Professional Documents
Culture Documents
Manning 1989
Manning 1989
To cite this article: Michael R. Manning & Joyce S. Osland (1989) The relationship between
absenteeism and stress, Work & Stress: An International Journal of Work, Health &
Organisations, 3:3, 223-235, DOI: 10.1080/02678378908251558
Taylor & Francis makes every effort to ensure the accuracy of all the information
(the “Content”) contained in the publications on our platform. However, Taylor
& Francis, our agents, and our licensors make no representations or warranties
whatsoever as to the accuracy, completeness, or suitability for any purpose of the
Content. Any opinions and views expressed in this publication are the opinions and
views of the authors, and are not the views of or endorsed by Taylor & Francis. The
accuracy of the Content should not be relied upon and should be independently
verified with primary sources of information. Taylor and Francis shall not be liable
for any losses, actions, claims, proceedings, demands, costs, expenses, damages,
and other liabilities whatsoever or howsoever caused arising directly or indirectly in
connection with, in relation to or arising out of the use of the Content.
This article may be used for research, teaching, and private study purposes. Any
substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,
systematic supply, or distribution in any form to anyone is expressly forbidden.
Terms & Conditions of access and use can be found at http://www.tandfonline.com/
page/terms-and-conditions
WORK & STRESS, 1989, VOL. 3, NO. 3, 223-235
MICHAEL R. MANNING
College of Business Administration and Economics,
New Mexico State University, USA
JOYCE S. OSLAND
Weatherhead School of Management, Case Western Reserve University, USA
Downloaded by [University of Liverpool] at 03:20 06 January 2015
This study explores the relationshp between stress and absenteeism with a non-managerial white collar
sample (n = 147). Various operationalizations of absence were employed representing, frequency,
hours, and length of absence (one day, two day, > two day). Stress was operationalized with variables
representing stressors and strain from both work and non-work domains. Results found small but
consistent relationships between prior absence (one day, > t w o day, and total absence) and many of
the stress measures (work events, work conditions, life events, life conditions, job satisfaction, strain
and negative affect). Two sets of variance were identified from the prior absence measures, short and
long-term, that were both related to stress. No relationship was observed between stress and
subsequent absence.
A recent computer search on stress and absenteeism yielded 32 pages of references with 99
articles. Most of these articles were located in trade journals o r magazines and strongly
advised managers to establish health promotion and stress management programmes to
remedy employee absenteeism. Only a handful of the articles were scholarly studies
empirically measuring stress and/or absenteeism. While the popular press commonly links
stress with absenteeism, there has been very little empirical research to support this
assumption. This study attempts to build on the small body of empirical literature exploring
the relationship between these two constructs.
The idea that work-related stress results in absenteeism was first suggested by Hill and
Trist (1955). They portrayed absence as withdrawal from the stress of work, thereby
allowing individuals to better adapt or cope with their social environment. Surprisingly,
other theoretical propositions connecting stress and absenteeism are not as well developed.
Steers and Rhodes’ (1978) theory appears to be the only other comprehensive absenteeism
theory that suggests stress plays a central role in withdrawal from work. They contend that
stress, arising from situational work-related factors, influences employee dissatisfaction with
their jobs. Job dissatisfaction,in turn, causes short term dispositional absence. Their model is
cyclical in that absenteeism also influences the job situation and pressures to attend.
This research was supported by grants from the W. K. Kellogg Foundation and Blue Cross and
Blue Shield of Ohio. Correspondence concerning this article should be addressed to Michael R.
Manning, Department of Management, College of Business Administration and Economics, New
Mexico State University, Las Cruces, New Mexico 88003-0001.
02674373/89 f3.00 01989 Taylor & Francis Ltd.
224 M . R . Manning andj. S. Osland
1. Empirical evidence
The empirical work of Hedges (1973) found high absence rates for labourers, service
workers, and operatives. Frankenhaeuser and Gardell (1976) found perceptions of stress and
absenteeism of lumbermen were highest among those whose work could be characterized as
skilled but low in autonomy. They hypothesized that work perceived as low in control and
autonomy creates stress which results in absenteeism.
