Developing a Comprehensive Model of Motivation and Performance

Author(s): Kae H. Chung

Source: The Academy of Management Journal, Vol. 11, No. 1 (Mar., 1968), pp. 63-73
Published by: Academy of Management
Stable URL: http://www.jstor.org/stable/255197 .
Accessed: 23/01/2011 10:03

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at .
http://www.jstor.org/action/showPublisher?publisherCode=aom. .

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

Academy of Management is collaborating with JSTOR to digitize, preserve and extend access to The Academy
of Management Journal.

http://www.jstor.org
 ai
Developing Comprehenosie Model of
Motivation and Performance*
The comprehensive model of motiva-
tion and performance was expressed by
the. formula: P =a AblNb2/b3Eb4, where A
is ability, N needs, I incentives, and E
expectancies. An empirical study, with,
a sample of 175 students, supported the
predictability of performance of the
model.

KAE H. CHUNG
Northern Illinois University
INTRODUCTION

Industrial psychologists and managerial practitioners have long been

interested in searching for motivational factors that lead to a high level
of motivation and performance. In this search, many scholars have realized
the importance of needs, incentives, and perceptional variables in analyz-
ing and explaining human motivation. However, there has been a tendency
for researchers to concentrate on one variable or a partial set of these
variables in their explanation of human motivation. As a result, while much
is known about the separate effect of these variables, little is known about
the nature of their simultaneous interactions. The tendency to stress one
particular class of determinants of motivation has led to the development
of such partial theories of motivation as need theory, incentive theory, and
perceptional theory.
This development of partial theories of motivation can be attributed
to the intensive studies within each specialized area of interest. However,
they are so ethnocentric that they have no universal applicability in the
understanding and analyzing of a general class of motivated behavior
which consists of a variety of motivational properties. The lack of a general
theory of motivation in industry handicaps managers, not only in gaining
an understanding of employee motivation, but in finding ways to approach
the problems of motivating employees.

*The author acknowledges the support for this research granted by the Graduate
School, Louisiana State University, in the form of a Graduate Fellowship during the summer
of 1967.
63
64 Academy of Management Journal March

The primary purpose of this study, therefore, is to develop a com-

prehensive model of human motivation which will include a broad class
of determinants of motivation. Various determinants of motivation are
derived from the partial theories of motivation. In this respect, the de-
velopment of a comprehensive model depends upon the development of
the partial theories. On the other hand, the partial theories, in this context,
become more meaningful if the effects and contributions of motivational
factors in each partial theory can be understood and evaluated within the
framework of the whole.
In the comprehensive model of motivation, the strength of motivation
(M) is hypothesized as a multiplicative function of the strengths of needs
(N), incentive (1), and expectancies (E): M=f(NxIxE); or more in-
clusively, M-= f(Ni x li x Ei). This formula is essentially similar to At-
kinson's formula: Ts=MsxPsxIs, where Ts is the tendency to achieve
success, Ps the strength of expectancy of success, and Is the incentive
value of success.' It is, however, different from some conceptuations in
determining the strengths of the major motivational variables. Atkinson's
theory of achievement motivation is mainly concerned with one particular
class of needs, the achievement need, mostly in risk-taking situations,
and the expectancy and incentive variables associated with the achieve-
ment need. Thus, his theory does not specify which motivational factors
should be included under each major variable. Consequently, it does not
consider the effects of other motivational factors in determining the
strengths and the resultant forces of the major motivational variables. In the
present motivation model, however, various needs such as physiological,
safety, affiliation, self-esteem, and self-actualization are taken into con-
sideration in determining the strength of motive to do a task.
As a means of determining the resultant force of the various needs
of an individualwhich pull him in different directions with different strengths,
Hicks introduced a technique of vector analysis.2 The resultant force can
be found by connecting the vectors end to end in a sequence, while keep-
ing the original lengths and directions the same as before. Although
vector analysis of this type may more closely approximate the reality of
motivational phenomena, it is at this point only a conceptual tool, for the
lengths and directions of the vectors cannot be identified.
To make the conceptual scheme operationally workable, however, it
is suggested that the system be operated on an unidimensional basis as an
approximation. For instance, as illustrated in Figure 1, a student's dominant
need may be to study, but he may also have other needs which are opposing

1John W. Atkinson, An Introduction to Motivation (New York: D. Van Nostrand Com-

pany, 1964), p. 242.
2HerbertG. Hicks, The Management of Organizations (New York: McGraw-Hill,1967),
pp. 40-43.
1968 Developing a Comprehensive Model 65

(socialization), partially opposing (money), neutral (self-respect), and

partially identical (self-actualization).

