You are on page 1of 12

RICHARD L.

OLIVER*

As a result of a lack of empirical investigotion, the variance in salesmen's


performance attributable to motivational constructs has not been estimated.
Vroomian expectancy theory v^as used to shov/that the motivational perceptions
attributed to a set of sales "incentives" by a sample of life insurance salesmen
were related to two performance criteria.

Expectancy Theory Predictions of Salesmen's


Performance

INTRODUCTION motivation. It is further assumed that other incentives


The determinants of salesmen's performance have or stimulators (e.g., contests, bonuses, conventions)
long intrigued researchers and practitioners alike. operate only to induce performance over and above
Unfortunately, little information has been amassed that which can be engendered from the basic plan.
in this area. Rather, the bulk of current knowledge While many of the managerial tools used by the sales
has been accumulated in the domain of sales aptitude administrator provide strong motivating forces, in in-
where the evidence is substantial. Based on results stances they are insufficient and additional incentives
reported in numerous studies, Ghiselli found that the are required. Most sales executives agree that a sound
average validity coefficient between sales proficiency compensation plan can be the strongest force to moti-
criteria and tests of intellectual abilities was .33 while vate salesmen [43, p. 430].
that for personality (including interest) tests was .30 Conclusions such as the one just cited are often
[15, p. 470]. Although these average coefficients are supported by reference to an early study by Haring
high relative to those obtained on other occupational and Myers [21]. Questionnaires, designed to probe
groups, the two aptitude measures taken together as the use of special incentives for salesmen, were mailed
independent predictors would not explain more than to members of the National Sales Executives repre-
20% of the variance in performance. Thus, over 80% senting approximately 8,000 companies. Of 542 re-
of the variance remains to be explained by other spondents, 396 answered a question pertaining to the
constructs. effectiveness of various incentives in stimulating the
The subject of motivation has also received exten- average salesman to better his usual performance. The
sive discussion in the sales management literature but researchers found that the basic compensation plan
little substantive knowledge has emerged. This may, was cited as the first ranked incentive by 243 respond-
in part, reflect the fact that authors usually subsume ents and concluded that "basic compensation is the
motivation under the broad category of compensation primary motivator of salesmen" [21, p. 158].
or under other financial remuneration schemes some- This conclusion is unwarranted for three reasons.
times classified as "stimulators" (see, for example, First, the authors have made inferences about the
[10, 35, 43, 44]). Writings on the subject frequently relative effectiveness of various "motivators" solely
contain an implicit assumption that the basic on the basis of superiors' reports. However, it has
compensation package is the primary regulator of been noted elsewhere [28, 34] that, when asked to
rank the importance subordinates attach to various
job factors, managers typically overemphasize the
*Richard L. Oliver is Assistant Professor in the Department of importance of pay. The results of the Haring and
Business Administration, University of Kentucky. He wishes to Myers [21] study may have been similarly biased.
thank Professors Shelby D. Hunt, Donald P. Schwab, and Herbert
H. Heneman, III of the University of Wisconsin for their counsel Second, the study provided no evidence of validity.
and assistance. While a number of possible incentives were identified
in the study, Haring and Myers fell short of providing
243

Journal of Marketing Research


Vol. XI (August 1974), 243-53
244 JOURNAL OF MARKETING RESEARCH, AUGUST 1974

empirical support for their assertion. A predictive In contrast to the paucity of theory and evidence
study testing the validity of the various "motivators" regarding the performance of sales personnel, consid-
would be required to substantiate their conclusion. erable research has been undertaken in the area of
Third, the results of one mail survey do not provide industrial psychology and significant progress has been
sufficient evidence for the authors' sweeping conclu- made in increasing the state of current knowledge
sion, especially in view of the paucity of empirical on employee productivity as a function of motivation
research on financial compensation. Opsahl and Dun- at both the theoretical and empirical levels. Much
nette observed that: of industrial motivation theory has direct applicability
(W)e know amazingly little about how money either to the motivation of sales personnel.
interacts with other factors or how it acts individually VROOMIAN EXPECTANCY THEORY
to affect job behavior. Although the relative literature
is voluminous, much more has been written about the Expectancy theory as applied to employee motiva-
subject than is actually known. Speculation, accompa- tion is generally attributed to Vroom [46], although
nied by compensation fads and fashions, abounds . . . its historical roots are in the works of Lewin [29]
[34, p. 94]. and Tolman [45] who postulated that organisms de-
Opsahl and Dunnette further noted that compensation velop cognitive expectancies regarding the outcomes
of behavior and consequently behave in a manner
as a primary motivator of organizational behavior is which is likely to result in preferred outcome states.
only one of many possible theoretical roles posited
in the literature. In short, no evidence exists to suggest Generally, expectancy theory posits that the mo-
that money is either a primary motivator or is primary tivational force experienced by an individual to select
on a hierarchy of motivators. one behavior from a larger set is some function of
To a lesser extent, the motivating qualities of non- the perceived likelihood that that behavior will result
in the attainment of various outcomes weighted by
monetary or psychological incentives have also been the desirability (valence) of these outcomes to the
prominently argued in the literature, although they person [25]. Thus, it is essentially a process theory
are usually suggested as adjuncts to the basic in that its focus is on the major classes of motivational
compensation package. Lists of psychological needs constructs and the manner in which they interact as
requiring satisfaction have been proffered by many opposed to detailing the specific outcomes or needs
writers in the sales management field (see, for example, that presumably motivate behavior [5]. As such it
[4, 14, 31]) but have not been subjected to empirical may help to integrate previous writings on the content
investigation. As a result of the lack of evidence in of sales motivation.
both the monetary and nonmonetary areas of motiva-
tion, little is known about how compensation and other Vroom's [46] theory hypothesizes that employee
possible rewards motivate salesmen to produce. job performance (P) is a function of the multiplicative
Preoccupation with the content of motivational interaction between motivation (M) and ability (A).'
schemes (e.g., pay, psychological needs) has led writ- Thus:
ers to ignore the cognitive processes by which behavior (1) P=f(MxA).
is initiated, directed, and continued. That is, theorists
have been content merely to suggest specific things The rationale for the multiplicative relationship is that
that motivate behavior rather than to delineate the if an individual is low on either performance compo-
processes by which major classes of variables interact nent, then his performance must be necessarily low
to produce behavior. Campbell, Dunnette, Lawler, as well.
and Weick summarized this point succinctly: Motivation, in turn, is hypothesized to be a function
of the multiplicative interaction of the valence of one's
A motivational theory is useful for making predictions performance goal / (Vp and the subjective probability
only to the extent that it specifies both content and or expectancy that one's efforts will result in the
process, that is, to the extent that it specifies the identity attainment of that performance goal (Ej). Thus:
of the important variables and the processes by which
they influence behavior [5, pp. 341-2]. (2) M=f(V,. x £ p .
As a result of the lack of a comprehensive motiva- A performance level is seeh as acquiring valence
tional model, the determinants of salesmen's motiva- only if it is perceived as leading to the attainment
tion remain essentially unknown. Without this knowl- of desired job-related outcomes such as pay or recog-
edge, one is unable to assess empirically the propor-
tion of variance in performance attributable to motiva-
tion. The implications for the sales manager are
' In the strict theoretical sense, Vroom, [46] posits a
substantial. In order to increase production through discrete level of motivation for each effort and performance
an optimal allocation of resources, the relative impact level. However, when effort is construed generically and
of motivation as opposed to ability and other factors one performance level is singled out as the criterion of
on performance must be known. interest, Vroom's conceptualization reduces to (1) [33].
EXPECTANCY THEORY PREDICTIONS OF SALESMEN'S PERFORMANCE 245

