You are on page 1of 25

Herzberg’s two-factor theory of job

attitudes: a critical evaluation and some


fresh evidence
by Toby 0. Wall and Geoffrey M. Stephenson
Department of Psychology, University of Nottingham

This evaluation of Herzberg‘s theory of motivation discusses its ambiguities and


the influence of the tendency for people to give ”socially desirable” answers.
Evidence from the authors’ experiments suggests that the theory is untenable as a
description of the structure of job attitudes and the determinants of job satisfaction
and dissatisfaction.The authors acknowledge that a policy of job enrichment derived
from the application of the theory would be likely to promote satisfaction and allay
dissatisfaction. A concise summary of their analysis and conclusions is given at the
end of the article.
OUTLINE OF THE PAPER tional variables may enhance or minimize any such
HERZBERG’S two-factor theory of job attitudes has tendencies.
been both influential and controversial. The aim Part 7 describes two contrasting studies carried
of this paper is to evaluate the two-factor theory, out by the authors in which a strong relationship
and to suggest some reasons for the controversy was found between “social desirability response
surrounding it. This controversy is due partly set” and responses consistent with the two-factor
to the ambiguity inherent in the theory; and partly theory.
to the respondents’ tendencies to give “socially Purt 8, the conclusion, discusses the practical and
desirable” answers-which has implications for theoretical implications of the findings describcd in
all job attitude research. Part 7.
The paper is divided into eight parts. Part I
describes briefly the aims of job attitude research. 1. THE AIMS OF JOB ATTITUDE RESEARCH
In Parts 2 and 3 the importance of Herzberg’s A person’s occupation governs many aspects of his
two-factor theory is discussed and the theory itself life: it determines his standard of living and his
is outlined. social status. Almost all men and most women are
Part 4 is a theoretical discussion of the two-factor members of the labour force at some time in their
theory. Some of the difficulties and ambiguities lives; indeed, most men spend one-sixth of their
inherent in the theory are considered in depth, and lives at work. Yet until relatively recently man’s
it becomes apparent that several different inter- relation to work had not been extensively studied.
pretations can be placed upon Herzberg’s theory. Job attitude research constitutes one approach
However, the interpretation which appears most to the study of the man/work relationship and is
justified is used in the subsequent evaluation. variously justified by different investigators. Three
In Part 5 studies designed to test the two-factor ostensible aims may be discerned : the “economic”,
theory are reviewed, and it is evident that conflicting “humanistic” and “theoretical” aims.
results abound. Economic aims were strongly represented in early
Part 6 offers an explanation of the two-factor industrial research, as exemplified by the work of
theory and of the conflict in research results in terms Taylor (191I)’ and Gilhreth (1911).~ Profitability
of respondents’ tendencies to give socially desirable
answers to job attitude questions. It is noted that ’ F. W.Taylor, Sritntijr Managtmtni, Harper, New York,
1911.
individuals vary in the extent to which they give F. B. Gilbreth, Morion Study, Van Nostnnd, New York,
socially desirable responses, and that several situa- 1911.

4’
Herzberg’s two-factor theory ofjob attitudes

and efficiency were their concern, as they were of satisfaction, and absenteeism and turnover (Fleish-
many early industrial psychologists such as man, Harris and Burtt, 1 9 5 5 ; ~Giese
~ and Ruter,
Miin~terberg~ who, in 1913 entitled his textbook 1949;” Harding and Bottenberg, 1961).12 Kerr,
Psyhology and Industrial Eficiency. Koppelmeir and Sullivan (1951),13 for example,
With respect to job attitude research the profit found that job satisfaction correlated -0.41 with
motive emerged in the tenet that a satisfied employee inexcused absenteeism. In skilled populations, in
will be a more productive employee-a proposition the long term, high absenteeism and turnover rates
for which modern writers of industrial relations have unfortunate economic implications. This is
texts have considerable affection (see, for example, less true in the short term for unskilled popula-
Tannenbaum, 19681.~It became the role of job tions.
attitude research to identify the sources of satis- The humanistic justification of job attitude
faction and dissatisfaction and, in the event of the research takes the not unreasonable view that as
latter, to provide remedies. Satisf‘v the workers, and society requires men to work so long and frequently,
they will work harder. Likert ( 1 9 6 1 ) ~McGregor
~ the least it can do is to make his time at work as
(1960)6and Argyris (1964)’ all make this assumption. pleasurable and as meaningful as possible. Viteles
Dunnette and Kirchner (1965, p. 1 7 1 ) ~have this (I953)I4 stressed the importance of this aim when
to say: “Likert, like Argyris and others, believes he wrote that, “the excellence of civilization is to be
that a modified theory of organization and manage- measured not only by its material achievements, but
ment is needed. Very simply, he believes that em- more by the life it affords emotionally for the average
ployee motivation is the key to overall organizational man”.
effectiveness and his theory strongly stresses tech- The fheoreticd aim of job attitude research is to
niques and methods which take better account of increase our understanding of man. Man at work
what people seek from their job and organizational provides a valuable source of evidence about his
environments.” The belief that a satisfied employee nature. Knowledge of the determinants of satisfac-
will also be a more productive employee is certainly tion and dissatisfaction at work can contribute
plausible, but it is not always substantiated by appreciably to motivation and personality theory
research. Brayfield and Crockett (19.55)~~for in particular, and to psychological theory in general.
example, in a review of the evidence, found little Whilst in principle the “economic1’, “humanistic”
justification for such an assumption. The validity and “theoretical” aims are separable, in practice
of the economic argument rests now on the fact they are not. As Kornhauser (1965)’~so bluntly
that there is a negative relationship between job says, “any single firm can work for the fulfilment of

’ H. Miinsterherg, Psychology and Industrial EBcicnry, l o E. A. Fleishman, E. F. Harris and H. E. Burtt, Leadership
Houghton Mifflin, Boston, 1913. and Supervision in industry, Columbus, Ohio State University,
‘ A. S. Tannenbaum, Social Psychology of the Work Organ- Bureau of Educational Research, 1955.
iration, Tavistock Puhlications, London, 1968. W. J. Giese and H. W. Ruter, “An objective analysis of
R. Likert, New Patterns of Manapcmmt, McGraw-Hill, morale”, Journal of Applied Psychology, 1949,33,421-427.
New York, 1961. ’ 2 F. D. Harding and R. A. Bottenberg, “Effect of personal

D. M. McGregor, Thc Human Side Cfhnttrprisc, McGraw- characteristics on relationships between attitudes and job
Hill, New York, 1960. performance”, Journal of Applied Psychology, 1961, 45,
’ C. Argyris, Integrating thc Indicidual and the Organization, 428-430.
Wiley, New York, 1964. 1 3 W. A. Kerr, G. Koppelmeir and J. J. Sullivan, “Absen-
M. D.Dunnette and W. K . Kirchner, Psychology Applied teeism, turno\*er, and morale in a metals fabrication factory”,
to Industry, Meredith Publishing Company, New York, Occupational Psychology, 1951,25, 50-55.
1965. M. S. Viteles, Motivation and Morale in Industry, 1953,
A. H. Brayfield and W. H. Crockett. “Employee attitudes Norton & Company, New York, 1953.
and employee performance”, Psychological Bulletin, 1955, 52, A. Kornhauser, Mental Health of the Industriaf Worker,
396-424. Wiley, New York, 1965.

42
non-economic values, including the well-being of its A theory which boasts such comprehensiveness,
employees, only to the extent that such objectives which has generated so much research and directly
are compatible with maintenance or increase of influenced industrial change, needs careful evalua-
overall efficiency” (p. 280). All the same, it is beconi- tion.
ing increasingly clear, as this article will argue,
that without an adequate theoretical framework, and 3. THE EMPIRICAL BASIS OF THE TWO-FACTOR
particularly without an adequate methodology, the THEORY
pursuit of the “economic” and “humanistic” aims Origins and central hypothesis
is an uncertain enterprise. The inspiration for the two-factor theory may be
found in a review of job attitude research in a book
by Herzberg, Mausner, Peterson and Capwell,
2. THE SIGNIFICANCE OF THE TWO-FACTOR
THEORY
published in 1957, entitled Job Attitudes :Research
and O p i n i ~ n . ’Commenting
~ on this book in 1966,
Herzberg’s motivation-hygiene theory of job at-
Herzberg writes: “The book was a saddening
titudes-the two-factor theory-adopts all three
experience, because the major conclusion, I felt,
aims (Herzberg, Mausner and Synderman, 1959).16
was that we could document almost any position
Essentially a theory of the structure of job attitudes,
one wishes to take with respect to what affected
Herzberg’s conclusions about the causes of satis-
people at work. It uas this feeling of hopelessness
faction and dissatisfaction may be used to increasr
that led to the formulation of a completely new
job satisfaction for humanistic reasons, and claims
approach to, and conceptualization of, job attitudes
are made that satisfaction so increased leads to
as reported in the follow-up book, The Motivation
improved performance at work.
to Work” (p. 148).~O
The impact of the two-factor theory on job
One finding did, however, emerge from the earlier
attitude research is unrivalled by any other approach.
publication, and was to form the basis of the two-
More hypotheses have been derived from this
factor theory. Herzberg ( 1959)2’ expressed it this
theory, and subsequeiitly tested, than from any
way: “The one dramatic finding that emerged in our
other theory, and it remains true of recent years
review of this literature was the fact that there was a
that the bulk of research on job attitudes has been
difference in the primacy of factors, depending
dominated by it. (See Smith and Cranny, 1968.)”
upon whether the investigator was looking for
This probably stems from its considerable appeal
for personnel managers and others actively engaged things the worker liked about his job or things he
in industry. In this country the theory has been disliked. T h e concept that there were some factors
increasingly a topic for debate in business circles that were ‘satisfiers’ and others that were ‘dis-
(for example, The Economist, 6 July 1970, and The satisfied was suggested by this finding. From it
Financial Times, 24 November 1969 and 17 April was derived one of the basic hypotheses of our 01%n
1970). Furthermore, suggestions purporting to be study” (p. 7).
derived from the two-factor theory have been put In this article we shall examine the central hypo-
into practice in the form of “job enrichment’’ thesis of the two-factor theory: that the perceived
(Paul, Robertson and Herzberg, 1969).’* determinants of job satisfaction are qualitatively
different from the perceived determinants of job
l b F. Herzberg, B. Mausner and B. Snyderman, The
dissatisfaction.
Motivation to Work, Wiley, New York, 1959.
I’ P. C. Smith and C. J. Cranny, “Psychology of men at l9 F. Herzberg, B. Mausner, R. 0. Peterson and D. F.
nork”, Aiznual Review oJ’Psychology, 1968, 19,467-496. Capwell, Job Attitudes : Reviem of Research and Opinion,
W. J. Paul, K. B. Robertson and F. Herzberg, “Job Pittsburgh: Psychological Service of Pittsburgh, 1957.
Enrichment pays off”, Harvard Business Reaicm, 1969 March/ l o F. Herzberg, B. Mausner and B. Snyderman, op. rit.
April, 61-79. Ibid.

