The Development of Goal Setting Theory

LEGACIES IN MOTIVATION SCIENCE

The Development of Goal Setting Theory: A Half Century Retrospective

Edwin A. Locke University of Maryland

Gary P. Latham University of Toronto

This chapter summarizes the authors’ joint development of the goal setting theory. The basic concept was based on more than 50 years of research and the formal theory has endured for 28 years (Locke & Latham, 1990). The theory was not developed through overgeneralization from only a few studies or by deduction but rather by induction. The inductions involved the integration of hundreds of studies involving thousands of participants. The theory initially focused solely on consciously set goals. To date, the goal setting theory has shown generality across participants, tasks, nationality, goal source, settings, experimental designs, outcome variables, levels of analysis (individ- ual, group, division, and organizational), and time spans. The theory identifies both mediators and moderators of goal effects. Numerous subsequent studies since 1990 have supported the main tenets of the theory. New findings have enlarged our knowl- edge of the relevant mediators and moderators as well as showing new applications (Locke & Latham, 2013). Among these discoveries are when to set learning rather than performance goals, the effect of goals primed in the subconscious on job performance, and that goal effects are enhanced by having people write at length about them.

Keywords: goal setting, motivation, theory building

The present authors independently discovered the importance of goal setting for significantly improving the performance of individuals and teams. We subsequently formed a research part- nership in 1974 (Latham & Locke, 1975) that has continued to the present day (e.g., Latham & Locke, 2018). In this chapter we describe our individual discoveries, our joint research that led to the development of the goal setting theory in 1990, and new developments to the theory since that time period.

Locke

I entered graduate school in the Department of Psychology at Cornell University in 1960.

My first course was Introduction to Industrial- Organizational (I-O) Psychology taught by Pa- tricia Cain Smith, who was among the earliest and most respected female industrial psycholo- gists. The textbook she assigned to the students in this course was Principles of Industrial Psy- chology that she had coauthored with Thomas A. Ryan (Ryan & Smith, 1954). A figure in a chapter in this book entitled Industrial Motiva- tion showed the results of an experiment con- ducted in England by C. A. Mace (1935). The figure (p. 397) showed that employees who were given a specific goal to attain each day on a computation task showed markedly greater im- provement than those who had been told to do their best. However, no statistical tests had been used to analyze the data. Nevertheless, this ap- peared to be a promising way to approach the topic of work motivation.1

To put this in context, I must digress with a discussion of the field of psychology at that

1 In the late 1960s, Locke met Mace in England. He was excited by Locke’s findings.

Edwin A. Locke, R. H. Smith School of Business, Uni- versity of Maryland; Gary P. Latham, Rotman School of Management and Departments of Psychology, Industrial Relations, and Nursing, University of Toronto.

Correspondence concerning this article should be ad- dressed to Edwin A. Locke, 2451 Norwalk Avenue, Los Angeles, CA 90041. E-mail: elocke@rhsmith.umd.edu

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

Motivation Science © 2019 American Psychological Association 2019, Vol. 5, No. 2, 93–105 2333-8113/19/$12.00 http://dx.doi.org/10.1037/mot0000127

93

 

 

point in time. The dominant philosophy under- lying empirical experiments was behaviorism, the doctrine that human action can be predicted, explained, and controlled without reference to consciousness. John B. Watson (1924), the founder of behaviorism, argued that because consciousness can be neither weighed nor mea- sured as with physical entities, it should not be considered as part of science. Only observable behavior should be studied. Behaviorists denied free will and asserted that psychology should embrace environmental determinism.2

The philosophy of behaviorism was accepted by most psychologists in Canada and the United States because it did not encounter much oppo- sition. For example, instinct theory turned out to be a dead end because people were not found to have inborn goals or knowledge. Superficially, the school of introspectionism seemed better than the study of instincts because it dealt with consciousness, but its core was narrowly fo- cused on reports of physical sensations. The success psychologists had in influencing animal behavior at that time through the use of rewards was considered to provide support for behavior- ism.

Ryan (1970) rejected behaviorism as inade- quate and asserted that conscious intentions should be studied. He was writing his book on intentions during my time in the Cornell doc- toral program (1960–1964). Thus, I was able to read drafts of his chapters.