nition, or to the blocking of undesirable outcomes this nature have not been forthcoming. Thus, a test
such as being terminated. The desirability of a job- of the separate effects of each set should, in part,
related outcome, k, is specific to the individual and resolve some long-standing issues regarding the effi-
constitutes his valence for that outcome (V^). Valence cacy of monetary, organizational, and psychological
is positive if the outcome is desirable, negative if outcomes.
the outcome is undesirable, and zero if one is indiffer- Of special import to the present investigation is
ent toward the outcome. One's perception of the the fact that no previous study has attempted to
degree to which performance at level / will result in validate expectancy theory on a sample of salesmen
or block the attainment of outcome k is termed the despite the great emphasis placed on motivation as
instrumentality^ of performance level j for outcome a determinant of performance in the sales management
k (7j^). Instrumentality is positive if performance re- literature. The purpose of the study reported here
sults in attainment of outcome fc, negative if it blocks is to provide such evidence.
attainment of outcome k, and zero if it has no effect
on attainment of outcome k. METHOD
The theory posits that the valence of a performance
level (Vj) is a function of the multiplicative interaction Subjects
of the valence of the fcth outcome and the instru- Subjects were all full-time life insurance agents of
mentality that performance level / will result in out- a medium-sized Midwestern life insurance company.
come k, summed over all (n) outcomes. Thus: A questionnaire tapping motivational perceptions was
sent to 99 full-time agents in late March, 1972. In
(5) V, = all, 95 (96%) of the full-time field responded to the
request. Of the 95 respondents, three were eliminated
Evidence as to the predictive validity of the theory because of incomplete and unusable questionnaires
in industrial settings is mounting [23, 32], although and three were terminated before the performance
the performance variance explained, while significant, period expired.
is low. Generally, performance measures are regressed In addition, nine "agency building" agent-managers
on the valences and instrumentalities of a priori were eliminated from the study because they were
determined outcomes, expectancy, and ability in addi- not sufficiently involved in production activities to
tive and/or interactive form. Tlie motivation and have meaningful performance goals. This resulted in
performance valence constructs are usually not mea- a final sample size of 80. The average subject was
sured directly because of operational difficulties. 38 years old, had been with the company 8 years,
While self-reported effort is sometimes used as a proxy and had 2 years of education at the college level.
for motivation, this variable can be seen as completely
determined by the antecedent constructs. Design
It should be noted here that the theory does not A cross-sectional nonexperimental design was em-
specify the content of the motivationally relevant ployed. Because of time and resource constraints,
outcome set. It has been shown elsewhere, particularly performance data were collected over the six-month
with regard to the formation of attitudes and also period of January through June, 1972. The question-
in simulated choice behavior, that only a small subset naire was administered midway through this time span.
of all possible affective and belief cognitions about While an annual performance period may have been
an object are "salient" [6, 19, 37]. This issue takes more meaningful in terms of reliability, analysis of
on added significance given the large number of the previous years production figures showed that
"motivators" suggested in the sales management liter- production for the first six months correlated .936
ature. The motivating qualities of various compensa- with the annual data.
tion plans, incentive type "stimulators" (e.g., sales
contests, conventions, production clubs) [20], and Measures
higher order psychological needs [47] have, at one Dependent variables. Two criteria were investigated.
time or another, been afforded prominent status in The first, six-month production volume, was used
the motivational literature. While at least one writer because of the pervasive nature of absolute volume
has argued that little evidence exists to support the as (he criterion in the insurance industry. The produc-
alleged potency of any one set [8], investigations of tivity of agents, agencies, companies, and the industry
itself is measured in terms of this factor. Production
data were supplied by the company.
^ Similarities between the instrumentality construct as used In an effort to overcome the problems introduced
here and as used in attitude models attributed to Rosenberg when comparing the performance of salesmen in
[37] and Fishbein [12] and also the possible isomorphism different territorial environments (for discussion, see
between expectancy and instrumentality-value theories have [18]), a second relative criterion was used. Asa matter
been elaborated by Jacoby [25]. of company policy, agents were encouraged to estab-
246 JOURNAL OF MARKETING RESEARCH, AUGUST 1974

Table 1 was based on three annual performance criteria.


