The need for work: Jahoda's latent functions of employment in a representative sample of the German population

Abstract

Marie Jahoda's latent deprivation model was tested with a representative sample of the German population (N = 998). As expected, employees reported high levels of time structure, social contact, collective purpose, and activity not only in comparison to unemployed persons but also in comparison to persons who are out of the labor force (i.e., students, homemakers, retirees). Even unskilled manual workers reported more access to these “latent functions” than persons without employment. For the fifth of Jahoda's dimensions, identity/status, no significant differences between employed persons and persons who are out of the labor force could be identified. However, unemployed persons reported less status than all other groups did. Thus, Jahoda's model was clearly endorsed for four of the five latent functions of employment and partly endorsed for the fifth function. All variables in the model correlated significantly with distress, as expected. Demographic correlates of the manifest and latent functions were also analyzed: Access to the latent functions was best among young men from higher social classes who lived in an intimate relationship in a comparatively large household with children. Copyright © 2009 John Wiley & Sons, Ltd.

Full text

Running head: The need for work: Latent functions of employment
The need for work: Jahoda’s latent functions of employment in a
representative sample of the German population
Karsten Paul & Bernad Batinic
Preprint
Paul, K. I. & Batinic, B. (2010). The need for work: Jahoda’s latent functions of employment
in a representative sample of the German population. Journal of Organizational Behavior, 31,
45-64. DOI: 10.1002/job.622

2
Abstract
Marie Jahoda’s latent deprivation model was tested with a representative sample of the
German population (N=998). As expected, employees reported high levels of time structure,
social contact, collective purpose and activity not only in comparison to unemployed persons
but also in comparison to persons who are out of the labor force (i.e. students, homemakers,
retirees). Even unskilled manual workers reported more access to these “latent functions” than
persons without employment. For the fifth of Jahoda’s dimensions, identity/status, no
significant differences between employed persons and persons who are out of the labor force
could be identified. However, unemployed persons reported less status than all other groups
did. Thus, Jahoda’s model was clearly endorsed for four of the five latent functions of
employment and partly endorsed for the fifth function. All variables in the model correlated
significantly with distress, as expected. Demographic correlates were also analyzed: Access to
the latent functions was best among young males from higher social classes who lived in an
intimate relationship in a comparatively large household with children.

3
Introduction
Marie Jahoda’s (1981, 1982, 1997) latent deprivation model states that employment has not
only the manifest function of earning a living but also five unintended or “latent” functions.
These latent functions correspond with basic human needs and help to sustain well being and
mental health. Although it was originally developed in the context of psychological
unemployment research and is usually cited in this context, Jahoda’s model can also be seen
as a general theory concerning the meaning of work to mental health that implies a very
positive view of employment. According to Jahoda, employed people should not only have
more access to the latent functions than unemployed people have, but also more access than
all other people who are not employed, e.g. homemakers, students, and retirees. Thus, in
Jahoda’s eyes, employment is an important means for mental health and satisfaction in
contemporary societies. The purpose of the present paper is to test these assumptions with a
representative sample of the German population.
Manifest and latent functions of employment and mental health
Based on her renowned research in Marienthal concerning the mental health effects of
unemployment, Marie Jahoda specified five aspects of employment that unemployed people
usually lack (Jahoda, Lazarsfeld, & Zeisel, 1933/1975; Jahoda, 1981, 1982, 1997). These
aspects are (1) time structure, (2) collective purpose, (3) social contact, (4) status, and (5)
activity. In her opinion, lack of these aspects is an important cause of the distress unemployed
people often experience (for recent meta-analyses on the mental health effects of
unemployment see Murphy & Athanasou, 1999; McKee-Ryan, Song, Wanberg, & Kinicki,
2005). Using Merton’s (1957) differentiation between manifest and latent functions of social
institutions she called these aspects the “latent functions” of employment (1981, p.189). They

4
are assumed to be unintended by-products of employment, as opposed to its “manifest”, i.e.
obvious and primary functions of earning a living and the production of goods.
According to Jahoda, the experience of time is shaped by social institutions such as school
and employment in modern societies. As a result, adults in such societies are in need of clear
time structures and the possibility to fill their days with planned activities. The lack of such a
structure impairs a person’s well-being: ”Days stretch long when there is nothing that has to
be done; boredom and waste of time become the rule” (Jahoda, 1982, p.22). According to
Jahoda people also cannot do without what she called collective purpose, i.e. the feeling of
being useful, of being needed by other people, the experience “that the purposes of a
collectivity transcend the purposes of an individual“ (1982, p.24). If this need for a meaning
in life is deprived, “a sense of purposelessness” (1982, p.24) takes place, resulting in distress.
Related to this is the need for “regularly shared experiences and contacts with people outside
the nuclear family” (1982, p.188). Such social contacts outside the family cannot be replaced
simply by intensifying family life, because they “provide more information, more opportunity
for judgement and rational appraisal of other human beings with their various foibles,
opinions, and ways of life” (1982, p.26). Thus, “enlarging the social horizon” (1982, p.25) is
also important to mental health according to Jahoda. Furthermore, the social status of a
person - largely resulting from the value system of the society he or she lives in – is essential
for the construction of one’s identity. People tend to see themselves in a similar way as others
see them, and even a low status - for example as a manual worker - is better than having no
status at all, which is the way unemployed people often experience their situation. Finally, the
level of activity is the fifth important determinant of a person’s mental health in Jahoda’s
model. Being active, even being active due to external forces such as the need to earning a
living, is better for a person’s psychological well-being than being passive. In sum, the five
latent functions are supposed to correspond to basic human needs which are “deep seated and
widespread”, i.e. important and ubiquitous (1982, p.85). Therefore, satisfaction or

5
dissatisfaction of these needs is expected to exert strong effects on psychological well-being.
Furthermore, employment is the only sufficient source of the latent functions according to
Jahoda. Not only unemployed persons, but also other persons who are out of employment
should experience a lack of the latent functions. She explicitly mentioned “housewives” and
“retired people” in this context (1981, p.189). She also stated that individuals are sometimes
able to find access to the latent functions from other sources than employment, but only to a
limited degree. Religion, for example, political activity, or leisure activities may provide some
of the functions, but not all of them and not in an intensity that is comparable to employment.
Therefore, all persons who are out of employment are likely to experience deprivation of
these latent functions and to suffer a diminished mental health. Jahoda speculates that even
jobs with unfavourable working conditions – as long as they are not completely dehumanising
- may be better than having no job at all: “Employment is psychologically supportive, even
when conditions are bad” (Jahoda, 1981, p.188). In sum, Jahoda had a remarkably positive
view of work and employment, an opinion that stimulated intense theoretical controversies
(see Fryer, 1986, 1992), as well as empirical studies that reporting evidence in conflict with
her contentious statement about the preferability of bad employment over unemployment
(O’Brien & Feather, 1990; Winefield, Winefield, Tiggemann, & Goldney, 1991). Whether there is
direct empirical evidence for the different aspects of her latent deprivation model will be
reviewed below.
Is access to the latent function necessary for mental health?
Several studies from Australia (e.g. Creed & Macintytre, 2001, Creed & Reynolds, 2001;
Feather & Bond, 1983; Hoare & Machin, 2006; Muller, Creed, Waters, & Machin, 2005;
Waters & Moore, 2002a), the United Kingdom (Evans & Banks, 1992; Haworth & Paterson,
1995; Haworth, & Evans, 1987; Kilpatrick & Trew, 1985; Miles, 1983), and the USA (Brief,

