An Evaluation of the Effects of Very Difficult Goals

Abstract

In 1968 the first cohesive theory on goal setting proposed that difficult goals produce higher levels of performance than easy goals and that specific goals produce a higher level of performance than “do your best” goals. While over 40 years of research supports this theory, there has been some discrepancy regarding the use of very difficult goals. This study was designed to examine the effects on performance of different levels of performance improvement goals and two different types of feedback, as feedback is often used in conjunction with goal setting. A group design was used with participants receiving one of two goal levels, and two different types of feedback. While no significant effects were found between the two types of feedback, goal level produced significant results in terms of performance and accuracy.

Full text

Full Terms & Conditions of access and use can be found at
http://www.tandfonline.com/action/journalInformation?journalCode=worg20
Journal of Organizational Behavior Management
ISSN: 0160-8061 (Print) 1540-8604 (Online) Journal homepage: http://www.tandfonline.com/loi/worg20
An Evaluation of the Effects of Very Difficult Goals
Kathryn M. Roose & W. Larry Williams
To cite this article: Kathryn M. Roose & W. Larry Williams (2018) An Evaluation of the Effects
of Very Difficult Goals, Journal of Organizational Behavior Management, 38:1, 18-48, DOI:
10.1080/01608061.2017.1325820
To link to this article: https://doi.org/10.1080/01608061.2017.1325820
Published online: 19 Jun 2017.
Submit your article to this journal
Article views: 313
View related articles
View Crossmark data

An Evaluation of the Effects of Very Difficult Goals
Kathryn M. Roose and W. Larry Williams
University of Nevada, Reno, Nevada, USA
ABSTRACT
In 1968 the first cohesive theory on goal setting proposed that
difficult goals produce higher levels of performance than easy
goals and that specific goals produce a higher level of perfor-
mance than “do your best”goals. While over 40 years of
research supports this theory, there has been some discre-
pancy regarding the use of very difficult goals. This study was
designed to examine the effects on performance of different
levels of performance improvement goals and two different
types of feedback, as feedback is often used in conjunction
with goal setting. A group design was used with participants
receiving one of two goal levels, and two different types of
feedback. While no significant effects were found between the
two types of feedback, goal level produced significant results
in terms of performance and accuracy.
KEYWORDS
feedback; goals; goal
setting; productivity
Goal setting is one of the most widely studied interventions in psychological
literature. As early as 1897, Bryan and Harter wrote about improving the
performance of telegraph operators simply by assigning a goal. Taylor (1911)
advocated for specific, difficult goals, and Drucker introduced the term
“management by objectives”in his 1954 book, The Practice of Management.
In 1968 Edwin Locke provided the first cohesive theory on goal setting,
stating that difficult goals produce higher levels of performance than easy
goals and specific goals produce higher levels of performance than “do your
best”goals (Locke, 1968, p. 157). Around 40 years later, Latham and Locke,
the two most prolific researchers of goal setting, reported that there had been
over 1,000 studies conducted on goal setting utilizing more than 88 different
tasks, with over 40,000 participants on various continents (Latham & Locke,
2006).
Early in his career, Locke (1968) lamented that the research on task
performance showed a “persistent neglect in experimental psychology of
the study of conscious factors in task performance”(p. 158), blaming this
neglect on the “doctrine of behaviorism”(p. 158), which focuses on
CONTACT Kathryn M. Roose kathrynroose@gmail.com University of Nevada, Reno, 1664 N. Virginia Street,
MS 285, Reno, NV 89557.
Color versions of one or more of the figures in the article can be found online at www.tandfonline.com/WORG.
JOURNAL OF ORGANIZATIONAL BEHAVIOR MANAGEMENT
2018, VOL. 38, NO. 1, 18–48
https://doi.org/10.1080/01608061.2017.1325820
© 2017 Taylor & Francis

observable behavior rather than mental states. According to Locke (1996),
“motivation is something inside the organism”(p. 117), and therefore a field
dominated by behaviorism is unequipped to deal with motivation. Locke
(1968) points to a shift in the field in the late 1960s, with psychologists
becoming “dissatisfied with the limitations placed upon research and theory
by the behaviorist dogma”(p. 158), leading to growing interest in studying
conscious goals, intentions and desires relating to task performance. Later,
Locke (1996) reports “the materialist approaches did not work”(p. 117), and
that dissatisfaction with the state of the field of psychology led to the
“cognitive revolution”(p. 117) in the 1970s and 1980s in terms of goal
setting research. In a 1981 review of the literature, Locke, Shaw, Saari, and
Latham assert, “the concept of goal setting falls within the broad domain of
cognitive psychology and is consistent with recent trends such as cognitive
behavior modification”(p. 125).
Locke and Latham’s contributions to early goal setting research have been
replicated and supported in countless studies throughout the years, including
the connection between goal difficulty and performance. In 2006 Locke and
Latham pointed to decades of goal setting research and asserted, “So long as a
person is committed to the goal, has the requisite ability to attain it, and does
not have conflicting goals, there is a positive, linear relationship between goal
difficulty and task performance”(Locke & Latham, 2006, p. 265). Specifically,
high goals produce higher levels of responding compared to low goals or no
goals.
In contrast, some have argued that goals imply a cap to performance.
Forexample,agoalof“do 20 pushups”specifies stopping when 20
pushups have been completed. Therefore, low goals result in lower pro-
ductivity even when the performer has the capacity for higher productiv-
ity, thus exaggerating the linear relationship between goal difficulty and
performance (Lorenzi, 1988). Lorenzi replicated Locke’s(1982)experi-
ment asking participants to find creative uses of a wire hanger. He
compared “do your best”goals to specific goals with low, medium, and
highdifficulty,aswellasconditionsinwhichparticipantsobservedthe
desired behavior of a model being reinforced, had their own desired
behavior reinforced, or did not observe reinforcement of another or
have their own behavior reinforced. Regardless of goal level or specificity,
performance in the reinforcement conditions was higher than in the
conditions without reinforcement. When controlling for reinforcement,
there was no statistical difference in performance between the “do your
best”conditions and the low, medium, and high goal conditions. Lorenzi
suggests that vague goals with reinforcement may be more effective than
specific goals without reinforcement, and that very difficult goals may be
more effective at increasing productivity when compared to low goals
when additional reinforcement is provided.
JOURNAL OF ORGANIZATIONAL BEHAVIOR MANAGEMENT 19

Goal Commitment
Locke, Shaw, Saari, and Latham (1981)arguethatwhile“Goal setting
might be called ‘stimulus control’by a modern behaviorist”(p. 126),
focusing on the stimulus or goal (an environmental event) ignores the
importance of goal commitment, which they propose is an important
moderatoroftheeffectsofgoals.Essentially,ifagoalisrejected,itwill
not regulate performance.
Goal commitment may be measured indirectly by asking questions such
as “How committed are you to attaining the goal?”(Locke, Latham, &
Erez, 1988, p. 24) or by assessing the discrepancy between assigned goals
and the personal goal the individual claims to be trying to attain (Locke
et al., 1988).Itmayalsobemeasureddirectly, by observation of behavior
following the assignment of the goal. In a comparison of direct and
indirect measures, Earley (1985a,1985b) found a high degree of correla-
tion between the two (.76 and .90). Locke et al. (1988) caution that for
performance to be taken as an indicator of goal commitment, other
variables must be controlled (e.g., goal level, ability). An issue with
measuring the effects of goal commitment is that in most studies goal
commitment is high, with minimal variation (Huber & Neale, 1986;
Locke, Frederick, Lee, & Bobko, 1984). Salancik (1977)proposesthat
when a goal is assigned, people are likely to commit to it because the
assignment suggests that the goal is attainable, and because not objecting
to the goal is essentially a form of consent.
In a meta-analysis of 83 studies, Klein, Wesson, Hollenbeck, and Alge
(1999) concluded that goal commitment is a strong moderator between
goal setting and performance, more so for difficult goals than easier
goals. In one example, Erez and Zidon (1984) found a positive linear
relationship between performance and goal difficulty when the subjects
were committed to the goal, but a negative linear relationship when
goals were rejected. Participants who reported the highest level of goal
commitment performed at very high levels, even in the face of impos-
sible goals. Erez and Zidon (1984) suggest that an “acceptance-rejection
threshold”(p. 77) exists at the point at which the linear relation between
goal difficulty and goal acceptance switches from positive to negative.
Others (e.g., Vance & Colella, 1990) have found high levels of perfor-
mance even after assigned goals were rejected. Mento, Cartledge, and
Locke (1980) suggest that some individuals may self-set goals in place of
rejected assigned goals, as they found in their study. Stedry (1960)
proposed that while individuals are more likely to reject difficult goals
than easier goals, those who accept a difficult goal will work harder to
attain that goal until they decide that it is impossible, at which point
they will lower the goal or abandon it altogether.
20 K. M. ROOSE AND W. L. WILLIAMS

