ee

6RFLDO�6XSSRUW�E\�)RXU�&DWHJRULHV�RI�:RUN�&ROOHDJXHV��5HODWLRQVKLSV�%HWZHHQ�$FWLYLWLHV� 6WUHVV�DQG�6DWLVIDFWLRQ $XWKRU�V���0RQLND�+HQGHUVRQ�DQG�0LFKDHO�$UJ\OH 6RXUFH��-RXUQDO�RI�2FFXSDWLRQDO�%HKDYLRXU��9RO�����1R�����-XO����������SS��������� 3XEOLVKHG�E\��:LOH\ 6WDEOH�85/��http://www.jstor.org/stable/3000094 . $FFHVVHG������������������

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

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

.

Wiley is collaborating with JSTOR to digitize, preserve and extend access to Journal of Occupational Behaviour.

http://www.jstor.org

This content downloaded from 129.174.21.5 on Mon, 26 Jan 2015 17:57:28 PM All use subject to JSTOR Terms and Conditions

JOURNAL OF OCCUPATIONAL BEHAVIOUR, Vol.6, 229-239 (1985)

Research note

Social support by four categories of work colleagues: Relationships between activities, stress and satisfaction

MONIKA HENDERSON and MICHAEL ARGYLE Department of Experimental Psychology, University of Oxford, South Parks Road, Oxford OXI 3UD,

England

SUMMARY

An approach to the study of social support at work is presented emphasizing nature and type of social support. One hundred employees in a range of occupations rated the frequency of engaging in 12 types of social and work-related activity with each of four work colleagues varying in degree of intimacy. They also rated levels of job satisfaction and work stress. The results confirm the usefulness of this approach and also suggest that type of work relationship is associated with different sorts of socially supportive activities, and that these affect stress but not satisfaction. These findings were discussed in terms of the interaction between type of support given, source of support, and outcome.

INTRODUCTION

The existence of socially supportive relationships has been shown to affect stressful life events and a variety of physical and psychological illnesses, either directly or by buffering the relationship between stress and health (e.g. Brown and Harris, 1978; Berkman and Syme, 1979; Campbell, Converse and Rodgers, 1976; Colletta, 1979; Gottlieb, 1981; Gore, 1978; Gove, 1972; Holahan and Moos, 1981; Wilcox, 1981). The issue of whether support directly reduces stress and health or whether support buffers the effect of stress on health has been well discussed in a number of writings (e.g. House, 1981). However, less work has been done to examine the nature of social support itself, such as the sorts of activities that comprise support, or the importance of different sources of social support.

There is evidence that source of social support is particularly important. House and Wells (1978) note that the major effect of social support on health derives from work supervisors and spouses, with relatively little effect for work colleagues, while perceived work stresses are correlated with support from both supervisors and co-workers, but not nonwork support. Caplan, Cobb, French, Harrison and Pinneau (1975) found strong associations between support and measures of dissatisfaction, anxiety and depression when this was provided by work supervisors and colleagues rather than by family sources, but no effect for physiological indicators. Reanalysis of this data by La Rocco, House and French (1980) shows that co-workers provide more buffering than supervisors or home support, especially for depression and somatic complaints. 0142-2774/85/030229-11$01.10 Received 11 June 1984 ? 1985 by John Wiley & Sons, Ltd. Revised 13 September 1984

This content downloaded from 129.174.21.5 on Mon, 26 Jan 2015 17:57:28 PM All use subject to JSTOR Terms and Conditions

230 M. Henderson and M. Argyle

In general, results from occupational settings tend to show that job-related stresses and strains are primarily affected by social support at work, although whether the strongest effect derives from colleagues or work supervisors varies, depending upon the occupational group. Indicators of health are less consistently correlated with work-related social support, and may be more dependent on other sources, such as family and friends.

There are two main kinds of activity shared with work colleagues, though the two are closely intertwined-cooperation over work, and primarily social activity. A large proportion of time at work may be spent in interaction with others. Stewart (1967) in a diary-record study of British managers found that one-third to 90 per cent of time at work was spent in interaction with others. Goldthorpe et al.'s (Goldthorpe, Lockwood, Bechhofer and Platt, 1968) study of industrial workers found that 47 per cent of workers spent 'a good deal' of time talking to workmates. Part of this social interchange involves games, comic routines, jokes and teasing, especially in repetitive and monotonous work (e.g. Roy, 1959). Such interchange functions to relieve boredom, and occurs regardless of patterns of the friendship network, although usually there is overlap between those who help each other and those who engage in social activities together (e.g. Roethlisberger and Dickson's study of the Bank Wiring Observation Room, 1939). Friends at work may also provide social support by being confidants, giving advice about stressful work situations, providing information and help with work issues directly, and providing moral support or being allies in difficult interpersonal situations at work.