6
Konovsky, Goodwin, & Link, 1995; Wanberg, Griffiths, & Gavin, 1997) have demonstrated
that the amount of access to the latent functions is correlated with various measures of mental
health among employed and unemployed persons, as predicted by Jahoda’s model. However,
apart from a few studies involving students (e.g. Evans & Banks, 1992; Creed & Evans,
2002), the existing studies were usually restricted to members of the labor force. This is
understandable because Jahoda’s model was developed in the context of unemployment
research. Yet, she explicitly stated that all members of modern societies are characterized by
“deep-seated” needs for the five latent functions (Jahoda, 1982, p.85). Thus, it would be
interesting to learn whether the association between the access to the latent functions and
mental health that was found among members of the labor force can also be found in the
general population, i.e. whether the need for the latent functions is really as ubiquitous as it is
implied in Jahoda’s theory. Furthermore, the model has been tested almost exclusively in
English-speaking countries up to now (for a Swedish exception from this rule see Isaksson,
1989). Thus, it would be a step towards a higher generalizability of the model to test the
correlation between latent functions and mental health in a non-anglo-saxonian cultural
context such as Germany. As can be seen below, knowledge about the validity of Jahoda’s
model is also of practical importance for German organizations, which will soon be
confronted with labor shortage.
Context
Despite the nearly complete breakdown of East-Germany’s economy and the considerable
problems of the reunification process following the seminal year of 1989, Germany is still an
economically successful country as measured by the gross national income per capita
(Worldbank, 2008). Traditionally, Germany is also characterized by a high level of social
protection. This can be seen in its position among the top ten in a list of 74 countries ranked

7
according to their expenditure for social security and welfare as a share of the gross domestic
product (Besley, Burgess & Rasul, 2003). Unemployment protection, for example, is
generous in terms of number of entitled persons as well as level and duration of financial
support (International Labour Office, 2000). Another important aspect of Germany’s current
situation pertains to the demographic structure of the population. This structure is rapidly
changing due to a low fertility rate of below 1,5 Children per woman (Bundesinstitut für
Bevölkerungsforschung, 2008). The process of first reduction and then stagnation of German
fertility rates is not a new phenomenon but started more than four decades in the past. As a
consequence, the last generations born in times of high fertility will soon reach retirement
age, to be replaced by cohorts with considerably fewer members. Thus, a severe labor
shortage is predicted, beginning already in 2015 (Fuchs & Dörfler, 2005; Fuchs & Söhnlein,
2005). This means that very soon German organizations will be forced to compete not only
with each other, but also with the generous social security system for the few remaining able
workers, rendering the problem of how persons can be motivated to take up employment an
important research question.
Is employment the main source of the five latent functions?
Some empirical evidence concerning the question of whether the latent functions of
employment are more prevalent among employed persons than among unemployed persons
has already been published. For example, recent Australian studies found that employed
persons reported better access to the latent functions than unemployed persons did (Creed &
Reynolds, 2001; Waters & Moore, 2002a). A British study found significantly higher levels of
activity, status, time structure, and social contact among the employed compared to the
unemployed, but found no significant differences for collective purpose (Miles, 1983). A
Swedish study found significantly higher levels of time structure, activity, and collective

8
purpose among employed single men in comparison to unemployed single men, but no
differences for social contact and status were identified (Isaksson, 1989). In sum, the
available empirical evidence supports the assumption that employed people have more access
to latent functions than unemployed people have, particularly time structure and activity. Yet,
some doubts remain for social contact, identity/status, and collective purpose.
Do the aforementioned empirical results help to validate the central part of Jahoda’s theory
that states that time structure, activity, social contact, collective purpose and identity/status
are “latent functions of employment”? We don’t think so, because the aforementioned studies
compared employed persons only with unemployed persons. However, unemployment is a
very special life situation with unique characteristics. For example, unemployed people suffer
more stigmatization than other social groups do (McFayden, 1995), possibly reducing their
sense of status and collective purpose. Uncertainty concerning the future has also been
described as a special characteristic of unemployment (Fryer, 1997), possibly influencing the
perception of time structure. These specific characteristics may not be shared by other groups
of persons who do not work, e.g. homemakers and students. Thus, in order to test whether
employment really is the single best provider of the five latent functions, the comparison of
employed persons with unemployed persons should be augmented by comparisons with
persons who are out of the labor force (OLF).
To our knowledge, a comparison with OLF-persons was done in only three studies up to now,
with mixed results: Creed, Muller, and Patton (2003) used an overall measure of the latent
functions. They found a significant difference between adolescents who were employed full-
time and adolescents who were studying nine months after high-school. The two other studies
measured each of the latent functions separately, giving a more detailed picture of the
differences between employed persons, unemployed persons, and OLF-persons. Henwood
and Miles (1987) examined a small sample of British persons from different social groups

9
with single item-measures. Similar to other studies, they found significantly higher levels
among employed persons in comparison to unemployed persons for all five dimensions.
Comparing employed women with housewives, the authors found significant differences only
for activity and social contacts, favouring the employees in both cases. Furthermore,
employed people scored significantly higher than retirees on social contact, collective
purpose, and time structure. In another study, Evans and Banks (1992) examined young
people from three British cities. Their results showed strong differences between unemployed
individuals and all other groups with regard to four of the five latent functions. The
unemployed persons always had the lowest scores. However, with only one exception the
results showed no differences between employed individuals and OLF-individuals. Only with
regard to collective purpose had employed youths significantly higher scores than those in
full-time education. In sum, the results from these studies appear to be inconclusive. Only in a
minority of comparisons did people out of the labor force differ from employees with regard
to their access to the latent functions of employment. This is an unexpected result, as they do
not work and should thus score lower on these dimensions according to Jahoda’s model.
The aforementioned studies were restricted to comparatively small sample sizes and in two
cases to a specific age group (youths). One study used single-item measures with unknown
reliability. Another study used only an overall measure of latent deprivation, prohibiting a
detailed analysis. Thus, more data concerning this research question are needed. Using a
representative sample of the general population would help to secure the generalizability of
the results. This is important because the great majority of samples that were used to test
aspects of Jahoda’s model have been convenience samples, leaving some doubts whether the
results really are valid for a country’s population as a whole.