JOB OUTCOMES Agents selling one million dollars in paid annual volume
were admitted to the "$1 Million" club, while those
Making the (Lives) club placing life insurance on 100 people in a calendar
Receiving more responsibility in my position year were awarded with membership in the "Lives"
Getting improvements in home office practices and proce- club. The top 16 agents yielding the greatest paid
dures annualized premium (a profitability measure) were
Having better working relations with my supervisor admitted to the "Premium" club. Finally, the conven-
Furthering my professional growth and development
Receiving more recognition and appreciation for my produc- tion qualifications were based on a combination of
tion efforts volume and commission. As has been noted, a test
Making the (Premium) club of the separate effects of this set should, in part,
Getting better agency office facilities resolve some long-standing issues concerning the be-
Feeling more secure in my job havioral implications of sales incentives over and
Feeling a greater sense of accomplishment from the work
I'm doing above the effect of the basic compensation plan.
Advancing within the company to a (higher) field management Intrinsic outcomes. Lawler [27] has drawn a dis-
position tinction between intrinsic and extrinsic outcomes.
Making greater use of my skills and abilities on my job
Having my family and friends view my job as having greater Intrinsic outcomes are self-bestowed and are thus more
prestige available to the individual while extrinsic outcomes
Getting better supervision must be mediated through another entity. In addition,
Feeling greater self-esteem from my job intrinsic outcomes are of higher order in terms of
Going to the company convention Maslow's [30] paradigm and conform to what sales
Having better working relations with the agents in my agency
Receiving more income from my job management writers refer to as "psychological"
Making the ($1 Million) club needs. The five intrinsic outcomes used in this study
Feeling a greater sense of self-fulfillment from my job included furthering one's professional growth, receiv-
ing a sense of accomplishment from one's job, the
satisfaction of using one's skills and abilities on the
lish a production objective or goal for the calendar job, furthering one's self-esteem, and gaining a sense
year. As agent goals presumably incorporated a of self-fulfillment.
number of situational factors (e.g., market potential, Most desirable outcomes. Researchers in organiza-
intensity of competition) and individual factors (e.g., tional behavior [17], social psychology [37], and
tenure), they were used to provide a common basis marketing [6, 19] have suggested that only a small
for making interagent production comparisons. Thus, set of the most salient or important outcomes are
a goal attainment criterion was established in this study used by individuals in behavioral decisions and attitude
by dividing an agent's six-month volume by his pro- formation. This position assumes that a motivationally
duction objective. Goals were obtained from the salient outcome set is peculiar to the individual. The
questionnaire. outcomes used here included the three most desirable
Outcomes. Twenty job outcomes were selected so outcomes as selected by the subject in a special section
as to include compensation, incentive-type awards of the questionnaire.
peculiar to the research setting, and a number of other All job outcomes. In early writings on expectancy
job dimensions identified by Herzberg et al. [24] as theory, no distinction was made between the salience
having implications for motivation. These outcomes of various sets of outcomes. Rather it was assumed
were aggregated into sets so that competing interpreta- that all outcomes had some behavioral implications
tions of a "motivationally salient" set could be tested. because if any did not, they would receive an assigned
The various outcome sets are listed in Table 1. valence or instrumentality weight of zero. A separate
Compensation. In addition to the strong nonempiri- test of all 20 outcomes chosen for the study was
cal emphasis on pay in the sales management literature, included to provide evidence for this notion.
some evidence exists to suggest that pay is a significant Valence. Outcome valences were measured on a
predictor of productivity when tested alone in an scale from extremely undesirable (-3) through neither
expectancy theory format [26, 36,39,40]. The potency desirable nor undesirable (0) to extremely desirable
of pay as a predictor may arise because it can be (+3). This scoring method was used for all analyses
seen as instrumental in the attainment of many other except those in which the three most desirable out-
needs, job related or otherwise [28]. The outcome, comes were used. Because considerable restriction
receiving more income from my job, was used for of range was expected, a procedure used by Goodman,
this variable. Rose, and Furcon [17] was applied. The most desirable
Incentives. Four production incentives specific to outcome was assigned a valence weight of 3, the second
the research setting were investigated. These included ranked outcome was given a weight of 2, and the
membership in three production clubs and invitation third outcome was given a valence score of unity.
to the company's annual convention. Club membership Instrumentality. Instrumentality items were phrased
EXPECTANCY THEORY PREDICTIONS OF SALESMEN'S PERFORMANCE 247