Goal Setting in Behavior Analysis
Fellner and Sulzer-Azaroff (1984) insist that an explanation of cognitive
states is not necessary to explain the effects of goal setting on behavior.
They propose that goals are stimuli that precede behavior. If responses to
goals are likely to be reinforced, the goal has discriminative control over the
response, increasing the likelihood that the response will be repeated in the
future. In addition, as goal attainment is frequently paired with positive
consequences or the removal of negative consequences, goals may function
as conditioned positive or negative reinforcing stimuli. Alternatively, goals
may exert control over behavior through rule-governance. A behavioral
analysis of goal setting emphasizes the identification of relevant environ-
mental variables rather than focusing on mental states or other inaccessible
variables (Fellner & Sulzer-Azaroff, 1984).
Malott’s(1993) discussion of rule-governed behavior is relevant to the
subject of goal setting as goals may be conceptualized as rules (Fellner &
Sulzer-Azaroff, 1984). Malott reviews the position of Mawhinney and Ford
(1977), who stated that for a rule to be a discriminative stimulus (S
D
),
specified responses will be reinforced or punished on some schedule when
the rule is present. However, Malott states that for a rule to function as an S
D
,
the absence of a rule should function as an S-delta, a stimulus in the presence
of which specified responses are less likely to be reinforced or punished. This
is not the case. He uses the example of setting a deadline. If a person has
four hours to complete a task, they might state the rule, “If I do not get to
work right now, I will miss the deadline and look bad”(Malott, 1993, p. 54).
In the absence of a rule statement to describe the contingency, the contin-
gency still exists, and therefore, the rule is not functioning as an S
D
. Rather,
he suggests that rule statements might function as conditioned establishing
operations (EOs) that establish noncompliance with the rule as a learned
aversive condition (Malott, 1993). He proposes that when one follows a rule,
starting behavior in the direction of that rule may be the first stage in
reducing the aversive condition; and finishing the task might lead to the
complete escape of the aversive condition.
According to Agnew (1997), goals do not always signal the availability of
reinforcement that was not present before the installation of the goal.
Instead, goals also function as EOs making achievement of the goal more
valuable due to the conditioned reinforcing consequences of goal achieve-
ment. Goals momentarily increase the reinforcing effectiveness of positive
feedback, evoking goal-directed behaviors, as they have been reinforced by
positive feedback in the past. She concludes, “It is not so much that feedback
is more available in the presence of goal setting (suggestive of the S
D
explanation), rather, feedback is more valuable in the presence of goal setting
(suggestive of the EO explanation)”(Agnew, 1997, p. 13).
JOURNAL OF ORGANIZATIONAL BEHAVIOR MANAGEMENT 21

Goal Setting and Relational Frame Theory
From a Relational Frame Theory (RFT; Hayes, Barnes-Holmes, & Roche,
2001) perspective, goal statements establish feedback as reinforcement for
goal directed behavior (O’Hora & Maglieri, 2006). Goals identify a specific
level of performance and establish a relationship between the level of per-
formance requested by the goal and current level of performance. When the
current level of performance is in a “less than”relationship with the level of
performance specified by the goal, each response that closes the gap reduces
the “less than”relationship, resulting in derived reinforcement for goal-
directed behavior, until the goal is achieved. If goal attainment is not
reinforced, the derived reinforcing effects of goal directed behavior and
feedback will dissipate, decreasing the likelihood of continued goal-directed
behavior (O’Hora & Maglieri, 2006).
Difficult Goals
As early as 1911, Taylor stated that goals should be set at a level of performance
just out of reach of the most capable worker. Locke recommended setting goals
that are “very hard—even outrageous”(Locke, 2001, p. 50). A challenge in
terminology arises when discussing goals that are on the highest end of the
goal setting spectrum. A term often used is “stretch goals”(e.g., Daniels, 2009;
Kerr & Landauer, 2004; Sherman, 1995; Sitkin, See, Miller, Lawless, & Carton,
2011), while others have focused on “unattainable goals”(e.g., Hatzigeorgiadis,
2006; Wrosch, Scheier, Miller, Schulz, & Carver, 2003). Sitkin and colleagues
(2011) provide a distinction between difficult goals and “stretch goals.”Whereas
difficult goals are seemingly attainable, stretch goals are goals “with an objective
probability of attainment that may be unknown but is seemingly impossible
given current capabilities (i.e., current practices, skills, and knowledge)”(p. 547).
In contrast, Latham and Seijts (1999)andLocke(1991) consider a difficult goal
one that is likely to be attained by around 10% of individuals that attempt them,
and Daniels (2009) defines stretch goals as “those that are attainable less than
10% of the time”(p. 36). Of course, using this criterion to set goals in experi-
mental studies is challenging, as it cannot be known if 10% of participants
achieved the goal until the study is complete.
Similarly, the use of the terms “attainable”and “unattainable”in relation
to goals is problematic, as attainable and unattainable may often only been
determined following performance, and even then one cannot always classify
the goal as such for all participants. For example, factors such as fatigue or
illness may prevent the attainment of a goal at a certain time, however, the
goal would be attainable in other situations. In addition, an attainable goal
for one person may be unattainable by another.
22 K. M. ROOSE AND W. L. WILLIAMS

Research on Difficult Goals
Most research has supported the linear relationship between goal difficulty
and performance even with very difficult goals (e.g., Latham & Yukl, 1975;
Locke, 1968; Locke, Mento, & Katcher, 1978). A meta-analysis indicates that
the effect size of goal difficulty on performance is between 0.52 and 0.82
(Locke & Latham, 1990), which indicates medium to high relations between
the two. Locke (1966) concluded that the higher the goal, the higher the level
of performance, even when the goal was so high that participants reached it
on less than 10% of opportunities. Later, Locke (1982) examined 14 levels of
goals across 1-minute sessions. Participants achieved the first goal, but no
goal thereafter. Performance did not decrease in response to increasing goals
even when the previous lower goal was not achieved. Locke admits that these
results would not likely sustain following repeated trials, or trials of longer
duration. Garland (1982) gave students adjectives and asked them to list
items that could be described by that adjective. There were three levels of
increasing difficulty, and results showed sustained or increasing perfor-
mance, even following repeated failure to reach the previously assigned
goals. Later Garland asserts, “increased motivation produced by a difficult
performance goal does not seem to dissipate under conditions in which the
goal is beyond the probable reach of most individuals to whom it is assigned”
(Garland, 1983, p. 29).
In a 1981 review of the literature, Locke and colleagues found overwhelm-
ing support for higher performance in response to higher goals, however,
they found a handful of studies that did not follow this trend. They proposed
that some goals were too difficult, so the participants gave up. This is
supported by behavior analytic research, as demands outside of the repertoire
of an individual will not evoke responses that are likely to be reinforced.
Further, when shaping a response, if the level of performance required for
reinforcement increases too drastically, the individual’s behavior is likely to
experience extinction (Catania, 2007).
Vroom’s(1964) valence-instrumentality-expectancy (VIE) theory predicts
that performance is affected by valence (anticipated satisfaction), instrumen-
tality (belief that performance will be rewarded), and expectancy (belief that
effort will lead to the performance that will be rewarded). As expectancy is
said to be linearly related to performance (all other variables being equal), an
apparent contradiction emerges when examining increasingly difficult tasks.
As difficulty increases, expectancy should decrease, leading to a decrease in
performance, resulting in an inverted-U-shaped relationship between goal
difficulty and performance such that performance increases as goals become
more challenging, but at a certain level of difficulty, performance will
decrease. This inverted-U-shaped relationship is similar to the relationship
proposed by Erez and Zidon regarding the “acceptance-rejection threshold”
JOURNAL OF ORGANIZATIONAL BEHAVIOR MANAGEMENT 23

(Erez & Zidon, 1984) in that valence, instrumentality, and expectancy may
have an effect on goal commitment.
Atkinson (1958) found support for the inverted-U relationship when he
gave participants a task, and told them that 1/20, 1/3, 1/2, or 3/4s of the top
performers would win a prize. The 1/3 and 1/2 groups performed the best,
and the 1/20 group performed the worst, reporting that they thought the
chance of reward was low. Put simply, effort was highest for moderately
difficult tasks, with less effort expended for very easy or very difficult tasks.
Similar results were seen by Motowidlo, Loehr, and Dunnette (1978), and
Erez and Zidon (1984), however, this model has found little additional
support. Garland (1983) suggests that an inverted-U relationship is more
likely to be found when an incentive is offered only for goal attainment (e.g.,
Atkinson, 1958; Mowen, Middlemist, & Luther, 1981). Garland (1984) dis-
cusses Locke’s view, obtained in personal communication, that “he (Locke)
believes exclusive rewards for goal attainment may undermine the positive
influence of difficult goals by suggesting that partial success is not worth-
while”(p. 80).
In an unpublished study, See, Heath, and Fox (2003)examinedthe
difference in performance when goals were just out of reach or far beyond
reach. They gave elite runners 10 seconds to run 100 meters or 200 meters.
While elite runners would be aware that these goals surpass the world
record (“100 Metres,”2016) the researchers found that those with the
easier goal expended more effort toward the goal, supporting the inverted-
U relationship. However, these participants were all elite runners, and the
researchers predicted that very difficult goals would motivate those with
higher abilities. Therefore, in their second study, they recruited partici-
pants of varying abilities to perform a strength task called a wall-sit with a
relatively low goal, then a relatively high goal. Their finding was that the
participants who were more successful without a goal performed better in
the higher goal condition, and the participants who were less successful
without a goal performed worse in the higher goal condition, suggesting
that there may not be a universal relationship between goal difficulty and
performance; rather, personal and environmental factors influence the
relationship. Other researchers have found that while difficult goals may
increase performance of the highest performers, a much larger group will
not be able to achieve the goal, leading to poor performance (e.g., Soman
& Cheema, 2004).
O’Hora and Maglieri (2006) caution against the use of goals that are
unattainable. They propose that if the subject is not getting closer to the
goal, and thus decreasing the “less than”relationship, the goal will not
effectively manage performance. Excessively difficult goals will prolong the
“less than”relationship and the temporal distance to the reinforcer. O’Hora
and Maglieri cite studies that show unrealistic goals may actually decrease
24 K. M. ROOSE AND W. L. WILLIAMS

