... 132

Table 7 - Regression results for Hypothesis 1: State Mastery orientation ...... 133

Table 8 - Regression results for Hypothesis 1: State

Performance orientation .................................................................... 134
Table 9 - Regression Results for Hypothesis 2 ............................................... 135
Table 10 - Regression results for Hypothesis 3: Exam goals ......................... 137
Table 11 - Regression results for Hypothesis 3: Learning goals ................... 139
Table 12 - Regression Results for Hypotheses 4, 5 and 6 .............................. 140
Table 13 - Regression Results for Hypothesis 7: Increase effort ................... 142
Table 14 - Regression Results for Hypothesis 7: Goal change ...................... 143
Table 15 - Regression Results for Hypothesis 7: Feedback Credibility .......... 144
Table 16 - Regression results for Hypothesis 8 .............................................. 145
Table 17 - Regression results for Hypothesis 9 .............................................. 146

Table 18 - Regression results for Hypothesis 10 ............................................ 147

LIST OF FIGURES

Figure 1 - Wrthin episode linkages ................................................................. 148
Figure 2 - Between episode linkages ............................................................ 149

Figure 3 - Summary of interaction between goal orientation and

goal-performance discrepancies .................................................. 150
Figure 4 - Four factor goal orientation measurement model .......................... 151
Figure 5 - Three factor mastery goal orientation and interest model ............... 152

Figure 6 - Interaction between self-efficacy and state performance
orientation ....................................................................................... 153
Figure 7 - Interaction between planned increase in effort and state mastery
orientation ....................................................................................... 154
Figure 8 - Interaction between feedback credibility and state performance

orientation ....................................................................................... 155

INTRODUCTION

Learning has never been as critical to organizations as it is today. Many of the
current trends in organizations depend on the ability and propensity of individuals to
Ieam. Some organizations strive to be Ieaming organizations, defined by Peter Senge
as “organizations where people continually expand their capacity to create the results
they tnriy desire and where people are continually Ieaming how to Ieam together”
(Senge, 1990, p. 3). Other companies are adopting the principles of continuous
Ieaming, where employees actively participate in expanding their skills, and Ieam from
one another, often in the context of autonomous work teams (Rosow 8 Zager, 1988).
Continuous Ieaming is at the heart of Corporate Quality Universities at organizations
such as Motorola, Intel, and Sprint (Meister, 1994).

Other Ieaming-based trends in organizations are directed at individuals. New
career management concepts suggest that, as organizations restructure in an effort to
increase flexibility, individuals need to take responsibility for their own career
development (9.9., Hall, 1987; Mirvis & Hall, 1994). Individuals will change jobs more
frequently, and will be required to Ieam new skills and work habits. To attain career
success, Mirvis and Hall suggest that most individuals will have to find and engage in
Ieaming activities of their own volition. Farr and Ford (1990) contended that individuals
will be increasingly asked to innovate in their work role, introducing new ideas,

processes and procedures to the job (Farr, Hofmann, & Ringenbach, 1993). Each of

2
these trends requires that organizations have the capability to identify individuals who
are predisposed to Ieam and improve themselves.

One concept which has been used to characterize the motivational tendencies
of individuals in Ieaming situations is goal orientation (e.g., Dweck, 1986; Dweck &
Leggett, 1988). Goal orientation has been defined as a general tendency to view
Ieaming opportunities in one of two ways: (a) mastery orientation or (b) performance
orientation. Mastery goal orientation focuses the individual on the material to be
Ieamed, and on increasing his/her competence in that area (Dweck, 1986). Mastery
oriented Ieamers tend to believe that one can increase one’s ability through effort, and
that greater effort leads to greater mastery (Nicholls, 1984). Performance orientation
focuses the individual on attaining positive judgments of his/her ability, and on avoiding
negative judgments of his/her competence (Dweck, 1986). Performance oriented
Ieamers tend to believe that ability and effort are inversely related; if an individual has
to apply a great deal of effort to a task, helshe must have lower abilities (Nicholls,
1984). Thus, the core elements of the goal orientation constnrcts are these beliefs
about effort, and the cognitive focus on judgments of success.

Mastery and performance orientation have been linked to distinct behavioral
patterns in achievement situations. Mastery orientation has been associated with an
adaptive response pattern (Dweck, 1986; Dweck & Leggett, 1988). The adaptive
response pattern promotes “the establishment, maintenance, and attainment of
personally challenging and personally valued achievement goals” (Dweck, 1986, p.
1040). The adaptive pattern is traditionally characterized by selecting challenging
tasks, and persistence in the face of obstacles and failure. Alternatively, performance

orientation has been linked with a maladaptive response pattern, defined as "a failure

3
to establish reasonable, valued goals, to maintain effective striving toward those goals,
or, ultimately, to attain valued goals that are potentially within one’s reach” (Dweck,
1986, p. 1040). This pattern is characterized by selecting less challenging tasks, and
responding to obstacles and failure with low persistence, and negative affect.

The linkage between the cognitive and affective elements of goal orientation
have been so frequently associated with these response patterns that, quite often, the
behavioral patterns have been treated as part of the definition of goal orientation. As
highlighted by Fisher, Delbridge and DeShon (1997), it is crucial in the study of goal
orientation to retain the distinction between the construct itself (i.e., the cognitive and
affective elements), and the outcomes associated with the construct (i.e., task choice,
persistence).

Researchers in a number of different fields, including developmental,
instructional, and industrial/organizational psychology, have investigated relationships
between goal orientation and the behavioral outcomes related to Ieaming. This
research has shown promising results, with goal orientation serving a significant
explanatory function in processes such as task choice, reactions to failure, cognitive
effort and engagement, goal setting, and feedback seeking. Research has found that
mastery oriented Ieamers had higher self-efficacy in complex training situations, which
led to greater generalization of Ieamed skills (Kozlowski, Gully, Smith, Nason, & Brown,
1995). Mastery oriented Ieamers are more likely to persist in the face of difficulty,
developing more sophisticated problem solving strategies, while performance oriented
learners tend to withdraw from tasks, and experience negative affect (Elliott & Dweck,
1988). Mastery oriented trainees are likely to engage in more metacognitive activities,

such as self-monitoring, diagnosis, and evaluation of their behavior (Smith, Ford,

4
Weissbein, & Gully, 1995) and put forth greater effort on Ieaming tasks (Fisher & Ford,
in press). Mastery oriented Ieamers tend to select more challenging tasks that offer
greater Ieaming opportunities, while performance oriented Ieamers tend to select tasks
that allow them to avoid public failure (Elliott & Dweck, 1988). VandeWaIle and
Cummings (1997) demonstrated that individuals with a high mastery orientation were
more likely to seek feedback on academic performance.

Theorists have suggested ways in which goal orientation may affect a variety of
other situations in organizations. Goal orientation may affect the way in which
individuals respond to feedback (Kluger & DiNisi, 1996), with mastery orientation
associated with a task oriented response, and performance orientation associated with
a self-oriented response on the part of the learner. Farr, et al. (1993) suggested that
mastery oriented Ieamers may be more motivated to attend training, may engage in
greater on-the-job Ieaming, and self-initiate more development experiences. In
addition, Farr, et al. (1993) described situations in organizations in which high
performance goal orientation would be useful, such as military operations or running a
nuclear power plant. In situations such as these, a focus on flawless performance is
cnrcial.

The concept of goal orientation, with its relationships to many important
achievement related behaviors such as task selection and reactions to feedback, has
the potential to complement a broader theory of motivated behavior in organizations:
self-regulation. Self-regulation theories, including goal-setting (Locke 8 Latham, 1990),
control theory (Carver 8. Scheier, 1982), and the integrated information
processing/resource allocation theory (Kanfer & Ackerrnan, 1989), share a focus on the

cognitive processes related to how individuals allocate time and effort to a range of

5
activities in attempts to reach various outcomes. These theories differ in the exact
mechanisms and individual difference variables purported to function in self-regulation.
Goal orientation is one of the individual difference variables recently suggested to
impact the self-regulation process (Farr, et al., 1993, Kanfer, 1990).

Self-regulation theories specify several key events in motivated, achievement
related situations that could be associated with goal orientation. For example, a critical
situation in the self-regulatory process is how an individual reacts to a discrepancy
between a standard and a behavior. How the individual reacts to the discrepancy is a
determining factor in future cognitive processes, and hence, future Ieaming and
performance. Farr, et al., (1993) suggested that goal orientation may affect individuals’
reactions to discrepancies, thus playing an important role in self-regulation. There are
other potential roles for goal orientation in the self-regulation process, such as an
antecedent of the choice of goals or standards.

Goal orientation may also be M by the self-regulation process as an
individual proceeds through the cycles of standard setting, performance, and reactions.
In fact, the self-regulatory focus on examining behavior across a series of events allows
the examination of a contentious aspect of goal orientation: its temporal stability.
There has been considerable variability in the literature concerning the stability of goal
orientation both across situations and over time. Some researchers have treated goal
orientation as a trait variable, representing stable individual differences. Other
researchers have used goal orientation as a state variable, representing situation to
situation fluctuations. A third option is to use goal orientation in both ways, as a kind of

‘hybrid' variable (Farr, et al., 1993, p. 199). More specifically, goal orientation may be a

6
stable individual difference that is susceptible to change under certain situational
conditions.

Thus, by investigating the role of goal orientation in the self-regulation process,
this research seeks to better understand the construct of goal orientation, and how the
construct functions over time. First, this research will examine how goal orientation
affects several stages of the self-regulatory process. It is expected that goal orientation
will impact both how individuals set standards for their behavior, and how they react to
differences between self-set standards and the actual behavior. It is well known that
individuals with high self-efficacy are likely to select more difficult goals for themselves
(Locke 8. Latham, 1990). Goal setting can be conceptualized as a process through
which one chooses the difficulty of one’s task. As goal orientation has been related to
choices regarding task difficulty, it follows that goal orientation should affect the
difficulty of self-set goals. Additionally, Locke and Latham (1990) and Kluger and
DiNisi (1996) have suggested that personal characteristics such as self-efficacy will
affect how individuals react to discrepancies. It is likely that goal orientation will also
affect reactions.

Second, this research will investigate the parameters under which an
individual’s goal orientation is likely to change as a result of individual cognitive
interpretations of self-regulatory events, and the effect of these changes on future
achievement. It is suggested that the consistency of the state, or variable, aspect of
goal orientation will be affected by multiple cognitive processes. First, changing
interest in the Ieaming situation will likely affect state goal orientation. As Ieamers
become more or less interested in a topic, their general motivational approach to

Ieaming about that topic should change. Further, discrepancies between self-set goals

7
and actual performance are expected to affect state goal orientation. These
discrepancies will likely focus learners” attention on different aspects of the situation,
which should change their general motivational approach to Ieaming in that situation.

It is important to investigate the consistency of goal orientation, as an empirical
examination of the consistency of goal orientation will help to better define the
construct itself. This research will shed light on whether goal orientation is
appropriately used as an individual difference variable, a characteristic easily affected
by situational factors, or both. Both state and trait goal orientation will be examined to
determine if a state conception of goal orientation contributes to the understanding of
self-regulatory behavior above and beyond the contributions of trait goal orientation.
One research implication of this investigation may be a better understanding of when
and how often goal orientation should be measured. If it is a stable trait, relatively
unaffected by changes in the Ieaming situation, goal orientation need only be
measured once to understand and predict and individual’s behavior. However, if it is
indeed susceptible to change as a result of situational influences, researchers would
have to select appropriate times for measurement.

The literature review that follows will begin by reviewing models of self-
regulation and goal setting, with a focus on reactions to goal discrepancies. Next, the
usage of goal orientation in developmental, educational, and Industrial/Organizational
psychology over the past several years is explored. The attributes of states and traits
are reviewed from the perspective of personality psychology, and the feasibility of goal
orientation as a member of either category is explored. The role of interest in the self-
regulatory model is explored. Finally, hypotheses are proposed for how goal

orientation contributes to the traditional, discrepancy reduction model of self-regulation,

8
and how goal orientation may change as a result of reactions to discrepancies and

changing interest in the Ieaming event.

REVIEW OF THE LITERATURE

Self-Regulation

Self-regulation is a fundamental process in Ieaming and control of behavior. As
described by Karoly (1993), self-regulation consists of the processes “that enable an
individual to guide his/her goal—directed activities over time and across changing
circumstances” (p. 25). While some basic types of Ieaming (e.g., operant conditioning
or social Ieaming) require little in the way of goal directed behavior, more complex
types of Ieaming such as skill acquisition are certainly goal directed, motivated activities
involving attention, practice, and allocation of effort (Kanfer, 1990). Thus,
understanding self-regulation is critical to understanding how individuals Ieam.

Specific self-regulatory theories have focused on different aspects of the self-
regulation process. Control theory focuses on general discrepancy reduction (Carver &
Scheier, 1982), while goal setting theory focuses on the motivational draw of goal
attainment (Kluger & DiNisi, 1996). Alternatively, Kanfer and Ackerman’s (1989)
Integrative Resource Model focuses on the attentional requirements of self-regulation,
and the importance of appropriately allocating attention during skill acquisition.
Regardless of the specific focus of each theory, the central feature of all self-regulatory
models is the negative feedback loop, or discrepancy reduction process (Carver 8.
Scheier, 1982).

There are three key steps in this discrepancy reduction process. First,
individuals must be aware of their own behavior. This awareness primarily occurs
through self-monitoring (Karoly, 1993), as individuals determine what they are doing

and how well they are doing it. Self-awareness at the input stage does not involve an

1 0
in-depth examination of one’s own behavior, rather a temporary shifting of attention to
a particular aspect of behavior (Carver & Scheier, 1982). Behavioral awareness can
also occur through external input, as with feedback from an instnrctor or supervisor.

Secondly, individuals make comparisons between the current behavior, as
detected through self-monitoring, and some kind of behavioral standard. The purpose
of this comparator function (Carver & Scheier, 1982) is to determine any discrepancy
between the current situation and the desired situation. The comparator serves not
only to identify the existence of a discrepancy, but also identifies the direction of the
discrepancy (i.e., positive or negative).

In the third step of the process, individuals attempt to reduce the sensed
discrepancy between current performance and the standard of comparison. Individuals
may attempt to reduce the discrepancy through cognitive or behavioral reactions
(Keman & Lord, 1994). Discrepancies can be reduced by changing either the behavior
or the standard. Thus, one individual could improve his/her performance so that it
meets or exceeds the original standard, while another could mentally or behaviorally
withdraw from the situation, effectively removing the previously existing standard
(Carver & Scheier, 1990).

The standards against which individuals compare their behaviors are generally
considered to be hierarchically organized. Carver and Scheier (1982) contended that
an individual has many standards of comparison (goals) at any one time. The top of
the hierarchy is principle centered, representing what the person wants to be. The
middle range levels of the hierarchy contain task level goals, such as driving to a
location or studying for a test. Each superordinate level provides a standard for the

next lower level. The lowest levels in the hierarchy represent physical movements

1 1
required to attain the middle level goals. This linkage implies that many standards
(goals) are set without using extensive decision making processes (Lord 8. Levy, 1994).

Goals affect behavior by directing attention and effort to the task, and
maintaining attention and effort over time (Locke & Latham, 1990: Karoly, 1993). Goal
setting research has consistently reported that adoption of difficult, specific goals
results in superior performance by individuals (e.g., Locke & Latham, 1990).
Additionally, when a discrepancy between the current state and the goal is detected,
the goal guides the individual in selecting an appropriate behavior to reduce the
discrepancy.

The concept of the standard, or goal, is critical to self-regulation. One aspect of
standards and goals that has not received much research attention is the process by
which these standards or goals are set. In the bulk of self-regulation and goal setting
research, the standards, or goals, examined have been assigned by an external
source, such as a supervisor or an experimenter. For such goals to be effective,
individuals must be committed to that goal (e.g., Hollenbeck & Klein, 1987). Potential
lack of commitment to an assigned goal is a primary problem with this type of goal.

When goals are set by the individual him/herself, some degree of goal
commitment can be assumed. Wright and Kacmar (1994) suggested that individuals
would not select a goal unless they were somewhat committed to it. However, Locke
and Latham (1990) contend that self-set goals are not consistently related to higher
goal commitment than assigned goals, particularly when experimental subjects are
explicitly asked to set goals when they may not have done so on their own. In addition
to the potential for greater goal commitment, the study of self-set goals has the benefit

of allowing free variation in the goal selected (Karoly, 1993). Perhaps the most

12
desirable result of focusing on self-set goals, however, is the opportunity to examine
the antecedents of goals over time (Carver & Scheier, 1982).
MGM—ls

When considering self-set goals instead of assigned goals, the antecedents of
goal selection must be considered. Campion and Lord (1982) discovered that initial
goals were related to ability and past performance. Goals for subsequent performance
episodes were determined by a combination of these factors and the discrepancy
between the goal and actual performance on the previous episode. Campion and Lord
found that both the size and the direction of the discrepancy affected future goals.
Small discrepancies did not affect the setting of goals to the same degree as did larger
discrepancies.

Mono and Baker (1992a) proposed a model of the determinants and
consequences of self-set goals. They contended that self-efficacy, an individual’s
overall judgment of his or her performance capability on a given task, and performance
valence both affect self-set goals (Locke & Latham, 1990; Gist & Mitchell, 1992). Self-
efficacy judgments are based on such factors as ability, prior performance, attributions,
planned effort, and anticipated situational constraints. Empirically, it has been shown
that individuals with high self-efficacy tend to set or select more difficult goals (Locke &
Latham, 1990).

Self-set goals, in turn, affect performance, discrepancy evaluation, and causal
attributions (Mone & Baker, 1992a). These three latter factors then affect self-efficacy
and self-set goals in the next performance episode. Mono and Baker (1992b) tested
the propositions in that model, examining the self-efficacy, self-set goals, and

attributions of over 400 students across three performance episodes (exams). In each

1 3
of the three performance episodes, self-efficacy was a significant predictor of goal
difficulty, even after controlling for prior performance. Additionally, exam performance
was positively related to future self-efficacy following episodes 1 and 2.

Thomas and Mathieu (1994) tested a model similar to that of Mono and Baker,
examining the effects of causal attributions on the goal setting process across two
performance episodes (exams) in a college classroom setting. They found that self-
efficacy influenced self-set goals, which in turn affected exam performance. Self-
efficacy increased across performance episodes for those students with high stability
attributions.

Individuals are likely to consider higher order goals when setting task specific
goals. Individuals have multiple goals at any one point in time, which are hierarchically
ananged (Carver & Scheier, 1982; Lord & Levy, 1994). The lower goals must be
achieved in order to achieve the higher goal. In turn, the standards for the lower goals
are partly determined by higher level goals. This effect was demonstrated through a
surprising finding in the Campion and Lord (1982) study. Campion and Lord had
predicted that negative feedback would result in a reduction of exam goals. A subset
of the students in the Campion and Lord (1982) study reacted in the opposite manner,
Mg their goal for the next exam. This effect is known as a compensation strategy
(Locke & Latham, 1990), where students raise the proximal (exam) goal in order to
maintain feasibility of attaining the distal (course) goal. Consistent with the arguments
of Farr, et al., (1993), the students who raised their goal for the next exam may have
had high self-efficacy. Students did tend to lower course goals if the test feedback was

extremely negative or repeated. Thus, it appears that distal goals, in addition to

14
individual characteristics, must be accounted for when predicting the difficulty of self-
set goals.

Distal goals do not exert a consistent effect on proximal goals. As the proximal
goal and the distal goal become closer in time, the relationship changes. One effect
noted by Campion and Lord (1982) is that as the distal goal drew nearer, it had less of
an effect on the proximal goal. For most students, as the end of the academic quarter
approached, their course grades (the distal goal) were already determined. Their grade
on the last exam had a negligible effect on their final course grade. The grade goals
set for the last exam reflected this knowledge. Students also received no feedback on
the last exam, because the quarter was over. Because of this compression effect,
Campion and Lord did not even analyze much of the data from the last time period.
_l3_e_actions t_o Goal—Performance Discrepancies

Although goal setting has been shown to positively affect performance, the act
of setting a goal does not guarantee attainment of that goal. Furthermore, individuals
may perform in a manner that exceeds their goals. This discrepancy between the goal
and the performance is an important motivational stimulus in self-regulatory theories. In
control theory, the system continuously acts to keep discrepancies within reasonable
limits (Lord 8. Levy, 1994). Therefore, the actions taken by an individual to reduce the
discrepancy are often of interest.

Reactions to discrepancies can impact not only how an individual feels about
his/her performance, but how helshe approaches future performance opportunities, or
episodes. Performance episodes are defined as “distinguishable periods of time over
which performance accrues and is reviewed” (Mathieu & Button, 1992, p. 1759).

Kluger and DiNisi (1996) outlined several possible reactions to feedback, including

15
reactions both within and across performance episodes. The three types of
discrepancy reactions discussed below are changing the effort put forth, changing the
self—set goal, and discounting the feedback.

Change effort. The first strategy for dealing with a goal-performance
discrepancy is to change the effort dedicated to the task. This is a cross-episode
reaction. in which individuals can apply more or less effort to the task in the next
performance episode than in the previous episode. This is the approach most closely
associated with goal setting research. Locke and Latham (1990) suggested that
individuals with high self-efficacy and high goal commitment will increase their effort
rather than decrease their goal. If the gap is positive (performance is higher than the
goal), the individual may decrease their effort. A similar effect has been posited in
equity theory, where individuals who perceive their inputs are greater than their rewards
will decrease their input (e.g., Adams, 1963).

Change goal. A second strategy for dealing with a goal-performance
discrepancy is to change the goal. This strategy is also a cross-episode reaction, in
which an individual can set a higher or lower goal than in the previous episode. If a
person has failed, or otherwise received negative feedback, he or she may lower the
goal (Karoly, 1993). Lowering goals may help the individual attain the desired state of
discrepancy reduction, but is associated with several negative personal outcomes
including decreased motivation, decreased self-esteem, and lower task enjoyment
(Karoly, 1993). Alternatively, if a person has exceeded a goal, or received positive
feedback, he or she may raise the goal for the next episode. Both of these goal
change effects were demonstrated by Campion and Lord (1982), who found that

students who met or exceeded their goals tended to raise their goals for the next exam,

16
while students who failed to meet their goals tended to lower their goals. Students who
failed to meet their goals repeatedly were especially likely to lower their goals.

Individual characteristics and perceptions also affect goal change. Individuals
are less likely to change the goal in either direction if they are strongly committed to the
goal (Locke & Latham, 1990). Lowering the goal is less likely for individuals who have
very high self-efficacy (Locke & Latham, 1990), or who have attributed the negative
discrepancy to an external cause (Thomas & Mathieu, 1994; Mona & Baker, 1992).

Reject feedback. A third strategy for dealing with a discrepancy, particularly a
negative discrepancy, is to simply reject the feedback. This is a within episode
reaction, but has indirect impacts on future performance episodes, as it typically leads
to maintenance of the current goals, effort, and strategies (Locke & Latham, 1990). If
feedback is not accepted, it cannot stimulate changes. One rationale for rejecting the
feedback is lack of credibility of the feedback source (Ilgen, Fisher & Taylor, 1979). If
an individual perceives that the source of the feedback (i.e., teacher, supervisor) is not
honest, reliable, or otherwise tnrstworthy, the individual could attribute negative
feedback to the incompetence of the source, rather than to his/her poor performance.
Feedback is also likely to be rejected when the recipient does not feel the feedback is
an accurate reflection of his or her performance (Ilgen, et al., 1979).

Summary of SeiIf-Regula_ti_gp_

Self-regulatory theories of behavior all center around the negative feedback
loop, or the discrepancy reduction process. Individuals compare their performance with
their desired standards of performance, and make judgments concerning how to deal
with any discrepancy between the standard and actual performance. There are several

different ways in which individuals can respond to goal—perfonnance discrepancies,

17
including changing effort, changing goals, and discounting feedback. The choices
individuals make regarding their responses to discrepancies affect behaviors, such as
future goals and persistence in difficult situations, that impact Ieaming and
performance. Thus, it is critical to the study of motivated behavior in organizations to
understand why individuals make the choices they make.

Self-regulation theories, particularly control theory, have traditionally viewed
self-regulation as a somewhat automatic, universal process. These theories contend
that all people generally proceed through the same steps of the process in the same
way. Variations in the self-regulation process, such as differences in standards or
reactions to discrepancies, are viewed as products of experience (Lord, 8 Levy, 1994),
or features of the situation, such as the size of the discrepancy (Keman 8 Lord, 1990).
Consequently, few individual difference variables have been empirically studied in the
context of self-regulation. Self-efficacy is one notable exception, as research has
consistently demonstrated strong, positive relationships between self-efficacy and goal
setting (e.g., Locke 8 Latham, 1990). More recently, theorists have begun to look
beyond self-efficacy and explore the effects of other individual difference constructs on
the self-regulation process (Kluger 8 DiNisi, 1996).

Goal orientation is an individual difference variable that potentially plays an
important role in self-regulatory processes. Previous research investigating goal
orientation suggests that this constnrct is related to task choice, allocation of effort, and
continuation of effort in difficult situations; all events related to self-regulation. More
specifically, goal orientation may affect self-set goals within performance episodes, and
may be affected by reactions to goal related feedback between performance episodes.

These ideas are explored further in the next section.

18

_G__oanOrimtgtion grd Self-Regu_la_tm

One important individual difference variable that has not been fully investigated
in a self-regulation context is goal orientation. Goal orientation is integrally related to
the cognitive processes in self-regulation, such as goal setting, that occur over time. In
addition to goal setting, it is likely that goal orientation will affect how individuals react to
discrepancies between their goals and their performance. In this section, potential
roles for goal orientation in self-regulation are explored.

Gail orient_a__tion a_nd Sel_f-set Goafi

Campion and Lord (1982) suggested that individual difference characteristics,
such as need for achievement and locus of control, might affect self-set goals. Another
characteristic which will likely affect self-set goals is goal orientation. Dweck (1989)
suggested that individuals with a high mastery orientation tend to follow an adaptive
response pattern, selecting challenging but achievable tasks. Those with a high
performance orientation tend to follow a maladaptive response pattern, selecting tasks
on which they are certain to avoid failure. This proposition suggests that a mastery
orientation leads to more difficult self-set goals, while performance orientation leads to
the selection of easier goals.

Farr, et al., (1993) also hypothesized relationships between goal orientation and
self-set goals. They suggested that goal orientation is likely to affect both the difficulty
of the goal, and the referent (internal vs. external) of the goal. Individuals with a high
mastery orientation should want to improve over previous performance, while those

with a performance orientation should want to perform better than other people.