in terms of subjective probabilities. Subjects were model and the incremental variance explained was
asked to scale their beliefs of the effect of attaining tested for significance. If the increment was signifi-
their production goal on the occurrence of each of cant, the interaction was concluded to have made a
the outcomes. The scale ranged from no effect (0) contribution to the explained variance above that of
to certain to occur (1). Intermediate values included the main effects.
low probability (.25), 50-50 chance (.5), and high
probability (.75).^ Hypotheses
Expectancy. For the expectancy measure, each The research hypotheses investigated were as fol-
subject was asked to indicate the chances in ten that lows:
his efforts would lead to the attainment of his produc- Hypothesis 1: Aggregate valence and instrumentali-
tion goal.
Ability. While the literature on the ability of sales- ty
men [15, 18] strongly suggests that both intelligence (a) P = a + p , 2 V . + P ^ S / j ,
and personality inventories are predictive of sales (b) P = a + p, 2 V, + p , 2 I,, + p3 2 VJ^,
success, the company unfortunately did not administer
Hypothesis 2: Performance valence and expectancy
standard versions of these tests to its agents. Only
a recently introduced "in-house" intelligence test was (a) P = a + p . i : V,7,.,+ P2E,.
available. To confound matters, test scores were (b) P = a+^,lV,I^,+ ^^Ei
obtained under both employee and applicant conditions
and were not available for roughly 10% of the sample.
As the company was not willing to have additional Hypothesis 3: Motivation and ability
tests administered to its field, the scores on the (a) P = a + p . E j S VJ^, + p, A
"in-house" test were used as a sole measure of ability. (b) P = a + p , E , . S V , r
The mean score was substituted for missing scores.
ANALYSIS RESULTS
While Vroom's [46] model is stated in terms of
multiplicative interactions between variables, re- Hypothesis 1
searchers [13, 23] have raised the issue of whether Simple correlations between aggregate valence,
the multiplicative model contributes any variance instrumentality, their interaction, and the performance
beyond that explained by the main effects in additive criteria are presented in Table 2. Aggregate valence
combination. In order to provide evidence relating figures for the most desirable outcome set are not
to this issue, the contribution of a multiplicative presented because of the artificial weighting scheme
interaction above that of the main effects was assessed used. Two-tailed tests were used for the aggregate
in terms of the incremental variance explained in the valence and instrumentality figures while one-tailed
criteria [7]. tests were used for the interaction term. Directionally
The criteria were first regressed on the components testing the interaction term but not its components
of the additive model under investigation and the derives from the theory. While both positive and
variance explained in each criterion was observed. negative valence scores may be related to a motivated
Next, the multiplicative interaction of components was state if the corresponding instrumentality term is of
introduced in the model as a third variable. The the same sign, the interaction term should be related
coefficient of determination yielded by this second to high motivation only if the valence-instrumentality
model was then compared to that obtained in the first product is positive. A negative product would be
associated with reduced motivation.
'Negative instrumentality responses (corresponding to a It is apparent from inspection of Table 2 that only
blocking effect) were not included in the study for two the incentive outcomes predicted the performance
reasons. First, it was felt that the inclusion of four corre- criteria. Aggregate valence, instrumentality, and the
sponding negative items would make the instrumentality interaction for this outcome set were highly correlated
section more burdensome and difficult to comprehend. with the volume criterion although only aggregate
Instrumentality is a rather abstract concept requiring condi- valence and the interaction were correlated with the
tional probabilistic judgments. In order to minimize the goal attainment criterion. While the instrumentality
possibility of subject frustration, efforts were made to term for the compensation outcome predicted the
facilitate completion of this section. Second, it was difficult volume criterion, the correlations between the interac-
to imagine goal attainment as having a blocking effect on tion term and both criteria were effectively zero. Thus,
any of the outcomes. Because of the autonomous nature
of the agents' job and high importance placed on production, it appears that only perceptions regarding the incentive
even peer relations were not expected to be adversely outcomes were effectively related to agent motivation
affected. Thus, it was hoped that the omission of negative in this study.
values would not truncate the range of probable responses. Results obtained from regressing the criteria on an
248 JOURNAL OF MARKETING RESEARCH, AUGUST 1974

Table 2
CORRELATIONS BETWEEN THE PERFORA(\ANCE CRITERIA AND AGGREGATE VALENCE, INSTRUMENTALITY, AND THEIR
INTERACTION FOR FIVE JOB OUTCOME SETS

Aggregate
Aggregate Aggregate valence times
valence instrumentality instrumentality
Goal Goal Goal
Outcome set Volume attainment Volume attainment Volume attainment
Compensation -.036 .068 .225" .112 .022 055
Incentive .433" .238" .429" .148 .480'' 201 °
Intrinsic -.150 -.070 .093 .070 -.105 - 032
Most desirable — — .153 .090 .172 102
All outcomes .043 -.018 .168 .070 .112 .048

"ps .01.
Table 3
RESULTS OBTAINED WHEN THE PERFORAAANCE CRITERIA WERE REGRESSED O N TWO AGGREGATE VALENCE AND
INSTRUMENTALITY MODELS (HYPOTHESIS 1)

Volume Goat attainment


Outcome
set Model R R^ p:R^ pAR^ R R^ pR^ pAR^
Compensation a .271 .073 n.s. .113 .013 n.s.
b .283 .080 n.s. n.s. .136 .019 n.s. n.s.
Incentive a .470 .220 .01 .239 .057 n.s.
b .484 .234 .01 n.s. .240 .058 n.s. n.s.
Intrinsic a .225 .051 n.s. .129 .016 n.s.
b .266 .071 n.s. n.s. .136 .018 n.s. n.s.
Most desirable" a .153 .023 n.s. .090 .008 n.s.
b .185 .034 n.s. n.s. .113 .013 n.s. n.s.
All outcomes a .194 .038 n.s. .114 .013 n.s.
b .201 .041 n.s. n.s. .144 .021 n.s. n.s.
"2 V,, term omitted.

additive combination of (a) aggregate valence and unique contribution of the individual terms, a phenom-
instrumentality and (b) aggregate valence, instru- enon also observed in a study of attitude formation
mentality, and their multiplicative interaction are [42].
shown in Table 3. The results show that, when using This result could be explained by cognitive consis-
the incentive set to predict the volume criterion, a tency principles. As predicted by balance theory [22],
significant multiple correlation coefficient was ob- the agents may have oriented their instrumentality
tained although it was not significantly higher than
the simple correlation obtained between the criteria
and the components taken separately. However, no Table 4
other outcome set predicted this criterion nor was CORRELATIONS BETWEEN AGGREGATE VALENCE,
any outcome set related to the goal attainment criteri- INSTRUMENTALITY, AND THEIR INTERACTION
on. In addition, the contribution of the interaction
term over and above that of its components was Aggregate Aggregate
nonsignificant in every test for both criteria. Aggregate valence with instrumentality
valence with the with the
The pattern of results above can best be explained Outcome set instrumentality interaction interaction
by the correlations between the aggregate terms pre- Compensation .441 .847 .665
sented in Table 4. Valence and instrumentality were Incentive .685 .902 .822
highly correlated as were the component terms with Intrinsic .402 .878 .714
their interactions. It appears that the high multicollin- Most desirable .978
earity of the predictor variables obscured any possible All outcomes .677 .885 .811
EXPECTANCY THEORY PREDICTIONS OF SALESMEN'S PERFORMANCE 249