performance when compared to easier goals (e.g., Audia, Locke, & Smith,
2000; Earley & Perry, 1987).
Kerr and LePelley (2013) report that business literature has emphasized
the utility of stretch goals and underreported cases in which stretch goals
were counterproductive or ineffective. A well-known example of a reported
success of a stretch goal is the invention of the bullet train. At one time it
took more than 6 hours by train from Tokyo to Osaka, Japan. Rather than
being tasked with a goal of reducing the time to 6 hours, which would have
resulted in small, incremental changes, Japanese executives gave engineers a
goal of 3.5 hours. This seemingly impossible goal resulted in engineers
reexamining every aspect of train travel, resulting in the bullet train. Jack
Welch, at that time the head of General Electric (GE), used this example to
support the use of stretch goals at GE. Sitkin and colleagues (2011) give a
similar account of stretch goals, stating that they “serve as jolting events that
disrupt complacency and promote new ways of thinking and acting”(p. 545).
While situations such as the bullet train are examples in which ingenuity
could have extreme effects, they have been used in work environments to
support the use of very difficult goals in situations in which a physical
performance ceiling exists, and no amount of ingenuity or disruption in
complacency would be sufficient to achieve the goal.
Quality of Work
Many studies on goal setting fail to account for the effect of difficult goals on the
quality or accuracy of performance. One might predict that as demand for speed
or output increases, quality of work might suffer. In an unpublished master’s
thesis, Isley (2007) found that increasing a goal on completing simple math
problems led to an increase in incorrect responses and responses to skip ques-
tions, with an insignificanteffect on correct responses. Similarly, Bavelas and Lee
(1978) found that higher goals resulted in responses “farther from ideal”(p. 219).
In real-world organizational settings, Daniels (2009) warns, “When sys-
tematic positive reinforcement is lacking from goal-based systems, employee
efforts are driven by negative reinforcement”(p. 37). In other words, once
employees give up on attaining reinforcement for achieving goals, they are
likely to perform at a level just high enough to avoid aversive consequences,
which may be lower than their baseline performance.
Schweitzer, Ordóñez, and Douma (2004) caution that using unattainable
goals may lead to unethical behavior. Ordóñez, Schweitzer, Galinsky, and
Bazerman (2009) write about employees of Sears, Roebuck and Co.’sautorepair
staff overcharging customers and completing unnecessary repairs in order to
meet unrealistic sales goals set by management in the 1990s. In the 1960s Ford
Motor Company promised to produce a car under 2,000 pounds to be sold for
under $2,000. The result was the Ford Pinto, which made it to car dealerships
JOURNAL OF ORGANIZATIONAL BEHAVIOR MANAGEMENT 25

without being safety tested. In a more recent example, Wells Fargo set a goal of
eight accounts per customer, resulting in employees setting up accounts for
customers without their knowledge. Welsh and Ordóñez (2014) suggest that
when managers focus more on goal attainment than on how goals are attained,
the resulting environment may facilitate unethical behavior to achieve those
goals.
Feedback
Feedback and goal setting are often used together in research and in practice
(e.g., Garland, 1982; Wilk & Redmon, 1997). In a 1981 review, Locke points
out that there is evidence that providing feedback along with setting goals
leads to better performance than either goals or feedback alone. Prue and
Fairbank (1981) define performance feedback as information given to indi-
viduals regarding the quality of their performance. Rummler and Brache
(1995) define feedback as information that tells performers how well they
are doing; while Daniels (1994) defines it as information given to performers
that allows them to adjust their performance.
As for the function of feedback, there have also been a variety of
suggestions. Feedback is simply a consequence of behavior, with positive
feedback strengthening the behavior and negative feedback weakening the
behavior (Balcazar, Hopkins, & Suarez, 1985); it serves an instructional
purpose, functioning as an S
D
(Locke et al., 1981);itfunctionsasanEO,
increasing the value of consequences for desired performance (Agnew,
1997; Duncan & Bruwelheide, 1986); or it evokes rule-governed behavior
(Agnew, 1997).
Fellner and Sulzer-Azaroff (1984) propose that the combination of goal
setting and feedback is effective because with the goal functioning as an
antecedent and feedback functioning as a consequence, the behavior may
come under stimulus control. In other words, the behavior is differentially
reinforced by the feedback in the presence of a stimulus (the goal). Goal setting
on its own may be ineffective because the corresponding response is not
reinforced, and goal attainment may not function as a conditioned reinforcer
for certain individuals due to their history of reinforcement. Further, due to its
presentation with other reinforcers, feedback might function as a conditioned
reinforcer (Kang, Oah, & Dickinson, 2005). From the RFT perspective, goal
statements establish feedback as reinforcement for goal directed behavior, and
for goal directed behavior to persist, the individual must receive feedback
indicating that they are closing the gap between their performance and the
goal performance (O’Hora & Maglieri, 2006). One would expect that the effects
of feedback on individuals would vary depending on each individual’s personal
learning history, and as Houmanfar and Hayes (1997) point out, research does
support this assertion (e.g., Balcazar et al., 1985; Duncan & Bruwelheide, 1986).
26 K. M. ROOSE AND W. L. WILLIAMS

Persistence
If an individual is engaging in goal-directed behavior in the absence of, or
with a low probability of reinforcement, one might say they are behaving
with persistence with respect to the goal, or exhibiting resistance to extinc-
tion. Early in the goal setting literature, Locke and Bryan (1969) found that
difficult goals on addition problems not only lead to initial increases in
performance, but this effect was sustained over prolonged work periods. In
another early study, when comparing “do your best”goals and difficult goals,
Locke and Bryan (1967) found that those with difficult goals initially
reported more interest in the task, and while interest declined for both
groups throughout the task, it declined much faster for the “do your best”
group. As reported boredom was negatively correlated with performance,
Locke and Bryan suggested that higher goals make work tasks more inter-
esting, and therefore people are more likely to work harder and longer than
those who are simply told to “do your best.”Further studies have shown that
subjects with difficult goals spent more time working and less time resting
during an anagram task than those with easy goals (Sales, 1970), subjects with
goals chose to work more trials on a maze task than those without goals
(Singer, Korienek, Jarvis, McColskey, & Candeletti, 1981), and subjects with
difficult goals compressed a hand dynamometer longer than those told to “do
your best”(Hall, Weinberg, & Jackson, 1987).
Recent Research
In a study by Tammemagi, O’Hora, and Maglieri (2013), participants were
exposed to an analog data entry task, first with no goal, then a low goal
followed by a high goal, or vice versa. They predicted that in the absence of
reinforcement for meeting the high goal, performance would decline. Their
results showed greater increases in performance during the high goal condi-
tion than during the low goal condition, and only 38% of participants
exhibited a negative trend in performance within the high goal condition
even in the absence of goal attainment. They concluded that their results did
not support the assertions of O’Hora and Maglieri (2006) who would have
predicted that in the continued absence of reinforcement for goal attainment,
that goal directed behavior would decrease.
Tammemagi’s doctoral dissertation (2012) included four additional
studies that were not included in this publication (Tammemagi et al.,
2013). In one of these studies (Study 4), participants were exposed to the
same experimental task for a no-goal baseline, then four 12-minute
experimental sessions with goals set at 160% of their own baseline per-
formance. Only 20% of participants showed a negative trend in perfor-
mance, even while never reaching the assigned goal. The low performing
JOURNAL OF ORGANIZATIONAL BEHAVIOR MANAGEMENT 27

group had the highest percentage of participants to show a decreasing
trend in performance (35%). Tammemagi suggests that low performance
during baseline may have been an indicator that those participants would
persist less when assigned an unattainable goal. Potential explanations
given are that the low performing participants had more room to
improve, or that the goal set at 160% of their baseline performance was
not actually a difficult goal, and therefore, performance did not increase
as predicted by the linear relationship between goal difficulty and
performance.
Summary and Specific Aims
The conclusions reached by Locke in 1968 have, for the most part, gone on to
be supported in research throughout several decades. Hundreds of studies
have found that more difficult goals produce higher levels of responding than
less difficult goals, and that specific goals produce higher levels of responding
than “do your best”goals.
The aim of this study was to assign difficult goals and two types of
feedback to participants and measure persistence in performance, along
with correct and incorrect responses, and increases in performance over
the no-goal baseline condition. Goals were set at either 150% or 175% of
each participant’s baseline performance, and 13-minute sessions were used,
similar to the procedures used in Tammemagi et al. (2013). The goals used in
this study were intended to be challenging and potentially just out of reach of
most participants, as seen in the study by Tammemagi and colleagues.
Feedback was displayed throughout experimental sessions in one of two
forms. One displayed the percent of goal completion (Feedback 1); the other
compared the participant’s progress to the progress that would be required at
that time in the session to meet the goal by the end of the session (Feedback 2).
Feedback 2 was designed based on the conceptualization of the “less than”
relationship between goals and current performance as described by O’Hora
and Maglieri (2006). Due to the high goals, Feedback 2 was likely to indicate to
participants that they were not decreasing the “less than”relationship between
their performance and the goal performance. A choice condition was used to
assess preference for Feedback 1 or 2.
Persistence was measured by use of a “Start Over”button. This button allowed
participants to leave a session, and to start over with a new session in an attempt
to meet their goal, thus extending the duration of their work sessions. Goal
commitment was assessed to analyze its relationship with performance. A five-
item scale was used that the authors propose is “unidimensional and equivalent
across measurement timing, goal origin, and task complexity”(Klein, Wesson,
Hollenbeck, Wright, & DeShon, 2001,p.33).
28 K. M. ROOSE AND W. L. WILLIAMS