In this same time period, I was reading the work of philosopher Ayn Rand (for a summary see Peikoff, 1991). I learned that consciousness had the status of an axiom, that is, a concept that along with other axioms such as existence, formed the basis for all knowledge and are self-evident to perception and thus irrefutable. Consequently, I knew that behaviorism, includ- ing Skinner’s (1953) operant model and envi- ronmental determinism, was wrong in principle. As a result, I wrote a number of critiques of behaviorism (e.g., Locke, 1971) and its alleged applicability to management (Locke, 1977). Ul- timately the behaviorist doctrine failed because of its inadequacy as an explanation of human action.

Another aspect of Ryan’s book proved espe- cially useful to me. He reviewed the most com- mon, contemporary attempts at dealing with the issue of motivation, aside from the behaviorist’s concept of reinforcement. He noted the substan-

tial Freudian influence on some psychologists such as Henry Murray. Murray (1938) viewed motivation as stemming from an individual’s unconscious motives or general needs as mea- sured by projective tests. These were often poor predictors of job performance.3 The thrust of Ryan’s book is an approach to motivation that includes conscious, task, and situationally spe- cific intentions on the premise that these are the most direct regulators of an individual’s actions.

Armed with philosophical confidence, I did my dissertation on goal setting, under Ryan’s and Smiths’ supervision. Following in Mace’s (1935) footsteps, I added task and goal variety and, most importantly, statistical tests. The lab- oratory experiments I conducted supported Mace, and I was able to publish my results. I was subsequently hired by the American Insti- tutes for Research where the director of the Washington office, Ed Fleishman, helped me get a grant from the Office of Naval Research. With my research assistant, Judy Bryan, I con- ducted additional laboratory experiments on the goal-performance relationship. Two years later I was hired by the University of Maryland De- partment of Psychology. There I continued my research on goal setting.

Here it is worth pointing out why a large number of researchers proceeded to conduct goal setting experiments after seeing these re- sults. Goal setting is a technique that, if properly implemented, works, and it works reliably to increase an individual and a team’s perfor- mance. Researchers like their experiments to succeed because it is difficult to publish null findings.

Frederick Herzberg’s motivator-hygiene the- ory (Herzberg, Mausner, & Snyderman, 1959) was a competing theory to goal setting. Herz- berg’s theory asserted that extrinsic aspects of the job, that is, hygiene variables (e.g., an em- ployee’s pay) cause job dissatisfaction but not job satisfaction. The theory claimed that intrin-

2 Some psychologists who adhered to this model viewed the study of consciousness in psychology as a regression to mysticism.

3 McClelland, Atkinson, Clark, and Lowell (1953) sub- conscious need for achievement motive predicts entrepre- neurship, but it does so no better than self-reports of con- scious achievement motivation (Collins, Hanges, & Locke, 2004). For recent data on this issue, see Howard (2013) and Locke (2015).

94 LOCKE AND LATHAM

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

 

 

sic aspects of the job, that is motivating vari- ables (e.g., such as the work itself) influence job satisfaction but not dissatisfaction. This theory was based on only two studies and did not focus directly on performance. The methodology used in both studies was the critical incident tech- nique (CIT), a technique originally developed for conducting a job analysis, not for identifying sources of job satisfaction (Flanagan, 1954). Research that used different methodologies did not replicate Herzberg’s findings (e.g., Locke, 1976).

Another competitor to goal setting in the 1960s–1980s was the expectancy theory (Vroom, 1964). This theory states that an em- ployee’s decisions are made by multiplying va- lence (value) of a Task � Expectancy of per- forming a Task Effectively � Instrumentality to the individual for doing so, and the product of these in turn leads to making choices. This theory was developed through deduction. How- ever, it turned out that people do not typically make choices by performing such multiplica- tions and Vroom later admitted this fact (Latham, 2012).

Another motivation theory, the prospect the- ory, is based on Kahneman and Tversky’s re- search. Heath, Larrick, and Wu (1999) and Wu, Heath, and Larrick (2008) claimed that the pros- pect theory parsimoniously explains goal set- ting results when, in fact, few if any of its claims about goals and goal setting theory are correct (Locke & Latham, in press). The claims of prospect theorists include misstatements about goal setting such as: (a) the theory does not take into account reference standards, even though a goal is a standard for self-evaluation; (b) it cannot explain the effect on performance of setting a difficult-to-attain goal; (c) claiming goal difficulty and specificity are the same; (d) failing to take into account the theory’s moder- ators (e.g., goal commitment) and goal determi- nants (e.g., values, self-efficacy); and (e) an over- reliance on paper people experiments in which an individual’s actual performance is not mea- sured.

