I‘I _

 

1"” LIBRARY
\ :- Michigan State
310 University

 

 

 

This is to certify that the
thesis entitled

DISENTANGLING THE EFFECTS OF GOAL DIFFICULTY
AND SPECIFICITY: NORMATIVE CALIBRATION AND
SELF-REGULATORY PROCESSES

presented by
Goren Kuljanln

has been accepted towards fulfillment
of the requirements for the

MA. degree in PsychologL

WA

’ ' ’ Male/Prefe s Slgneture
3 0 Zw 6'

U
Date

MSU is an afimetlve—acfion, equal-opportunity employer

 

—.-l-l-.-O-.‘n-Q-A------_n-.—u-o-t-Q-I-.-‘-I-I-i-‘—.-.-.-C-.-.-‘---.-.-.-.-.-L-s.-.-A-

-._4--_e_t.-.-

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K1/ProjIAcc8Pres/ClRC/DateDue.indd

 

DISENTANGLIN G THE EFFECTS OF GOAL DIFFICULTY AND SPECIFICITY:
NORMATIVE CALIBRATION AND SELF -REGULATORY PROCESSES

By

Goran Kuljanin

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
MASTER OF ARTS
Department of Psychology

2008

ABSTRACT

DISENTANGLING THE EFFECTS OF GOAL DIFFICULTY AND SPECIFICITY:
NORMATIVE CALIBRATION AND SELF -REGULATORY PROCESSES

By

Goran Kuljanin

This study examines the effects of goal difficulty and specificity as conceptualized in
goal setting theory (Locke & Latham, 1990a). Three dominant self-regulation theories
are reviewed first. Then, a close examination of experimental design assumptions in goal
setting theory is undertaken. This is followed by a discussion of goal setting theory’s
current explanation of the effects of goal difficulty and specificity, which is integrated
with the self-regulation literature. The results of the present study suggest that there is
still much to learn about the effects of attributes of goals on task performance as the
results do not follow predictions derived fiom goal setting theory. Implications of

present findings for future research are discussed.

ACKNOWLEDGEMENTS

Thanks to my advisor, Steve Kozlowski, for providing invaluable guidance
throughout the research process of the present study. Working with him has made me a
better researcher.

Thanks to my committee members, Rick DeShon and Kevin Ford, for providing
guidance in improving the research in the present study.

Thanks to my family and friends for all of the loving support they have always

provided me.

iii

TABLE OF CONTENTS

LIST OF TABLES ....................................................................................................... vi
LIST OF FIGURES .................................................................................................... vii
INTRODUCTION ........................................................................................................ 1
Theoretical Foundation ..................................................................................... 2
Self-Regulation Theories in General .................................................... 2
Social Cognitive Theory ....................................................................... 3
Control Theory ...................................................................................... 5
Goal Setting Theory .............................................................................. 6
Recap of Self-Regulation Theories ....................................................... 9
Critique of Goal Setting Experiments ................................................. 10
Goal Setting Theory Guidelines to ‘Good’ Experimental Design
and their Critique ................................................................................ 1 1
Goal Setting Experiments in General and their Critique .................... 13
Integration ....................................................................................................... 1 9
Model Development and Hypotheses ............................................................. 25
Model and Independent Variables ...................................................... 25
Goal Difficulty ........................................................................ 25
Goal Specificity ...................................................................... 26
Normative Information ........................................................... 26
Hypotheses .......................................................................................... 29
The Effects of the Independent Variables on Task
Performance ............................................................................ 30
The Effects of the Independent Variables on Process
Variables ................................................................................. 34
The Effects of the Process Variables on Task Performance 39
METHOD ................................................................................................................... 41
Design ............................................................................................................. 41
Task ................................................................................................................. 41
Participants ...................................................................................................... 41
Procedure ........................................................................................................ 42
Manipulations ................................................................................................. 42
Goal Difficulty .................................................................................... 42
Goal Specificity .................................................................................. 43
Calibration ........................................................................................... 44
Measures ......................................................................................................... 44
Ability ................................................................................................. 44
Goal Commitment ............................................................................... 44
Self-efficacy ........................................................................................ 45
Personal Goals .................................................................................... 45

iv

Effort ................................................................................................... 45

Task Performance ............................................................................... 45
RESULTS ................................................................................................................... 46
The Effects of the Independent Variables on Task Performance ................... 47
The Effects of the Independent Variables on Process Variables .................... 51
The Effects of the Process Variables on Task Performance ........................... 57
DISCUSSION ............................................................................................................. 59
Goal Difficulty and Goal Specificity .............................................................. 59
Effects on Task Performance and Task Performance Variability ....... 60

Effects of Goa] Commitment, Self-efficacy, Personal Goals, and
Effort ................................................................................................... 63
Normative Calibration .................................................................................... 64
Effects on Task Performance and Task Performance Variability ....... 65
Effects of Self-efficacy, Personal Goals, and Effort ........................... 66

Relationship between Goal Commitment, Self-efficacy, Personal Goals,
Effort, and Task Performance ......................................................................... 67
Synthesis of Results ........................................................................................ 67
Limitations and Future Research .................................................................... 69
Conclusion ...................................................................................................... 72
APPENDIX A: Consent and Debriefing Forms ......................................................... 74
APPENDIX B: Measures ............................................................................................ 77
REFERENCES ........................................................................................................... 81

LIST OF TABLES

Table 1. Current explanations of why specific, difficult goals lead to higher
task performance than vague, difficult goals ............................................... 19

Table 2. Presentation of normative task performance information ............................ 44

Table 3. The number of participants who were removed from each condition
as a result of not following instructions ....................................................... 47

Table 4. The final number of participants in each condition ..................................... 47

Table 5. Means, standard deviations, correlations, and reliabilities of all
variables ....................................................................................................... 58

vi

LIST OF FIGURES

Figure 1. Heuristic model of the propositions of GST ............................................... 30
Figure 2. Goal difficulty by goal specificity interaction on task performance .......... 49

Figure 3. Goal difficulty by goal specificity interaction on task performance
without norms ............................................................................................ 50

Figure 4. Goal difficulty by goal specificity interaction on task performance
with norms ................................................................................................. 50

Figure 5. Goal specificity by norms interaction on standard errors of task

performance ............................................................................................... 51
Figure 6. Goal difficulty by goal specificity interaction on personal goals ............... 53
Figure 7. Goal difficulty by goal specificity interaction on effort ............................. 54
Figure 8. Goal difficulty by norms interaction on self-efficacy ................................ 54

Figure 9. Goal difficulty by goal specificity interaction on personal goals
without norms ............................................................................................ 56

Figure 10. Goal difficulty by goal specificity interaction on personal goals
with norms ............................................................................................... 56

Figure 11. Time by norms interaction on personal goals ........................................... 57

vii

INTRODUCTION

Goal Setting Theory (GST) has been one of the most researched theories in
industrial/organizational psychology, and the effects of setting goals have been one of the
most valuable applications of industrial/organizational psychology knowledge used in
organizations (Locke & Latham, 1990a). The two most commonly investigated findings
in GST deal with the relation between attributes of goals and task performance. One
finding is that there is a linear relationship between the difficulty of goals and task
performance such that more difficult goals lead to higher task performance (Locke &
Bryan, 1967). A second finding is that specific, difficult goals lead to higher task
performance than vague, dimcult goals (e.g., “do your best”; Locke, Shaw, Saari, &
Latham, 1981). However, despite the fact that the effects of setting goals have been
replicated in numerous laboratory and field experiments (for a review see Latham &
Locke, 1991; Locke & Latham, 1990b; Locke & Latham, 2002), goal setting research has
not addressed well the underlying mechanisms that account for these effects. This study
will closely examine and elaborate the explanation offered by GST on why specific,
difl'rcult goals lead to better task performance than vague, difficult goals.

The current explanation as to why specific, difficult goals lead to better task
performance than vague, difficult goals rest on goal specificity (Locke & Latham, 1985).
For GST, this is a natural explanation as goal difliculty may potentially be equal across
the two conditions (see Locke & Latham, 1990a, pp. 29-31), and therefore, goal
specificity must be responsible for the effect. However, GST is not clear about how goal

specificity influences the motivational process. In fact, the large majority of goal setting

studies have only investigated the simple effect between specific, difficult goals and
vague, difficult goals, while only a handful of studies have attempted to tease apart the
effects of goal difficulty and goal specificity on process variables (e.g., personal goal
setting) and task performance (e.g., Locke, Chah, Harrison, & Lustgarten, 1989).

The purpose of the present study is to disentangle the effects of goal difficulty and
goal specificity and to better articulate how goal difficulty and goal specificity exhibit
1mique and interactive effects on the self-regulatory process that can account for effects
on task performance. Since GST is one among a class of self-regulatory models, a brief
overview of human self-regulatory systems is provided first, followed by more detailed
analyses of three dominant self-regulatory theories - Social Cognitive Theory (SCT),
Control Theory (CT), and GST - to establish the theoretical foundation. Moreover, given
that GST is an inductive theory, it is important to closely examine the experimental
paradigm that has been most frequently utilized and its design assumptions because of the
possible role it plays in the explanation as to why specific, difficult goals lead to better
task performance than vague, difficult goals. Thus, after discussing the properties of three
dominant self-regulatory theories, a close consideration of GST experimental design
assumptions is undertaken. Finally, an integrated model and an alternative explanation as
to why specific, difficult goals lead to better task performance than vague, dificult goals
are offered.

Theoretical Foundation
Self-Regulation Theories in General
There are many theories that share the same self-regulatory elements, and hence,

these theories have come to be labeled as theories of self-regulation (Karoly, 1993).

Some of the self-regulation theories include SCT, CT, and GST. Generally, the self-
regulatory framework is composed of several key components. First, an individual sets a
goal that she would like to achieve. Then, the individual takes action (i.e., behaves) to
achieve that goal. A reiterative process is set into place by receiving feedback (e.g.,
knowledge of score information) regarding the effects of one’s behavior (e.g., task
performance). Then, the individual examines the discrepancy between his goal and the
feedback information. This is followed by a reaction to the discrepancy where the
individual assesses his satisfaction with the current performance level. The reactions lead
to a reevaluation of the current goal and the current goal is adjusted to be higher, lower,
or kept at the current level, while effort is adjusted accordingly as well. Finally, when the
individual is given another chance of taking action (e. g., performing on a given task) the
self-regulatory process begins anew.

The important things to note from the self-regulatory process is that an individual
needs to have goals relevant to particular behaviors (e.g., task performance) and needs to
receive feedback on those particular behaviors in order for the goals to affect future such
behaviors. The opportunity to have multiple instances to exhibit behaviors is something
that is stressed in the three self-regulatory theories discussed below.

Social Cognitive Theory

SCT states that perceived future events are utilized as present motivators and
regulators of behavior (e.g., task performance; Bandura, 1991). The self-regulatory
process is described by three phases. Self-observation, also known as the self-monitoring
sub-function, is the initial phase that provides information for the later two phases such as

providing realistic information for setting goals and evaluating one’s progress toward

those goals. Two factors that affect self-observation are the temporal proximity of
feedback and how informative the outcome feedback is (Bandura, 1991).

Self-evaluation, also known as the judgmental sub-function, is the process of
comparing information obtained in the self-observation phase to personal goals or
external standards. It is important to note that three pieces of information are required to
evaluate one’s own performance: the attained performance level (i.e., receiving
informative feedback), one’s personal standards, and the performance of others (Bandura,
1991). These three pieces of information are best explained by Bandura’s (1991)
example:

“For example, a student who achieves a score of 115 points on an examination,

and whose aspiration is to be in the upper 10% of a certain group, would have no

basis for meaningful self-appraisal without knowing how others have performed”

(p.254).

The last phase of the self-regulation process in SCT is self-reaction. In the task
performance domain of goal setting studies, performance standards regulate task
performance by deploying self-incentives to reward satisfactory task performance. In
other words, people are motivated by the anticipated satisfaction of the goals that they
desire. Bandura (1991) argues that motivational effects come from the evaluation of
one’s own behavior and not the goals that people set. Furthermore, in order to provide a
basis for self-reactive influences, two conditions must be satisfied. The first condition is
that if a goal is adopted, then one needs to know how one is doing. The second condition

is that if one knows how one is doing, then one needs to have a goal. Simply stated, self—

reactive influences will have a motivational impact if and only if one has adopted a goal
and one knows how one is doing (i.e., performing).

Self-efficacy, one’s beliefs regarding one’s capabilities to perform on a given
task, is central to the self-regulatory process described in SCT (Bandura, 1991).
Generally, those who are more self-efficacious will set higher goals and be committed to
these goals (Wood & Bandura, 1989). Furthermore, performance discrepancies are
interpreted differently depending on the level of one’s self-efficacy. If one is of low self-
efficacy, then one generally gives up in the face of failure. However, if one is of high
self-efficacy, then performance discrepancies are met with more exerted effort (Bandura
& Cervone, 1986). Additionally, self-efficacy is construed as a dynamic construct that is
updated as experience is acquired and information is updated (Gist & Mitchell, 1992).
The role of social comparison, such as that induced by normative information, can have
an efi‘ect on self-efficacy, which in turn, affects task performance (Bandura, 1991).

To summarize, SCT of self-regulation requires multiple opportrmities to perform
in order for the three self-regulatory sub-functions to be useful. The crucial components
of this self-regulatory system are the availability of feedback information and
performance standards. Feedback information is used in the self-monitoring sub-function
of SCT, while performance standards are used in the judgmental sub-function of SCT.
Both, feedback information and performance standards, are required to provide a basis
for self-reactive influences.

W
CT details the process of self-regulation via a negative feedback loop (Miller,

Galanter, & Pribram, 1960; Carver & Scheier, 1982; Carver & Scheier, 1990).

Generally, it is conceived that there is an input which is compared to some standard,
either externally or internally set. An output (i.e., a behavior) is produced which is aimed
at reducing the discrepancy between the standard and task performance. This process
iterates until the discrepancy between task performance and the standard has been
eliminated. Once a discrepancy is eliminated, it is proposed that an individual can focus
her attention on other goals (Carver & Scheier, 1998). That is, in this conceptualization,
it can be considered that there also exists a positive feedback loop such that once a goal is
achieved for one task, the individual moves on to another task. Finally, the negative
feedback loop can be operative at various levels of the goal hierarchy (Carver & Scheier,
1 982).

If the individual continually fails to reduce the performance discrepancy, then the
individual will tend to disengage from the task (Carver & Scheier, 1981). This is a
somewhat different prediction fiom SCT since one’s disengagement from a task in SCT
depends on self-efficacy. Ifone’s self-eflicacy is high, then, in the presence of a
performance discrepancy, the individual will exert more effort to perform better.
However, if one is of low self-efficacy, then one will tend to disengage from the task.

The point of convergence of CT with SCT is that it includes assigned or personal
goals and feedback is necessary in order to adjust action (i.e., task performance). Again,
this is another example of a self-regulatory system where multiple opportunities,
feedback, and performance standards are necessary to affect task performance.