43
Herzberg’s two-factor theory ofjob attitudes

Herzberg’s original study (1959)22 accomplishment in terms of consequent recognition


“Think of a time when you felt exceptionally good of his worth by the firm, and the factor of recognition
or exceptionally bad about your present job or any would be recorded at the second level. In the original
other job you may have had. . . . Tell us what study these first- and second-level determinants of
happened.” job satisfaction presented a similar picture. In
This instruction was given to 203 Pittsburgh 1966 Herzberg2‘ commented that the more objective
engineers and accountants, and their replies con- first-level factors should take precedence over the
stituted the data for Herzberg’s original study. second-level factors (p. 96). Studies testing the
Details of both “good” and “bad” times were two-factor theory have concentrated on first-level
obtained, and interviewers were instructed to put factors, and in practice the distinction between
additional open-ended questions to elicit reasons for first- and second-level factors is difficult to main-
their remembered attitudes. As many occasions, or tain. For these reasons, only first-level factors will
“sequences”, as respondents were able to recall were be considered here.
recorded. In all the 203 subjects offered a total of
476 sequences. The content of the interviews was Results
analyzed in an attempt to identify factors which the The results of Herzberg’s first study confirmed the
subjects believed had determined their attitude. basic hypothesis. Five first-level factors emerged
Dunnette and Kirchner (1965)~’comment on this as major determinants of job satisfaction, and were
method as follows: relatively unimportant as determinants of job dis-
“. . . a more fundamental contribution ofthe study satisfaction. These factors were achievement, advance-
is that the job factors so identified were allowed to ment, recognition, responsibility and work itself:
emerge from descriptions of actual job situations ,Henberg argued that these factors were all inrrinsir
rather than being based exclusively on responses incentives to the performance of work itself, deserv-
to checklists or sets of statements developed ahead ing of the title motivators. In contrast, the factors of
of time by the investigator. The job factors derived work conditions, supervision, inter-personal relations
by Herzberg’s classification are more likely, there- and company policy and administration did not
fore, to reflect things in the job environment leading emerge as major determinants of job satisfaction,
to employees’ approach and avoidance behaviours” but were strong determinants of job dissatisfaction.
(P. ‘52). Herzberg called these factors hygiene factors, as
At this point we must draw attention to Herz- they represented the work environment. Such factors,
berg’s distinction between first- and second-level he reasoned, were extrinsic to the performance of
factors. The idea is that a factor may emerge at two work itself.
levels. First-level factors are inferred from “ob- These results suggested to Herzberg that the deter-
jective” elements of the situation which are seen by minanis of j o b saiisfaction are separate and distinct
the subject as determining “good” or “bad” feelings. from the determinants of job dissatisfaction. The
At the second-level, on the other hand, factors are occurrence of, for example, advancement or recog-
deduced from the subject’s interpretation of a situa- nition may lead to satisfaction with one’s job, but
tion, in response to the question, “What did these their absence does not lead to dissatisfaction.
events mean to you?”. Hence, a subject who cites Similarly, the quality of work itself may be a factor
successful accomplishment of a task as a source of in satisfaction, but not dissatisfaction. Conversely
satisfaction is credited with the achievement factor a hygiene factor, like poor company policy and
at the first level. He may proceed to interpret his administration, is a potent source of dissatisfaction,
F. Herzberg, B. Mausner and B. Snyderman, op. cir. “ F . Herzberg, Work and the Nature of Man, World
*’M. D. Dunnett and W. K. Kirchner, op. cir. Publishing Company, New York, 1966.

44
but good company policy and administration is not in the system used for ranking factors on “im-
an important source of job satisfaction. Satisfaction portance”.
and dissatisfaction are thus envisaged as separable An alternative method of measuring the import-
dimensions, each determined by the operation of ance of factors would be to weight them according
quite distinct sets of factors. to the number of factors listed as determinants of a
This view contrasts sharply with the traditional given attitude. Thus a factor described as a sole
and commonsense view that any factor may contrib- determinant would score I, as one of two deter-
ute to an individual’s single “level” of satisfaction: minants would score 0.5, as one of three would
where “good” a factor will be satisfying, where score 0.33, and so on. It is not possible from Herz-
“bad” it will be dissatisfying. Moreover, this two- berg’s data to ascertain to what extent his failure to
factor theory invites comparison with the views of adopt some such weighting system biased his results
personality theorists like Maslow ( 1 9 4 3 ) ~whose
~ in a direction favourable or unfavourable to his
distinction between low-level “deficit” motives theory.
like hunger and bodily comfort, and higher-level
“growth” motives such as “self-actualization” The coding of factorsz8
parallels Herzberg’s distinction between “hygiene” House and Wigdor (1967)” suggest that Herzberg’s
and “motivating” factors. coding system is not completely determined by the
data, but in addition requires interpretation by the
4. THE AMBIGUITIES OF THE TWO-FACTOR coder. Vroom (1964)~”made a similar observation.
THEORY These investigators suggest that the way factors
Measuring the importance of factors were coded might reflect more the coder’s hypo-
The central hypothesis of the two-factor theory theses than the respondents’ own perceptions.
presupposes reliable measurement of the “import- Inadvertent bias may have crept in at both the
ance” of job factors. In practice, the importance of a interviewing and coding stages. More importantly,
factor was equated with the frequency of its occur- it is possible that the definitions of some of the factors
rence in sequences. Thus “responsibility” appeared led to their being necessarily associated with satis-
in 23 per cent of the sequences in relation to job faction or dissatisfaction. For example, it is quite
satisfaction and in only 6 per cent in relation to likely that the event of advancement will result in
job dissatisfaction. This method of analysis gives satisfaction, but “demotion” is rare and lack of
equal weight to every factor mentioned, irrespective advancement is a “non-event”. The coding system
of whether it was perceived as a partial or total is thus tautologous to some extent. Another illustra-
determinant of the job attitude reported. One tion is warranted. In a recent piece of research one
might, however, suggest that a factor viewed as one respondent gave lack ofresponsibility as a determinant
of several determinants of an attitude is of less of satisfaction. It is incongruent in this case to code
importance than a factor which appears as the sole the source of satisfaction as the factor “respon-
determinant of an attitude. It is conceivable, taking sibility”. On the other hand, one cannot code this
the example of responsibility (above), that in relation response as “responsibility” leading to dissatisfac-
to job satisfaction it occurred, on average, with tion-which is implied-because this is inadmissible
three other factors, whereas, in relation to job 26 For coding system see F. Herzberg et a!., The Motivation
dissatisfaction it may always have occurred by to Work, op. cit., pp. 44-49 and 143-146.
itself, This possibility should be acknowledged I’ R. J. House and L. A. Wigdor, “Herzberg’s dual-factor
theory of job satisfaction and motivation: a review of evidence
and a criticism”, Personnel Psychology, 1967, 20, 369-389.
I s A. H. Maslow. “A theory of human motivation”, Psyrho- V. H. Vroom, Work and Motivation, Wiley, New York,
logiral Review, 1943,50,37*396. 1964.

45
Herzberg’s two-factor theory of job attitudes

when the task is to code what are manifestly deter- Interpretation two
minants of satisfaction. The coding SyJtem presup Motivators: Occur more frequently in satis-
poses how eachfactor will occur, and for some persons fied as opposed to dissatisfied
these presuppositions will be incorrect. This will sequences
result in either the factor not being coded, or in Hygiene Factors: Occur more frequently in dis-
misrepresentation in the coding. satisfied as opposed to satisfied
sequences
The second interpretation invites us to envisage
What is t h e principal hypothesis o f t h e two-
factor theory? that each person has a certain number of motivators
to distribute between “satisfaction” and “dis-
The basic hypothesis states merely that motivators
satisfaction”. In what proportions will he distribute
are more important as determinants of job satis-
faction and less important as determinants of job them-and similarly, and quite separately, the
hygiene factors? In the second interpretation the
dissatisfaction, and that hygiene factors show the
relative importance of the two classes of factors does
reverse relationship. It is never clearly stated what
not become an issue. Each class is dealt with separ-
it is that each of these classes of factors is more
ately with respect to its distribution between the two
important than. Interpretations have frequently
kinds of sequence, and the prediction for one class
differed. Two restatements of the basic hypothesis
is the reverse of the prediction for the other.
are possible. Ifonefocuses upon the relative importance
of motivators and hygiene factors in (a) satisjied and
(6) dissatiJfied sequences, the prediction is that: Interpretations one and tmo compared. The first
interpretation focuses on sequences of a particular
motivators are more important than hygiene kind and makes predictions about the relative contri-
factors as determinants of job satisfaction and bution of motivators and hygiene factors within
hygiene factors are more important than motivators each. The tno interpretations are not equivalent.
as determinants of job dissatisfaction. This leads to: The results of a study by Hinrichs and Mischkind
(1967)~~ illustrate this point. Their results clearly
Interpretation one support interpretation two, concerning the distribu-
In satisfied sequences: motivators occur more tion of first, hygiene factors, and secondly,
frequently than hygiene motivators, between “positive” (satisfied) and
factors “negative” (dissatisfied) sequences. They report as
dissatisfied sequences : hygiene factors occur follows:
more frequently than “6)A peater proportion of hygiene factors arc mentioned
motivators as negative a.s opposed to p i t i v e muons. (p < o.ooor).
(b) A greater proportion of motivator factors are mentioned
If, on the other hand, one focuses upon the as positive as opposed to negative reasom (p < O.OOOI).”
distribution of (a) motivators and (6) hygiene factors
between satiy5ed and dissatiyied sequences, the The authors go on to say, however, that, “It
prediction is that: should be noted that hygiene factors account for
significantly greater than 50 per cent of mentions as
motivators are more important as determinants of either positive or negative reasons ( p pos. <0.005;
job satisfaction than they are as determinants of p neg. <O.OOOI).”Interpretation one is undoubtedly
job dissatisfaction and
hygiene factors are more important as determi- ~

19J.
~~

R. Hinrichs and L. A. Mischkind, “Empirical and


nants of job dissatisfaction than they are as deter- theoretical limitations of the twwfactor hypothesis of job
minants of job satisfaction. This leads to: satisfaction”, ~ownalofApplie~Psycholo~, 1967,gr,191-zoo.