10
Demographic correlates of the latent functions
Jahoda’s theory includes the notion that it is necessary for human beings to satisfy five basic
needs in order to live a psychologically healthy life. This is a general assumption with far-
reaching implications that might be interesting beyond the unemployment context. Knowing
which social groups have more or less access to the latent functions could help to identity
groups of persons who are at risk for elevated distress. This is a question that has not received
much scrutiny up to now. Underlid (1996) found that the proportion of time filled with
various activities was larger among cohabiting people than among singles. She also found a
significant negative association between socializing and age in her sample of unemployed
Norwegians. No associations with gender were found. Creed and Macintyre (2001) also found
a significant negative correlation between age and social contact. They found no effects of age
on the other latent functions specified in Jahoda’s model. They also found no gender
differences with regard to these functions. Muller, Creed, Waters and Machin (2005) found
weak but significant negative correlations between age and social contact and status,
respectively. No significant correlations with gender were detected. Finally, Creed and
Watson (2003) found that younger unemployed persons had more social contacts, higher
status but less time structure than older age groups. Again, no significant main effects for
gender were identified.
In sum, age appears to be more strongly associated to Jahoda’s latent functions than gender or
marital status. Especially, the positive effect of young age on the frequency of social contacts
was replicated several times. Yet, many combinations between latent functions and
demographic variables have not been studied at all up to now and those studies that were done
did not lead to unequivocal results. Furthermore, all studies have been limited to samples of
unemployed persons, leaving open the question of whether the associations found are unusual

11
and can be seen as distinctive features of the unemployment situation or whether they simply
parallel associations that can also be found in the general population.
Aims of the study
The aim of the present paper is (1) to test whether employed persons have significantly better
access to the latent functions as measured by the ACE-scales not only in comparison to
unemployed persons but also in comparison to persons who are out of the labor force; (2) to
test whether there exist significant and meaningful correlations between psychological health
and Jahoda’s manifest and latent functions in a representative sample of the German
population; (3) to identify demographic correlates of Jahoda’s manifest and latent functions in
the general population.
Method
Participants
Participants were 998 Germans from all parts of the country, including Eastern Germany,
52% of whom were female; the age range was 14 - 91 (M = 47.5). The sample was
representative of the German population with regard to age, sex, occupation, household size,
geographic region, and size of village/town. The data were collected in the course of a
regularly conducted multiple-topic-survey (“Classic-Bus”) in cooperation with a well-known
German market research company (GfK AG). The sampling method applied was multi level
quota-sampling. The computer-assisted face-to-face interviews were conducted in November
2005 in the homes of the participants by professional interviewers.

12
Measures
All scales that were used in the present investigation were translated into German with the
translation-backtranslation method and the help of a native English speaker.
Latent functions: Due to the high financial costs of conducting a representative population
survey, it was an absolute necessity to use scales that were as short as possible for the present
study. Therefore, we chose Evans’ (1986) “access to categories of experience” (ACE) scales
as measure of access to the latent functions of employment. These scales consist of 15 items,
3 for each dimension. The instrument is based on items generated by Miles (1983) and was
explicitly constructed with the aim of assessing Jahoda’s five latent functions. Responses
were scored on a seven-point scale, using end points of “completely agree” to “completely
disagree”.
Although several studies have successfully used the ACE-scales in the past (e.g. Water &
Moore, 2002ab), the factorial structure has recently been disputed. One study found support
for the intended five-factor structure (Evans & Banks, 1992), but another study was less
successful (Creed & Machin, 2003). Therefore, a confirmatory factor analysis was run on the
five ACE-subscales with data generated from the present study. First, a visual inspection of
the intercorrelation matrix showed that three items did not correlate with the other items in the
expected way. Therefore, these three items were excluded. Then the confirmatory factor
analysis was run with the remaining twelve items. The results supported the expected five-
factor structure of these items: Chi2(44) = 277,36***, IFI = .97, CFI = .97, RMSEA = .07,
SRMR = .06.
The final scale for time structure consisted of two items (”People often rely on me to turn up
at the right time”, “I very rarely ever need to be punctual”, second item reverse scored).
Internal reliability of this scale was alpha = .54. The scale for social contact consisted of three
items (”Most days I meet quite a range of people”, “I see a lot of my friends and workmates”,
“I don’t get to meet many people regularly”, last item reverse scored). Reliability was alpha =

13
.69 for this scale. The final scale for collective purpose consisted of two items (”At this time
in my life I feel I’m making a positive contribution to society at large”, “I’m dong things that
need doing by someone”), alpha = .51. The scale for status also consisted of two items
(“Sometimes I feel like I’m on the scrapheap”, “I sometimes feel that people are looking
down on me”, both items reverse scored). Internal reliability of the status scale was alpha =
.68. The scale for activity consisted of three Items (”My time is filled with things to do”,
“Things I have to do keep me busy most of the day”, “Time often lies heavy on my hands”,
last item reverse scored). Internal consistency of the activity scale was alpha = .79.
Manifest function: The manifest function of employment was operationalized with three items
from a scale for financial strain used by Creed and Macintyre (2001) and Ullah (1990) (“Are
you often not able to do the things you like to do because of shortages of money?”, “Are you
often not able to do the things you need to do because of shortages of money?”, “Do you have
serious financial worries?”). Again the responses were scored on a seven-point scale from
“completely agree” to “completely disagree”. High scores indicate a good financial situation.
The internal consistency for the present sample was alpha = .88. The correlation with
household income was r = 0.38 (p < .001), supporting the validity of this scale.
Validation study: Since the reliabilities of the ACE-scales were less than satisfactory, possibly
instigating doubts concerning the validity of the scales used here, we conducted a web-based
validation study. The data-base was a large (n=1069) convenience sample of Germans that
consisted primarily of employees (76,9%) and was not representative to the German
population. Therefore, this sample was not used to test our main hypotheses but served only
as basis for our validation study.
We used the Latent and Manifest Benefits (LAMB) scales developed by Muller, Creed,
Waters, and Machin (2005) in order to validate the ACE-scales and the financial-strain items
taken from Creed and Macintyre (2001). The LAMB-scales, like the ACE-scales, were
specifically constructed to measure Jahoda’s manifest and latent functions of employment.

14
They consist of six items per function. However, we excluded three items from the status-
subscale, because these items are exclusively concerned with helping behavior, a rather
indirect and problematic operationalization of status in our opinion.
The results of the validation study were as follows: Convergent validity was large or very
large for the manifest function (r = .81, p < .001), for social contact (r = .61, p < .001), for
activity (r = .68, p < .001) and for collective purpose (r = .57, p < .001). With regard these
functions, we also found unequivocal evidence for discriminant validity, since with only one
exception the convergent validities were always larger than the correlations between the
respective function of employment and any of the other functions, whether they were
measured with items from the same questionnaire (heterotrait-monomethod correlations) or
with items from the other questionnaire (heterotrait-heteromethod correlations; Campbell &
Fiske, 1959).
For status, convergent validity was of medium-to-large size (r = .40, p < .001). In 16 out of 20
comparisons this correlation was larger than the correlation between status and any of the
other functions. Thus, there was also some evidence for discriminant validity. Overall, the
evidence endorsing the validity of the status/identity measures was weaker than for the
aforementioned functions of employment. This might have been the result of differences
regarding the reference group used in the LAMB- and ACE-scales for status: All of the
LAMB-items refer to a person’s closest social network (“friends” and “people around me”),
while the ACE-items are more general, referring to the “society” or simply universally to
“people”. In our eyes, the operationalization of the ACE-scale is closer to Jahoda’s original
intention.
For time structure, convergent validity was only of medium size (r = .29, p < .001). However,
in 18 out of 20 comparisons this medium-size correlation was larger than the correlation
between time structure and any of the other functions. Thus, evidence for discriminant
validity was good.