Goal 8mm Theory
As noted earlier, the formulation of GST was most clearly stated in Locke and

Latham (1990) when they summarized the research findings ofhundreds ofgoal setting

studies. The basic tenet of the theory is that goals lead to action (e.g., task performance).
Furthermore, it is stated that the attributes of goals have differential effects on task
performance (Austin & Vancouver, 1996). The two most researched attributes of goals
have been goal difficulty and goal specificity, and there are two main findings from goal
setting experiments that have been replicated numerous times (Locke & Latham, 1990a).
First, goals that are high in difficulty for the given task lead to better task performance
than goals that are of easy goal difficulty. Meta-analytic findings indicate that the effect
size (d) is about .58 (Wood, Mento, & Locke, 1987). The second common finding in
GST experiments is that goals of high specificity and difficulty (e.g., precisely stated
quantitative goals) lead to higher task performance than goals of high difficulty but low
specificity (e.g., ‘do your best’). A meta-analysis of this common finding indicates that
the effect size is around .43 (Wood et al., 1987).

Given that these findings relate to goals that are assigned to individuals, the
process of how the assigned goals affect task performance is the remaining part of GST.
The assignment of goals affects personal goals that are set by individuals, which, in turn,
affect task performance. That is, the effect of an assigned goal on task performance is
mediated by personal goals. In addition, GST stipulates that in the presence of feedback,
which is simply knowledge of score, the effect of goals on task performance is stronger
than in the absence of feedback. Hence, feedback moderates the relationship between
goals and task performance. It needs to be noted that in most goal setting experiments
that the high goal difficulty and specificity group receives feedback in the form of actual
task performance results, whereas the high goal difficulty and low specificity group either

does not receive any feedback regarding task performance results or receives feedback on

task performance results where each performance trial is of a varying length as compared
to previous and firture performance trials. Notice that this experimental design element is
inconsistent within a self-regulatory framework, because in order to provide necessary
elements of self-regulation both goals and clear feedback information need to be
available to the individual (Bandura, 1991).

Two other variables that moderate the relationship between goals and task
performance are goal commitment and task complexity. Goal commitment, from the
earliest goal setting studies, has been proposed to be a moderator of goal difficulty and
task performance. A meta-analysis has shown that this moderation effect explained about
3% of the variance in task performance (Donovan & Radosevich, 1998). This should not
be considered to be a trivial amount of variance explained given that it is an interaction
effect. The importance of a significant interaction is that main effects cannot be
interpreted without considering the interaction. Goal specificity has been proposed and
demonstrated to be a determinant of goal commitment (Hollenbeck & Klein, 1987;
Wright & Kacmar, 1994). Wright and Kacmar (1994) demonstrated that specific goals
had significantly higher goal commiment than non-specific goals.

In terms of task complexity, it has been demonstrated that the effects of level of
goal difficulty on task performance is stronger when tasks of a simple nature are utilized.
However, this does not mean that the non-significant relationships are found on more
complex task. On the contrary, the effect sizes are still in the medium range: d = .41
when comparing specific, difficult goals versus vague, difficult goals (Wood et al., 1987).

The effect size on tasks that are less complex is d = .77 (Wood et al., 1987).

Finally, GST includes self-efficacy as an important process variable in linking
assigned goals to task performance. The theory states that assigned goals will affect self-
efficacy, which in turn, affects personal goals and task performance (Locke & Latham,
1990a). The inclusion of self-efficacy as an important variable in the model links GST
with SCT.

To summarize, GST incorporates a self-regulatory framework in that it contains
the necessary elements (e.g., feedback) of self-regulation and proposes relationships with
key self-regulatory processes (e.g., personal goals). However, goal setting studies have
consistently treated specific, difficult goal conditions differently from vague, difficult
goal conditions in that only the former received clear score information (i.e., feedback).
Feedback information is an important design element in self-regulation as it is expressed
in all three theories of self-regulation that both feedback and goals are necessary for self-
regulation to affect behaviors (e.g., task performance).

Recap of Self-Regulation Theories

SCT, CT, and GST all contain similar components of the self-regulatory process.
That is, all include some conceptualization of a goal, discuss the importance of feedback,
and discuss the adjustment of personal goals to increase task performance. The following
section focuses in on an examination of the design assumptions that have been utilized in
GST. Again, the design assumptions are of utmost importance to GST as the results from
experiments have built the theory. If there are problematic or questionable design

assumptions, then the processes and explanations offered by GST come into question.

Critique of Gold Setting Experiments

GST to its authors is a motivational technique that can be used to increase task
performance (Bryan & Locke, 1967). However, GST is an inductive theory which
implies that the results of experiments guided theoretical propositions. That is, the
theoretical propositions offered by GST were built experiment by experiment. In fact,
the authors of GST, Locke and Latham (1990), most completely described the theoretical
propositions of the theory afier about 25 years of intensive research in this area, although
excluding the research that was conducted since the beginning of the twentieth century
(e.g., Taylor, 1911; Lewin, Dembo, Festinger, & Sears, 1944). Given that GST was built
from the ground up, it is important to examine the design assumptions that have been
utilized to build the theory because the experimental results will have direct implications
for the theoretical propositions and explanations of GST. If the designs of goal setting
experiments are not carefully thought out, then the common effects that are found in GST
could be a result of the design assumptions rather than the theoretical propositions and
explanations espoused in GST. Given that, one of its authors has been quick to criticize
any experiment that does not follow his guidelines to good experimental design for goal
setting studies (Locke, 1991; Locke, 1994; Locke & Latham, 1990a). These guidelines
may have an unintended effect in that the same design assumptions are recommended to
be utilized to direct research on GST, and therefore, potential mechanisms may be
missed. Thus, this section contests some of these guidelines, and furthermore, critiques

other aspects of the experimental design paradigm that have been used in GST.

10

Goal SettinLTheorv Guidefinps to ‘Good’ Experimental DerignAnd their Critique

One of the principal guidelines to good experimental design in goal setting theory
is to not let ‘do your best’ participants set their own goals (Locke & Latham, 1990a;
Locke, 1991; Locke, 1994). These participants should either receive no feedback
regarding their performance or they should be given feedback based on varying lengths
of work periods (Locke, Cartledge, & Koeppel, 1968). However, this provides unequal
conditions between the specific, difficult goal group and vague, difficult goal group.

Feedback is an integral part to assessing one’s progress toward his goals whether
the goals are assigned or self-set (Bandura, 1991; Carver & Scheier, 1982; Locke &
Latham, 1990a). Naturally, the lack of feedback information in ‘do your best’ conditions
in goal setting studies does not give participants the opportunity to self-monitor. Notice
that for a ‘do your best’ goal condition it would seem natural to provide feedback
information to participants as participants would be able to evaluate their best on the
metric of task performance for the given task. That is, if feedback information is
provided to ‘do your best’ participants, then participants would be able to use their
previous task performance results as standards to evaluate a ‘do your best’ goal.

Giving an assigned goal, whether it is a specific, difficult goal or a ‘do your best’
goal, requires an interpretation by the participants of what that goal means given the task
that they are asked to work on. Those in the specific, difficult group have to find out
what the assigned numeric goal means on the given task. Is this a goal that can be
achieved at first attempt on the task or is this a goal that will require several performance
attempts to achieve? prarticipants given the ‘do your best’ goal are not given feedback

information regarding task performance, then these participants have no method of

11

translating their externally set goal, ‘do your best,’ to an internally set goal standard. GST
proposes that if participants do not receive feedback there will be little or no further
effect of goals fiorn that of the initial effect the goals had on task performance (Locke &
Latham, 1990a), which is in alignment with other self-regulatory theories (e.g., see SCT
in Bandura, 1991). This is because feedback makes clear the range of task performance
for a particular task. Again, notice that the issue of translating an externally set goal to an
internally set goal is also present in specific, difficult goal conditions of goal setting
studies. Ifspecific, difficult goal participants were not given feedback on their
performance results, then they too would not be able to interpret the meaning of their
assigned goal relative to the metric of task performance.

The issue of translating externally set goals to internally set goals is important to
examine. Note that most tasks in GST require some sort of quantitative output as the
measure of task performance (Locke & Latham, 1990a). Hence, the specific, difficult
group is given an advantage in that their assigned goal is already in the form of a
quantitative measure whereas the ‘do your best’ group starts out with the qualitative goal
to ‘do your best.’ The ‘do your best’ group has to take their goal and translate it into a
quantitative score before they can start striving for that difficult goal. Providing feedback
information will give the ‘do your best’ goal group the opportunity to translate their
externally set goal to an internally set quantitative goal.

Locke (1991) goes on to advise that the second step in good goal setting
procedure is to measure the personal goal set by the participants themselves in response
to the assigned goal that was given to them According to GST, goal difliculty is what is

affecting task performance (Locke et al., 1989). Given that, and if personal goals are

12

measured, then it would be interesting to examine the level of goal difliculty that the ‘do
your best’ group sets. If the ‘do your best’ group is setting easy goals, then, according to
GST, the specific, diffith group will outperform the ‘do your best’ group as the specific,
difficult group has a more difficult goal. Note also that knowledge of score information
influences personal goal setting, which has been shown to be more highly correlated with
task performance than assigned goals (Locke, 1994).

The third rule of good procedure for goal setting studies is making sure that the
specific, difficult group actually receives difficult goals such that no more than 10% of
subjects can reach them (Locke, 1991). Notice that the assigned goal for the specific,
difficult group is in relation to other people. In other words, it is a normative standard.
That is, most goal setting studies have compared a specific, diffith assigned goal, set in
relation to a normative reference, to a ‘do your best’ goal where people were not given an
equivalent feedback system to assess the meaning of their assigned goal. Even if the ‘do
your best’ goal group is given feedback, then the comparison of the specific, difficult
goal group to the ‘do your best’ goal group would reflect a comparison between a
normatively referenced assigned goal and a self-referential assigned goal (N aylor &
Ilgen, 1984).

Goal Setting Exmriments in General and their Critique

There are several other questions to be answered about the experimental design
elements of goal setting studies that are not part of the guidelines offered by Locke and
Latham (1990), Locke (1991), and Locke (1994). The questions pertain to the types of

tasks that have been used, the statistical analyses that have been done, and the lack of

13

discussion that a difficult, specific goal is a combination of two goal attributes (i.e.,
difficulty and specificity).

The most common types of tasks that have been used in the goal setting literature
have been what most researchers would classify as simple tasks (see Wood et al., 1987,
Figure 1). In fact, four of the six most commonly used tasks in goal setting literature at
the time of Locke and Latham’s (1990) review were “listing nouns, objects, uses,”
“arithmetic/computation,” “assembly,” and “perceptual speed.” These tasks do not
always lend themselves to a clean goal difficulty and goal specificity manipulation. For
instance, in the several reaction time tasks that have been used with GST, it is usually
required for the difficult goal group to respond quickly to some stimulus, whereas the
easy goal group would respond more slowly. The problem is that it is not easy to assess
which assigned goals are, in fact, most difficult. In the reaction time tasks that have been
used, participants have been asked to respond as fast as the given reaction time (i.e.,
assigned goal; Locke et al., 1989). However, it is a bit arbitrary to determine which goal
is, in fact, more difficult. Is a response time of .3 seconds more difficult than a response
time of .5 seconds? Participants see a red light, and then they depress a key to respond.
Locke et al. (1989) operationalized more difficult as responding more quickly. However,
it can certainly be argued that waiting to respond to hit the exact target of .5 seconds is
more difficult than waiting to respond to hit the exact target of .3 seconds. If this is the
case, then better task performance (i.e., reaction time) is arbitrary as well. The most
plausible conclusion that can be made is that asking participants to respond to a stimulus

leads to diyfirent task performance and not better task performance.

14

As noted previously, most goal setting studies have not looked at the independent
effects of goal difficulty and specificity. In one of the few studies to treat them as distinct,
Locke et al. (1989) conducted a 3 (goal specificity) x 3 (goal difficulty) study to separate
the effects of goal specificity and goal difficulty on a reaction time task. In Study 1, the
same participants were used across goal difficulty levels whereas different participants
were used across goal specificity levels. The researchers first ran those in the vague
specificity condition across all goal difficulty levels. Participants were alternatively
asked to respond as fast as possible, moderately fast, and slowly. The order of
presentation of goal difficulty levels was counterbalanced. To set the other two levels of
goal specificity, the researchers used the performance of the vague specificity group
across the three goal difficulty levels. Medium specificity was considered to be non-
overlapping ranges of reaction times across the goal difficulty conditions whereas high
specificity was considered to be the middle point of these ranges across the goal difficulty
conditions except that the difficult, specific condition was set at the 90th percentile of the
vague, difficult condition (see Table l in Locke et al., 1989).

Given the prior discussion on utilizing reaction time tasks to investigate GST
effects, quicker reaction times are not necessarily more difficult. Waiting longer to
depress a key in specified seconds can be considered to be more difficult than waiting for
a shorter time. Examining the means and standard deviations across goal difficulty in the
vague specificity group shows that there was considerable overlap in the distributions of
the three levels of goal difliculty. However, examining the means and standard
deviations across goal difficulty in the high specificity group shows that there was still

much overlap in the distributions of three levels of goal difficulty. The researchers argue

15

that specificity decreases the variance in performance, but it is difficult to disambiguate
what these results mean due to the ambiguities in the task in terms of goal difficulty. It
could just as well be interpreted that as the reaction time was specified to be slower the
standard deviation increased. In other words, the higher standard deviation for the slower
times could have been a result of the fact that people had trouble (i.e., difficulty) judging
the appropriate time to depress the key. In fact, the results indicate that not only was
there a difference in variance across specificity levels, but also, a difference in variance
across difficulty levels in that slower reaction times led to greater variance (see Table 3).
Furthermore, there was also a difference across goal difficulty but not goal specificity
when assessing intro-individual variability in reaction times in the same direction as that
of the inter-individual variability analysis. This gives more credence to the idea that
participants had a difficult time judging when to press the key for the slower time versus
the faster time. Locke et al. (1989) ponder how goal ‘difficulty’ affected performance in
this task. They proposed that the fast reaction time (high goal difficulty) group could
only respond too slowly where as the slow reaction time (low goal difficulty) group could
respond too slowly or too fast with respect to the assigned reaction time. In other words,
the slow reaction time group had a more difficult time of achieving their goal.

Finally, the specific, difficult to ‘do your best’ comparison was marginally
significant. The researchers set the goal for the specific, difficult group at the 90th
percentile of reaction time for the vague, diflicult (‘do your best’) group. This analysis
may be conformded as it would have been more appropriate to select a specific, difficult

goal from data that were not part of the analysis of that study.

16

In Study 2 of Locke et al. (1989) participants had to list improvements to the
business program at their university. This time the design was completely between-
subjects of 3 (goal specificity) x 3 (goal difficulty). The other difference was that the
specific, difficult group was assigned a goal at the mean of the vague, difficult group.
Given this, it was not surprising that there was no significant difference between the
specific, difficult group and the vague, difficult (list a large number of improvements)
group. In fact, the vague, diflicult group had a higher mean than the specific, difficult
group, although not significantly so. It is interesting to note that feedback was built into
the task (as the participants could see the number of improvements they had listed), and
therefore, the ‘do your best’ group had guidance on how they were performing. The
pattern of means in this experiment is such that there was an interaction between goal
difficulty and goal specificity. The relationship of goal difficulty on task performance
was strongest in the high specificity condition in that higher goals led to better task
performance.