Work relationships may, however, be major sources of stress in themselves (Kelly, 1982; Marshall and Cooper, 1979; Argyle and Henderson, 1985). Argyle and Furnham (1983), in a study of different relationships, found that work colleagues loaded moderately highly on a general factor of emotional conflict, while work superiors had a high loading, second only to that of marital partners on this factor. Work associates loaded low on the first factor of satisfaction (instrumental reward), and also on the second (social and emotional support). However, the correlation between overall satisfaction and conflict was significant and positive, and higher for work associates and work supervisors than for all other relationships except siblings and opposite-sex friends.

It is not surprising that studies of work relationships report inconsistent findings, since work colleagues may include quite different relationships, ranging from close friendship to active dislike. These distinctions are rarely made in studies of work relationships. A useful categorization, particularly when studying social support, is according to the level of intimacy, and it is hypothesized that these different categories reflect differences in the nature of support and cooperation given. In this study work peers have been divided into four categories:

(1) Social friends: i.e. friends in the normal sense who are known through work and seen at social events outside the work setting. Studies have shown that up to 25 per cent of friends are made in this way.

(2) Friends at work: i.e. friends who interact together over work or socially at work, but who are not invited home or and do not engage in joint leisure activities outside the work setting. Goldthorpe et al. (1968) found that 55 per cent of manual workers reported having at least one close friend at work, but only 28 per cent met these friends outside the work setting.

This content downloaded from 129.174.21.5 on Mon, 26 Jan 2015 17:57:28 PM All use subject to JSTOR Terms and Conditions

Social Support at Work 231

(3) Workmates: i.e. people at work seen simply through formal work contacts and with whom interactions are relatively superficial and task-oriented, and not characterized by either liking or dislike.

(4) Conflict relations: i.e. work colleagues who are actively disliked.

Not only the source but the nature of support given may affect the usefulness of social support in mitigating stress. Cohen and Hoberman (1983) conclude that the major benefits derive from the availability of a confidant and the opportunity for positive self-evaluation by others. Argyle and Henderson (1985) describe six types of benefits available from relationships, including tangible or practical help, informational help, social integration and affirmation. In the work situation, social support may be both piactical/tangible, informational, or socio-emotional. Burke, Weir and Duncan (1976) found that managers were approached by their subordinates primarily with directly or indirectly work-related matters (76 per cent of all problems) compared to other problems such as dealing with personal life or relationships (10 per cent) or physical or emotional health (2 per cent). Kaplan and Cowen (1981) report similar results with industrial foremen, who spent an average of 2.5 hours each week discussing moderate or serious problems of their subordinates. The most common were work-related problems such as job dissatisfaction and difficulties with other employees, but there were also a large number of more personal problems such as marital and sexual difficulties, nervousness and depression, and problems with children, drugs or alcohol. Social support in the occupational setting may therefore be emotional or instrumental, work or nonwork related. However, a few studies in the occupational setting have examined type of social support as a relevant variable.

The aim of the current study is to assess the usefulness of distinguishing type and source of social support when examining the effect of social support on job stress and satisfaction. In this exploratory study, self-report measures were used with a heterogenous sample of employees, and assessment of changes over time based on retrospective self-ratings. A prospective study using more detailed assessment of job stress and satisfaction over time would provide a more rigorous test of specific hypotheses derived from the findings of this preliminary investigation. The following research questions were investigated:

(1) Categories of work relationships determined by degree of intimacy will differ in the nature and frequency of activities engaged in at work. High-intimacy work relationships will involve a higher frequency of activities across a range of activities, both social and work-related, while disliked work colleagues will be avoided in social but not work-related activities.

(2) Job satisfaction is associated with the presence of intimate work colleagues, and can be predicted from the pattern of activities engaged in with these individuals.

(3) Work stress is correlated with the frequency and nature of activities with different categories of work colleagues.