15
In sum, this validation study endorsed the validity of the ACE-scales and the financial-strain
items taken from Creed and Macintyre (2001) as measures for Jahoda’s manifest and latent
functions of employment. Support was very good for the manifest function, for social contact,
for activity, and for collective purpose and at least mixed for status/identity and time
structure.
Psychological well-being/distress: Psychological well being was measured with four items
from the German version (Linden, Maier, Achberger, Herr, Helmchen, & Benkert, 1996) of
the General Health Questionnaire (Goldberg & Hillier, 1979). We selected those four items
that showed the highest part-whole correlations during our pre-tests (“Have you recently been
feeling unhappy and depressed?” “Have you recently felt you couldn't overcome your
difficulties? “Have you recently been losing confidence in yourself?” “Have you recently
been feeling reasonably happy, all things considered?”). The internal consistency of this
short-form of the GHQ for the present sample was alpha = .80.
Employment status: Employment status was measured with a single question asking
participants to categorize themselves in one of seven groups. Three groups had at least some
access to employment: (1) Full time employees, (2) part time employees, and (3) young
persons in vocational training/apprentices. Members of four other groups had no or only very
limited access to employment: (4) pupils and students, (5) homemakers, (6) retirees, and (7)
unemployed people.
Demographics: Occupation level, income, level of education, age, sex, marital status, number
of children in the household, size of village/town, and whether the respondent lived in Eastern
or Western Germany were measured with standard survey questions.
Results
As expected, significant differences between employed persons and unemployed persons
could be identified for all the variables specified in Jahoda’s model. For most of these

16
variables significant differences between employed persons and OLF-persons were also
found. A finer-grained analysis comparing the seven employment-status subgroups also
generated interesting results.
Persons with access to employment showed larger mean scores for time structure than OLF-
persons and unemployed persons. With d = 1.02 and d = 0.95 these differences were large and
highly significant. Analyses conducted on the more specific level of the seven subgroups
showed that the full-time and part-time employees reported significantly more access to time
structure than homemakers, retirees, and unemployed persons. Full-time employed persons
also reported significantly more time structure than students (see table 1, note that sharing the
same subscript indicates a significant mean difference in this table; e.g. full-time employees
and students share an “a” in the first line, meaning that they differ with regard to time
structure). The small group of persons in vocational training reported significantly larger
scores than retired and unemployed persons. Furthermore, students had significantly larger
scores than retirees. In sum, the results for time structure clearly endorsed Jahoda’s theory.
Employed persons reported significantly more access to social contact than OLF-persons and
unemployed person. With d = 0.66 the difference was of medium-to-large size for OLF-
persons, and with d = 0.96 it was large for unemployed persons. At the more specific level of
the seven employment status groups an age effect could be detected that influenced the results
of this analysis: The two younger groups (apprentices and students) had particularly high
scores significantly different from all other groups except full-time employees. Among the
adults, full time employees reported significantly more social contact than homemakers,
retired persons, and unemployed persons, while part-time employees reported higher scores
than retirees and unemployed individuals. Thus, Jahoda’s model again was supported,
although age obviously had a slightly biasing influence on the results for this variable.
The pattern of results for collective purpose again endorsed Jahoda’s model: The difference
between employed persons and OLF-persons was highly significant and of medium-to-large

17
size (d = 0.65). With d = 0.82 the difference between employed persons and unemployed
persons was also significant and of large size. The analysis on the subgroup level showed that
full-time employees experienced significantly more collective purpose than all of the non-
employed groups, i.e. students, homemakers, retirees, and the unemployed. Part-time
employees experienced significantly more collective purpose than retired and unemployed
persons. Young persons in vocational training experienced significantly more collective
purpose than unemployed people. In sum, employment was clearly associated with the
amount of access to this latent function.
For activity, large differences were found between employed persons and OLF-persons (d =
0.95) as well as between employed persons and unemployed persons (d = 1.17), again
favoring employed persons. At the subgroup level, the largest mean score was found for full-
time employed persons. This group was significantly busier than students, homemakers,
retirees, and unemployed persons. Part-time employed respondents reported significantly
more activity than retirees and unemployed persons. Youths in vocational training also
reported more activity than retirees. We also found differences between the non-employed
groups: Retired people reported particularly low scores of activity, lower than all other groups
except for the unemployed. Unemployed persons reported lower scores than homemakers. In
sum, the results again endorsed Jahoda’s model with clearly higher scores of activity among
employed persons compared to OLF-persons and unemployed persons. Retirement and
unemployment seem to be associated with a particularly passive lifestyle.
For Identity/status the results were less clear than expected. Though employed persons
reported considerably more status than unemployed persons did (d = 0.84), there was no
significant difference between employed persons and OLF-persons. (d = 0.09). On the level of
the seven subgroups, unemployed individuals reported significantly less status than all other
groups. However, there were no significant differences between employee groups and OLF-
groups. Thus, there was only weak support for Jahoda’s model with regard to the status-

18
variable. Unemployment obviously plays a very specific role here, distinct from being out of
the labor force. A deprivation of status is only associated with being unemployed, but not
with being a homemaker, a student, or a retiree.
With regard to the manifest function, i.e. the perceived financial situation, a large difference
between employed und unemployed persons emerged with employed persons having a much
better financial situation than unemployed persons, as expected (d = 1.07). However, there
was no significant difference between employees and OLF-persons (d = -0.06). At the
subgroup-level very low scores were found for students and for the unemployed. Both groups
differed significantly from the other groups except the small sample of youths in vocational
training. The most favorable mean scores were found for homemakers and retired persons.
Thus, for the financial situation no clear differentiation between employed and OLF-groups
could be detected.
For distress we found a significant difference between employed persons and unemployed
persons of medium size (d = -0.58). However, with d = -0.11 the difference between
employed person and OLF-persons was small and only marginally significant. On the
subgroup level, only one significant difference emerged, with unemployed persons reporting
significantly more distress than employed persons.
We conclude that for four of the five latent functions our results clearly endorsed Jahoda’s
model: Employed persons reported more access to the latent functions of time structure,
collective purpose, social contact, and activity, than unemployed persons and also more than
persons out of the labor force. For the identity/status-dimension the model was only partly
endorsed: Employees reported more perceived status than the unemployed, but there was no
difference between employees and people who were out of the labor force.
For the manifest function no differences between employed persons and OLF-persons
emerged, which is an unexpected result. Furthermore, we found only a small difference
between employed persons and OLF-persons with regard to distress. This finding is not in

19
good agreement with Jahoda’s model, as the low scores with regard to the latent functions that
were found among the OLF-persons are supposed to lead to elevated levels of distress among
this group.
The seven employment status groups analyzed here differed considerably with regard to their
demographic characteristics. Homemakers, for example were nearly exclusively female, while
the majority of full-time employees was male. Furthermore, several demographic
characteristics were significantly related to the latent and manifest functions (see table 4). In
order to check whether these group differences with regard to demographics influenced our
results, we conducted covariance-analyses using several demographic variables as covariates.
We used all those demographic variables as covariates that were significantly related to at
least one of the latent or manifest functions (cf. Becker, 2005). The group differences with
regard to the latent functions were reduced in size, but remained significant in these analyses.
Overall the results were very similar to the analyses without controls (a documentation of
these results is available from the first author).
We also selected the subgroup of unskilled manual workers from the group of all employed
persons and compared them with the non-employed persons. This was done because unskilled
manual work is usually regarded as less preferable than skilled manual work or non-manual
work. Thus, this comparison might shed some light on the question of whether employment is
always a better provider of the five latent functions than non-employment.
The results showed that unskilled manual workers reported significantly more access to time
structure, social contact, collective purpose, status, and activity than unemployed persons.
They also reported significantly more access to time structure, social contact, and activity
than OLF-persons (see table 2). The effect sizes were smaller than the effect sizes for the
comparisons including the whole group of all employed persons. Only weak differences that
were not significant at the 95% level were found for distress, though.