1 9

As discussed above, self-efficacy is a strong predictor of the difficulty of self-set
goals. Individuals with high self-efficacy tend to set more difficult goals. This
relationship between self-efficacy and self-set goals may be moderated by goal
orientation. Farr, et al., (1993) suggested that the correlation between mastery and
self-set goals will be positive, regardless of the level of self-efficacy. However, they
also suggested that performance oriented people with high self-efficacy will set difficult
goals, while performance oriented people with low self-efficacy will set easier goals.
Similarly, Dweck (1986) contended that the degree to which the maladaptive response
pattern, or the tendency to select less challenging tasks, and respond to obstacles and
failure with low persistence and negative affect, is associated with performance
orientation depends on confidence. If an individual with a high performance orientation
has very high confidence, he or she may very well display behaviors associated with
the adaptive response pattern, such as selecting more difficult tasks.
_Go_a_l Orientation aid GflI-Perfonngnce Discrepancies

The degree to which individuals choose to increase their effort, or even continue
to put forth effort, in the face of difficulty is integrally related to goal orientation. It has
been demonstrated that mastery oriented individuals will persevere in the face of
failure, and that these individuals attempt to learn from their mistakes. Performance
oriented individuals, on the other hand, tend to quit in the face of failure (Dweck, 1986,
1989). They feel that mistakes are not an acceptable part of the Ieaming process, and
would rather stop working on the task than risk making mistakes.

This effect is explained by Kluger and DiNisi (1996) through the manner in
which feedback is interpreted. Performance oriented individuals seek social rewards

based on their task performance. Thus, they are primed to interpret feedback as

20
personal feedback, regardless of the intended feedback cue. When they receive
feedback, the effort that was previously directed toward task performance is channeled
into the maintenance of self-esteem and impression management tactics. As
suggested by Kluger and DiNisi, this redistribution of effort would lead to diminishing
resources for the original task, and eventual quitting of task performance. Mastery
oriented individuals strive to learn the task, and improve their performance over
previous levels. They are primed to interpret feedback at the task level. When they
receive feedback, they use that information to generate and test hypotheses regarding
task performance. When a suitable hypothesis is found to correct the feedback
situation, learning occurs, and task performance improves.

Individual characteristics lead people to interpret feedback in a certain manner,
either as self or as task feedback (Kluger 8 DiNisi, 1996). High mastery orientation
should lead individuals to select a reaction that allows them to Ieam as much as
possible and increase their competence. High performance orientation should lead
individuals to select a reaction that allows them to protect the ego, and maintain
appearances of competence.

Goal orientation research has typically been concerned with only negative
feedback (where the goal was higher than the actual performance), and the
implications of goal orientation on task continuation (Farr, et al, 1993). A broader view
of reactions to goal-performance discrepancies should include positive discrepancies,
where the performance exceeds the goal. Reactions to goal discrepancies beyond
continuation of effort also need to be examined. Thus, the three different kinds of
reactions discussed above will be examined in relation to goal orientation, considering

both positive and negative discrepancies.

21

Change Effort. Goal orientation has been demonstrated to affect the amount of
effort put forth in Ieaming tasks (Fisher, 1995). People with a high mastery orientation
tend to devote greater effort to Ieaming. Mastery orientation is associated with a belief
in a positive relationship between effort and ability (Dweck 8 Leggett, 1988).
Consequently, a high mastery orientation should lead people to respond to goal-
perforrnance discrepancies by increasing effort. Conversely, individuals with a high
performance orientation will not choose to increase effort. They believe that effort is
inversely related to ability (Dweck, 1986). In addition, since they are likely to interpret
negative feedback as self-directed (rather than as task feedback), they will divert
attentional resources from Ieaming to self maintenance activities (Kluger 8 DiNisi,
1996). High performance oriented people may also choose to decrease the total effort
devoted to the task in order to protect their ego. This process is also known as self-
handicapping, or defensive evaluative avoidance (Dweck 8 Leggett, 1988: Karoly,
1993)

Goal Change. Dweck’s (1986) notion of the adaptive and maladaptive response
patterns provides linkages between goal orientation and goal change. She suggested
that mastery oriented individuals, who tend to adopt the adaptive response pattern, will
strive to maintain and attain personally challenging goals. Mastery oriented individuals
also demonstrate persistence in the face of obstacles and failure. Thus, mastery
oriented individuals will be less likely to change their goals after failure. In fact, Kluger
and DiNisi (1996) suggested that if individuals who have received positive feedback
see the opportunity to attain other self goals, they will raise their performance standard.
A high mastery orientation should lead people to value the attainment of more

challenging goals than they originally set for themselves. Thus, individuals with a high

22
mastery orientation may respond to a positive goal-performance discrepancy by raising
their goals.

Alternatively, individuals with a high performance orientation, who tend to follow
the maladaptive response pattern, often fail to maintain effective striving toward their
goals, and respond to obstacles and failure with low persistence. Thus, highly
performance oriented people would tend to lower their goals when provided with
negative feedback. In addition, highly performance oriented people fail to attain valued
goals that are potentially within their reach. This suggests that even when given
positive feedback that exceeds their goals, highly performance oriented people would
not choose to raise their goals.

Reject Feedback. A highly mastery oriented individual may have this latter
reaction to positive feedback. This type of individual may receive objective feedback
from an external source indicating he or she performed quite well on exam, but still feel
as if he or she has not really Ieamed the material. In this case, discounting the
feedback could lead to additional effort. In the case of a negative goal-performance
discrepancy, individuals with a high mastery orientation are likely to accept the
feedback, and use it as diagnostic information for how they can improve their
performance in the next episode (Kluger 8 DiNisi, 1996). Rejecting feedback can also
serve as an ego maintenance function, as the individual attributes the negative
feedback to a characteristic of the feedback source (Karoly, 1993). It is likely that
highly performance oriented individuals could take this approach to negative feedback,

as they desire to protect the ego.

23
Summary of 603 Orientgtion i_r3 Self-Regulation

Goal orientation potentially plays an important role in the self-regulation
process. First, goal orientation is expected to affect the kinds of goals set by
individuals. A high mastery orientation is typically related to selecting difficult tasks,
while a performance orientation tends to be related to selecting easier tasks (Dweck,
1986). Self-efficacy may moderate the relationship between performance orientation
and goal difficulty, as performance oriented people with high self-efficacy set difficult
goals (Dweck 8 Leggett, 1988; Farr, et al., 1993). Second, goal orientation may affect
how individuals respond to discrepancies between their goals and their performance.

It is also possible that goal orientation could be a_f_f_e_ctgq by the self-regulation
process over time. Other individual difference variables have been demonstrated to
change during the self-regulation cycle. For example, self-efficacy can change
dramatically over multiple performance episodes (e.g., Mitchell, Hopper, Daniels,
George-Falvy, 8 James, 1994). If a researcher measured self-efficacy once at the
beginning of a ten trial Ieaming experiment, and tried to use that measure of self-
efficacy to explain goal setting and performance on the tenth trial of an experiment, he
or she would likely draw misleading conclusions concerning the relationship of self-
efficacy and goal setting.

Thus, research has shown that self-efficacy can, and does, change across
performance episodes as individuals perform, receive feedback, and react to that
feedback. A similar effect may also be present with goal orientation. The discrepancy
between the goal and performance will serve to highlight performance or mastery
goals. For example, failing to reach a grade goal could highlight the importance of

grades in future performance episodes.

24

Unfortunately, there appears to be no published research that has used multiple
measurements of goal orientation to investigate this possibility. In the next section,
literature that has examined the issue of the stability of goal orientation over time and in
response to different situations is reviewed, along with various operationalizations and
measurements of goal orientation. Trait theory is reviewed and integrated with goal
orientation literature to explore issues sunounding the stability of goal orientation. The
guidelines of lnteractionist personality theory are then used to build a conceptual case

for the functionality of the goal orientation concept in self-regulation.

Operationalization of Goa‘l Orientation

 

Goal orientation has been operationalized and measured quite differently by
researchers (Farr, et al., 1993; Dweck, 1989). Many of these differences are related to
the disagreement over the conceptual underpinnings of goal orientation, as well as the
relationship between mastery and performance orientation. Some researchers treat the
two dimensions of goal orientation, mastery and performance, as endpoints on a single
continuum. This perspective has been implied in much of the work by Dweck and
colleagues (e.g., Dweck, 1986; Dweck 8 Leggett, 1988). This treatment stems from the
belief that implicit theory of ability is an important antecedent of goal orientation.
Individuals can either believe that ability is malleable and can be improved, which is
suggested to lead to a mastery orientation, or that ability is fixed, which is suggested to
lead to a performance orientation (Dweck 8 Leggett, 1988).

This continuum approach to goal orientation has led researchers to measure

goal orientation by classifying individuals as either mastery oriented or performance

25
oriented (Elliott 8 Dweck, 1988; Dweck, 1989). Wrthin this approach, goal orientation
has been inferred from measures of attributions (Dweck 8 Leggett, 1988), or from
modified version of other scales, including the Intellectual Achievement Responsibility
Scale, the Motivational Orientation Scales, and the Science Activity Questionnaire
(Meece, Blumenfeld 8 Hoyle, 1988).

Other researchers treat these two dimensions as independent constructs (e.g.,
Button, Mathieu, 8 Zajac, 1996; VandeWalle 8 Cummings, 1997). Researchers in this
latter group contend that individuals can be oriented toward both mastery and
performance. For example, graduate students presenting papers at professional
conferences strive both to improve their presentation skills, and to favorably impress
colleagues in the audience. Empirically, relationships between goal orientation and
other variables tend not to be minor opposites of one another, as one would expect if
mastery and performance orientation were endpoints on a continuum (Button, et al.,
1996). These results suggest that goal orientation can be legitimately treated as two
constructs, rather than end-points on a continuum.

Researchers in this second group have typically measured goal orientation by
measuring the strength of both mastery and performance orientation. Contrary to
Dweck’s original conceptualization of mastery and performance orientations as
endpoints on a continuum, researchers taking this approach to goal orientation have
conceptualized mastery and performance orientations as distinct attributes that co-exist
in all individuals. Button, et al., (1996) and VandeWalle (in press) have each
developed measures for goal orientation, based on different conceptualizations and

dimensions. Much of this research has demonstrated a low (non-significant) correlation

26
between measures of the two concepts (Button, et al., 1996; Fisher, 1995; Smith, et
al., 1995).

With this wide range of manipulations and measurements of goal orientation, it
is not surprising that there is little consensus on the consistency and causal structure of
goal orientation. Dweck and colleagues (e.g., Dweck, 1986; Dweck 8 Leggett, 1988)
have suggested that goal orientation is strongly related to one’s innate theory of ability.
If an individual believes that ability is fixed, he or she is likely to have a performance
orientation. If an individual believes that ability is not fixed, and can be increased
through effort, he or she is likely to have a mastery orientation (Dweck 8 Leggett, 1988;
Dweck, 1989). This argument implies that goal orientation is internally caused, and
therefore consistent over time. However, the predominance of studies in which goal
orientation is manipulated suggests that goal orientation may change over time as a
result of situational influences. Researchers have set up competitive or cooperative
reward structures, introduced an audience or evaluator (Elliott 8 Dweck, 1988), or
presented subjects with game vs. test instructions on a task (Dweck, 1989). Each of
these manipulations is intended to highlight one kind of goal or the other, thus affecting
subjects’ behavior.

The operationalization of goal orientation, as with any psychological construct,
must follow from theory. The inconsistent usage of goal orientation may be the result
of an unclear theory of the construct, as illustrated by studies that have treated goal
orientation as both a trait and a state without careful explication of the theory behind
such usage. To help clarify the issues surrounding the usage of goal orientation,

general trait theory will be reviewed. Following the discussion of general trait theory, a

27
comprehensive review of the goal orientation literature in light of trait theory will be
presented.
Trait Theou

Trait theory has been considered the defining perspective on personality (Buss,
1989). Simply put, trait theory suggests that individuals vary on an unspecified set of
relatively stable and enduring dispositions, or traits (Mischel, 1968). These traits cause
individuals to vary on their patterns of behavior over time and across situations. Traits
are used by laypersons and psychologists alike to predict and explain behavior. Trait
theory suggests that traits can be directly inferred from a person’s behavior. Traits are
often measured with self-report questionnaires which require individuals to report on
their attitudes, beliefs and behavioral patterns. Traits are also measured with
observational questionnaires, which require individuals to report on the behavioral
tendencies of others (Hogan, 1991).

A complementary personality concept is the state. States are temporary
psychological conditions caused by external stimuli or situations. These states than
lead individuals to act in a manner specific to the situation. For example, Spielberger
(1977) defined state anxiety as the specific psychological condition of a person in a
strong situation, which leads to certain behavioral and physiological reactions. Trait
anxiety, on the other hand, was the general predisposition to react in an anxious
manner to a variety of situations. States are also measured using self-report and
descriptive questionnaires.

Considerable controversy has surrounded the state-trait paradigm. Critics have
questioned the usefulness of traits, the conceptual definition of traits and states, and

the measurement systems designed to capture states and traits. How do we tell the

28
difference between a state and a trait? Can the same label refer to both? In the next
section, both conceptual and psychometric issues regarding traits and states will be
reviewed.

Conceptu_al Issues. Allen and Potkay (1983) suggested that research on traits
and states has been hampered by the reality that some psychologists prefer to be free
of the constraints of having to distinguish conceptually between states and traits, and
use whichever measure “works best to predict behavior” (Allen 8 Potkay, 1983, p.
1089). Certainly, while this criticism is not true of all psychologists, a firm conceptual
distinction between state and trait is crucial to clearly define the constnrcts of interest,
inform research methodology, and clarify measurement. The conceptual distinction
should include information on the consistency of the construct, as consistency is the
hallmark of a trait. Other distinguishing characteristics of states and traits which should
be discussed include causal origin, temporal duration, and frequency of occurrence
(Chaplin, John 8 Goldberg, 1988). The labels used to represent states and traits must
also be considered.

One aspect of conceptually distinguishing between states and traits is to
consider the causal antecedents of each. Hans Eysenck and Raymond Cattell, early
trait researchers, both believed that traits are at least partially caused by genetics
(Cattell, 1983). Evidence of a hereditary component in traits such as extraversion,
neuroticism, and anxiety supports the trait theory contention of internal, rather than
strictly situational causes of behavior (Eysenck 8 Eysenck, 1980). Both Eysenck and
Cattell also advocated the use of the state, defined as a temporary psychological

condition caused primarily by situational factors.

29

Fridhandler (1986) has also contended that the causes of states and traits
differs. States are caused by extemal events or situations. Traits are internally
caused, and their source is more difficult to trace. States mediate the effects of
situations on behavior, while traits affect behavior through the effects of many complex
causal factors such as genetics and upbringing on behavior (Fridhandler, 1986).

Temporal duration is another distinguishing factor between traits and states.
Traits are stable over moderate periods of time, but could change over the course of a
person’s lifetime. Cattell contended that traits should not only be stable within a person,
but the structure of traits should be stable across persons and cultures (Cattell, 1983).
Alternatively, states are brief psychological circumstances (e.g., moods and emotions)
which affect behaviors and responses. The visible duration of states and traits differs.
States are continuous in their brief duration. Traits are not displayed continuously,
even though they are still a characteristic of the individual. One can describe a person
as generally happy (trait), even when the situation, such as a funeral, requires a sober
response (Fridhandler, 1986). States are more concrete and visible than traits. It is
fairly easy to spot a person who is in an unhappy state. To determine if an individual
possesses a trait, however, requires multiple observations.

Mischel (1968, 1969) criticized trait theory on the basis of consistency. While
consistency has been demonstrated in cognitive aspects of personality, such as field
dependence, conceptual tempo, and cognitive style (Mischel, 1969), Mischel argued
that behavioral consistency had not been adequately demonstrated in the less
cognitive areas of the personality domain. He cited an average correlation of .30 found
between trait measures and the behaviors displayed by an individual. Mischel then

contended that the appearance of behavioral consistency was the result of cognitive

30
organizing strategies, suggesting that regardless of actual discontinuity in behavior, the
human mind searches for and perceives regularity (Mischel, 1968).

Several personality psychologists have contended that Mischel’s original
definition of behavioral consistency was too nanow (Eysenck 8 Eysenck, 1980;
Hampson, 1988). Instead of trying to predict the occurrence of a single, specific
behavior from a trait measure, researchers should look for ”consistency at the
intervening-variable level rather than at the behavioural level” (Eysenck 8 Eysenck,
1980, p. 202). That is, researchers should consider a broader range of behaviors as
confirming evidence of a trait. There are both methodological and conceptual reasons
why consistency should not be expected at the single behavior level. First, a single
behavior is a one item measure, and as such, is not reliable. Correlations are
substantially attenuated by low reliability, and without a reliability estimate, the
correlations cannot be corrected to estimate the true relationship. Secondly, a trait
represents behavioral consistency over time. Observation of a single behavior, or even
two behaviors, is not enough to make a judgment of behavioral consistency. Thus,
behavioral consistency should be examined in the context of the trait underlying a
range of related behaviors.

More recently, Mischel and colleagues have proposed a theory of personality
that allows for broad classification of situational similarity (Mischel 8 Shoda, 1995;
Shoda, Mischel 8 Wright, 1994). The Cognitive-Affective Personality System (CAPS)
recommends examining behavioral consistency across situations based on their
psychological features (peer approval, authority punishment), rather than their nominal
features (e.g., a party, a meeting). The theory then suggests that there are stable

individual differences in how people relate to different psychological features. In CAPS,

31
a state is a pattem of mental activation at a given time that changes readily when
activated situational features change. The theory does not specifically mention traits.
However, ideographic analyses have demonstrated stable profiles of if...then
statements, or “behavioral signatures” (Shoda, et al., 1994, p. 674) that could be
interpreted as traits. These if...then statements are developed through life experiences,
and they affect how social information is processed.

Much of the work under the umbrella of trait theory has taken the lexical
approach to traits. The lexical approach uses person descriptor words culled from a
standard dictionary (Hogan, 1991: Block, 1995). These adjectives are pared down into
a reasonable set, usually using some rational method which is specific to the
researcher. People are then asked to rate themselves and others on these adjectives.
The main assumption behind the lexical approach is that all the words required to
describe personality already exist in the natural language. Because of this foundation,
Allen and Potkay (1981) asserted that the layperson should be able to reliably identify
words that represent states and traits, and that a given descriptor should be used for
eithera state or a trait, not both.

Chaplin, John and Goldberg (1988) suggested that although the distinction
between traits and states is not discrete, it is still meaningful. In their schema-based
framework, individuals hold prototypical images of traits and states, defined by a cluster
of related characteristics. The prototype characteristics are stability, causality, duration,
situational scope, and frequency (Chaplin, et al., 1988). None of the related
characteristics is sufficient or necessary to categorize a particular person descriptor as
either a trait or a state. Person descriptors that are exemplars of neither the trait nor

state schema can be classified as either a state or a trait, while more perfect exemplars

32
are classified more consistently. This framework addresses the labeling concem of
Allen and Potkay (1981), providing an explanation for inconsistent classifications.
Fridhandler (1986) also contended that there is no reason the same label cannot be
applied to corresponding states and traits, and suggested that Allen and Potkay
advanced no reasonable arguments to support the opposite position.

lnteractionist perspective. The guidelines presented above focus on the
differences between traits and states, and how to distinguish between the two. These
guidelines do not, however, limit the usage of any personality concept to strictly a trait
or a state. As emphasized by Chaplin, et al., (1988), many characteristics are
adequately classified as a trait and a state. This real'rzation is consistent with a general
trend in personality psychology to simultaneously consider traits and states as
important antecedents of behavior.

The discipline of personality psychology has not only recovered, but has
benefited, from Mischel’s criticisms of the trait concept (Kenrick 8 Funder, 1988;
Hogan, 1991). From Mischel’s original emphasis on the situational causes of behavior,
and others’ counterarguments that traits are the cause of behavior, general consensus
has arisen that both the strict trait theory and the strict situational approach treat
behavior as overly simplistic (Olweus, 1977). Studies can be designed that support
either perspective, but these studies demonstrate little other than a firm understanding
of experimental research design (Hampson, 1982; Buss, 1989). In the interactionist
approach, both the individual’s dispositional characteristics and the extemal situations
that individual encounters are essential determinants of behavior (Magnusson 8 Endler,

1977; Murtha, Kanfer 8 Ackerrnan, 1996). Mischel’s CAPS theory is consistent with the

33
interactionist approach, emphasizing the role of both situational features and inter-
individual differences in producing behavior.

Measurement issues. Beyond the conceptual difficulties with states and traits,
inadequate measurement has further muddled the personality scene. The next section
provides a review of issues that are particularly relevant to the measurement of states
and traits, and links the issues to personality theory.

Personality instruments that measure both states and traits have been criticized
severely. Allen and Potkay (1981) strongly criticized the practice of retaining the basic
structure of a trait instrument and changing only the instructions to create a state
instrument. An example of this practice is using a list of person descriptors with one set
of instructions (describe yourself in general) to measure a trait, and using that same list
with a different set of instructions (describe yourself at this moment) to measure a
state. Users of a trait or state instrument should not be able to “simply declare” that
their instrument measures a state or a trail.

The multipurpose use of a single instrument is tied to the practice of using a
measure of state to derive an index of trait, where traits are defined and measured as a
function of states over repeated measurements. There are several issues to consider
with this measurement approach. Allen and Potkay (1981) highlighted the problem of
identifying when a trait exists. It is unknown over how many state measurements a
researcher should aggregate to create a trait measure.

Another concern with aggregating states to measure a trait is the response set
that might result from repeated measurements over a short period of time (Allen 8
Potkay, 1981). If individuals answer the same questionnaire every day for two weeks,

their responses may be biased. Response consistency may be a function of memory

34
instead of true consistency over time. Fridhandler (1986) supported the position that
state measures should not be summed or averaged to created trait measures, since
traits and states are caused by different sources (internal vs. external).

Trait and state scales should demonstrate different patterns of reliability
coefficients (Zuckennan, 1983). Both state and trail scales should have high intemal
consistency reliability. However, state scales should have low test-retest reliability,
unless the situational factors present at the second test administration are highly similar
to those at the first administration. Trait scales should always have high test-retest
reliability. Thus, trait scores should not change as the situation changes, but state
scores should change in response to different situations. Stagner (1977) has cited
reliabilities for trait measures ranging from .7 to .8.

There are two primary ways in which the validity of trait and state measures
should be investigated. First, because a major component of the definition of trait is
consistency, test-retest reliability must be considered a part of construct validity
(Cronbach 8 Meehl, 1955). Trait instmments must display high test-retest coefficients
to demonstrate the consistency required of a trait. Secondly, Zuckennan (1977, 1983)
suggested guidelines for discriminant and convergent validity correlation patterns. Trait
and state tests which measure the same person descriptor (i.e., anxiety), should
correlate to a low degree. Trait tests should demonstrate convergent validity with other
tests of the same trait, and these validity coefficients should be higher than coefficients
relating that trait with other measures of the corresponding state.

Example: State and trait anxiety. Spielberger (1977) discussed research on
anxiety that provided support for the distinguishability of states and traits in the

interactionist approach. He described a study in which subjects were given both state

35
and trait measures of anxiety. The between-subjects manipulation involved an anxiety
inducing stimulus (a balloon exploding in the subject’s face). The state measure taken
after the manipulation displayed changes for those subjects in the high anxiety
condition, and no change for those in the low anxiety condition. The trait measure
showed no change for either group. In addition, those subjects in the high anxiety
condition who had higher trait anxiety showed a greater increase in state anxiety.
Spielberger (1977) concluded that the anxiety trait implies a greater susceptibility to
influences from situations.

The trait is a general construct that predisposes an individual to respond in a
given manner across time and situations. Traits are used by laypersons and
psychologists alike to predict or explain an individual’s behavior. The state is a more
localized, temporary psychological condition which is caused by an external stimulus or
situation. The interactionist perspective holds that an individual who possess a trait
may not display the predicted behavior if the situational cues are strong enough to
create a temporary psychological condition that will override whatever trait that
individual might possess. Further, Mischel’s CAPS theory of personality suggests that
behavior is affected by individual response patterns to psychological features of
situations. These more recent perspectives in personality psychology advise against a
limiting classification system for individual difference variables where constructs are
either states o_r traits. Rather, researchers should consider the psychological features
of situations that may interact with individual differences to affect behavior. Some
personality concepts may be best classified as state or trail, but many concepts have

characteristics of both.

36

Summau of state and trail characteristics. The review of trait theory suggests
several guidelines for the appropriate measurement of states and traits. Attempts to
measure traits as aggregates of state measures overlook the different causal origin,
duration, and consistency of traits and states. According to the interactionist
perspective, trait measures must provide an indication of trait strength. Trait measures
must be subjected to empirical investigation regarding longitudinal consistency. While
this sort of test-retest procedure is typically treated as an indication of reliability, in this
case it is also an indication of validity, as consistency across time is a defining
characteristic of any psychological trait (Cronbach 8 Meehl, 1955). Validity must also
be demonstrated in more traditional ways, such as convergent and discriminant
correlations with known measures. Finally, strength of the situation, in terms of
psychological impact, must be considered.
_Go_a_l Orientation Revisited

As detailed in the previous section, there are many conceptual and
measurement issues involved in the determination of whether a construct can and
should be treated as a trait, a state, or both. In this section, goal orientation will be
reviewed in light of the conceptual and measurement issues highlighted in the review of
trait theory in an effort to determine how goal orientation can best be treated.

Conceptual Issues. Dweck contended that goal orientation is a consequence
of an individual’s implicit theory of ability (e.g., Dweck, 1986, 1989). Individuals who
believe that intelligence is fixed tend to have high performance orientations. Individuals
who believe that intelligence is malleable tend to have high mastery orientations.
Button, et al., (1996) found that theory of ability was positively correlated with mastery

orientation (r = .419), and negatively correlated with performance orientation (r = -.157).

37
High scores on the theory of ability measure represented an incremental theory of
ability and low scores represented an entity theory of ability. Surprisingly, goal
orientation does not appear to be related to actual ability. Button, et al., (1996) found
non-significant or small correlations between goal orientation and college GPA (.182
with mastery; -.004 with performance), and goal orientation and SAT scores (.028 with
mastery; -.113 with performance). Fisher and Ford (in press) found non-significant
correlations between goal orientation and scores on a test of general cognitive ability
(.08 with mastery; .04 with performance).

Hokoda and Fincham (1995) empirically investigated the origins of goal
orientation. They categorized children as either mastery oriented or performance
oriented using the Intellectual Achievement Scale. These children were then observed
interacting with their mothers in achievement situations. Hokoda and Fincham found
that the mothers of mastery oriented children displayed different patterns of behavior
than the mothers of the performance oriented children. While statements regarding the
causal nature of this relationship cannot be made as a result of this study, it does
appear that at the very least, parental behaviors will reinforce children’s goal
orientation.