perceptions so as to coincide with valence perceptions and attitudes to bring them in accord with their
or vice versa. On the one hand, behavior may be behavior [3].
seen as instrumental only for valent outcomes. That As a corollary to the relationships suggested above,
is, no thought may be given to unimportant outcomes a third hypothesis may be tenable. Perhaps the valence
whether performance is instrumental for their attain- -^ instrumentality relation reflects the state of nature
ment or not. As an alternative hypothesis, outcomes for some outcomes, while the instrumentality -* va-
may be desirable only to the extent that they are lence relation best models reality for other outcomes.
attainable. That is, only a dreamer would desire Based on the results presented here, these hypotheses
something wholly outside of his grasp. This argument are worthy of investigation.
is consistent with recent thinking in attitude theory
where it is believed that people change their beliefs Hypothesis 2
Simple correlations between the expectancy variable
and the performance criteria are presented in Table
5. Significant correlations were obtained for both
Table 5 criteria although the magnitude of the goal attainment
CORRELATIONS BETWEEN EXPECTANCY, ABIUTY, AND relationship was higher. Clearly performance was a
THE PERFORMANCE CRITERIA function of expectancy perceptions.
The correlations between the motivation term and
1 the performance criteria are presented in Table 6.
1. Volume .262° .065 One-tailed tests were assumed. Both the incentive and
2. Goal attainment .309» .029 most desirable outcome sets predicted the criteria
3. Expectancy — .087 significantly. However, with the exception of the
4. Ability
relation between the volume criterion and the motiva-
°p < .01, one-tailed test. tion term attributable to the incentive set, it appears
that the significant results derived largely from the
Table 6 expectancy measure. Note further that the multiplica-
CORRELATIONS BETWEEN THE MOTIVATION VARIABLE tive combination of expectancy and performance va-
AND THE PERFORAAANCE CRITERIA lence for the incentive set did not appreciably increase
the correlation obtained with the performance valence
Goal term alone.
Outcome set Volume attainment The potency of the expectancy measure is reflected
Compensation .131 .171 in the pattern of results obtained when the criteria
Incentive .502" .259=" were regressed on an additive function of performance
Intrinsic -.025 .060 level valence, expectancy, and their interaction
Most desirable .269" .245" reported in Table 7. Contrary to the earlier findings,
All outcomes .174 .135
the goal attainment criterion was significantly predict-
.05. ed for all outcome sets. However, the multiple cor-
.01.

Table 7
RESULTS OBTAINED WHEN THE PERFORMANCE CRITERIA WERE REGRESSED O N TWO EXPECTANCY AND
PERFORAAANCE VALENCE MODELS (HYPOTHESIS 2)

Volume Goal attainment


Outcome
set Model
Compensation a .262 .069 n.s. 311 .096 .05
b .313 .098 .05 n.s. 344 .119 .05