46
TABLE 1 Diagrammatic representations of interpretations one and t w o (portraying hypothetical
data)

A. Interpretations one and two verified

Dissatisfaction Satisfaction
II

Absolute
frequency
, , I

300 250 200 150 100 50


, '4 00 50 100 150 200 250 300
, Absolute
frequency

B. Interpretation one verified: Interpretation two 'refuted'

Dissatisfaction Satisfaction
II

Absolute
frequency
, I I I I I Absolute
frequency
300 250 200 150 100 50 0 0 50 100 150 200 250 300
Satisfaction
BUT-
As discussed above: confirmation of
interpretation one implies confirmation
of interpretation two where the factors I I I I I

given as determinants of satisfaction and 75 60 45 30 15 15 30 45 60 75


dissatisfaction are calculated separately
on a percentage basis. Percentage frequency Percentage frequency

C. Interpretation two verified: Interpretation one refuted

Dissatisfaction Satisfaction
I I

Absolute
frequency
, I ,
300 250 200 150 100 50
I I rl
'1 , ,
00 50
I

100 150 200 250 300


Absolute
frequency

Hygiene factors

1 1 Motivators

47
Herzberg’s two-factor theory ofjob attitudes

refuted, for in both positive and negative sequcnces, hygiene factor will be more important to dis-
hygiene factory are more important than motirators. satisfaction than it is to satisfaction. This is related
Which interpretation is the more important ? to interpretation two of the two-factor theory. As
Interpretation two does not necessarily imply with interpretation two, this prediction about
interpretation one (except in the event of equal specific factors will be true, generally, where state-
numbers of motivators and hygiene factors being ment one is verified and the numbers of reasons for
mentioned), and interpretation one does not imply satisfaction and dissatisfaction are equated. However,
interpretation two, except ir. the event of an equal interpretation one will not follow, necessarily, from
number of factors being mentioned in both positive this third interpretation unless similar numbers of
and negative sequences. However, in the interview motivators and hygiene factors occur.
itself the spotlight was turned separately onto
positive and negative sequences, and the relative Interpetation four. This again concerns specific
number of hygiene factors and motivators was the factors. It is that each and every motivator is more
“dependent variable” of the investigation. If more important to satisfaction than any hygiene factor;
factors were elicited in one or the other instance, and that each hygiene factor is more important to
for whatever reason, allowance can be made for this dissatisfaction than any motivator (House and
difference by expressing the relative contribution of Wigdor, 1967;” Hinton, 1968).’~ This elaboration
motivators and hygiene factors to a sequence of statement one is unwarranted from Herzberg’s
(positive or negative) in terms of percentages. own work. It suggests that if the factors are ranked
When this reasonable procedure is followed, it in order of importance, separately for satisfaction
always happens that interpretation one implies and dissatisfaction, then, in the case of satisfaction
interpretation two. In fact Herzberg (1966)’~adopts the top positions will be monopolized by all the
interpretation one when evaluating relevant studies: motivators, whereas, in the case of dissatisfaction,
‘I. , . the frequency of the motivators during high all the hygiene factors will be at the top.
job attitude feelings and low job attitude feelings Let us take a practical example. Imagine that,
is contrasted with the frequency of the hygiene as in Herzberg’s original study, six motivators and
factors during these opposite states of affect’’ ten hygiene factors emerge from the data; further-
(p. 99). Herzberg’s analyses of supportive studies more that the motivators account for 40 per cent of
show that he equates frequency with “percentage the reasons for dissatiJfction and the hygiene factors
frequency”. Following Herzberg’s lead interpreta- account for 60 per cent of the reasons for dissatis-
tion one is adopted here as the most appropriate faction. With respect to interpretation one this
interpretation of two-factor theory. result would confirm the two-factor theory.
However, each motivator, if the six motivators
Interpretation three. A third interpretation is con- contributed equally to dissatisfaction, would have a
cerned with specific factors. (See Henberg, 1959, value of 40% + 6 = 6*7%, whilst each hygiene
p. 72;’’ and 1966,p.gg;’~andWhitsettand Winslow, factor would have a value of 60% i 10 = 6.0%.
1967.)” The prediction is that each and every A rank ordering of motivator and hygiene factors,
motivator will be more important to satisfaction in this example, would result in all the motivators
than it is to dissatisfaction, and that each and every being ranked as most important, and all the hygiene
factors as least important, to dissatkfaction.
F. Herzberg, Work and the Nature of Man, op. F i t .
I’ F. Herzberg ct al., The Motivation to Work, op. cit.
’* F. Herzberp, Work and the Nature of Man, op. rit. ’‘R. J. House and L. A. Wigdor, op. cit.
D. A. Whitsett and E. K. Winslow, “An analysisofstudies B. I,. Hinton, “An empirical investigationof the Herzberg
criticalofthe motivation-hySienetheory”,Personnel Psyrhology, methodologyand two-factor theory”, OrganierrtionalBehaziour
19679 20,391-415 and Human Pn/brmance,, 1968.3,286-309.

48
With respect to the rank ordering of specific in Whitsett and Winslow’s paper-Herzberg does
factors the two-factor theory does not make the not make very clear what he means by job satisfac-
predictions which interpretation four suggests. tion. He mentions the “perception of a peak or
Nevertheless, if interpretation one is verified, then valley in his feelings about his job” (1959,p. 23),
there will be a tendency for hygiene factors, as a class “overall feelings about his job” (1959, p. 23),
of factors, to be ranked less important in relation “feeling good” and “feeling bad” (1959,p. 22),
to satisfaction and more important to dissatisfaction “high” and “low” attitudes or feelings (1959,p.
than motivators. 6o), and so on. Many investigators interpret these
phrases as referring to “overall job satisfaction”
Other difficulties of the two-factor theory (Ewen, Smith, H u h and Locke, 1966,)~ for
Two papers (House and Wigdor, 1967;)~ Whitsett example). Yet Whitsett and Winslow (1967)state
and Winslow, 1967)” published independently that “The theory does not, and purposely does not,
of each other and in juxtaposition, demonstrate make statements about overall job satisfaction”
the differing interpretations that can be placed (Pa 395).
upon the two-factor theory and upon research What in fact is meant by “overall job satisfaction” ?
designed to evaluate the two-factor theory. Both We shall take it as meaning an individual’s expressed
papers reviewed the theory and the research it attitude towards his work as a mholc with respect
has engendered and came to opposite conclusions to his like OY dislike of it. It is taken as a global
concerning its usefulness. Yet both papers reviewed assessment of like or dislike of a working situation.
essentially the same evidence. Operationally overall job satisfaction might be
House and Wigdor offer criticisms of the meth- measured on an item such as that used by Hinrichs
odology and coding procedure used in Herzberg’s and Mischkind (1967):39 “Now, considering every-
original study. In addition, they cite several pieces thing, how would you rate your present satisfac-
of research which failed to support the theory. tion as an XYZ employee? Completely Satisfied
Whitsett and Winslow, concluding that the two- o I 2 3 4 5 6 7 8 g 10CompletelyDissatisfied.”
factor theory is useful, point to the limitations and Against Whitsett and Winslow’s assertion, Herz-
inadequacies of research which is critical of the berg ct al. in their original study reveal that overall
theory. They make the point that a principal job satisfaction is the target. This is particularly
measure of a theory is the extent to which it generates true in relation to “long-range sequences of events”
research. where each respondent is asked to recall “a period
The controversy which surrounds the two-factor of time during which his overall feelings about his
theory, and which is exemplified by these two papers, job were consistently high or low, despite possible
is in part due to the ambiguities inherent in H e r s fluctuations of feeling or even minor inversions
berg’s exposition of his theory. We have already within these periods. For example, a worker who
considered the differing interpretations which may on the whole was quite unhappy in his job could
be placed upon the basic hypothesis, but other occasionally achieve real feelings of satisfaction at
ambiguities exist. For example, Herzberg persists in the completion of an assignment or from a passing
describing his results in terms of the uni-polarity word of praise. This, however, does not minimize
of two classes of factors-yet his results show, at the fact that he could report a long period of time
best, only a tendency towards this. A more con- as being one in which his basic fundamental attitude
servative statement would be more reasonable.
To take a second example-one which is apparent ” R. B. Ewen, P. C. Smith, C. L. H u h and E. A. Lockc,
“An empirical test of thc Herzbcrg two-factor theory",
36 R. J. House and L. A. Wisdor, op. cir. 3 d of AppriCd Psychoiogy, 1966, 50, HI-550.
I’ D. A. Whitsen and E. K. Winslow,op. cit. l9 J. R. Hinrichs and L. A. Mischkmd, op. tit.