The lack of discussion on the interaction or combinative nature of specific,
diffith goals is surprising. A core finding fiom goal setting studies is that more difficult
goals lead to better task performance than less difficult goals. However, if a ‘do your
best’ goal is potentially the same level of goal difficulty as a specific, difficult goal (see
Locke & Latham, 1990a, pp. 29-31), then it cannot be that goal difficulty is what is
contributing to differential task performance between specific, difficult and vague,
diflicult goal groups. Locke et al. (1989) say that goal specificity does not have an affect
on task performance, but rather that it affects task performance variability. However, a

natural explanation is that goal difficulty and goal specificity have an interactive effect on

17

task performance as there should be a steeper positive slepe fiom low to high goal
difficulty when goal specificity is high as compared to when goal specificity is low.

A final commentary on goal setting experiments is that many studies have been
cross-sectional or have treated the data as cross-sectional. In other words, even if
participants were performing a task for several trials, the data were averaged across trials
and these means were analyzed. This is important because treating the data as repeated
measures could address some of the questions of the process of how the core findings of
GST occur through mediating variables such as personal goals and self-efficacy.

The critique of goal setting experiments is important because different
experimental designs allow for different theoretical conclusions regarding the effects of
goals on task performance in an inductive theory such as GST. If a general self-
regulatory fiamework (i.e., feedback, goals, and multiple performance instances) is
utilized, then knowledge of score information is important for the vague, difficult goal
group. Otherwise, vague, difficult goals should not have any effect on task performance
beyond the initial effect (Locke & Latham, 1990a; Bandma, 1991). Note further that if
people are given the opportunity to view their knowledge of score information and be
aware of normative information regarding task performance, then one can choose
whether to set personal goals around a self-referential point or a normative point. If one
does not have knowledge of score information or some other source (e.g., normative task
information) that can be used as a reference point, then it would be difficult for an
individual to understand the metric of task performance, and thus, their self-regulation

would be impeded.

18

Integration

This section will pull together the key conclusions fiom the critique of GST
experimental design assumptions, interpret these conclusions with respect to the common
self-regulatory framework, and highlight how this study will bridge the gap between
current GST design assumptions and explanation of GST effects with a general self-
regulatory design and the alternative explanation that the self-regulatory design offers.

Table 1 cites explanations across several publications by the authors of GST as to
why specific, difficult goals lead to better task performance than vague, difficult goals.
The offered explanation rests on the notion that the ambiguity of vague, difficult goals
(e.g., “do your best”) allows individuals to interpret many levels of task performance as
their best (Locke & Latham, 1990a; Locke & Latham, 2002). However, given the
critique of experiments on GST discussed above, other experimental designs can provide
further clarification for the observed effects.

Table 1. Current explanations of why specific, difficult goals lead to higher task
performance than vague, difficult goals.
Reference Quote
Locke & Latham “Specific goals direct activity more effectively and reliably than
(1985), pp.206- vague or general goals (Locke, Mento, & Katcher, 1978). Telling
207 someone to ‘do as well as you can’ is an ambiguous statement in
that it does not make clear exactly what the person is to
do. . .Contrary to what many believe, people do not do their best
whentheirgoal istodojustthat. Thisisbecauseamindsetof
doing one’s best is not specific enough to produce maximum
performance.”
Locke, Chah, “Logically one would expect that specificity, divorced fiom level,
Harrison, and would affect the degree of variability in performance as opposed to

Lustgarten level of performance across individuals. Vague goals can be

(1989), pp.271- interpreted in many different ways by different people; as goals

272 become more specific the leeway for interpretation is progressively
reduced.”

 

19

 

Locke, Chah,
Harrison, and
Lustgarten

(1989), p.284

Locke & Latham,
(1990), pp. 30-31

Latham& Locke
(1991), p.215

Locke & Latham
(2002), p.706

“The explanation must be that because do your best is a high yet
vague goal, it can be interpreted in different ways depending on the
context. . .Norrnally some interpretations of do your best will be
more challenging than others. Making high level goals specific
reduces this variance in interpretation and “biases” it toward the
high side, i.e., ensures that all or most individuals will interpret the
goal to mean that a very high level of performance is called for.
Further the high degree of specificity makes it harder for any given
individual to redefine the meaning of the goal during performance,
since performance lapses that would be acceptable under a do best
goal would be seen as failures under a specific, high goal. ..Goal
specificity serves a directive firnction by indicating exactly what
acceptable performance consists of, while goal level serves an
energizing function by mobilizing the effort required to attain the
required performance.”

“ Our explanation is that the ambiguity of difficult goals (this refers
to “do your best” goals) allows people to give themselves the
benefit of the doubt in evaluating their performance; thus, a wide
range of performance levels may be interpreted as being compatible
with doing one’s best.”

“The consistent superiority of the former (specific, difficult goals)
is attributed to the fact that vague goals are compatible with many
different outcomes, including ones that are lower than the person’s
actual best...Maximum effort is not aroused under a do best goal.
This is because the ambiguity inherent in doing one’s best allows
people to give themselves the benefit of the doubt in evaluating
their performance.”

“In short, when people are asked to do their best, they do not do so.
This is because do-your-best goals have no external referent and
thus are defined idiosyncratically. This allows for a wide range of
acceptable performance levels, which is not the case when a goal
level is specified.”

First, if a self-regulatory fi'amework is taken to experimental design, then one missing

component in most goal setting studies has been the lack of clear knowledge of score

information for the ‘do your best’ goal group. Vague, difficult goal participants have

been given either feedback on performance trials of varying length or have not been

given feedback at all (Locke & Latham, 1990a). The three dominant theories of self-

regulation, which includes GST, all emphasize the importance of feedback. In fact, GST

2O

researchers argue that goals without feedback do not have any effect on task performance
(Erez, 1977; Iocke & Latham, 1990a). Thus, goal setting studies have not given vague,
difficult goal participants an important component of regulating their behavior. Notice
that the design assumptions of GST were made to prevent vague, difficult goal
participants fiom setting personal goals as setting personal goals was considered to alter
the goal specificity manipulation (Locke & Latham, 1990a). At the same time, though,
Locke et al. (1989) found that providing feedback to vague, difficult goal participants did
not alter their results (Locke et al., 1989). Then, to disentangle the effects of goal
difficulty and goal specificity, it would be important to investigate at what level vague,
difficult goal participants set personal goals. Although, GST does not indicate the
difference in difficulty levels between specific, difficult goals and vague, difficult goals,
operationally the difficulty levels have been treated as being the same (e.g., Locke et al.,
1989). However, this has not been clearly investigated as of yet and measuring personal
goals will indicate if the equivalent goal difficulty level assumption is tenable.

If personal goals are measured and multiple opportunities are given to perform,
then the explanation offered by GST as to the observed effect comes into question. That
is, given enough performance trials with knowledge of score information, participants
would not waiver in their interpretation of what a ‘do your best’ goal is, but rather, they
would be aware that their highest achieved performance score up to a particular trial, is at
least, their best performance on the given task. Thus, the variability of interpretation of
what a ‘do your best’ goal means for any particular individual will be constrained in a
self-regulatory framework of experimental design because feedback allows participants

to discover the range of task performance.

21

In terms of assigned goals, there is a difference in the referent standard between
assigned goal conditions. A specific, difficult goal has typically been set to the 90th
percentile of task performance (Locke & Latham, 1990a). That is, the specific, difficult
goal is a normative referent. On the other hand, a vague, difficult goal (e.g., ‘do your
best’; ‘list as many as you can’) refers to the self when it is assigned. If feedback is not
provided, then it is difficult for individuals to assess how they are performing with
respect to their self-referenced goal. Additionally, if personal goals are not measured for
the vague, difficult goal group, then it is unknown where on the scale of task performance
the vague, difficult goal falls. That is, it is not known if the level of goal difliculty will
be the same as that for the specific, difficult goal set at the 90th percentile of task
performance. Thus, the different referent standards may be an explanation as to the
observed effect as the two goal conditions may not be of equivalent goal difficulty.

The tasks utilized by Locke et al. (1989) confounded (Study 1; reaction time task)
or constrained (Study 2; listing task) the effect of goal difficulty on task performance.
Thus, the independent and interactive roles of goal difficulty and goal specificity are not
easily discernible from those studies. Given the assumptions of GST, goal difficulty and
goal specificity should interact to affect task performance as the simple efl‘ect between
high and low goal difliculty is greater when goal specificity is high as compared to a
smaller simple effect when goal specificity is low. This interaction effect is not
mentioned in GST, and Locke et al. (1989) decline to elaborate on it in their studies.
However, such an explanation provides a clearer interpretation of the role of goal

specificity in affecting task performance.

22

Despite the espoused mediating role of self-efficacy and personal goals in GST,
most analyses in goal setting studies have collapsed across time. Thus, GST researchers
have not fully accounted for the effects of assigned goals on self-regulatory processes.
This is an important assessment to investigate the lasting effects of assigned goals. For
example, it is expected that individuals given goals of high difficulty will set higher
personal goals than individuals given goals of low difficulty. However, as individuals in
the low goal difficulty condition achieve their assigned goals, they may increase their
personal goals, whereas individuals in the high goal difficulty condition may lower their
personal goals if they cannot achieve their assigned goals. In the presence of feedback
individuals told to do their best may incrementally adjust their personal goals, perhaps,
eventually outperforming the specific, difficult goal group as time passes. Thus, it is
important to investigate the effects of assigned goals over time on process variables that
have shown to affect task performance.

Individuals need social information to assess their achievements and progress
(Bandura, 1991). That is, information regarding how others have performed the
particular task at hand gives another opportunity for an individual to assess his
performance levels. Indeed, everyday achievements are assessed on a relative scale. For
instance, organizations promote individuals who are performing highly relative to their I
peers. In goal setting studies, specific, difficult goals are taken fi'orn the 90"I percentile of
task performance. Thus, a standard is given that only 10 percent of individuals have
achieved. However, as mentioned previously, vague, difficult goals are self-referenced.
So, it would be interesting to investigate what personal goals individuals will choose

when given vague, difficult goals in the presence of normative information. Additionally,

23

the presence of normative information allows vague, difficult goal participants to be on
the same relative scale as the specific, difficult goal participants. That is, vague, difficult
goal participants will be calibrated to a normative scale of task performance versus
simply being calibrated to a self-referenced scale of task performance. In fact, this study
sets out to demonstrate that calibration moderates the relationship between goal
difficulty, goal specificity, and task performance.

The design assumptions in GST leave out potentially important theoretical
mechanisms that can provide explanations for why specific, difficult goals lead to better
task performance than vague, difficult goals. Immediately, it is evident that one
potentially important theoretical mechanism that can be gained from a broader self-
regulatory framework is being able to calibrate one’s performance goals to a standard of
performance. That is, the lack of knowledge of score information in the vague, difficult
goal condition prohibits these participants fiom calibrating their goals (e.g., “do your
best”) to some referential standard.

The explanation offered here rests on the notion that participants have trouble
calibrating vague, difficult goals to the task that they are perfomring on, and hence, their
task performance sufiem compared to the specific, difficult goal group. Notice, though,
that knowledge of score information for the vague, difficult goal group only calibrates
individuals to a self-referential standard of performance. In other words, the combination
of a vague, difficult goal (e.g., “do your best”) and knowledge of score information
calibrates an individual to their previous task performance results. However, in the
presence of normative task performance information, an individual in the vague, difficult

goal group can be calibrated to other individuals’ performance results on the given task

24

(Bandura, 1991). In this case, the performance standard is not self-referential task
performance, but rather, the performance standard is normative task performance. Thus,
the ambiguity of the vague, difficult goal is not related to one’s interpretation of what that
goal means, as it is proposed in the current explanation by Locke and Latham (2002), but
rather, in one’s translation of such a goal into a task performance standard.
Model Development and Hypotheses

Model and Independent Variables

Goal Difficulfi. The difficulty of a goal, if the goal is quantitative, is assessed on
a normative continuum where the majority of people performing a given task can achieve
easy goals and only a small minority of people can achieve difficult goals. Thus, when
quantitative goals are assigned to individuals, these goals are on a normative difficulty
continuum. Goal setting research has consistently demonstrated that goal difficulty for
quantitative goals has a linear relationship to task performance within achievable
performance levels (Locke, 1982). The difficulty of qualitatively stated goals has
generally been assessed on an implicit continuum of difficulty for a given task (e.g.,
Locke et al., 1989). If the task was to list improvements that can be made to a university,
then an easy goal could be to list a “small” number of improvements where a difficult
goal could be to list a “large” number of improvements to the university. Although there
has been much less research on qualitatively stated goals than quantitatively stated goals,
the research that exists indicates that the relationship between goal difficulty and task
performance is still linear (Locke & Bryan, 1968; Locke, Mento, & Katcher, 1978).
Locke and Latham (1990) propose that the effect of goal difficulty on task performance is

mediated by personal goals and self-efficacy.

25

Goal Sgificity. Goal specificity serves as a proxy for the clarity of a goal. That
is, a point goal is very specific, a range of quantitative goals is somewhat specific, and a
verbal goal is not very specific. There have been few studies that have examined the
independent effects of goal specificity in the goal setting literature. Two studies that did
separate goal specificity fiom goal difficulty argued that goal specificity does not affect
task performance per se, but rather, it affects the variance of task performance
(Experiment 1 and Experiment 2 fiom Locke et al., 1989). Locke and Latham have made
the same argument in theoretical works (Locke & Latham, 1990a; Latham & Locke,
1991). Research by Wright and Kacmar (1994) indicates that goal specificity can affect
goal commitment such that specific goals lead to higher goal commitment than non-
specific goals. Furthermore, Wright and Kacmar (1994) demonstrate that goal specificity
is related to personally set goal change such that specific goals lead to smaller goal
changes than non-specific goals.

Normative Information. Normative information displays where different levels of
task performance fall in the distribution of task performance of a group of people who
have performed the task. Normative information effects have been argued to be
important because normative information can result in social comparisons, which relates
to the setting of goals around those social comparisons (Bandura, 1991; Mitchell,
Rothman, & Liden, 1985; Festinger, 1954). As Bandura (1991) points out, behavior is
more easily self-regulated when there are clear indicators of adequacy. Normative
information regarding task performance provides an individual the opportunity to assess
where her task performance falls relative to other people who have performed the task,

and thus, provide an opportunity to assess her task performance on a normative scale for

26

the task. Normative information more clearly defines the metric of task performance for
an individual than does only receiving feedback regarding one’s past performance. This
is because normative information gives the range of possible outcomes that have been
achieved by others whereas only receiving feedback regarding one’s past performance
gives the range of outcomes achieved by a particular individual. Bandura and Jorn'den
(1991) also demonstrated that different social comparison conditions have differential
effects on self-regulatory processes such as goals, self-efficacy, and self-satisfaction.
Those participants who experienced progressive mastery (i.e., this group was
outperformed by a comparison group early in the study, but then performed better than
the comparison group later in the study) demonstrated the greatest gains in self-efficacy,
goals, self-satisfaction, and task performance over three time assessments.