This content downloaded from 129.174.21.5 on Mon, 26 Jan 2015 17:57:28 PM All use subject to JSTOR Terms and Conditions

232 M. Henderson and M. Argyle

METHOD

Subjects

One hundred employed men and women took part. This included 63 males and 37 females, with a mean age of 27.2 (S.D. = 10.2, range = 18 to 60). Subjects were recruited from a large processing firm in Northern England, as well as friends and families of these subjects. Questionnaires were distributed by a male postgraduate student who was also an employee of the firm at this time. He was instructed to recruit a heterogenous sample, selecting across all occupational categories, ages, and both sexes, ensuring that no single occupational group was over-represented in the sample. Due to the preponderance of process workers at this plant, approximately one fifth of the questionnaires were distributed to employed friends or family of other subjects. Approximately one quarter of the total sample fell into each of the occupational categories used (professional 25 per cent, skilled 25 per cent, semi-skilled 22 per cent and unskilled 28 per cent). Females tended to be over-represented in the skilled and males in the semi-skilled occupations, but this was not a significant difference (Z2(3) = 6.90, p > 0.05). The average length of time spent in the current job was 60 months (S.D. = 71.0), with a range of 1 to 372 months.

Questionnaire

Each subject was asked to select a similar status member in the immediate work group, (i.e. not a work superior nor an individual directly supervised) in each of the four categories defined below:

A: People at work who you get on very well with, who you see socially outside of work, and whom you would call good friends. B: People at work who you get on well with and spend time with socially at work (at lunch and coffee breaks) but whom you do not see as friends outside work. C: People at work you get on reasonably well with, but just as a workmate-not someone you spend much social time with either at work or outside work. D: People at work you don't get on very well with.

If there were no members of the immediate work group in a particular category, subjects rated the other target categories only. If the subject worked predominantly alone, only the global work ratings defined below were made.

Subjects rated how often he/she engaged in 12 discrete activities (see Table 2) with each of the four individuals using a 5-point scale of 1 (never or very rarely), 2 (once or twice a week on average), 3 (once or twice a day on average), 4 (several times a day on average) and 5 (nearly all the time at work). In addition, each subject was asked to make four self-ratings:

(i) Current job satisfaction. (ii) Job satisfaction when first starting the current job.

This content downloaded from 129.174.21.5 on Mon, 26 Jan 2015 17:57:28 PM All use subject to JSTOR Terms and Conditions

Social Support at Work 233

(iii) Current feelings of job stress. (iv) Job stress when first starting the current job.

Both satisfaction ratings were made on a 9-point bipolar scale, ranging from 1 ('love it') through 3 ('quite like it'), 5('don't mind it'), 7('don't like it much') to 9 ('hate it'). Stress ratings were both made on a 5-point scale, ranging from 1 (not at all stressful) to 5 (very stressful). Ratings (ii) and (iv) were included primarily to allow the assessment of change over time by subtracting level of current stress and satisfaction from the rated level of initial stress and satisfaction respectively. Although retrospective and therefore subject to bias, change measures based on these ratings should be less susceptible to bias than asking subjects to rate change over time directly.

Each questionnaire was prefaced by a standard introduction stressing confidentiality. Detailed instructions and examples were included. Questionnaires were given out and collected by the same male experimenter who also emphasized that all responses would remain anonymous.

RESULTS

Most subjects report at least one member of the immediate work group for each of the four work categories. Ten per cent of the sample report working predominantly alone, so have no immediate work group member in any category. Sixteen per cent work with one other individual only, and a further 19 per cent with 2 or 3 others.

The effect of having a particular category of work relationship in the immediate work group on self-rated satisfaction and stress was examined by separate analyses of variance using the presence or absence of each work category as the independent variable. Change scores were also calculated by subtracting current satisfaction and stress from initial levels of each variable. These results are presented in Table 1.

The only significant difference occurs for current level of job stress in the person A work category. The presence of at least one high-intimacy work colleague in the immediate work group is associated with significantly less self-rated current stress than is the absence of any such work colleague, but is not related to work satisfaction nor to changes in level of stress or satisfaction over time.

Activity ratings compared across work categories

The frequency of engaging in the specified activities was compared across target persons using Chi-squared. On 7 of the 11 activities, small cell size in rating category 5 resulted in 15 to 40 per cent of the cells with expected frequency less than 5. For these activities, the two highest frequency ratings ('several times a day' and 'nearly all the time') were combined into a single category. These results are summarized in Table 2.

Only one category failed to discriminate significantly between work categories. In the majority of cases, the ordering of frequency is from person A to person D, and this is significant for all pairwise comparisons on total satisfaction scores.