49
4
Herzberg’s ttvo-factor theory of job attitudes

towards the job was deep dissatisfaction” (1959, twelve companies, and aged between 60 and 65.
p. The fact that Herzberg ez al. found that the These managers recalled periods of satisfaction
classes of reasons given for overall like or dislike of and dissatisfaction sampled from a life-time of
the job were different in the two cases does not mean work. Strong support was found for the two-factor
that they are not talking about overall satisfaction theory. Motivators appeared eight times as fre-
as it has been defined above. Whitsett and Winslow quently as hygiene factors in relation to qtisfaction,
erroneously concentrate on interpretation two- whilst hygiene factors were twice as frequently
“job attitudes must be looked at twice-once to see mentioned in relation to dissatisfaction as were
if the needs fulfilled by the hygiene factors are indeed motivators. Myers ( 1 9 6 4 )interviewed
~~ 50 scientists,
fulfilled, and again to see if the needs fulfilled by the55 engineers, 50 supervisors, 75 male technicians
motivator factors are met” (1967, p. 396).“ Yet and 52 female hourly-paid assemblers. T h e two-
these authors comment that: “It is evident that man factor theory was supported for all five groups of
lives his life as a unity and consequently combines Myer’s population, though differences in the
his feelings of satisfaction and dissatisfaction.” patterning of factors did occur. Myers made the
It should by now be evident that the two-factor interesting observation that “workers become dis-
theory is concerned with overall satisfaction as satisfied” when their “opportunities for meaningful
defined above. It is assumed that each respondent achievement are eliminated and they become
can recall periods during which his overall assessment sensitized to their environment and begin to
of his work situation is one of like or dislike. find fault”. This suggests that although the re-
ported determinants of dissatisfaction are hygiene
factors the cause of dissatisfaction lies in the motiv-
CAN HERZBERG‘S RESULTS BE VALIDATED?
ators.
Supportive studies
Herzberg ( 1 9 6 5 ) ~replicated
~ his original results
Interpretation one, that motivators are more on a population of Finnish supervisors. I n this
important than hygiene factors to job satisfaction study, as in his original study, he found that an
and hygiene factors are more important than analysis of the results in terms of age and education
motivators to job dissatisfaction, is taken as a valid “revealed no meaningful alterations in the findings”
interpretation of the two-factor theory. Interpreta- (1966, p. 102):~ This last finding supports Herz-
tion one implies interpretation two, in a case where berg’s claim that his theory is appropriate for all
the number of factors related separately to satis- working populations. Herzberg (1966) reports
faction and dissatisfaction is equated. Interpretation studies by Clegg, on 58 county argicultural adminis-
one is also taken as referring to two classes of factor, trators; Walt, on 50 professionally qualified women;
not to specific factors. Gendel, on I 19 housekeeping employees; and by
a. Srudies using the same methodology Anderson on 95 nurses and engineering service
employees. All these studies, using essentially the
Several studies, using Herzberg’s original meth- same methodology as Herzberg’s original study,
odology, have replicated his results. Saleh (1963, support his results. This is so for populations of
1 9 6 4 ) ~ ~interviewed
’~’ 85 managers, drawn from differing ages, sex, nationalities, occupational levels
and, by implication, social classes and salary levels.
°.F. Herzberg et al., The Motiiwtion to Work, op. cit.
D. A. Whitsett and E. K. Widow, op. cit.
41
‘* S. Saleh, Report of “Pre-retirement study”, in Henherg, M. S. Myers, “Who are your motivated workers?”,
1966. Hatward Buiness Rm’c~v,1964, Jan-Feb, 73-88.
4 3 S . Saleh and J. Otis, “Sources of job satisfaction and F. Henberg, “The motivation to work among Finnish
their affects on attitudes toward retirement”, in Henberg, supcrvisors”, Personnel Psychology, 1965, I 8,393-401.
1966. 46 F. Henberg, Work and the Nature of M a n , op. cit.
b. Other nrethodologirs Two studies considered satisfaction alone, and
Schwartz (1959)” used a questionnaire which was the prediction from the two-factor theory that
similar to Herzberg’s (1959) interview schedule, motivators would contribute more to satisfaction
on 373 third-level supervisors. He found that the than would hygiene factors. Halpern (1966)’’
factors most significantly related to satisfaction were obtained ratings Of four motivators, and four
achievement, and recognition of achievement (motiv- hygiene factors, and of overall job satisfaction
vators); company policy and administration (hygiene from 93 graduates. Motivators were more strongly
factor) was found to be the principal reason for related to satisfaction than hygiene factors.
dissatisfaction. Fine and Dickman (1962, reported in Herzberg
Schwartz, Jenusaitis and Stark (1963)‘~patterned 1966) tested and Cm~firmeda similar hypothesis
a questionnaire after the Herzbergian interview on a population of engineers, physicists, technicians,
schedule and administered it to I I I male supervisors secretaries and clerical e m p b ’ m . Unfortunately,
in electric and gas utility companies. In relation to neither of these two studies considered the im-
satisfaction they found that motivators appeared portance of these factors to dissatisfaction. It is
with seven times the frequency of hygiene factors; possible that motivators may also have been more
the hygiene factors favoured dissatisfaction, corn- important than hygiene factors to job dissatisfac-
pared to the motivators, by a ratio of 3 to 2. tion.
Hahn ( 1 9 5 9 ) ~analysed
~ questionnaires obtained Weissenberg and Gruenfeld (1968)’~later found,
from 800 officers in the U.S. Air Force. The major from the questionnaire responses of 92 civil service
Source of satisfaction was from factors similar to supervisors, that satisfaction with motivators
Herzberg’s categories of achievement and respon- awounted for 49 per cent of the variance of overall
sibility, whilst dissatisfaction resulted from super- satisfaction, whilst satisfaction with the hygiene
vision and job context factors which did not contri- factors aCcounted for only 36 Per cent. While
bute to job satisfaction. both hygiene factors and motivators contributed to ’
Friedlander and Walton (1964)sO proposed, on overall satisfaction, it Was predominantly motivators
the basis of the two-factor theory, that the reasons that did so. Once again, the relationship of motiv-
for remaining with a company would be motivators ators and hygiene factors to dissatisfaction was not
rather than hygiene factors and the reasons for rePofled. Studies by P ~ c z e on l 78 engineers, and
leaving would be predominantly hygiene factors by Gibson on 1,700 manufacturing company
rather than motivators. Eighty-two scientists and employees, are quoted by Herzbert? (I966Is3 as
engineers in an army research and development supporting interpretation one of the two-factor
laboratory were interviewed. Motivators were given theory.
in relation to hygiene factors in a ratio of almost The studies which support Herzberg have shown
3 to I as reasons for staying, whilst hygiene factors that his results can be generalized to different
occurred in a ratio of 4 to I as reasons for leaving. Populations. They have been confirmed across
populations
- *
which differ in occupation. class., sex.,
41 P. Schwartz, Attitudes of Aliddle M a n a ~ m n r tPersonnel, age and national characteristics. They tend, however,
American Institute for Research, Pittsburgh, 1959. to use methodologies similar to those of the original
H. A. Schwartz, E. Jenusaitis and M. A. Stark. “Moti-
vational factors among supervisors in the utility industry”,
Personnel Psychology, 1963, 16,45-53. s t G . Halpern, “Relative contributions of motivator and
49 C. Hahn, Dimensions of Job Satisfaction and Career hygiene factors to overall job satisfaction”, Journal ofApplied
Motivation, unpuhlished manuscript, 1959 (in House and Psychology, 1966,50, xg8-zoo.
Wigdor, 1967)23. P. Weissenberg and L. W. Gruenfeld, “Relationship
so F. Friedlander and E. Walton, “Positive and negative between job satisfaction and job involvement,” Journal o j
motivations toward work”, Adniinistrativr Science Quarterly, Applied Psychology, 1968,52,469-473.
1964,9,194-207. sJ F. Herzberg, Work and the Nature of Man, op. cit.
Herzberg’s two-factor theory ofjob attitudes