Meyer and Gellatly (1988) in Study 1 demonstrated that assigned goals affected
perceived norms (i.e., student’s belief regarding the average performance of other
students on the task), which, in turn, influenced performance expectancy and
performance valence. This is evidence that assigned goals can serve as indicators of what
other people have achieved on a given task. That is, in fact, a specific, difficult goal is an
indicator of what other pe0ple have achieved on a given task because specific, difficult
goals are set at the 90'h percentile of task performance. In Study 2, Meyer and Gellatly
(1988) manipulated assigned goals (i.e., easy, specific vs. difficult, specific vs. no goal)
and normative information on a brainstorming task such that people in the low norm
condition were told that the average score on the given task was 1 point higher than the
easy, specific assigned goal and people in the high norm condition were told that the

average score on the given task was the same as the difficult, specific assigned goal. The

27

third norm group was not given norm information. The results indicated that the
performance norm manipulation influenced perceived norms (measured as in Study 1)
whereas assigned goals did not affect perceived norms. Furthermore, assigned goals did
affect personal goals in Study 1, but assigned goals did not affect personal goals in Study
2. The direct effect of perceived norms to personal goals was not tested in their path
analytic models. Although, the correlations between perceived norms and personal goals
were r = .73 and r = .71 in Study 1 and Study 2, respectively, whereas the correlations
between assigned goals and personal goals were r = .59 and r = .29 in Study 1 and Study
2, respectively. The correlation between manipulated performance norms and personal
goals in Study 2 was r = .42. These relationships indicate the relationship between
normative information and personal goals and how assigned goals are perceived
normatively by participants.

Other results regarding perceived normative information indicate that manipulated
normative information affects task performance when it is more proximal than assigned
goals and that assigned goals affect task performance when it is more proximal variable
than normative information (Barley & Erez, 1991). Garland (1983) manipulated assigned
goals (i.e., easy, specific vs. difficult, specific) and normative information such that
participants were told that 95%, 50%, or 5% of previously tested students were able to
beat the assigned goals. The results indicated that the there was interaction between
assigned goals and normative information that explained incremental variance to the
main effects on personally set goals. The interaction indicated that participants in the
easy, specific goal condition set increasingly higher personal goals as normative

information indicated higher standards of goal attainment by other participants whereas

28

participants in the difficult, specific goal condition decreased their personal goals as
normative information indicated higher standards of goal attainment by other
participants.

The above studies (i.e., Garland, 1983; Meyer & Gellatly, 1988; Earley & Erez,
1991) manipulated the perception of normative information such that the normative
information was manipulated around the assigned goal. That is, in all three studies
participants were not told veridical information regarding the task performance of other
people who have performed the tasks in the experiments. This study will include
conditions where people will have access to veridical normative information regarding
task performance.

W

This section is divided into three parts. First, the hypotheses between the three
independent variables (i.e., goal difficulty, goal specificity, and normative information)
and task performance will be proposed as the focus of most goal setting studies has been
on the relationship between assigned goals and task performance. Then, a second section
will propose hypotheses between the independent variables and mediating variables of
the relationship between the independent variables and task performance. This is
followed by hypotheses regarding the effects of the process variables on task
performance. Figure 1 presents a heuristic model of the hypotheses proposed in the
present study. However, the main purpose of the present study was not to test this model
as a whole via structural equations modeling, but rather more directly test how assigned
goals afi‘ect task performance. Thus, Figure 1 should simply be used as a guide to

understand how this process is carried out.

29

 

’ Self-

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Efficacy
Goal Personal Effort Task
Difficulty Goal Setting Performance
Normative
Information

 

 

 

 

 

Goal
Specificity

 

 

 

Figure 1. Heuristic model of the propositions of GST

The Effects of the Indgmndent Variables on Task Performance
GST has argued and goal setting studies have consistently demonstrated that high

goal difficulty leads to higher task performance than low goal difliculty (Locke et al.,
1981; Garland, 1982; Garland, 1983; Locke & Latham, 2002). Locke and colleagues
conducted several experiments to demonstrate that the relationship between goal
difficulty and task performance was linear rather than curvilinear as it was proposed by
other researchers (Locke, 1966; Locke & Bryan, 1968; Locke, 1982; see Steers & Porter,
1974 also).

Hypothesis 1 : Goal difi‘iculty afiects task performance such that high goal

difiiculty leads to higher task performance.

The role of goal specificity independent of goal difficulty has been much less

studied in goal setting studies. Locke et al. (1989) argued that goal specificity affects

30

variability of task performance and found statistical support for their assertion across two
experiments. Wright and Kacmar (1994) provided additional empirical support that non-
specific goals are associated with higher task performance variability than specific goals.

HypothesisZ: Goal specificity aflects task performance variability such that high

goal specificity reduces task performance variability.

Most studies in the goal setting literature have not separated the effects of goal
specificity and goal difficulty on task performance. Rather, high goal specificity, high
goal difficulty are often compared to low goal specificity, high goal difficulty (e.g., “do
your best” goals). Notice that this is a comparison between two cells in a four cell
orthogonal design. The statistically significant difference between high goal specificity,
high goal difficulty and low goal specificity, high goal difficulty can be referred to as a
simple effect (Keppel & Wickens, 2004). However, given the assumption in the goal
setting literature that goal specificity does not affect task performance, the difference
between the high goal specificity, low goal difficulty and low goal specificity, low goal
dificulty should not be statistically significant as the driver of task performance should
be goal difficulty and not goal specificity. The lack of a simple effect between the latter
two cells indicates that an interaction exists between goal specificity and goal difficulty
because the simple effects of one independent variable (i.e., goal specificity) is not the
same at all levels of the other independent variable (i.e., goal difficulty; Koppel &
Wickens, 2004). Thus, the prediction that high goal specificity, high goal difficulty leads
to higher task performance as compared to low goal specificity, high goal difficulty is an
interaction whereby goal specificity moderates the relationship between goal difliculty

and task performance.

31

Hwothesis 3: Goal specificity will moderate the relationship between goal
difliculty and task performance such that high goal specificity leads to higher task
performance when goals are of high drfiiculty than when goals are of low
difliculty more than when goals are of low specificity.

If the manipulations of most goal setting studies are examined closely, it can be
seen that comparative groups do not only differ on the types of goals (i.e., specific,
difficult vs. ‘do your best’) that they are given, but these goals also differ in the way that
they are determined. Specific, difficult goals are set at the 90th percentile of task
performance (e.g., Locke et al., 1989). Meyer and Gellatly (1988) argue that assigned
goals can be used as normative reference standards. That is, the assigned goal will serve
as a norm for evaluating the level of performance that is attainable. In fact, a specific,
difficult goal is a normative reference point that refers to the fact that only 10 percent of
people who perform on this task score higher than the assigned goal. On the other hand,
a ‘do your best’ goal will be perceived by participants as a self-referential goal. That is,
assigned goals in GST have not only been different in terms of specificity, but have also
been different in terms of the referent standard. Giving ‘do your best’ participants
normative information regarding task performance could serve to put the two goals on
equal grounds. Social information processing theory would suggest that social
comparisons are made when normative information is present (F estinger, 1954). Hence,
when participants are given normative information they will tend to compare themselves
to others.

Goal setting studies have assumed that ‘do yom best’ is on the same level of goal

difficulty as specific, difficult goals (Locke & Latham, 1990a). If that is the case, then

32

participants who are given normative information with a ‘do your best’ goal would be
expected to select goals close to the 90th percentile of task performance. On the other
hand, if normative information is combined with specific, difiicult goals no significant
increase in task performance should be observed because the assigned goal level is close
to the highest that people can score on this task. Additionally, it is expected that task
performance variability will be affected by normative information as the variability in the
personal goals individuals choose will decrease in the presence of normative information
which equates the specific, difficult goal group with the vague, diffith goal group.

Hwothesis 4: Normative information will moderate the relationship between goal

difficulty, goal specificity, and task performance such that the goal difi‘iculty by

goal specificity interaction will only be present in the absence of normative
information.

Hypothesis 5: Normative information will moderate the relationship between goal

specificity and task performance variability such that goal specificity will only

affect task performance variability in the absence of normative information.

To summarize, the hypotheses in this section demonstrate the established findings
in GST, clarify the role of goal difficulty and specificity on task performance, and
propose that a difl‘erence in the reference standard could explain the effects as well. That
is, hypotheses 1 and 2 test the GST findings that goal difficulty has a linear effect on task
performance and that goal specificity affects task performance variability. Hypothesis 3
clarifies the fact that GST, in fact, proposes that goal difficulty and goal specificity
combine to affect task performance. Finally, hypotheses 4 and 5 suggest that the results

could be due to differences in the referent standards of the assigned goals. That is, the

33

assigned goals calibrate individuals to different referent standards, and thus, the
differential effects are observed. The next section will examine the process by which the
effects of the independent variables are expected to manifest.
The Effects of th_e Independent Variables on Process Variables

GST and goal setting studies have posited and demonstrated that the effects of
assigned goals on task performance are mediated by personal goal setting and
performance expectancy or self—efficacy (Garland, 1983; Locke & Latham, 1990b; Locke
& Latham, 2002; Meyer & Gellatly, 1988). That is, goals that are of high difficulty will
lead to higher personal goals. Additionally, the correlation between personal goals and
task performance has shown to be higher than the correlation between assigned goals and
task performance (e.g., Meyer & Gellatly, 1988). Remember, though, that most goal
setting studies do not measure the personal goals of the low goal specificity condition,
which has generally been the ‘do your best’ goal group. However, since some goal
setting studies (e.g., Locke & Bryan, 1968; Locke, Mento, & Katcher, 1978) have
demonstrated that non-specific, difficult goal groups outperform non-specific, easy goal
groups, then the expected mechanism should be the same as that for when goal difficulty
is assessed with specific goals. That is, irrespective of goal specificity, goals that are of
high difficulty will lead to higher personal goals being set. Since effort is the major
mediator between personal goals and task performance, a similar effect should be found
on efl‘ort as well.

HJpothesis 6: Goal difficulty affects personal goal setting such that high goal

dzfliculty leads to higher personal goal setting than low goal dtfiiculty.

34

Hwothesis 7: Goal difliculty affects eflort such that high goal dtfliculty leads to

higher eflort than low goal dtfi'iculty.

Hollenbeck and Klein (1987) indicated that goal specificity could affect goal
commitment such that more specific goals would lead to higher goal commitment.
Wright and Kacmar (1994) provided empirical support for this assertion as specific goals
led to higher goal commitment than non-specific goals. However, in the present study
individuals in the low goal specificity conditions will receive feedback on their
performance levels. Thus, it can be argued that individuals’ goal commitment toward
their low specificity goal (e.g., ‘do your best’) will be high as their goal will have been
translated in terms of performance levels on the task at hand. On the other hand,
individuals receiving a specific goal may have higher overall commitment to their goal
than those receiving low specificity goals as the former have a constant standard to
evaluate their performance whereas a ‘do your best’ goal changes depending on previous
performance levels.

Hypothesis 8: Goal specificity affects goal commitment such that high goal

specificity leads to higher goal commitment than low goal specificity.

Interestingly, using Meyer and Gellatly’s (1988) studies, Locke and Latham
(1990) argue assigned goals portray normative information that suggests the expected
performance level for an individual. That is, in Meyer and Gellatly’s (1988) Study 1,
assigned goals affected perceived norms which affected self-efficacy. Thus, Locke and
Latham (1990) argue that assigned goals have a positive effect on self-efficacy. Note,

however, the computed positive correlation in Locke and Latham (1990) comes from

35

studies that only looked at goal difficulty, so it is unclear if goal specificity has any effect
on self-efficacy.

Hypothesis 9: Goal difi‘iculty affects self-efi‘icacy such that high goal difliculty

leads to higher self-efiicacy than low goal difi'iculty.

The interaction of goal specificity and goal difficulty should also be present on
personal goal setting and effort in the same way this interaction is described for task
performance because personal goals and effort are the major mediators between assigned
goals and task performance (Locke & Latham, 1990a). This is to say that the specific,
difficult goal group will set the highest goals and display more effort as compared to the
other goal groups. Note, however, that even though GST expects assigned goals to affect
personal goals, whether this relationship exists is not known for low specificity goals. As
generally, personal goals were not measured when individuals were given low specificity
goals.

Hypothesis 10: Goal specificity will moderate the relationship between goal

difi‘iculty and personal goal setting such that high goal specificity leads to higher

personal goal setting when goals are of high difi'iculty than when goals are of low
difi‘iculty more than when goals are of low specificity.

Hypothesis 11: Goal specificity will moderate the relationship between goal

difliculty and effort such that high goal specificity leads to higher eflort when

goals are of high dzfiiculty than when goals are of low dtfiiculty more than when
goals are of low specificity.

The presence of normative information has shown to affect levels of aspirations

(i.e., personal goals; Frank, 1941). Perceived norm information and manipulated

36

perceived norm information has shown to affect performance expectancy (Meyer &
Gellatly, 1988), self—efficacy, personal goals, and task performance (Erez & Barley,
1991). Bandura and Jourden (1991) showed that individuals that received a progressive
mastery normative information manipulation reported the greatest gains in self-efficacy
over three time assessments. Additionally, those individuals who were told that they
consistently outperformed others reported gains in self-efficacy. Given that goal
difficulty is expected to affect self-efficacy, then normative information may moderate
this relationship as those given low goals will see that higher performance levels have
been achieved by a majority of previous task performers. Thus, individuals given low
goals will expect to achieve higher performance levels in the presence of normative
information.

Hypothesis 12: Normative information will moderate the relationship between

goal dtfi‘iculty and self-efi‘icacy such that goal difficulty will affect self-efi‘icacy

only in the absence of normative information.

As previously explained, Garland (1983) demonstrated that an interaction
occurred between goal difliculty and normative information regarding those assigned
goals. The interaction was such that those individuals who were given a goal of low
difficulty set higher personal goals across each of the normative information conditions
whereas those individuals who were given a goal of high difficulty set lower personal
goals across each of the normative information conditions where the normative
information increasingly indicated that fewer previous individuals had beaten that
assigned goaL Locke and Latham (1990) offered an explanation for this interaction

effect. They write that high normative information (i.e., 5% of other people beat the

37

assigned goal) could have been viewed as a challenge by individuals given a goal of low
difficulty whereas individuals given a goal of high difficulty could have believed that
their goal was too high, and thus, individuals with low assigned goals set higher personal
goals whereas individuals with high assigned goals set lower personal goals. This
crossover interaction (see Keppel & Wickens, 2004) is rather interesting. Lewin, Dembo,
Festinger, and Sears (1944) also offer a similar explanation proposed by Locke and
Latham (1990). Individuals who performed above the average of their group of
comparison tended to decrease their personal goals whereas individuals who performed
below the average of their group of comparison tended to increase their personal goals.
Given that effort is the major mediator between personal goals and task performance, it is
expected that the effects on effort would be similar.

Hypothesis 1 3: Normative task performance information will aflect personal goal

setting such that when normative task performance information is present

individuals will set high personal goals.

Hypothesis I 4: Normative task performance information will affect effort such

that when normative task performance information is present individuals will

display higher effort.