More direct research on the relationship between one aspect of stress, anxiety, and
absenteeism has been conducted by Bemardin (1977), Pocock et al. (1972) and Sinha (1963).
All three of these research efforts found a moderate relationship (about r = 0.20) between
self-reported anxiety and absenteeism. In addiiton, Parkes (1983) reported a correlation of
0.15 between frequency of absence and affective state for student nurses.
Empirical research on the relationship between work stress and self-reported
absenteeism has been undertaken by Spector et al. (1988) and Margolies et a f . (1974). With a
sample of 191 female secretaries, Spector et al. (1988) reported only one significant
Downloaded by [University of Liverpool] at 03:20 06 January 2015
correlation (0.17) between perceptions of role ambiguity and frequency of sick day absences
over the past three months. The stress measures of autonomy, workload, constraints, and
interpersonal conflict were not significantly correlated with absence frequency. Margolies
et al. (1974) also measured self-reported frequency of absence, this time with a national
sample of 1,496 working individuals. They found small but significant correlations for
underutilization (0.09), overload (0.06), resource inadequacy (0.06), and non-participation
(0.11).
Probably the two best empirical studies on the work stress-absenteeism relationship
were conducted by Gupta and Beehr (1979) and Jamal (1984). Both studies operationalized
absence through company records. Gupta and Beehr (1979) found that role ambiguity, role
overload, underutilization of skills, and resource inadequacy all correlated, 0.13 to 0.16,
with absence frequency in the subsequent month. Since n o correlation existed between their
stress measures and the prior month’s absence, Gupta and Beehr (1979) suggested that job
stress is a predictor or cause of future work absence. Jamal (1984) reported the highest
correlations found in the literature between stress and absence frequency during a
subsequent four month period. With a sample of 440 nurses, the correlations reported were:
role ambiguity (0.34); role overload (0.27); role conflict (0.23), and resource inadequacy
(0.37). By and large, the correlations derived from the literature support a stress-
absenteeism relationship but of small magnitude.
2. Absence measures
Any attempt to cast further light on the stress-absenteeism relationship raises
methodological issues related to measuring absence. Historically, absence has been
operationalized by the total number of hours absent per unit of time. Recently, researchers
have suggested that different lengths of absenteeism (e.g., one day, two day, short term,
long term, etc.) and different forms of measuring absence (e.g., hours and frequency) might
correlate differently with other variables (Chadwick-Jones et al. 1982) and with stress in
particular (Fogler and Belew 1985).
Chadwick-Jones et al. (1982) proposed that one-day absences are usually voluntary and
may represent what Staw and Oldham (1978) term an ‘unauthorized vacation’. Hence,
Fogler and Belew (1985) have hypothesized that one day absences might be the best
conceptual measure for stress research. Lengthier absences are more likely involuntary
unavoidable absences, such as actual illnesses. To the extent that physical illnesses are stress-
induced, the associated long term absences should also relate to measures of stress. Further
Absenteeism and stress 225
frequency, rather than hourly indices of absenteeism, that give equal weight to all absences
regardless of length, may be the most effective measure for stress research. This position is
buttressed by Muchinsky’s (1977) finding that frequency of absence is the most reliable
absence index.
The study tests the general hypothesis that there is a positive relationship between our
measures of stress and employee absence. Furthermore, based on the conceptual work of
Fogler and Belew (1985), the study examines whether different patterns of absence (e.g.,
length of absence and frequency/hourly approaches) relate differently to our stress measures.
Results of the study should indicate which absence measures might be most useful when
conducting stress research. O u r hope is that investigations like this will yield both a better
understanding of the stress-absenteeism relationship and encourage innovative approaches
to theory development.
4. Methods
4.1. Procedures
Surveys were completed by participants at theirjob site during normal working hours. The
attendance data were gathered for twelve months prior to the survey administration and for
twelve months afterwards. Managers were responsible for keeping track of non-union
employees’ attendance. After receiving attendance information from individual personnel
files, the researchers coded the data.