Method
Participants
Twenty-four participants took part in the study. Participants were under-
graduate students at a four-year university, recruited through an online
recruiting system, public posting of fliers in academic buildings and tutoring
centers, and announcements in undergraduate psychology classes.
Participants earned course credit for participating in the study. All partici-
pants gave consent to participate, and all procedures were approved by the
Institutional Review Board.
Apparatus and Setting
The experimental task was designed with Visual Basic 6.0 to simulate a data
entry task that might be used in a medical setting. Responses were made using a
typical keyboard and mouse. The program presented all stimuli and recorded
all responses. A Samsung personal laptop computer was used in a laboratory
room in the Department of Psychology in a university setting. The total
duration of the study was a maximum of 150 minutes per participant. Data
were collected via the Visual Basic program, output into an Excel spreadsheet.
Experimental Task
The data entry task that was used was developed to simulate typical electrocardio-
gram (ECG) data. The simulation was created by Maglieri (2007), and has since
been modified for additional studies (e.g., Smith, 2013; Tammemagi et al., 2013),
and was modified again for this study. The screen contains fictional medical
information that would be recorded following an ECG reading, populated using
a randomizing formula. Participants were required to use the patient’sgenderand
QT interval reading to determine if the QT interval was “Below Range,”“Within
Range,”or “Above Range,”and click the corresponding radio button. Next,
participants used the patient’s age and heart rate to determine if the heart rate
was “Below Avg.,”“Average,”or “Above Avg,”and click the corresponding radio
button. Finally, the participant clicked the “submit”button. Correct responses
were recorded at the bottom of the screen throughout the task. Each work session
lasted for 13 minutes. At the end of the session, a summary screen was presented
to the participant, showing them the number of correct and incorrect responses
for each session. See Figure 1 for a screen shot of the experimental task.
Dependent Variables
Data were recorded on correct responses and incorrect responses, and raw
scores were used to calculate a percent increase over baseline responding, and
JOURNAL OF ORGANIZATIONAL BEHAVIOR MANAGEMENT 29

Figure 1. Screen shot of experimental task and Feedback 1.
Figure 2. Screen shot of Feedback 2.
30 K. M. ROOSE AND W. L. WILLIAMS

accuracy in responding across sessions. Persistence was measured by the use
of the “Start Over”button, which added time to the experimental session, as
this indicates continued work toward the goal following unsuccessful
attempts and lack of reinforcement. Goal commitment was also assessed.
Independent Variables
The independent variable of goal difficulty was presented in two levels: (a) a
goal set at 150% of baseline performance, and (b) a goal set at 175% of baseline
performance. The independent variable of feedback was presented in two
formats: (a) feedback indicating the participant’s percent of goal completion
(Feedback 1), and (b) feedback indicating the participant’s percent of goal
completion, plus feedback indicating whether the current rate of performance
will be sufficient to reach the goal by the end of the time limit (Feedback 2).
Research Design
This experiment utilized a group design with two goal levels and two types of
feedback. The participants were randomly assigned to a 150% goal, a 175%
goal, or a no-goal (control) condition. All experimental participants were
exposed to both types of feedback and a choice condition. The order of the
feedback was counterbalanced across participants (ABCX[150% Goal], ABCX
[175% Goal], ACBX[150% Goal], ACBX[175% Goal]). Eight control partici-
pants completed four no-goal conditions (AAAA).
Procedure
Training
The experimenter used screen shots of two different patient records to walk
the participant through the task. The experimenter completed the first record
while describing the process out loud. The participant was asked to complete
the second record while describing the process out loud. Next, the participant
started the program and completed one record while the experimenter
watched, or two records if the first was completed incorrectly. The partici-
pant was given 5 minutes to practice while the experimenter left the room.
Following the completion of training, the researcher asked the participant if
they had any questions about the work task.
Baseline (A)
The researcher told the participant, “you will now complete the same task for
13 minutes. Do your best.”The task appeared, and the participants
JOURNAL OF ORGANIZATIONAL BEHAVIOR MANAGEMENT 31

completed the data entry task as described above. Participants saw a running
total of the number of correct responses per session at the bottom of the
screen. At the end of the session, the computer program generated a total
score. No additional feedback was given.
Control (A)
Participants in this condition were told, “you will now complete the same
task for 13 minutes. Do your best.”The condition was otherwise identical to
the baseline condition.
Percent of Goal Feedback/Feedback 1 (B)
Eight participants received a goal of 150% of their baseline performance and
eight received a goal of 175% of their baseline performance. The goal was
communicated to the participants in this way: “You completed X responses
in the 13 minute baseline session. For this session, your goal will be Z% of that
score, which is Y responses in 13 minutes,”with “X”indicating each partici-
pant’s baseline performance, “Z”indicating 150% or 175%, and “Y”indicating
the goal for the experimental condition. Next, the experimenter stated, “There
are a couple differences this time around. First, you will see a new feedback box
on the bottom of your screen. This will show you your progress towards the
goal as a percent. For example, if your goal is 100 correct responses, and you
have completed 10, you will see ‘Your Progress: 10%.’In addition, you will see a
‘Start Over’button at the bottom of the screen. You may use the ‘Start Over’
button as many times as you want if you are not satisfied with your progress
towards the goal, but you don’thaveto.Ifyouusethe‘Start Over’button, your
time and responses will reset to zero and you will have a new 13-minute session
to try to reach your goal. Do you have any questions?”
Participants answered yes or no questions about goal commitment after
the assignment of their goal, and after each use of the Start Over button (if
applicable). The questions used to measure goal commitment were: (a) It’s
hard to take this goal seriously, (b) Quite frankly, I don’t care if I achieve
this goal or not, (c) I am strongly committed to pursuing this goal, (d) It
wouldn’t take much to make me abandon this goal, and (e) I think this is a
good goal to shoot for (Klein, Wesson, Hollenbeck, Wright, & DeShon,
2001).
Feedback 1 was displayed on the work screen throughout the session (see
Figure 1 for a screen shot of Feedback 1). At the end of the session, the
computer program generated a total score and displayed it on the screen. No
additional feedback was given.
32 K. M. ROOSE AND W. L. WILLIAMS

Current Performance Versus Goal Performance/Feedback 2 (C)
The goal for this condition was calculated and explained to each participant in the
same way as described in the (B) condition. Next, the experimenter stated, “There
are a couple differences this time around. First, you will see two new feedback
boxes on the bottom right of your screen. The first box will show you your
progress towards the goal as a percent. For example, if your goal is 100 correct
responses, and you have completed 10, you will see ‘Your Progress: 10%.’The
feedback box below it will show you what percent of the goal you SHOULD have
completed in order to meet your goal by the end of the session. For example, if you
are halfway through the session with a goal of 100, and you have completed 40
correct responses, you will see ‘Your Progress: 40%’and ‘Good Progress: 50%.’If
your progress is sufficient to meet the goal, it will be shown in green, if your
progress is not sufficient to meet the goal, your progress will be shown in red. In
addition, you will see a ‘Start Over’button at the bottom of the screen. You may
use the ‘Start Over’button as many times as you want if you are not satisfied with
your progress towards the goal, but you don’t have to. If you use the ‘Start Over’
button, your time and responses will reset to zero and you have a new 13-minute
session to try to reach your goal. Do you have any questions?”
Goal commitment was measured as described above. Feedback 2 was
displayed on the work screen throughout the session (see Figure 2 for a
screen shot of Feedback 2). At the end of the session, the computer program
generated a total score. No additional feedback was given.
Choice (X)
The goal for this condition was calculated and described to the participants
in the same way as described in conditions (B) and (C). Participants were
reminded about the two different types of feedback, and were asked to
0
20
40
60
80
100
120
140
BL 1 2 3
Raw Score
Session
Control Participants - Raw Scores
Figure 3. Raw scores for control participants (n= 8).
JOURNAL OF ORGANIZATIONAL BEHAVIOR MANAGEMENT 33