Hwothesis I 5: Normative task performance information will moderate the

relationship between goal difi‘iculty, goal specificity, and personal goal setting

such that the goal drfi‘iculty by goal specificity interaction will only be present in

the absence of normative task performance information.

Hypothesis 1 6: Normative task performance information will moderate the

relationship between goal difi‘iculty, goal specificity, and eflort such that the goal

38

difficulty by goal specificity interaction will only be present in the absence of

normative task performance information.

Goal setting studies have typically averaged across performance trials when
investigating the relationship between assigned goals and task performance (e.g., Locke
et al., 1989). However, as Garland (1983) suggests, if individuals are aware that their
assigned goal is too difficult, then they will tend to set lower personal goals. Similarly, if
individuals are aware that their assigned goals are too easy, then they will tend to set
higher personal goals. Thus, in the presence of normative information, it is expected that
all individuals, whether they are given easy or diflicult goals to begin with, will converge
upon similar personal goals over time. This provides an explanation as to why normative
information will negate the interaction of goal difliculty and specificity on personal
goals, and thereby, on task performance as welL That is, normative information
calibrates the goal difficulty by goal specificity conditions to the same referent standard.
Once individuals are calibrated by normative informative they can better assess
achievable performance levels.

Hypothesis 1 7: In the presence of normative information individuals fiom each of

the goal difficulty by goal specificity conditions will converge to the same

personal goals over time.
The Effects of the Process Vaflbjgs on '[a_sk Performing

Self-emcacy has been proposed to affect personal goal setting and task
performance in GST (Bandura & Jourden, 1991; Barley & Erez, 1991). Individuals that
believe they have the capabilities to perform well on a given task will indicate so by

setting high personal goals. Furthermore, individuals’ belief that they can perform well

39

also relates directly to task performance (Bandura, 1986; Bandura, 1991). High self-
efficacious individuals put forth the necessary effort to achieve high task performance.
Personal goal setting has been the most investigated mediating variable between assigned
goals and task performance in goal setting studies. Naturally, higher personal goals have
led to higher task performance (Garland, 1982; Garland, 1983; Garland, 1984; Meyer &
Gellatly, 1988). GST suggests this relationship will be carried out through efl‘ort (Locke
& Latham, 19903).
Hypothesis 18: Self-eflicacy will aflect personal goal setting such that higher self-
efficacy leads to higher personal goal setting. Self-efi‘icacy will afirect eflort
setting such that higher self-efficacy leads to higher effort. Self-eflicacy will afi'ect
task performance such that higher seif—efi‘icacy leads to higher task performance.
Personal goal setting will aflect efi'ort such that higher personal goals will lead to
higher effort. Personal goal setting will aflect task performance such that higher
personal goals will lead to higher task performance. Finally, effort will aflect task

performance such that higher eflort leads to higher task performance.

40

METHOD

Design
The present study utilized a fully-crossed 2 x 2 x 2 factorial design with repeated

measures. The first factor is goal difficulty, which includes two levels: high goal
difficulty and low goal difficulty. The second factor is goal specificity, which includes
two levels: high goal specificity and low goal specificity. Finally, the third factor is
calibration, which includes two levels: normative task performance information is present
and normative task performance information is not present.
M

The task that was used in this study was a version of TANDEM (Dwyer, Hall,
Volpe, Cannon-Bowers, & Salas, 1992), which is a computer-based radar tracking
simulation. In the version of TANDEM that was used for this experiment participants
monitored twenty-two contacts on the radar screen. Participants had to make four
decisions in total for each contact. Each of the first three decisions had three answer
choices, and the last decision had two answer choices. Correct decisions were based on
selecting'the correct answer choice based on the cue for that particular sub-decision.
Thus, there were three cues to learn regarding the contacts. Participants received 100
points for correct decisions and lost 100 points for incorrect decisions.
Participants

There were 526 psychology undergraduates from a large Midwestern rmiversity in
this study who completed the study for course credit. There were 317 females (60 %),

and the average age of the sample was 19.50 years old.

41

Procedure

This study took about two hours to complete: thirty minutes of learning contact
cues on the Web and 1.5 hours in the laboratory. Participants were randomly assigned to
one of the eight conditions (see below) once they came to the laboratory. Given that skill
acquisition is not of interest in this study, participants were trained to quickly remember
the cue sets for each decision. Upon signing up for the study online and at the beginning
of the laboratory session, participants were instructed on how to make correct decisions
to process targets and how different cues relate to different answer choices.

The cycle of reviewing the task cues manual, performing TANDEM, and
receiving feedback on one’s performance was followed for each of the ten trials. The
task cues manual was reviewed for up to 1.5 minutes for the first trial and thirty seconds
for the nine trials that followed, each performance trial was 4 minutes, and performance
feedback was reviewed after each trial. prarticipants were in the normative information
conditions, then they also saw the normative performance information on their
performance feedback screens after each trial. Following performance feedback,
participants were reminded of their assigned goal. Goal commitment, personal goals,
self-efficacy, and effort were then assessed.

Manipulations

Goal difficulty. The level of goals assigned to participants was based on pilot
data that used the exact specifications of TANDEM described above. Goal setting
studies have traditionally set the high goal difficulty level at the 90‘“ percentile oftask

performance (Locke et al., 1989; Locke & Latham, 1990a). However, there has not been

42

a traditional percentile level defined for lower goal difficulties. Wright and Kacmar
(1994) utilized the 10th percentile as an easy goal condition, and for the sake of symmetry
the same criterion was employed. In this study, 1700 points was the 90th percentile of
task performance and 800 points was the 10th percentile.

In terms of setting the difficulty level of non-specific goals, there is not much
guidance in the goal setting literature as to how to assign such goals. Typically, these
goals have been adjusted for the type of task that will be utilized (e.g., Locke et al., 1989;
Wright & Kacmar, 1994). The expression “get a high score” was assigned to represent
high goal difficulty since this was most appropriate for the task that was utilized in this
study. Similarly, the expression “get a low score” was assigned to represent low goal
difficulty. Wright and Kacmar (1994) used the similar expressions for their high and low
non-specific goal conditions.

Goal mificity The specificity of the goal was operationalized by giving
participants either a quantitative goal or a qualitative goal. Participants in the high goal
specificity conditions were told their goal is to obtain a numeric score set at the
appropriate goal difficulty. On the other hand, participants in the low goal specificity
conditions were told their goal in verbal form as described above. The exact phrases were
as follows:

1a. Easy, specific goal: “Your assigned goal is to get a score of 800 points.”

lb. Difficult, specific goal: “Your assigned goal is to get a score of 1700
points.”

2a. Easy, non-specific goal: “Your assigned goal is to get a low score.”

2b. Difficult, non-specific goal: “Your assigned goal is to get a high score.”

43

Calibration. Calibration was specified by giving participants normative task
performance information or not giving participants normative task performance
information. Normative information was presented to individuals on the same screen as
when they observed their feedback. The distribution of task performance was obtained
from pilot data where individuals were operating under a “do your best” goal. The
normative information was presented as displayed in Table 2.

Table 2. Presentation of normative task performance information.

You can use the table below to compare your score with how others have scored on this
simulation.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Comparison Bttrn Bttm Bttm Bttm Avg Top Top Top Top Top

10% 20% 30% 40% 40% 30% 20% 10% Score
Percentile 10 20 30 40 50 60 70 80 90 100
Score 800 1000 1100 1200 1300 1400 1500 1600 1700 2000+
Measures

The laboratory part of this study included nine measurement time points after
each trial. The variables of interest were he goal commitment, self-efficacy, personal
goals, effort, and task performance. Locke and Latham (1990a) recommend that all of
these variables be measured when nmning a goal setting experiment.

Ability. Locke and Latham (1990a) recommend that participants’ abilities be
controlled for on the task. Hence, participant’s ability was assessed by performance on
the first performance trial. The first performance trial was considered to be an
appropriate estimate of participants’ natural ability on the task as none of them had
previous experience with this task.

Goal commitment. The measure of goal commitment was based on items fi'om
Hollenbeck, Klein, O’Leary, and Wright (1989). This measure assesses a participant’s

commitment or valence to the assigned goal. An example item was, “I think my

 

ASSIGNED goal is a good goal to shoot for.” The average Cronbach’s alpha across trials
was .70. Appendix B presents the complete measure used in the present study.

Self-efficacy. Self-efficacy was assessed by three items that have been previously
used in research utilizing this task (e.g., DeShon, Kozlowski, Schmidt, Milner, &
Weichmann, 2004). Self-efficacy could not be measured as suggested by Locke and
Latham (1990) because this would give participants in all conditions information on the
range of performance. This would confound one of the benefits of having normative
information. An example item was, “I am certain that I can manage this simulation.” The
average Cronbach’s alpha across trials was .91 . Appendix B presents the complete
measure used in the present study.

Personal goals Participants were asked to type in the goal that they are striving
for after being presented with feedback from the previous trial. Specifically, participants
were asked: “What is yoru' PERSONAL goal for the next trial?”

Egg. Effort was assessed by three items that have been previously used in
research utilizing this task (e.g., DeShon, et al., 2004). An example item was, “I will put
a great deal of effort into my performance on this simulation in the next trial.” The
average Cronbach’s alpha across trials was .96. Appendix B presents the complete
measure used in the present study.

Task Mormance. A participant’s performance score was based on the number of
correct and incorrect decisions that he or she made. One hundred points were received

for correct decisions and one hundred points were lost for incorrect decisions.

45

RESULTS

The results will be presented in the order of the presentation of the hypotheses.
The first set of hypotheses pertained to the effects of assigned goals and normative
calibration on task performance. The second set of hypotheses pertained to how the
effects of assigned goals and normative calibration may be carried out through process
variables. And finally, process variables were linked to task performance. All measures
(i.e., goal commitment, self-efficacy, personal goals, and task performance) were
averaged across trials 5 through 10. These trials were chosen because the distribution of
task performance remained the same starting with trial 5 across all conditions. Thus, the
results of analyses on task performance were not a result of task learning. Participants
were informed that negative scores do not count toward achieving their assigned goals.
Thus, any participant that scored negative 300 or less in any of these trials was removed
from all analyses as they did not follow instructions. A score of negative 300 was chosen
because by trial 5 all participants had learned how to play the simulation, and thus, were
purposely not following instructions if they processed three or more contacts incorrectly.
This assertion is supported by the fact that 439 (84%) participants answered correctly all
13 questions of the knowledge test the last time they took it before trial 3 and another 52
(10%) participants answered 12 questions correctly. Thus, Table 3 describes the number
of participants who were removed fi'om each condition as a result of not following
instructions. The final sample size was 480 people. Table 4 presents how these 480
people were distributed across conditions. All of the dependent variables were analyzed

using a fully-crossed 2 (Goal Difficulty) x 2 (Goal Specificity) x 2 (Normative

46

Calibration) model with performance from the first trial as a covariate in the analyses on

 

 

task performance.
Table 3. The number of participants who were removed fiom each condition as a result
of not following instructions.
No Norms Norms
Goals Non-specific Specific Non-specific Specific
High Difficulty 7 5 l 2
Low Difficulty 9 5 9 8

 

Table 4. The final number of participants in each condition.

 

 

No Norms Norms
Goals Non-specific Specific Non-specific Specific
High Difficulty 58 68 67 64
Low Difficulty 50 60 56 57

 

The Effects of the In_d£pendent Variables on Tas_k Performapcg

Hypothesis 1: The first hypothesis pertains to one of the central findings from
GST and it stated that individuals given goals of higher difficulty would have higher task
performance than individuals given goals of lower difficulty. The analysis found that
high goal difficulty, (M = 1569.96, SE = 24.65), led to higher task performance,
(F(1,471)= 143.17, p < .001, (02 = .23), than low goal difficulty, (M = 1136.64, SE =
26.48). Thus, hypothesis 1 was supported.

Hypothesis 2: The second hypothesis stated that individuals given goals of higher
specificity would have lower task performance variability than individuals given goals of
lower specificity. Locke et al. (1989) utilized Scheffe’s procedure to test for differences
in variance (W iner, 1971). The analysis reported here replicates that same procedure as
reported in Locke et al. (1989). Participants within each condition were randomly
assigned to four groups. Then the variance of the averaged task performance scores was

computed for each randomly assigned group. These variances were then converted to

47

logs. Thus, for this analysis, each condition had a sample size of 4. A one-way AN OVA
found that high goal specificity, (M = 4.98, SE = 0.07), led to lower task performance
variability, (F (1,30) = 9.11, p < .001), compared to low goal specificity, (M = 5.30, SE =
0.07).

The homogeneity test of variance is a second method of testing this hypothesis.
For this analysis, the original averaged task performance scores were analyzed across the
two goal specificity conditions, and it was found that the error variances of the two goal
specificity conditions were not equal (F (1 ,478) = 93.45, p < .001) such that those who
received goals of high specificity (SE = 32.88) had lower error variance compared to
those who received goals of low specificity (SE = 34.14). Thus, hypothesis 2 was
supported.

Hypothesis 3: The third hypothesis gives information on whether the goal
difficulty effect is differential across goal specificity and it stated that goal specificity
would moderate the relationship between goal difficulty and task performance such that
high goal specificity would lead to higher task performance when goals are of high
difficulty than when goals are of low difficulty more than when goals are of low
specificity. The analysis found that goal specificity did moderate, (F(1,471) = 108.41 , p
< .001 , (02 = .19), the relationship between goal difficulty and task performance.
However, the pattern of the interaction was not according to our expectations. Goal
specificity moderated the relationship between goal difficulty and task performance such
that only low goal specificity led to higher task performance when goals were of high

difficulty than when goals were of low difficulty. This effect is shown in Figure 2. The

48

95% confidence intervals indicate that only the low difficulty, non-specific goal group

was different fiom the other three goal groups.

 

l Goal Difficulty by Goal Specificity Interaction on Task
Performance

1800.00

1600.00 -

1400.00 ~> - ,

1200.00 . +Non-spcciflc
100000 - Specific
800.00

600.00 -

400.00 .

 

Task Performance

Low High
Goal Difficulty

 

Figure 2. Goal difficulty by goal specificity interaction on task performance.

Hypothesis 4: The fourth hypothesis indicated that there would be a boundary

 

condition on the goal difficulty by goal specificity interaction and it stated that normative
information would moderate the relationship between goal difficulty, goal specificity, and
task performance such that the goal difficulty by goal specificity interaction would only
be present in the absence of normative information. The analysis found that normative
information did moderate, (£11,471) = 3.32, p = .07, 002 = .01), the relationship between
goal difficulty, goal specificity, and task performance. However, the pattern of the
interaction was not according to our expectations. Instead, normative information simply
increased task performance for each combination of goal difficulty and specificity. This
effect is shown in Figures 3 and 4. The 95% confidence intervals indicate that in each
normative information condition only the low difficulty, non-specific goal group was

different fiom the other three goal groups.

49

 

1800.00

Task Performance

 

Goal Difficulty by Goal Specificity Interaction on Task

1600.00 ~
1400.00 1
1200.00 ,. *
1000.00 .
800.00 -~
600.00 ~
400.00 -

Performance without Norms

-o-Non-spccific
Specific

 

Low High
Goal Difficulty

 

- , , . , , in n , , , , l
Figure 3. Goal difficulty by goal specificity interaction on task performance without
norms.