While there is evidence that goal orientation has origins consistent with a trait
model, many researchers have treated goal orientation as a state, manipulating it with
strong laboratory situations. Kozlowski, et al., (1995) gave subjects explicit instructions
regarding Ieaming or attaining a numerical goal to emphasize one orientation or the
other. Elliott and Dweck (1988) highlighted the value of performance goals by telling
subjects their performance was being filmed. Boyle and Klimoski (1995) used early

failure or success manipulations in a training context to induce performance or mastery

38
goals, respectively. While few of the researchers using goal orientation manipulations
performed manipulation checks to ensure goal orientation was affected, most of the
studies supported the predicted relationships. Thus, it appears that these researchers
have been successful in manipulating subjects’ goal orientation in specific settings.

An assumption of experimental manipulation is that goal orientation can be
altered, and as such, may function as a state. Experimental conditions increase the
behavioral similarity of subjects within a condition. The within group variance is
expected to be minimal, and between group variance is maximized. This experimental
paradigm assumes that subjects’ natural orientation can be overcome, and produce the
experimentally desirable orientation. This paradigm depends on inconsistency of
individual characteristics. However, experimental manipulations tend to produce
relatively weak effects as opposed to real life situations (Sackett 8 Larson, 1990;
Kerlinger, 1986), due to lack of subject motivation, as well as ethical concerns for
experimental subjects. Consequently, experimental manipulations create goal
orientations that may be inconsistent with the subjects’ usual orientation, and are likely
to be of brief duration.

Evidence surrounding the causal antecedents, consistency, and duration of goal
orientation suggests it can be meaningfully treated as either a trait or a state. There
are long term factors related to either genetics or early upbringing that affect an
individual’s trait goal orientation. For instance, the work by Hokoda and F incham (1995)
suggests that children’s goal orientation will remain stable over time, as they are
consistently reinforced by maternal behavior patterns. However, there is also
substantial empirical evidence suggesting that situations can be created that will induce

individuals to behave in a manner not consistent with their trait goal orientation. This

39

evidence indicates that strong situations can create temporary psychological conditions
(motivational approaches) that are not necessarily congruent with an individual’s trait
goal orientation.
_G_9__a;l Orienmmpi the lptegctionist Persm

Since both situational influences and individual differences have been shown to
affect achievement behavior, perhaps goal orientation should be approached from the
interactionist perspective (Dweck 8 Leggett, 1988). Button, et al., (1996) characterized
goal orientation as “a somewhat stable individual difference variable that may be
influenced by situational characteristics” (p. 9). This statement is consistent with the
interactionist perspective that goal orientation may be treated as both a trait constmct
and a state construct, depending on the situation. When the situation offers no cues
concerning goal orientation, an individual’s predisposition to one orientation or the other
will be the primary determinant of behavior. In the context of a strong situation, a
motivational mediator variable, state goal orientation, becomes the direct antecedent of
behavior. When the influential situation is over, behavior is again influenced primarily
by the trait. A strong predisposition will be more resistant to situational cues than will a
weak predisposition (Dweck 8 Leggett, 1988).

Dweck and Leggett (1988) did not recommend use of goal orientation as either
a state or a trait, but suggested that the interactionist perspective might provide the
most useful understanding and prediction of behavior. In fact, Dweck (1989)
distinguished between manipulating actual goal orientation and manipulating the
salience and value of performance goals. The latter manipulation would alter state goal

orientation, without changing the enduring personality characteristic.

40

Boyle and Klimoski’s (1995) work is consistent with the interactionist approach
to goal orientation. They attempted to influence state goal orientation, defined as the
goal orientation adapted by participants in a self-directed training session, by creating
early failure and success conditions. Success on a preliminary task was expected to
induce a state mastery orientation, while failure on that task was expected to induce a
state performance orientation. Trait goal orientation was also expected to influence
state goal orientation. The success manipulation was effective, and correlated with
state mastery orientation (r = .24) and state performance orientation (r = -.20). The
failure manipulation had no effect on state goal orientation. Trait goal orientation was a
significant predictor of state goal orientation.

In order to advance the study of goal orientation, the measurement model must
be consistent with the theory (Magnusson 8 Endler, 1977). In the interactionist
perspective, researchers must clearly consider the strength of the trail, as well as the
strength of situational characteristics. These two factors both serve as antecedents of
the temporary state, which then leads to behavior. The following section provides a
review of current measurement of goal orientation, using the guidelines presented
above. Suggestions will be made for how to measure goal orientation in an
interactionist framework.

Measurement Issues

Instruments. Button, et al., (1996) presented a validation study of two scales
measuring trait goal orientation. The goal orientation scales, as well as measures of
other constructs believed to be associated with goal orientation, were administered to
four independent samples of college students. Button, et al. supported a two factor

structure (mastery goal and performance goal) for the goal orientation data over a one

41
factor solution. Button, et al., (1996) examined the consistency of the structure of goal
orientation across samples. However, a validity study of a personality trait should
include an examination of the consistency of individual responses over time or across
situations (Cronbach 8 Meehl, 1955). Thus, the reported validation study failed to
examine a critical component of the trait concept.

VandeWalle (in press) designed three scales measuring goal orientation
specifically in the work setting; mastery orientation, avoid (performance) orientation,
and prove (performance) orientation. The scales were designed to measure goal
orientation as a domain specific trait. VandeWalle conducted a confirmatory factor
analysis that demonstrated excellent fit with the three scales across four different
samples. A distinguishing feature of both the Button, et. al., (1996) and VandeWalle
scales is the capability to measure the strength of goal orientation. Indeed, Button, et
al. suggested that their measure of goal orientation corrects some early flaws of
measurement, such as one-item measures which could not provide an indication of the
strength of the goal orientation.

Meece, Blumenfeld and Hoyle (1988) measured what appeared to be state goal
orientation over six occasions, but then aggregated the measures to create one index
of goal orientation. The questions were phrased in reference to a particular Ieaming
event. In contrast, the Button, et al. (1996) scales are phrased as general statements.
The painrvise correlations between the measures at the six different times ranged from
.44-.83, and the consistency coefficients ranged from 53-64. Meece, et al., (1988)
concluded that these scores were similar enough to aggregate the goal orientation
scores over time. Thus, although they measured goal orientation in a state fashion,

they analyzed their data using a trait conceptualization of goal orientation. Given the

42
criticisms levied by Allen and Potkay (1981) and Fridhandler (1986) concerning the use
of averaged state measures to create a trait measure, this measurement method
appears inappropriate.

Reliabilgy' . Internal consistency reliability estimates for both trait and state
scales have been reasonably good. The Button, et al., (1996) scales have consistently
demonstrated coefficient alpha estimates ranging from .75-.85. Internal consistency
reliability estimates for all three of VandeWalle’s (in press) scales were over .80. The
state goal orientation scales used by Button, et al. and Boyle and Klimoski (1995)
demonstrated coefficient alpha estimates ranging from .70 to .74 for the mastery scale,
and from .81 to .83 for the performance scale.

VandeWalle (in press) conducted test-retest reliability analyses on his work-
specific trait goal orientation scales. The test-retest coefficients over a three month
time period were .66, .57, and .60, respectively. Schuerger, Zarella and Hotz (1989)
reported an average test-retest coefficient of .73 for 52 studies using test-retest
intervals of less than one year on major personality trait inventories (e.g., the Sixteen
Personality Factor Questionnaire [16-PF], the California Personality Indicator [CPI], and
the Edwards Personal Preference Schedule). Thus, the test-retest coefficients
reported by VandeWalle are lower than the reported average in the personality
literature. However, it is the only known attempt to provide test-retest data for a goal
orientation instrument. The relationship of these scales with more general measures of
goal orientation remains to be examined.

1aIi_d'gy_. A primary concern in the validation of state and trait measures is the
discriminability of two measures that share the same label and tap corresponding

aspects of a variable. Button, et al., (1996) included state measures of goal orientation

43

in one of the four samples. The corresponding state and trail scales displayed strong
positive correlations (mastery r = .506, performance r = .477). Correlations between
the alternate state and trail scales were non-significant (trait mastery and state
performance r = .098; trait performance and state mastery r = -.109). According to the
guidelines proposed by Zuckennan (1977, 1983), the correlations between
corresponding state and trait measures should be low to moderate. However, if there
are no strong situational influences present, as was the case in the Button, et al, study,
higher correlations between the state and trail scales seem appropriate in light of
interactionist theory. The two state scales displayed a low, positive correlation (r = .16).
In addition to the correlations, Button, et al. performed a confirmatory factor analysis on
the state and trait scales, and found a four factor solution, reflecting the conceptual
difference between state and trait scales, fit better than the two factor solution that
collapsed across the state and trait scales.

In their study of the effects of goal orientation in a self-directed training context,
Boyle and Klimoski (1995) measured both the trait and the state goal orientations at
different points in the study, using the same scales used by Button, et al., (1996). Trail
measures were completed at the beginning of the study, while the state measures were
taken upon completion of a self-directed Ieaming task. Results of the Boyle and
Klimoski study indicated that trait and state measures of the corresponding goal
orientation were positively related (mastery r = .35, p < .001; performance r = .40, p <
.001), and the measures were not related to the opposite goal orientation. This result is
consistent with the guidelines above. State mastery orientation was affected by an
early training success/failure manipulation as expected, while state performance

orientation was not affected. Of the four measures, only trait performance orientation

44

was significantly related to the primary outcome of interest; Ieaming as a result of
training (r = -.22, p < .01). This study provides some evidence for the existence of
state goal orientation. Unfortunately, test-retest reliability was not examined.

Measurement of goal orientation i_n the intiactionist perspective. As
demonstrated by Spielberger’s ( 1977) work on state and trait anxiety, the amount any
one individual is affected by a goal manipulation is subject to empirical study. When
manipulating goal orientation, both state and trait must be taken into account. One
cannot manipulate the state, and assume the resulting goal orientation is the same
across study participants. Those who are high on the trait are less likely to be changed
by the situational manipulation. Those who are low on the trait are likely to be more
susceptible to the manipulation. Both trait strength and situational strength affect
behavior. This interaction can create confusion in laboratory studies (Buss, 1989).
Imagine a researcher wants to manipulate aggression in his subjects. Subjects who
are low in the trait of aggression may become mildly aggressive. Subjects who are
naturally high on aggression may become extremely aggressive. Consequently, the
manipulation is not affecting subjects equally. In other situations, a manipulation may
affect only those who are high on a trait or only those who are low on the trait (Buss,
1989). A failure to measure and account for the trait level, as well as to check the
effects of one’s manipulation in the presence of such interactions will add to the enor
term in data analysis (Buss, 1989). This effect supports the explicit use of both states
and traits in the study of behavior, as long as both are adequately measured.

However, manipulation checks are vital in research considering both traits and states.

45
ML!

The above sections have provided a discussion of trait theory, and associated
measurement specifications for state and trait scales. These topics were then related
to goal orientation research. This review has highlighted the importance of further
investigation in the following areas: (1) Can state and trait goal orientation be
empirically distinguished? Measurement of these concepts must follow both goal
orientation theory and state-trait theory to ensure appropriate measures are being
used. Additionally, test re-test analysis must be conducted to investigate the relative
consistency of state and trait goal orientation. (2) If state goal orientation changes
over time and across situations, which situational features cause the changes? As
researchers and practitioners, we need to identify the kind of situations, such as
aspects of the self-regulation cycle and situation specific interest, that interact with trait
goal orientation to produce a state goal orientation that affects self-regulatory
behaviors. Suggestions for the parameters and stability of the trait would be useful. A
complete theory of goal orientation should include relationships such as these. Indeed,
Button, et al. (1996) called for further study of the situational aspects of goal
orientation. (3) Finally, does the measurement of state goal orientation increase our
understanding of goal-directed behavior? Researchers have manipulated state goal
orientation (e.g., Kozlowski, et al., 1995), but the effects of naturally occurring changes
in state goal orientation have not been examined.

In conclusion, whether the two dimensions of goal orientation, mastery
orientation and performance orientation, are appropriately treated as consistent traits,
situationally induced states, or both is an empirical question that has yet to be

answered. Evidence exists to support each position. Goal orientation appears to be a

46
construct that falls in the boundary category between state and trait (Chaplin, et al.,
1988). A more useful area of exploration may come from taking an interactionist
perspective on goal orientation. Instead of forcing a choice between a state or trail
representation of goal orientation, research must investigate the conditions under which
situations lead to certain response patterns, and the degree to which certain individuals
are affected by situational influences (e.g., Dweck 8 Leggett, 1988). The temporal and
cross-situational consistency of these constructs must also be tested more rigorously.
This is an area of research which could contribute significantly to our understanding of
behavior in organizations, as well as in achievement oriented situations in general.

The next section reviews potential antecedents of state goal orientation.
According to the interactionist perspective, both internal traits and extemal situations, or
the cognitive representations of situations, must be considered. The primary internal
trait expected to affect state goal orientation is trait goal orientation. Situational factors
that will be considered are domain specific interest and goal-performance
discrepancies. Two types of antecedents will be considered. First, variables expected
to affect state goal orientation within a single performance episode will be discussed.
Second, variables expected to affect state goal orientation between performance
episodes (e.g., results from time 1 affecting state orientation at time 2) will be
discussed.

Antecedents of St_a_te Goa_I Orientgtjpg
Within Episode Antecedents

Trait goal orientation. The most direct antecedent of state goal orientation is

expected to be trait goal orientation. The interactionist perspective suggests that the

two major determinants of a psychological state are the individual’s predispositions and

47
strong situational factors. Thus, in absence of strong situational factors, trait goal
orientation would be the prime determinant of trait goal orientation.

George (1996) proposed similar relationships among trait and state affect. She
defined trait concepts of positive and negative affectivity as individual dispositions to
experience positive and negative moods. These dispositions were suggested as one of
two determinants of state affectivity, defined as fluctuating and changing emotions and
moods. Individuals high on positive affectivity are more likely to experience positive
moods, and individuals high on negative affectivity are more likely to experience
negative moods (George, 1996).

Boyle and Klimoski (1995) supported this form of relationship for goal
orientation. They found that trait goal orientation was a significant predictor of state
goal orientation. Trait mastery orientation was a significant predictor of state mastery
orientation, and trait performance orientation was a significant predictor of state
performance orientation.

um Relationships between interest and Ieaming goal orientation have
been suggested by several researchers. Interest is defined as “a relatively enduring
preference for certain topics, subject areas, or activities (Schiefele, 1991, p. 302).
Interest contains both a feeling related component (enjoyment, positive affect) and a
value-related component (personal significance). Schiefele suggests that interest is a
content-specific concept, and that interest is not a personality characteristic. Interest is
a directive force which can explain students’ choice of area in which they exhibit
intrinsic motivation. Schiefele’s treatment of interest is very similar to Pintrich, Smith,

Garcia, and McKeachie’s (1991, 1993) notion of task value. Task value is one concept

48

representing the value component of motivation, and measures judgments of how
interesting, useful, and important the content of a specific course is to students.

Empirical studies have consistently revealed a relationship between interest and
goal orientation. However, there is no consistency as to the direction of the
relationship. Whether goal orientation appears to affect interest or interest appears to
affect goal orientation depends on the perspective of the researcher. For example,
Butler (1987) suggested that the goal orientation adopted by students in a given
situation would affect their interest in the task. She manipulated goal orientation by
providing different kinds of feedback. One group received individualized comments
regarding their performance. A second group received grades based on a normal
distribution of performance. The comments were intended to highlight task aspects of
the situation, thus inducing a mastery orientation. The grades were intended to
highlight comparative aspects of the situation, thus inducing a performance orientation.
Interest was measured with a self-report questionnaire completed after three sessions
of task performance. Butler found that students in the comments condition reported
greater interest and enjoyment in the task than the students in the grades condition.

Boyle and Klimoski (1995) suggested that both trait and state mastery
orientation would lead to greater interest in the Ieaming task. They found that task
interest was related to state mastery orientation (r = .50), but not to trait mastery
orientation (r = .15). Interest was not significantly correlated with either trait or state
performance orientation. In addition, they reported several two-way interactions among
the trait and state goal orientations that significantly predicted task interest.

Alternatively, Schiefele (1991) suggested that interest in a specific topic partially

determines the strength and nature of the motivational orientation adopted in a specific

49
Ieaming situation. An interested person naturally wants to Ieam more about, or
become involved with, the topic in which he or she is interested. Highly interested
students have also been found to work harder on tasks, and process information more
deeply. Topic-specific interest is also related to the use of more complex Ieaming
strategies, such as elaboration and critical thinking (Schiefele, 1991, Pintrich, et al.,
1991, 1993). Schiefele presented a model in which state goal orientation is affected by
both the trait goal orientation and interest in the topic of study. High interest leads a
person to adopt a mastery orientation in that specific situation.

Likewise, Meece, et al., (1988) treated goal orientation as a state, hypothesizing
that goal orientation would be affected by student interest in science. They did find that
student interest in science, measured at the beginning of the course, was positively
related to mastery orientation, and negatively related to performance orientation
(standardized maximum likelihood path estimates .33 and -.15, respectively). Both
mastery (.63) and performance (.17) goal orientation, in turn, positively affected
cognitive engagement during Ieaming.

Between Episode Antecedents

flagpof perfonn_a_nce—goal discrepancies. Between performance episodes,
state goal orientation may change in response to the goal-performance discrepancy. If
the goal-performance discrepancy is negative, the salience of the grading aspect of the
situation will increase. This is the desired consequence of many of the manipulations
designed to alter goal orientation (Dweck 8 Leggett, 1988; Heyman 8 Dweck, 1992).
Wofford and Goodwin (1990) suggested that negative feedback causes people to

break out of habits, and use non-automatic processes. Similarly, negative feedback

50
may cause people to react in ways that are inconsistent with their usual patterns. in
other words, this negative feedback could cause a state change.

This hypothesis is consistent with the interactionist perspective on goal
orientation (Dweck 8 Leggett, 1988). When an individual initially approaches the task,
behavior is predominantly guided by the trait. As the situation changes, and certain
aspects of the situation are highlighted (e.g., grades), the trait loses influence. The
situation has created a stronger state, which then influences further behavior, in this
case, self-set goals. In the subsequent performance episode, self-set goals would then
change in response to changes in state performance orientation. Wrth an increased
state performance orientation, the self-set goal should change. This relationship may
be moderated by the individual’s self-efficacy (Farr, et al., 1993).

Effects of Changing Interest. The effects of unmet goals on state mastery
orientation may operate primarily through interest. As discussed above, interest is
expected to affect state mastery orientation. Interest in a subject area may change as
a result of performance experiences in that area. Schiefele (1991) postulated that
success strengthens interest in an area. Failure may also decrease interest, but the
effect may not be immediate. Repeated failure may be required to lessen an
individual’s interest in an area.

Interest may also change independently of goal-performance discrepancies.
Lane, Newman, and Bull (1988) discovered that interest changes over the duration of a
course. They measured interest four times within a study that examined the
relationship between interest and advance organizers; at the beginning of the course,
after the advance organizer, after instruction, and following a retention test two weeks

after instruction. They found that interest did change across these time periods. The

51
changes were partially due to the advance organizer, which was found to increase
student interest for those who were initially low on interest. Lane, et al., also suggested
that interest might change as a result of increased familiarity with a subject area. They
suggested that students who know very little about an area might be initially interested
in the area, but after Ieaming more about that area, might realize that they are actually
not interested. Thus, if interest changes between performance episodes, as suggested
by Schiefele (1991), than state mastery orientation may very well change between
performance episodes, as a result of changing interest.
Summary of St_a_te Ggoaj Orientgion Ap_teced_e_n_t_s_

There are several influences on state goal orientation, both within and between
performance episodes. First, the trait of goal orientation is an important influence on
the state adopted in any given situation. If the situation is weak, and does not
emphasize mastery or performance goals, the trait will predominate. However, if the
situation does emphasize one type of goal over the other, the temporary psychological
state will be strong enough to influence behaviors. Second, topic specific interest is
suggested to be a situational factor that will highlight the mastery goal. Many
researchers have found a relationship between interest and mastery orientation, but the
direction of that relationship is unclear. Course specific interest is suggested to affect
state mastery orientation. Third, discrepancies between goals and performance are
suggested to affect state goal orientation. Negative discrepancies will increase the
salience of grades, thus increasing state performance orientation. Effects of
discrepancies on state mastery orientation are largely expected to occur through
interest. Negative feedback will decrease interest in the course, resulting in a decrease

in state mastery orientation.

52

A MI of Goa_I Orientation in the Self-Regulation Process

A model describing the role of goal orientation in the self-regulation process is
presented in Figure 1. A second model detailing the relationships between
performance episodes is presented in Figure 2. The hypotheses in the first section,
within episode linkages, have been derived from the model presented in Figure 1. This
first set of hypotheses suggests relationships within a single performance episode.
Consequently, none of the variables themselves (i.e., state goal orientation) are
expected to change in the model displayed in Figure 1. Those between episode
relationships are depicted in Figure 2, and are detailed in the between episode
hypotheses section. Each linkage in the figures is labeled with the corresponding
hypothesis number. Two control variables, ability and general self-efficacy, are
represented in the model. Both ability and general self-efficacy are expected to directly
affect academic self-efficacy and distal goals. Ability and general self-efficacy will be
included in the data analysis as control variables.

Wrthin Episode Linkages

Antecedents of state goal orientation. Figure 1 contains the model depicting the
effects of goal orientation on goal setting and discrepancy reactions. The first pair of
linkages in the model shows the direct relationship between trait goal orientation and
state goal orientation. The primary antecedent of state goal orientation is trait goal
orientation. Boyle and Klimoski (1995) found that mastery trait orientation significantly
predicted mastery state orientation, while performance trait orientation significantly

predicted performance state orientation.

53
Hypothesis 1: Each trait orientation will be positively related to the corresponding
state orientation.

Also affecting the state mastery orientation at this stage of the model is course
specific interest. In the absence of situational influence, trait and state goal orientation
would be identical. Therefore, situational antecedents of state goal orientation must be
considered. According to Mischel’s CAPS theory, the psychological features of a
situation affect states through individual interpretations of social information. Many
aspects of a college course (e.g., content, professor, grading) are likely to provide
social information that students will encode and process. This processing, in concert
with other individual attributes and experiences, would result in interindividual variability
in course specific interest. Thus, interest would function as a psychological
representation of the situational features. Schiefele (1992) has suggested that interest
is likely to affect the goal orientation adopted by individuals. Boyle and Klimoski (1995)
have found that area-specific interest is related to state mastery orientation, but not to
state performance orientation.

Hypothesis 2: Interest in the course will interact with trait mastery orientation to
affect state mastery orientation. Interest will have a greater effect on state
mastery orientation when trait mastery orientation is weak.

Antecedents of SSH-Sm. The next set of linkages in the model portrays the
antecedents of self-set goals. First, state goal orientation directly affects the goals set
for exams. Goal orientation will affect the difficulty of the goal as well as the content of
the goal. Dweck has suggested that goal orientation may affect the both type and
difficulty of the goal chosen. Mastery orientation is reflected in a desire to improve

oneself, and a comparison of current performance with past performance.

54
Performance orientation results in a desire to prove oneself, and to focus on normative
evaluation methods.

Because of these two varieties of goals, type of goal and difficulty of goal, self-
set goals will be operationalized in two ways. The traditional treatment of grades as
self-set goals will be used to measure grade goals (Campion 8 Lord, 1982; Locke 8
Bryan, 1968). However, students may set goals in their courses other than grades.
Mono and Baker (1992a) suggested that measures of students’ subjective Ieaming
goals would provide additional insight into the goal setting process. This kind of
measure logically meshes with the treatment of goal orientation. A mastery orientation
leads individuals to focus on Ieaming, while performance orientation leads individuals to
focus on the normative aspects of the situation. Attempting to elicit self-set goals by
asking only about grades severely restricts goal choice, particularly for those individuals
high on mastery orientation. Therefore, both grade goals and Ieaming goals are
considered in the model. Given the inherently subjective nature of the Ieaming goals
and the self-report Ieaming perceptions, the Ieaming goals are included more for
increased understanding than for rigorous hypothesis testing.

Research has demonstrated that a strong mastery orientation is associated with
a preference for difficult tasks, while a strong performance orientation tends to be
associated with a preference for easier tasks (Dweck, 1986). State mastery orientation
has been linked to the selection of tasks where individuals have the opportunity to Ieam
a great deal (Dweck, 1986). Individuals are more likely to Ieam a great deal when they
set more difficult goals for themselves in courses. A strong mastery orientation will also

be associated with setting a Ieaming goal.

55
Hypothesis 3: State mastery orientation will have a direct, positive effect on the
difficulty of the grade goal set for the upcoming performance episode.

Individuals who have high performance orientation tend to select less difficult
tasks, as they seek to avoid failure and negative evaluations (Dweck, 1986). However,
Farr, et al., (1993) suggested that this relationship will be moderated by the individual’s
self-efficacy. A high performance orientation does not lead to the typical helpless
pattern if the individual has high self-efficacy. Individuals who have low self-efficacy
and a high performance orientation will set lower goals, while those who have high self-
efficacy and a high performance orientation will set higher goals. An individual with
high self-efficacy and a strong performance orientation will tend to set difficult goals,
because he or she enjoys demonstrating to others that he or she can succeed on a
difficult task.

Hypoflresis 4: Self-efficacy and state performance orientation will interact to
affect the difficulty of the goal set for the next performance episode. More
specifically, individuals with a high state performance orientation and high self-
efficacy will set more difficult goals. Individuals with a high state performance
orientation and low self-efficacy will set easier goals.

In addition to goal orientation, distal goals and self-efficacy are expected to
affect self-set goals. Several self-regulation researchers have postulated that goals are
ananged in a hierarchical fashion (Lord 8 Levy, 1994). The goals at the higher levels
provide a standard for the goals at the lower levels. Thus, more distal goals provide a
reference point for the setting of more proximal goals. Distal goals have a direct effect
on the difficulty of self-set goals. While this effect is likely to change over time

(Campion 8 Lord, 1982), the grade a student is striving toward for the entire course will

56
affect the exam grade goal he or she will set within a given performance episode. In a
classroom setting, Campion and Lord (1982) found that exam goals were set in
reference to the goal students had set for the entire course.
Hypothesis 5: Distal goals (goals for the entire course) will have a positive, direct
effect on self-set proximal goals (exam goals).

Self-efficacy also has a direct effect on self-set goals. It is well known that high
self-efficacy is related to individuals setting more difficult goals for themselves (Locke 8
Latham, 1990). This effect has been supported in research by Mone and Baker
(1992b) and Thomas and Mathieu (1994), who found that self-efficacy was related to
goal difficulty in each of the performance episodes in their studies.