FIGURE 1
Student Needs on an Unidimensional Scale

Self -respect
Socialization
-____________ Study
> 2 3 4 5
5 4 3 2 1

Money Self-actualization

When these needs are operated on a five-point scale, the resultant force
of his various needs can be found by subtracting the sum of the strengths
of the opposing needs (or -4) from the sum of the strengths of positive
needs (or +7). This will then leave three points for his need to study. The
resultant forces for incentives and expectancies also can be found in this
fashion. This operational technique leads to the postulate that the relation-
ship between motivational factors within a major motivational variable (e.g.,
needs, incentives, and expectancies) is additive.
As the major concern of management in dealing with employees is
to find the factors that lead to a high level of performance of employees,
the secondary purpose of the study is to relate the comprehensive model
of motivation to the model of job performance. In the model of per-
formance, performance (P) was hypothesized as a multiplicative function
of ability (A) and motivation (M): P=f(AXIM). This formula was originally
hypothesized by Maier3 and partially tested by some scholars. For in-
stance, Vroom studied the relationships between ability and performance
scores of supervisors who were classified as high, moderate, and low in
motivation. He found fairly high positive correlations between the ability
and performance of supervisors high in motivation, generally low positive
correlations for those moderate in motivation, and zero or slightly negative
correlations for those low in motivation.4 Lawler also partially supported
the hypothesis. In his study of civil service personnel, he found that when

'N. R. F. Maier, Psychology in Industry (2nd ed.; Boston: Houghton-Mifflin, 1955),

pp. 203-204.
WVictorH. Vroom, Some Personality Determinants of the Effects of Participation (Engle-
wood Cliffs, N.J.: Prentice-Hall, 1960), pp. 50-59.
66 Academy of Management Journal March

high ability managers were viewed, a clear relationship between attitudes

(or motivation) and performance was apparent. However, when low ability
individuals were considered, a similar relationship did not exist.5 These
studies imply that the use of an ability variable will improve the correlation
between ability and performance.
But previous studies in this problem did not specify which variables
of performance should be included under both motivation and ability.
Thus, the author attempts to specify the subvariables of motivation by in-
tegrating the comprehensive model of motivation into the performance
model as expressed by the formula:
P=f(AXNXI XE),
or more inclusively,
P=f[AXY,(Ni xlix Ei)].
As the contribution of each major variable on the level of performance
may differ from the others, a further attempt is made to differentiate the
contribution by assigning a coefficient of power to each variable as
follows:
P = aAblNb2lb3Eb4.

This formula represents a general predictive model of motivation and per-

formance in the present study.

METHODOLOGY

Descriptions of the Measures

In order to support the comprehensive model of motivation and per-
formance, an empirical study was undertaken. The data were obtained
from college students whose performances were measured in terms of
final scores and whose abilities were represented by their college
entrance exam scores. The sample of the study included 175 students
who were registered in an introductory management course at Louisiana
State University during the spring semester of 1967. The motivational
measures used in the study were obtained from questionnaires which were
designed to measure the strengths of needs, incentives, and expectancies.
Each student was asked to answer the questions by checking the ap-
propriate alternatives on a five-point scale. Relative scores ranged from
one, representing a weak strength of a variable, to five, representing a
strong strength. Information obtained from the university records in-
cluded both college entrance and final exam scores of the students.
The motive-to-study is the strength of needs that impels the student
towards a goal or a set of goals implicit in the study. The motive-to-study
"Edward E. Lawler, III, "Ability as a Moderator of the Relationship between Job
Attitudes and Job Performance," Personnel Psychology, IXX,No. 2 (Summer, 1966), 153-164.
1968 Developing a Comprehensive Model 67

was derived by summing the responses of the student to those questions

which were designed to measure the degree of his needs for studying.
The motive-to-study is the tendency to offset the motive-to-study and to
prefer a non-study environment. The avoidant score was obtained by sum-
ming the responses of the student to those questions which were de-
signed to measure the extent he felt to offset the positive motive-to-study.
Consequently, the resultant score for the motive measure was derived by
subtracting the avoidant score from the motivant score. The resultant forces
for incentives and expectancy measures were also obtained in this fashion.
The total motivation score was found by multiplying the resultant
scores for the need (N), incentive (1), and expectancy (E) measures: M=
N x IX E. Alternative motivation measures were also obtained by the fol-
lowing formulas: (1) M=(N + I) x E and (2) M = N + I+ E. The total motiva-
tion score which demonstrated the best predictability of performance
among these three alternatives was adopted in the model of performance.
The scores in the performance measure, which were represented by
student final exam scores, were correlated with the products of multiplica-
tion between the motivation and ability scores. In addition an alternative
hypothesis, or P=A+M, was correlated with performance and compared
with the multiplicative model.