 

1800.00

Task Performance

 

Goal Difficulty by Goal Specificity Interaction on Task

1600.00 -
1400.00 -
1200.00
1000.00 -
800.00 ~
600.00 -
400.00 -

Performance with Norms

-O-Non-spccific

- Specific

 

Low High
Goal Difficulty

 

Figure 4. Goal difficulty by goal specificity interaction on task performance with

norms.

Hypothesis 5: The fifth hypothesis stated that normative information would

 

moderate the relationship between goal specificity and task performance variability such

that goal specificity would only affect task performance variability in the absence of

normative information. Following Locke et al. (1989), a two-way AN OVA found that

normative information did not moderate, (F(1 ,28) = 0.90, p = .35), the relationship

50

between goal specificity and task performance variability. However, the homogeneity
test of variance did find that the error variances of the conditions were not equal
(F(7,472) = 22.29, p < .001). The pattern of the standard errors did follow the

expectation, and it is shown in Figure 5.

 

Goal Specificity by Norms Interaction on Standard Errors of

 

 

 

Task Performance
W
E
a 50.00 ~ ~ . . __________. _ ..
E
g 49.00 ,
1: 48-00 ' -0-NONorms
'5 47.00 . - ~ —- - -.-. ,, “-“Norms
1- .. .
"6 46.00 . \
8
g 45.00 ,,
I“
g Non-specific Specific
“g GoalSpecificity
3

Figure 5. Goal specificity by norms interaction on standard errors of task
performance.

The Effects of the Ind ndent Variables on Process Variables

 

Hypgthesis 6: The sixth hypothesis stated that goal difficulty would affect
personal goal setting such that high goal difliculty would lead to higher personal goal
setting than low goal difficulty. The analysis found that high goal difiiculty, (M =
1696.27, SE = 28.26), led to higher personal goal setting, (F(1,467) = 196.1 1, p <.001, (02
= .30), compared to low goal difficulty, (M = 1115.23, SE = 30.38). Thus, hypothesis 6
was supported.

Hyp_othesis 7: The seventh hypothesis stated that goal difficulty would affect
effort such that high goal difficulty would lead to higher effort than low goal difficulty.
The analysis found that high goal difiiculty, (M = 3.86, SE = 0.05), led to higher effort,

51

(F(1,472)= 5.66, p <.05, 002 = .01), compared to low goal difficulty, (M = 3.68, SE -=
0.06). Thus, hypothesis 7 was supported.

Hypgthesis 8: The eighth hypothesis stated that goal specificity would affect goal
commitment such that high goal specificity would lead to higher goal commitment than
low goal specificity. The analysis did not find that high goal specificity, (M = 3.45, SE =
0.05), led to higher goal commitment, (F(1,472) = 1.61, p = .20, (02 = .00), compared to
low goal specificity, (M = 3.35, SE = 0.06). Thus, hypothesis 8 was not supported.

Hypothesis 9: The ninth hypothesis stated that goal difficulty would affect self-
efficacy such that high goal difficulty would lead to higher self-efficacy than low goal
difficulty. The analysis did not find that high goal difficulty, (M = 4.19, SE = 0.04), led
to higher self-efficacy, (£10,472) = 0.10, p = .75, (02 = .00), compared low goal difficulty,
(M = 4.20, SE = 0.04). Thus, hypothesis 9 was not supported.

Hypgthesis 10: The tenth hypothesis stated that goal specificity would moderate
the relationship between goal difficulty and persomrl goal setting such that high goal
specificity would lead to higher personal goal setting when goals are of high difficulty
than when goals are of low difliculty more than when goals are of low specificity. The
analysis formd that goal specificity did moderate, (m ,467) = 96.57, p <.001, (02 = .17),
the relationship between goal difficulty and personal goal setting. However, the pattern
of the interaction was not according to our expectations. Goal specificity moderated the
relationship between goal difficulty and personal goal setting such that only low goal
specificity led to higher personal goal setting when goals were of high difficulty than
when goals were of low difi'rculty. This effect is shown in Figure 6. The 95%

confidence intervals indicate that the low difficulty, non-specific goal group was difi‘erent

52

from the other three goal groups, and the low difficulty, specific goal group was different

from the high difficulty, specific goal group.

 

Goal Difficulty by Goal Specificity Interaction on Personal
Goals

1800.00
1600.00
1400.00 -
1200.00 -
1000.00
800.00 ,
600.00 -
400.00 -

____.._ ._M _._._._ ___

-O-Non-specific

Specific

Personal Goals

 

 

. Low High
| Goal Difficulty

 

Figure 6. Goal difficulty by goal specificity interaction on personal goals.

Hypothesis 11: The eleventh hypothesis stated that goal specificity would
moderate the relationship between goal difficulty and effort such that high goal
specificity would lead to higher effort when goals are of high difficulty than when goals
are of low difficulty more than when goals are of low specificity. The analysis found that
goal specificity did moderate, (F(1,472) = 3.9l,p <.05, 032 = .01 ), the relationship
between goal difficulty and effort. However, the pattern of the interaction was not
according to our expectations. Goal specificity moderated the relationship between goal
difficulty and effort such that only low goal specificity led to higher effort when goals
were of high difficulty than when goals were of low difficulty. This effect is shown in
Figure 7. The 95% confidence intervals indicate that the low difficulty, non-specific goal

group was different fiom the high difficulty goal groups.

53

 

Goal Difficulty by Goal Specificity Interaction on Effort

3.9 a
3.8 " '
3.7 -
3.6 .
3.5 -

—O—Non-specific I

Effort

Specific

 

Low High
Goal Difficulty

 

 

Fig“; 7—'G°—31 “ms; SPfTSPOCifiCitY interaction on effort." _

Hypothesis 12: The twelfth hypothesis stated that normative information would
moderate the relationship between goal difficulty and self—efficacy such that goal
difficulty would affect self-efficacy only in the absence of normative information. The
analysis found that normative information did marginally moderate, (F(1,472) = 3.43, p
= .07, (1)2 = .01), the relationship between goal difficulty and self-efficacy. This effect is

shown in Figure 8.

 

Goal Difficulty by Norms Interaction on Self-Efficacy

4.35 -

i

l

l 4.30 -

l :7 425

I g . +NONorms
z - orms
3 4.15 . \

4.10 -

l 4.05 -

l Low High
l

|

 

Goal Difficulty

 

Figure 8. Goal difficulty by norms interaction on self-efficacy.

54

Hypothesis 13: The thirteenth hypothesis stated that normative information
would affect personal goal setting such that when normative information is present
individuals would set higher personal goals. The analysis found that normative
information, (M = 1490.17, SE = 28.95), led to higher personal goal setting, (F(1,467) =
16.56, p = .001, 002 = .03), compared to no normative information, (M = 1321.32, SE =
29.72). Thus, hypothesis 13 was supported.

Hypothesis 14: The fourteenth hypothesis stated that normative information
would affect effort such that when normative information was present individuals would
display higher effort. The analysis found that normative information, (M = 3.77, SE =
0.05), did not lead to higher effort, (F(1,472) = 0.00, p = .98, 032 = .00), compared to no
normative information, (M = 3.77, SE = 0.06). Thus, hypothesis 14 was not supported.

Hypothesis 15: The fifteenth hypothesis stated that normative information would
moderate the relationship between goal difficulty, goal specificity, and personal goal
setting such that the goal difficulty by goal specificity interaction would only be present
in the absence of normative information. The analysis found that normative information
did not moderate, (F(1,467) = 2.24, p =.14), the relationship between goal difliculty, goal
specificity, and personal goal setting. However, the pattern of the interaction was similar
to the three-way interaction found on task performance. This pattern is shown in Figures

9and 10.

55

 

l Goal Difficulty by Goal Specificity Interaction on Personal
l Goals without Norms

1800.00
1600.00 ~

1400.00 -- ~ ~

1200.00 - -o-Non-5pecific
1000.00 - . , Specific
800.00 - . ..

; 600.00 -

F 400.00

l Low High
Goal Difficulty

Personal Goals

 

Figure 9. Goal difficulty by goal specificity interaction on personal goals without
norms.

 

Goal Difficulty by Goal Specificity Interaction on Personal
Goals with Norms

1800.00

1600.00

1400.00 -

1200.00 ~

1000.00 »-
800.00 -
600.00 -
400.00 .

» Low High

I! Goal Difficulty

t

Figure 10. Goal difficulty by goal specificity interaction on personal goals with

norms.

-o—Non-specific

 

Specific

Personal Goals

 

 

Hyp_othesis 16: The sixteenth hypothesis stated that normative information would
moderate the relationship between goal difficulty, goal specificity, and effort such that
the goal difficulty by goal specificity interaction would only be present in the absence of

normative information. The analysis found that normative information did not moderate,

56

(F (1,472) = 0.76, p =.39), the relationship between goal difficulty, goal specificity, and
effort. Thus, hypothesis 16 was not supported.

Hypothesis 17: This hypothesis stated that in the presence of normative
information individuals from each of the goal difficulty by goal specificity conditions
would set similar personal goals over time. A one-way repeated-measures AN OVA
indicated that the time by condition interaction, (F (15,1 160) = 0.39, p = .98), was not
significant. Thus, hypothesis 17 was not supported. Figure 11 clearly displays a lack of

convergence between the different conditions.

 

Time by Norms Interaction on Personal Goals

1

l

i

i 2000.00 - ,, ~ ~~ ~

3 1800.00 - “or i :41- —

1600.00 - Ker—e a ““1

1400.00 , , ,

1200.00 -

1000.00 -
800.00 -- /
600.00 - ~
400.00 -

E 1 2 3 4 5 6

l

 

 

—o—High Specific
.» . .. High Non-specific
+Low Specific

 

Personal Goals

Low Non-specific

 

Time

 

i , , W , A
Figure 11. Time by norms interaction on personal goals.

The Effects of £15 Process Variables on Task Performance

Hmthesis 18: This hypothesis stated that self-efficacy would be positively
related to personal goals, effort, and task performance; personal goals would be positively
related to effort and task performance; and effort would be positively related to task
performance. The correlations in Table 5 indicate that self-efficacy was positively
related to personal goals (r = .24, p < .001), effort (r = .43, p < .001), and task

performance (r = .23, p < .001), and personal goals was positively related to effort (r =

57

.33, p < .001), and task performance (r = .89, p < .001). Finally, effort was positively

related to task performance (r = .27, p < .001).

Table 5. Means, standard deviations, correlations, and reliabilities of all variables.

 

 

M SD 1 2 3 4 5 6 7 8
1.Goal

Difficulty 0.54 0.50 -

2.Goal

Specificity 0.52 0.50 -.01 -

3.Norms 0.51 0.50 .00 -.05 -

4.Goal ,,

Commitment 3.42 0.89 .23 .05 -.01 .95

5'86”" 4.19 0.64 -.02 .10* -.04 .32* .94

efficacy

35:1?“ 1436 608 .46* .33* .12* .15* .24* .97

7.Effort 3.78 0.86 .10* .07 .00 56* .43* 33* .95
“as“ 1378 546 .39* .32* .ll“ .11* .23* .89* .27* .96
Performance

 

Note. *p < .05. N = 480; except for personal goals where N = 475. l = High goal
difficulty, specific goals, and norms; 0 = Low goal difficulty, non-specific goals, and no
norms. The reliabilities on the diagonal are test-retest reliabilities across trials 5 through

10.

58

DISCUSSION

This study was interested in testing goal setting theory fiom a broader self-
regulatory perspective (Karoly, 1993). More specifically, this study was interested in
understanding the relationship between goal difficulty, goal specificity, and task
performance. Goal setting theory stipulates that goal difficulty positively affects task
performance, while specific goals lead to lower task performance variability than non-
specific goals. However, there have only been a few studies (e.g., Locke et al., 1989;
Wright & Kacmar, 1994) that have separately examined the effects of goal difficulty and
specificity. One of the major questions raised in this study was if these effects are clearly
understood because many goal setting studies had several limitations (e.g., differential
feedback across assigned goal conditions) in their research design. This study attempted
to investigate goal setting effects without these limitations. Hypotheses were derived
from GST, but some of the results did not follow expectations. Thus, the results in this
study serve to point out that goal setting effects are not clearly understood, and therefore,
more research is needed. The results ofthis study can serve to inform future research.

Goal Difficulty and Goal, Specificity

One set of hypotheses described effects of goal difficulty and specificity on task
performance, task performance variability, goal commitment, self-efficacy, personal
goals, and effort. Given that most of goal setting research has investigated the effects of
goals on task performance, this will be discussed first, which will then be followed by a

discussion of the results on process variables.

59

Effects on Ta_s_k Perfongnce and Ta_sk Performgnce Valuing

The results confirmed expectations of a main effect of goal difficulty on task
performance such that goals of higher difficulty led to higher task performance than goals
of lower difficulty. However, this first result was averaged over goal specificity
conditions. The goal setting literature indicates that there will be positive effects of goals
on task performance whether or not these goals are specific or non-specific (Locke &
Lathanr, 1990a). However, GST also suggests that specific, difficult goals will lead to
higher task performance than non-specific, difficult goals. Thus, these propositions
suggest that investigating the interaction between goal difficulty and specificity is useful
in order to understand these effects.

In this study, goal specificity was found to be a moderator of the relationship
between goal difficulty and task performance. However, the pattern of the interaction
was not according to our expectations which again were derived fiom GST. As
demonstrated in Figure 2, when goals were non-specific it was the case that goal
difficulty positively affected task performance. However, when goals were specific, then
low and high goal difficulty affected task performance similarly. The first thing to note
from this pattern of interaction is that the specific, dificult goal group did not differ in
task performance fi'om the non-specific, difficult (i.e., “do your best”) goal group. As
mentioned previously, most goal setting studies only gave feedback to the specific,
difficult goal group (e.g., Locke et al., 1981). In the present study, both goal groups
received feedback. Thus, when individuals are given the opportunity to self-regulate by

noting their feedback, then the results of the present study suggest that these two goal

60

groups will not differ. This implies that feedback may account for this effect. The
second thing to note about this interaction is that the positive effect of goal difficulty on
task performance only occurs when goals are non-specific, otherwise the effects are
similar. Given the rarity of organizations wanting individuals to get a low score (i.e.,
non-specific, low difficulty goal manipulation), it might be concluded that in contexts
that allow people to self-regulate, people will tend to perform similarly no matter what
their initial assigned goal (e.g., specific, low difficulty goal manipulation).