study. With respect to individual factors, Herz- “company policy and administration”, was selected
berg (1966,pp. gg and 124)” tests his theory using to represent hygiene factors. These two factors
interpretation three-a motivator is expected to be accounted for nearly 75 per cent of the variance
more strongly associated with satisfaction than dis- found in satisfaction scores. The authors observed
satisfaction, and a hygiene factor is expected to that “Motivators accounted for nearly three-and-
show the reverse relationship. Herzberg found one-half times as much variance (0.57) as did
(1966,pp. 124-125)~~ a 93 per cent correct con- hygiene (0.17), indicating that, when looking at
firmation of this hypothesis. However, in that inter- those classes of variables, workers view motivators
pretation three only implies interpretation one as the prime source of their satisfaction. This con-
where the number of motivator and hygiene factors clusion is further strengthened by the observation
are equated, and that in the studies reviewed motiv- that low, or even moderate levels of achievement
ators characteristically occurred in greater numbers lead to reported feelings of dissatisfaction, regardless
than hygiene factors, this is not a legitimate test of of the level of hygiene.” In other words, “workers
the two-factor theory. who do not feel as if they are accomplishing much
on the job will be dissatisfied, even though condi-
Non-supportive studies tions surrounding the job are good”; the motivator
Studies supportive of the two-factor theory generally “achievement” is therefore more of a satisfier and
have provided evidence which permitted a test of dissatisfier than the hygiene factor “company
interpretation one to be made. Studies critical of policy and administration”. However, Lindsay
the two-factor theory have, however, demonstrated et al.’s study only considers one motivator and one
greater diversity of interpretation. For this reason, hygiene factor: furthermore, these two factors
non-supportive studies will be considered in three differ in overall importance. The results of their
categories : (a) those studies providing evidence study can only be considered as non-supportive of
which test interpretation one; (b) those studies which interpretation one of the two-factor theory to the
test interpretation two; and (c) those studies which extent to which the two chosen factors are repre-
test other aspects of the two-factor theory, such as sentative of motivators and hygiene factors respec-
the validity of its method and its generality. tively.
Dunnette, Campbell and Hake1 (1967)~’factor
a. Studies which test interpretation one of the two- analysed job factors obtained from 133 store
factor theory managers, 44 secretaries, 129engineers and research
Lindsay, Marks and Gorlow (1967)~~ administered scientists, 89 sales clerks, 49 salesmen and gz
a questionnaire to 162 professional and 162 non- army reservists and night-school students. Three
professional employees of an aerospace research motivators and one hygiene factor emerged as the
and development company. Two hundred and most important determinants of satisfaction and
skventy of the 289 completed questionnaires were dissatisfaction. Whilst motivators were found to be
used in their statistical analyses, nineteen being more important than hygiene factors in relation
discarded at random for ease of statistical calculation. to satisfaction, they were also more important in
One job factor, “achievement”, was selected to relation to dissatisfaction.
represent motivators, and a second job factor, Hinrichs and Mischkind (1967)~~ administered
a questionnaire to 613 technicians. The question-
14 F.Hcrzbcrg, Work and the Nature of Man, op. tit. I’M. D. Dunnette, J. P. Cpmpbell and M. D. Hnkel,
Ibid. “Factors conmbuting to job satisfactionand job dissatisfaction
I6 C. A. Lindsay, E. Marks and L. Gorlow, “The Hvzbcrg in six occupationd groups”, Orgmkational Behaviour and
theory: a critique and reformulation”, Journal of Applied Human Pmfmnce, 1961,a, 143-174.
PWholoU, 1967,545 330-369. J. R. Hinrichs and L. A. Mischkind, op. tit.
naire was concerned with reasons for present satis- to 80 evening class students who were also full-
faction with a current job situation. Their results time employees. He found that motivators were
refute interpretation one of the two-factor theory the major determinants of dissatisfaction as well
whilst confirming interpretation two. With respect as satisfaction. Lahiri and Srivastva (1967)~~ had
to interpretation one, they found that “Hygiene 93 Indian middle-management employees rate
factors account for significantly greater than 50 per thirteen motivators and thirteen hygiene factors for
cent of the mentions as either positive or negative the importance of each to both satisfaction and dis-
reasons (p pos. <0.005; p neg. <0~0001).~’ It is satisfaction. This the respondents did for both their
interesting to note that Hinrichs and Mischkind present jobs and imaginary jobs. Imaginary jobs
found hygiene factors more important overall than were the best and worst jobs each respondent
motivators; other studies have consistently obtained could imagine himself doing. In all cases, for
the reverse result. satisfaction and dissatisfaction, the mean rating of
Wernimont (1966)~~ tested four hypotheses the motivators was higher than that for the hygiene
derived from the two-factor theory. The first and factors, but there was no significant difference
second of his hypotheses conformed to interpretation (with one exception-motivators were significantly
one whilst his third and fourth hypotheses paralleled greater for satisfaction in imaginary jobs).
interpretation two. Fifty accountants and 82 All the studies described above refute interpreta-
engineers were given a questionnaire which consisted tion one of the two-factor theory. Moreover, all
predominantly of forced-choice items, but also those studies used questionnaires for obtaining
contained a few free-choice items. On the forced- data and all (except one, Hinrichs and Mischkind,
choice items both accountants and engineers 1967)~~ refuted interpretation one because motivators
endorsed a higher proportion of motivators than were found to be more important to dissatisfaction
hygiene factors when describing past satisfying than were hygiene factors.
and past dissatisfying experiences. Thus interpreta- b. Stirdies which test interpretation two of the two-
tion one is refuted, for the major determinants of .factor theory
dissatisfaction were motivators rather than hygiene
factors. On the free-choice items similar results were Interpretation two focuses separately on motivators
and hygiene factors. It proposes that motivators
obtained, though they more nearly approximated to
those predicted by interpretation one. The motivators will be more significantly related to satisfaction
than they will be to dissatisfaction, and that hygiene
increased in importance over the hygiene factors as
reasons for satisfaction and were less dominant as factors will demonstrate the reverse relationship.
reasons for dissatisfaction. Wernimont did, however, It has already been argued that this second
interpretation is not, in itself, a legitimate statement
find some support for interpretation two of the two-
of the two-factor theory. However, it should be
factor theory. For accountants the motivators were
significantly more strongly related to satisfaction considered for two reasons. First, studies testing
than to dissatisfaction, whilst the hygiene factors interpretation two have been seen as tests of the
showed the reverse relationship. The results were two-factor theory, and have been accepted as
similar for the population of engineers, but did not evidence for or against the theory. Secondly, when
interpretation two isrejiited, and where the importance
reach an acceptable level of statistical significance.
of reasons for satisfaction and dissatisfaction are
Friedlander ( 1964)~~ administered questionnaires
equated, then no support for interpretation one is
~~

L9 P. F. Wernimont, “Intrinsic and extrinsic factors in job 6 1 D. K. Lahiri and S. Srivastva, “Determinants of satis-
satisfaction”, Journal qf Applied Prychofogy, 1966,50, 41-50. faction in middle-management personnel”, Journal of Applied
6o F. Friedlander, “Job characteristics as satisfiers and dis- Psychology, 51,3,~j4-~65.
satisfiers”, Journal of Applied Psychology, 1964, 48, 388-392. 6z J. R. Hinrichs and L. A. hlischkind, op. cir.

53
Herzberg’s two-factor theory ofjob attitudes

also implied. However, where interpretation two studies refute interpretation two. However, the
is confirmed this has no implications for the accuracy differential importance of all factors to satisfaction .
of interpretation one (except where the numbers of and dissatisfaction must influence these results, and
motivators and hygiene factors are equal-which is an interesting finding in itself. Why do respondents
in practice has not occurred). rate the positive aspect of factors as very important
Two studies offer evidence supporting interpreta- to their satisfaction yet rate the negative aspects of
tion two. Wernimont (1966)~~ and Hinrichs and the same factors as less important to their dis-
Mischkind (1967):~ whilst obtaining evidence satisfaction? Is it that they are less willing, or
critical of interpretation one, found support for less able to identify the sources of their dissatisfac-
interpretation two. Both these studies present their tion?
results in a form which equates the importance of Ewen, Smith, Hulin and Locke (1966)~’and
the factors for satisfaction and dissatisfaction. H u h and Smith (1967)~’place an interpretation
Both found that motivators were more strongly upon the two-factor theory which is an extreme
related to satisfaction than to dissatisfaction and that form of interpretation two. They propose that the
hygiene factors showed the reverse relationship. theory states that motivators will have no relation-
These results need qualifying, however. Wernimont ship to dissatisfaction, and that hygiene factors will
found support at a statistically acceptable level have no relationship to satisfaction. This interpre-
only for his population of accountants. For his tation is presumably derived from Herzberg’s
population of engineers the differences were in the theoretical discussion of the absolute unipolarity of
predicted directions, but did not reach significance. motivators, and of hygiene factors. Whilst this
Hinrichs and Mischkind found support for interpre- interpretation is legitimate from Henberg’s theo-
tation two with 455 respondents who were “high” retical discussion, it is clearly not supported even
on overall satisfaction ; however, I 58 respondents by his own evidence. Only a tendency towards uni-
who were h low^' on overall satisfaction found motiv- polarity was observed.
ators and hygiene factors almost equally important H u h and Smith (1967) found that the two
to satisfaction and dissatisfaction. motivators which they considered were rated more
Two further studies offer evidence critical of important to satisfaction than they were to dis-
interpretation two, whilst at the same time refuting satisfaction. T h e single hygiene factor these investi-
interpretation one (Lahiri and Srivastva, 1967;~’ gators chose was unfortunate; it was “pay”, which
Friedlander, 1964):~ Both these studies asked as Herzberg commented may be considered as
their respondents to rate factors for their importance both a motivator and a hygiene factor depending
to satisfaction and dissatisfaction separately; and upon its implications. It was found that pay was
both found that all factors were rated more im- also more strongly related to satisfaction than to
portant with respect to satisfaction than with respect dissatisfaction (as were all five of the factors Hulin
to dissatisfaction. Not surprisingly, they found and Smith considered). Ewen et al.69 considered
that, whilst motivators were more important to
satisfaction than they were to dissatisfaction, hygiene
6’ R. B. Ewen, et al., op. cit.
factors demonstrated the same relationship. In 6* C. L. Hulin and P. A. Smith, “An empirical investigation
that hygiene factors were rated as more important of two implications of the two-factor theory of job satis-
to satisfaction than to dissatisfaction, these two faction”, Journal of Applied Psychology, r967, 51,396-402.
69 Graen (1966) re-analysed Ewen rt al.’s data and showed
that only ao per cent of the variance in overall satisfaction
63 P. F. Wernimont, op. rit. had been accounted for by the three factors chosen. See
64 J. R. Hinrichs and L. A. hlischkind, op. rit. G. B. h e n , Addendum to “An empiricaltest of the Hnzbeq
6’ D. K. Lahiri and S. Srivastva, op. rit. two-factor theory”, Journal of Applied Psychology, 1966, 50,
66 F. Friedlander, op. tit. 551-555.

54
the same three factors as Hulin and Smith (1967). in the factor matrix”. Super (1962)’~found low
In that both these studies misinterpreted the two- or negative correlations between different intrinsic
factor theory (with respect to our interpretation items as sources of satisfaction, suggesting that these
one), used pay as the sole representative of hygiene items were independent. Friedlander (1963)~’~
factors and considered only three factors, the however, found that intrinsic items were positively
conclusions that the authors reach are of doubtful and significantly correlated. Friedlander suggested
validity. Furthermore, Hulin and Smith found that that differences, such as those between Super’s
the positive aspects of factors were more strongly study and other factor analyses, “might well be due
related to satisfaction than their negative aspects to the two separate approaches to the problem,
were to dissatisfaction. This may be related to the ie values z~ersussources of satisfaction”.
fact that whereas the satisfaction scale was a six- Whether Herzberg’s coding of reasons for satis-
point scale, in practice their dissatisfaction scale faction and dissatisfaction represents factors in the
became a three-point scale. factor-analytic sense is doubtful ; whether they need
In general the studies related to interpretation to or not is also debatable.
two are inconclusive. They have methodological Hinton ( 1 9 6 8 ) presents
~~ results which he suggests
weaknesses, or where they refute interpretation two cast doubt upon the methodology and validity of the
their implications for interpretation one are not two-factor theory. He demonstrates, on a population
clear, as the relationship of all factors to satisfaction of 143 full-time students, that the same method with
is not the same as the relationship of factors to dis- different recall periods, and different methods of
satisfaction. ratings, lead to differing rank orders of the Herzberg
factors. However, Hinton adopts interpretation four
c. Additional studies critical of the two-factor theory of the two-factor theory, and consequently his
Do “motivators” form a distinctive grouping conclusions are subject to the defects discussed
separate from “hygiene factors” ? Several studies earlier. A re-analysis of Hinton’s data offers support
have been undertaken to measure the extent to for interpretation one of the two-factor theory.
which the categories Herzberg proposes are “factors” However, Hinton does show that if different
in the factor-analytic sense (Ewen, 1964;~’Malik- methods of investigation (eg different recall periods)
nowsky and Barry, 1965;~’Graen, 1 9 6 6 ; ~and ~ are used, then the rank order of motivators and
Friedlander, 1963).~’Quinn and Kahn (1967)’~ hygiene factors will vary.
point out that none of these “factor analyses of
job satisfaction measures has yielded any clear Conclusions drawn from studies of the two-
first or second order factors which correspond to factory theory
Herzberg’s hypothesized two dimensions and It is apparent that the two-factor theory is contro-
simultaneously account for the bulk of the variance versial. Some investigators quote in support of the
theory studies which others regardasnon-supportive.
R. B. Ewen, “Some determinants of job satisfaction: a Differing interpretations have been placed upon the
study of the generality of Henberg’s theory”, Journal of
Applied Prychology, 1964,48, 161-163. two-factor theory, and the validity of these inter-
M. R. Maliknowski and J. R. Barry, “Determinants of pretations has itself become a subject of controversy.
work attitudes”, Journal of Applied Psychology, 1965,49,446-
45’.
’* G. B. Graen, op. cir. ’1 D. E. Super, “The structure of work values in relation
” F. Friedlander, “Underlying sources of job satisfaction”, to status, achievement, interests and adjustment”, Journal of
Journal of Applied Psychology, 1963,47,246-250. Applied Psychology, 1962, 46, 231-139.
74 R. P. Quinn and R. L. Kahn, “Organizational psycho- ’ 6 F. Friedlander, “Underlying Sources of Joh Satisfaction”,
logy”. in Annual Review of Psychology, 1967, 18, 437-466. op. cir.
Annual Reviews Inc., California. ” B. L. Hinton, op. ( i f .