Incentive a .498 .248 .01 331 .109 .05


b .503 .253 .01 341 .117 .05

Intrinsic a .290 .084 .05 314 .099 .05


b .294 .086 n.s. 322 .104 .05

Most desirable a .290 .084 .05 312 .097 .05


b .291 .085 n.s. 314 .099 .05

All outcomes a .269 .072 n.s. 309 .096 .05


b .276 .076 n.s. 330 .109 .05
250 JOURNAL OF MARKETING RESEARCH, AUGUST 1974

relation coefficients were not significantly larger than pooling test scores obtained under both conditions
the simple correlation between expectancy and the as was done in this study may also have contributed to
goal criterioij. Because the correlation between the the low correlation between ability and performance.
volume criterion and expectancy was of a lower In view of these problems, the reader is cautioned
magnitude, the results using the volume criterion were not to interpret the findings reported here as discon-
attenuated for all outcome sets except the incentive firming evidence for the role played by ability as a
group where the coefficient of determination was determinant of performance. Rather, these results
highly significant. Almost 25% of the volume criterion should further underscore the necessity of using reli-
variance was explained when expectancy and the able and valid instruments in a conscientious selection
S V^/,fc term were regressed in additive combination. program.
When the interaction between expectancy and per- In summary, a "motivationally salient" outcome
formance valence was introduced in the regression, set was identified using expectancy theory constructs.
its contribution to the explained variance was nonsig- Valence and instrumentality perceptions of the four
nificant for all outcome sets using both criteria. In incentive-type outcomes used by the company under
fact, addition of this term dropped the multiple cor- study were strongly related to productivity and, to
relation coefficient below the .05 significance level a lesser extent, goal attainment. No other outcome
for two of the outcome sets when regressed on the set was consistently related to the criteria. While the
volume criterion. On the other hand, it increased the general expectancy measure was also correlated with
predictive power of the compensation constructs al- both performance criteria, the ability measure was
though this increase was not significant. not. When the volume performance criterion was
The reason for the nonsignificant increase in predict- regressed on an additive combination of expectancy
ability when the interaction term was introduced can and the S V^I^^ term for the incentive set, almost
again be attributed to the inherent multicollinearity 25% of the variance in performance was explained.
between interaction terms and their components. Both Given the paucity of empirical evidence on the role
the expectancy measure and the performance valence of motivation in salesmen's performance, these results
term were significantly correlated with their interac- will hopefully encourage further research in this area.
tion. Thus the additive (a) models appear to be at
least as predictive as higher order models incorporating DISCUSSION
interaction terms. The general predictive superiority of motivational
perceptions attributed to the incentive outcome set
Hypothesis 3 over those of other sets has significant implications
Simple correlations between the ability measure and for sales management. First, this result shows that
the performance criteria are shown in Table 5. The expectancy theory may be useful in identifying which
statistics show that ability as measured in this study of many possible outcomes in a sales situation are
was not correlated with the criteria. In fact, the ability effective motivators and which are unrelated to per-
measure made no further predictive contribution when formance. This information will permit the sales man-
combined either additively or multiplicatively with the ager to shift the resources at his disposal to the most
other variables. For this reason, further discussion productive incentive programs and to improve or phase
of the findings is not presented here. out those with little effect.
The lack of correlation between ability and perfor- Second, the finding that various sales "stimulators"
mance should be viewed only in the context of this were, in fact, related to productivity supports in part
study, however. As the reader may recall, three what had been heretofore accepted as "lore" in the
problems were evident with the available scores for sales management literature. While this result might
the "in-house" intelligence test used for the ability be attributed to spurious association in that a number
measure. First, the test was of uncertain origin; hence of outcome sets were tested against the criteria,
its reliability and validity were essentially unknown. incentive outcomes are generally recognized as having
Second, the technique of substituting mean predictor motivational implications. The great frequency with
scores for missing scores was used. Generally, this which these devices are used [20] attests to the
will attenuate the magnitude of a correlation unless popularity of this belief.
the corresponding criterion values are also at the mean. The poor results obtained using the other outcome
The third problem concerns using scores from tests sets are noteworthy. In particular, the valence times
administered under both employee and applicant instrumentality product for the compensation outcome
conditions. Rothe [38] has shown that the mean of did not yield significant results, although instrumental-
a test distribution of applicant scores tended to be ity alone was significantly correlated with the volume
higher than the mean of scores for employees who criterion. While this finding would appear to dispute
were asked to take the test for validation purposes the central postulate of many theories of sales motiva-
and attributed this to a higher level of "test-taking tion, it could be attributed solely to a restriction of
incentivation" on the part of the applicants. Thus, range in the valence of pay measure. However, the
EXPECTANCY THEORY PREDICTIONS OF SALESMEN'S PERFORMANCE 251

Standard deviation of this variable (LIO) was not performance and can be loosely classified as those
significantly smaller than the mean standard deviation factors which influence behavioral norms (i.e., what
of all 20 outcomes (L32) or of the 4 incentive outcomes one should do in a situation). Jacoby hypothesizes
(1.25). Thus it appears that the desirability of furthering that individuals can respond to questions concerning
one's income was, in fact, randomly related to both the "significance" of a particular input and also the
performance criteria in this study. "likelihood that that input will influence selection"
The low validities attributed to the most desirable of a behavioral alternative. Jacoby's revised model
outcome set may have arisen from the possibility that can be stated as:
outcome desirability does not necessarily connote
outcome instrumentality although, as Table 4 shows,
there is a marked tendency for these constructs to
be correlated. Conceivably, some outcomes could be
highly desired but be relatively unattainable. Expec-
tancy theory would predict that outcomes of this nature where:
would have little motivational impact.
M = motivation to perform (at level j),
The poor results obtained with the intrinsic outcome S( = significance of the Jth input,
set suggest that outcomes of this nature may be Ly = likelihood that input i will lead to selecting
motivationally dysfunctional. In particular, agents performance level j ,
more desirous of intrinsic rewards showed a slight V^ = valence of outcome k, and
tendency to be poorer performers. Thus, until further Ij^ = instrumentality of performance level / for
evidence is collected, it is suggested here that the attaining outcome k.
motivational implications of higher order psychological
needs have been overemphasized in the literature. Hopefully future tests of expectancy theory predic-
Finally, the notion that all job-related outcomes are tions will profit from incorporation of the two theoret-
motivationally salient was not supported. Rather, as ical refinements suggested here.
a number of writers [3, 12, 37] have suggested in Apart from basic extensions of the theoretical
other contexts, only a small subset of salient cognitions model, the relationships reported in this study may
were related to the sales behavior investigated in this have been attenuated by improper specification of
study. That is, while an individual may harbor many the combinatorial mode between variables. Expec-
beliefs about the effect of his performance on various tancy theory assumes that an individual mentally
outcomes and be able to elicit them when asked, he multiplies valence and instrumentality to arrive at a
may only act on a limited number. motivational force deriving from his desire for or
aversion to each outcome. Yet no evidence exists
Conceptual Refinements on how individuals actually process valence and
A number of conceptual modifications may help instrumentality perceptions. Conceivably this process
to improve upon the performance variance explained may be linear as opposed to multiplicative or may
in this study. Campbell et al. [5] and Jacoby [25] be of some other interactive form. A growing body
have suggested promising theoretical refinements to of research is emerging in the area of clinical judgment
the basic expectancy model. Campbell et al. posit [16] that may provide insights to this issue.
a hierarchy of outcome levels and suggest that per-
formance is only instrumental for some (lower level) Methodological Refinement
outcomes (e.g., pay). Higher level outcomes are seen Bass and Wilkie [2] and Wilkie and Pessemier [48]
as contingent not on performance, but on the attain- have recently drawn attention to a potential problem
ment of lower level outcomes. Thus the motivational that may arise when cross-sectional regression is
force attributed to a higher level outcome is a function performed on multi-attribute attitude models. Unless
of its desirability, the probability that lower level the very strong assumption of homogeneity across
outcomes will result in attainment of the higher need, respondents is met, the results obtained in tests of
and the desirabilities and instrumentalities of the lower such models may be attenuated because various seg-
level outcomes. While this approach may be a more ments of the sample perceive different rankings of
accurate representation of the relationship between attribute importance [41] or display divergent re-
outcomes residing at various levels and performance, sponse sets [2]. These problems can be solved by
problems may be encountered in specifying the partic- performing an intra-individual analysis if independent
ular need hierarchy for each individual. variable data are collected on more than one criterion
Jacoby [25] has distinguished between the motiva- (e.g., preferences for competing brands). Criterion
tional force attributed to outcomes of behavioral acts rankings can then be compared to rankings predicted
and that attributed to antecendent (input) values. Input by the attitude model for each individual taken
values are those which "push" an individual to per- separately. Bass and Wilkie have shown that the use
form at a certain level apart from the outcomes of of intra-individual analysis improved upon the number
252 JOURNAL OF MARKETING RESEARCH, AUGUST 1974