Locke et al. (1989) did not find an interaction between goal difficulty and
specificity on task performance in Study 1, but they did find one in Study 2. Study 1 had
a small sample size (N = 16) in each condition, and thus any interaction pattern would
have been difficult to detect statistically. The interaction pattern in Study 2 indicated a
stronger slope from low to high goal difficulty in the specific goal condition as compared
to the non-specific goal condition. Thus, the interaction pattern was different than fiom
the one found in this study as there was only a positive slope from low to high goal
difficulty on task performance when goals were non-specific. Locke et al. (1989)
indicate that the listing task they utilized in Study 2 may not have been the best task to
investigate the goal to task performance relationship as it may have limited task
performance variability. Additionally, the instructions given to participants in Locke et
al. (1989) specific goal conditions asked participants to list “exactly 2 ways, 3 ways, or 4
ways” (p. 280), while in this study we asked participants “to get a score” of 800 or 1700
points. The differences in the slopes fiom low to high goal difficulty when goals are
specific may be explained by participants interpreting these sets of instructions

differentially. The individuals in Study 2 of Locke et al. (1989) may have interpreted

61

their instructions strictly. That is, they achieved their assigned goal and did no more. On
the other hand, the individuals in the present study may have viewed their assigned goals
as a marker to surpass and achieve higher scores. This would also suggest that the
positive slopes fi'om low to high goal difficulty when goals are non-specific, observed in
both studies, may be a result of individuals interpreting their assigned goals strictly.
Since these goals are non-specific individuals would have to pick some referent point as
their goal.

Task performance variability was investigated using the method specified by
Locke et al. (1989). It was found that high goal specificity led to lower task performance
variability compared to low goal specificity. Post-hoc analyses indicated that goal
difficulty also affected task performance variability such that high goal difficulty led to
lower between-person task performance variability compared to low goal difficulty.
These results replicate findings fi'om previous research (Locke et al., 1989; Wright &
Kacmar, 1994). Additionally, post-hoe analyses indicated that the two-way interaction
between goal difficulty and goal specificity was significant such that there was a strong
negative slope from non-specific to specific goals when goal difficulty was low, but no
slope when goal difficulty was high. Just like task performance, the significant
interaction on task performance variability comes fiom the low difficulty, non-specific
goal condition as compared to the other three conditions. If these results are coupled with
the explanations provided for the interaction effect on task performance, then these
results suggest that people in the low difficulty, non-specific goal group did not converge
to the same degree as the other groups on their referent points. That is, the greater

between-person variance for the low difficulty, non-specific goal group is probably a

62

result of those with norms performing around a score of 800 points, while those without
norms performing around a score of 600 points. The performance of the three goal
groups did not change as much (see Figures 3 and 4).
Effects on Goal Commitmenf. Self-efficacy. Personal Goals. and Effort

The effects of assigned goals on task performance are considered to be explained
by goal commitment, self-efficacy, personal goals, and effort (Locke & Latham, 1990a).
Thus, a second set of hypotheses predicted relationships between assigned goals and
these variables. It was not found that goal specificity affected commitment to the
assigned goal. However, post-hoe analyses indicated that goal difficulty did affect goal
commitnrent such that high goal difficulty led to higher commitment to the assigned goal
than low goal difficulty. Additionally, an interaction between goal difficulty and goal
specificity was found such that the positive slope was steeper from low to high goal
difficulty when goals were non-specific as opposed to specific goals. Given that this
measure captures a valence component of the assigned goals, the goal difficulty main
effect result indicates goals of lower difficulty were less valued than goals of higher
difficulty. The interaction result indicates that the difference in how much assigned goals
were valued was greater when there were non-specific goals as opposed to specific goals.

It was not found that goal difliculty affected self-efficacy. However, post-hoc
analyses indicated that goal specificity did affect self-efficacy such that self-eflicacy was
higher when goals were specific as opposed to non-specific. The interaction between
goal specificity and difficulty on self-efficacy indicated that when goals were of low
difficulty there was a steep positive slope fiom non-specific to specific goals, while the

slope was flat when goals were of high difficulty. Thus, the effect of goal specificity

63

occurs when goals are of low goal difficulty. Given knowledge of their task
performance, these results may be interpreted as indicating that the low difficulty,
specific goal group was confident in that they could do well in the simulation as they had
exceeded their original assigned goal, while the low difficulty, non-specific goal group
never exceeded their assigned goal, and thus, were less confident in their performance.
This additionally implies that the high difficulty goal groups did not differ in their self-
efficacy because they were not exceeding their high expectations.

It was found that goal difficulty affected personal goal setting and effort such that
high goal difficulty led to higher personal goal setting and effort compared to low goal
difficulty. The interaction between goal difficulty and specificity on personal goal setting
and effort followed the same pattern as that found on task performance. As demonstrated
in Figures 6 and 7, when goals were non-specific it was the case that goal difficulty
positively affected personal goal setting and effort. However, when goals were specific,
then low and high goal difficulty affected personal goal setting and effort similarly.

Thus, these results suggest that the interaction effect found on task performance is carried
through personal goals and effort.
Normative Calibration

The interaction effects of goal difficulty and specificity were expected to be only
present in the absence of normative information as it was proposed that normative
information would make participants behave similarly across the goal difficulty by
specificity conditions. The results with respect to task performance and task performance

variability will be discussed first, followed by the effects on process variables.

Effects on Task Performance and Task Perform_ance Valafilifi

Normative information did moderate the relationship between goal difficulty, goal
specificity, and task performance. However, the pattern of the interaction was not
according to expectations. As demonstrated in Figures 3 mid 4, the pattern of the goal
difficulty by goal specificity interaction on task performance does not change across
normative information conditions. Instead, the means of each of the goal difficulty by
goal specificity conditions increases, especially the low difficulty non-specific goal
group. Interestingly, the mean of this group in the presence of norms was around the
assigned low specific goal (i.e., 800 points). Additionally, these results are interesting as
the low difficulty non-specific goal group did not increase their performance to the levels
of the other groups even though they rmderstood that their performance levels were at the
10th percentile level. Thus, they are restraining their performance levels, while the low
difficulty specific goal group is not as this group achieves performance levels comparable
to the remaining two groups. This again may be explained by participants in different
goal conditions interpreting their instructions differently where the low difficulty, non-
specific goal group is interpreting as a strict benchmark. In the presence of norms they
may have interpreted the lowest percentile score (i.e., 800 points) as their benchmark, and
thus, the difference in this goal group across normative information conditions.

The analysis on task performance variability did not find that normative
information moderated the relationship between goal specificity and task performance
variability. However, the homogeneity test of variance indicated that the error variances
of these conditions were not equal. Figure 5 indicates that the pattern of the standard

errors in these conditions was according to expectations. That is, standard errors were

65

similar across the goal specificity conditions when norms were presented, but the
standard error for the specific goal group was lower than the standard error of the non-
specific goal groups when norms were not presented. This suits the previous discussion
on task performance variability. When norms were present the non-specific individuals
were better able to converge to striving for the same score.

Effects on Self-efficacy. Person_a_l Goals and Effort

 

Another set of hypotheses predicted how normative information affects the
mediators of the relationship between assigned goals and task performance. It was found
that norms marginally moderated the relationship between goal difliculty and self-
efficacy. That is, when norms were not presented the slope fi'om low to high goal
difficulty was negative, but when norms were presented the slope was positive. Thus,
when norms were not present, the people in the low goal difficulty group believed they
could do well on the simulation as compared to people in the high goal difficulty group
probably because they were exceeding their assigned goal, but when norms were present
the self—efficacy of the low goal difficulty group was lower as compared to the high goal
difficulty group as norms gave each of these groups material for interpreting their
assigned goals.

It was found that normative information affected personal goal setting such that
those presented norms set higher personal goals compared to those not presented with
norms. Additionally, it was expected that normative information would moderate the
relationship between goal difficulty, goal specificity, and personal goal setting. This
three-way interaction was not significant. However, the pattern of the interaction

followed the pattern of this interaction on task performance. That is, those individuals

66

who were presented norms tended to set higher personal goals across all four goal
difficulty by specificity conditions. This suggests that the pattern of the three-way
interaction may be explained by the personal goals that people in the different conditions
are setting. Finally, it was hypothesized that individuals would set similar personal goals
over time across the normative information conditions. However, as demonstrated in
Figure 10, the low difficulty, non-specific goal group never raised their personal goals to
match the personal goals set by the other three groups. Again, this provides an
explanation as to why this group performed the worst and that they may have set a
referent for each achieving their goal and not doing anything beyond it.

The analysis on effort did not find that norms positively affected effort or the
three—way interaction. Thus, this suggests that the effect of normative information on
task performance is only operating through personal goal setting.

Whip between Gig] Commitmeng Self-efficacy, Personal Goals, Effort, and Task
Performance

The final set of hypotheses pertained to the relationships between the process
variables and task performance. As indicated in Table 5, all of the correlations between
these variables were positive. This suggests that the effects of goal difficulty, goal
specificity, and norms on task performance are mediated by these process variables
giving support to the heuristic model presented in Figure 1.

Syngesis of Results

The results of the present study did not always follow expectations. GST states

that goal difficulty positively affects task performance, while goal specificity restricts

task performance variability (Locke & Latham, 1990). This study, though, found that

67

goal specificity also positively affected task performance. The interaction of goal
difficulty and specificity indicated that goal difficulty has differential effects on task
performance when goals are non-specific, while goal difficulty does not have differential
effects on task performance when goals are specific. In fact, only the mean of the low
difficulty, non-specific goal group is different fiom the means of the other three groups.
This same pattern is seen on personal goal setting and effort, which are the two main
mediators of the relationship between assigned goals and task performance. Thus, the
pattern of the interaction on task performance may be explained by the similar patterns of
the interaction exhibited on personal goal setting and effort. Normative information did
not account for this interaction as expected The most important effect of normative
information was to adjust the self-efficacy expectations of the different goal groups.

The results found in the present study do not correspond to findings in other goal
setting research (Locke & Latham, 1990a). For one, there was not a difference between
the specific, difficult goal group and the non-specific, difficult goal group. This
difference may potentially be explained by feedback. That is, most goal setting studies
only provide feedback to the specific, difficult goal group. However, in the present study
both goal groups were provided feedback, and thus, both goal groups were able to self-
regulate. The interaction pattern of goal difficulty and specificity on task performance
and mediating variables differed fiom that found in previous research. Locke et al.,
(1989) found that the goal difficulty effect was stronger when goals were specific.
However, the present study found that the goal difficulty effect exists only when goals
were non-specific. An explanation for this difl'erence may be the way participants

interpreted their assigned goals. In the present study, individuals receiving a non-specific

68

goal may have been interpreting their goals as standards to achieve, while individuals
receiving specific goals may have been interpreting their goals as standards to Stu-pass.
This explanation requires that the non-specific goal group had to set their standard in
some way. When there were no norms, individuals were solely basing their standards on
their own performance, and when there were norms, individuals could base their
standards on the norms. This explanation is supported by the results found in this study.
That is, when there were norms, individuals in the low difficulty, non-specific goal group
adjusted their goals and task performance to the lowest percentile score in the norms.
Examining this goal group’s goals over time showed that they never raised their personal
goals to those of the other assigned goal groups. That is, they chose a standard and did
not vary fiom it.

Limitations and Future Resear_£l_r_

In the present study, it is clear in terms of task performance, the low difficulty,
non-specific goal group behaved differently fiom the other three goal groups. The
present study told participants that negative scores did not count toward achieving their
assigned goal. This was done as a result of initial data collection where participants in
the low difficulty, non-specific goal group were achieving high negative scores when
participants were not told anything about interpreting their assigned goal. The
adjustment in instructions was made because other goal setting studies have used tasks
where the scale of task performance only included positive scores (e.g., Locke et al.,
1989). This is something that needs to be explored firrther as the instructions of negative
scores not counting toward achieving one’s assigned goal were most useful for the non-

specific, low difficulty goal group in interpreting their assigned goal. However, it would

69

be expected that the goal difficulty by specificity interaction would be even more
dramatic as the non-specific, low difficulty goal group would perform on the negative
side of task performance. All of this points again to the idea that the assigned goal
groups are interpreting their goals differentially. That is, one possible explanation for the
pattern of the results in the present study may be a result of how individuals in different
goal conditions are interpreting their assigned goals as either standards to achieve or
standards to surpass.

One way to examine this explanation may be done by the following manipulation:
the goal instructions utilized in the present study may be used versus instructions that
inform participants to get their assigned goal (e.g., “get a low score”) or higher. Given
the present results, any effect would only be expected on the low difficulty side of non-
specific goals. That is, in the group that would receive the instructions utilized in the
present study, it is expected that the goal difficulty by specificity interaction would like
the interaction displayed in Figure 2. However, in the group that would receive the other
instructions, it is expected that the goal difficulty by specificity interaction would not be
present as all four groups would achieve similar task performance. This finding would
add credence to the idea that individuals in the non-specific goal conditions view their
assigned goals as standards to achieve versus standards to surpass.

Another way to examine this explanation may be done by investigating a similar
effect on specific goals only by the following manipulation: the goal instructions utilized
in the present study versus instructions that indicate to achieve exactly the specified score
(e.g., “get exactly a score of 800 points”). Naturally, it would be expected that the

specific, high difficulty goal group will perform similarly across the two sets of

70

instructions. However, if the specific, low difficulty goal group performs as instructed,
then they would perform worse than in the present study. That is, they would get exactly
a score of 800 points and achieve no more compared to the 1600 points they achieved in
the present study. Thus, this finding would additionally indicate that the present results
may be explained by whether participants interpret their goals as performance standards
to achieve versus standards to surpass.

The discussion of present findings indicates another area of future research. One
of the key variables not manipulated in the present study was feedback. That is,
individuals in all conditions were presented with feedback as it was argued that previous
research had prevented self-regulation in non-specific goal conditions by not giving
individuals in those conditions feedback. However, knowledge of results has been
presented as one of the explanations for differences between present findings and
previous research (e.g., Locke et al., 1989). Although previous research has examined
the effects of assigned goals and the presence or absence of feedback (e.g., Becker,
1978), there have been no studies that have examined the effects of goal difficulty and
specificity separated fiom each other and the presence or absence of feedback. In the
present study, individuals that were presented with norms had the choice of choosing
their personal goals based on one’s own performance or based on the normative
information. A study designed to separate calibration based on one’s own performance
versus based on norms may also provide insight to the present findings. Therefore, a
follow-up study may utilize the conditions of the present study where feedback is not
presented to any condition. GST indicates that assigned goals will not have an impact on

task performance beyond initial effects when there is no feedback (Locke & Latham,

71

1990a). Given this, it might be expected that task performance would decrease for all
assigned goal conditions, but even more so for the specific, low difficulty goal group as it
would not be expected for this group to perform beyond the initial effect of this goal
(presumably around 800 points). Thus, the goal difficulty by specificity interaction
would disappear when there is no feedback if individuals in the other goal conditions
decrease their task performance by smaller margins. Additionally, it would be expected
that if one were to compare the non-specific, difficult goal group without feedback to the
specific, difficult goal group with feedback (i.e., the specific, difficult goal group fiom
the present study), then the methodology of most goal setting studies would be replicated
and one should find an effect. These results would suggest that feedback plays an
important role in explaining the effects of assigned goals on task performance.
Conclusion

The present study was interested in taking a closer look at the effects of assigned
goals on task performance and the explanations provided for these effects. Perhaps the
most definitive conclusion of the present study is that our understanding of how different
goal attributes differentially affect task performance is not entirely clear, and thus, more
research is required as the pattern of findings in the present study did not follow
predictions derived fiom GST. However, testable explanations have been proposed to
continue further work in this area. This work, in turn, will surely lead to a more complete
understanding of how different types of goals affect task performance than is currently

known.