55
Herzberg’s two-factor theory of job attitudes

educational groups” (p. 411). Henberg (1959~’


All these conflictsunderline the ambiguities inherent and 1965):~ Schwartz (1959),8’ and Schwartz
in the theory. et 01. (1963):~ Wernimont (1966):’ Myers (1g@),9O
A few observations are warranted, however, on Saleh (1964)’’ and Lindsay et al. ( 1 9 6 7 ) ~all ~
the basis of studies testing the twa-factor theory. confirmed this finding. Hinrichs and Mischkind
(1961)~~ alone, on a population of technicians,
Methodology. One of the most persistent criticisms found hygiene factors to be more important overall
of the two-factor theory is that it is method-bound. than motivatore. Several of these studies have,
Quinn and Kahn (1967)’~point out that “Those however, shown that motivators are more important
studies which best support the Henberg theory are than hygiene factors to respondents at lower
those methodologically most similar to Herzberg’s occupational levels, as well as to those at higher
work” (p. 456). The crucial question to answer, occupational levels.
with respect to this criticism is: what is it in Hen- Centers and Bugental(1966)~~ found that white-
berg’s methodology that ensures support for his collar employees placed more importance upon
theory, yet is lacking in other methodologies which motivators and less importance upon hygiene
yield results inconsistent with the two-factor factors than did blue-collar employees.
theory? The relative importance of motivators and
hygiene factors to different occupational levels
Occupational level and overall importance of motiv- has been well documented. However, relatively
ators and hygiene factors. A consistent finding hsls few studies, within the context of a single study,
been that motivators are more important than have considered whether the twefactor theory is
hygiene factors overall. Whitsett and Winslow supported to a greater or lesser extent by one
(1967)’~ make this point clearly: “There is one occupational level than by another. Is it possible
consistent finding” (Burke, 1966;”~Centers and that the two-factor theory is applicable to one
Bugental, 1966;~’Ewen, Smith, Hulin and Locke;8z occupational level but inappropriate for another 1
Friedlander, 1 9 6 4 ; ~ ~Graen, 1966;’~) “which The relative importance of motivators, compared
focuses on the issue of the relative importance of with hygiene factors, is an interesting result. It has
motivator and hygiene factors. These studies report, implications with respect to conventional personnel
in general, that motivators appear more important programmes where the emphasis rests primarily
in rankings of factors in job attitudes than the hygiene upon hygiene factors in preference to motivators.
factors. One possible explanation of this result be- This is possibly because hygiene factors are more
comes evident in examining the populations used tangible and open to amelioration.
in these studies. The subjects, for the most part,
have been drawn from higher occupational and Recall period. Several investigators have suggested
that the use of lengthy recall periods might bias
R. P. @inn and R. L. Hahn, op. cir.
79D. A. Whitsett and E. K. Winslow, op. Fit. F. Henberg e i a/., The Morivrrrion ?o Work, op. dt.
OD R. J. Burke, “Arc Henberg’s motivators and hygiencs F. Huzberg, “The motivation to work amongst Finnish
midimensional ?”,~oumcrlof Applied Psychology, 1966, 50, supervisors”, op. cit.
317-312. I7 P. Schwnrtz, op. cit.
R. Centers and D. E. Bugend, “Intrinsic and extrinsic 88 €1. A. Schwartz et al., op. cir.
job motivations among different segments of the working P. F. Wemimont, op. cir.
population”, ~ 0 u m o lof Applird Psychology, 1966, 50, 193- 90 M.S.Myers, 0). rir.
197. S. Saleh, op. cir.
R. B. Ewen, et d.,op. cit. C A. Linday ct sl., op. rit.
F. Friedlander, “Job Characteristics”,op. cit. J. R. Hinrichs and L. A. Mischkind, 09.rir.
G. B. Graen, op. cit. R. Centers and D. E. Bugental, op. dt.
results of job attitude research. The respondents probability of their being endorsed. The correlation
in Herzberg’s original research, and most of the he found was +0.87, the more desirable the item the
research supportive of the two-factor theory, were greater was the likelihood of it being accepted as
asked to recall a period (of exceptional satisfaction true to oneself. From such beginnings tests of
or dissatisfaction) which occurred in any job they “social desirability response set” have been con-
may have held. Does selective memory cause a con- structed and validated. These tests consist of items
sistent bias, and would this bias result in support describing highly virtuous, if unlikely, behaviour.
for the two-factor theory? Certainly two studies Naturally the items have high social desirability
concerned with present job attitudes failed to confirm values. On some personality tests such items are
the two-factor theory (Hinrichs and Mischkind, employed as “lie detectors”. Crowne and Marlowe
1967;~’Lahiri and Srivastva, 1 9 6 7 ) . ~ ~ ( 1 9 6 4 ) ~believe
~ that a good assessment can be
made of the extent to which individuals respond to
6. SOCIALLY DESIRABLE RESPONSES : a possible these tests in a socially desirable manner, and hence
explanation of the conflicting results of research of their “need for social approval”.
related t o the two-factor theory Individuals vary in the extent to which they
This topic is of importance to job attitude research respond in a socially desirable manner. It is not
in general. A question frequently asked of interview really possible, however, to determine whether
and questionnaire research is: does an individual higher scores on social desirability tests are obtained
report things as he really sees them, or does he report because respondents are deliberately distorting
the things he thinks he is expected to report? A their replies, inadvertently distorting their replies,
related question is:, do individuals perceive and or because they do behave or at least actenzpt to
remember objectively, or do they distort events, behave in the manner they report they do. Whatever
even to themselves, in a manner which reflects the reason for giving socially desirable responses
favourably upon them? In other words, how great is individuals vary in the extent to which they do so,
the tendency of respondents to give socially desirable and their level of “social desirability response set”
answers, and how will this affect job attitude (as measured by a social desirability scale) is related
research ? to their behaviour, verbal or otherwise, in other
The answers to these questions usually stress social situations (Crowne and Marlowe, 1 9 6 4 ) . ~ ~
the confidentiality of the research situation. It is
argued that, because the research is confidential, Situational determinants of socially desirable
respondents have no reason to respond in a socially responses
desirable manner, What they say in the interview Apart from the strength of an individual’s approval
or questionnaire can have no repercussions. How- motive-measured by his tendency to give, or not to
ever, whilst confidentiality, if fully accepted, re- give, socially desirable responses-other factors are
moves one of the reasons for giving socially desirable likely to increase or minimize the occurrence of
responses, it does not remove them all. such responses.
Social desirability as I personality characteristic
a. Confidentiality. The extent to which a research
Edwards (1953)97‘Orrelated the ratings for the situation is perceived as confidential will affect the
desirability of Personality with the responses given. Individuals who fear, or believe,
95 1. R. Hinrichs and L. A. Mischkind, op. rit. that what they say might get back to their superiors
96 b.K.Lahiri and S. Srivastva, op. tit.
q7 A. L. Edwards, “The relationship between the judged 98 D. P. Crowne and D. Marlowe, The Approval Motive,
desirability of a trait and the probability that the trait will be Wiley, New York, t964.
endorsed”, Journal of Applied P@ologY, 1953,37, 90-93. 99 Ibid.

57
Herzberg’s two-factor theory ofjob attitudes

or other people important to them, will tend to for example, asked j o b applicants to rate various job
avoid saying things which could be interpreted as characteristics with respect to their importance.
reflecting badly on them. The ratings were obtained in the job application
situation. It seems reasonable to assume that the
b. Research methodology. Different research method- ratings obtained from each individual rellected
ologies may tend to minimize or maximize socially as much the importance he felt he should ascribe
desirable responses. This point is related to the to the job characteristic (so as not to prejudice his
perceived confidentiality (point a) of the research chances of selection) as the felt importance of the
situation. An anonymous questionnaire may be factor itself.
perceived as more confidential than a face-to-face Because investigators have largely ignored this
interview; the questionnaire can be unnamed whilst issue many of the factors described above, which
the confidentiality of the interview will be based may be important in determining the level of
solely upon trust. socially desirable responses, have not been con-
sidered. The description of most studies is not
c. Formality of the research sitriation. The degree of sufficiently explicit to permit any assessment of the
formality may also affect the extent to which socially extent to which socially desirable responses will have
desirable responses are given. Respondents in a occurred. This has also resulted in ignoring the
formal and “correct” social situation may be more problem of which (if any) factors affect socially
likely to give formal and socially “correct” re- desirable responses, and in what manner they affect
sponses. attitude research.