4.2. Respondents
Survey respondents were recruited from a manufacturing company employing 350
individuals. Subjects were asked to participate voluntarily in a larger research project
focusing on the effectiveness of health promotion programmes in work settings. Three
hundred and three employees participated in the larger research project. However after the
attendance data was ‘cleaned’, 147 respondents were eligible for the present study. It was
necessary to eliminate 56 managers whose attendance was not systematically taken and
approximately 100 production employees who were union members. Although the
production workers’ data were highly accurate due to the use of a timeclock, their union
contract stipulated that one day absences were not reimbursed. This policy did not apply to
non-union members. Therefore, the production employees were dropped from this study
because this policy likely affected the incidence of one day absences. Subjects were primarily
white collar workers in non-managerial support positions.
4.3. Measures
The study operationalized the following sets of variables: stressors; strain and well-being
measures; and categories of absence.
226 M . R . Manning and J . S . Osland
participants to rate on a 7-point scale (‘dissatisfied’to ‘satisfied’) how satisfied they were with
ten specific aspects of their jobs (Manring 1979).
Strain was measured by an instrument constructed by Adams (1981). This questionnaire
consists of26 items on a 5-point response format (‘rarely or never’ to ‘quite frequently’) that
sampled the content domains of depression, anxiety and somatic complaints. The Affect
Rating Scale (Sipprelle et al. 1976) provided a measure of general affective state. T w o scales
from this instrument were utilized: (1) Positive affect and (2) Negative affect. The final self-
report measure, Quality of Life (Medical Datamation Inc. 1982), assessed one’s general life
and emotional well-being.
4.3.3. Absenteeism
Absence was divided into both total hours and frequency of absence for the following
categories: one day, two day and more than two day absences. One and two day absences
were classified as short term absence, while greater than two day absences were considered
long term absence. T w o total absence measures, for frequency and hours, were computed
by adding short term and long term absence figures. All of these categories represent
absences coded by the company as illness. Absence due to other reasons such as personal
business, family emergencies, medical appointments, disability and tardiness, were coded
separately and not included in this analysis.
To make allowances for possible seasonal variations, the absence data were accumulated
for the twelve months before and after the completion of the survey. The totaled absence
figures were divided by the number of months the subjects had been employed. Therefore,
absence figures represent monthly averages. As the study progressed, discussions with
company officials revealed that on infrequent occasions, some managers did not record short
one or two hour absences if an employee worked compensatory non-paid overtime.
Because of this, it is possible that the attendance data for the one day absence category may
be under-reported.
5. Results
5.1. Characteristics of sample
The demographic nature of the 147 participants is quite heterogeneous. Fifty men and 97
women participated in the study. There were 122 whites, 24 blacks, and one Asian. The
mean age was 35.8 years (ranging from 21 to 62). Eighty-four individuals were married, 28
were either separated, widowed, or divorced and 33 had never married. The sample
A bsenteeism and stress 227
included 66 college graduates, 47 with some college, 32 with high school diplomas, and one
had not graduated from high school. Total family income ranged from less than $12,000 to
more than $60,000 with a median income in the $30,000 to $39,000 category.
measures. These results indicate similar stable mean absence patterns when comparing the
frequency and hourly measures for the subsequent and prior year.
lv
lv
W
Standard
Stress variables Mean deviation 1 2 3 4 5 6 7 8 9 10
Work stressors
1 . Work events 262.12 162.92 -
2. Work conditions 2.21 0.59 0.49 (0.77)'
Life stressors
3. Life events 201.27 165.41 0.28 0.34 -
4. Life conditions 2.13 0-62 0.28 0.36 0.33 (0.75)
Work strainlwell-being
5. Job related tensions 2.67 1.01 0.52 0.80 0.35 0.46 (0.90)
6. Job satisfaction 4.95 1.23 -0.36 -0.77 -0.34 -0.33 -0.70 (0.88)
Life strainlwell-being
7. Strain 2.02 0.66 0.39 0.52 0.44 0.62 0.57 -0.51 (0-88) %
8. Negative affect 2.93 0.93 0.30 0.43 0.37 0.48 0.44 -0.51 0.71 (0.83)
9. Positive affect 4.84 0.86 -0.10"' 0.20 -0GI"" -0.22 -0.20 0.33 -0.35 -0.47 (0.79) ?
10. Quality of life 5.49 0.82 -0.31 -0.45 -0.17 -0.39 -0.50 0.55 -0.62 -0.73 0.65 (0.93) $
3
Table 2. Means, standard deviations and intercorrelationsfor Year I and Year 2 absence measures.