Hypothesis 6: Self-efficacy will have a direct, positive effect on the difficulty of the
goal set for the upcoming performance episode.

Antecedents of cflscregapcy reactions. Three possible reactions to
discrepancies are represented in the model; changing the goal, changing effort, and
discounting feedback. The discrepancy between the self-set goal and actual
performance affects the reaction to the discrepancy. For example, negative feedback
tends to be accepted less readily than does positive feedback (Ilgen, et al., 1979).

The reaction an individual has to a discrepancy between his/her goal and
performance is also dependent on individual differences including goal orientation
(Kluger 8 DiNisi, 1996). A strong mastery orientation will lead individuals to react in a
manner which will assist further Ieaming, such as increased goals, increased effort, and
accepting negative feedback. A strong performance orientation will lead people react
in a manner which will assist in ego protection, such as decreasing goals, decreasing

effort, and rejecting negative feedback.

57
Goal orientation theory and research suggests that mastery and performance

orientation affect the amount of effort an individual is willing to devote to a task (Dweck,
1986; Fisher, 1995). Goal orientation also affects the type of task individuals will
choose, where high mastery orientation is associated with more difficult tasks, and
performance orientation is associated with simpler, less risky tasks (Dweck 8 Leggett,
1988). This argument can be extended to the selection of future goals. Finally, goal
orientation is also suggested to affect the degree to which individuals accept the
feedback given to them, where performance oriented people reject negative feedback
in order to defend the ego, while mastery oriented people are more likely to reject
positive feedback. (see Figure 3 for a summary of hypothesis 7).
Hypothesis 7: State goal orientation will moderate the relationship between
discrepancies and reactions (change in effort, change in goals, and rejecting
feedback).

H7a: In a positive goal-performance discrepancy situation, a high state

mastery orientation will be associated with maintenance of effort, feedback

acceptance, and an increase in goal difficulty. A high state performance

orientation will be associated with a decrease in effort, feedback acceptance,

and goal maintenance.

H7b: In a negative goal-performance discrepancy situation, a high state

mastery orientafion will be associated with increased effort, feedback

rejection, and maintenance of goal difficulty. A high state performance

orientation will be associated with decreased effort, feedback rejection, and a

decrease in goal difficulty.

58

Between Episode Linkages

Several of the variables in this model are expected to change from one
performance episode to the next. These changes will affect the kinds of goals that are
set in the next performance episode. First, the discrepancy between the goal and
performance will affect several variables in the model. Interest will be directly affected
by the discrepancy. If students meet or exceed their exam goal, they are likely to
become more interested in the class (Schiefele, 1991). Course specific interest is likely
to change over the semester, as students find out more about the subject area (Lane,
et al., 1988). Interest will also change as a result of performance feedback (Schiefele,
1991). Positive feedback tends to increase interest, while negative feedback,
particularly repeated negative feedback, tends to decrease interest.
Hypothesis 8: Performance-goal discrepancies at time t will be directly related to
interest in the course at time t + 1. Specifically, a positive discrepancy (grade
higher than goal) will be associated with an increase in interest. A «win
discrepancy (grade lower than goal) will be associated with a decrease in interest.

There is a similar relationship between performance and self-efficacy. Self-
efficacy judgments are based on a number of attributes, including past perfomlance
(Locke 8 Latham, 1990). Meeting or exceeding goals increases self-efficacy to perform
the task again in the future (Locke 8 Latham, 1990; Mono 8 Baker, 1992b). When
performance is below self-set goals, self-efficacy will decrease. When performance
exceeds self-set goals, students will become more confident in their ability to perform
well.
Hypothesis 9 : Performance-goal discrepancies on exam goals at time t will be

directly related to self-efficacy at time t + 1. Specifically, a positive discrepancy

59
(grade higher than goal) will be associated with an increase in self-efficacy. A
negative discrepancy (grade lower than goal) will be associated with a decrease in
self-efficacy.

Goal-perfonnance discrepancies are also suggested to affect state goal
orientation in the subsequent performance episode. Consistent with the experimental
manipulations designed to affect goal orientation, an event that makes grade goals
especially salient to the student will be positively related to a change in state
performance orientation (e.g., Dweck 8 Leggett, 1988; Heyman 8 Dweck, 1992).
Events such as failure have been shown to increase the salience of grades in a variety
of situations, which in turn increase state performance orientation (Boyle 8 Klimoski,
1995). The interactionist perspective must be addressed in this hypothesis as well.
Situational factors have more influence on a state when the corresponding trait is weak
(Magnusson 8 Endler, 1977).

Hypothesis 10 : Negative discrepancies on grade goals will cause state
performance orientation to increase. Positive discrepancies on grade goals will
cause state performance orientation to decrease. Negative discrepancies will
have a greater effect on state performance orientation when trait orientation is
weak.

Reseich Settipg

Several researchers have suggested that the academic setting may be the ideal
setting for studying goal setting processes over time (Mono 8 Baker, 1992a; Thomas 8
Mathieu, 1994; Campion 8 Lord, 1982). First, the setting is relatively natural. Students
would be performing the work and obtaining feedback with or without the study.

Second, students’ performance is typically independent of one another. Third, there

60
are multiple performance episodes. Most college courses include a number of exams,
each of which requires Ieaming, performance, and the receipt of feedback. The
constructs in this research will be operationalized as variables in an academic setting.
Thus, the proximal goals will be exam grade goals, and the distal goals will be course
grade goals. Performance will be measured with actual exam grades. The discrepancy
between the grade and the goal will be the simple difference between the two
measures. Goal-performance discrepancy is expected to function as a psychological
feature (e.g., Mischel 8 Shoda, 1995) of the classroom situation that would impact state
goal orientation.

\Mlile the college classroom is a relatively natural setting for conducting goal
setting research, there are concerns about the motivational effects of eliciting goals
from students. Wright, O’Leary-Kelly, Cortina, Klein, and Hollenbeck (1994), presented
a taxonomy of self-set goals. They suggested that there are three main goal
generators, or reasons for setting goals; dispositional traits, request by an authority
figure, or reaction to an assigned goal. These three sources interact to create specific
goal situations. Trait generated goals with no extemal request are called natural goals.
Goals generated by a request when there is no disposition for setting that goal are
called induced goals. Goals generated spontaneously that co-occur with an external
goal request are labeled reinforced goals. The purpose of this taxonomy is to
encourage researchers to think about the different types of self-set goals, and how their
source can affect the motivational consequences of that goal.

It is possible that the self-set goals in this study will be influenced goals, or goals
brought about by the researcher. It is also possible that the self-set goals will be

naturally occurring, and thus will be reinforced goals. Students may very well have

61
standards for their performance in courses, verbalized or not, as evidenced by students
who are displeased by their grades and attempt to have them raised after the fact.
Students may not typically verbalize these standards as goals, but many students do
indeed have naturally occurring standards. This issue of induced goals is particularly
relevant with the relationship between mastery orientation and grade goals. Given the
definition of mastery orientation, where students are motivated by Ieaming rather than
external standards, a high mastery orientation should not be linked with the
spontaneous setting of grade goals. However, when subjects are asked to set grade
goals, a high mastery orientation should be associated with the setting of difficult goals.
This induced goal may or may not affect the student’s behavior, depending on the
valence of that goal. To deal with this issue, goal valence will be measured for grade
goals. If an individual reports a low valence for a grade goal, it is likely that it was an
induced goal, and will not have the same impact on behavior as would a natural or

reinforced goal.

METHOD

S_arpple Characteristics

Participants. Participants in this study were 434 undergraduate students in
various psychology courses at Michigan State University. The students received partial
course credit for their participation. Complete data across three waves of data
collection were available for 354 participants. Of those students who provided
complete data, 68.6% (n = 243) were recnrited through the Department of Psychology
subject pool. These students were primarily first and second year students enrolled in
introductory psychology courses. To obtain a more representative sample of
psychology students, the other third of the sample was drawn from upper level
psychology courses such as statistics and research methods.

MsticaL Power. An important consideration in behavioral research is the
statistical power for detecting various effect sizes (Cohen 8 Cohen, 1983). The sample
size of 354 in this study provides acceptable power. Wrth a sample of 350, the power
to detect an effect size of .20 (p <.01) is .89. A sample of 350 allows for power of
greater than .995 (p <.01) with correlations of .30. Regarding changes in R2 in a
regression analysis with 5 independent variables, each contributing an R2 of .04 to a
total R2 of .40, power of .90 can be reached with a sample size of 250 (Cohen 8
Cohen, 1983). Thus, the final sample size of 354 allowed acceptable levels of power

for hypothesis testing.

62

63
Measures

This section details the measures used in the study, including current reliability
estimates. All scales used in this research were used with a 5 point Likert scale, where
5 represents Strongly Agree, and 1 represents Strongly Disagree. The full surveys are
presented in Appendices A and B.

Int_e[asp Interest in the course was measured with a 4 item scale (see Appendix
C) based on the task value scale of Pintrich, et al., (1991, 1993). The original scale
had six items, and a coefficient alpha of .90. Two items were deleted based on their
low structural equation parameter estimates with the scale as a whole (Pintrich, et al.,
1991). The coefficient alpha reliability estimates for the three waves of this study
ranged from 91-92. A sample item on the interest scale is, “I am very interested in the
content area of this course.”

Self-effiw was measured with two scales; one for general self-efficacy (GSE),
and one for academic self-efficacy (ASE). GSE was measured with an eight item
version of Sherer, et al’s (1979) general self-efficacy scale. A sample item for GSE is “I
rely on myself to accomplish my goals.” GSE was measured only at Wave 1, and the
coefficient alpha reliability estimate was .73. ASE was measured with a 5 item scale
(see Appendix D) based on the self-efficacy scale of Pintrich, et al. (1993). The
coefficient alpha reliability estimates for three waves of this study ranged from .90-.91.
A sample item for the self-efficacy scale is, “I am certain I can master the skills being
taught in this class.”

Trait goal orientation was measured with two eight item scales (see Appendix E)
by Button, et al. (1996). These scales treat mastery and performance orientation as

distinct constructs. These scales have typically resulted in reasonable fit to a two factor

64
structure, with a very small correlation between the scales. Coefficient alpha reliability
in the first wave of data collection was .75 for the mastery scale, and .76 for the
performance scale.

State goal orientation was measured with two six item scales (see Appendix F)
based on Boyle and Klimoski (1995). Similar to the trait goal orientation measures,
these scales treat mastery and performance orientations separately. An exploratory
factor analysis performed by Boyle and Klimoski revealed a lack of simple structure.
Therefore, several items were deleted from the scale. New items were added to make
the scale more directly relevant to an academic situation. Finally, items were reworded
to focus the respondent on the immediate situation. The original scale depended on
the instructions to frame the scale as responses to the situation. The new items were
intended to provide that direction on an item-level basis (Allen 8 Potkay, 1981). The
revised version of the scale contained 12 items, six for mastery and six for
performance. Coefficient alpha reliability estimates for the state mastery scale ranged
from .70-.76, and the estimates for the state performance scale ranged from .76-.78. A
sample item from the state mastery scale is “I want to really understand the material in
this section.” A sample item from the state performance scale is “I am eager to show
how much I know about the material in this part of the class.”

Self-set grade goals were measured with four items, two for distal goals and two
for proximal goals (see Appendix G). Students were asked to indicate the grade they
were actually trying for, and the lowest acceptable grade (e.g., Locke 8 Bryan, 1968;
Campion 8 Lord, 1982). Goals set in the first wave were higher than goals set in later

waves, as found by Thomas and Mathieu (1994).

65

Self-set goals have typically been measured in classroom settings using these
types of questions (e.g., Locke 8 Bryan, 1968; Campion 8 Lord, 1982; Thomas 8
Mathieu, 1994). However, since mastery orientation may lead people to set goals other
than grade goals, subjective Ieaming goals were also measured. Questions were
included regarding the importance of Ieaming in the next section of the course. These
items are found in Appendix G. The coefficient alpha reliability estimate for the self-set
Ieaming goals scale ranged from .7883. From a construct validity perspective, it was
expected that state mastery orientation would be positively related to this scale, and
state performance orientation would be negatively related to this scale.

Because the reported self-set goals may be affected by the request to set such
goals (Wright, et al., 1994), a measure of goal valence was needed. The personal
importance of a self-set goal will help indicate which of the goals were induced, and
which were reinforced. While no specific hypotheses were made about induced and
reinforced goals, the valence measure will help to further understand behavioral
pattems associated with goal orientation. Goal valence was measured with three items
reflecting the importance of grade goals and Ieaming goals. Because of low item-total
correlations, the third item on the scale was not used In analyses. A sample item is “It
is very important to me to attain the grade goal I have set for the next exam.” The
coefficient alpha reliability estimate for this scale ranged from .58 to .68 across waves.

Goal-perfomnce cflscmm were measured by calculating the difference
between the actual performance and the goal for each performance episode. Actual
exam performance was obtained from the instructor of each course. Grades were
assigned values of 4.0, 3.5, 3.0, 2.5, 2.0, 1.5, 1.0, and 0 according to each instnrctor’s

grading scale. According to "gen, et al., (1979), feedback can come from the self as

66
well as external sources. Consequently, subjective Ieaming perceptions were
measured with several self-report questions (see Appendix G). A sample question is
“Regardless of my exam performance, I feel that l Ieamed a lot in the previous section
of the class.” The coefficient alpha reliability estimate, calculated on items 2 and 3 of
this scale, ranged from .73 to .78 across waves. Because of low item-total correlations,
item 1 from this scale was examined separately.

_Re_actions to feedb_ack_ were measured with several items intended to capture
the four reactions outlined by Kluger and DiNisi (1996); change in effort, change the
goal, abandon the goal, and discount the feedback. The goal change reaction was
measured with the goal setting items. A goal change score was computed for grade
goals and for course goals by subtracting the goal for the latter wave from that of the
previous wave. Change in effort was measured in three ways. First, several questions
regarding the time students devoted to the class were asked. Subjects were asked to
estimate the number of hours they studied for the previous exam (Campion 8 Lord,
1982), and how many hours they planned to study for the next exam. Additionally, they
were asked how many class sessions they missed in the previous performance
episode. Second, participants were asked if they intended to increase the amount of
effort they put into the class.

Finally, a three item version of the Mental Workload scale (Fisher, 1995) was
used to determine the amount of mental effort put forth in Ieaming the material for the
previous section of material. The original six item version had an internal consistency
reliability of .87. The internal consistency estimates for the three item version in this

study ranged from .77 to 80 across the three waves. A sample item is “I had to work

67
very hard to Ieam the material for the previous exam.” See Appendix H for the effort
items.

Feedback acceptance was measured with items based on the feedback source
credibility and feedback accuracy measures used by Podsakoff and Farh (1987).
Source credibility and feedback accuracy are the two primary factors involved in
whether individuals choose to accept or reject feedback (Ilgen, et al., 1979). If the
feedback is perceived as highly accurate, and the feedback giver has high credibility,
feedback is more likely to be accepted. In this study, two subscales (source credibility
and feedback accuracy) were used to measure feedback acceptance, with two items
for each subscale. The coefficient alpha reliability estimates for three waves of this
study ranged from .76 to .81 for the source credibility scale, and from .85 to .89 for the
feedback accuracy scale. A sample item for source credibility is “I feel the instnrctor of
this course is very knowledgeable about the course content.” A sample item for
feedback accuracy is “My exam grade reflected my true performance.” The items are
presented in Appendix H.

Procedure

Students completed Survey 1 during the first two weeks of class in the fall
semester. Participants initially signed up for a time to meet with the researcher in a
computer laboratory at Michigan State University in groups of 1429 students. At this
introductory session, the researcher explained the nature of the project, and asked
participants to sign a consent form. Participants were assured that their course
instructor would not see their responses to any of the survey items.

The researcher then described the two format options for completing surveys.

On their consent form, participants were asked to choose the e-mail option or the pencil

68
and paper option. Those who selected the pencil and paper option (n = 92) provided a
local mailing address for future surveys. They then completed the pencil and paper
version of Survey 1. The participants who selected the e-mail option for the study (n =
342) were given a brief training session on the use of the student e-mail system.
Those who did not have e-mail accounts were assisted to set up an account.
Participants provided the researcher with their e-mail addresses. A research assistant
immediately sent Survey 1 to the participants, as the researcher provided instruction on
how to complete and return the surveys on the e-mail system.

All participants completed Survey 1 before leaving the introductory session.
This sunrey consisted of the self-efficacy, and trait and state goal orientation measures.
Several demographic and academic background questions were also included in
Survey 1 (see Appendix A). These questions asked participants to indicate the number
of psychology courses previously taken, reasons for taking the current course, and their
current college grade point average (GPA).

Because first-year students did not have a current college GPA, students were
asked to provide SAT or ACT scores. Test scores were also obtained from the
University for each participant. When possible, the University provided ACT scores
were used. It is likely that the self-report test scores were accurate, as the correlation
between University provided and student self-report test scores was extremely high (r =
.94). In addition, the means for self-report (23.3) and University provided (23.2) ACT
scores were very close. Several students had SAT scores instead of ACT scores. In
these instances, a concordance table provided by the American College Testing
Service was used to translate the SAT score into an ACT score. Standardized test

scores (ACT) were used as control variables when testing the hypotheses.

69

Immediately prior to each exam in the relevant psychology courses, students
were asked to estimate the effort they put forth in the course in the time leading up to
the current exam. Students completed a brief, five-item paper and pencil questionnaire
containing time spent and mental workload items. Following each of the exams in their
respective courses, students completed Survey 2, consisting of the state goal
orientation, self-set goal, self-efficacy, interest, and reaction measures (see Appendix
B). In order to obtain an estimate of the test-retest reliability of the trait orientation
scales, trait orientation was also measured on each occasion. Surveys 2-4 were sent
to the participants one week after the exams in their respective courses. This delay
period ensured that each of the participants has actually received the feedback, taking
class absences into consideration. Subjects were instructed to answer the sunrey only
after they had received exam feedback. The two versions of the survey were sent out
such that participants would receive the survey at approximately the same time,
regardless of survey format.

The two sets of survey responses were compared to check for differences by
administration mode. While individual studies have suggested that computerized
testing results in slightly more or less impression management, the balance of studies
suggests that mode of test administration does not affect scores (Hough 8 Schneider,
1996). Non-cognitive measures administered by computer and by pencil and paper do
appear to be equivalent on both the number of factors and the factor loadings (King 8
Miles, 1995). In the current study, few differences were found between respondents in
the two administration mode groups. Students in the e-mail group scored significantly
higher on the trait mastery orientation scale, _F_ (1, 352) = 9.43, p <.01. The mean

score for the e-mail group was 4.14, SD = .44, and the mean score for the paper and

70
pencil group was 3.96, SD = .50. It is likely that this difference is related to self-
selection rather than measurement. Individuals who are high in trait mastery orientation
are theoretically expected to be willing to Ieam new things. It makes sense that people
high in trail mastery orientation would be more willing to Ieam how to use the e-mail
system in this study.

Missing Data Given the longitudinal nature of this study, it is not surprising that
there was substantial missing data. Precautions were taken to minimize the occurrence
of missing data. Reminder messages were sent to participants two days prior to the
distribution of the first post-exam survey. Reminders were also sent out the day
following the distribution of the surveys. Regardless, there were participants who did
not return one or more surveys. Three waves of data, including Survey 1, were
required for classification as a complete set of data. Incomplete data sets were still
used in the calculation of psychometrics for the data set. However, only complete data
sets were used for hypothesis testing.

In the first wave, 434 subjects completed surveys. Of these, 92 (21.2%)
selected the paper option, and 342 (78.8%) selected the e-mail option. In the second
wave, the sample size was reduced to 380. In the third wave, the sample was further
reduced to 371. However, subjects who had not returned Survey 2 were given the
opportunity to complete Survey 3, and many did. Therefore, all subjects who had
missed either Survey 2 or Survey 3 were removed from the final data set. The final
sample size used for between wave analyses was 354, resulting in an 81.6%
completion rate. In the final sample there were 70 subjects using paper surveys

(19.8%) and 284 subjects using e-mail surveys (80.2%). These percentages do not

71
vary substantially from the percentages in the first wave, suggesting that administration

mode was unrelated to subject mortality.

RESULTS

Data Anaysis Strategy

The data analysis for this study began with a series of exploratory and
confirmatory factor analyses to determine the factor structure of the measures used. It
was expected that the trait and state goal orientation data would be best represented
by a four factor model, with each a priori scale defining one factor. This factor structure
was expected because of the different causal antecedents suggested for the trait and
state variables. The trail items were intended to measure a dispositional construct
caused by multiple, interrelated factors in an individual’s background and experience.
The state items, conversely, have a specific reference to a situation, and should
correlate more strongly with other state items than with the corresponding trait items
because of the causal antecedents. This correlational pattern was expected even in
the first wave of data, as interest was expected to impact the relationship between trait
mastery orientation and state mastery orientation. Once the final scales were
determined, interscale correlations were examined. Next, a series of hierarchical
regressions were run to test the within-wave hypotheses. Hierarchical regressions
were also used to test the between-wave hypotheses.
EBCILOI' Apalvses

_Gpaj orientiion scalar; Principal components exploratory factor analyses
including state and trait mastery orientation items and state and trait performance

orientation items were conducted as a preliminary investigation of factor stmcture.

72

73
Four factors were extracted. The eigenvalues for the four retained factors were 5.4,
3.8, 1.7, and 1.6 (the next highest eigenvalue was 1.2). An oblique rotation was used.
The factor solution across all three waves of data revealed a lack of simple stmcture.
State mastery items 1 and 11 consistently loaded highly on the factors representing
both state and trait mastery. One of the trait performance items (item 7) loaded
consistently on the factors representing both trait and state performance. Coefficient
alpha reliability estimates were in the acceptable range for the trait mastery, trait
performance, and state mastery scales for all three waves (.75 and above). Coefficient
alpha estimates were lower for the state performance scale, ranging from .71-.74
across the three waves.

Based on these analyses, four items were deleted from the goal orientation
scales; two from the state mastery, one from state performance, and one from trait
performance. Deletion of an item from the established trait performance scale of
Button, et al. (1996) was deemed appropriate because this scale had never been
analyzed in concert with the state scale used in the current study. State performance
item 2 was removed from the scale because of its low corrected item-total correlation
with the state performance scale (.19 -.21 across all three waves). Removal of this item
increased the reliability estimate for the scale in all three waves (.76-.78). Coefficient
alpha reliability estimates for the revised goal orientation scales are presented in Table
5. All remaining analyses were performed with the revised scales.

One additional exploratory factor analysis was conducted on the Wave 1 data to
examine the primary self-report scales used in the study: trait and state mastery
orientation, trait and state performance orientation, interest, and academic self-efficacy.

The principal factors method was used to extract six factors. The eigenvalues for the

74
six factors were 7.21, 3.86, 2.29, 2.04, 1.60, and 1.24 (the next highest eigenvalue was
1.14). An oblique rotation was used. The six major self-report variables demonstrated
a clean factor structure. All items loaded most strongly on the factor represented by
the appropriate scale. For example, all five academic self-efficacy items loaded most
strongly on factor one, and no items from other scales loaded most strongly on that
factor. The same was true for the remaining five factors. Factor two was represented
by the trait performance orientation items, factor three by the interest items, factor four
by the trait mastery orientation items, factor five by the state performance orientation
items, and factor six by the state mastery orientation items. The interfactor correlations
ranged from .00 (factors 2 and 3) to .41 (factors 1 and 4).

The next step of the analysis was to perform several confirmatory factor
analyses. A four factor model of the goal orientation scales was subjected to a
confirmatory factor analysis using AMOS 3.6 for Windows (see Figure 4). Wrth the
Wave 1 data, this model achieved a moderate level of fit, if = 731.022, fl = 246,
p<.05; root mean residual (RMR) = .046, Goodness-of-Fit Index (GFI) = .87, root mean
square enor of approximation (RMSEA) = .067. Factor loadings for the Wave 1 data
are presented in Table 1. At Wave 2, the model fit decreased somewhat (x2 = 850.140,
a; = 246, p < .05; RMR = .052, GFI = .83, RMSEA = .081). At Wave 3, the fit did not
change appreciably from Wave 2 (x2 = 823.533, at = 246, p < .05; RMR = .053, GFI =
.83, RMSEA = .080). Each model was tested using the trait goal orientation measures
from that wave. This allows for a more stringent test of the factor stnrcture. If only
Wave 1 measures of trait goal orientation had been used, time could have been

partially responsible for the reduced covariation among scales.

75

Given the lack of simple structure apparent in the exploratory analyses on the
goal orientation scales, this level of model fit is quite reasonable. In the exploratory
analysis, several of the items loaded on multiple factors. This is reflected in the
modification indices. For example, including a link between trait performance item 4
and the state performance scale would improve model fit. In addition, the four factor
model fits significantly better than a nested two factor model, in which all mastery
orientation items load one factor and all performance orientation items load on one
factor. The nested 2 factor model using Wave 1 data does not fit as well as the 4
factor model (38 = 1146.352, d_f = 248, p < .05: RMR = .062, GFI = .795, RMSEA =
.091). The x2 decreases by 415.33 with the move to the four factor model, with a
decrease of only 2 degrees of freedom. This change in x2 is highly significant.
Although the four factor model fits only moderately well, the results of the confirmatory
factor analysis do generally support the distinction between the trait and state scales.

Goa_I orientation and interest. Because of high correlations between the state
mastery and interest scales, a confirmatory factor analysis was conducted. A three
factor model including state and trait mastery orientation and interest was fit to the
data. The model fit the Wave 1 data reasonably well (at: = 368.276, at: = 101, p <.05;
RMR = .039, GFI = .90, RMSEA = .078). At Wave 2, the model fit decreased slightly
(x2 = 416.853, _tfi = 101, p <.05; RMR = .044, GFI = .87, RMSEA = .091), and did not
change appreciably at Wave 3 (x2 = 394.754, g = 101, p <.05; RMR = .047, GFI = .88,
RMSEA = .089). The modification indices did not suggest the addition of any
parameters between the interest scale and any of the state mastery orientation items.
A two factor model (one factor combining state mastery and interest, and one factor

representing trait mastery orientation) was fit to the data, and fit and less well than did

76
the three factor model at Wave 1 (x2 = 638.124, d_f = 103, p <.05; RMR = .067, GF I =
.79, RMSEA = .118). Thus, the confirmatory factor analysis supports the separation of
the interest and state mastery orientation scales, regardless of the correlation between
the two scales.