choose which feedback to receive in their final session in this way, “You have
now had the opportunity to experience two types of feedback. One tells you
what percent of your goal you have completed, and the other gives you your
progress compared to the progress needed to meet your goal by the end of
the session. Which feedback would you prefer for the final session?”Goal
commitment was measured as described above. The feedback chosen by the
participant was displayed on the screen as previously described.
Participants were asked if they need a break between each session. The
experimenter remained inside the room, on the other side of a cubicle wall.
At the conclusion of the study, the experimenter debriefed the participant on
the goals of the study, and thanked the participant for their participation.
Due to the potential use of the Start Over button, a maximum time of
2.5 hours was set for the experiment. All participants received 2.5 hours of
credit for participation whether they used the Start Over button or not.
Results
Raw score results for control participants, 150% goal participants, and 175%
goal participants may be found in Figures 3,4, and 5, respectively. A
technical problem affected one control participant. During the second ses-
sion, the session ended unexpectedly at 5.2 minutes. The software was reset
and the second and third experimental (control) sessions were completed.
This resulted in an extra 5.2 minutes of practice before the final two sessions.
150% Goal Participants
Four participants chose Feedback 1 in the choice condition, and four chose
Feedback 2. Six participants chose the feedback they had just experienced,
while two chose to revert to the first feedback they experienced. The average
0
20
40
60
80
100
120
140
BL123
Raw Score
Session
150% Participants - Raw Scores
Figure 4. Raw scores for 150% goal participants (n= 8).
34 K. M. ROOSE AND W. L. WILLIAMS

increase over baseline was 159% for sessions with Feedback 1, and 163% for
sessions with Feedback 2.
The Start Over button was used by five participants a total of 24 times,
adding a total of 50.7 minutes of work. In the Feedback 1 condition, the Start
Over button was used seven times by four participants, adding 28.0 minutes
of work. In the Feedback 2 condition, the Start Over button was used 17
times by five participants, adding 22.7 minutes of work.
175% Goal Participants
Four participants chose Feedback 1 during the Choice condition, three chose
Feedback 2, and one quit the study before the Choice condition. Four
0
20
40
60
80
100
120
140
BL123
Raw Score
Session
175% Participants - Raw Scores
Figure 5. Raw scores for 175% goal participants (n= 8).
0
20
40
60
80
100
120
BL123
Raw Score
Session
Average Raw Scores by Group
Control 150% 175%
Figure 6. Average raw scores by group.
JOURNAL OF ORGANIZATIONAL BEHAVIOR MANAGEMENT 35

participants chose the feedback they had just experienced, while three chose
to revert to the first feedback they experienced. The average increase over
baseline was 135% for sessions with Feedback 1, and 143% for sessions with
Feedback 2.
The Start Over button was used a total of 17 times by four participants,
adding 79.8 minutes of work time. In the Feedback 1 condition, the Start
Over button was used by four participants a total of nine times, adding
47.6 minutes of work time. In the Feedback 2 condition, the Start Over
button was used by four participants a total of eight times, adding 32.2 min-
utes of work time.
Participant 357 used the Start Over button three times in the Feedback 1
condition, adding 4.4 minutes to that condition. Participant 357 then worked
for 7.5 minutes in the Feedback 2 condition, used the Start Over button,
worked for 6.5 minutes in another session of the Feedback 2 condition, then
used the Start Over button again. At that time, the participant requested to
quit the experiment. He made comments regarding the goal being too high,
and that he was trying to be accurate, but was getting frustrated by starting
over after errors.
Group Results
The control group’s increases averaged 123%, 131%, and 137% over baseline
across sessions; the 150% goal group’s increases averaged 159%, 169%, and
180% over baseline across sessions; and the 175% goal group’sincreases
averaged 125%, 136%, and 144% over baseline across sessions. An ANOVA
was completed on the independent variable of Goal Level. The main effect for
Goal Level was significant F(2,20) = 4.16, p= .031. In addition, a one-way
ANOVA examining the main between-subject effect of Goal Level was com-
pleted. The mean value of the increases over baseline for the 150% goal
condition (M= 169.32%, SD = 42.00%) was significantly higher than the control
participants (M= 130.35%, SD = 17.26%), and the 175% goal participants
(M= 135.13%, SD = 23.20%). See Tables 1 and 2for ANOVA results and for
means, standard deviations, and sample sizes. See Figure 6 for raw scores by
group and Figure 7 by percent increases by group.
Seven of eight participants in the 150% goal group met their goal by the
end of the experiment. Five met the goal in the first experimental session, one
more in the second session, and one more in the third session. One of the
175% goal participants met the goal in the third experimental session. Of the
150% goal participants, six increased across all sessions. Of the seven 175%
goal participants who completed the experiment, three increased across all
sessions.
Feedback 2 produced slightly higher increases over baseline, with the 150%
group increasing 159% with Feedback 1 and 163% with Feedback 2, and the
36 K. M. ROOSE AND W. L. WILLIAMS

175% group increasing 135% with Feedback 1 and 143% with Feedback 2.
Across all participants, Feedback 1 produced increases of 147% and Feedback
2 produced increases of 53%. A paired samples t-test for feedback type was
not significant, t(14) = 0.45, p= .662. See Table 3 for the t-test results.
The control group’s accuracy improved from 93.7% during baseline to an
average of 98.5% during the experimental conditions. The 150% goal group’s
accuracy increased from an average of 93.8% during baseline to an average of
96.1% during the experimental conditions. The 175% goal group’s accuracy
decreased from the baseline average of 90.8% to an average of 88.1% during
the experimental conditions. An ANOVA was completed to determine if goal
Table 1. One-Within One-Between ANOVA for Goal Level.
Source df SS MS F p η
2p
Between-subjects
Goal level 2 21380.87 10690.43 4.16 .031 0.29
Residuals 20 51404.09 2570.20
Within-subjects
Within factor 2 3812.18 1906.09 26.35 < .001 0.57
Goal level within factor 4 96.17 24.04 0.33 .855 0.03
Residuals 40 2893.86 72.35
Note. ANOVA = analysis of variance.
Table 2. Means and Standard Deviations for Goal Level.
Goal level Session 1 Session 2 Session 3 Row Average
Control 122.90 (14.59) 130.85 (17.22) 137.29 (18.72) 130.35 (17.26)
150 158.99 (42.24) 169.28 (43.11) 179.68 (43.70) 169.32 (42.00)
175 124.81 (25.45) 136.19 (17.82) 144.39 (24.58) 135.13 (23.20)
Column average 136.04 (33.27) 145.84 (32.91) 154.20 (35.42) 145.36 (34.21)
Note. Standard deviations are in parentheses.
Figure 7. Average increase over baseline by group.
JOURNAL OF ORGANIZATIONAL BEHAVIOR MANAGEMENT 37

level had a significant effect on accuracy. The main effect of group level was
significant (2,20) = 4.71, p= .021. The mean accuracy of the 150% group
(M= .96, SD = .04) and the control group (M= .97, SD = .05) were both
significantly greater than the 175% goal group (M= .89, SD = .10). See Tables
4and 5for ANOVA results, and for means, standard deviations, and sample
sizes. See Figure 8 for graphed results.
Five participants in the 150% group used the Start Over button 24 times,
adding a total of 50.7 minutes of work time. Four participants in the 175%
goal group used the Start Over button 17 times, adding a total of 79.8 minutes
of work time. When broken down by feedback type, eight participants used
the Start Over button a total of 16 times during the Feedback 1 condition,
adding a total of 75.6 minutes of work time, and eight participants used the
Start Over button a total of 25 times during the Feedback 2 condition, adding
a total of 54.9 minutes of work time.
Goal Commitment
Goal commitment was measured as a score out of five possible points at
the start of each experimental session, and after each use of the Start
Over button, therefore each participant had a different number of scores
for goal commitment. As predicted by Locke et al. (1984), goal
Table 3. Paired Samples t-Test for Feedback.
FB2TOT FB1TOT
M SD M SD t p d
152.71 34.23 147.81 41.17 0.45 .662 0.13
Note. Degrees of Freedom for the t-statistic = 14. drepresents Cohen’sd. M = mean; SD = standard
deviation.
Table 4. One-Within One-Between ANOVA for Accuracy by Goal Level.
Source df SS MS F p η
2p
Between-subjects
Group 2 0.13 0.06 4.71 .021 0.32
Residuals 20 0.27 0.01
Within-subjects
Within factor 3 0.01 0.00 0.86 .467 0.04
Group within factor 6 0.02 0.00 1.34 .252 0.12
Residuals 60 0.12 0.00
Table 5. Means and Standard Deviations for accuracy ANOVA.
Group Baseline Session 1 Session 2 Session 3 Row Average
150 0.94 (0.08) 0.96 (0.02) 0.96 (0.02) 0.96 (0.03) 0.96 (0.04)
175 0.90 (0.10) 0.87 (0.13) 0.88 (0.09) 0.89 (0.09) 0.89 (0.10)
0 0.94 (0.09) 0.99 (0.01) 0.99 (0.01) 0.98 (0.01) 0.97 (0.05)
Column average 0.93 (0.09) 0.94 (0.09) 0.95 (0.07) 0.95 (0.07) 0.94 (0.08)
38 K. M. ROOSE AND W. L. WILLIAMS