Absence Standard
measures Mean deviation 1 2 3 4 5 6 7 8 9 10
Frequency
Short term
1. One day 0.13/0.13' 0.15/0.14 (0-40)*
2. Two day 0.03/0.03 0.06/0*06 0.30/0.26 (0.25)
3. Total short
term 0.16/0.17 0.16/0.17 0.9510.94 0.5910.57 (0.44)
Long term
4. > T w o day 0.02/0.04 0.02/0*05 0.12*/0.20 0.03*/0.11* 0.11*/0.21 (-044)*
5. Total absence 0.17/0.18 0.18/0.18 0.9310.91 0.57/0*55 0.98/0.97 0.31/0.46 (0.44)
Hours
Short term
6. One day 0.90/0.94 0.97/1.08 0.97/0-99 0,30/0*20 0.91/0.91 0.05*/0.24 0.86/0.88 (0.40)
7. Two day 0.49/048 0.87/0.84 0.28/0.20 0.99/0.99 0.61/0.56 0.23/0.15 0.6210.54 0.29/0.26 (0.21)
8. Total short
term 1.3811.42 1.4911.53 0.7710.78 0.8210.75 0.9410.96 0.18{0.25 0,9210.93 0.82/0.85 0.78/0.73 (0.42)
Long term
9. Two day 0.70/0.64 1.83/1.49 0.01*/0.10* 0.06/0.17 0.03*/0.16 0.6910.74 049/0.34 0.03*/0.18 0.02*/0.12* 0.03*/0.19 (-0.05)*
10. Total absence 2.0812.06 240/2.34 0.2610.47 0.3210.50 0.33/0'59 0.70/0.69 0.4810.72 0.53/0.67 0.50/0.56 0.64/0.78 0.7910.77 (0.18)
to 0-18 (total hours). Hours and frequency measures for long term absence appear to have no
stability across the two time periods. This is not surprising since extended illness may not be
a consistent yearly event.
n
a
4.
Table 3. Pearson correlations between frequency absence measures and stress variables. 4
cd
-4
Work stressors
Work events 0.15* 0.08 0.04 - 0.08 0.14* 0.04 0.19* 0.16* 0.08
Work conditions 0.14* 0.00 0.06 0.00 0.14* 0.00 0.06 0.16* 0.01
Life stressors
Life events 0.06 0.02 -0.02 0.20** 0.04 0.12 0.08 0.22* 0.06
Life conditions -0.06 -0.03 0.09 0.13 - 0.02 0-07 0.04 0.14* -0.01
Work strainlwell-being
Job related tensions 0.09 - 0.08 0.03 -0.09 0.09 -0.10 0.13 -0.01 0.11 - 0.09
Job satisfaction -0.18* -0-02 -0.10 -0.02 -0.19** -0.02 -0-06 0.03 -0.19* -0.01
Life strain/well-being
Strain 0-16* 0.01 0.05 - 0.02 0.15* 0.00 0.13 0.03 0.17* 0.01
Negative affect 0.16* 0.01 0.03 0.01 0.14* 0.01 0.08 0.03 0.15* 0.02
Positive affezt 0.05 - 0.05 - 0.07 - 0.01 0.02 - 0.05 0.04 - 0.01 0.02 -0.05
Quality of life - 0.06 0.01 -0.08 0.01 -0.08 0.02 0.01 -0.03 -0.07 0.01
* p <0.05, * * p <0.01.
Downloaded by [University of Liverpool] at 03:20 06 January 2015
Table 4. Pearson correlations between hour absence measures and stress variables.