of correct preference predictions from a cross-sec- promising theories can be evaluated against the re-
tional analysis of the same data by 15% [2, p. 267]. maining variance attributed to deterministic processes.
Although the data collected in this study did not
permit intra-individual analysis in that the motivational
perceptions of various performance levels were not REFERENCES
investigated, small samples of sales personnel from
the same company selling to a regional market such 1. Bass, Frank M. "The Theory of Stochastic Preference
as that used in this study may be "homogeneous and Brand Switching," Institute Paper No. 415, Institute
for Research in the Behavioral, Economic, and Manage-
enough" for cross-sectional analysis [48] when ment Sciences, Krannert Graduate School of Industrial
compared to national consumer samples used in many Administration, Purdue University, 1973.
studies of brand preference. However, the possibility 2. and William L. Wilkie. "AComparative Analysis
remains that the results obtained here may have been of Attitudinal Predictions of Brand Preference," Journal
attenuated by regressing across heterogeneous re- of Marketing Research, 10 (August 1973), 262-9.
spondents. To overcome this problem infuture studies, 3. Bem, Daryl J. Beliefs, Attitudes, and Human Affairs.
researchers are advised to investigate the motivational Belmont, Calif.: Brooks/Cole, 1970.
perceptions attributable to various performance levels 4. Burton, William L. "There's More to Motivating Sales-
rather than focusing on one (e.g., goal attainment) men than Money," in Taylor W. Meloan and John M.
as was done here. For example, subjects could be Rathmeli, eds.. Selling: Its Broader Dimensions. New
York: Macmillan, 1960, 62-5.
asked to scale the instrumentalities of falling short 5. Campbell, John P., Marvin D. Dunnette, Edward E.
of, making, and exceeding their goal for each outcome. Lawler, III, and Karl E. Weick, Jr. Managerial Behavior,
The predicted performance level having the highest Performance, and Effectiveness. New York: McGraw-
2 Vf^Ij^ score could then be compared to the actual Hill, 1970.
performance level for each salesman taken separately. 6. Churchill, Gilbert A., Jr. "Linear Attitude Models: A
Study of Predictive Ability," Journal of Marketing
Research, 9 (November 1972), 423-6.
Sources of Unexplained Variance 7. Cohen, Jacob. "Multiple Regression as a General Data-
Analytic System," Psychological Bulletin, 70 (December
While the multiple correlation obtained when the 1968), 426-43.
volume criterion was regressed on expectancy theory 8. Cotham, James C , III. "Comments on: 'Sales Contests
constructs was highly significant, the theory as tested as a Motivating Device,' " in Thomas R. Wotruba and
here explained only 25% of the performance variance. Robert M. Olsen, eds.. Readings in Sales Management:
Three sources of unexplained variance were not tested Concepts and Viewpoints. New York: Holt, Rinehart
in this study that may have increased the coefficient and Winston, 1971, 297-300.
of determination measurably. The first encompasses 9. Cravens, David W., Robert B. Woodruff, and Joe C.
a number of situational variables including market Stamper. "An Analytical Approach for Evaluating Sales
potential, territory workload, and the company's effort Territory Performance," Journal of Marketing, 36 (Jan-
and experience [10, 11,35] that may appreciably affect uary 1972), 31-7.
a salesman's productivity. While an attempt has been 10. Davis, Kenneth R. and Frederick E. Webster, Jr. Sales
Force Management. New York: The Ronald Press, 1968.
made to include these variables in a multivariate model 11. Dodge, H. Robert. Field Sales Management: Text and
[9], the initial results have not been productive. Cases. Dallas: Business Publications, 1973.
Second, the potential contribution to the perfor- 12. Fishbein, Martin. "Attitude and the Prediction of Beha-
mance variance explained by pure measures of ability vior," in Martin Fishbein, ed.. Readings in Attitude
was not adequately tested. Had scores from a valid Theory and Measurement. New York: John Wiley and
selection battery been available, intelligence and per- Sons, 1967, 447-92.
sonality constructs may well have been predictive. 13. Gavin, James F. "Ability, Effort, and Role Perception
As suggested in the introduction to this article, stan- as Antecedents of Job Performance," The Experimental
dard tests of this nature have demonstrated validity Publication System, Issue 5, Manuscript 190A. Wash-
on sales occupations. ington, D.C.: American Psychological Association, 1970,
1-25.
Third, much of a salesman's decision to produce 14. Gaythwaite, Myles S. "Motivating Salesmen: Some
may be stochastic. Bass [1] has demonstrated that Ends and Means," in Albert Newgarden, ed.. The Field
brand switching may be a stochastic process. It is Sales Manager. New York: American Management
also conceivable that a salesman's performance may Association, 1960, 161-85.
not be entirely deterministic because of a "stochastic 15. Ghiselli, Edwin E. "The Validity of Aptitude Tests in
element in the brain" which influences his production Personnel Selection," Personnel Psychology, 26 (Winter
decisions. This may be especially true for sales oc- 1973), 461-77.
cupations where a high degree of autonomy is present. 16. Goldberg, Lewis R. "Simple Models or Simple Pro-
cesses? Some Research on Clinical Judgments," Ameri-
Future researchers may wish to provide an estimate can Psychologist, 23 (July 1968), 483-96.
of the performance variance remaining after stochastic 17. Goodman, Paul S., Jerry H. Rose, and John E. Furcon.
processes have been accounted for so that potentially "Comparison of Motivational Antecedents of the Work
EXPECTANCY THEORY PREDICTIONS OF SALESMEN'S PERFORMANCE 253