This content downloaded from 129.174.21.5 on Mon, 26 Jan 2015 17:57:28 PM All use subject to JSTOR Terms and Conditions

234 M

. H enderson and M

. A rgyle

T able

1.

M ean

ratings of

w ork

stress

and

satisfaction

according to

presence or

absence of

each

type of

w ork

colleague

M ean

scores

Person A

Person B

Person C

Person D

present

absent

present

absent

present

absent

present

absent

n = 61

n = 39

n = 70

n = 30

n = 62

n = 38

n 54

n 46

Satisfaction: C

urrent

level

3.80

4.05

3.75

4.23

3.97

3.80

3.88

3.92

C hange

-1.00

- 1.53

-1.26

-1.47

- 1.49

- 1.08

-1.35

- 1.30

Stress: C

urrent

level

2.24

3.05*

2.47

2.77

2.45

2.72

2.50

2.62

C hange

-1.10

- 1.33

- 1.14

- 1.00

- 1.34

-0.75

- 1.27

-0.94

*p <

0.01.

T his content dow

nloaded from 129.174.21.5 on M

on, 26 Jan 2015 17:57:28 PM A

ll use subject to JST O

R T

erm s and C

onditions

Social Support at W ork

235

T able

2.

Frequency of

activity

according to

type of

w ork

colleague

R ating

in

w hicli

highest

percentage of

each

w ork

Full

C ollapsed

category

accrues*

A ctivities

df = 12

df = 9 Person A

Person B

Person C

Person D

1.

H elping

each

other

w ith

w ork

_ 70.3c

4.5

3 2 1

2.

D iscussing

w ork

71.3c

- 5 4 2 1

3.

C hatting

casually

93.3c

- 5 4 3 1

4.

H aving

an

argum ent

of

disagreem ent

- 18.8'

1 1 1 1

5.

T eaching

or

show ing

the

other

person

som ething

about

w ork

- 20.2

2 2 1 1

6.

Joking

w ith

the

other

person

112.6c

- 5 4 2 1

7.

T easing

him /her

35.O c

- 5 3 2 1

8.

D iscussing

your

personal

life

- 38.3c

4.5

1 1 1

9.

D iscussing

your

feelings or

em otions

- 42.4c

2 1 1 1

10.

A sking

or

giving

personal

advice

- 58.5c

4.5

1 1 1

11.

H aving

coffee,

drinks, or

m eals

together

65.1c

- 5 3 3 1

12.

C om

m ittee

w ork,

or

sim ilar

discussion

at

w ork

- 11.8

1 1 1 1

* 1 =

N ever

or

very

rarely; 5 =

N early

all

the

tim e;

4.5 =

C ollapsed

rating of 4 + 5.

T his content dow

nloaded from 129.174.21.5 on M

on, 26 Jan 2015 17:57:28 PM A

ll use subject to JST O

R T

erm s and C

onditions

236 M

. H enderson and M

. A rgyle

T able

3.

Significant

regression

equations

predicting

changed

stress

levels

from

frequency of

activities

Predicted

D ependent

Significance of

Item s

w ith

significant

beta

w eights

variable

variables

function

p R2

B eta*

p

Item

C hange

in

job

A ctivities

< 0.01

0.28

1.

H elping

each

other

w ith

w ork

0.48

< 0.01

stress

w ith

person C

3.

C hatting

casually

0.29

< 0.05

6.

Joking

w ith

the

other

person

0.41

< 0.01

7.

T easing

him /her

0.52

< 0.01

8.

D iscussing

personal

life

0.29

< 0.05

10.

A sking

or

giving

personal

advice

-.44

< 0.01

A ctivities

<0.10 > 0.05

0.21

2.

D iscussing

w ork

0.50

< 0.01

w ith

person D

3.

C hatting

casually

-.26

< 0.05

7.

T easing

him /her

.30

< 0.01

8.

D iscussing

personal

life

-.35

< 0.01

*Positive

betas

indicate

that

increased

frequency of

activities is

associated

w ith

reduced

stress.

N egative

betas

indicate the

opposite

relationship.