72

APPENDICES

73

APPENDIX A

Consent and Debriefing Forms
Informed Consent
Project Title: Decision Making Training
Investigators' Names: Goran Kuljanin & Dr. Steve Kozlowski

Description and Explanation of Procedure:

This research study contains two parts: an online questionnaire and an experiment
in a laboratory. You have by now completed the online questionnaire. In this part of the
research study you will be re-introduced to the simulation, perform the simulation, and
complete measures that are related to the simulation. This study is about computer-based
training, learning, and performance. You will learn how to assess the attributes of radar
contacts and to decide what action should be taken.

Those not interested in this research can seek other alternatives and research
studies for subject pool credit by consulting their instructor or the Department of
Psychology subject pool web site.

Estimated time required: 30 minutes for the online questions [1 Psychology subject
pool credit]. 120 nrinutes for completing the experiment [4 Psychology subject pool
credits].

Risks and discomforts: None anticipated

Benefits: You gain experience with simulation-based training, which is increasingly
used in organizations, and you will learn about the process of conducting psychological
research.

Agreement to Participate:

Participation in this research study is completely voluntary. You also give
permission to the experimenters to assess or verify your ACT/SAT score fiom University
Register. You have been fully informed of the above-described procedure with its
possible benefits and risks. You are flee to withdraw this consent and discontinue
participation in this project at any time without penalty. For those participants who are
receiving course credit for participation, if you choose to withdraw fiom the study prior
to its completion, you will receive credit for the time you have spent in the study (1 credit
per 30 minutes).

The investigator will be available to answer any questions you may have. If, at
any time, you feel your questions have not been adequately answered or you want to
discuss the research, please contact the investigators (Goran Kuljanin, 432-7069; Steve
Kozlowski, 353-8924). Ifyou have questions or concerns regarding your rights as a study

74

participant, or are dissatisfied at any time with any aspect of this study, you may
contact—anonymously, if you wish— Peter Vasilenko, Ph.D., Director of Human
Research Protections, (517)355-2180, fax (5 l 7)432-4503, e-mail irb@nsu.edu, mail 202
Olds Hall, Michigan State University, East Lansing, MI 48824-1047.

You must 18 years of age to participate in this research study. If you agree to
participate, please enter the information requested at the bottom of this form. The reason
you are asked for your name and P11) is to ensure that you receive credit for participating
in the study. Participants' identity data will be kept on a secure server belonging to Steve
Kozlowski until the end of data collection (approximately 3 months) and only the
investigators will have access to the data. Your identity will not be associated with your
responses for any data analyses. Your confidentiality will be protected to the maximum
extent allowable by law.

If you do not consent, then tell the experimenter that you will not be participating in
this research study.

75

Debriefing Form

Thank you for your participation in this investigation. The study in which you
just participated was designed to examine how individuals self-regulate in training-based
simulations. If you are interested in obtaining more information, please contact Goran
Kuljanin at his e-mail address: kuljanin@msu.edu.

The questions you answered have no right or wrong answers and have no
significance other than to help us understand how assigned goals affect task performance.
We will de-identify the data at end of data collection (i.e., end of F all semester 2007), and
we will keep the data confidential at all times. All of the information we collected will be
used to understand how individuals self-regulate in training-based simulations. Your
help will be invaluable in this regard.

If you have any questions about this study or would like to receive a copy of the
results when they are complete, please notify the investigator now. We have tried to
make your experience in this study as interesting for you as possible and are open to
suggestions for improvements. If you have any additional questions about the study or
your involvement in it, contact one of the Principal Investigators listed below. If you have
any questions or concerns regarding your rights as a study participant, or are dissatisfied
at any time with any aspect of this study, you may contact - anonymously, if you wish -
Peter Vasilenko, Ph.D., Director of the Human Subject Protection Programs at Michigan
State University, by phone: (517) 355-2180, fax: (517) 432-4503, email: irb@msu.edu, or
regular mail: 202 Olds Hall, East Lansing, MI 48824.

76

APPENDIX B

Measures

Demomphics:

1]. What is your gender?
-Female

-Male

2]. What is your age?

Declarative Knowledge
The following is a knowledge test about what you have just learned.

1].If a Response is Given, what is the Intent of the contact?
-Military

-Hostile

-Civilian

-Peaceful

2].A submarine may have which of the following characteristics?
-Speed 30 knots

-Speed 40 knots

-Speed 20 knots

-Speed 60 knots

3].Ifa contact’s Intelligence is Private, what Class does this suggest for the contact?
-Air

-Civilian

-Military

-Peaceful

4].Ifa contact’s Speed is 40 knots, what Type is this contact?

-Air

-Surface

-Civilian

-Military

5].If a contact's Intelligence is Platform, what Class does this suggest for the contact?
-Sub

-Civilian

-Military

-Hostile

6].If a Response is No Response, what is the Intent of the contact?
-Surface

-Hostile

-Civilian

-Peaceful

77

7]. If a contact's Speed is 27 knots, what Type is this contact?

-Surface

-Military

-Sub

-Peaceful

8]. Ifthe Intent of a contact is Peaceful, what is the Final Engagement decision?
-Shoot

-Clear

9]. If the Intent of a contact is Hostile, what is the Final Engagement decision?
-Shoot

-Clear

10]. What is the correct order of decisions for each contact?

-Type, Class, Intent, Final Engage

-Class, Type, Intent, Final Engage

-Intent, Class, Type, Final Engage

-Class, Intent, Final Engage, Type

11]. What is the zoom function used for?

-To locate contacts that are not in the current viewing range

-The zoom fimction is not used

12]. How is your score computed for each contact?

-Gain 100 points for making each of the four decisions correctly, and lose 100 points

for making each of the four decisions incorrectly

-Gain 200 points for making each of the four decisions correctly, and lose 200 points

for making all four decisions incorrectly

-Gain 100 points for making all four decisions correctly, and lose 100 points for

making any one of the four decisions incorrectly

-Gain 200 points for making all four decisions correctly, and lose 200 points for

making any one of the four decisions incorrectly

13]. Do scores below zero count toward achieving your goal?
-No
-Yes

Goal Commitment
This set of questions asks you to rate your commitment to the ASSIGNED goal you

just saw. Thinking about your ASSIGNED goal, please use the scale to make your
ratings for these questions.

1]. Quite frankly, I don’t care if I achieve my ASSIGNED goal or not.
2]. I am strongly committed to pursuing my ASSIGNED goal.
3]. I think my ASSIGNED goal is a good goal to shoot for.

78

Self-efficacy
This set of questions asks you to describe how you feel about your capabilities for

performing the simulation.

1]. I am confident I can meet the challenges of this simulation.

2]. I am confident in my understanding of how information cues are related to
decisions.

3]. I am certain that I can manage this simulation.

Personal Goal
This question asks you to set a PERSONAL goal for the next trial.

1]. What is your PERSONAL goal for the next trial?

Effort
This set of questions asks you about how much effort you will put into the simulation
during the next trial.

1]. I will put a great deal of effort into my performance on this simulation in the next
trial.

2]. I will try very hard to perform well on this simulation in the next trial.

3]. I will work very intensely on this simulation in the next trial.

79

REFERENCES

80

REFERENCES

Austin, J. T., & Vancouver, J. B. (1996). Goal constructs in psychology: Structure,
process, and content. Psychological Bulletin, 120, 338-3 75.

Bandura, A. (1986). Social foundations of thought and action: A social cognitive
theory. Englewood Cliffs, NJ: Prentice Hall.

Bandura, A. (1991). Social cognitive theory of self-regulation. Organizational Behavior
and Human Decision Processes, 50, 248-287.

Bandura, A., & Cervone, D. (1986). Self-evaluative and self-emcacy mechanisms
governing the motivational effects of goal systems. Journal of Personality and
Social Psychology, 45, 1017-1028.

Bandura, A., & Jourden, F. J. (1991). Self-regulatory mechanisms governing the impact
of social comparison on complex decision making. Journal of Personality and
Social Psychology, 60, 941-951.

Becker, L. J. (1978). Joint effect of feedback and goal setting on performance: A field
study of residential energy conservation. Journal of Applied Psychology, 63, 428-
433.

Bryan, J. F., & Locke, E. A. (1967). Goal setting as a means of increasing motivation.
Journal of Applied Psychology, 51, 274-277.

Carver, C. S., & Scheier, M. F. (1981). Attention and self-regulation: A control-theory
approach to human behavior. New York: Springer-Verlag.

Carver, C. S., & Scheier, M. F. (1982). Control theory: A useful conceptual framework
for personality-social, clinical, and health psychology. Psychological Bulletin,
92, 1] 1-135.

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

Carver, C.S. & Scheier, M.F. (1998). On the self-regulation of behavior. New York:
Cambridge University Press.

DeShon, R. P., Kozlowski, S. W. J., Schmidt, A. M., Milner, K. R., & Wiechmann, D.
(2004). A multiple-goal, multilevel model of feedback efl‘ects on the regulation
of individual and team performance. Journal of Applied Psychologx 89, 1035-
1 056.

81

 

 

Donovan, J. J ., & Radosevich, D. J. (1998). The moderating role of goal commitment on
the goal difficulty-performance relationship: A meta-analytic review and critical
reanalysis. Journal of Applied Psychology, 83, 308-315.

Dwyer, D. J., Hall, J. K., Volpe, C., Cannon-Bowers, J. A., & Salas, E. (1992,
September). A performance assessment task for examining tactical decision making
under stress (Spec. Rep. No. 92-002). Orlando, FL: Naval Training Systems Center,
Human Factors Division.

Barley, P. C., & Erez, M. (1991). Time-dependency of goals and norms: The role of
cognitive processing on motivational models. Journal of Applied Psychology, 76,
71 7-724.

Erez, M. (1977). Feedback: A necessary condition for the goal setting-performance
relationship. Journal of Applied Psychology, 62, 624-627.

F estinger, L. (1954). A theory of social comparison. Human Relations, 7, 117-140.

Garland, H. (1982). Goal levels and task performance: A compelling replication of some
compelling results. Journal of Applied Psychology, 67, 245-248.

Garland, H. (1983). Influence of ability, assigned goals, and normative information on
personal goals and performance: A clmllenge to the goal attainability assumption.
Journal of Applied Psychology, 68, 20-30.

Gist, M. E., & Mitchell, T. R. (1992). Self-efficacy: A theoretical amlysis of its
determinants and malleability. Academy of Management Review, I 7, 183-21 1.

Hollenbeck, J. R., & Klein, H. J. (1987). Goal commitment and the goal-setting process:
Problems, prospects, and proposals for future research. Journal of Applied
Psychology, 72, 212-220.

Hollenbeck, J. R., Klein, H. J ., O’Leary, A. M., & Wright, P. M. (1989). Investigation
of the construct validity of a self-report measme of goal commitment. Journal of
Applied Psychology, 74, 951-956.

Karoly, P. (1993). Mechanisms of self-regulation: A systems view. In L. W. Porter &
M. R. Rosenzweig (Eds) Annual Review of Psychology (vol. 44; 29-52). Palo
Alto, CA: Annual Reviews, Inc.

Keppel, G., & Wickens, T. D. (2004). Design and analysis: A researcher’s handbook.
Upper Saddle River, NJ: Prentice Hall.

Latham, G. P., & Locke, E. A. (1991). Self-regulation through goal setting.
Organizational Behavior and Human Decision Processes, 50, 212-247.

82

Lewin, K., Dembo, T., Festinger, L., & Sears, P. S. (1944). Level of aspiration. In J.
McVHunt (Ed.), Personality and the behavior disorders, Vol. 1. New York:
Ronald.

Locke, E. A. (1966). The relationship of intentions to level of performance. Journal of
Applied Psychology, 50, 60-66.

Locke, E. A. (1982). Relation of goal level to performance with a short work period and
multiple goal levels. Journal of Applied Psychology, 67, 512-514.

Locke, E. A. (1991). Problems with goal-setting research in sports-and their solution.
Journal of Sport & Exercise Psychology, 8, 31 1-316.

Locke, E. A. (1994). Comments on Weinberg and Weigand. Journal of Sport & Exercise
Psychology, 1 6, 212-215.

Locke, E. A., & Bryan, J. F. (1968). Goal setting as a determinant of the effect of
knowledge of score on performance. American Journal of Psychology, 81, 398-406.

Locke, E. A., Cartledge, N., & Koeppel, J. (1968). Motivational effects of knowledge of
results: A goal-setting phenomenon? Psychological Bulletin, 70, 474-485.

Locke, E. A., Chah, D. 0., Harrison, S., & Lustgarten, N. (1989). Separating the effects
of goal specificity from goal level. Organizational Behavior and Human Decision
Processes, 43, 270-287.

Locke, E. A., & Latham, G. P. (1985). The application of goal setting to sports. Journal
of Sport Psychology, 7, 205-222.

Locke, E. A., & Latham, G. P. (19903). A theory of goal-setting and take performance.
Englewood Cliffs, NJ: Prentice Hall, Inc.

Locke, E. A., & Latham, G. P. (1990b). Work motivation and satisfaction: Light at
the end of the tunnel. Psychological Science, 1, 240-246.

Locke, E. A., & Latham, G. P. (2002). Building a practically usefirl theory of goal setting
and task motivation: A 35-year odyssey. American Psychologist, 5 7, 705-717.

Locke, E. A., Mento, A. J ., & Katcher, B. L. (1978). The interaction of ability and
motivation in performance: An exploration of the meaning of moderators.
Personnel Psychology, 31, 269-280.

Locke, E. A., Shaw, K. M., Saari, L. M., & Latham, G. P. (1981). Goal setting and task
performance: 1969-1980. Psychological Bulletin, 90, 125-152.

83

Meyer, J. P., & Gellatly, I. R. (1988). Perceived performance norm as a mediator in the
effect of assigned goal on personal goal and task performance. Journal of Applied
Psychology, 73, 410-420.

Miller, G.A., Galanter, E., & Pribram, K.H. (1960). Plans and the structure of behavior.
New York: Holt, Rinehart, & Winston.

Mitchell, T. R., Rothman, M., & Liden, R. C. (1985). Effects of normative information
on task performance. Journal of Applied Psychology, 70, 48-55.

Naylor, J. C., & Ilgen, D. R. (1984). Goal setting: A theoretical analysis of a
motivational technology. Research in Organizational Behavior, 6, 95-140.

Steers, R M., & Porter, L. W. (1974). The role of task-goal attributes in employee
performance. Psychological Bulletin, 81 , 434-452.

Taylor, F. W. (1911). Scientific management. The principles of scientific management.
Harper & Row (pp. 30—48; 57-60).

Wood, R., & Bandura, A. (1989). Social cognitive theory of organizational management.
Academy of Management Review, I 4, 361 -3 84.

Wood, R. 13., Mento, A. J., & Locke, E. A. (1987). Task complexity as a moderator of
goal effects: A meta-analysis. Journal of Applied Psychology, 72, 416-425.

Wright, P. M., & Kacmar, K. M. (1994). Goal specificity as a determinant of goal

commitment and goal change. Organizational Behavior and Human Decision
Processes, 59, 242-260.

84

EEEEEEEEEEEEEEEEEEEEEEEEEEEEE

llHflllllllljflflllllllzlljwllflllllllllll