20
In sum, Jahoda’s theory was partly endorsed by these results: Even a type of employment
which most people would like to avoid, such as unskilled manual work, provides more access
to the latent functions than unemployment or being out of the labor force. However, this
better access did not lead to clear group differences with regard to mental health.
With regard to the latent functions of employment as predictors of mental health, our data also
endorsed Jahoda’s model (see table 2, below diagonal). All correlations between distress and
manifest and latent functions were significant, although the majority was only of small size.
With r = -.49, identity/status was the latent function that had the strongest association with
distress. The correlation between distress and social contact was the second largest and of
medium size (r = -.27). The correlations between distress and time structure (r = -.13), activity
(r = -.18) and collective purpose (r = -.16) were of small size (Cohen, 1977). Contrary to
Jahoda’s prediction, the manifest function, i.e. the measure for a person’s financial situation,
was strongly correlated with distress (r = -.45).
We also conducted a multiple regression analysis to examine the predictive strength of the
manifest and latent functions when the influence of the other variables specified in Jahoda’s
model was controlled (see table 3). In this regression, only status, financial situation, and
social contact emerged as significant predictors of distress, while the regression coefficients
for time structure, collective purpose and activity were low and not significant.
Thus, we can conclude that Jahoda’s model was endorsed by our data insofar as all latent
functions of employment were significantly associated with mental health. However, most of
the correlations were of small or medium size. Only identity/status was strongly correlated
with distress. Contrary to Jahoda’s predictions, financial situation also had a strong influence
on distress.

21
We were also interested in the question of whether access to the manifest and latent functions
of employment varies for different levels of important demographic variables. The results
showed that access is higher in the higher social classes, as all functions correlated positively
with level of education and household income (see table 4). Age was negatively related to
time structure, activity, collective purpose, and social contact, but was positively related to the
person’s financial situation. For sex, only weak associations with the latent and manifest
functions were found: Women reported slightly less social contact, less activity, and less
collective purpose than men. Having an intimate relationship was associated with a good
perceived financial situation, with much activity, high status, and much collective purpose.
Household size and whether there were children in the household correlated positively with
time structure, social contact, and especially with activity, but not with identity/status.
Household size also correlated with collective purpose. With regard to the perceived financial
situation, household size and having children had negative effects. The size of the
town/village participants lived in had no effect on the variables in Jahoda’s model. East
Germans perceived their financial situation to be worse than West Germans and reported a
lower status. This is in good accordance with a frequently expressed feeling among East
Germans that they are the “losers” of the German reunification. We can conclude that access
to Jahoda’s latent functions was best among younger males from higher social classes who
lived in an intimate relationship in large households with children. The best financial situation
was found among older West-German persons from higher social classes who lived in an
intimate relationship in a comparatively small household with no children. Thus, the
demographic correlates differed markedly for the two types of functions (latent vs. manifest)
specified in Jahoda’s model.

22
Discussion
The results of the present study with a representative sample of the German population
endorsed Jahoda’s model in two important points: Employed persons generally reported more
access to four of the five latent functions than persons who were not employed. This includes
not only unemployed people, but also OLF-persons such as students, retirees, and
homemakers. This result is consistent with Jahoda’s notion that employment is the best
provider of the latent functions in modern societies. Furthermore, all latent functions were
significantly associated with distress, as predicted by Jahoda, and explained a significant
proportion of the variance in mental health.
However, there were also results that did not endorse the model. First, OLF-persons did not
report less access to status than employed persons did. This is in good agreement with the two
older studies conducted by Henwood and Miles (1987) and Evans and Banks (1992) who also
were not able to demonstrate such a difference. Thus, employment obviously is not the single
best provider of the experience of status in modern societies, as Jahoda predicted. Status can
be gained by several - equally successful - means, for example through the homemaker role.
This suggests that the low sense of status among unemployed persons may not be the result of
a lack of employment, but may be a specific characteristic of the unemployment situation.
This assumption is in good agreement with research that shows that attitudes toward the
unemployed have always, throughout history, been rather negative (Kelvin & Jarrett, 1985),
and remain negative today (Breakwell, Collie, Harrison, & Propper, 1984). In other words:
Unemployed people are a special group that suffers stigmatization, probably causing their low
sense of status. It is not the lack of employment that is responsible for this experience.
Another unexpected finding was that the alleged “manifest” function of employment did not
clearly differentiate between employed and OLF persons. The two groups with the best
financial situation were two OLF-groups, i.e. homemakers and retirees. However,
homemakers usually depend on the income of their husbands and retirees benefit from the

23
money they contributed to the old age pension scheme when they still were employed. Thus,
the positive assessments of the current financial situation of these two groups can possibly
been seen as indirect consequences of employment.
Furthermore, the manifest function, i.e. a person’s financial situation, was strongly correlated
with mental health and explained nearly as much variance in this variable as all latent
functions together (20% vs. 26%). This result also does not endorse Jahoda’s model, because
Jahoda assumed the latent functions of employment to be clearly more important for mental
health than the manifest function. Thus, the present results appear to support critics of the
latent deprivation model who claimed that Jahoda underestimated the role of poverty in
explaining the negative psychological effects of unemployment (e.g. Fryer, 1986). However,
this conclusion should be treated with some caution because the low reliabilities of the ACE-
scales may have biased this comparison (see below).
Another unexpected finding was the lack of a clear distinction between employed and OLF-
persons with regard to mental health. Full-time employed persons felt considerably less
distress than unemployed persons, which is a finding that is in good agreement with meta-
analytic results concerning the mental health effects of unemployment (McKee-Ryan et al.,
2005). However, employed persons did not show a clear advantage in comparison to OLF-
persons, as one would expect from Jahoda’s model. Thus, although OLF-individuals
experienced a lack with regard to time structure, social contact, collective purpose, and
activity, and although these four variables were significantly correlated with mental health,
this did not result in reduced health scores among the OLF-Persons. In case of retirees and
homemakers their good financial situation may have acted as a protective factor. Another
possible explanation for this result could be that Jahoda’s model is misspecified with regard to
the number and the kind of latent functions included. Possibly, there are other functions which
are relevant with regard to mental health? Warr’s vitamin model for example (Warr, 2007)
includes nine environmental variables with only limited overlap with Jahoda’s five latent