Latham

In the fall of 1967, I began my pursuit of a master’s degree in I-O psychology at the Geor- gia Institute of Technology. The I-O faculty, most of whom had served in the military as

psychologists during World War II, instilled in students the desire to make a difference as sci- entist-practitioners. I was assigned to work with William Ronan as his research assistant. Ronan was a former student of Flanagan (1954), the developer of the critical incident technique noted above.

In the summer of 1968, Ronan was hired by the American Pulpwood Association to find ways to increase the productivity of pulpwood crews. Organizations such as International Pa- per were dependent on these crews for timber to make paper. Ronan accepted the consulting as- signment on the condition that I be hired as a research assistant for that summer to initiate and implement a project that would be of sufficient quality to serve as a master’s thesis. I chose to conduct a job analysis, using the CIT, to iden- tify behaviors that differentiate effective from ineffective pulpwood crews. Among my discov- eries was that effective crews set specific pro- duction goals for the amount of wood that they would harvest in a day and/or week.

During the second year of my master’s pro- gram, Ronan was factor analyzing the results of questionnaires that had been completed by for- esters regarding their observations of pulpwood crews. Fortuitously, the questionnaire included an item on setting a weekly goal for the amount of wood to be harvested. This item loaded on the same factor as an objective measure of crew productivity (i.e., cords per employee hour). Thus, two different methodologies, the CIT and a questionnaire, yielded the same conclusion: There is a goal-performance relationship.

The very day in 1969 that I presented the results of my thesis to the American Pulpwood Association, the organization hired me to be its first staff psychologist. On Saturdays I fre- quently returned to the Georgia Tech library where I perused the quarterly issues of Psycho- logical Abstracts. In issue after issue, I read abstracts of journal articles that described labo- ratory experiments in which people who had been assigned a specific, high goal brainstormed more ideas, solved more arithmetic problems, made more tinker toys, and so forth than did those who had been randomly assigned to a placebo (control) condition in which they had been urged to do their best. Racing into his office the following week, I excitedly ex- claimed, “Dr. Ronan, an I-O psychologist by the name of Locke has found. . . .” Suddenly, we

95DEVELOPMENT OF GOAL SETTING THEORY

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

 

 

had the beginning of a theoretical framework for conducting our research and explaining our inductively obtained results (e.g., Locke, 1968).

Based on Locke’s laboratory experiments, Sid Kinne, a PhD forester who reported directly to Georgia Kraft Company’s CEO, convinced him to allow the two of us to conduct a field experiment in which we randomly assigned pulpwood crews to an experimental condition in which they were assigned a specific, high- productivity goal or to a placebo condition in which they were urged to do their best. The importance of the latter exhortation to these crews was highly relevant for them because they were paid on a piece-rate basis. In the very first week and throughout this 3-month experi- ment, the crews that had specific, high goals outperformed those in the control condition. The goals provided the crews with a purpose, a sense of challenge, and feelings of accomplish- ment for otherwise tedious work. Consequently, in addition to significant increases in weekly productivity, job attendance soared in the goal- setting condition because cutting trees was now meaningful to the crews. The goal was a stan- dard for self-assessing their effectiveness (Latham & Kinne, 1974).

By this time Ronan was far more than my thesis supervisor; he was a trusted mentor to me. Sadly, the psychology department’s empha- sis in that time period was solely on industrial psychology and human factors engineering. There was no organizational psychology faculty. Thus, rather than return to Georgia Tech to pursue a PhD, I entered the doctoral program in psychology at the University of Akron (Akron, OH) in 1971. Two things had attracted me to that department. First, only a PhD in I-O psy- chology was offered; second, the department had recently hired Gary Yukl, a rapidly rising star in I-O psychology.

The assigned readings in Yukl’s doctoral sem- inar included Locke’s research. Hence, I showed Yukl the technical reports on goal set- ting published by the American Pulpwood As- sociation for its client companies. He immedi- ately urged me to send copies to Locke. Within the week Locke sent me a letter in which he suggested that I submit them for publication, which I did (Latham & Kinne, 1974; Ronan, Latham, & Kinne, 1973). The timing was per- fect because that was the time period in which Locke’s goal-setting experiments were being

criticized for lack of external/ecological validity (e.g., Heneman & Schwab, 1972; Hinrichs, 1970).

Unbeknownst to me, the global forest prod- ucts company, Weyerhaeuser, was aware of my research conducted at the American Pulpwood Association as well as my progress in the doc- toral program at the University of Akron. In November 1972, they offered me a job starting in June 1973 as their first staff psychologist. I immediately accepted the offer when the com- pany informed me that upon joining them I could choose any topic for my dissertation and that I would be given all the resources I would need to complete it. The reason for this won- derful offer was the results of my goal-setting research involving pulpwood crews.