Criteria for Selecting the Measures

As a basis for validating the motivational measures, the following
criteria were employed.6 First, each item in the measures has to fit the
concept for which it was used. A few questions which seemed to confuse
students were dropped from the measures. For those researchers who
want to use this type of questionnaire, it is suggested that the semantic
differential technique be used to test the meanings of the questions that
will be used in the measures.
Second, each question within a scale has to indicate a generally
positive intercorrelation with items in the same scale and a negative in-
tercorrelation with items in its avoidant scale. The reason for this criterion
was to differentiate positive motive-to-study factors from negative ones. The
magnitude of intercorrelations indicates the degree of homogeneity among
the items. High intercorrelations among items reflect high reliability and
considerable overlap among items in a measure; low intercorrelations, on
the other hand, reflect unreliability of the measure and a different nature
of the items. In -this study, those items which had statistically insignificant
intercorrelations at .05 confidence level were eliminated from the measures.
Finally, each positive item had to show a generally positive correla-
tion with the performance score; conversely, each negative item had to

"The first two criteria were adopted from Bernard P. Indik, "Measuring Motivation to
Work," Personnel Administration (Nov.-Dec., 1966), p. 42.
68 Academy of Management Journal March

show a generally negative correlation with it. This criterion was adopted
to distinguish the predictivity of performance of each item. Accordingly,
those items which did not show significant correlation at .05 confidence
level were eliminated from the measures.

Limitations of the Study

One of the major difficulties in constructing a comprehensive model
of motivation and performance is to deal with the introspective nature
of the concepts of needs and expectancies, as these concepts are not
directly testable. For this reason, some of the behaviorists, namely phys-
iological and revealed preference behaviorists, do not appreciate the
use of introspective concepts. These groups, therefore, advocate that
scientific knowledge must be derived from directly observable and hence
irrefutable data.7
However, their criticism of the use of the introspective concepts
seems partly unreasonable for the following reasons: First, although the
concepts of "needs" and "expectancies" cannot be directly observable,
they have observable consequences.8 For instance, a person who feels
hungry searches for food. Thus, any information or measures obtained
from these introspective concepts can be tested by comparing the measures
with the observable consequences; hence they too tend to be irrefutable.
Second, as motivational study is aimed at understanding and not merely de-
scribing observed phenomena,9 the concepts of needs and expectancies are
admissible as valid points for theory development.
Another major difficulty was to select a sample of subjects which
would represent the whole population. Theoretically, it may be desirable to
include various subsamples of the universe. In this respect, the findings
of the present study may only be applicable to the particular class of the
universe - students. However, the methodological approach used in the
present study can be utilized in analyzing other types of goal-achieving
behavior. Furthermore, it seems desirable to develop a comprehensive
model for each particular class of the universe, for each group may re-
quire some particular attention unique in each situation.

RESULTS
In this section we will first consider the data concerning the effects
of the major motivational measures on motivation and performance. Then
the findings relevant to the testing of the hypothesis will be presented.

7See Tapas Majumdar, The Measurement of Utility (London: MacMillan, 1958), pp.
78-111.
8lbid., pp. 78-79.
9Ibid., p. 78.
1968 Developing a Comprehensive Model 69

Characteristics of Motivational Measures

The motive measures show correlations of .73 (P<.O1) with motiva-
tion and .56 (P<.O1) with performance (see Table 1). When the motive
measures were compared between high, middle, and low performer groups,
78.2 per cent of high performers scored over eleven on the motive scale and
81.5 per cent of the low performers scored under ten.10 The differences
of the motive measures between high and middle performers (t=1.61,
P=.06) and between middle and low performers (t=1.34, P=.08) are
moderate. The difference between high and low performers (t=3.46,
P<.01) is very significant. These results suggest that the motive scale is
a reliable predictor of motivation and performance. Also, the data can be
interpreted as evidence that the stronger the need or motive to study, the
higher the level of motivation and performance. The students in the
sample ranked the strength of their needs to study, on a five-point scale,
in the following order: (1) job security (4.2), (2) self-esteem (3.9), (3)
economic necessity, (3.8), (4) new knowledge (3.6), and (5) friendship (2.8).