commitment measures in this study were high and consistent. The
average commitment scores of all participants on all sessions was 3.67
out of 5, with 12 out of 16 participants averaging a score of 4 or higher.
Theaveragecommitmentscoresforthe150%andthe175%groupwere
identical (4.04). Pearson correlations between each participants’goal
commitment and final increase over baseline, number of times using
the Start Over button, and amount of time added by using the Start Over
button were −0.273 (p= 0.306), 0.158 (p= 0.559), and 0.117 (p=0.666),
respectively.
Low Versus High Performers
When the experimental participants are divided into the top eight performers
and bottom eight performers (based on baseline performance), the resulting
groups include even numbers across goal assignments. The bottom perfor-
mers increased 185%, 195%, and 206% across sessions with the 150% goal,
and 111%, 129%, and 139% across sessions with the 175% goal. The top
performers increased 133%, 144%, and 153% across sessions with the 150%
goal, and 133%, 142%, and 148% across sessions with the 175% goal. The
150% goal condition produced higher increases in responding than the 175%
goal for all participants in all sessions except for one that was identical
(Session 1 for the high performers). See Figure 9 for graphed results.
Four of the bottom eight performers used the Start Over button a total of
14 times, adding a total of 31.2 minutes of work time. Five of the top eight
0.00%
10.00%
20.00%
30.00%
40.00%
50.00%
60.00%
70.00%
80.00%
90.00%
100.00%
BL 1 2 3
Accuracy
Session
Average Accuracy by Group
Control 150% 175%
Figure 8. Average accuracy by group.
JOURNAL OF ORGANIZATIONAL BEHAVIOR MANAGEMENT 39

performers used the Start Over button a total of 27 times, adding a total of
99.3 minutes of work time.
Discussion
This study was designed to further examine the effects of very difficult goals
on performance and persistence. The results of the current study support
previous research (e.g., Atkinson, 1958; Erez & Zidon, 1984; See et al., 2003)
that has found goal difficulty and performance to be linearly related until the
goal becomes so difficult that performance stops improving or even declines.
Greater increases in performance were found in the lower goal condition
when compared to the higher goal, and only 37% of participants with the
higher goal showed an increasing trend across sessions. While only one of the
175% goal participants met their goal in one experimental session, three
150% goal participants increased to 175% of their baseline performance or
higher in seven experimental conditions, well above their assigned goal.
Further, the 150% goal participants were more likely to increase or remain
steady across sessions than the 175% goal participants.
According to O’Hora and Maglieri (2006), feedback must indicate that the
“less than”relation between the goal and current performance is decreasing
for goal-directed behavior to persist. This assertion was examined through
the use of Feedback 2, which was designed to provide feedback to partici-
pants that they were not going to attain their goal. While Feedback 2
produced slightly higher increases in responding than Feedback 1, the dif-
ference was not statistically significant. Feedback 2 may have prevented some
50.0%
70.0%
90.0%
110.0%
130.0%
150.0%
170.0%
190.0%
210.0%
230.0%
123
Percent Increase over Baseline
Session
Average Increase over Baseline by High and
Low Performers
Low Performers Low Goal Low Performers High Goal
Hi
g
h Performers Low Goal Hi
g
h Performers Hi
g
h Goal
Figure 9. Average increase over baseline by high and low performers and goal level.
40 K. M. ROOSE AND W. L. WILLIAMS

participants from experiencing this closing of the “less than”gap, as the
feedback alerted them that their performance would not be sufficient to meet
the goal.
Use of the Start Over button was meant to assess whether and for how
long participants would continue to persist toward their goals. There was
no added benefit for participants to use the Start Over button, and no
instruction on differential consequences for reaching or not reaching the
goal that might incentivize participants to use the Start Over button.
While one would expect that undergraduate students would have no
reason to extend their time participating in an experiment, 9 of 16
experimental participants extended their sessions by using the Start Over
button adding a grand total of 1.84 hours to the 16 experimental sessions.
The Start Over button was used more frequently with the 150% goal,
however, more time was added through the use of the Start Over button
during the 175% goal condition. These results may be explained by the fact
that the 150% goal participants hit their goal more frequently than the 175%
goal participants, meaning that use of the Start Over button was not
necessary.
The Start Over button was used more frequently during the Feedback 2
condition, but added more time during the Feedback 1 condition. The
increased number of uses with shorter duration during the Feedback 2
condition may be explained by the fact that most participants received feed-
back indicating that they were not going to reach their goal very early on,
leading to more frequent use after shorter periods of time. In contrast, it
would likely take longer for participants to determine that they were not
going to reach their goal when Feedback 1 was displayed. The data may lend
support to this explanation, as average latency to use the Start Over button
was 263 seconds for Feedback 1 and 128 seconds for Feedback 2.
The measure of Goal Commitment was not found to be useful or pre-
dictive of performance. While this concept is widely measured and used in
goal setting research in the cognitive psychology literature, the analysis of
performance in this study did not benefit from an analysis of goal commit-
ment. In this study, many participants were found to have similar goal
commitment scores, with dissimilar performance on the work task. Further,
when answers to the goal commitment questions appeared to describe the
performance of the participant, it made the measure unnecessary, as the
performance needed no further description.
In addition to group results, it is important to consider that each partici-
pant entered the study with unique histories of reinforcement and experience
in similar computer tasks. The effects of this can be seen in the wide range of
baseline performances (M= 64.13, SD = 16.92, n= 24). The high/low
performer analysis was informed by a similar analysis performed by See
et al. (2003) who found that different goal levels had different effects on
JOURNAL OF ORGANIZATIONAL BEHAVIOR MANAGEMENT 41

participants considered high and low performers based on no-goal perfor-
mance. Similar results were seen in this study. The highest increase in
performance was achieved by the low performers with the low goal, while
the lowest increase was achieved by the low performers with the high goal.
Whereas the high increase of the low performers with the low goal can be
explained by having more room to improve, the low performers with the
high goal did not exhibit the same gains. Challenges that arise in the high/low
analysis are that the participant that dropped out of the study would have
been included in this data set (low performer/high goal), and therefore this
group included three participants instead of four. In addition, one of the
three participants in this group was the only participant that performed
below baseline on all three experimental sessions (52.1%, 77.1%, 77.1%).
When this participant’s data is removed from the data set, the two remaining
participants average 122.1%, 126.3%, and 153.2%, still resulting in the two
lowest increases in the first two experimental sessions, but the third session is
higher than the high performers with the high goal. Therefore, these results
should be interpreted with caution. By use of the Start Over button, low
performers with the low goal added 9.8 minutes of work, low performers
with the high goal added 9.5 minutes of work, high performers with the low
goal added 41.0 minutes of work, and high performers with the high goal
added 59.4 minutes of work. The high performers persisted at the task longer
than the low performers, working the longest in the high goal condition.
While both the low and high performers performed better in response to the
lower goal than the higher goal, the difference for the high performers was
very slight, and much larger for the low performers. A real world implication
of this finding may be that employees who are currently performing at high
levels may be less susceptible to any negative effects of difficult goals, while
employees performing at low levels may be more susceptible to those nega-
tive effects and would benefit more from a more gradual shaping of their
performance.
This study has several limitations. First is the small sample size and lack of
an a priori power analysis to determine an adequate sample size for statistical
significance. Rounsaville, Carroll, and Onken (2001) promote a “Stage Model
of Behavioral Therapies research”in which Stage I is comprised of pilot
testing, which may lead to larger-scale randomized controlled trial research.
This model recommends utilizing the effect size from pilot studies to deter-
mine the sample size needed in large scale follow-up research, thereby saving
resources by only following up on studies that have promise. By this model,
the current study may inform a follow-up study with a larger sample size,
based on the statistically significant results. Additional limitations include the
inclusion of a second independent variable (feedback type), short session
lengths, and the use of individual goals. Although a statistical analysis
indicates that the differential effect of the two feedback types was not
42 K. M. ROOSE AND W. L. WILLIAMS

statistically significant, future research may utilize only one independent
variable. In terms of external validity, the length of the sessions and the
experiment in general may be too short to support a transfer of the findings
to real world work situations. Future research may utilize longer sessions, or
more sessions over time. While individualized goals, like those set in this
study, may be realistic in some settings, it is possible that the response effort
of setting individual goals for a large amount of employees may be unrealistic
for some managers or organizations.
Another potential limitation is the addition of the Start Over button. A
2.5-hour maximum duration was set; therefore, participants arrived to the
experiment planning to stay for 2.5 hours and to receive 2.5 credits for their
participation. While seven of the experimental participants did not use the
Start Over button, thus completing the study 1.5 hours early, but still
receiving 2.5 hours of credits; other participants may have been more likely
to use the Start Over button due to their original time commitment.
Finally, this study did not include all of the variables that would affect
goal-directed behavior in the real world. For instance, failure to meet a goal
set by management may result in some type of punishment, while this study
did not include a punishment component. In addition, meeting goals in a
work situation may result in reinforcement in the form of praise, promo-
tions, and/or monetary incentives, or negative reinforcement by avoiding
aversive conditions promised or expected for not meeting a goal.
Despite these limitations, these findings warrant a follow-up study with a
larger sample size and refined methodology. In practical settings, it would be
important for managers to know boundaries and components of effective
goal setting. Despite research against their use, stretch goals continue to be
used in work settings, as some employers believe that the highest goals will
produce the highest levels of responding. These results indicate that this may
not always be so, and that difficult goals may lead to other concerns, such as
decreases in accuracy. In addition to negative effects on accuracy, another
concern with difficult goals is their potential to lead to unethical behavior
(e.g., Ordóñez et al., 2009; Schweitzer et al., 2004). While there were no
apparent opportunities to engage in unethical behavior in this study, research
has cautioned that this type of unethical behavior may occur in real world
situations in response to very challenging goals.
While some have argued that the linear relationship between difficulty and
performance has been exaggerated due to goals setting a cap on performance
(Lorenzi, 1988), such a phenomenon was not seen in this study. Of the nine
participants that met their goal, none stopped after reaching it. All but one of
the participants that reached their goal before the final session went on to
perform even better in following session(s). However, in work environments,
and/or over longer periods of time, a capping effect of goals may be seen, and
this effect should be studied independently.
JOURNAL OF ORGANIZATIONAL BEHAVIOR MANAGEMENT 43