One day Two day Total short term Long term Total absence
Stress variables Year 1 Year 2 Year 1 Year 2 Year 1 Year 2 Year 1 Year 2 Year 1 Year 2
Work stressors
Work events 0.12 0.06 0.04 -0.10 0.10 - 0.01 0.18* 0.16* 0.21 ** 0.10
Work conditions 0.09 -0.01 0.05 -0.02 0.09 -0.02 0.1 1 0.01 O.t4* - 0.01
Life stresson
Life events 0.22** 0-06 0.03 -0.03 0.16* 0.03 0.17* 0.04 0-23** 0.03
Life conditions 0.18* -0-06 -0.04 0.09 0.10 0.00 0.07 0.03 0.11 0.02
Work strainlwell-being
Job related tensions 0.07 -0.09 0-02 -0.11 0.06 -0.12 0.11 -0.05 0.12 -0.11
Job satisfaction -0.15* -0.01 -0.07 0.01 -044* 0.00 -0.03 0.06 -0.11 0.04
Life strainlwell-being
Strain 0.17* 0.01 0.05 -0.03 0.14* -0.04 0.15* -0.04 0.20** -0.04
K
Negative affect 0.15* 0.00 0.02 0.00 0.11 0.00 0.10 -0.04 0.15* -0.04 P
Positive affect 0-01 - 0-04 -0.06 0.00 -0.03 -0.02 0.03 0.03 0.01 0.01
Quality of life - 0.06 0.03 - 0.07 0.04 - 0.08 0.04 -0.01 0.02 -0.06 0.05
*p<0.05, **p<O.Ol.
0
F
3
a
Absenteeism and stress 233
one day frequency is partialled from the relationship. This is not surprising, since one day
frequency and total absence frequency appear to be very similar measures (r = 0.93).
At first glance, the correlations with total absence hours (Yl) are very similar to the
results for total absence frequency. But contrary to the findings with frequency measures,
significant relationships still existed when one day or long term hour measures were held
constant. This suggests that the total hours-stress correlation is composed of two sets of
absence variance each related to stress: one day and long term absence. These results also
indicated no significant correlations between Y2 total absence measures and the stress
variables.
6. Discussion
Our primary research question addressed the nature of the relationship between stress and
absenteeism. Results indicated significant but small correlations between many of the stress
Downloaded by [University of Liverpool] at 03:20 06 January 2015
measures and prior absence. Except for the single significant correlation between work
events and long term absence, no relationships were observed between stress and the
subsequent year’s absence. These findings appear to run counter to the commonly held view
that work absence is a coping response to stress. They also seem to contradict Gupta and
Beehr’s (1979) and Jamal’s (1984) research which suggests job stress is a predictor of future
work absence. One possible explanation is that periods of stress are short with relatively
immediate reactions.
Explaining the nature of cause and effect is not possible with studies employing
correlational designs. However, our results may suggest the need to develop alternative
views ofthe relationship between stress and absenteeism. For example, our data appear to be
consistent with the view that absenteeism is a cause of stress, rather than the reverse. Clegg
(1983) contends that absenteeism is often flawed by researchers’ refusal to consider
alternative hypotheses. His research indicates that attendance impacts job satisfaction,
representing a departure from the traditional view that affect causes behaviour. This is also
consistent with the work of Ross and Fletcher (1985), who found that people who perceive
themselves missing work for trivial reasons conclude that this absence is due to job
dissatisfaction. Ross and Fletcher (1985) suggest that behaviour is affecting cognition rather
than vice versa.
If the explanation that absenteeism causes stress has merit, future research might try to
identify the conditions under which the causal arrow flies in this direction. Could a variable
such as job involvement determine whether stress predicts absenteeism or vice versa? For
example, if you are very involved in your work, absence might well be distressing; in
contrast, if your job is not a central life interest, absence might be a means of coping with
stress.
Causal detection with stress measures is further complicated by the possibility that any
measure of stress may be a generalization of past experiences. If this is the case, the point of
time at which the measurement of stress is taken may mean little more than the point from
which a generalized response, influenced by past experiences, starts. If so, it is hard to
identify which occurred first, stress or the absence. This explanation is indeed possible with
two ofour measures (Work Events and Life Events) since respondents were asked to identify
the occurrence of specific events within the past year. However, this logic is not as clearly
supported with our measures that asked for sentiments pertaining to ‘current status’, ‘how
you feel in general’, and ‘how frequently things are true for you these days’. Presumably the
latter questions would pertain to the past few weeks prior to survey administration. Does
stress that originated months before influence individuals’ responses to these stress measures?