d. The investigator’s status. The perceived status of Social desirability and the two-factor theory
the investigator may affect whether or not socially
desirable answers are given. Perhaps a high-status It is plausible to suggest that the extent to which
questioner is more likely to elicit more socially research supports or negates the two-factor theory
desirable responses than would a lower-status person. is a function of the degree to which the occurrence
of socially desirable responses have been maximized
e. Ego-involvement in the research situation. The or minimized. Indirect evidence for this comes
extent to which the questions asked are concerned from the observation that studies supporting the
with topics of personal importance to the respondent two-factor theory tend to use a face-to-face inter-
will influence the social desirability of his replies. view situation. Information concerning the per-
Questions concerning topics which are unimportant ceived status of the interviewer, the perceived
are unlikely to prompt socially desirable response; confidentiality of the research situation and the
topics which are central to a respondent’s life, for formality of the research situation is at best inex-
instance his marital relations, are more likely to plicit in studies on the two-factor theory. It is
elicit distorted views. possible that the face-to-face research situation,
interacting with one or more of the situational
Social desirability and job attitude research variables, tends to maximize the occurrence of
socially desirable responses. Research which is non-
Job attitude research has largely ignored the pos-
supportive, characteristically using a confidential
sibility that the results obtained reflect as much, or
questionnaire, may have minimized socially desirable
more, what individuals think they should say as
responses.
opposed to what they really feel. Jurgensen ( 1 9 4 8 ) , ~ ~ ~
loo C. E. Jurgensen, “What job applicants look for in a Blum, Experimental fndurrriul Psychology, Prmtice-Hall Inc.,
company”, Prrsonrrl Pjyrhology, 1948, I (reprinted in hi. L New York, 1952).
Kahn (1961),~O’House and Wigdor ( 1 9 6 7 ) ’ ~ ~attributed to defensive processes within individuals.
and Hinrichs and Mischkind ( 1 9 6 7 ) ‘ ~have ~ pro- It was predicted that individuals interviewed in a
posed that the results of the two-factor theory are highly firma1 and highly ego-involved situation
a function of social desirability. Vroorn (1964)”~ would be more likely to attribute their past highly
made this point effectively when he wrote (p. 129) satisfving experiences to those factors which Herzberg
“It is . . . possible that obtained differences between called motivators, and attribute their past highly
stated sources of satisfaction and dissatisfaction dissatisfying experiences to those factors Herzberg
stem from defensive processes within the individual labelled hygiene factors, than would individuals
respondent. Persons may be more likely to attribute interviewed in a less ,formal and less ego-im~olz*ed
causes of satisfaction to their own achievements situation.
and accomplishments on the job. On the other To test this hypothesis, two contrasting groups
hand, they may be more likely to attribute their dis- of respondents were asked questions similar to those
satisfaction not to personal inadequacies or deficien- used in Herzberg’s original study.
cies, but to factors in the work environment: ie The experimental group answered the questions
obstacles presented by company policies or super- during the course of a formal job selection interview
vision.” at a given company. Acceptance or rejection of the
respondent’s application for the desired position
within the company clearly depended upon his
7. THE SOCIAL DESIRABILITY HYPOTHESIS
IN RELATION TO THE TWO-FACTOR THEORY:
performance in the interview; it was in his own
interests to show himself in the best possible light.
RESEARCH FINDINGS
The control group consisted of employees of the
The discussion so far concerning the relationship same company, who answered the same questions
of social desirability to the two-factor theory has during the course of an informal interview. The
been entirely at a theoretical level. No empirical interviewer in this situation attempted to ensure
evidence has yet been offered to substantiate the that the respondent was fully at ease. The employee
argument. It has not been established whether the knew that nothing depended upon his performance
situation variables considered affect the extent to in the interview. The interview was confidential
which the two-factor theory is confirmed or refuted ; and the interviewer was known to be independent
nor has it been shown whether the level of an of the management of the company. It was stressed
individual’s approval motive is related to the that the respondent should be as accurate as possible
answers he gives to Herzberg-style interview in his replies. In short, the individual had less to
questions. Two studies which test the social desir- gain by responding in a socially acceptable manner.
ability hypothesis empirically will be described The experimental group comprised fourteen male
below. respondents. The control group comprised a further
fourteen respondents who were individually matched
An experimental field-study
as closely as possible for occupational level and age
Wall, Stephenson and Skidmore ( 1 9 7 0 ) investi- ~~~ from a pool of fifty-five respondents (see Wall et ui.,
gated the hypothesis that Herzberg’s results can be 1970, for further methodological details).
Table 2 presents the results of that study in a form
lo’ R. L. Kahn, “Job factors, attitudes and effects. Review
of F. Herzberg, B. Mausner and B. Snyderman-The Motiv-
comparable to that used by Herzberg et al. (1959,
ation to Work”, Contemporary Psychology, 1961,6, 9-10. p. 7 2 ) . The length of each bar represents the per-
lo2 R. 1.House and L. A. Wigdor, op. cir. centage frequency of occurrence of a given factor
R. Hinrichs and L. A. Mischkind, op. cit.
103 J.
lo‘V. H. Vroom, op. cit.
IO’T. D. Wall, G. M. Stephenson and C. Skidmore, experimental field study”, British Journal of Social and
“Ego-involvement and Henberg’s two-factor theory : an Clinical Psychology (in press).

59
Herzberg's two-factor theory ofjob attitudes

TABLE 2 Percentage of each factor appearing in high and low attitude sequences
Experimental
group Dissatisfaction Satisfaction

Factors I l I r l l I
Recognition C
Responsibility
Work itself C Motivators
Achievement

Work conditions i
Company policy and admin. I
Supervision
Security
II Hygiene factors
Interpersonal relations I
Personal life
Salary c
Summary
i40i2oiw80 60 40 20 o o M 40 60 aoiooi20140
Motivators
Hygiene factors

Control group Dissatisfaction Satisfaction

70 60 50 40 30 20 10 0 0 10 20 30 4050 60 70
Factors
Recognition
Respondbility
Work itself Motivators
Achievement
Possibility of growth
Work conditions

kl
Company policy and admin.
Supervision
Security Hygiene factors
Interpersonal relations
Personal life
Salary
Summary
14012010080 60 40 20 0 0 20 40 60 80100120140

Motivators
Hygiene factors

in relation to the number of sequences elicited. whilst the control group did not. In the experi-
(Note that percentages total more than 100, since mental group, motivators (recognition, respon-
more than one factor can occur in each sequence.) sibility, etc.) are positively associated with high job
Quite clearly, the experimental group yielded attitudes (satisfaction), whilst the hygiene factors
results similar to those reported by Herzberg ct al., (supervision, work conditions, etc.) occur mainly in

60
Responsibility
Work itself
Achievement
r I
Possibility of growth I . - I J
Work conditions I
Company policy and admln.
Supervision 1
Security
Inter-personal relations 1 I
Personal life
Salary C I

Control group Dissatisfaction Satisfaction

600 500 400 300 200 100 0 0 100 200 300 400 500 600
Factors
Recognition
Responsibility
Work itself
Achievement
Possibility of growth
Work conditions
Company policy and admin.
Supervision

--
Security
Inter-personal relations
Personal life
Salary Um Summary
12001000 800 600 400 200 0 0 200 400 600 800 1000 1200

Adjusted for 12 sequences


connection with low job attitudes (dissatisfaction). An alternative method of presenting the results is
These trends are very much less marked in the more legitimate. I n the previous method no allow-
control group. It is interesting to note that advance- ance is made for the fact that some individuals
ment, a motivator, was not mentioned by any of the yield several factors whilst others yield only one.
subjects. Verbose respondents may influence the results

61
Herzberg’s two-factor theory of job attitudes

disproportionately. The alternative method, see were simultaneously evaluated. The hypotheses
Table 3, used controls for the number of factors tested the extent to which support, or non-support,
offered. Equal weight is attributed to each person’s of the two-factor theory was related to the respon-
total contribution rather than to each factor. In dents’ age, social class, overall job satisjaction,
practice this means that factors are given scores neuroticism, extraversion, and level of approval
depending upon whether they were the sokperceived motive (ie the tendency to give socially desirable
reason for the feelings reported, one of two reasons, responses). Furthermore, the relationship between
one of three, one offour or one 0ffi.e reasons. Thus, each of these variables, and support, or non-support,
where the factor was the sole reason it scored for the two-factor theory, was considered separately
“IOO”, where one of two “50”, one of three “33+”, for three different recall periods. These recall
one of four “25” and one of five “zo”. periods represented present satisfactions and dis-
Again, the experimental group yields results satisfactions, satisfactions and dissatisfactions in
similar to those reported by Herzberg, whereas the year prior to the interview, and satisfactions
the control group does not. These differences are and dissatisfactions in a previous occupation.
statistically significant at a high level.
The results quoted above suggested very strongly
Procedure
that ego-defensive processes do, indeed, affect the
responses individuals give to job attitude questions. A sample of seventy-seven male employees, of a
In general, we concluded that the results of job medium-sized chemical process company, was
attitude research depending solely upon evidence interviewed. This sample represented 30 per cent of
acquired in a job application situation (eg Jurgensen, those who were both involved in the manufacture
1948)106 must be viewed with extreme caution. of the company’s main products and who had worked
Such work should make every effort either to mini- for the company for at least one year. Each inter-
mize the effect of defensive processes or include in viewee, having stated his age and occupation and
its design a means of assessing and partialling out having assessed his level of overall job satisfaction,
their effects. In particular, this study suggested that was asked to describe in his own words: (a) the
statements by employees which are in accord with determinants of his present satisfactions and dis-
Herzberg’s two-factor theory of job motivation are, satisfactions; (b) the determinants of his most
to a large extent, an artefact of ego-defensive pro- satisfied and most dissatisfied periods at work during
cesses. In the light of this result further investigations the course of the year prior to the interview; and
were undertaken with a larger sample to assess more (c) the determinants of his most satisfied and
directly the relationship between such individual dissatisfied periods at work during the course of
processes and the tendency towards “Herzbergian- previous employment. Each interviewee completed
ism”. an Eysenck Personality Inventory (Form B), and
a Marlowe-Crowne Social-Desirability Scale. The
Further investigations Eysenck Personality Inventory yielded assessments
Subsequent to the experimental field-study de- of each individual’s level of neuroticism (as opposed
scribed above, a more comprehensive research to stability) and extraversion (as opposed to intro-
project was ~ompleted.’~’Several hypotheses, version). The Marlowe-Crowne Social-Desirability
suggested by prior research and discussion of the Scale measured the strength of each respondent’s
two-factor theory and other job attitude research, need for social approval.
The results demonstrated that age, overall job
Io6 C. E. Jurgenscn, op. cir.
satisfaction, neuroticism and extraversion were all
lo’ A paper by Wall (in preparation) describes the procedure unrelated to the extent to which the two-factor
of this study in detail and the more important results. theory was supported or not supported. The single