24
functions. These nine “vitamins” are also assumed to be predictors of mental health. It would
be fascinating to test whether these nine variables are better predictors of well being than
Jahoda’s five latent functions are. Another aspect that might also play a role here are values
and belief systems, as has recently been proposed in the incongruence model of
unemployment distress (Paul & Moser, 2006). This model focuses on the fit between a
person’s level of employment commitment and his or her employment situation. Whenever
these two variables do not match, i.e. are incongruent to each other, distress is provoked.
Unemployed persons usually are strongly committed to employment, but since they do not
have a job, these high levels of commitment can cause symptoms of distress (Paul & Moser,
2006). What about people who are out of the labor force? Information concerning
employment commitment among OLF-persons is rare, but some evidence exists that they
report comparatively low levels of employment commitment (Banks & Henry, 1993; Hannan,
ORianin, & Whelan, 1997), possibly protecting them from the kind of incongruence that
characterizes unemployed people. In other words: The negative mental health effects of the
lack of the latent functions among OLF-persons might be ameliorated by a good match
between their employment-related value system and their current life situation. However, this
interpretation remains tentative as long as there do not exist more empirical data concerning
the level of employment commitment of OLF-persons.
In sum, Jahoda’s model with its very positive view of employment found considerable
support in the present study: The latent functions were shown to be relevant for mental health
in the general population and were shown to be more prevalent among employed persons, not
only in comparison to unemployed persons, but also in comparison to OLF-persons.
However, the question why these differences did not translate into clearer mental health
advantages for employees remains unanswered for now.
The present findings have implications for several areas of practice in the field of work and
organizational psychology. Measures for unemployed people, for example, should always be

25
designed in a way that provides as much access to the latent functions as possible. Retirement
counseling, a field that will grow in the future due to the demographic changes in most
industrialized countries, can also benefit from the knowledge about the importance of
Jahoda’s latent functions. Finding ways to substitute them during retirement could be a
helpful element of such counseling. Furthermore, German organizations will probably be
confronted with a labor shortage in the near future (see context section). (Re)activating
potential workers who are presently out of the labor force will be vital for any organization
then. Emphasizing the psychological importance of Jahoda’s latent functions could be very
helpful in any such attempt at motivating non-employed persons to take up jobs. This may be
true not only for German organizations but also for organizations from other welfare states
with a changing demographic structure.
A limitation of the present study may be seen in the low reliabilities of the ACE-scales,
particularly collective purpose and time structure. Although the confirmatory factor analysis
clearly supported the assumed factor structure of the ACE-scales and although the validity
study endorsed their convergent and discriminant validity, the low reliabilities could possibly
have influenced the results of the present study by attenuating the effect sizes. Therefore, a
closer examination of the possible consequences of this limitation is necessary.
Effect sizes - d-values as well as correlations – are biased downward by measurement error,
making it less likely to find significant differences between groups or to find significant
relationships between variables when measures with low reliabilities are used (Hunter &
Schmidt 1990, Schmidt & Hunter, 1996). Thus, with more reliable scales, the mean
differences between employed, unemployed and OLF-groups as well as the correlations
between latent functions and mental health would have been larger in the present study. In
other words: With regard to our most central research questions (Is employment really the
main provider of the latent functions? Is a lack of the latent functions associated with

26
diminished mental health?), the bias that is introduced by the low reliabilities is a
conservative one, one that reduced the chances to find support for the hypotheses. Therefore,
our main conclusions are not threatened by the low reliabilities of the ACE-scales as we
found the expected group differences and correlations despite the downward bias. The
correlations between latent functions and demographic variables were also biased downward
by the low reliabilities. Thus, it might be possible that there are more significant associations
between latent functions and demographics than we have detected. A more serious problem
that results from the low reliabilities concerns the comparison of the effects of latent vs.
manifest functions on mental health. We were not able to demonstrate that the latent functions
are better predictors of mental health than the manifest function is, as proposed by Jahoda.
However, the reliability of the measure of financial difficulties was much better than the
average reliability of the measures of latent functions. Thus, the correlations between mental
health and latent functions had a stronger downward bias than the correlation between mental
health and the manifest function. With more reliable scales the latent functions might have
proved themselves to be better predictors of health than the manifest function. Thus, there
remain some doubts with regard to our conclusions concerning the importance of the latent
versus the manifest functions of employment. In sum, we think that the use of these short
ACE-scales was a good compromise between the practical (and financial) constraints of a
large-scale study with a representative population sample and the wish to use measures that
are as reliable as possible.
Another limitation of the present study can be seen in the use of self-report measures and the
possible consequence of monomethod bias. However, since the GHQ is one of the best
validated measures of mental health that is available and since our validation study endorsed
the discriminant validity of the ACE-scales, we are confident that the possible distortion of
the results caused by this problem was limited.

27
A final limitation of the present study is its reliance on cross-sectional data. This prohibits
causal interpretations of the findings and makes it impossible to rebut reverse-causation
interpretations of the results. For example, it might be possible that certain values of Jahoda’s
latent functions determine an individual’s employment status. Persons with a strong tendency
toward activity might possibly choose to be employed more often than persons who are
generally more passive. Such an effect of reverse-causation could have led to the group
differences we identified in the present investigation. Furthermore, it is not yet clear whether
the latent functions really influence mental health, as predicted by Jahoda, or whether the
opposite is true. Impaired mental health might possibly lead to perceptions of diminished
access to the latent functions. In the latter case, deprivation of the latent functions could not
be seen as a mediator of the mental health effects of non-employment, but only as an
epiphenomenon of reduced mental health, whatever the cause of this reduced mental health
may be. The few existing longitudinal studies that included measures of the latent functions of
employment do not give sufficient answers to these questions (Creed, Muller, & Patton, 2003;
Waters & Moore, 2002b) and should be ameliorated by additional longitudinal investigations of
Jahoda’s fascinating model.

28
Literature
Banks, M. H., & Henry, P. (1993). Change and stability in employment commitment. Journal of
Occupational and Organizational Psychology, 66, 177–184.
Becker, T. E. (2005). Potential problems in the statistical control of variables in organizational
research: A qualitative analysis with recommendations. Organizational Research Methods, 8,
274-289
Besley, T., Burgess, R., & Rasul, I. (2003). Benchmarking government provision of social safety nets.
Social Protection Discussion Paper Series No. 0315 Retrieved May 21, 2008, from
http://siteresources.worldbank.org/SOCIALPROTECTION/Resources/SP-Discussion-
papers/Safety-Nets-DP/0315.pdf
Breakwell, G. M., Collie, A., Harrison, B., & Propper, C. (1984). Attitudes towards the unemployed:
Effects of threatened identity. British Journal of Social Psychology, 23, 87-88.
Brief, A. P., Konovsky, M. A., Goodwin, R., & Link, K. (1995). Inferring the meaning of work from
the effects of unemployment. Journal of Applied Social Psychology, 25, 693-711.
Bundesinstitut für Bevölkerungsforschung (2008). Die demographische Lage in Deutschland 2006
[The demographic situation in Germany in 2006]. Retrieved May 21, 2008, from
http://www.bibdemographie.de/cln_050/nn_750440/SharedDocs/Publikationen/DE/Do
wnload/Demolage/Demolage2006.html
Campbell, D.T., & Fiske, D.W. (1959). Convergent and discriminant validation by the
multitrait-multimethod matrix. Psychological Bulletin, 56, 81-105.
Cohen, J. (1977). Statistical power analysis for the behavioral sciences. New York: Academic Press.
Creed, P. A., & Evans, M. A. (2002). Personality, well-being and deprivation theory. Personality and
Individual Differences, 33, 1045-1054.
Creed, P. A., & Machin, M. A. (2003). Multidimensional properties of the access to categories of
experience scale. European Journal of Psychological Assessment, 19, 85-91.