MASIBW oriejntatjon and Ieaming goals. State mastery orientation and self-set
Ieaming goals were strongly correlated in both Waves 2 and 3 (r = .58 and .69
respectively, p <.01). A confirmatory factor analysis was conducted to determine if
these variables were statistically distinguishable. A two factor model with all state
mastery items loading on one factor and all Ieaming goal items on the other factor was
tested, using Wave 2 data. The data fit the model well (x2 = 54.133, d_f = 13, p <.05;
RMR = .033, GFI = .96, RMSEA = .092). A one factor model combining the two scales
did not fit the data as well (98 = 74.049, d_f = 14, p <.05; RMR = .036, GFI = .94,
RMSEA = .107). Thus, these two scales were used as separate variables in the data
analysis.

Aaaction to feedback scalp; The two reaction to feedback subscales, source
credibility and feedback accuracy, were subjected to a confirmatory factor analysis to
determine if these scales should be used as one scale or as two separate scales. The
variance of the enor term associated with the second indicator on the feedback
accuracy scale was negative (-.052). Consequently, this error term was set to zero.
With the error term set to zero, the two factor model, using Wave 2 data, fit the data
well (38 = .402, d! = 2; RMR = .009, GFI = .99, RMSEA = .00). A one factor model
combining the two scales did not fit the data as well (38 = 155.848, d_f = 2, p <.05; RMR

= .163, GFI = .85, RMSEA = .451).

77

Because of the negative enor term in the two factor model, a principal
components factor analysis was also conducted to examine the factor structure. All
factors with an eigenvalue greater than 1 were extracted, resulting in two factors which
were then subjected to an oblique rotation. This analysis suggested two distinct
factors, the first representing feedback accuracy (eigenvalue = 2.3), and the second
representing source credibility (eigenvalue = 1.1). The eigenvalue for the third factor
was .36. All factor loadings for the appropriate factor were greater than .83, and none
of the loadings for the alternate factor were greater than .15. Thus, it was determined
that these scales should be used separately.

Cone:l_a_tions

Wrthin-wave correlations are presented in Tables 2 - 4. lntemal consistency
estimates, where appropriate, are presented in the diagonal of each matrix. All
correlations were computed using the final sample of 354 subjects. lntemal
consistency estimates were calculated on the full sample within each wave. The Wave
1 trait measurements were used in each correlation matrix. Theoretically, the trait
should not change across measurement periods within a relatively short time period of
three months.

_Go_al orientat_idns. Correlations support the distinction between trait mastery
and performance orientations (r = -.04, ns). At each wave, state mastery and state
performance were significantly conelated (r = .33, .31, and .33, p < .01). This is
consistent with the view that states are caused by some extemal source. These
correlations suggest that both states are affected by a common source. One common
causal antecedent may have been interest in the course. It was expected that state

mastery would be positively associated with interest. This relationship did occur, with

78
correlations between state mastery and interest ranging from .50 to .53 across three
waves. State performance orientation was also related to interest, with correlations
ranging from .09 to .24.

Correlational analyses were used to estimate the test-retest reliability of the trait
and state goal orientation scales (see Table 5). The mastery state scale had test-retest
coefficients of .59 and .55 between one wave, and .43 between two waves. The
performance state scale had test-retest coefficients of .65 and .67 between one wave,
and .53 between two waves. These estimates are somewhat higher than would be
expected for a state, given variation in the situation. The mastery trait scale had test-
retest coefficients of .60 and .65 between one wave, and .52 between two waves. The
performance trait scale had test-retest coefficients of .70 and .74 between one wave (4-
6 week time span), and .61 between two waves (3 month time span). According to
Schuerger, Zarella and Hotz (1989), the average test-retest reliability coefficient for
major personality inventories such as the 16-PF and CPI is .73. Following this
guideline, the performance trait scale demonstrates acceptable test-retest reliability at
4-6 weeks. The three month test-retest coefficients are similar to those obtained by
VandeWalle (in press). All test-retest estimates for trait scales are higher than the
estimates for the state scales. However, the difference is not as large as anticipated.

_Ggl orientation and self-set goals. Each of the goal orientation variables
showed a different pattern of correlations with self-set goals. In Wave 1, course goal
and exam goal were positively correlated with trait and state mastery orientation, with
the trait correlations of higher magnitude. Both grade goals had a slight positive
correlation with state performance orientation, and a non-significant correlation with trait

performance orientation. In Waves 2 and 3, both grade goals were still positively

79
correlated with both mastery orientations, but the correlations with state mastery
orientation were of higher magnitude. The correlations between grade goals and state
performance orientation were non-significant at Wave 2, and were slightly positive at
Wave 3. The correlations between trait performance and grade goals were negative in
Waves 2 and 3.

In Waves 2 and 3, Ieaming goals were measured in addition to the grade goals.
Leaming goals were strongly correlated with state mastery orientation in both waves
(r = .58, .69 respectively, p <.01). Leaming goals had a moderate, positive correlation
with both trait mastery and state performance orientations, and a non-significant
correlation with trait performance orientation. To test the possibility that this pattern of
correlations may be due to the influence of interest on Ieaming goals and both state
orientations, interest was partialed out of the correlations between state orientations
and Ieaming goals. The correlations were reduced somewhat, but the relationship
between the state orientations and Ieaming goal was still present (state mastery Wave
2 partial r = .47, p <.01; state performance Wave 2 partial r =.23, p <.01).

Goal valence. Goal valence was negatively related to the difficulty of self-set
goals, both for exams and the entire course (see Tables 2—4). Thus, participants who
set more difficult goals tended to view them as less important. Goal valence was
positively correlated with both trait performance orientation and state performance
orientation, and negatively correlated with both state and trait mastery orientation and
interest. These correlations with goal orientation support theory that suggests that
performance oriented individuals tend to place greater emphasis on visible outcomes
such as grades. The normal distribution of the goal valence scores within each wave

suggests that many of the participants were setting induced goals rather than natural or

80
reinforced goals (Wright, et al., 1994). Therefore, the effects of goal valence were
considered when testing hypotheses including self-set goals.

Exam performance. As expected, there was a moderate, positive correlation
between exam goals and exam grades across all three waves. The correlations ranged
from .35 in Wave 1 to .45 in Wave 3 (p <.01). Course goals and exam grades were
also positively correlated, with the correlation increasing across the three waves. At
Wave 1, the correlation was .26 (p <.01). At Wave 2, the correlation increased to .37,
and to .49 at Wave 3. These correlations decreased somewhat after partialing out the
measure of cognitive ability, ACT scores, but were all still significant.

Academic self-efficacy was also related to exam performance. Wrthin each
wave, academic self-efficacy was positively correlated with exam scores, even with
cognitive ability partialed out. Exam scores were also related to self-efficacy in the next
wave. The partial correlation between Exam 1 scores and self-efficacy in Wave 2 was
.34 (p <01), and the partial correlation between Exam 2 scores and self-efficacy in
Wave 3 was .29 (p <.01). After accounting for ACT scores, there were no significant
correlations between general self-efficacy and exam scores.

Exam scores were generally uncorrelated with the goal orientation and interest
measures (see Tables 2 - 4), with the exception of trait performance orientation. In
Waves 1 and 2, trait performance orientation was negatively correlated with exam
grades (-.23, -.17). After partialing out ACT scores, the correlations decreased to -.18
and -.10. In the third wave, both interest and state mastery orientation displayed a
slight, positive correlation with exam scores.

Exam performance was also related to some of the reactions to goal-

performance discrepancies. Exam scores were negatively related to individuals’ plans

81
to increase their effort in the next section of the course (after exam 1 r = -.32; after
exam 2 r = -.31; p <.01). Exam scores were positively related to individuals’ judgments
of the credibility of the feedback source both within and between waves. Within waves,
exam scores were positively correlated with source credibility (r = .36 for exam 1 and
.24 for exam 2, p <.01). Between waves, the source credibility judgment for the first
exam was positively associated with the performance on the second exam (r = .19, p
<.01).

Effort measures. The three variables measured on the pro-exam survey,
hours studied, number of class sessions missed, and perceived mental workload, were
minimally related (see Table 6). At Wave 3, hours studied was positively related to
mental workload (g = .18, p < .01). Also at Wave 3, classes missed was negatively
related to mental workload (I = -.16, p < .01). Number of classes missed was
consistently related to exam scores. Correlations ranged from -.20 to -.22 , all
significant at p < .01. Thus, the more classes students missed, the lower their exam
grade. Additionally, there was a negative relationship between state mastery
orientation and the number of classes missed in the second and third waves (r = -.20
and -.21 respectively, p <.01). Consistent with Dweck’s hypothesis that mastery
orientated individuals would feel that effort is related to increased ability, individuals
with a higher state mastery orientation tended to miss fewer classes, which appears to
have resulted in superior exam performance.

Hmthesis Testing: Wrthin Wave
All hypotheses were tested using the final sample of 354 students. Hierarchical
regression was used to test the hypotheses. General self-efficacy and academic ability

(ACT scores) were entered as control variables. The effects of trait goal orientation

82
were examined in each analysis before testing the effects of state goal orientation. By
testing the hypotheses in this manner, a clearer determination of the incremental value
of state goal orientation can be made.

Hymahesis 1. Hypothesis 1 suggests that each trait goal orientation is
positively related to the corresponding state goal orientation. This hypothesis was
tested separately for each dependent variable: state mastery and state performance
orientation (see Tables 7 and 8). For state mastery orientation, the hierarchical
regression started with the entry of general self-efficacy and ACT scores in step 1.
Trait mastery orientation was entered in step 2. For Waves 2 and 3, the state mastery
score from the previous wave was entered next. In Wave 1, general efficacy, ACT
scores, and trait mastery all were significant predictors of state mastery orientation.
General efficacy and trait mastery positively affected state mastery, while ACT scores
had a negative effect (b = -.13, p <.05). These relationships held in Wave 2, with the
addition of a significant positive beta weight for state mastery (rs =.47, p <.01). In the
Wave 3 analysis, general efficacy, trait mastery orientation, and state mastery had
significant effects on state mastery at Wave 3.

Because academic self—efficacy was consistently correlated with trait mastery
orientation (r = .42 - .45), an additional step was added to each of the hierarchical
regressions described above. Academic self-efficacy was a significant, positive
predictor of state mastery orientation in each wave. However, state mastery orientation
remained a significant predictor with the effects of academic self-efficacy partialed out.
Although academic self-efficacy is a significant predictor of state mastery orientation, its

effects are not redundant with those of the previous wave state mastery orientation.

83

For the prediction of state performance orientation, academic self-efficacy, ACT
scores, and general self-efficacy were all entered on the first step of the hierarchical
regression. Trait performance orientation was entered next, followed by the state
performance orientation of the previous wave (in Waves 2 and 3 only). At each wave,
academic self-efficacy and trait performance orientation were significant predictors of
state performance orientation. ACT scores predicted state performance orientation in
Wave 3 (I3 = -.14, p <.05). In Waves 2 and 3, the previous state performance
orientation was a significant predictor of state performance after trait performance
orientation was taken into account.

From these analyses, it can be concluded that Hypothesis 1 was supported. In
all three waves, state mastery was predicted by trait mastery orientation. The effect of
trait orientation on state orientation decreased steadily over time, presumably as the
situation became more salient. Likewise, state performance was predicted by trait
performance orientation at each wave, with the beta weight decreasing overtime.

Hyghesis 2. Hypothesis 2 suggested that interest and trait mastery orientation
would interact to predict state mastery orientation. To test this hypothesis, interaction
terms between trait mastery orientation and interest were created. Trait mastery at
Wave 1 was used for each interaction term, and interest at each wave was used. The
hierarchical regression analysis was performed by entering general self-efficacy and
ACT scores on the first step. On the second step, trait mastery orientation was
entered, followed by academic self efficacy, state mastery orientation (in Waves 2 and
3) and interest on the third, fourth and fifth steps. Finally, the interaction between trait

mastery and interest was entered (see Table 9).

84

In each wave, both general self-efficacy and ACT scores were significant at step
1. General self-efficacy had a strong, positive beta weight (.41), while ACT scores had
a negative beta weight (rs = -.13, p<.01). At the second step, trait mastery orientation
added significantly to the prediction of state mastery orientation. In Waves 2 and 3, the
state mastery orientation from the previous wave added significantly to the prediction of
state mastery orientation. In fact, both trait and state mastery orientation explained
unique variance in the dependent variable in Waves 1 and 2. In Wave 3, the beta
weight for trait mastery became non-significant once state mastery was added to the
equation. Academic self-efficacy and interest also added significantly when entered in
their respective steps. The interaction between trait mastery orientation and interest
was significant only in the third wave analysis (IS = .74, p<.01). Thus, hypothesis 2 was
supported in Wave 3 only. In Waves 1 and 2, trait mastery orientation and interest
acted independently to predict state mastery orientation. The effect of trait mastery
orientation on state mastery orientation weakened overtime. In Wave 3, the direct
effect of trail mastery was no longer significant, and the interaction became significant.

Hygthesis 3. To test the hypothesisithat state mastery orientation would have
a direct, positive effect on goal difficulty, exam goals were regressed on general self-
efficacy, ACT scores, trait mastery, interest, and state mastery. Trait mastery and
interest were included in the analysis as control variables, given their relationship with
state mastery orientation. For Wave 1, general self-efficacy, ACT scores, and mastery
trait orientation were significant predictors of exam goals. In Wave 2, ACT scores had
a significant, positive beta weight in the first step. .In the second step, interest had a
significant beta weight ([5 =20, p <.01), but trait mastery did not. In the third step, state

mastery had a significant effect on the exam goal (I3 =.16, p <.05). In Wave 3, only

85
ACT scores and interest predicted the exam goal. Thus, it appears that mastery
orientation became progressively less predictive of grade goals. At Wave 1, grade
goals were predicted by trait mastery. At Wave 2, they were predicted by state
mastery. At Wave 3, they were predicted by neither mastery orientation variable.

Because of the relationship between academic self-efficacy and state mastery
orientation, academic self-efficacy was added as a final step in the regression analyses
for Hypothesis 3. In each wave, academic self-efficacy was a significant predictor of
exam goals (see Table 10). The test of the hypothesis as stated provided partial
support for the positive effects of state mastery orientation on exam goals. Academic
self—efficacy, however, demonstrated much larger, more consistent effects for the
prediction of exam goals.

The effect of mastery orientation on self-set Ieaming goals was also tested by
regressing self-set Ieaming goals on general self-efficacy, ACT scores, trait mastery,
interest, state mastery, and academic self-efficacy (see Table 11). In both waves,
general self efficacy, interest, and state mastery orientation had a positive effect on
Ieaming goals. ACT scores had a negative effect on Ieaming goals in Wave 2
(I3 = -.19, p <.01). Trait mastery orientation predicted Ieaming goals in Wave 2 (3 =
.15, p <.05), but not in Wave 3. Academic self-efficacy had little effect on Ieaming
goals in Wave 2, and a positive effect in Wave 3 (IS = .12, p <.01). Given the nature of
the Ieaming goals scale, this analysis was exploratory. However, it does appear that
mastery orientation is much more strongly related to the difficulty of self-set Ieaming
goals than it is to the difficulty of specific grade goals for exams.

Hypothesis 4. Hypothesis 4 suggested that academic self-efficacy and state

performance orientation would interact to predict exam goals. More specifically, It was

86
hypothesized that individuals with a high state performance orientation and high
academic self-efficacy would tend to set higher exam goals. To test this hypothesis,
exam goals were regressed on general self-efficacy, ACT scores, trait performance
orientation, state performance orientation, academic self-efficacy, and the interaction of
state performance orientation and academic self-efficacy (see Table 12).

In Wave 1, general efficacy, ACT scores, and academic self-efficacy
significantly predicted goal difficulty. State performance orientation did not have a
direct effect, but the interaction of academic self-efficacy and state performance was
significant (8 = -1.10, p <.01). Results were similar for Wave 2, with ACT scores and
academic self-efficacy directly affecting goal difficulty, and a significant academic self-
efficacy X state performance interaction (8 = -.77, p <.01). In Wave 3, ACT scores and
academic self-efficacy were still significant. State performance orientation was also
significant (3 = .12, p <.05). The interaction was not significant in Wave 3. The Wave
1 interaction is depicted in Figure 6. This figure indicates that when self-efficacy was
high, there was a negative relationship between state performance orientation and the
grade goal set for exam 1. When self-efficacy was low, there was a positive
relationship between state performance orientation and the grade goal set for exam 1.
Therefore, the form of the interaction does not support hypothesis 4.

Hypothesis 5. Hypothesis 5 suggested that distal goals (goals for the entire
course) would positively affect the difficulty of exam goals. This hypothesis was tested
by adding an additional step containing the participants’ course goals to the analysis for
hypothesis 4 (see table 12). In each wave, the course goal was a significant predictor
of the specific exam goal, adding significantly to the R2 (AR2 = .40 - .42) even after

considering seven other variables. These results support Hypothesis 5.

87

Hymnesis 6. This hypothesis, which suggested that academic self-efficacy
would have a direct, positive effect on exam goals, was incidentally tested in the
regression analysis for Hypothesis 4 (see Table 12). Results indicated that academic
self—efficacy did have a direct effect on exam goal difficulty. Beta weights for academic
self-efficacy ranged from .26 in Wave 1 to .41 in Wave 2 (p <.01). These results
support Hypothesis 6.

Given the correlational relationships found between goal valence (the
importance of attaining the self-set goal) and many of the variables of interest, goal
valence was added to the regression equation presented in Table 12. Goal valence
was entered as the last direct effect in the analysis. It did not add significantly to the
prediction of goals. Therefore, after considering the effects of the main variables of
study, goal valence did not impact the difficulty of the self-set goals. Goal valence was
not considered in the remainder of the analyses.

Hmmsis 7. This hypothesis suggests that state goal orientation would
moderate the relationship between goal-performance discrepancies and reactions to
feedback. This hypothesis was tested in two ways; first by dichotomizing each variable
using a median split and conducting a series of chi square analyses as suggested in
Figure 3, and secondly with hierarchical regression.

The chi-square tests indicated that after a negative performance discrepancy in
Wave 1, individuals with a high mastery orientation were more likely to increase effort
than were individuals with a low mastery orientation (3:? = 4.74, g = 1, p <. 05). In
addition, after a negative performance discrepancy in Wave 2, individuals with a high
performance orientation were less likely to reject feedback than were individuals with a

low performance orientation (x2 = 3.77, d = 1, p <.05).

88

Several other trends were apparent in the data. For example, after a positive
performance discrepancy in Wave 2, individuals with a low performance orientation
were less likely to reject feedback than were individuals with a high performance
orientation of = 2.99, d_f = 1, p <.10). However, these trends were not significant, as
analyses concerning the positive discrepancy group suffered from low power as a result
of the relatively small number of individuals in this group (Wave 1, n = 48; Wave 2, n =
34). The small number of individuals with positive discrepancies is partially the result of
range restriction. If a participant set a goal of 4.0 for the exam, it would be impossible
for him/her to have a performance discrepancy, as 4.0 is the highest grade possible.

To further investigate the relationships proposed in Hypothesis 7, and to more
directly test the interactions, separate regressions were performed for each of the three
reactions to feedback; goal change, effort change, and feedback credibility. The goal-
perforrnance discrepancy was entered on the first step, followed by the trait goal
orientations. The state goal orientations were entered on the third and fourth steps,
followed by the interaction between the discrepancy and the state goal orientations on
steps five and six.

Results for the planned increase in effort (see Table 13) indicate that in both
waves, the discrepancy between one’s goal and one’s performance had a strong,
positive effect (8 = .32 and .29, p <.01). Trait goal orientations did not significantly
impact planned effort. State mastery orientation affected planned effort in Waves 1
and 2 (IS = .14, .15 respectively, p <.01). In contrast, state performance orientation did
not have a significant effect. In Wave 1, the interaction between state mastery
orientation and the discrepancy added significantly to the prediction of planned effort

(AR2 = .02, p <.05). Figure 7 shows that after a negative discrepancy, people with a

89
high state mastery orientation planned to increase their effort during the next section of
the course. After a positive discrepancy, people with a high state mastery orientation
were less likely to plan to increase their effort. This interaction partially supports
Hypothesis 7. No other interaction terms added significantly to the prediction of planned
effort.

Regarding changes in goals, the discrepancy between the goal and exam
performance was again a significant predictor between both waves (see Table 14). In
Wave 1, state mastery orientation had a positive effect on future goals. No other
variables significantly affected goal change.

Feedback credibility was also significantly affected by the goal-performance
discrepancy in both waves (see Table 15). Trait performance orientation was a
significant, negative predictor of feedback credibility in both Wave 1 and Wave 2 (13 = -
.25, -.31 respectively, p <.01). After considering the trait goal orientations, state
mastery orientation was a significant predictor in both waves. Finally, the interaction
between state performance orientation and the discrepancy was significant (13 = .77, p
<.01) in Wave 2. Figure 8 shows that after a negative discrepancy, people with a high
state performance orientation tended to reject feedback. After a positive discrepancy,
people with a high state performance orientation were more likely to accept the
feedback. This interaction supports Hypothesis 7.

These results indicate partial support for Hypothesis 7. The discrepancy
between the students’ goals and their actual performance was a robust predictor of all
three reaction variables across both time periods. State mastery orientation was also a

consistent predictor of all three dependent variables. However, the interaction terms

90
were generally not significant. Of the six interaction terms, only two were significant.
The form of these interaction terms generally supported Hypothesis 7.
Hypothesis Testidg;_Between Waye

Hmthesis 8. This hypothesis suggested that perfonnance-goal discrepancies
at one time period would be directly related to interest in the course at the next time
period. Hypothesis 8 was tested by regressing interest at Waves 2 and 3 on general
self-efficacy, ACT scores, interest from the previous wave (as a covariate), and the
discrepancy between the exam goal and exam performance. This discrepancy variable
was created by subtracting the goal from the performance. Thus, a negative
discrepancy score reflects an exam score lower than the participant’s exam goal.

Results of the hierarchical regression (see Table 16) indicate that general self-
efficacy, interest from the previous wave, and the discrepancy score all affected
interest at the next time period. The beta weight for the discrepancy score was
negative between Waves 1 and 2, and between Waves 2 and 3 (I3 = -.11, -.10
respectively, p <.01). These results support Hypothesis 8.

HMhesis 9. This hypothesis suggested that perfonnance-goal discrepancies
at one time period would be directly related to academic self-efficacy at the next time
period. This hypothesis was tested by regressing academic self-efficacy from Waves 2
and 3 on general self-efficacy, ACT scores, academic self-efficacy from the previous
wave (as a covariate), and the discrepancy between the exam goal and exam
performance (see Table 17). Results indicated that ACT scores, general self-efficacy,
academic self-efficacy from the previous wave, and goal-performance discrepancies all

affected academic self-efficacy. The regression weights for the discrepancy were

91
negative between Waves 1 and 2, and between Waves 2 and 3 (I3 = -.22, -.12
respectively, p <.05), These results support Hypothesis 9.

Hymthesis 10. This hypothesis suggested that negative goal-performance
discrepancies would cause state performance orientation to increase, while positive
goal-performance discrepancies would cause state performance to decrease.
Hypothesis 10 also suggested that goal-performance discrepancies would interact with
trait performance orientation to affect state performance. This hypothesis was tested
by regressing state performance orientation from Waves 2 and 3 on trait performance
orientation, state performance from the previous wave, the discrepancy between exam
goal and exam performance, and the interaction of trait performance and the
discrepancy score. Results of the hierarchical regression (see Table 18) showed that
while both trait and state performance orientation predicted the dependent variable,
neither the goal discrepancy nor the interaction term were significant predictors of state

goal orientation. Thus, Hypothesis 10 was not supported.

DISCUSSION

The current study was designed to examine the effects of state and trait goal
orientation over time on the self-regulatory processes of goal setting and reacting to
goal-performance discrepancies. Several of the hypothesized relationships were
supported, both within and between performance episodes. Mastery orientation, both
state and trail, positively affected exam goals and Ieaming goals, but both performance
orientations were generally unrelated to exam goals. Some evidence was found for a
moderating relationship of goal orientation between discrepancies and reactions. This
study also investigated properties of trait and state goal orientation. Results
demonstrated that trait goal orientation was a significant predictor of state goal
orientation. Interest affected state mastery orientation, although the hypothesized
interaction between interest and trait mastery orientation was not supported.
Implications of specific findings are discussed below.

Role of Goal Orientation manterest i_n Self-regulation

A major goal of this study was to develop a richer, more detailed model of self-
regulation that integrated important individual difference variables. Traditionally, self-
regulation models have considered limited individual variables such as ability and past
performance (e.g., Campion 8 Lord, 1982; Locke 8 Latham, 1990). Recent research

demonstrating the complex relationships between goal orientation and numerous

92

93
Ieaming outcomes (e.g., Kozlowski, 1995; Fisher 8 Ford, in press) has pointed toward
the inclusion of goal orientation in a broader self-regulation framework. The models
developed in this study suggest that state goal orientation is involved in self-regulation
in two ways (see Figure 1). First, state goal orientation was expected to affect the
difficulty of self-set goals. Previous literature has linked goal orientation with task
choice (e.g., Dweck, 1986; Dweck 8 Leggett. 1988). Individuals with high mastery
orientation tend to choose more difficult tasks that provide the opportunity to Ieam new
concepts. Individuals with a high performance orientation tend to choose tasks that
allow that individual to avoid appearances of poor performance (Dweck, 1986).

Consequently, in the current study, state mastery orientation was expected to
be positively related to goal difficulty, and state performance orientation was expected
to interact with self-efficacy to predict goal difficulty. In this study, task difficulty was
operationalized with participants’ goals for exam grades in psychology courses.
Participants who set a difficult goal for themselves were considered to have chosen a
more difficult task. Exam grades were then used as indicators of Ieaming.

The second role hypothesized for state goal orientation in Figure 1 is that state
goal orientation should affect how individuals react to discrepancies between their
goals and performance feedback. Mastery goal orientation has been linked with an
adaptive response pattern in achievement situations, while performance orientation has
been linked with a maladaptive pattern (Dweck, 1986; Dweck 8 Leggett, 1988).
Building on the previously identified response patterns, individuals with a high state
mastery orientation were expected to have more adaptive responses to negative

feedback, while individuals with a high state performance orientation were expected to

94
protect their ego. resulting in the use of less adaptive responses to goal-performance
discrepancies.

In the remainder of this discussion section, the results of the study are reviewed
in light of the expectations outlined above and depicted in Figure 1. The discussion
begins with a review of the effects of the individual difference variables (goal
orientation, self-efficacy, and interest) on goal setting. The effects of these variables
on reactions to discrepancies are described next, followed by a discussion of the
implications of this study for state and trait measurement of goal orientation. Finally,
study limitations and implications are discussed.