Weyerhaeuser’s senior management had been astounded that something so simple and straightforward as setting a specific, high goal could have such a positive effect on an employ- ee’s and a team’s performance. “Doesn’t every- one set goals?” was the question commonly asked of me. The answer was and is yes, but the goals are almost always general or vague in nature and thus have little or no effect on a person’s behavior. Remaining a bit skeptical yet intrigued by my findings, they asked whether “my methodology” could be tweaked to further inspire productivity.

As a doctoral student, I was impressed by Rensis Likert’s (1967) research on principles that he had labeled a system 4 style of leader- ship, namely encouraging employee participa- tion in decision making, goal setting, and devel- oping a supportive relationship with subordinates. Thus, my response to the question of tweaking goal setting determined my choice of my doc- toral dissertation. I compared the effect of as- signed versus participatively set goals on per- formance versus a do-your-best condition (Latham & Yukl, 1975). This led to program- matic research involving about 17 subsequent field and laboratory experiments on this topic. I found that the crews that participated in setting their performance goals had the highest produc- tivity. Moreover, they set higher goals than those who had been assigned goals by the crew supervisor. The goal setting theory states that high goals lead to higher performance than easy goals (Latham & Locke, 2018; Locke & Latham, 1990).

96 LOCKE AND LATHAM

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

 

 

Among my other findings in that time period was that goals improve performance in the ab- sence of monetary incentives. For example, they did so for unionized loggers (Latham & Baldes, 1975) and truck drivers (Latham & Saari, 1982),4 who were paid by the hour. They also did so for scientists/engineers, most of whom possessed graduate degrees. Consistent with the goal setting theory, those individuals who were urged to do their best despite receiv- ing praise, public recognition, or a monetary bonus performed no better than those in the control group. The scientists/engineers who par- ticipated in the goal-setting process had the same level of goal commitment as those who had been assigned goals. But, as was the case with the loggers (Latham & Yukl, 1975), the difficulty level of the participatively set goals was higher than the goals that had been assigned by a manager. For the same reason, as was the case with the loggers, job performance was highest in the participative goal condition (Latham, Mitchell, & Dossett, 1978). Subse- quent laboratory experiments (e.g., Latham & Saari, 1979; Latham, Steele, & Saari, 1982), as well as an earlier field experiment involving word processing operators (Latham & Yukl, 1976), showed that when goal difficulty be- tween conditions is the same, the performance of those with assigned versus participatively set goals does not differ.

Locke and I met at the 1974 annual meeting of the American Psychological Association. As noted earlier, we coauthored our first paper to- gether a year later (Latham & Locke, 1975). Based on data from the American Pulpwood Association, we discovered that pulpwood crews, paid on a piece rate basis and restricted to cutting wood to 2–3 days a week cut as much as they normally did in a 5-day week. The restricted number of days had become a spe- cific, challenging time frame (i.e., a goal).

Theory Building

We did not begin our research with theory building in mind. Because of Herzberg, we were acutely aware of the dangers of premature the- orizing (Locke, 1976, 2007; Locke & Latham, 2005). As we noted earlier, Herzberg’s database was very small, his methodology was dubious, and his results could not be replicated using different and sounder methodologies.

In 1968, Locke published an article entitled “Toward a Theory of Task Motivation and In- centives” based primarily on the results of his early goal-setting experiments and the ideas of Mace. But there was insufficient empirical evi- dence at that time to build a formal theory. However, the evidence supporting the develop- ment of a theory soon began to accumulate based on hundreds of studies conducted by our- selves and others. Thus, in 1990 we published a book that presents the goal setting theory based on approximately 400 studies (Locke & Latham, 1990).

A good theory must be based on a clear defi- nition of its concept or concepts (Locke, 2003).

Key concepts require careful measurement. We discovered empirically that the best goal measure was: “What is the minimum score you would be satisfied with?” (An untested alterna- tive would be, “What is the lowest score you would not be dissatisfied with?”). Previously such measures had not been used. Usually peo- ple had simply been asked to try for X.