According to unsolicited verbal reports, participants in the 175% goal
condition were more likely to set personal goals rather than strive for the
assigned goal, and were more likely to make comments regarding their goal.
When a top performer was given a 175%, the participant stated, “I shouldn’t
have worked so fast!”Another participant with the high goal reported that he
did not use the Start Over button because the goal was “not attainable.”
When a 175% goal was given to participant 357, his response was “I only got
57, now you want me to get 100?”After completing the first experimental
session, and attempting the second experimental session twice, he requested
to quit the study, saying, “I’m trying so hard to be accurate, but I keep
making mistakes. It’s too frustrating.”While there would certainly be other
contingencies in place in an actual work setting (e.g., paycheck, job stability,
coworkers), the potential for task abandonment in the face of very difficult
goals should be considered by employers.
While goal setting has been studied for decades in a variety of fields,
this study indicates the need for continued research on the use of
difficult goals. When setting goals in organizational settings, employers
should be aware of potential break points, or goals set too high for
employees to reach, which may fail to maintain goal-directed behavior.
While goals have been shown to be effective at increasing performance
in a variety of studies, if used incorrectly goals may lead to decreased
performance, or performance that does not meet desired standards.
Further research may shed light on optimal methods of goal setting in
organizations, for high performers, low performers, and all employees.
The answer may be a combination of a variety of goal levels, and
managers may consider each individual’s baseline performance when
selecting goals when possible. Continued research on these topics may
assist in the development of more specific recommendations for goal
setting in work environments.
Acknowledgments
This article is adapted from a thesis submitted in partial requirement for a Master of Arts
degree in psychology from the University of Nevada, Reno. The authors thank Ramona
Houmanfar and Steve Rock for their contributions to this study, and the Aubrey Daniels
Institute for their research support.
References
100 metres. (2016, August 4). In Wikipedia, the free encyclopedia. Retrieved August 6, 2016,
from http://en.wikipedia.org/w/index.php?title=100_metres&oldid=732968301
Agnew, J. L. (1997). The establishing operation in organizational behavior management.
Journal of Organizational Behavior Management,18(1), 7–19. doi:10.1300/J075v18n01_02
44 K. M. ROOSE AND W. L. WILLIAMS

Atkinson, J. (1958). Towards experimental analysis of human motivation in terms of motives,
expectancies and incentives. In J. Atkinson (Ed.), Motives in fantasy, action and society (pp.
288–305). Princeton, NJ: Van Nostrand.
Audia, P., Locke, E. A., & Smith, K. (2000). The paradox of success: An archival and a
laboratory study of strategic persistence following a radical environmental change.
Academy of Management Journal,43, 837–853. doi:10.2307/1556413
Balcazar, F., Hopkins, B. L., & Suarez, Y. (1985). A critical, objective review of performance
feedback. Journal of Organizational Behavior Management,7(3–4), 65–89. doi:10.1300/
J075v07n03_05
Bavelas, J., & Lee, E. S. (1978). Effects of goal level on performance: A trade-off of quantity
and quality. Canadian Journal of Psychology,32(4), 219–240. doi:10.1037/h0081692
Bryan, W. L., & Harter, N. (1897). Studies in the physiology and psychology of the telegraphic
language. Psychological Review,4,27–53. doi:10.1037/h0073806
Catania, A. C. (2007). Learning (interim 4th ed.). Cornwall-on-Hudson, NY: Sloan.
Daniels, A. C. (1994). Bringing out the best in people. New York, NY: McGraw-Hill.
Daniels, A. C. (2009). Oops!: 13 management practices that waste time and money (and what
to do instead). Atlanta, GA: Performance Management Publications.
Drucker, P. F. (1954). The practice of management. New York, NY: Harper & Row.
Duncan, P. K., & Bruwelheide, L. R. (1986). Feedback: Use and possible behavioral functions.
Journal of Organizational Behavior Management,7(3–4), 91–114. doi:10.1300/
J075v07n03_06
Earley, P. C. (1985a). Influence of information, choice and task complexity upon goal
acceptance, performance, and personal goals. Journal of Applied Psychology,70, 481–491.
doi:10.1037/0021-9010.70.3.481
Earley, P. C. (1985b). The influence of goal setting methods on performance, goal acceptance,
self-efficacy expectations, and expectancies across levels of goal difficulty. Paper presented at
the meeting of the American Psychological Association, Los Angeles, CA.
Earley, P. C., & Perry, B. (1987). Work plan availability and performance: An assessment of
task strategy priming on subsequent task completion. Organizational Behavior and Human
Decision Processes,39, 279–302. doi:10.1016/0749-5978(87)90025-2
Erez, M., & Zidon, I. (1984). Effect of goal acceptance on the relationship of goal difficulty
to performance. Journal of Applied Psychology,69(1), 69–78. doi:10.1037/0021-
9010.69.1.69
Fellner, D. J., & Sulzer-Azaroff, B. (1984). A behavioral analysis of goal setting. Journal of
Organizational Behavior Management,6(1), 33–51. doi:10.1300/J075v06n01_03
Garland, H. (1982). Goal levels and task performance: A compelling replication of some
compelling results. Journal of Applied Psychology,67(2), 245–248. doi:10.1037/0021-
9010.67.2.245
Garland, H. (1983). Influence of ability, assigned goals, and normative information on
personal goals and performance: A challenge to the goal attainability assumption.
Journal of Applied Psychology,68(1), 20–30. doi:10.1037/0021-9010.68.1.20
Garland, H. (1984). Relation of effort-performance expectancy to performance in goal setting
experiments. Journal of Applied Psychology,69(1), 79–84. doi:10.1037/0021-9010.69.1.79
Hall, H. K., Weinberg, R. S., & Jackson, A. (1987). Effects of goal specificity, goal difficulty,
and information feedback on endurance performance. Journal of Sports Psychology,9,43–
54. doi:10.1123/jsp.9.1.43
Hatzigeorgiadis, A. (2006). Approach and avoidance during task performance in young men:
The role of goal attainment expectancies. Journal of Sports Sciences,24(3), 299–307.
doi:10.1080/17461390500188645
JOURNAL OF ORGANIZATIONAL BEHAVIOR MANAGEMENT 45

Hayes, S. C., Barnes-Holmes, D., & Roche, B. (2001). Relational frame theory: A post-
Skinnerian account of human language and cognition. New York, NY: Plenum Press.
Houmanfar, R., & Hayes, L. J. (1997). Effects of feedback on task completion, time distribu-
tion and time allocation of graduate students. Journal of Organizational Behavior
Management,18(1), 69–91. doi:10.1300/J075v18n01_05
Huber, V. L., & Neale, M. A. (1986). Effects of cognitive heuristics and goals on negotiator
performance and subsequent goal setting. Organizational Behavior and Human Decision
Processes,38, 342–365. doi:10.1016/0749-5978(86)90005-1
Isley, S. D. (2007). Toward a functional approach to goal setting (Unpublished master’s thesis).
University of North Texas, Denton, TX.
Kang, K., Oah, S., & Dickinson, A. M. (2005). The relative effects of different frequencies of
feedback on work performance. Journal of Organizational Behavior Management,23(4),
21–53. doi:10.1300/J075v23n04_02
Kerr, S., & Landauer, S. (2004). Using stretch goals to promote organizational effectiveness
and personal growth: General Electric and Goldman Sachs. The Academy of Management
Executive,18(4), 134–138. doi:10.5465/AME.2004.15268739
Kerr, S., & LePelley, D. (2013). Stretch goals: Risks, possibilities, and best practices. In E. A.
Locke & G. P. Latham (Eds.), New Developments in Goal Setting and Task Performance,
(pp. 21–31). New York, NY: Routledge.
Klein, H. J., Wesson, M. J., Hollenbeck, J. R., & Alge, B. J. (1999). Goal commitment and the
goal-setting process: Conceptual clarification and empirical synthesis. Journal of Applied
Psychology,84(6), 885. doi:10.1037/0021-9010.84.6.885
Klein, H. J., Wesson, M. J., Hollenbeck, J. R., Wright, P. M., & DeShon, R. P. (2001). The
assessment of goal commitment: A measurement model meta-analysis. Organizational
Behavior and Human Decision Processes,85(1), 32–55. doi:10.1006/obhd.2000.2931
Latham, G. P., & Locke, E. A. (2006). Enhancing the benefits and overcoming the pitfalls of
goal setting. Organizational Dynamics,35(4), 332–340. doi:10.1016/j.orgdyn.2006.08.008
Latham, G. P., & Seijts, G. H. (1999). The effects of proximal and distal goals on performance
on a moderately complex task. Journal of Organizational Behavior,20, 421–429.
doi:10.1002/(SICI)1099-1379(199907)20:4<421::AID-JOB896>3.0.CO;2-#
Latham, G. P., & Yukl, G. A. (1975). A review of research on the application of goal setting in
organizations. Academy of Management Journal,18, 824–845. doi:10.2307/255381
Locke, E. A. (1966). The relationship of intentions to level of performance. The Journal of
Applied Psychology,50(1), 60–66. doi:10.1037/h0022937
Locke, E. A. (1968). Toward a theory of task motivation and incentives. Organizational
Behavior and Human Performance,3, 157–189. doi:10.1016/0030-5073(68)90004-4
Locke, E. A. (1982). Relation of goal level to performance with a short work period and
multiple goal levels. Journal of Applied Psychology,67(4), 512–514. doi:10.1037/0021-
9010.67.4.512
Locke, E. A. (1991). Problems with goal-setting research in sports—And their solution.
Journal of Sport and Exercise Psychology,13(3), 311–316. doi:10.1123/jsep.13.3.311
Locke, E. A. (1996). Motivation through conscious goal setting. Applied and Preventive
Psychology,5(2), 117–124. doi:10.1016/S0962-1849(96)80005-9
Locke, E. A. (2001). Motivation by goal setting. In R. T. Golembiewski (Ed.), Handbook of
organizational behavior (pp. 43–56). New York, NY: Marcel Dekker, Inc.
Locke, E. A., & Bryan, J. F. (1967). Performance goals as determinants of level of performance
and boredom. Journal of Applied Psychology,51(2), 120–130. doi:10.1037/h0024471
Locke, E. A., & Bryan, J. F. (1969). Knowledge of score and goal level as determinants of work
rate. Journal of Applied Psychology,53,59–65. doi:10.1037/h0026736
46 K. M. ROOSE AND W. L. WILLIAMS