Effects igaal orientation and interest on goals. As hypothesized, both state
and trait mastery orientation positively affected exam goals. In the first wave, trait
mastery orientation predicted exam goals. In the second and third waves, state
mastery orientation affected exam goals, and trail mastery orientation did not. In
addition to setting specific grade goals for each exam, participants were asked to
indicate their goals for Ieaming. Subjective Ieaming goals were included in this study
because individuals with a high mastery orientation should focus on Ieaming rather
than on grades. Exam grades, while used as indicators of Ieaming, inherently lead to
comparisons with others. Students with a high mastery orientation should be less
focused on grades, and are likely to be focused on goals that specifically reference
Ieaming rather than grades. The data suggested that this is the case. State mastery
orientation was strongly related to Ieaming goals in both Waves 2 and 3 (see Table 11).
In fact, the relationship between state mastery orientation and Ieaming goals was much

stronger than the relationship between state mastery orientation and grade goals.

95

Perhaps the most surprising finding in this study concerned the relationship
between state performance orientation and goal setting. Previous goal orientation
literature had suggested that individuals with a strong performance orientation
preferred easier tasks, consequently setting lower goals, in order to avoid negative
judgments of their ability (e.g., Dweck 8 Leggett, 1988). Later scholars (e.g., Farr, et
al., 1993) suggested that this relationship may be moderated by self-efficacy. They
suggested that individuals with a high performance orientation would indeed prefer less
challenging tasks (i.e., easier goals) if they had low self-efficacy. However, individuals
with high performance orientation and high self-efficacy would prefer more challenging
tasks (i.e., set more difficult goals).

The results of this study indicated that neither state nor trait performance
orientation had a significant, direct effect on exam goals in the first two waves. State
performance orientation was a positive, significant predictor of exam 3 goals. The
interaction between academic self-efficacy and state performance orientation was
significant in Waves 1 and 2. However, the form of the interaction was opposite that
which was predicted. Individuals with high state performance orientation and low
academic self-efficacy tended to set more difficult goals, while individuals who were
high in both characteristics tended to set lower goals. This effect was not large, and
decreased steadily across waves, until it was non-significant at wave three.

This hypothesis suggesting the interaction between state performance
orientation and self-efficacy was based on the suggestions of Dweck (1986, 1989) and
Farr, et al. (1993). These researchers focused primarily on the positive aspects of an
individual having high performance orientation and high self-efficacy. They contended

that high self-efficacy would allow individuals to overcome the problems associated with

96
high performance orientation, and to exhibit behaviors more consistent with the
adaptive response pattern. Dweck (1986, 1989) did, however, suggest an alternate
course of events. She suggested that individuals who had high performance
orientation and low self-efficacy could either set low, easily attainable goals, or could
set excessively difficult goals, safe with the knowledge that failure to attain such difficult
goals does not signify low ability. Perhaps this latter scenario was the case in this
sample of students. Those individuals with low self-efficacy may have been setting
unrealistically high goals as a defense mechanism. These results indicate that the
relationship between performance orientation and goal setting warrants more research
attention.

Both academic self-efficacy and interest had larger effects on exam goal
difficulty than did either mastery or performance orientation. High self-efficacy has
consistently been found to lead to the selection of more difficult goals (e.g., Locke 8
Latham, 1990). Thus, this finding is consistent with previous literature. Self-efficacy
had a minimal impact on Ieaming goals, however. Self-efficacy appears to be a
predictor of more traditional, specific, measurable goals like exam grades, while
mastery orientation is a better predictor of more subjective, lntemal goals.

Participants’ interest in course played a greater role in the self-regulation model
than was hypothesized. The direct effect of interest on goal difficulty in Waves 2 and 3
was unexpected. As shown in Figure 1, interest was hypothesized to affect goals only
indirectly through state mastery orientation. However, the students in this sample who
were more interested in the course tended to set higher goals, regardless of their state
mastery orientation. This effect increased over time as students became more familiar

with the course and could make more informed judgments about their interest. The

97
variation in interest also increased over time, as the standard deviation of interest
increased from .80 to .96 from Wave 1 to Wave 3. Interest was also strongly related to
participants’ Ieaming goals.

These results suggest that self-set goals are indeed related to individual
difference variables other than the frequently studied variables of self-efficacy and past
performance. Self-efficacy and past performance are excellent indicators of whether or
not an individual feels that he or she can achieve a certain performance level.
Individual difference variables such as goal orientation and interest are more indicators
of whether or not and individual wants to participate in the Ieaming activity, and if he or
she wants to Ieam. Thus, this study expands the predictors of self-set goals from those
variables found in the traditional expectancy-valence model of motivation to those
found in intrinsic models of motivation (Kanfer, 1990).

The relationships among the goal orientation and interest variables lead us to
greater understanding of the process through which individuals set their own goals.
This, in turn, helps us understand important outcome variables such as exam
performance. Exam scores were generally uncorrelated with state and trait mastery
orientation, state performance orientation, and interest, and were negatively correlated
with trait performance orientation in Waves 1 and 2 (see Tables 2-4). Consistent with
volumes of previous research, goals and self-efficacy were both strongly correlated with
exam scores (e.g., Locke 8 Latham, 1990; Thomas 8 Mathieu, 1994; Mono 8 Baker,
1992b)

Although only trait performance orientation was directly correlated with exam
performance, mastery orientation and interest affected exam performance through their

relationships with goal setting. Mastery orientation and interest were both positively

98
related to exam goal difficulty, which was then strongly related to exam performance.
Both of these variables have indirect effects on exam performance through their
relationship with self-set goals. This relationship is similar to the results of a previous
study which indicated that goal orientation affected test scores through effort and
attention (Fisher 8 Ford, in press). Goal orientation tends to affect Ieaming outcomes
through several motivational processes such as goal setting and attention allocation
rather than directly affecting test scores.

Effects of goal onenddtion on reactions. The second role of state goal
orientation in self-regulation depicted in Figure 1 is the impact on reactions to
performance discrepancies. Three types of reactions were examined in this study:
change in effort, change in goals, and feedback acceptance. It was expected that state
goal orientation would interact with the actual goal-performance discrepancy to predict
individuals’ reactions. Individuals with a high mastery orientation were expected to
react in an adaptive manner, increasing or maintaining effort levels and goal difficulty in
situations of both positive and negative feedback. These individuals were
hypothesized to accept feedback in the negative discrepancy situation, but not
necessarily in the positive discrepancy situation. Individuals with a high performance
orientation were expected to react in a less adaptive manner, striving to protect their
ego at all times by minimizing effort, lowering goals in the face of a negative
discrepancy, and rejecting negative feedback (see Figure 3).

Consistent with the suggestions of Farr, et al., (1993) and Kluger and DiNisi
(1996), reactions to performance discrepancies were indeed related to both state and
trait mastery goal orientation. Participants’ planned increase in effort for the next

section of the course was partially explained by state mastery orientation, and in Wave

99
1, the interaction between state mastery and the discrepancy. Examining the plot of
the interaction (see Figure 7) reveals that those with a high state mastery orientation
planned to increase their effort after a negative performance discrepancy, but not when
they encountered a positive performance discrepancy. This finding supports the
hypotheses depicted in Figure 3.

While trait mastery orientation had no effect on planned increase in effort,
participants with a high state mastery orientation tended to plan an increase in effort for
the next performance episode. Participants’ goal change was primarily accounted for
by the performance discrepancy, but there was a positive, direct effect for state
mastery orientation on goal change between waves 1 and 2. Thus, individuals with a
high state mastery orientation tended to plan an increase in effort and raise their goals
between performance episodes.

Changes in effort and goals were related to exam performance. As discussed
above, goal difficulty was positively associated with exam scores. However, planned
increase in effort was negatively associated with exam scores. Participants who
reported planning to increase their effort following exam 2 tended to score lower on
exam 3 (r = -.24). Thus, even though they planned to increase their effort, their
performance did not improve. This effect could be a result of poor study strategies.
Participants who reported a high degree of workload while studying also tended to
score lower on exams (see Table 6). It appears that these students were applying
effort, but may have misdirected that effort.

Feedback credibility was also affected by goal orientation. State mastery
orientation was positively associated with feedback credibility after waves 1 and 2,

while trait performance orientation had a significant, negative impact on feedback

100
credibility. These findings support previous suggestions (e.g., Kluger 8 DiNisi, 1996)
that individuals with a high performance orientation will find it more difficult to accept
and acknowledge the accuracy of negative performance feedback than those
individuals low in performance orientation, and consequently, with lower ego-protection
needs. Rejecting negative performance feedback can serve an ego-protection
function. Instead of recognizing that their performance was substandard, these
individuals reported that the feedback did not accurately describe their performance.

The moderator analysis predicting feedback credibility perceptions also
supported this ego-protection concept. In Wave 3, the interaction between state
performance orientation and goal-performance discrepancy was significant as
expected. Participants with a high state performance orientation who experienced a
positive discrepancy felt the feedback was credible. Those with a low state
performance orientation perceived the feedback as less credible when they
experienced a positive discrepancy, and were more likely to accept the feedback when
they experienced a negative discrepancy. Thus, when the direction of the discrepancy
was taken into account, it became even more clear that a high performance orientation
was associated with a tendency to protect the ago. A high performance orientation was
associated with rejecting only negative feedback, while individuals with a low
performance orientation felt such feedback was more credible.

Previous research indicates that feedback that is negative and provided by a
source that lacks credibility is least likely to be accepted by the feedback receiver
(Ilgen, Fisher 8 Taylor, 1979). Individuals are unlikely to correct mistakes made in
earlier performance episodes if they do not attend to and accept the negative feedback

provided. Consequently, the current findings conceming feedback credibility suggest

101
that individuals with a high performance orientation are likely to continue engaging in
the maladaptive behavioral patterns identified by Dweck (1986; Dweck 8 Leggett,
1988), as they are most likely to reject feedback.

Ilgen, Fisher and Taylor (1979) were among the first to suggest that individual
differences influence reactions to feedback. Kluger and DiNisi (1996) suggested that
individual differences influence reactions to feedback by directing attention to the self
rating than to improving performance. The results of this study support this aspect of
Kluger and DiNisi’s feedback intervention theory, as individuals with a high
performance orientation tended to reject negative feedback. In attempts to protect their
ego in the short term, individuals with high state performance orientation may overlook
more long-tenn performance improvement strategies, such as changing goals or
increasing effort.

To summarize, state and trait goal orientation did affect self-regulatory
processes in this study. Self-set goals were positively affected by both state and trait
mastery orientation. Participants with a high state mastery orientation tended to display
more adaptive, functional responses to goal-performance discrepancies, as they were
more likely to plan to increase their effort in the next performance episode and increase
their goals. State mastery orientation was also associated with greater acceptance of
feedback, as represented by exam scores. The primary effect of state performance
orientation in this study was a negative association with feedback credibility.
Participants with a high state performance orientation tended to report that exam
scores lower than their goals were actually an inaccurate reflection of their
performance. These results are generally consistent with Dweck’s (1986) notion of the

adaptive response pattern. A strong mastery orientation was associated with

102
responses to goal-performance discrepancies that would lead to improved performance
in the next performance episode.

It was beneficial in this study to include measures of state goal orientation in
addition to trait goal orientation in the examination of self-regulation. State orientation
generally impacted self-regulatory outcomes m statistically accounting for the effects
of trait orientation. Further, trait goal orientation tended to affect outcomes in the first
wave, and state goal orientations had greater effects in waves 2 and 3. As situational
influences affected participants’ state goal orientation, state mastery and state
performance became more influential toward self-regulatory reactions and behaviors.
These findings were made possible by the repeated measurement of state goal
orientation with a tool based on clear definitions of state and trait goal orientation that
was designed according to state measurement guidelines.

Stataand Trait Goal Orientaion

In the past several years, the goal orientation literature has begun to examine
both state and trait conceptualizations of the constnrct. However, researchers have
typically failed to clearly distinguish between state and trail. For example, researchers
have defined goal orientation in a trait-like manner, but then proceeded to manipulate it
(e.g., Elliott 8 Dweck, 1988; Dweck 8 Leggett, 1988). It is important to distinguish
between state and trait goal orientation because these variables will impact human
behavior in different ways. Trait goal orientation is the tendency to respond to a variety
of achievement situations in a similar manner, and is the result of complex factors such
as genetics and upbringing. State goal orientation is the immediate response to a
particular achievement situation, and is the result of an individual’s trait goal orientation

plus situational factors. The failure of many researchers to conceptually clarify the

103
difference between state and trait goal orientation can cause misinterpretation of
research findings, and makes it difficult to interpret the body of goal orientation
research as a whole. Once a researcher distinguishes conceptually between state and
trait goal orientation, the measurement of the variables must clearly follow from the
definition of those variables.

Despite the importance of clearly distinguishing between state and trait,
research attempting to distinguish between state and trait goal orientation has been
somewhat limited by the available measurement tools. The state goal orientation
measure used in the past (e.g., Button, et al., 1996; Boyle 8 Klimoski, 1995) does not
meet several of the standards for instruments designed to measure personality states
(Zuckennan, 1977, 1983; Allen 8 Potkay, 1981). Most importantly, the test-retest
reliability of the previous scale was virtually indistinguishable from the test-retest
reliability of the trait scales (Fisher, et al., 1997). Test-retest reliability is a critical
aspect of the distinction between trait and state (Cronbach 8 Meehl, 1955). Therefore,
this study strove to improve the measurement of state goal orientation.

State and trail goal orientation were both measured in this study to investigate
how goal orientation changes as a result of changing situational influences over time
(see Figure 2). Consistent with interactionist theory of personality (Magnusson 8
Endler, 1977; Mischel 8 Shoda, 1995), goal orientation theory suggests that goal
orientation is a relatively enduring individual characteristic that may be susceptible to
temporary changes due to situational influences (Button, et al., 1996). Up to this point,
research has not examined naturally occurring situations that affect the state aspect of

goal orientation.

104

Measurement issues. This study did provide evidence for an improved
measurement system for goal orientation. A new state scale was constructed in this
study that adheres to guidelines for the measurement of personality states. The items
on each measure were related to the conceptual definitions of mastery and
performance orientation. However, the trait items focused on general tendencies, while
the state items focused on cognitive reactions to specific situations encountered in the
Ieaming environment. The state instrument was not dependent on instructions to orient
individuals to the appropriate frame of mind (Allen 8 Potkay, 1981), rather, the framing
was accomplished at the individual item level. A confirmatory factor analysis supported
the distinction between the state and trail measures of goal orientation, as a two-factor
model combining the state and trait measures did not fit as well as a four factor model
that separated the state and trail measures.

These improvements in the state goal orientation scales resulted in better
psychometric qualities than found with previous scales. Test-retest statistics
demonstrated greater consistency over time for the trait measures than the state
measures. Previous studies have found similar, and occasionally identical, test-retest
coefficients for state and trail scales of goal orientation (Fisher, et al., 1997). In
addition, lntemal consistency estimates were higher for the state scales developed in
this study than demonstrated for previous state scales. Consequently, the state scales
used in this study represent an important improvement in the measurement of state
goal orientation, thus allowing more accurate comparisons between state and trait.

ConstruLct issues. The improved state goal orientation scales facilitated the
examination of how goal orientation functions over time. As shown in Figure 2, state

goal orientation was expected to change across episodes of the self-regulation process

105
as a result of participants’ changing interest in the course. State mastery orientation
did change over time, while state performance orientation was more stable. The test-
retest estimates for state performance orientation were consistently higher than the
estimates for state mastery orientation.

Much of the expected fluctuation in state mastery orientation appears due to the
strong relationship between state mastery orientation and interest. As interest changed
between waves, as a result of perfonnance—goal discrepancies, state mastery
orientation changed as well. Surprisingly, the effects of interest on state mastery
orientation were primarily direct effects. The hypothesized interaction between trait
mastery orientation and interest was not significant in the first two waves, even though
both trait mastery orientation and interest exerted strong, direct effects on state
mastery orientation.

This finding contradicts interactionist theory which suggests that situational
influences such as course-specific interest will be more likely to affect the state if the
trait is weak (Magnusson, 1977). These results suggest that some situational
influences, such as course-specific interest, affect state goal orientation even when trait
goal orientation is strong. This result does not suggest that behavior is strictly specific
to the situation, but rather suggests that the situation can affect all individuals in a
similar manner. These results suggest that manipulations of goal orientation could be
expected to affect the goal orientation of all subjects. The results also emphasize the
need to measure state goal orientation in the presence of any situational influence
expected to affect goal orientation. Even if all individuals have a high mastery
orientation, for example, their specific reaction to the situation will be affected by the

features of the situation.

106

State performance orientation was also related to course-specific interest, but to
a lesser degree. Changes in state performance orientation across waves were
primarily accounted for by changes in academic self-efficacy. As demonstrated by the
data supporting Hypothesis 9, large perfonnance-goal discrepancies predicted lower
academic self-efficacy in the next wave. These changes were reflected in state
performance orientation, although, as noted above, state goal orientation was more
stable than state mastery orientation.

The fluctuations in state goal orientation demonstrate the importance of
measuring goal orientation at the appropriate time in a research study. Researchers
may measure individuals’ traits at the beginning of a study and expect that
measurement to affect behavior throughout the study. If participants receive feedback
repeatedly over several experimental trials, the results of this study suggest that
individuals’ state goal orientation will change, and begin to impact behavior. In this kind
of study, the researcher should consider examining state goal orientation at several
points in the study.

It is important to remember that the current study investigated changes in state
goal orientation across time, but within the same domain. One criticism of the goal
orientation literature is that the bulk of the research has been conducted within
laboratory and academic settings. The degree to which goal orientation might be
consistent within a person but across domains is unknown (VandeWalle, in press). It
certainly appears likely that as an individual’s interest varies across different activities,
state goal orientation would also vary across activities. One person may have low state
performance orientation for college psychology courses, but have extremely high state

performance orientation for golf or tennis. Thus, even though state performance

107
orientation was rather stable in this sample, it may be less stable across domains,
particularly outside of the educational domain.

The data consistently demonstrated a positive relationship between state
mastery orientation and academic self-efficacy. This relationship was not
hypothesized. However, even after accounting for both trait mastery and state mastery
orientations in the previous wave, academic self-efficacy was a robust predictor of state
mastery orientation. Alternatively, Kozlowski, et al., (1995) found that both trait mastery
orientation and a mastery goal manipulation positively affected self-efficacy in a self-
directed Ieaming experiment. In the current study, academic self-efficacy also had a
positive relationship with state performance orientation. Thus, individuals who were
confident in their ability to Ieam the course material tended to have both high state
mastery and performance orientations. Future work should continue to examine the
relationship between self-efficacy and goal orientation.

Overall, the results supported fewer of the hypothesized relationships involving
performance orientation than those involving mastery orientation. One potential
explanation for this lack of results is that performance orientation may actually consist
of two dimensions (e.g., Elliot 8 Church, 1997; VandeWalle, in press). Elliot and
Church suggested that performance orientation contains both approach and avoidance
motivations. The approach motivation is represented by the desire to do better than
others, while the avoidance motivation is represented by the fear of doing poorly.
VandeWalle named these dimensions perfonnance—prove and perfonnance-avoid. An
exploratory factor analysis (Elliot 8 Church, 1997) and a confirmatory factor analysis

(VandeWalle, in press) statistically supported the division of performance orientation

108
into these two factors. If performance orientation should indeed be treated as two
separate factors, combining these factors could mask results.

The performance orientation scale used in this study was not separated into
prove and avoid subscales because the items in the Button et al (1996) scale were not
written to tap these two subdimensions of performance orientation. The items cannot
be cleanly divided into the two subscales. Many of the Button et al items refer to
lntemal motivations rather than focusing on other individuals. For example, the items
refer to the happiness and enjoyment derived from performing well rather than a desire
to prove one’s ability to others. Thus, this potential explanation for non-significant
results regarding performance orientation could not be directly tested.

A broader concern with the two-factor conceptualization of performance
orientation is the measurement system used by proponents of that view. The scales
appear to be inconsistent with goal orientation theory. Many of the items on these
scales involve the choice of tasks. For example, VandeWalle’s scale asks respondents
if they would choose to avoid engaging in certain kinds of activities. Measuring a
motivational tendency in this way is misleading, as it focuses on the outcome of the
construct (task choice) instead of the construct itself (Fisher, et al., 1997). Using task
choice as an indicator of goal orientation also reinforces a bipolar view of goal
orientation. If one chooses to avoid certain tasks, one cannot also choose to prove
his/her ability to others on those tasks. The use of task choice as a measurement tool
promotes the finding of distinct factors, but is inconsistent with the current view that the
different “parts” of goal orientation are independent but related concepts. Thus,
although performance orientation may indeed have two distinct subscales, existing

measures do not allow researchers to adequately determine differences between them.

109

Survey Administaation Mada

Another contribution of this study was the investigation of alternative survey
administration modes. Traditional paper-and-pencil surveys were used, as well as
electronic versions of the same surveys administered using the student e-mail system.
Response rates for the two modes were similar, although slightly higher for the e-mail
respondents. Over three waves, the response rate for e-mail respondents was 83%,
while the response rate for paper-and-pencil surveys was 76%. In addition, the surveys
appeared to be roughly equivalent across the administration modes. Differences in
means and standard deviations between the two groups were minimal. The only
significant difference in group means was on the trait mastery orientation scale in the
first wave, with the e-mail group having a higher mean. This difference may reflect a
willingness by the individuals with a high mastery orientation to Ieam and use new
concepts, including technology. The use of the non-traditional survey administration
mode improved the three-wave response rate, without impacting results. However, it
must be noted that the improved response rate cannot be wholly attributed to
administration mode, as participants were not randomly assigned to e-mail and paper-
and-pencil groups. It is possible that the participants who were more dedicated to the
study and were more likely to respond tended to choose the e-mail mode.
Study Limitations

This study did have several limitations that require caution in interpreting and
generalizing the results. First, it is difficult to disentangle measurement issues from
construct validation issues. If we find that test-retest for the trait scale is poor, does this
mean that there is not a trait by that name, or does it mean that our instrument is poor?

Questions such as this cannot be answered in the current study. A second

1 10

measurement limitation of the study is that the state goal orientation measures were
developed for this study, and their psychometric properties were previously untested.
Three of the twelve state goal orientation items were deemed of poor quality and were
discarded, resulting in shorter scales than desired.

An additional limitation of the study was the variability in the participant sample.
While inclusion of a broader range of students could be viewed as a strength of the
study, allowing greater generalizability, this variability was a potential threat to lntemal
validity. The students who participated in the study came from six different psychology
courses. Differences across these courses, such as difficulty of the course, the slightly
different timing of exams, and the inclusion of different grading mechanisms such as
homework and written projects in addition to the exams could all have affected the
results of the study. Furthermore, year in college was associated with course.
Students who were enrolled in the statistics and measurement courses tended to be
upper level students, while those enrolled in the introductory course tended to be first
or second year students.
Implications

First, this study suggests that goal orientation is indeed an individual difference
variable that plays a role in the self-regulatory process. Of particular interest are the
effects of goal orientation on reactions to goal-performance discrepancies. State
mastery orientation predicted planned increases in effort, increases in goal difficulty,
and realistic acceptance of performance feedback. Thus, it appears that educators and
trainers could assist students and trainees in reacting positively to performance

discrepancies by creating situations that promote a state mastery orientation.

1 1 1

The strong, positive relationship between interest and state mastery orientation
has implications for the role of interest in organizations. Interest has traditionally played
a role in career counseling efforts, as represented by the use of the Strong interest
Inventory (Sll) and the Kuder Occupational Interest Survey (Dawis, 1991). Scores on
the SII have been demonstrated to predict occupational membership, and to a lesser
extent, job performance. The link between interest and state mastery orientation
reinforces the use of interest in career selection, as people with a high mastery
orientation may be more successful, as they may be more likely to Ieam and develop
throughout their careers (Min/is 8 Hall, 1994).

Further, this study suggests a new mechanism for managers for promoting a
mastery orientation in their employees. Given the strong, positive relationship between
interest and state mastery orientation, perhaps managers in organizations should
consider employees” interests before distributing work assignments. The results of this
study and others suggest that employees who are highly interested in their work will
have higher state mastery orientation, and thus may devote greater effort to the work
(Fisher 8 Ford, in press), be more willing to seek feedback on their work (VandeWaIle
8 Cummings, 1997), and generalize skills Ieamed in training to the workplace
(Kozlowski, et al., 1995). Few studies of goal orientation have actually been conducted
in work organizations. Much work is needed in this area to determine if the beneficial
effects of mastery orientation discovered in the laboratory and the classroom will
generalize to the workplace.

The results of this study also have implications for the further study of goal
orientation. Dweck (1989) suggested there was an important difference between

manipulating an individual’s goal orientation (trait), and manipulating the salience and

1 12
value of different types of goals (state). The actual goal orientation, or trait, is relatively
stable, and changes little in the short term. The salience and value of different types of
goals, or state, can be affected by environmental factors. Researchers must remember
that it is not the nominal presence of environmental factors, such as goal
manipulations, that change state goal orientation (Shoda 8 Mischel, 1995), but rather
the individual’s interpretation of the environment in light of his or her experiences.
Thus, all individuals will not respond to environmental influences in the same manner,
or to the same degree. For example, course specific interest is an integration of the
features of the situation with the individual’s interpretation of the situation in light of
his/her preferences and previous experiences. There is no such thing as an objectively
interesting situation. Similarly, not all individuals will find the same situation motivating,
or find the same goals appealing.

The results of this study support Dweck’s view concerning the difference
between state and trail goal orientation, and suggest that both state and trait concepts
are important to consider in the explanation and prediction of goal directed behavior.
Several of the hypothesized relationships were supported for state mastery orientation,
but not for trait mastery orientation. Had state mastery orientation not been examined,
the role of goal orientation in self-regulation would have appeared much smaller and
less interesting.

W

In summary, this study has demonstrated several points about the role of goal
orientation in self-regulation, as well as the measurement properties of goal orientation.
Goal orientation, both trait and state, affected goal setting and reactions to goal-

performance discrepancies. State goal orientation was distinguished from trait goal

1 13
orientation in a theoretically dr'wen measurement model. Trait and state goal
orientation were found to have different effects on self-regulatory behaviors. This
evidence, as well as the finding that interest in the course affected state mastery
orientation over time, supports the theoretical position that goal orientation can be
fruitfully conceptualized and operationalized as b_ot_h a state and a trait. Interest
appears to be a significant factor in state goal orientation, and may be a key concept in
understanding changes in state goal orientation over time and across situations.
Future research examining the role of goal orientation in goal directed, achievement
related behavior should carefully consider both the environmental and the person-

driven aspects of goal orientation.

APPENDICES

114

APPENDIX A

Survey #1

Your responses to all questionnaires in this research will be kept confidential, and will
be seen only by the experimenters. Your instructor (professor and/or teaching
assistants) will not see any of your responses to any of the questionnaires used in this

research.