T his content dow

nloaded from 129.174.21.5 on M

on, 26 Jan 2015 17:57:28 PM A

ll use subject to JST O

R T

erm s and C

onditions

Social Support at Work 237

Predicting job satisfaction and stress

Current levels of job satisfaction and job stress were separately regressed onto each of the sets of activity ratings associated with each category of work relationships. Ratings were treated as interval data for this analysis. None of the four equations for predictingcurrent level satisfaction reached significance, and only activities with person A significantly predicted current levels of job stress at p = 0.05. Only two activities showed significant beta weights-joking (-62, p < 0.01) and discussing feelings or emotions (+41, p < 0.01). Apparently, more frequent discussion of personal feelings with high-intimacy work colleagues predicts more stress, while increased frequency of joking with this category of work relationship is associated with lower levels of job stress. Change scores were separately regressed onto activity ratings for each work category. Change in satisfaction was not significantly predicted by any function. One of the four equations for change in stress was significant, while a second showed a trend at p < 0.10 > 0.05. Both tended to predict only a small proportion of the variance. Details of these are given in Table 3.

Increased frequency of mutual help with work, chatting casually, joking, teasing and discussing personal life with person C predicted a reduction in stress over time. Higher frequency of asking or giving personal advice with this low-intimacy colleague is associated with an increase in stress. For person D (disliked colleague), higher frequency of discussing work predicts reduced stress but an increase in the frequency of chatting casually, teasing, or discussing personal life with this category of work colleague is associated with increased levels of job stress over time. The association of job stress and activities therefore operates in different ways for the two categories. High frequency of engaging in both work and social (but not highly intimate) activities with person C is associated with reduced stress, while for person D, high frequency of work-related activities and low incidence of social activity is related to reduced stress.

To summarize, the reported frequencies of social and work-related activities differ according to type of work colleague, as predicted. However, this does not appear to affect job satisfaction in any way. Work stress, on the other hand, seems to relate to the frequency and nature of such activities, although not in a consistent manner: current stress levels are significantly predicted by certain activities with high-intimacy colleagues while change in stress over time appears to be linked only to interaction with low-intimacy and disliked colleagues.

DISCUSSION

The results of this study have confirmed the notion that different types of work relationships, as assessed by level of intimacy, are associated with different types and frequency of interaction and exchange, both in work-related and social activities. In addition, there is some evidence that the various types of work relationship are differentially effective as sources of social support.

The existence of at least one high-intimacy colleague in the immediate work group is associated with lower stress. This is consistent with studies showing that the effects of social support are often a function of a single source rather than the

This content downloaded from 129.174.21.5 on Mon, 26 Jan 2015 17:57:28 PM All use subject to JSTOR Terms and Conditions

238 M. Henderson and M. Argyle

cumulative effect of a number of sources (House, 1981). The nature of social suppott given by such colleagues also significantly predicts current stress, and is the only relationship to do so. However, only two activities significantly contribute-joking and discussing personal feelings, and these operate in opposite directions with more joking in low stress and more frequent discussion of personal problems in high stress situations. Interaction with high-intimacy colleagues may therefore reflect coping mechanisms, or ways of dealing with high stress.

However, reductions in level of stress over time are associated more with the sorts of interaction carried out with low-intimacy and disliked colleagues. Frequency of asking or giving personal advice is in fact associated with inversed stress. It appears that overall, friendly but not intimate interactions with low-intimacy relationships at work are linked to a reduction in stress.

Type of social support also appears to operate differently on stress. Cobb (1976) describes three types of social support-emotional, esteem, and network. Others (e.g. House, 1981; Gottlieb, 1981) distinguish between emotional and instrumental support. Many of the activities employed in the current study involve instrumental help. Activities such as asking or giving personal advice (10) or discussing feelings or emotions (9) comprise emotional support, while network support would include such low-intimacy social activities as chatting casually (3) or having coffee, drinks or meals together (11). It is these elements of network support that appear relevant in stress reduction (at least with low-intimacy colleagues) while emotional support is associated more with high stress situations. It is possible that such support is used primarily as a coping mechanism, while instrumental and network support provide those aspects of social support that operate directly on stress in the work setting.

It has been noted earlier in this paper that job-related stresses are primarily affected by sources of social support at work, while health appears to be affected more by nonwork sources such as family and friends. In the current study high-intimacy work colleagues are defined as friends who are also seen outside work. Social support of this kind may therefore be more relevant in providing the sort of emotional support that affects general health and well-being, but not work-related stresses and strains. However, the instrumental and network support offered by low-intimacy work colleagues directly affects work-related stress, but may not have the general effect on health and well-being provided by other sources of social support.