Locke, E. A., Frederick, E., Lee, C., & Bobko, P. (1984). Effect of self-efficacy, goals and task
strategies on task performance. Journal of Applied Psychology,69, 241–251. doi:10.1037/
0021-9010.69.2.241
Locke, E. A., & Latham, G. P. (1990). A theory of goal setting & task performance. Englewood
Cliffs, NJ: Prentice-Hall, Inc.
Locke, E. A., & Latham, G. P. (2006). New directions in goal setting theory. Current
Directions in Psychological Science,15(5), 265–268. doi:10.1111/j.1467-8721.2006.00449.x
Locke, E. A., Latham, G. P., & Erez, M. (1988). The determinants of goal commitment. The
Academy of Management Review,13(1), 23–39.
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,32,
269–280. doi:10.1111/j.1744-6570.1978.tb00446.x
Locke, E. A., Shaw, K. N., Saari, L. M., & Latham, G. P. (1981). Goal setting and task
performance: 1969-1980. Psychological Bulletin,90, 125–152. doi:10.1037/0033-
2909.90.1.125
Lorenzi, P. (1988). Underestimated effects of goals and rewards: A systematic replication.
Journal of Organizational Behavior Management,9(2), 59–71. doi:10.1300/J075v09n02_05
Maglieri, K. A. (2007). Evaluation of performance under various pay systems. Dissertation
Abstracts International: Section B: The Physical Sciences and Engineering,68(3–B), 1914.
Malott, R. W. (1993). A theory of rule-governed behavior and organizational behavior
management. Journal of Organizational Behavior Management,12(2), 45–65. doi:10.1300/
J075v12n02_03
Mawhinney, T. C., & Ford, J. D. (1977). The path goal theory of leader effectiveness: An
operant interpretation. Academy of Management Review,2, 398–411.
Mento, A. J., Cartledge, N. D., & Locke, E. A. (1980). Maryland vs. Michigan vs. Minnesota:
Another look at the relationship of expectancy and goal difficulty to task performance.
Organizational Behavior and Human Performance,25, 419–440. doi:10.1016/0030-5073(80)
90038-0
Motowidlo, S. J., Loehr, V., & Dunnette, M. D. (1978). A laboratory study of the effects of
goal specificity on the relationship between probability of success and performance.
Journal of Applied Psychology,63(2), 172–179. doi:10.1037/0021-9010.63.2.172
Mowen, J. C., Middlemist, R. D., & Luther, D. (1981). Joint effects of assigned goal level and
incentive structure on task performance: A laboratory study. Journal of Applied Psychology,
66, 590–603. doi:10.1037/0021-9010.66.5.598
O’Hora, D., & Maglieri, K. A. (2006). Goal statements and goal-directed behavior. Journal of
Organizational Behavior Management,26(1–2), 131–170. doi:10.1300/J075v26n01_06
Ordóñez, L. D., Schweitzer, M. E., Galinsky, A. D., & Bazerman, M. H. (2009). Goals gone
wild: The systematic side effects of overprescribing goal setting. Academy of Management
Perspectives,2,6–17. doi:10.5465/AMP.2009.37007999
Prue, D. M., & Fairbank, J. A. (1981). Performance feedback in organizational behavior
management: A review. Journal of Organizational Behavior Management,3(1), 1–16.
doi:10.1300/J075v03n01_01
Rounsaville, B. J., Carroll, K. M., & Onken, L. S. (2001). A stage model of behavioral therapies
research: Getting started and moving on from stage I. Clinical Psychology: Science and
Practice,8(2), 133–142. doi:10.1093/clipsy.8.2.133
Rummler, G. A., & Brache, A. P. (1995). Improving performance: How to manage the white
space on the organization chart. San Francisco, CA: Jossey-Bass.
Salancik, G. R. (1977). Commitment and the control of organizational behavior and belief. In
B. M. Staw, & G. Salancik (Eds.), New directions in organizational behavior (pp. 1–54).
Chicago, IL: St. Clair.
JOURNAL OF ORGANIZATIONAL BEHAVIOR MANAGEMENT 47

Sales, M. (1970). Some effects of role overload and role underload. Organizational Behavior
and Human Performance,5, 592–608. doi:10.1016/0030-5073(70)90042-5
Schweitzer, M. E., Ordóñez, L., & Douma, B. (2004). Goal setting as a motivator of
unethical behavior. The Academy of Management Journal,47(3), 422–432. doi:10.2307/
20159591
See, K., Heath, C., & Fox, C. (2003). Motivating individual performance with challenging goals:
Is it better to stretch a little or a lot? (Unpublished manuscript). New York University, New
York, NY.
Sherman, S. (1995). Stretch goals: The dark side of asking for miracles. Fortune Magazine,12
(10), 231–232.
Singer, R. N., Korienek, G., Jarvis, D., McColskey, D., & Candeletti, G. (1981). Goal setting
and task persistence. Perceptual and Motor Skills,53, 881–882. doi:10.2466/
pms.1981.53.3.881
Sitkin, S. B., See, K. E., Miller, C. C., Lawless, M. W., & Carton, A. M. (2011). The paradox of
stretch goals: Organizations in pursuit of the seemingly impossible. Academy of
Management Review,36(3), 544–566. doi:10.5465/amr.2008.0038
Smith, G. S. (2013). Exploring the predictive validity of implicit relational assessment procedure
(IRAP) with respect to performance in organizations (Unpublished doctoral dissertation).
University of Nevada, Reno, NV.
Soman, D., & Cheema, A. (2004). When goals are counterproductive: The effects of violation
of a behavioral goal on subsequent performance. Journal of Consumer Research,31,52–62.
doi:10.1086/383423
Stedry, A. C. (1960). Budget control and cost behavior. Pittsburgh, PA: Graduate School of
Industrial Administration, Carnegie Institute of Technology.
Tammemagi, T. (2012). Dynamic analyses of goal-directed behavior (Unpublished doctoral
dissertation). National University of Ireland, Galway, Ireland.
Tammemagi, T., O’Hora, D., & Maglieri, K. A. (2013). The effects of a goal setting interven-
tion on productivity and persistence in an analogue work task. Journal of Organizational
Behavior Management,33(1), 31–54. doi:10.1080/01608061.2013.758001
Taylor, F. W. (1911). Principles of scientific management. New York, NY: Harper & Brothers.
Vance, R. J., & Colella, A. (1990). Effects of two types of feedback on goal acceptance and
personal goals. Journal of Applied Psychology,75(1), 68–76. doi:10.1037/0021-9010.75.1.68
Vroom, V. (1964). The motivation to work. New York, NY: Wiley.
Welsh, D. T., & Ordóñez, L. D. (2014). The dark side of consecutive high performance goals:
Linking goal setting, depletion, and unethical behavior. Organizational Behavior and
Human Decision Processes,123(2), 79–89. doi:10.1016/j.obhdp.2013.07.006
Wilk, L. A., & Redmon, W. K. (1997). The effects of feedback and goal setting on the
productivity and satisfaction of university admissions staff. Journal of Organizational
Behavior Management,18(1), 45–68. doi:10.1300/J075v18n01_04
Wrosch, C., Scheier, M. F., Miller, G. E., Schulz, R., & Carver, C. S. (2003). Adaptive self-
regulation of unattainable goals: Goal disengagement, goal reengagement, and subjective
well-being. Personality and Social Psychology Bulletin,29(12), 1494–1508. doi:10.1177/
0146167203256921
48 K. M. ROOSE AND W. L. WILLIAMS