MSU Student ID #:

 

Please write your answer to each of the items in the blank to the left of the item.

1.

 

 

 

 

 

 

Sex: 1 = Male 2 = Female
Year in College (1-5)

How many other college level psychology courses have you taken (do not
count courses you are currently taking)?

How many college level courses in math or statistics have you taken,
including psychology statistics courses?

What is your current GPA (if this is your first semester, skip this question)?

What was your GPA last semester (if this is your first semester, please give
your high school GPA)?

What were your SAT scores? Math: Verbal: Total:
(If you did not take the SAT, please report your total ACT score.)

Why are you taking your current psychology class (write letter for all
reasons that apply):

it fulfills a requirement

the content seemed interesting

it will be useful to me in other courses

it is an easy elective

it was recommended by a friend or counselor

it fit into my schedule

What is the minimum acceptable grade for you for the first exam in this
class (on 4.0 scale)?

. What is the minimum acceptable grade for you for the entire course (on 4.0

scale)?

11.

12.

1 15
What is the grade you are actually trying for on the next exam in this class
(on 4.0 scale)?

What is the grade you are actually trying for in the entire course (on 4.0
scale)?

Please use this scale to answer the following items. Write the number of the response
that best represents your opinion in the blank to the left of the item.

1 9 a . 5

Strongly Disagree Neutral Agree Strongly
Disagree Agree

13.
14.
15.
16.

 

 

21.
22.
23.

 

 

 

24.

 

25.
26.
27.

 

 

28.
29.

 

17.
18.
19.
20.

The opportunity to do challenging work is important to me.

I do my best when I am working on a fairly difficult task.

I try hard to improve on my past performance.

Men I have difficulty solving a problem, I enjoy trying different approaches
to see which one will work.

The opportunity to Ieam new things is important to me.

The opportunity to extend the range of my abilities is important to me.

I prefer to work on tasks that force me to Ieam new things.

When I fail to complete a difficult task, I plan to try harder the next time I
work on it.

The things I enjoy the most are the things I do the best.

I feel smart when I can do something better than most other people.

I like to be fairly confident that I can successfully perform a task before I
attempt it.

I am happiest when I perform tasks on which I know that I won't make any
errors.

I feel smart when I do something without making any mistakes.

I prefer to do things that I can do well rather than things that I do poorly.
The opinions others have about how well I can do certain things are
important to me.

I like to work on tasks that I have done well on in the past.

If I can’t do something the first time, I keep trying until I can.

30.
31.
32.
33.

 

35.
36.

 

 

 

 

 

 

43.

 

 

46.
47.

 

48.
49.
50.

 

51.
52.
53.

 

 

 

,3!

116

When I set important goals for myself, I rarely achieve them.

I avoid facing difficulties.

When I have something unpleasant to do, I stick with it until I am finished.
When I decide to do something, I go right to work on it.

. When trying to Ieam something new, I soon give up if lam not initially
successful.
I rely on myself to accomplish my goals.

I do not seem capable of dealing with most problems that come up in my
life.

. I find this course very interesting.
38.
39.
40.
41.
42.

I think the course material in this class is useful for me to Ieam.

I like the subject matter of this course.

I am very interested in the content area of this course.

It is more important to really Ieam the material than to get the grade I want.
It is OK if I don’t know much about the material at the end of the class as
long as I get the grade I want.

It is very important to me to attain the grade goal I have set for the next

exam.

. It is important to me to Ieam a lot in the next section of the course.
45.

I really want to increase the amount I know about the material in this
course.

I am confident I can understand the basic concepts taught in this course.
I am confident I can do an excellent job on the assignments and tests in this
course.

I am confident that I will do well in this class.

lam certain I can master the skills being taught in this class.

lam certain I can understand the most difficult material presented in the
readings for this course.

I am eager to get started Ieaming the material in this section.

I hope I don't ask any stupid questions during this part of the class.

I intend to learn as much as I can in this part of the class.

I am eager to prove to others how good I am at the content of this class.

 

55.
. I look fonrvarid to mastering the challenging material in this part of the course.
57.
58.
59.
60.
. l have little desire to learn new things in this part of the class.
62.

 

 

 

 

 

 

117

I want to really understand the material in this section.

I wonder how my score on the next exam will compare with others.

I am eager to show how much I know about the material in this part of the class.
I want to appear competent in this class.

I want to do better than others on the next exam.

If I don’t understand the material in this section right away, I will keep trying until
I do understand it.

118

APPENDIX B

Survey 2

Your responses to all questionnaires in this research will be kept confidential, and will

be seen only by the experimenters. Your instructor (professor and/or teaching

assistants) will not see any of your responses to any of the questionnaires used in this

research.

MSU Student ID #:

Please use this scale to answer the following items. Write the number of the response

that best represents your opinion in the blank to the left of the item.

1 2 3 4 5

 

Strongly Disagree Neutral Agree Strongly
Disagree Agree

10.

I wish I had Ieamed more in the previous section of the course.
In the last section of the course, I increased the amount I know about the
material in this course.
Regardless of my exam performance, I feel that l Ieamed a lot in the previous
section of the class.

I feel that my exam score is an accurate evaluation of my performance.
My exam grade reflected my true performance in the course.

I feel the instructor of this course is very knowledgeable about the course
content.
The instructor of this course is qualified to evaluate my performance in this
class.

How many hours do you plan to study between now and the next exam?

I plan to study more for the next exam than I did for the last exam.

In general, I plan to work harder in the next part of this course than I did in
the last part.

119

__11. The opportunity to do challenging work is important to me.

__12. I do my best when I am working on a fairly difficult task.

_13. I try hard to improve on my past performance.

__14. When I have difficulty solving a problem, I enjoy trying different approaches

to see which one will work.

_15. The opportunity to Ieam new things is important to me.

_16. The opportunity to extend the range of my abilities is important to me.
_17. I prefer to work on tasks that force me to Ieam new things.

____18. When I fail to complete a difficult task, I plan to try harder the next time I

work on it.

19. The things I enjoy the most are the things I do the best.
20. I feel smart when I can do something better than most other people.
21. I like to be fairly confident that I can successfully perform a task before I

attempt it.

22. I am happiest when I perform tasks on which I know that I won't make any

errors.

23. I feel smart when I do something without making any mistakes.
24. I prefer to do things that I can do well rather than things that I do poorly.
25. The Opinions others have about how well I can do certain things are

26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.

 

 

 

 

lllll

important to me.

I like to work on tasks that l have done well on in the past.

I find this course very interesting.

I think the course material in this class is useful for me to Ieam.

I like the subject matter of this course.

I am very interested in the content area of this course.

I am eager to get started Ieaming the material in this section.

I hope I don‘t ask any stupid questions during this part of the class.

I intend to Ieam as much as I can in this part of the class.

I am eager to prove to others how good I am at the content of this class.
I want to really understand the material in this section.

I look fonivard to mastering the challenging material in this part of the course.

I wonder how my score on the next exam will compare with others.

120

I am eager to show how much I know about the material in this part of the

class.

. I want to appear competent in this class.

. I want to do better than others on the next exam.

I have little desire to Ieam new things in this part of the class.

. If I don’t understand the material in this section right away, I will keep trying

until I do understand it.

. I am confident I can understand the basic concepts taught in this course.

. I am confident I can do an excellent job on the assignments and tests in this

course.

. I am confident that I will do well in this class.
. lam certain I can master the skills being taught in this class.

. I am certain I can understand the most difficult material presented in the

readings for this course.

. What is the minimum acceptable grade for you for the next exam in this

class (on 4.0 scale)?

What is the minimum acceptable grade for you for the entire course (on 4.0
scale)?

What is the grade you are actually trying for on the next exam in this class
(on 4.0 scale)?

What is the grade you are actually trying for in the entire course (on 4.0
scale)?

It is important to me to Ieam a lot in the next section of the course.

. I plan to Ieam more in this next section than I did in the last section.

. I really want to increase the amount I know about the material in this course.
. It is more important to really Ieam the material than to get the grade I want.

. It’s OK if I don't know much about the material at the end of the class as

long as I get the grade I want.

. It is very important to me to attain the grade goal I have set for the next

exam.

FERN?"

121

APPENDIX C

Course-specific Interest Scale
(Pintrich, et al., 1991)

I find this course very interesting.
I think the course material in this class is useful for me to Ieam.
I like the subject matter of this course.

I am very interested in the content area of this course.

9959.”?

122

APPENDIX D

General Self-efficacy Scale

If I can’t do something the first time, I keep trying until I can.

When I set important goals for myself, I rarely achieve them.

I avoid facing difficulties.

When I have something unpleasant to do, I stick with it until lam finished.

When I decide to do something, I go right to work on it.

When trying to Ieam something new, I soon give up if I am not initially successful.
I rely on myself to accomplish my goals.

I do not seem capable of dealing with most problems that come up in my life.

Academic Self-efficacy Scale
(Pintrich, et al., 1991)

I am confident I can understand the basic concepts taught in this course.
I am confident I can do an excellent job on the assignments and tests in this course.
I am confident that I will do well in this class.

lam certain I can master the skills being taught in this class.

lam certain I can understand the most difficult material presented in the readings

for this course.

123

APPENDIX E

Trait Goal Orientation Scales
(Button, Mathieu 8 Zajac, 1996)

Trait Mastery Orientation

1.

99".”

.°°.“.°’.°‘

The opportunity to do challenging work is important to me.

I do my best when I'm working on a fairly difficult task.

I try hard to improve on my past performance.

When I have difficulty solving a problem, I enjoy trying different approaches to see
which one will work.

The opportunity to Ieam new things is important to me.

The opportunity to extend the range of my abilities is important to me.

I prefer to work on tasks that force me to Ieam new things.

When I fail to complete a difficult task, I plan to try harder the next time I work on it.

flail Performance Orientation

1.

S”

The things I enjoy the most are the things I do the best.

2. I feel smart when I can do something better than most other people.
3.
4

. I am happiest at work when I perform tasks on which I know that I won't make any

I like to be fairly confident that I can successfully perform a task before I attempt it.

errors.

I feel smart when I do something without making any mistakes.

6. I prefer to do things that I can do well rather than things that I do poorly.

The opinions others have about how well I can do certain things are important to
me.

I like to work on tasks that l have done well on in the past.

124

APPENDIX F
State Goal Orientation Scale

1. I am eager to get started Ieaming the material in this section. (M)*

2 I hope I don't ask any stupid questions during this part of the class. (P)*

3 I intend to Ieam as much as I can in this part of the class. (M)

4 I am eager to prove to others how good I am at the content of this class. (P)

5. I want to really understand the material in this section. (M)

6 I look forward to mastering the challenging material in this part of the course. (M)

7 I wonder how my score on the next exam will compare with others. (P)

8 I am eager to show how much I know about the material in this part of the class. (P)

9 I want to appear competent in this class. (P)

10. I want to do better than others on the next exam. (P)

11. l have little desire to Ieam new things in this part of the class. (M)*

12. If I don’t understand the material in this section right away, I will keep trying until I
do understand it. (M)

Note: M = mastery orientation, P = performance orientation. Items marked with an

asterisk (*) were deleted from the final version of the scales.

125

APPENDIX G

Self-set Goal Scales

Self-set grade goals (Locke 8 Bryan, 1968)

PPN.‘

What is the minimum acceptable grade for you for the next exam?
What is the minimum acceptable grade for you for the entire course?
What is the grade you are trying for on the next exam?

What is the grade you are trying for in the entire course?

Self-set learning goals (Mono 8 Baker, 1992)

1.
2.
3.

It is important to me to Ieam a lot in the next section of the course.
I plan to Ieam more in this next section than I did in the last section.

I really want to increase the amount I know about the material in this course.

Goal Valence

. It is more important to really Ieam the material than to get the grade I want.

It’s OK if I don‘t know much about the material at the end of the class as long as I
get the grade I want.

It is very important to me to attain the grade goal I have set for the next exam.

Subjective learning perceptions

1.
2.

I wish I had Ieamed more in the previous section of the course.

In the last section of the course, I increased the amount I know about the material in
this course.

Regardless of my exam performance, I feel that l Ieamed a lot in the previous

section of the class.

126
APPENDIX H

Discrepancy Reaction Scales

Time Spent (Campion and Lord, 1982)

1. Estimate the number of hours you studied for this class during the last section.
How many class sessions did you miss during the last section of the course?
How many hours do you plan to study between now and the next exam?

I plan to study more for the next exam than I did for the last exam.

9:59.”

In general, I plan to work harder in the next part of this course than I did in the last
part.

Mental Workload (Fisher, 1995)
6. I had to work very hard to Ieam the material for the last test.
7. Leaming the material for this class was easy.

8. Learning the material for the previous section required a lot of mental activity.

Feedback Accuracy (Podsakoff 8 Farh, 1987)
9. I feel that my exam score is an accurate evaluation of my performance.

10. My exam grade reflected my true performance in the course.

Feedback Source Credibility (Podsakoff 8 Farh, 1987)
11. I feel the instructor of this course is very knowledgeable about the course content.

12. The instructor of this course is qualified to evaluate my performance in this class.

127
APPENDIX I
Table 1
Factor Loadings for Confinnatorv Factor Model of Trail and State Goal Orientation:

Epdr-Factor Model: Wave 1 Data

 

 

Unstandardized
Sca_|e
Item Factor Loading g
Trait Mastepy
mm .499 .039
TMOZ .377 .047
TMO3 .307 .035
TMO4 .449 .047
TMO5 .356 .032
TMO6 .302 .027
TMO7 .440 .036
TM08 .396 .035
Trait Performance
TPO1 .368 .047
TPOZ .503 .040
TPO3 .418 .045
TPO4 .585 .049
TP05 .548 .039
TPO6 .542 .041
TP08 .412 .034
State Mastem
State3 .554 .033
State5 .531 .033
State6 .593 .044
State12 .366 .032
State Performance
State4 .793 .046
State7 .464 .050
State8 .655 .039
State9 .449 .039
State10 .504 .047

 

Note. it = 434.

128

.3. v m a 285ch 2.855e- 95 2. N 825980
$53508 00.500 n 00.0 00-50w 60:00:05 8:950:08 :0: n =8... .E0t0n_
.:0=0E0:0 00:0.E0t0: 0:80 u 23w .E0t0n_ 005000.: E0: 0:0 00 0005 00 028w F onm 0:0 .600 _- onw .300 00.500
:0. 520 .0: 0:0 mofiEzmm ._0:000_o 05 :_ 02:000.: 0:0 00fiE=m0 3:206:00 _0E0E_ 0:20 2065000 .vmmu: .202

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I 5. m? 8.- 8. o..- 8. 0.... 8. 8. B. 8. 88 <8 2839‘ .NF
83 2. 8. 8.- 8.- 3..- 8.- 9.- 9.- 9.- 8.- E. 00.8 883280.:
65 em. 3.- 8. 5. 8.- 8.- B.- u? 8.- em. «3. get-toad
63 0.. 8. cm. 8.- 9. 3. 5. mm. 8. v0...” anew-toe .0
at 00. 0F. 8. 8.. 8.. 8. we. 0.... :4. gap-bones. .0
No: on. 8. 0e. 8. 8. 0.... E. 8.... anew-cones. :-
83 8.- we. 2. 2. 8. «0. 6.0 access .0
I 8. on. 3. «N. N: 8.8 280 r $an .0
I k. 3. «a. 3. who .80 P sea .4
I 0.. 5. 8. who .80 8:50 .0
85 8. 9.. 5a 6850 .883 .N
63 E. E.» 8850-80 a
9 : 2 0 0 e 0 0 e m N F 00 _2

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3030:0225 F 0>0>>

N 030..-

129

.3. v m 0 30055.0 2.0020320 0:0 0 F. N 30:00:00

.N 00:0.0> .000 n N.0>.0 .N .000 05500. n N.0:.... $55520 00500 n N 00.0 .:0=2:0:0 3:030:00 :0: u On...-

60:25.8 00:02:30.. 220 .r. N 000 60:030.:0 3200.: :0: u 0.2... 60:030.:0 3200.: 220 u N 0.20 .2000 N 803 n N w
..000 N onm n .000 Nw ..000 00.300 n .000 0 .00-5000.: E2. 0:0 0.0 0005 00 .2000 N E0xm. 0:0 ._000 N E0xm. ._000 00500
:2 :020 8: 0.0 0208:0m. .3000? 05 :_ 02:002.. 00 023500 0:20.300 .0E2:_ 0:0.0 30.05000 .000": .202

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

00.. 50.. m.... h... 9... or... 00. Nov 2.. o0. 0N. 0N. t. 00.0 F.0N ...0< .3.
60.. 00.. oo. VNr VN. m... 00.- NN... N0.- oor 0...- w...- 9..- hm. 00.N N.0>.0 .0.-
350 N0. 00. ..o. 0N. N. 00. ..v. 0...- 00. 00. VN. 00. No;‘ N.0Er. .NF
6mg 5. 3. mo. 3. 0N. 00. Nr 0...- 0...- mo.- tr. 00.». NtmEn. .:

65% N0.. 0...- no. N... ..0. 2. 0N. 0N. VN. cm. 50.0 N620 .0
8kg 0N. v0... Nor V...- t... 0...- 0...- 0...- 00. N: On...- .m
65.0 no. ..0. mo. 00.. 00. 50. NF. 00. No.0 N Cam .0
65% m0. 0N. 00. Nr. 2.. N0. 8. 520 05.... K
SD 00. No. VN. 0N. NV. 50. v0.0 N020 .0
300 00. 0N. RN. 3. 00. 00.0 N #82:. .m
I 00. N0. VN. 0F... 00; Nm .v
it 05. 00. mm. 5.0 .000 Nm .0
I NV. .0. ~00 .000 0 .N
4.04 05. N00 N “ms—.00 ._.

0? NF 3 0.. m 0 h 0 0 v 0 N .. am 5.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

30:90:00.2... N 0..>r0>>

0 0.00...

 

130

.5. v m 0 83:80.0 2.00.0020 0.0 0... N 08.00.0800
.>....0.00.0 00500 n 0 .020 80:080.:0 00:08:00.2. ..0... u On... 80.008000
00:0:80000 220 n 0000 80.208000 02008 ..0: u 0...: 80.008000 52008 220 u 00.20 .0500 0 828 u 0.000 00

..000 0 80x0 u .000 mm ..000 00500 n .000 0 .0050008 80.. 0:0 0:0 000:. 00 .2000 0 80x0 0:0 ..000 0 80.0 ..000 00500
..0. :0>.0 5: 0.0 0208.00 ..0:000.0 05 :. 0080005 00 0208200 0:200:00 .0808. 0.5.0 80.05000 .000": .202

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0o.- 50.- 0T 5?. 0...- 9.- 00. No.- ..... o0. 0N. 0N. 5... 0Y0 ...0N ._.0< .3.
A00. mar 50.- 08- 0... 5o. 5...- 3r 0v.- 3.- 08- 5...- 0N.- 50. 00.N 0.020 0..
.00.. 00. 0N. Nor N0. 5. 00. 00. No.- 0... 0... ..v. N5. 00.0 0.0:... .N..
30.. 0... Po. 0F. oo. 00. 0F. 58. 00... 0o.- #0. 0. 0.20 0tmc0_n. .3
2.0.. ..0.- No. 50. 0p. 0F. 00. 0N. ..N. V0. 0. 05.0 0.00:0 .0?
A05. 0N. .0.- No. 00.- No. 08- .1..- 08- #0. N80 On; .0
.05.. 5o. 00. 0N. o... 0... v... 0... 50. 00.0 0000 .0
A05. 0N. 0... or. or. 2.. 0N. 0.... .....v 0.2... .5
403 N0. 0F. FN. N. 00. 00. 00.0 00.20 .0
30.. 0... .VN. 5N. 0*. 00. 50.0 00.0.0. .0
I 00. 00. 5N. 0F... 0v.N 00 .v
-I 05. 00. 00. ..00 .000 00 .0
l- 9. 00. N00 .000 0 .N
a. ..0. 00.0 00.9.00 .—
05 N.. E. or 0 0 5 0 0 v 0 N .. 00 .2

 

 

03.00.088.08. 0 0>M>H

v 0.00.5

 

131

 

..0:000.0 05 :. 0080005 0:0 0208000 2500.0: >0:20.0:00 .0808. .000": .202

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

00.. E. .0. 0o. 8.- 8.- 0 .0: 8:05.200 .0
.00.. on. 2.- 00.- S.- N .0: 8:05.280 .0
.05. 2.- 2.- 3.- F .0: 8:05.200 .0.
000 8. 00. 0 .0; 0200.2 .0
at 00. N .0; 0200.2 .0
.05. F ..0:- 000.0: .F
0 0 v 0 0 F
0020.00.00 .000 «.00-_-
85. no. 00. 00. .0... E. 0 20.0 8:080:00 .0
.05. 00. 0o. .0. 0.. 0 80.0 8:08.800 .0
.05. B. .0. 00. v 20.0 8:056:00 .0.
.05. 00. 0... 0 20.0 0200.2 .0
35. 00. N 20.0 0200.2 .N
.05. F 80.0 02005. .F
0 0 v 0 N P

 

 

 

 

 

 

 

 

 

8.0.5.8 _000 20.0

 

 

 

0%00 :0..080.:0 .00 .0. 0208000 ...00._0. 0020300...

0 0.00...

132

..Hh

.3. v m .0 E0055? 2.0020300 0.0 3. N 0.5.3.230 .82. 0:0 0.0 00.n0.:0> .050 ..0 600.0..o>> .o. 00000.0 05 E
020000... 0.0 020E=00 5:20.008 .0520. 030.0 E00E000 .0 0>0>> .o. 03 0:0 .N 0>0>> .0. SN .0 0>0>> .o. 00w u c ”202

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

l.- mo. Nv. 3.- N? no.- ON.- ..N... 3... ON. mo. no. 2... mvN a 0.8m mem .NF
I. mm. D... 2.- :r 00... ON.- 9..- N... 3. no. mwé mmé No.8w mew .2.
I. 3.- m...- mNr «UN. 9... NN.- mo. mo. no. Nwé mNN F 0.8m 593 d_-
CNQ mm. mm. 9... mo.. m...- 3. or. mo. an. no.0 2.00.0.5; .m
Ace-v 3.. mo.- mor no.- N_.. 3.. 3. mm. vwd Nuwo_x.o>> .m
3mg 00. we. ac. to. oo. 3.. vw. we.” Puma—v.3; N
I.- on. we. 2... mar nor va 3N 90000.2 .0
I. #0. oo... nor 3.. oNN hwé N0000=2 .m
I. NV- 0.... ho.. mm... om. 3000.2 .v
I.- mm. on. 3. ..F mN.mF ”0.30... .m
I E. NENF mvdw N050... .N
.l VN.N_. no.3. 00.30.... ._.
NF 3 or m m h o n v m N _. 0w 5.

 

 

 

 

 

.0co_.M_0..oo 0.80 :58 0:0 tot0 0>m>!>.:00.s.0m

o 030...

1 33
Table 7
Regression resglts forfivpothesis 1: State Mastery orientation

DV = State Mastery Wave 1

 

 

Step # IV a. R AR2
1 .41** .42 .1 8“
Gen. Efficacy -.13*
ACT
2 TMO .39“ .53 .1 1**
3 Acad. Efficacy1 .26“ .60 .05“

 

DV = State Mastery Wave 2

 

 

Step # IV rs R AR2
1 .34 .11“
Gen. Efficacy .31**
ACT -.15*
2 TMO .34" .44 .03"
3 SMO1 .47“ .60 .16**
4 Acad. Efficacy2 .24" .63 .05"

 

DV = State Mastery Wave 3

 

 

Step # IV a. R AR’
1 .27 .07“
Gen. Efficacy .27“
ACT -.04
2 TMO .24“ .34 .04“
3 SMOZ .46“ .54 .17**
4 Acad. Efficacy3 .30“ .60 .07“

 

N_ot§. All beta weights are from the step at which the variable was entered; n=354.
Acad. Efficacy - academic self-efficacy, TMO = trait mastery orientation, SMO = state
mastery orientation.

*p <.05. **p <.01.

134
Table 8
Regression rmlts for mythesis 1: Stat_3 Performa_nce orientation

DV = State Performance Wave 1

 

Step at N rs R AR2

 

1 .26 .07”
Acad. Efficacy1 .27"
Gen. Efficacy -.06
ACT -.10

2 Trait Perform. .36” .44 .13“

 

DV = State Performance Wave 2

 

Step # IV r:- R AR2

 

1 .16 .02“
Acad. Efficacy2 .14“
Gen. Efficacy -.06

ACT -.10
2 TPO .28“ .32 .08“
3 SPO1 .61” .64 .31M

 

DV = State Performance Wave 3

 

Step # IV rs R AR2

 

1 .24 .06“
Acad. Efficacy3 .23“
Gen. Efficacy -.05

ACT -.14*
2 TPO .28“ .37 .08“
3 SP02 .62M .70 .351W

 

m. All beta weights are from the step at which the variable was entered; n=354.
Acad. Efficacy = Academic self-efficacy, TPO = trait performance orientation, SPO =
state performance.

*p <.05. “p <.01.

 

135
Table 9
Regression Resglts for Hypothesis 2

DV = State Mastery Wave 1

 

 

 

 

 

Step # N rs R AR2
1 .42 .18“
Gen. Efficacy .41**
ACT -.13*
2 Trait Mastery .39” .53 .1 1**
3 Acad. Efficacy1 .26“ .60 .05“
4 lnterest1 .38” .68 .13“
5 Trait mastery X
interest1 .1 1 .68 .00
DV = State Mastery Wave 2
Step # lV rs R AR2
1 .34 .1 1**
Gen. Efficacy .31**
ACT -.15*
2 Trait Mastery .34“ .44 .08"
3 State Mastery .47“ .59 .16“
3 Acad. Efficacy2 .24“ .63 .05“
4 lnterest2 .27” .67 .05“
5 Trait mastery X
interest2 .14 .67 .00

 

Note. All beta weights are from the step at which the variable was entered; n=354.

*p <.05. **p <.01.

136

Table 9, continued

DV = State mastery 3

 

 

Step # IV rs R AR2
1 .27 .0?“
Gen. Efficacy .27“
ACT -.04
2 Trait Mastery .24** .34 .04“
3 State Mastery2 .46“ .54 .17**
4 Acad. Efficacy3 .30“ .60 .07“
5 lnterest3 .32“ .66 .07“
6 Trait mastery X
interest3 .74 .70 .05“

 

Note. All beta weights are from the step at which the variable was entered; n=354.
*p <.05. **p <.01.