TABLE 1
Intercorrelations Between Motivational
Measures, Ability, and Performance
(N = 175)

Measures IM EM Ml M2 M3 A P

Motive .41 .38 .72 .70 .76 .13 .56

Incentive -.08 .49 .38 .43 .14 .34
Expectancy .68 .79 .69 .16 .47
Motivation 1 .96 .84 .16 .63
Motivation 2 .88 .19 .63
Motivation 3 .20 .59
Ability .46

M1=NxIxE; M2=(N+l)xE; M3=N+I+E

The incentive measures show correlation of .50 (P<.O1) with motivation

and .34 (P<.O1) with performance. Although the correlation coefficients are
statistically significant, this is the least of all the motivational measures.
When the incentive measures are compared between high, middle, and low
performer groups, 58.2 per cent of the high performers scored over 9 and
75.2 per cent of the low performers under 8.11 Although the difference of
incentive measures between high and middle performers (t=.28, P<.25)
is statistically insignificant, the differences between middle and low per-

"'Themotive scores were ranged from one through twenty.

"The incentive scores were ranged from one to eighteen.
70 Academy of Management Journal March

formers (t=1.68, P=.05) and between high and low performers (t=2.15,
P=.02) are significant. These findings indicate that the incentive measures
can also be a predictor of motivation and performance. The students in
the sample ranked the importance of various needs, on a five-point basis,
as follows: (1) grade (4.1), (2) employment (4.0), (3) interesting subject
(3.2), (4) teacher's instructions (3.1), and (5) group effect (2.9).
The expectancy measures show correlation of .68 (P<.01) with moti-
vation and A47(P<.01) with performance. When these measures are com-
pared among high, middle, and low performer groups, 75.0 per cent of
high performers scored over 11 on the expectancy scale and 72.4 per cent
of the low performers under 10.12The differences of the expectancy mea-
sures between high and middle performers (t=.65, P=.25) and between
middle and low performers (t=.80, P=.18) are relatively insignificant, but
the difference between high and low performers (t=1.26, P=.11) is rela-
tively significant. These results also indicate that the expectancy measures
can be useful as a predictor of motivation and performance, but with a
wide variation. A possible reason for the variation may be due to the fact
that people are subject to various personality variables. Nevertheless, the
expectancy scale is a highly reliable predictor of motivation (r=.76,
P<.01) and of performance (r=.39, P<.01). Overall these findings can be
interpreted as evidence for the thesis that an individual is motivated to
the extent that he perceives his goal attainment is probable. The students
in the sample ranked perceptual variables, on a five-point scale, in the
following order: (1) perception of the subject (3.4), (2) perception of his
ability, (3.0), (3) perception of his fellow students (2.6), and (4) perception
of the teacher (2.4).
When we investigate the relationships among the motivational mea-
sures, the following points are identified: first, the motive measure is
positively intercorrelated with the measures of incentive and expectancy.
The results seem to support the thesis: (1) the stronger the need or motive
to perform a task, the higher the incentive value of the task a person is
undertaking; (2) the stronger the need or motive to perform a task, the
more a person considers the task likely to be achieved. Second, the in-
centive measure is, although insignificant, negatively intercorrelated with
the expectancy measure. Especially in low and middle performer groups,
the inverse relationship is more clear and significant (t= -.49, r= -.20,
P<.01). The result can be interpreted as evidence for supporting the thesis
that the subjective incentive value of a task is inversely related to the
subjective probability of attaining the task, for people tend to overesti-
mate (or underestimate) the outcomes which are difficult (or easy) to
achieve.

'2Theexpectancy scores were ranged from one to twenty-two.

1968 Developing a Comprehensive Model 71

Joint Effects of the Measures

The characteristics of the motivational measures are capable of
predicting the level of motivation and performance with a limited capacity.
However, the main thesis of this paper was that the study of the joint
effects of the motivational variables would provide a better predictor of
motivation and performance than the individual measures independently
and, likewise, that the study of the joint effects of motivation and ability
would show a better predictor of performance. The relationships of the
joint effects are investigated by the multiplicative and additive methods.
The multiplicative motivational measure, or Ml=NXIXE, as shown
in Table 1, shows a correlation of .64 (P<.O1) with performance, while its
alternative measures, (1) M2=(N+I)xE and (2) M3=N+I+E, show cor-
relations of .63 and .59 respectively. Although the magnitudes of the
correlations of the joint measures are statistically significant and bigger
than those of the individual measures, the differences between the three
alternative measures are nominal. These results indicate that there are
no significant differences between the alternative joint measures. Never-
theless, the results strongly support the thesis that the joint measures
provide better predictors than the individual measures independently.
In studying the joint effects between motivation and ability, the multi-
plicative measure of motivation is adopted as a major determinant of
performance, for it is the better among the alternatives. This motivation
measure shows a greater correlation with performance (r=.64) than that
of ability (r=.46). Generally, motivation measure is more positively re-
lated to the level of performance than the ability measure. However, when
the differences are compared between high and low performer groups,
as shown in Table 2, the ability measure is more positively related to the
level of performance for the low performers, while motivation is better
correlated to the level of performance of the high performers.