234 M . R. Manning a n d ] . S . Osland
If the answer to this question is yes, then stress as measured here could have occurred prior to
any absence. But, it is still confusing why our stress measures were not significantly
correlated with subsequent absence. To untangle this complex causal web, researchers will
need to employ longitudinal investigations measuring stress and absence at numerous points
in time.
O u r second research question addressed the issue, ‘Do different forms of measuring
absence relate differently to stress measures?’ In addition, the study sought to identify which
absence measures might be most useful when conducting stress research. To explore these
issues we employed five measures of absence duration (one day, two day, short term, long
term and total absence) and two indexing approaches (frequency and hours). Results from
this investigation indicated that all our categories of prior absence had some correlation with
the measures of stress, except for the two day category. These indices did not differ in the
magnitude of correlation with stress. Intercorrelations among the absence measures showed
that one day absence is virtually the same measure as total short term for both hours and
Downloaded by [University of Liverpool] at 03:20 06 January 2015
frequency. These short term measures also completely dominate total absence frequency.
Therefore, one day frequency, one day hours, short term frequency, short term hours and
total absence frequency are virtually interchangeable measures. All these measures tap into a
source of variance created by one day absence and therefore conceptually reflect the
dynamics of this type of absence.
This discussion brings us back to the results of this study and to the question: W h y is one
day absence associated with stress? Is the one day absence a necessary mental health day to
cope with the mounting pressures ofcomplex life? Or, do those forced to miss a day ofwork
find themselves so overwhelmed when they return by the backlog of work, the pressure
from co-workers, clients and superiors that stress results? Obviously, further research is
needed to explain the connection here.
O u r analysis also uncovered another source of absence related to stress: long term
absence hours. Again, we can ask why long term absence is associated with stress. Does long
term absence create even more pressure when one returns to work than one day absence?
Does illness beget stress or is it a result of a demanding life?
Even though others have advocated the use of frequency measures in absenteeism, we
believe that it is important to utilize a measure that is sensitive to length of absence duration,
especially for the longer periods of absence where there is large variability in the hours index
and almost no variance in the frequency index. A measure of long term hours, like that
employed in this study, is an important measure to use in stress research.
Although we employed different operationalizations of absence in this investigation, we
did not establish stronger relations between stress and absenteeism than has already been
reported. While our correlations were significant, they were very small. Perhaps this is due
to the fact that there is little relationship between these two constructs or, perhaps more
complicated relationships exist than we have been able conceptualize.
O u r objective in undertaking this investigation was to determine whether the stress-
absence relationship was worth investigating and if so, what directions may be indicated for
future research. W e conclude that further investigation appears to have merit and two
directions seem plausible. First, we encourage stress researchers to conduct longitudinal
studies with other variables, such as job involvement or commitment, so that we can begin
to develop more conclusive evidence about the causal relationship between stress and work
absence. This research might clarify when absenteeism is a coping response or a creator of
stress or both. Second, we suggest that research incorporate more controlled designs to
determine the underlying connections between short term absence and stress as well as long
term absence and stress. What different conceptual interpretations are needed to understand
Absenteeism and stress 235
stress and its connections with short and long term absence? W h a t other variables effect the
stress-absenteeism relationship? Perhaps when w e know the answers t o these questions,
empirical research will have supported, refuted or modified the popular press' claim that
stress management reduces absenteeism.
References
ADAMS, J. D. 1981, Health, stress and your lye style (NTL Institute, Arlington, Virginia).
BERNARDIN, H. J. 1977, The relationship of personality variables to organizational withdrawal,
Personnel Psychology, 30, 17-27.
CHADWICK-JONES, J. K., NICHOLSON, N. and BROWN,C. 1982, Social psychology of absenteeism
(Praeger, New York).
COCHRANE, R. and ROBERTSON, A. 1973, The life events inventory: a measure of the relative severity
of psychosocial stressors,Journal of Psychosomatic Research, 17, 135-139.
CLEGG, C. W . 1983, Psychology of employee lateness, absence, and turnover: a methodological
critique and an empirical study, Journal of Applied Psychology, 68, 88-101.
FOGLER, R. and BELEW, J. 1985, Nonreactive measurement: a focus for research on absenteeism and
Downloaded by [University of Liverpool] at 03:20 06 January 2015