Performance of Scientists and Engineers," Journal of tivational Antecedents," unpublished doctoral disserta-
Applied Psychology, 54 (December 1970), 491-5. tion. University of Wisconsin—Madison, 1973.
18. Guion, Robert M. Personnel Testing. New York: Mc- 34. Opsahl, Robert L. and Marvin D. Dunnette. "The Role
Graw-Hill, 1965. of Financial Compensation in Industrial Motivation,"
19. Hansen, Fleming. Consumer Choice Behavior: A Cogni- Psychological Bulletin, 66 (August 1966), 94-118.
tive Theory. New York: The Free Press, 1972. 35. Phelps, D. Maynard and J. Howard Westing. Marketing
20. Haring, Albert and Malcolm L. Morris. Prizes, Awards Management, third edition. Homewood, 111.: Richard
for Motivation. New York: Sales and Marketing Execu- D. Irwin, 1968.
tives-International, 1968. 36. Porter, Lyman W. and Edward E. Lawler, III. Manage-
21. Haring, Albert and Robert H. Myers. "Special Incen- rial Attitudes and Performance. Homewood, 111.: Richard
tives for Salesmen," Journal of Marketing, 18 (October D. Irwin, 1968.
1953), 155-9. 37. Rosenberg, Milton J. "Cognitive Structure and Attitudi-
22. Heider, Fritz. The Psychology of Interpersonal Rela- nal Effect," Journal of Abnormal and Social Psycholo-
tions. New York: John Wiley and Sons, 1958. gy, 53 (November 1956), 367-72.
23. Heneman, Herbert H., Ill and Donald P. Schwab. 38. Rothe, Harold F. "Distributions of Test Scores of
"Evaluation of Research on Expectancy Theory Predic- Industrial Employees and Applicants," Journal of Ap-
tions of Employee Performance," Psychological Bulle- plied Psychology, 31 (October 1947), 480-3.
tin, 78 (July 1972), 1-9. 39. Schuster, Jay R., Barbara Clark, and Miles Rogers.
24. Herzberg, Frederick, Bernard Mausner, and Barbara "Testing Portions of the Porter and Lawler Model
Block Snyderman. The Motivation to Work, second Regarding the Motivational Role of Pay," Journal of
edition. New York: John Wiley and Sons, 1959. Applied Psychology, 55 (June 1971), 187-95.
25. Jacoby, Jacob. "Consumer and Industrial Psychology: 40. Schwab, Donald P. and Lee D. Dyer. "The Motivational
Prospects for Construct Validation and Mutual Contri- Impact of a Compensation System on Employee Per-
bution," in Marvin D. Dunnette, ed.. The Handbook formance," Organizational Behavior and Human Per-
of Industrial and Organizational Psychology. Chicago: formance, 9 (April 1973), 215-25.
Rand McNally, forthcoming. 41. Scott, Jerome E. and Peter D. Bennett. "Cognitive
26. Lawler, Edward E., III. "Ability as a Moderator of Models of Attitude Structure: 'Value Importance' is
the Relationship Between Job Attitudes and Job Per- Important," Proceedings. American Marketing Associa-
formance," Personnel Psychology, 19 (Summer 1966), tion, 1971, 346-50.
153-64. 42. Sheth, Jagdish N. andW. Wayne Talarzyk. "Perceived
27. . "Job Attitudes and Employee Motivation: Instrumentality and Value Importance as Determinants
Theory, Research, and Practice," Personnel Psychology, of Attitudes," Journal of Marketing Research, 9 (Febru-
23 (Summer 1970), 223-37. ary 1972), 6-9.
28. Pay and Organizational Effectiveness: A Psy- 43. Stanton, William J. and Richard H. Buskirk. Manage-
chological View. New York: McGraw-Hill, 1971. ment of the Sales Force, third edition. Homewood, 111.:
29. Lewin, Kurt. The Conceptual Representation and the Richard D. Irwin, 1969.
Measurement of Psychological Forces. Durham, N.C.: 44. Still, Richard R. and Edward W. Cundiff. Sales Man-
Duke University Press, 1938. agement, second edition. Englewood Cliffs, N.J.: Pren-
30. Maslow, A. H. "A Theory of Human Motivation," tice-Hall, 1969.
Psychological Review, 50 (July 1943), 370-96. 45. Tolman, Edward C. Purposive Behavior in Animals and
31. McDermid, Charles D. "How Money Motivates Men" Men. New York: Century, 1932.
in Robert F. Gwinner and Edward M. Smith, eds.. Sales 46. Vroom, Victor H. Work and Motivation. New York:
Strategy: Cases and Readings. New York: Appleton- John Wiley and Sons, 1964.
Century-Crofts, 1969, 321-32. 47. Whitmore, E. "What Really Motivates Salesmen," Sales
32. Mitchell, Terence R. and Anthony Biglan. "Instru- Management, (October 7, 1960), 115-8.
mentality Theories: Current Uses in Psychology," Psy- 48. Wilkie, William L. and Edgar A. Pessemier. "Issues
chological Bulletin, 76 (December 1971), 432-54. in Marketing's Use of Multi-Attribute Attitude Models,"
33. Oliver, Richard L. "Expectancy Theory Predictions of Journal of Marketing Research, 10 (November 1973),
Salesmen's Performance and Determinants of the Mo- 428-41.

You might also like