TABLE 2
Correlations of Ability, Motivation,
and Alternative Performance Measures with
Actual Performance Among Various Performers
(N = 175)

Measures (1) High (2) Middle (3) Low Total

Ability .38 .36 .48 .46

Motivation .37 .35 .26 .64
P1=AXM .62 .45 .31 .66
P2=A+M .61 .41 .26 .63
72 Academy of Management Journal March

The multiplicative performance measure, or P=AXM, shows a cor-

relation of .66 (P<.01) with performance, while its alternative measure,
or P=A+M, shows a correlation of .63. Unfortunately, the difference be-
tween the two alternatives is not enough to support the multiplicative hy-
pothesis. However, the multiplicative method shows slightly better pre-
dictors of performance, throughout the various performer groups, than
those of the additive method.
Theoretically, it was also expected that the multiplicative perform-
ance measure would significantly improve the predictability of perform-
ance. The results of the study only slightly support the point. In high and
middle performer groups, however, the multiplicative performance mea-
sures (r=.62 and r=.45) demonstrate much better predictabilities of per-
formance than ability (r=.38 and r=.36) and motivation (r=.37 and r=.35)
respectively. Furthermore, the overall multiplicative measure is at least a
better predictor of performance than those of the ability, motivation, and
additive measures. Considering all those results, the study also supports
the multiplicative performance model.
As the findings generally support the multiplicative nature of motiva-
tional and performance models in the study, they may also support the
multiplicative nature of the comprehensive model of performance: P=
f(A x N x I x E). When a further attempt is made to differentiate the degree
of contribution of each major variable on the level of performance, the
constant for the model and the coefficients of power for the variables are
found as follows: P=48.5A 1.34N 1.19I 1.23E 1.24. This formula is a pre-
dictive model of performance of students when their academic activities
in class are measured on a 200-point scale. The coefficient of correlation,
or the degree of association, between the product of the model and the
actual level of performance of students, R =.66 (R2=.45), is statistically
significant. This correlation coefficient indicates that the model is capable
of predicting performance of students. However, to be applicable in its
present form in practical situations will require improvement in the mea-
sured used, for the size of correlation is not enough to insure a reliable
prediction of performance.

SUMMARY AND CONCLUSION

The major theses of the study are (1) that the study of joint interactions
among the three motivational variables would provide a better prediction
of motivation and performance than individual variables independently
and (2) that the study of joint interactions between motivation and ability
would provide a better tool of predicting performance. The findings of
the present study strongly supported the major theses.
Concerning the attempt to define the functional relationships be-
tween the major variables, the results of the study slightly support the mul-
1968 Developing a Comprehensive Model 73

tiplicative nature of the motivation and performance formulas: (1) M=

f(N X I X E) and (2) P = f(A x M), or P = f(A x N x I x E). Assigning weighting
factors for each variable, the comprehensive model was expressed by the
formula: P=aAblNb2lb3XEb4, where a is the constant and bl - b4 are
the coefficients of power for the variables. The constant and coefficients
can be specified for a particular class or a set of classes of population.
At this moment, the reliability of the comprehensive model in pre-
dicting the level of performance is relatively low for practical uses. How-
ever, by improving the measures in the future study we can hopefully
develop a highly reliable predictive model of performance.

ANNOUNCEMENT: BACK ISSUES

In the December, 1966 issue on page 376 reference was made

to negotiations then under way for the handling of back issues of
the Academy of Management Journal. Arrangements have now been
completed. Interested individuals may secure back issues from:

Johnson Reprint Corporation

111 Fifth Avenue
New York, New York 10003

Single copies of Volumes 1 through 5 are priced at $4.00 per number.

Copies of issues in Volumes 6 through 9 are priced at $3.00 per num-
ber. The Academy, through its Journal, will no longer sell back
issues.