62
Thus, at tne time this research project was THE WORK OF DECK RATINGS
undertaken the great Cunard Company, the most The special gang
famous name in British shipping, had sold five of The “special” men were at the peak of the informal
its passenger vessels (though the Queen Elizabeth hierarchy of deck ratings. They worked on specific
was a post-dated sale) and were reduced to a passen- tasks with only loose supervision by the bo‘sun.
ger fleet of three vessels. Only two 20 000 ton vessels, In the main they were men H-hohad served on board
the Carmania and Franconia, operating as year- the ship for long periods and they tended to be
round cruise ships and the Queen Elizabeth, to be older than the average deck hand. This geronto-
replaced at the end of 1968 by the new Q.E.2, re- cratic system also applied to other gangs, and with
mained. The fate of Cunard was symbolic. T h e few exceptions the older men were allocated to the
demise of the great passenger liners of the North better jobs. A seaman of mature age is more likely
Atlantic was the end of an era. Less dramatic, but to stay put in one ship and not drift from ship to
more important from the viewpoint of the present ship. The “Queens”, operating on a regular service,
study, it meant the loss of job opportunity for some were good ships for such men with family respon-
3,000 merchant seamen. sibilities. T h e short absence from home made it
possible to keep in touch with the domestic situa-
tion.
RESEARCH PERSPECTIVE: THE “Q.E.“ DECK There were eleven seamen in the special gang,
CREW one of these helped the 1st officer with clerical work,
The Queen Elizabeth carried a crew of 1,190 officers two had charge of deck stores, four were deck men--
and men divided into the three departments of working on their own initiative on passenger decks,
Deck, Engine and Catering. Within each department two had the permanent task of keeping ships’
there was an hierarchy of officers and petty-officers life-boats in order, one was handyman to the
with ratings at the bottom of the authority scale. carpenter and a “spareman” did odd jobs in the
The author worked as an able seaman-as a skilled passenger accommodation.
rating in the deck department-and the ratings of
this department are the main focus of the present Day gands)
paper. Other shipboard personnel were seen from There were two day gangs on theQ.E. but for normal
the perspective of the deck crew. routine purposes they were handled as one unit.
There were forty-five ratings employed on the Within these gangs there were two or three informal
Queen Elizabeth, excluding boy ratings and quarter- specials; unlike the specials they had no regular
masters. This latter group are helmsmen in the specified task but they were recognizable by the
British mercantile marine, they worked mainly way jobs were allocated to them and by the kind of
around the navigation bridge and apart from the work they were given.
other deck ratings. The bo’sun had charge of the At 08.00 each morning the deck ratings on day
day-to-day running of the deck crew under the work mustered outside their cabins and a bo’sun’s
orders of the Senior 1st Officer. Two bo’sun’s mate called a name roster and allocated jobs.
mates were next in the formal hierarchy; they were Members of the formal special gang were normally
responsible for the actual face-to-face supervision told to “carry on” without any task instruction. A
of ratings. Ratings were organized in three “watches” few others, the informal specials, had jobs allocated
or “gangs” : the day gang, night gang, and the special by the bo’sun. During the muster call they were
gang. There were also a small number of ratings who dismissed by a standard form of words-“the
tended to be selected for individual tasks without bo’sun wants to see you”. If caught sight of later in
supervision, and these we shall descrihe as “informal the day they would he seen to be doing the more
specials”. pleasant unsupervised tasks. Selection of seamen
Hrezberg’s two-factor theory ofjob attitudes

sibility, or work itself. I n contrast, hygiene factors that individuals with a strong tendency to give
are extrinJic to the performance of work, they are socially desirable or approved responses answer
related to the interpersonal relations and company job attitude questions in a manner more predictable
supervision, work environment. Work conditions, from the two-factor theory than do individuals
policy and administration, are a11 examples of whose “need for social approval” is less marked.
hygiene factors. The principal hypothesis of the
Implications for the two-factor theory
two-factor theory is that motivators and hygiene
factors are differentially related to job satisfaction The two research studies described above (Wall
and job dissatisfaction. The theory proposes that and Stephenson, 1g71,’~*Wa11,’09) have clear
motivators are more important as determinants of implications for Herzberg’s two-factor theory.
satisfaction and less important as determinants of They strongly suggest that the research results upop
dissatisfaction, and that hygiene factors demonstrate which Herzberg’s theory is based are a function of
the reverse relationship. “social desirability”. Individuals with a strong need
It is clear, however, from research which has for social approval, or individuals in a situation
attempted to evaluate the two-factor theory, that requiring highly socially acceptable responses,
the principal hypothesis is open to several inter- were shown to give results highly consistent with the
pretations. The most legitimate interpretation is two-factor theory. However, individuals with a
that motivators are more important than hygiene lesser need for social approval, or in a situation which
fnctors as determinants of job satis/hction; and did not necessarily demand socially acceptable
hygienefactors are more important than motivators as responses, gave results which were much less con-
determinants Ofilissatirfaction.This restatement of the sistent with the two-factor theory. Furthermore, it
two-factor theory we have named interpretation one. was found that the use of a lengthy recall period
Empirical tests of hypotheses derived from the tended to maximize the relationship between
two-factor theory have yielded conflicting results. social desirability and the two-factor theory.
This conflict may, to some extent, be due to the Consequently, as a description of the structure of
differing interpretations which have been placed job attitudes and of the determinants of satisfaction
upon the theory. However, it is noticeable that and dissatisfaction, the two-factor theory is not
those studies which support the two-factor theory tenable.
tend to use a methodology very similar to that used Implications for job attitude research in general
in the study upon which the two-factor theory is and for practical appllcatlonr
based. Research which refutes the two-factor theory Whilst the research on the social desirability hypo-
has tended to use dissimilar methodologies. It is thesis has focused upon the two-factor theory in
suggested that the extent to which research supports particular, it has much wider implications. Other
or negates the two-factor theory is a function of job attitude studies are likely to have obtained results
the degree to which socially desirable responses which were also systematically distorted by respon-
have been encouraged by the methodology em- dents’ need for social approval. Research which may
ployed. Two studies by the authors demonstrate have enhanced any tendency towards the giving of
that a strong relationship does exist between re- socially desirable responses (ie Jurgensen, 1948)’lo
sponses predictable from the two-factor theory should be viewed with extreme caution. Research
and the need to respond in a socially acceptable which ignores this issue should also be considered
manner. One of these studies shows that individuals with care.
in a highly ego-involved situation respond to job
attitude questions in a manner predictable from the T. D. Wall and G. M. Stephenson, op. tit.
T. D. Wd, Social Desirability and the Hrrzberg Two-
tao-factor theory, whilst individuals in a less ego- Fartar Theory :A Field Study (in prcss).
involved situation do not. The other study indicates C.E.Jurgensen, op. cir.

64
What is one to conclude? We have shown that industrial relations. Traditionally, personnel policies
the results of some attitude research are related to have concentrated upon ameliorating hygiene
the respondents’ needs for social approval-are factors. There is nothing in the present research to
we then to dismiss all such research as untenable ? suggest, that this policy is mistaken.
Are we also to conclude that job attitude research However, personnel policies may have over-
cannot yield knowledge of the reasons for satis- emphasized the importance of hygiene factors.
faction and dissatisfaction ? A more constructive Our results have shown that employees tend to
response, which in the long term will yield more exaggerate the importance of hygiene factors as
accurate information, is to examine critically job determinants of dissatisfaction, for social reasons.
attitude research methodology and initiate a more This provides the rationale for Myers’ (1964)”’
comprehensive assessment of the relationship be- suggestion that “When opportunities for meaningful
tween the need for social approval and research achievement are eliminated (employees) . . . become
results. In particular, aspects of the research situ- sensitized to their environment and begin to find
ation which are likely to enhance socially desirable fault.” It is not surprising that hygiene factors are
responses to job attitude questions should be a picked on; they are both more tangible and are
subject of research in themselves. Besides high traditionally the overt causes of dissatisfaction.
ego-involvement and the personality of the re- Job enrichment has been proposed as a method
spondent, other variables, such as the status of the of improving job satisfaction (Paul, Robertson and
interviewer or the formality of the research situation, Herzberg, 1969).”* This involves “building into
may be equally important. people’s jobs, quite specifically, greater scope for
We need to determine which variables, in relation personal achievement and its recognition, more
to which methodology, are likely to promote socially challenging and responsible work, and more
desirable responses to job attitude questions. Having opportunity for individual advancement and growth.
established the important variables research may then It is concerned only incidentally with matters such
be designed to eliminate, minimize, or estimate their as pay, working conditions, organizational structure,
efects. For example, in the present case we have communication and training . . . ” Job enrichment
shown that the tendency to support the first inter- was justified as the basis of Herzberg’s two-factor
pretation of Herzberg’s two-factor theory is strongly theory, but is very loosely tied to it. Paradoxically,
related to a respondent’s need to give what he although our results show the evidence for the two-
considers to be acceptable answers to the questions factor theory, per se to be largely a function of the
posed. This makes difficult any assessment of the “need for social approval” they vindicate the policy
relative importance of motivator and hygiene of job enrichment as likely to promote satisfaction
factors to job attitudes. However, the effect of social and allay dissatisfaction.
approval is most apparent, as Wall’s study demon- Acknowledgement : Financial support fiom the
strates, in a respondent’s reasons for periods of Scott Bader Commonwealth and from the Social
dissatisfaction. When explaining such times, he Science Research Foirndntion is grnteful@ acknom-
tends to underestimate the influence of motivators. ledged by the authors.
The implication is that motivators are, in fact,
of considerable importance to both satisfaction and I L L M. S.Myers, op. cit.
dissatisfaction, at least in the present climate of IIZ W.J. Paul, K. B. Robertson and F.Herzbcrg. up. ci/.

You might also like