29
Creed, P. A., & Macintyre, S. R. (2001). The relative effects of deprivation of the latent and manifest
benefits of employment on the well-being of unemployed people. Journal of Occupational
Health Psychology, 6, 324-331.
Creed, P. A., & Muller, J., & Patton, W. (2003). Leaving high school: The influence and consequences
for psychological well-being and career-related confidence. Journal of Adolescence, 26, 295-
311.
Creed, P. A., & Reynolds, J. (2001). Economic deprivation, experiential deprivation and social
loneliness in unemployed and employed youth. Journal of Community & Applied Social
Psychology, 11, 167-178.
Creed, P. A., & Watson, T. (2003). Age, gender, psychological wellbeing and the impact of losing the
latent and manifest benefits of employment in unemployed people. Australian Journal of
Psychology, 55, 95-103.
Evans, S. T. (1986). Variations in activity and psychological well-being in unemployed young adults.
Unpublished dissertation, University of Manchester.
Evans, S. T., & Banks, M. H. (1992). Latent functions of employment: Variations according to
employment status and labor market. In: Verhaar & Jansma (eds.) On the myteries of
unemployment (p. 281-295). Dordrecht: Kluwer.
Feather, N. T., & Bond, M. J. (1983). Time structure and purposeful activity among employed and
unemployed university graduates. Journal of Occupational Psychology, 56, 241-254.
Fryer, D. (1997). Agency restriction. In: N. Nicholson (Ed.), The Blackwell encyclopedic dictionary of
organizational psychology (p. 12). Oxford: Blackwell.
Fryer, D. (1986). Employment deprivation and personal agency during unemployment: A critical
discussion of Jahoda’s explanation of the psychological effects of unemployment. Social
Behaviour, 1, 3–23.

30
Fryer, D. (1992). Psychological or material deprivation: Why does unemployment have mental health
consequences? In E. McLaughlin (Ed.), Understanding unemployment (pp. 103–125). London:
Routledge.
Fuchs, J. & Dörfler, K. (2005). Projektion des Arbeitsangebot bis 2050: Demographische Effekte sind
nicht mehr zu bremsen [Prediction of labor supply until 2050: Demographic effects cannot be
retarded any longer]. IABKurzbericht, 11/2005, 1-5.
Fuchs, J. & Söhnlein, D. (2005). Vorausschätzung der Erwerbsbevölkerung bis 2050 [Forecasting the
economically active population until 2050]. IABForschungsbericht, 16/2005, 1-51.
Goldberg, D. P., & Hillier, V. F. (1979). A scaled version of the General Health Questionnaire.
Psychological Medicine, 9, 139-145.
Hannan, D. F., O’Riain, S., & Whelan, C. T. (1997). Youth unemployment and psychological distress
in the Republic of Ireland. Journal of Adolescence, 20, 307–320.
Haworth, J. T., & Evans, S. T. (1987). Meaningful activity and unemployment. In D. Fryer & P. Ullah
(Eds.), Unemployed people. Social and psychological perspectives (pp. 241-267). Milton
Keynes, England: Open University Press.
Haworth, J. T., & Paterson, F. (1995). Access to categories of experience and mental health in a
sample of managers. Journal of Applied Social Psychology, 25, 712-724.
Henwood, F., & Miles, I. (1987). The experience of unemployment and the sexual division of labor.
In: D. Fryer & P. Ullah (eds.) Unemployed people: Social and psychological perspectives (pp.
94 – 110). Milton Keynes, England: Open University Press.
Hoare, P. N., & Machin, M. A. (2006) Maintaining wellbeing during unemployment. Australian
Journal of Career Development, 15, 19-27.
Hunter, J. E. & Schmidt, F. L. (1990). Methods of Meta-Analysis. Correcting Error and Bias in
Research Findings. New York: Sage Publications.
International Labour Office (2000). World labour report 2000: Income security and social protection
in a changing world. Geneva, Switzerland: International Labour Office.

31
Isaksson, K. (1989). Unemployment, mental health and the psychological functions of work in male
welfare clients in Stockholm. Scandinavian Journal of Social Medicine, 17, 165-169.
Jahoda, M. (1981). Work, employment, and unemployment. Values, theories, and approaches in social
research. American Psychologist, 36, 184-191.
Jahoda, M. (1982). Employment and unemployment: A social-psychological analysis. Cambridge,
England: Cambridge University Press.
Jahoda, M. (1997). Manifest and latent functions. In N. Nicholson (Ed.), The Blackwell encyclopedic
dictionary of organizational psychology (pp. 317-318). Oxford, England: Blackwell.
Jahoda, M., Lazarsfeld, P. F., & Zeisel, H. (1975). Die Arbeitslosen von Marienthal (1933)[Marien
thal: The sociography of an unemployed community]. Konstanz, Germany: Verlag für
Demoskopie.
Kelvin, P., & Jarrett, J. E. (1985). Unemployment: Its social psychological costs. Cambridge, England:
Cambridge University Press.
Kilpatrick, R., & Trew, K. (1985). Life-styles and psychological well-being among unemployed men
in Northern Ireland. Journal of Occupational Psychology, 58, 207-216.
Linden, M., Maier, W., Achberger, M., Herr, R., Helmchen, H., & Benkert, O. (1996). Psychische
Erkrankungen und ihre Behandlung in Allgemeinarztpraxen in Deutschland. [Psychiatric
diseases and their treatment in general practice in Germany]. Nervenarzt, 67, 205-215.
McFayden, R. G. (1995). Coping with threatened identities: Unemployed people’s self-
categorizations. Current Psychology: Development, Learning, Personality, Social, 14, 233-
257.
McKee-Ryan, F. M., Song, Z., Wanberg, C. R., & Kinicki, A. J. (2005). Psychological and physical
well-being during unemployment: A meta-analytic study. Journal of Applied Psychology, 90,
53-76.
Merton, R. K. (1957). Social theory and social structure. Chicago: Free Press of Glencoe.

32
Miles, I. (1983). Adaptation to unemployment? (SPRU Occasional Paper Series, No. 20). Brighton,
United Kingdom: University of Sussex, Science Policy Research Unit.
Muller, J. J., Creed, P. A., Waters, L. E., & Machin, M. A. (2005). The development and preliminary
testing of a scale to measure the latent and manifest benefits of employment. European
Journal of Psychological Assessment, 21, 191-198.
Murphy, G. C., & Athanasou, J. A. (1999). The effect of unemployment on mental health. Journal of
Occupational and Organizational Psychology, 72, 83-99.
O'Brien, G. E. & Feather, N. T. (1990). The relative effects of unemployment and quality of
employment on the affect, work values and personal control of adolescents. Journal of
Occupational Psychology, 63, 151-165.
Paul, K. I., & Moser, K. (2006). Incongruence as an explanation for the negative mental health effects
of unemployment: Meta-analytic evidence. Journal of Occupational and Organizational
Psychology, 79, 595–621.
Schmidt, F. L., & Hunter, J. E. (1996). Measurement error in psychological research: Lessons from 26
research scenarios. Psychological Methods, 1, 199-223.
Ullah, P. (1990). The association between income, financial strain and psychological well-being
among unemployed youths. Journal of Occupational Psychology, 63, 317-330.
Underlid, K. (1996). Activity during unemployment and mental health. Scandinavian Journal of
Psychology, 37, 269-281.
Wanberg, C. R., Griffiths R. F., & Gavin, M. B. (1997). Time structure and unemployment: A
longitudinal investigation. Journal of Occupational and Organizational Psychology, 70, 75-
95.
Waters, L. E. & Moore, K. A. (2002a). Reducing latent deprivation during unemployment: the role of
meaningful leisure activity. Journal of Occupational and Organizational Psychology, 75, 15-
32.