 

Table 10

137

Regression results for lirpothesis 3: Exam oals

DV = Exam goal 1

 

 

 

 

 

Step # IV rs R AR2
1 .33 .1 1**
Gen. Efficacy .29“
ACT .15**
2 .40 .05“
Trait Mastery .25*
Interest2 .03
3 State Mastery2 .03 .40 .00
4 Acad. Efficacy2 .23” .44 .03“
DV = Exam goal 2
Step # IV rs R AR2
1 .28 .08“
Gen. Efficacy .07**
ACT .27”
2 .35 .04“
Trait Mastery .03
lnterest2 .20“
3 State Mastery2 .16* .37 .02*
4 Acad. Efficacy2 .37** .48 .10“

 

Note. All beta weights are from the step at which the variable was entered; n=354.

*p <.05. **p <.01.

138
Table 10, continued

Regression resglts for hygthesis 3: Exam goals

DV = Exam goal 3

 

 

Step # N r:- R AR2

1 .28 .08“
Gen. Efficacy .04“
ACT .28“

2 .35 .05“
Trait Mastery .08
lnterest2 .20“

3 State Mastery2 .09* .36 .01

4 Acad. Efficacy2 .27“ .42 .05“

 

Note. All beta weights are from the step at which the variable was entered; n=354.
*p <.05. **p <.01.

 

139
Table 11
Regression resglts for Hymthesis 3: Learning goals

DV = Learning goal, Wave 2

 

 

Step # IV r:- R AR2
1 .25 .06**
Gen. Efficacy .17“
ACT -.19**
2 .47 .16“
Trait Mastery .15*
Interest2 .36“
3 State Mastery2 .48“ .60 .14**
4 Acad. Efficacyz -.02 .60 .00

 

DV = Learning Goal Wave 3

 

 

Step # IV r2- R AR2
1 .18 .03”
Gen. Efficacy .17**
ACT -.07
2 .55 .27**
Trait Mastery .08
Interest2 .51“
3 State Mastery2 .59** .72 .23“
4 Acad. Efficacy2 .12** .73 .01“

 

Note. All beta weights are from the step at which the variable was entered; n=354.
*p <.05. **p <.01.

 

140
Table 12

Regression Results for Hypotheses 4. 5 a_nd 6

DV = Exam goal 1

 

 

Step # lV rs R AR2
1 .33 .1 1**
Gen. Efficacy .14“
ACT .29“
2 Acad. Efficacy .26“ .41 .06**
3 Trait Perform. -.05 .41 .00
4 State Perform. .08 .41 .00
5 State Perf. X
Acad. Efficacy1 -1.10* .44 .02“
6 Course Goal .71** .78 .42**

 

DV = Exam goal 2

 

 

 

Step # IV rs R AR2
1 .28 .08**
Gen. Efficacy .07
ACT .27**
2 Acad. Efficacy2 .41“ .48 .15**
3 Trait Perform. -.06 .48 .00
4 State .01 .48 .00
Perform.2
5 State Perf.2 X
Acad. Efficacy2 -.77 .49 .01*
6 Course Goal2 .72“ .80 .40“

 

Note. All beta weights are from the step at which the variable was entered; n=354.
*p <.05. **p <.01.

141
Table 12, continued

Regression results for Hypotheses 4, 5 and 6

DV = Exam goal 3

 

 

Step # IV rs R AR2
1 .28 .08“
Gen. Efficacy .04
ACT .27**
2 Acad. Efficacy2 .33“ .41 .09**
3 Trait Perform. -.08 .42 .01
4 State Perform.2 .12* .44 .01*
5 State Perf.2 X
Acad. Efficacy2 -.27 .44 .00
6 Course Goal2 .74** .78 .41“

 

Note. All beta weights are from the step at which the variable was entered; n=354.
*p <.05. **p <.01.

 

1 42
Table 1 3

Regression Resrits for Hypothesis ritcmgse effort

DV = Planned increase in effort, Wave 2

 

 

 

 

 

Step # IV rs R AR2
1 Discrep1 (D1) .32“ .32 .10**
2 .33 .01
TMO .09
TPO .04
3 SMO1 .14* .35 .02*
4 SPO1 .03 .36 .00
5 SMO1 x D1 -.94* .38 .02*
6 SPO1 x D1 .31 .38 .00
DV = Planned increase in effort, Wave 3
Step # IV 13 R AR2
1 Discrep2 (D2) .29“ .29 .09**
2 .30 .00
TMO .03
TPO -.04
3 SM02 .15“ .33 .02”
4 SP02 .00 .33 .00
5 SM02 x DZ -.69 .34 .01
6 SP02 x D2 -.31 .35 .00

 

Note. All beta weights are from the step at which the variable was entered; n=354.
*p <.05. **p <.01.

 

143
Table 14

_R_egression Results for Hypothesis 7: Goal change

DV = Goal change, Wave 1 to Wave 2

 

 

Step # IV a. R AR2
1 Discrep1 (D1) .34“ .34 .11**
2 .35 .01
TMO -.08
TPO -.02
3 SMO1 .12* .36 .01*
4 SPO1 -.02 .36 .00
5 SMO1 x D1 -.00 .36 .00
6 SPO1 x D1 -.56 .37 .01

 

DV = Goal change, Wave 2 to Wave 3

 

 

Step # lV rs R AR2
1 Discrep2 (DZ) -.26** .26 .07**
2 .27 00
TMO -.05
TPO -.03
3 SM02 -.04 .27 .00
4 SP02 .06 .28 .00
5 SM02 x DZ .13 .28 .00
6 SPOZ x DZ -.23 .28 .00

 

m. All beta weights are from the step at which the variable was entered; n=354.
Discrep = Goal-performance discrepancy, TMO = trait mastery orientation, TPO = trait
performance, SMO = state mastery . SPO = state performance.

*p <.05. **p <.01.

 

144
Table 15
B_e_gression Resglts for Hypothesis 7: Feedback Credibil'gy

DV = Feedback Credibility, Wave 2

 

 

Step # IV rs R AR2
1 Discrep1 (D1) .32" .32 .10"
2 .42 .07**
TMO .08
TPO -.25**
3 SMO1 .16** .44 .02**
4 SPO1 -.1o .45 .01
5 SMO1 x 01 .55 .45 .01
6 SPO1 x D1 -.07 .45 .oo

 

DV = Feedback Credibility, Wave 3

 

 

 

Step # IV rs R AR2
1 Discrep2 (DZ) .13* .13 .02*
2 .33 .10**
TMO .05
TPO -.31**
3 SMOZ .12* .35 .01*
4 SP02 .01 .35 .00
5 SM02 x DZ .46 .36 .00
6 SP02 x DZ .77* .38 .01*

 

_t10_t_e. All beta weights are from the step at which the variable was entered; n=354.
Discrep = Goal-performance discrepancy, TMO = trait mastery orientation, TPO = trait
performance, SMO = state mastery , SPO = state performance.

*p <.05. **p <.01.

 

 

 

 

 

 

145
Table 16
B_eg;ession results forfiypothesis 8
DV = Interest Wave 2
Step # IV rs R AR2
1 .17 .03“
General Efficacy .16“
ACT .06
2 Interest .69“ .71 .47**
3 Discrepancy1 -.1Z** .72 .01**
DV = Interest Wave 3
Step # rv rs R AR2
1 .13 .02
General Efficacy .04
ACT .13*
Z Interest2 .71** .72 .50“
3 DiscrepancyZ -.10** .73 .02M

 

Note. All beta weights are from the step at which the variable was entered; n=354.
*p <.05. **p <.01.

 

Table 17

Regression results for Hypothesis 9

DV = Academic Self-efficacy Wave 2

146

 

 

 

 

 

Step # IV r3 R AR2
1 .37 .14**
General Efficacy .34”
ACT .13*
Z Acad. Efficacy 1 .62“ .67 .32**
3 Discrepancfl -.ZZ** .70 .04“
DV = Academic Self-efficacy Wave 3
Step # IV rs R AR2
1 .37 .13”
General Efficacy .33”
ACT .15*
Z Acad. Efficacy 2 .69” .73 .40“
3 DiscrepancyZ -.12** .74 .01M

 

Note. All beta weights are from the step at which the variable was entered; n=354.

Acad. Efficacy = Academic self-efficacy

*p <.05. **p <.01.

147

Table 18
Regression results for Hypothesis 10

DV = State performance orientation Wave 2

 

 

Step # IV rs R AR2
1 TPO .28“ .28 .06**
2 SPO .61“ .64 .33**
3 Discrep1 (D1) -.02 .64 .oo
4 TPO x D1 -.22 64 .oo

 

DV = State performance orientation Wave 3

 

 

Step # IV rs R AR2
1 TPO .25“ .25 .00"
2 SPOZ .65“ .67 .39“
3 Discrep2 (DZ) -.03 .68 .00
4 TPO x 02 .03 .68 .oo

 

1%. All beta weights are from the step at which the variable was entered; n=354.
Discrep = Goal-performance discrepancy, TMO = trait mastery orientation, TPO = trait
performance, SMO = state mastery , SPO = state performance.

*p <.05. **p <.01.

 

..F-m“ .-
‘ I

 

 

148

 

00000.0... 0000.00 0.5.5 In... 0.: .u.

 

 

 

 

 

 

 

0.000000. 00000.0.
00:0 000000.

.000 000000.
000.2000”.
50002005

 

 

 

 

50002005
.000

000 0.000000... ‘lr

00000-00000

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.

.0220.

 

 

 

 

 

 

00020000
000200.. I
0000200000. 20.0 2.. 0000000000. 00.0
v...
- .000 0.0.0 “£00m
0000.0 0...
80x0.
0%wa ‘10:! 68.00.60
0.E0000<
_ 4/ 0.2 0.0802
0
00.00.00i 0 I 00020000
0200.2 22m. ‘1 020.05. 00....

 

 

 

 

149

 

 

 

 

 

00000.0... 0000.0m. 000200m. .mllml0... _...

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_ 2
00020000 00:200..
. 00000-00000 205 I 000050000”. 00.”
_ .
.............................. .
. 0
0.000000. “00000.0 ¢ F .000 .205
. - 50002005 $00.0
JMWM MWHMHW-r _80 EQXOV afloEOrh—Om
‘1 000 0.000000... ‘1. 0.000 A
00000000. 0000000000 .00-0040 0.E0000<
50002005 > ‘
u . / 0: . t . .
_ _ llllllllllllllllll |\ \
.
_ 00020000
_ 0002000Q All . .
_ .0 2 00.2 22m 0200.2 ..0...
_ .
_
_ IIIIIIIIIIIIII .0...- IT 00.0.0.

 

 

 

 

 

 

 

150
‘ High State. mastoiy * 7 High State-Performance
‘ Orientation}. f Orientation. - _ ‘ .
PositiveDiscrepancy   Change effort(0) r V Change effort (-)
(Performance ~=F‘Goal) . Reject feedback (+) Reject feedback (-)

f N V if I, Change goal (+) Change 908|(0)
Negative Discrepancy -* Change effort (+) Change effort (-)
(Perforrnance< Goal) ‘ Reject feedback (-) Reject feedback (+)

’ ’ ’ ¥ “ Change goal (0) Change goal (-)

 

 

 

 

Figure 3. Summary of Interaction Between State Goal Orientation and Goal-

Performance Discrepancies

 

 

.11

151

Trait
Mastery
' -, nta 'c

        
    
  

031 fl
0 I
1 Orientation
0 fl-
0‘ m

   

 

edbnnano-
'1 ' t- zl

 

 

mm 133 mm mt’Po
1° fl" 9 0 fl .9. .3.

Figure 4. Four factor goal orientation measurement model.

152

. Q) ”I 1 $5 : :21
’5’1 1 @1 *1 1&1 5’ Q

mm mm 00mm M

 

 

    

Orientation

 

 

 

 

 

M intrst2 intrst3 m

 

 

 

 

0 52

Figure 5. Three factor mastery goal orientation and interest model.

Exam goal 1

153

 

 

 

 

 

State Performance Orientation

+Self-eflicacy = -1
+Self-efticacy = 1

 

 

 

Figure 6. Interaction between self-efficacy and state performance orientation.

 

 

 

154

 

Planned Increase In effort

 

 

State meeury ortentatlon

 

+ Discrepancy = -1
+ Discrepancy = 1

 

 

 

Figure 7. Interaction between planned increase in effort and state mastery orientation.

 

 

155

 

Feedback credlblllty

 

 

State performance orientation

 

+ Discrepancy I -1
+Discrepancy = 1

 

 

 

Figure 8. Interaction between feedback credibility and state performance orientation.

 

 

LIST OF REFERENCES

 

 

156

LIST OF REFERENCES

Adams, J. S. (1963). Towards an understanding of equity. Journal of Abnormal
and Social PsycholpgyI 12, 365-381.

Allen, B. P., & Potkay, C. R. (1981). On the arbitrary distinction between states
and traits. Joumgl of Personalpy‘ and Socia_l Psychology. 41, 916-928.

Allen, B. P., & Potkay, C. R. (1983). Just as arbitrary as ever: Comments on
Zuckerman’s rejoinder. Journal of Personalgy' and SocflPsychologv. 44, 1087-1089.

Austin, J. T., & Klein, H. J. (1996). Work motivation and goal striving. In K.R.
Murphy (Ed.) Individual Differences and Belgvior ip Organizations, 209-257. San
Francisco: Jossey-Bass.

Boyle, K. A., & Klimoski, R. J. (1995). Toward an understanding of goal
orientation in a training context. Paper presented at the Tenth Meeting of the Society
for Industrial and Organizational Psychology. Orlando, FL.

Button, S. B., Mathieu, J. E., & Zajac, D. M. (1996). Goal orientation in
organizational research: A conceptual and empirical foundation. Organizational
Behavior _a_nd Human Decision Processes.

Carver, C. S, & Scheier, M. F. (1982). Control theory: A useful conceptual
framework for personality-social, clinical, and health psychology. Psychological
Bulletin 92, 111-135.

 

Carver, C. S, & Scheier, M. F. (1990). Origins and functions of positive and
negative affect: A control—process view. Psychological ReviewI 97, 19-35.

Cattell, R. B. (1983). Structured Persona_lity-Leaming Theory: A Wholistic
Multivariate Resegrch Approach. Praeger Publishers: New York, New York.

Chaplin, W. F., John, O. P., 8. Goldberg, L. R. (1988). Conceptions of states
and traits: Dimensional attributes with ideals as prototypes. Journal of Personalr_ty' ang
Socia_l Psychom, 541-557.

Cohen, J., & Cohen, P. (1983). Applied multiple regpessionlmnelgtions analysis
for the behaviorgj sciences (2nd ed.). Hillsdale, NJ: Lawrence Earibaum Associates.

 

 

157

Cronbach, L. J., & Furby, L. (1970). How should we measure “change” - or
should we? Psychol_ogical BulletinI 74, 68-80.

Dawis, R. V. (1991). Vocational interests, values, and preferences. In M.D.
Dunnette and L. M. Hough (Eds) Handbook of Industrial and Organizational
Psycholgy, 2nd Ed., (\_/ol. 2), 834-871, Consulting Psychologists Press: Palo Alto, CA.

Dweck, C. S. (1986). Mental processes affecting Ieaming. American
Psychologist. 41 . 1 040-1 048.

Dweck, C. S. (1989). Motivation. In A. Lesgold and R. Glaser (Eds.),
Foundations for a Psycthgy of Education. Lawrence Earlbaum Associates: Hillsdale,
NJ.

Dweck, C. S., & Leggett, E. L. (1988). A social-cognitive approach to motivation
and personality. Psychological Review. 95. 256-273.

Elliot, A. J.. & Church, M. A. (1997). A hierarchical model of approach and
avoidance achievement motivation. Journal of Personally and Socia_l Psychology. 72,
21 8-232.

Elliott, E. S., & Dweck, C. S. (1988). Goals: An approach to motivation and
achievement. Jouma_l of Personality and Social Psychology. 54, 5-12.

Eysenck, H. J. (1960). The Stmcture of Human Personality. John Wiley and
Sons: New York.

Eysenck, M. W. & Eysenck, H. J. (1980). Mischel and the concept of
personality. British Journal of Psycholggy, 71, 191-204.

Farr, J. L., Hofmann, D. A., & Ringenbach, K. L. (1993). Goal orientation and
action control theory: Implications for Industrial and Organizational Psychology. In C.L.
Cooper and LT. Robertson (Eds.), lntemational Review of Industrial and Organizational
Psycholpgy. Vthley and Sons.

 

Farr, J. L., & Ford, C. M. (1990). Individual innovation. In MA. West and J.L.
Farr (Eds.), Innovation and Creativgy' at Work (pp. 63-80). New York: John VVIIey and
Sons.

Fisher, S. L. (1995). The role of cognitive effort and Ieaming outcome demands
in skill acquisition and Ieaming. Unpublished master’s thesis, Michigan State
University.

Fisher, S. L., and Ford, J. K. (in press). Differential effects of Ieamer effort and
goal orientation on two Ieaming outcomes. Person_nel Psychology.

158

Fisher, S. L., Delbridge, KD, & DeShon. R. P. (1997). Here today, gone
tomorrow? The stability of goal orientation. Paper presented at the Twelfth Annual
Meeting of the Society for Industrial and Organizational Psychology. St. Louis, MO.

Fridhandler, B. M. (1986). Conceptual note on state, trait, and the state-trait
distinction. Journal of Personalgy' and Socia_l Psychology. 50, 169-174.

George, J. M. (1996) Trait and state affect. In K.R. Murphy (Ed.) Individual
Differences a_nd Behavior i_n Organizations, 145-174. Jossey-Bass: San Francisco.

Hall, D. T. (1987). Careers and socialization. Journal of ManagementI 13, 301-
321.

Hampson, S. E. (1988). The Constrflion of Persona_lity: An Introguction (2nd
Ed). Routledge: New York.

Heyman, G. D., & Dweck, C. S. (1992). Achievement goals and intrinsic
motivation: Their relation and their role in adaptive motivation. Motivation and
Emotion, 16, 231-247.

Hollenbeck, J. R. & Klein, H. (1987). Goal commitment and the goal-setting
process: Problems, prospects, and proposals for future research. Journal of Applied
Psychology.lg, 212-220.

Hough, L. M., & Schneider, R. J. (1996). Personality traits, taxonomies, and
applications in organizations. In K.R. Murphy (Ed.) Individual Differences and Behavior
in Organizations, 31-88. Jossey-Bass: San Francisco.

Ilgen, D. R., Fisher, C. D., & Taylor, M. S. (1979). Consequences of individual
feedback on behavior in organizations. Journal of Applied PsychologyL64, 349-371.

Kanfer, R. (1990). Motivation theory and industrial and organizational
psychology. In M.D. Dunnette and L. M. Hough (Eds.) Handbook of Industrial and
Organizationgfi’sychologyI 2nd Ed.I (got. 1). 76-170, Consulting Psychologists Press:
Palo Alto. CA.

Kanfer, R., 8. Ackerman, P. (1989). Motivation and cognitive abilities: An
integrative aptitude treatment interaction approach to skill acquisition. Journal of
Applied Psychology, 74, 657-690.

Karoly, P. (1993). Mechanisms of self-regulation: A systems view. Annual
Review of Psycholofl, 44. 23-52.

Keman, M. C., & Lord, R. G. (1990). Effects of valence, expectancies, and
goal-performance discrepancies in single and multiple goal environments. Journal of

Applied PsychologyI 75, 194-203.

 

159

King, W. C., Jr., & Miles, E. W. (1995). A quasi-experimental assessment of the
effect of computerizing noncognitive paper-and-pencil measurements: A test of
measurement equivalence. Journal of Applied PsyghologyI 80, 643-651.

Kluger, A., & DiNisi, A. (1996). The effects of feedback interventions on
performance: A historical review, meta-analysis, and a preliminary feedback
intervention theory. Psychological BulletinI 119,

Kozlowski, S. W. J., Gully, S. M., Smith, E. M., Nason, E. R., & Brown, K. G.
(1995, May). Learning orientation and Ieaming objectives: The effects of sequenced
mastery goals and advance organizers on performance, knowledge, meta-cognitive
structure, self-efficacy, and skill generalization for complex tasks. Paper presented at
the Tenth Conference of the Society for Industrial and Organizational Psychology,
Orlando, FL.

Lane, D. S., Newman, D. L., & Bull, K. S. (1988). The relationship of student
interest and advance organizer effectiveness. Contemporay Ed_ucationa_l Psychology,
1;, 15—25.

Locke, E. A., 8 Bryan, J. F. (1968). Goal-setting as a determinant of the effect
of knowledge of score on performance. Joumal of Applied PsychologyI 81, 398-406.

Locke, E. A., & Latham, G. P. (1990). A theory of goal setting and task
performance. Englewood Cliffs, NJ: Prentice Hall.

Lord, R. G., & Levy, P. E. (1994). Moving from cognition to action: A control
theory perspective. Applied Psychology: An lntemational Review, 43, 335-398.

Magnusson, D., & Endler, N. S. (1977). Interactional psychology: Present status
and future prospects. In D. Magnusson and N.S. Endler (Eds.), Personal'g at the
Crossroags: Cgryent lssgg§_in lnteractiopai Psychology, 3-36. Hillsdale, NJ: Lawrence
Earlbaum Associates.

Meece, J. L., Blumenfeld, P. C., 81 Hoyle, R. H. (1988). Students’ goal
orientations and cognitive engagement in classroom activities. Journal of Educational
Psychology, 80, 514-523.

Meister, J. C. (1994). Corporate Quality Universities. New York: Irwin.

Mirvis, P. H., & Hall, D. T. (1994). Psychological success and the boundaryless
career. Jouma_| of Organizational Behavior. 15, 365-380.

Mischel, W. (1968). Personal'gy and Assessment. New York: Wiley.

Mischel, W. (1969). Continuity and change in personality. American
PsychologistI 24 (11), 1012-1018.

160

Mischel, W., 8. Shoda, Y. (1995). A cognitive-affective system theory of
personality: Reconceptializing situations, dispositions, dynamics, and invariance in
personality structure. Psychological Review. 10;, 246-268.

Mone, M. A., 8. Baker, D. D. (1992a). Cognitive, affective, and behavioral
determinants and consequences of self-set goals: An integrative, dynamic model.
Human Perfonnayrcefi. 213-234.

Mone, M. A., & Baker, D. D. (1992b). A social-cognitive, attributional model of
personal goals: An empirical evaluation. Motivation and Emotion 16, 297-321.

 

Murtha, T. C., Kanfer, R., and Ackerrnan, P. L. (1996). Toward an interactionist
taxonomy of personality and situations: An integrative situational-dispositional
representation of personality traits. Journal of Personalgy' and Social Psychology. 71.
193-207.

Nicholls, J. G. (1984). Achievement motivation: Conceptions of ability,
subjective experience, task choice, and performance. Psychological Review. 91, 328-
346.

OIweus, D. (1977). A critical analysis of the “modern” interactionist position. In
D. Magnusson and NS. Endler (Eds.), Personal'gy at the Crossrofls: Cgrlent Issues in
lnteractionaj Psychology, 221-235. Hillsdale, NJ: Lawrence Earlbaum Associates.

Pintrich, P. R., Smith, D. A. F., Garcia, T., 8. McKeachie, W.J. (1991). A manual
for the use of the Motivated Strategies for Learning Questionnaire (MSLQ). Technical
Report No. 91-B-004, University of Michigan, Ann Arbor, MI.

Pintrich, P. R., Smith, D. A. F., Garcia, T., 8. McKeachie, W. J. (1993). Reliability
and predictive validity of the Motivated Strategies for Learning Questionnaire (MSLQ).

Educational and Psycholpgical MeasurementI 53, 801-813.

Podsakoff, P. M. 8. Farh, J. L. (1987). Effects of feedback sign and credibility on
goal setting and task performance. Organizationaj Behgflor and Human Decision
Processes, 44, 45-67.

Schiefele, U. (1991). Interest, Learning and Motivation. Educational
Psychologi_st, 26. 299-323.

Schuerger, J. M., Zarella, K. L., & Hotz, A. S. (1989). Factors that influence the

temporal stability of personality by questionnaire. Journal of Personal'fl and Social
Psychology, 56, 777-783.

Senge, P. M. (1990). The Fifth Discipline. New York: Doubleday.

Shoda, Y., Mischel, W., & Wright, J. C. (1994). Intraindividual stability in the
organization and patterning of behavior: Incorporating psychological situations into the

161

ideographic analysis of personality. Journal of Personaly' and SociaiPsycholgy, 67,
674-687.

Spielberger, C. D. (1977). State-trait anxiety and interactional psychology. In D.
Magnusson and N. S. Endler (Eds.), Personally at the Crossggds: Current Issues ig
Intergdionajfisycholggy, 173-184. Hillsdale, NJ: Lawrence Earlbaum Associates.

Smith, E. 8., Ford, J. K., Weissbein, D. A., 8. Gully, S. M. (1995, May). The
effects of goal orientation, metacognition, and practice strategies on Ieaming and
transfer. Paper presented at the Tenth Meeting of the Society for Industrial and
Organizational Psychology, Orlando, FL.

Stagner, R. (1977). On the reality and relevance of trails. The Joumal of
Genefiral Psychology, 96, 185-207.

Thomas, K. M. 8. Mathieu, J. E. (1994). Role of causal attributions in dynamic
self-regulation and goal processes. Journal of Applied PsychologyJ9, 812-818.

VandeWalle, D. (in press). Development and validation of a work domain goal
orientation instrument. Educational and Psychological Measurement.

VandeWalle, D., and Cummings, L. L. (1997). A test of the influence of goal
orientation on the feedback seeking process. Journal of Applied Psychology, 82, 390-
400.

Wofford, J. C. & Goodwin, V. L. (1990). Effects of feedback on cognitive
processing and choice of decision style. Journal of Applied Psychology, 75. 603-612.

Wright, P. M. 8. Kacmar, M. (1994). Goal specificity as a determinant of goal
commitment and goal change. Organizational Behavior and Human Decision
Processes. 59. 424-260.

Wright, P. M., O'Leary-Kelly, A. M., Cortina, J. M., Klein, H. J., & Hollenbeck,
J. R. (1994). On the meaning and measurement of goal commitment. Journal of

Applied Psycholggy, 79, 795-803.

Zuckennan, M. (1977). General and situation-specific traits and states: New
approaches to assessment of anxiety and other constructs. In M. Zuckennan and CD.
Spielberger (Eds.), Emotions and Anxiety: New Concepts. Methods, and Applications.
Lawrence Earlbaum Associates: Hillsdale, NJ.

Zuckennan, M. (1983). The distinction between trait and state scales is not
arbitrary: Comment on Allen and Potkay’s ”On the arbitrary distinction between states
and traits.” Journal of Personalgy' and Socfl Psychology, 44, 1083-1086.

       

HICHIG

nN smre UNIV. LIBRQRIES
1111““ W W Ill WWIWWII“ llli "H W WI
531293017120050