Several meta-analyses had been done on the effects of goals on task performance (Locke & Latham, 1990, Tables 2-1 and 2-2). These anal- yses show that people with specific, challenging goals reliably outperform those with do-your- best goals because the latter type of goal is interpreted too subjectively. Moreover, the de- gree of goal challenge or difficulty is linearly related to performance, given sufficient skill or ability. We concluded that the most effective goals for increasing performance are those that are specific and difficult. With regard to goal specificity, we found that it alone does not nec- essarily lead to high performance because a goal can be both specific and easy to attain. We found that specific goals in and of themselves affect the variance in performance only to the degree that performance is controllable.

Often overlooked by subsequent researchers are our Appendixes C and D in Locke and Latham (1990), which present guidelines for conducting laboratory and field experiments. Ignoring these guidelines can lead to substandard performance and erroneous conclusions. For example, as- signing impossible goals in a laboratory in- creases performance because there are no pen-

4 We never had a union grievance filed by a logger or a driver over goal setting.

97DEVELOPMENT OF GOAL SETTING THEORY

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

 

 

alties for failure to attain them. Impossible goals can motivate in the short run if people try hard to attain them. In field settings, impossible-to- attain goals can lead to demoralization and pun- ishment. In organizations, goals should be chal- lenging yet attainable.

Generality. Inductive theory building re- quires evidence of generality. Although we had no formal theory of induction when we began our research, we reported evidence of generality across tasks (n � 88), participants (n � 40,000), countries (n � 7), outcome measures (n � 10), and time spans ranging from 1 min to several years,5 designs (experimental, correlational), settings (laboratory, simulation, field), and goal sources (assigned, self-set, participatively set). Studies also showed that goal setting could be used successfully with groups/teams, divisions, and even small organizations (e.g., Porter & Latham, 2013; Pritchard, Young, Koenig, Sch- merling, & Dixon, 2013).

Our focus on generality has implications for how the issue of replication might best be ad- dressed in the psychological sciences and maybe elsewhere. The emphasis in many dis- cussions of replication has been on exact repli- cation of single studies. But it is hard to know how a study using one task, one set of instruc- tions, one setting, one type of measure, one time span, and one class of participants and so forth will generalize. Our view is that generality is best achieved by replication with variation of the type used in our research program.

Mediators. Goal research showed that goal mediators include choice/attention, effort, and persistence. Goals were also found to motivate people to use existing strategies for goal attain- ment or to discover new ones (Seijts & Latham, 2005; Winters & Latham, 1996). Having rele- vant strategies for goal attainment is a fourth mediator.

An important finding of goal research on the opposite side of the mediator coin showed that self-set goals along with self-efficacy could me- diate the results of other motivators on perfor- mance (e.g., assigned goals, feedback, person- ality, incentives, job design, and leadership). This research was updated by Locke (2001).

Moderators. We identified four modera- tors of goal-performance effects. Feedback is critical to goal effects because it enables people to track progress so that effort and strategy can be adjusted to attain the goal. Goals and feed-

back work better together to increase perfor- mance than either one alone.

A second moderator is goal commitment. A goal that one is not committed to attain will not affect that person’s actions. The ultimate proof of commitment is action, but self-report scales can be useful (Klein, Cooper, & Monahan, 2013). Commitment is especially important when a goal is difficult to attain because the goal requires more effort and persistence when setbacks are inevita- bly experienced. Commitment is affected by val- ues, including incentives, and self-efficacy.

A third moderator, which as noted is also a mediator, is ability, namely knowledge or skill. People cannot attain goals if they do not know how to do so. This is an example of motivation and cognition working together (see Wood, Whelan, Sojo, & Wong, 2013). Perceptions by supervisors that the goals assigned to them by their managers are excessively difficult has been shown subsequently to be related to their abuse of employees (Mawritz, Folger, & Latham, 2014).

Situational factors. Situational factors, a fourth moderator, affect the goal-performance relationship. Goal-directed action may be facil- itated or hindered by environmental factors and the degree of support an individual receives (e.g., people, money, facilities).

Affect. Goals are by their nature something one values. Emotions are based on subconscious value judgments (Locke, 2009). Thus, goal attain- ment is related to affect (see Locke & Latham, 1990, Chapter 10). Numerous studies show that goal attainment is related to satisfaction. However, there is an apparent paradox. Difficult goals are less likely to be attained than easier goals, thus making satisfaction harder to experience. So why do people try to attain them? The explanation, provided by an experiment by Mento, Locke, and Klein (1992), is that attaining challenging goals is often the path to more internal and external ben- efits than easier goals (e.g., pride, educational cre- dentials, better job, higher pay).

"Is this question part of your assignment? We can help"

ORDER NOW