33
Waters, L. E. & Moore, K. A. (2002b). Self-esteem, appraisal and coping: A comparison of
unemployed and re-employed people. Journal of Organizational Behavior, 23, 593-604.
Warr, P. B. (2007). Work, happiness, and unhappiness. Mahwah, NJ: Lawrence Erlbaum.
Winefield, A. H., Winefield, H. R.; Tiggemann, M.; Goldney, R. D. (1991). A longitudinal study of
the psychological effects of unemployment and unsatisfactory employment on young adults.
Journal of Applied Psychology, 76, 424-431.
Worldbank (2008). GNI per capita 2006, Atlas method and PPP. Retrieved May 21, 2008, from
http://siteresources.worldbank.org/DATASTATISTICS/Resources/GNIPC.pdf

Table 1: Latent and manifest functions in different employment status groups -
Employed full
time (n=359)
Employed
part time
(n=127)
Vocational
training
(n=27)
Formal
education
(n=67)
Home
duties
(n=80)
Retirement
(n=265)
Unemployed
(n=73)
F
d
Employed vs.
OLF-persons
d
Employed vs.
unemployed
Time
structure
6.00 abcd
(1.14)
5.89 efg
(1.11)
5.97 hj
(1.22)
5.33 ak
(1.32)
5.01 bel
(1.43)
4.27 cfhkl
(1.59)
4.84 dgj
(1.50)
49.76***
1.02***
0.95***
Social
contact
5.25 abc
(1.04)
4.86 defg
(1.22)
5.90 dhik
(0.77)
5.60 elmn
(0.92)
4.37 ahl
(1.27)
4.11 bfim
(1.22)
4.10 cgkn
(1.34)
41.48***
0.66***
0.96***
Collective
purpose
4.91 abcd
(1.26)
4.72 ef
(1.20)
4.87 g
(1.27)
4.21 a
(1.31)
4.16 b
(1.40)
3.96 ce
(1.20)
3.80 dfg
(1.53)
18.81***
0.65***
0.82***
Status
5.66 a
(1.30)
5.48 b
(1.30)
5.58 c
(1.52)
5.33 d
(1.56)
5.66 e
(1.34)
5.48 f
(1.32)
4.48 abcdef
(1.66)
8.06***
0.09 n.s.
0.84***
Activity
5.47 abcd
(0.99)
5.13 ef
(1.14)
4.83 g
(0.68)
4.75 ah
(1.19)
4.93 bik
(1.10)
3.99 ceghi
(1.13)
4.10 dfk
(1.28)
54.46***
0.95***
1.17***
Financial
situation
4.87 ab
(1.43)
4.69 cd
(1.50)
4.34
(1.43)
3.83 acef
(1.56)
5.21 eg
(1.56)
5.05 fh
(1.45)
3.21 bdgh
(1.69)
21.26***
-0.06 n.s.
1.07***

36
Distress
1.61 a
(0.49)
1.67
(0.40)
1.56
(0.35)
1.69
(0.49)
1.62
(0.50)
1.70
(0.53)
1.90 a
(0.61)
3.93***
-0.11+
-0.58***
Note: Means in the same row that share subscripts differ at p ≤ .05 in the Scheffé-test; standard deviations in parentheses below the mean values; the dataset was weighted before
analysis to ensure representativeness to the German population; *** p ≤ .001, ** p ≤ .01; * p ≤ .05
Table 2: Latent and manifest functions in different employment status groups: Comparison between manual workers and OLF-persons /
unemployed persons (d-values)
Time
structure
Social
contact
Collective
purpose
Status
Activity
Financial
situation
Distress
Unskilled manual workers vs. OLF-persons
0.84***
0.53**
0.29+
-0.15 n.s.
0.51**
-0.40**
0.09 n.s.
Unskilled manual workers vs. unemployed persons
0.76***
0.77***
0.43*
0.49*
0.68***
0.63**
-0.31+
Note: Effect sizes are standardizes mean differences (Cohen’s d); the dataset was weighted before analysis to ensure representativeness to the German population; sample size manual
workers: n = 47; *** p ≤ .001, ** p ≤ .01; * p ≤ .05

Table 3: Intercorrelations between latent and manifest function of employment and distress
M
SD
1
2
3
4
5
6
7
Time structure (1)
5.32
1.52
---
Social contact (2)
4.78
1.28
.37***
---
Collective purpose (3)
4.44
1.34
.28***
.38***
---
Status (4)
5.48
1.39
.18***
.34***
.11***
---
Activity (5)
4.83
1.25
.41***
.41***
.41***
.26***
---
Financial situation (6)
4.72
1.57
.11***
.18***
.14***
.59***
.14***
---
Distress (7)
1.67
0.50
-.13***
-.27***
-.16***
-.49***
-.18***
-.45***
---
note: n = 998; the dataset was weighted before analyses; *** p ≤ .001, ** p ≤ .01 (two-tailed).

38
Table 4: Multiple regression analysis for the prediction of distress
Variable
B
SEB
ß
Time structure
0.00
0.01
0.00
Social contact
-0.04
0.01
-0.11***
Collective purpose
-0.02
0.01
-0.05
Status
-0.11
0.01
-0.30***
Activity
0.00
0.01
-0.01
Financial situation
-0.08
0.01
-0.25***
Note. R2 = 0.30; n = 998; the dataset was weighted before analyses; * p ≤ .05; ** p ≤ .01; *** p ≤ .001

39
Table 4: Correlations between manifest and latent functions and demographic variables
Time
structure
Social
contact
Collective
purpose
Identity
/ Status
Activity
Financial
situation
Age
-.35***
-.37***
-.16***
.05
-.28***
.27***
Sex (1=male, 2=female)
-.04
-.13***
-.07*
.-04
-.11***
-.04
Level of education
.22***
.16***
.17***
.12***
.20***
.13***
Intimate relationship? (0=no, 1=yes)
-.01
-.06
.10**
.13***
.20***
.24***
Household size
.17***
.17***
.11***
.03
.27***
-.07*
Number of children in Household
.13***
.10***
.04
-.01
.22***
-.12***
Household income
.17***
.26***
.26***
.28***
.34***
.39***
Size of village / town
.00
-.06
-.03
.05
-.05
.03
Geographic region (1=west, 2=east)
.03
-.05
01
-.15***
.03
-.19***
note: n = 817 - 998; the dataset was weighted before analyses; * p ≤ .05; ** p ≤ .01; *** p ≤ .001 (two-tailed)