Further studies are needed to examine the relationship between type and source of social support and psychological and physiological measures of stress and health. The approach used in the current study, which distinguishes between type and source of social support, is potentially useful in disentangling some of the conflicting findings obtained in studying the effects of social support on job-related satisfaction, stress, and health.

REFERENCES

Argyle, M. and Furnham, A. (1983). 'Sources of satisfaction and conflict in long term relationships', J. Marriage and the Family, 45, 481-931.

Argyle, M. and Henderson, M. (1985). The Anatomy of Relationships, Heinemann, London.

Numbers in parentheses refer to Activities listed in Table 2.

This content downloaded from 129.174.21.5 on Mon, 26 Jan 2015 17:57:28 PM All use subject to JSTOR Terms and Conditions

Social Support at Work 239

Berkman, L. and Syme, S. (1979). 'Social networks, host resistance and mortality: A nine year follow-up study of Alameda county residents', American Journal of Epidemiology, 109, 186-204.

Brown, G. and Harris, T. (1978). Social Origins of Depression, Tavistock, London. Burke, R., Weir, T. and Duncan, G. (1976). 'Informal helping processes in work settings',

Academic Management, 19, 370-37. Campbell, A., Converse, P. and Rodgers, W. (1976). The Quality of American Life, Sage,

New York. Caplan, R., Cobb, J., French, R., Harrison, R. and Pinneau, S. (1975). Job Demands and

Worker Health, U.S. Dept of Health, Education and Welfare publications. Cobb, S. (1976). 'Social support as a moderator of life stress', Psychosomatic Medicine, 38,

300-314. Cohen, S. and Hoberman, H. (1983). 'Positive events and social support as buffers on life

change stress: maximising the prediction of health outcomes'. Unpublished manuscript, University of Oregon.

Colletta, N. (1979). 'Support systems after divorce: incidence and impact', Journal of Marriage and the Family, 41, 837-846.

Goldthorpe, J., Lockwood, D., Bechhofer, F. and Platt, J. (1968). The Affluent Worker, Cumbridge University Press, Cambridge.

Gore, S. (1978). 'The effect of social support in moderating the health consequences of unemployment', Journal of Health and Social Behavior, 19, 157-165.

Gottlieb, B. (1981). Social Networks and Social Support, Sage, Beverly Hills. Gove, WV. (1972). 'The relationship between sex roles, martial status and mental illness',

Social Forces, 51, 34-44. Holahan, C. and Moos, R. (1981). 'Social support and psychological distress: a longitudinal

analysis', Journal of Abnormal Psychology, 90, 365-370. Homans, G. (1950). The Human Group, Routledge and Kegan Paul, London. House, J. (1981). Work Stress and Social Support, Addison-Wesley, Reading, Mass. House, J. and Wells, J. (1978). 'Occupational stress, social support and health'. In: McLean,

A., Black, G. and Colligan, M. (Eds) Reading Occupational Stress: Proceedings of a conference, DHEW (NIOSH) publication 78-140, 8-29. Reported in House, 1981.

Kaplan, E. and Cowen, E. (1981). 'Interpersonal helping behavior of industrial foremen', Journal of Applied Psychology, 66, 633-638.

Kelly, M. (1982). 'Stress among blue collar workers'. In: Kakabadse, A. (Ed.) People and Organization, Gower, Aldershot.

La Rocco, J., House, J. and French, J. (1980). 'Social support, occupational stress, and health', Journal of Health and Social Behavior, 21, 202-218.

Mann, F. and Baumgartel, H. (1953). Absences and Employee Attitudes in an Electric Power Company, Institute for Social Research, Ann Arbor, Michigan.

Marshall, J. and Cooper, C. (1979). Executives under Stress, Macmillan, London. Roethlisberger, F. and Dickson, W. (1939). Management and the Worker, Harvard

University Press, Cambridge, Mass. Roy, D. (1959). 'Banana time: job satisfaction and informal interaction', Human

Organization, 18, 158-168. Stewart, R. (1967). Managers and their Jobs, Macmillan, London. Wilcox, B. (1981). 'Social support in adjusting to marital disruption: A network analysis'. In:

Gottlieb, B. (Ed.) Social Networks and Social Support, Sage, Beverly Hills.

This content downloaded from 129.174.21.5 on Mon, 26 Jan 2015 17:57:28 PM All use subject to JSTOR Terms and Conditions

1. Cit r113_c113:1:
2. Cit r117_c117:1:
3. Cit r122_c122:1:
4. Cit r124_c124:1: