Content uploaded by Ken Shultz
Author content
All content in this area was uploaded by Ken Shultz
Content may be subject to copyright.
The influences on bridge
employment decisions
Chanjira Pengcharoen
Personnel Commission, Los Angeles County Office of Education, Downey,
California, USA, and
Kenneth S. Shultz
Department of Psychology, California State University, San Bernardino,
California, USA
Abstract
Purpose – Population aging, and changes in labor force participation among older adults, will have
tremendous impacts on the aging workforce. Thus it is imperative that the factors that influence
whether older workers will continue in their career employment, engage in bridge employment, or fully
retire, should be understood better. This paper aims to focus on these issues.
Design/methodology/approach – In the present study longitudinal data for 2,869 older workers
from the Health and Retirement Study (HRS) data set in the USA were used to examine the influence of
demographic (e.g. income), nonwork related factors (e.g. marital satisfaction), and work related factors
(e.g. job involvement) on late-life employment decisions over a ten year period from 1992 to 2002.
Findings – The results indicate a wide variety of factors impact employment decisions later in life.
Specifically, it was found that work related factors of job involvement and schedule flexibility, as well
as the nonwork related factors of certainty of retirement plans, attitudes toward retirement, and job
seeking self-efficacy all distinguished the various employment statuses (e.g. completely retired, partly
retirement, and not retired at all) of older workers over a ten year period.
Originality/value – This study shows that both individuals and organizations need to examine a
wide variety of factors when examining bridge employment decisions at the end of workers’ careers.
While most studies of bridge employment use cross-sectional data, this paper uses longitudinal data to
examine actual bridge employment decisions, rather that prospective desires or potentially faulty
after-the-fact retrospective accounts.
Keywords Retirement, Decision making, Older workers, United States of America
Paper type Research paper
1. Introduction
Retirement is a complex process that involves major economic, social, psychological,
and health-related factors (Kiefer and Briner, 1998; Beehr and Bennett, 2007). In
addition, there are an increasing number of retirement options from which older
individuals can choose. According to recent analyses by Cahill et al. (2006), having a
bridge job is becoming a more popular retirement option than ever. Bridge employment
refers to part-time or full-time temporary jobs that are held subsequent to leaving one’s
career employment (Ruhm, 1990; Shultz, 2003). Thus, it bridges the gap between full
time career employment and full time leisure retirement.
In addition, when older individuals decide to retire, they have to decide whether to
leave their career job for full-time retirement or accept bridge employment within or
outside their career industry (Feldman, 1994). Survey data from more than three
decades ago suggested that over half of all older workers have left their career jobs and
are participating in bridge employment by the age of 60, but less than one out of nine
The current issue and full text archive of this journal is available at
www.emeraldinsight.com/0143-7720.htm
IJM
31,3
322
International Journal of Manpower
Vol. 31 No. 3, 2010
pp. 322-336
qEmerald Group Publishing Limited
0143-7720
DOI 10.1108/01437721011050602
has fully retired (Doeringer, 1990; Ruhm, 1990). More recent data analyzed by Cahill
et al. (2006) indicates that this trend has continued to accelerate.
Further, participation in bridge employment has been shown to be positively related
to both retirement and life satisfaction (Kim and Feldman, 2000; Wang and Shultz,
2010), as well as physical and psychological health (Zhan et al., 2009). Thus, bridge
employment has become an important retirement option for many workers, suggesting
the need to better understand recent trends, as well as to explore how different factors
influence various retirement options (Wang et al., 2009).
There are several reasons older workers choose to engage in bridge employment.
Recent retirement trends suggest that many older individuals prefer to work in bridge
jobs at postretirement in order to maintain a steady level of income and earn sufficient
pension (Wang et al., 2008), but also for social and personal reasons (von Bonsdorff
et al., 2009). Additionally, changes in common labor force practices offering more
flexibility to workers (e.g. abolishment of mandatory retirement age) suggest the
importance of bridge employment participation among older workers (Alley and
Crimmins, 2007; Bass, 2005; Cahill et al., 2006; Henretta, 2001). Bridge employment also
provides an opportunity for older workers to fill the gap between full-time employment
and full-time retirement by allowing them to gradually adjust to the new lifestyles
(Abraham and Houseman, 2004; Ruhm, 1990; Shultz, 2003).
Over 70 percent of baby boomers in various surveys express interests in engaging
in bridge employment rather than choosing full-time retirement (AARP, 1998).
However, one’s intention to engage in bridge employment does not always reflect his or
her retirement outcome as it depends upon one’s interests in employment as well as the
ability to acquire alternative employment at post-retirement (Abraham and Houseman,
2004). To the extent that the individual’s expectations about retirement are not
consistent with his or her plans, it is likely to lead to failure in realization of those plans
(Henkens and Tazelaar, 1997). In addition, changes in circumstances (e.g. health
problems) may change retirement plans, affecting a plan realization (Abraham and
Houseman, 2004; Wang et al., 2008).
Weckerle and Shultz’s (1999) study examined various antecedents that may
differentially predict certain retirement intentions. They analyzed Wave I of the Health
and Retirement Study (HRS) dataset, using the factors that distinguished older workers
based on their desire to retire early, continue work, participate in bridge employment
within one’s career job, and participate in bridge employment outside one’s career job.
They found that older workers who desired to continue in their current job were much
more satisfied with their financial status, had more flexible jobs, and felt the decision to
retire would be voluntary. Conversely, those who desired early retirement had
anticipated few financial rewards and expressed lack of flexibility in one’s current
employment.
In another cross-sectional study, von Bonsdorff et al. (2009) found that older
workers who perceived the job market to be good, wanted to have better use of their
skills, and had fewer concerns about changes in their benefits, were more likely to want
to engage in bridge employment in a different field than to fully retire. Meanwhile,
older workers who had fewer nonwork interests, wanted to have better use of their
skills, and had more monetary desire, were more inclined to want to hold career bridge
job than retire. Finally, younger female employees who had fewer nonwork interests
were more likely to want to hold career bridge job instead of a bridge job in a different
Bridge
employment
decisions
323
field. Similar cross-sectional studies have espoused a variety of factors that influence
bridge employment decisions including health status, job tenure, having a working
spouse, and having dependent children at home (Kim and Feldman, 2000), being self
employed and attaining a college degree (Kim and De Vaney, 2005), as well as having
an entrepreneurial orientation (Davis, 2003).
Gobeski and Beehr (2009) used current employment, or retrospective accounts of
their most recent employment for those no longer working, to examine the factors that
predicted actual bridge employment behaviors of those who have already retired from
their career employment. They found that those who had experienced job strains were
more likely to engage in bridge employment in a different field, while those who had
high job specific skills or high intrinsically motivation characteristics in their career
jobs, were more likely to engage in career related bridge employment.
An important shortcoming of Weckerle and Shultz’s (1999), von Bonsdorff at al.’s
(2009), and Gobeski and Beehr’s (2009) studies were the use cross-sectional data. In
addition, Weckerle and Shultz, as well as von Bonsdorff et al, examined retirement
intentions (versus actual behavior), whereas Gobeski and Beehr examined
retrospective accounts of previous bridge jobs, as their outcome variable. Thus,
while a few recent studies have looked at actual retirement behaviors (e.g. Gobeski and
Beehr, 2009), some using longitudinal data (e.g. Wang et al., 2008), at present, we still
have a modest understanding of what the antecedents of such important post career
employment decisions are (Wang and Shultz, 2010).
1.1 Work and nonwork-related factors
Few researchers have explored the potential predictors of bridge employment despite
its important implications for organizations, individuals, and society (Bennett et al.,
2005; Wang and Shultz, 2010). Previously, work and nonwork-related variables were
used as a framework to understand bridge employment predictors (Beehr et al., 2000;
Gobeski and Beehr, 2009; von Bonsdorff et al., 2009; Wang et al., 2009). Although
work-related predictors generally have less impact on full-time retirement, they can
influence older workers’ decisions to retire (Beehr et al., 2000; Bennett et al., 2005). In
addition, there is support for a greater influence of nonwork-related factors as better
predictors of retirement age (Reitzes et al., 1998) and bridge employment (Bennett et al.,
2005; Wang et al., 2008) than work-related factors. This suggests that older workers are
more influenced to retire by factors outside their job and workplace than factors based
on the perception of their work situations (see Henkens, 1999, and also Henkens and
Leenders in this special issue). Together, work-related and nonwork-related factors
tend to produce greater effects on retirement decisions than considering either type
alone (Beehr et al., 2000; Wang et al., 2009).
In addition, previous studies have found a variety of personal variables that are
associated with the decision to retire and/or engage in bridge employment (Beehr, 1986;
Beehr et al., 2000; Brody and Shultz, 2006; Brown et al., 1996; Davis, 2003; Gobeski and
Beehr, 2009; Kim and Feldman, 1998; Wang, 2007; Wang et al., 2008). Accordingly, age,
gender, health condition, annual household income, and education level were controlled
for when we explored the various work and nonwork related predictors. Seven
retirement related factors were explored in this study in an attempt to better
understand retirement decision. More specifically, four work related factors (i.e. work
schedule flexibility, job satisfaction, job involvement, and job seeking self-efficacy),
IJM
31,3
324
and three nonwork related factors (i.e. certainty of retirement plans, familial and
marital satisfaction, and attitudes toward retirement) were examined as a means to
better understand, explain, and accurately predict their influences on late life career
and retirement decision, above and beyond personal demographic variables
(Szinovacz, 2003).
1.2 Hypotheses
After controlling for age, gender, health condition, annual household income, and
education level:
H1. Older workers whose organizations offer a less flexible work schedule are
more likely to choose bridge employment rather than other options.
H2. Older workers who are less satisfied with their jobs are more likely to choose
full-time retirement rather than other options.
H3. Older workers who are more involved with their current jobs are more likely
to participate in bridge employment rather than other options.
H4. Older workers who view themselves as more competent job seekers are more
likely to choose bridge employment participation rather than other options.
H5. Older workers who are more certain about their retirement plans are more
likely to choose full-time retirement rather than other options.
H6. Older workers who are more satisfied with their family and marriage are more
likely to choose full-time retirement rather than other options.
H7. Older workers who perceive retirement more negatively are more likely to
choose bridge employment participation rather than other options.
2. Method
2.1 Participants
Our study used participants from the Health and Retirement Study (HRS), consisting of
US workers from the initial sample of 12,654 in Wave I (1992) face-to-face interviews.
We used four inclusion criteria to select our participants:
(1) working at the time of data collection in 1992;
(2) employed at their (then) current job (in 1992) for at least ten years;
(3) at least 51 years of age in 1992; and
(4) not retired in 1992, but reported that they were completely retired, partly
retired, or continued career employment in Wave VI (2002) (Juster and Suzman,
1995).
Based on these specified criteria, 2,869 participants were included in our study.
2.2 Procedure
The HRS is conducted by the University of Michigan with support from the US
National Institute on Aging (NIA), surveying more than 22,000 Americans over the age
of 50 (in 1992) every two years. A more detailed description of the initial data collection
Bridge
employment
decisions
325
procedures can be found in Juster and Suzman (1995) and the HRS official web site
(http://hrsonline.isr.umich.edu/).
2.3 Measures
Demographic/control variables. Previous research findings (e.g. Bennett et al., 2005;
Brody and Shultz, 2006; Kim and Feldman, 1998) revealed various personal
demographic factors influence a worker’s retirement decision; therefore, our study
controlled for age, gender, health condition, annual household income, and education
level of the participants.
Work-related variables. A series of one-item measures was used to measure work
schedule flexibility, job satisfaction, job involvement, and job seeking self-efficacy (see
Table I for specific wording of items).
Nonwork-related variables. A three-items scale was used to measure certainty of
retirement plans (
a
¼0:83), as well as familial and marital satisfaction (
a
¼0:75),
whereas a one-item scale was used to measure attitude toward retirement (see Table I
for specific wording of items).
Criterion variable. Retirement decision is a categorical variable and is defined as the
degree to which one categorizes his or her work status in 2002 as retired. It was
measured by a question containing four options, with 1 (completely retired), 3 (partly
retired), 5 (not retired at all), and 7 (question not relevant to respondent; doesn’t work
Work schedule flexibility Not counting overtime hours, could you reduce the number
of hours in your regular work schedule?
Job satisfaction Please tell me how satisfied or dissatisfied you are with your
job at the current time?
Job involvement Thinking of your job, how much do you agree or disagree
that even if you didn’t need the money, you would probably
keep on working?
Job seeking self-efficacy Suppose you were to lose your job this month, what do you
think are the chances that you could find an equally good
job in the same line of work within the next few months?
Certainty of retirement plans How much have you thought about retirement?
How much have you discussed retirement with your
husband, wife or partner?
How much have you discussed retirement with your friends
or co-workers?)
Familial and marital satisfaction Please tell me how satisfied or dissatisfied you are with your
marriage at the current time?
Please tell me how satisfied or dissatisfied you are with your
family life at the current time?
Generally speaking, would you say that the time you spend
together with your spouse or partner is _____
[1 ¼extremely enjoyable to 4 ¼not at all enjoyable])
Attitude toward retirement When you think about the time when you and your
husband, wife, or partner will (completely) retire, are you
looking forward to it, are you uneasy about it, or what? With
1 (looking forward), 3 (pro-con), and 5 (uneasy). Option 3
was set as missing as we are only interested in option 1
(positive attitude toward retirement) and 2 (negative attitude
toward retirement)
Table I.
Item wording for
predictor variables
IJM
31,3
326
for pay or is homemaker; hasn’t worked for 10 or more years). In our study, retirement
decision option 1 (completely retired) refers to fully retired, retirement decision option 3
(partly retired) refers to participated in bridge employment, option 5 (not retired at all)
refers to continued employment with the same job reported in 1992, and option 7
(question not relevant to respondent) was set as missing.
3. Results
A multinomial logistic regression (MLR) analysis was performed through SPSS
NOMREG to assess prediction of group membership in one of the three criterion
categories of full-time retirement or completely retired, bridge employment or partly
retired, and continued career employment or not retired at all, first on the basis of five
demographic predictors and then after the addition of the seven work related and
nonwork related predictors.
The non-significant goodness-of-fit results suggest that with all variables contained
in the model, it shows an excellent fit, using Pearson criterion,
x2ð3,928;n¼1,982) ¼3,953.51, p.0:05, and Deviance criterion,
x2ð3,928;n¼1,982) ¼3,524.73, p.0:05. The Model Fitting Information tables
(see Table II) for demographic and predictor variables together,
(x2ð34;n¼1,982) ¼473.87, p,0:05), and demographic variables alone,
(x2ð10;n¼2,409) ¼369.80, p,0:05), were used to calculate the difference
between the two models in order to evaluate improvement in fit. The difference
between the two models revealed reliable improvement in the model with the addition
of the predictor variables (x2ð24;n¼427) ¼104.07, p,0:05). All three pseudo r
2
measures show more variance explained after adding the seven predictor variables to
the model that already contained the five demographic variables (see Table II).
The Likelihood Ratio Tests show a total of seven variables (three demographic and
four predictor variables) reliably distinguished older workers’ employment status (see
Table III). The three demographic variables were age (x2ð2;n¼1,982) ¼253.283,
p,0:05), health condition (x2ð2;n¼1,982) ¼10.15, p,0:05), and gender
(x2ð2;n¼1,982) ¼7.435, p,0:05), while the other four predictor variables were
job seeking self-efficacy (x2ð2;n¼1,982) ¼17.417, p,0:05), one out of three items
measuring certainty of retirement plans (x2ð2;n¼1,982) ¼10.607, p,0:05), job
involvement (x2ð2;n¼1,982) ¼22.707, p,0:05), and attitudes toward retirement
(x2ð2;n¼1,982) ¼9.299, p,0:05). The model is significantly degraded by the
removal of each of the seven variables. In addition to demographic variables, five out of
seven predictor variables reliably separated participants’ employment status. The five
R
2
Model ndf x2
Cox and
Snell Nagelkerke McFadden
Step 1 – Demographic variables 2409 10 369.80 *0.14 0.16 0.08
Step 2 – Demographic and predictor
variables 1982 34 473.87 *0.21 0.25 0.12
Model change 427 24 104.07 *0.07 0.09 0.04
Table II.
Model fitting information
between the two models
and model change
information
Bridge
employment
decisions
327
significant predictor variables were derived from each employment status comparison
(a total of three pairs).
3.1 Work-related variables
Work schedule inflexibility variable significantly predicted whether individuals consider
themselves partly retired or completely retired with individuals who could not reduce
their work schedule being 38.80 percent more likely to consider themselves completely
retired (x2ð1;n¼1,982) ¼5.181, p,0:05, Exp(B) ¼1.388), and 28 percent less
likely to be partly retired (x2ð1;n¼1,982) ¼5.181, p,0:05, Exp(B) ¼0.720) (see
Table IV). Work schedule inflexibility as a significant predictor of being completely
retired was supported. Thus, the results did not support H1.
Job satisfaction was not a significant predictor of older workers’ employment status,
thus H2 was not supported. However, job involvement reliably separated participants
who were partly retired, completely retired, and not retired at all. Specifically, for every
scale rating increase in individuals’ agreement on the statement “if you didn’t need the
R
2
Variable (step 2) x2
22 log
likelihood
Cox and
Snell Nagelkerke McFadden
Model 1
IncomeSQ 2.813 3527.542
EducationSQ 0.554 3525.284
AgeSQ 253.283 *3778.013
Health condition 10.150 *3534.879
Gender 7.435*3532.164
Model 2
Work schedule flexibility:
Item 1 5.254 3529.983
Item 2 3.056 3527.786
Job satisfaction:
item 1 0.436 3525.165
Job involvement
Item 1 22.707 *3547.436
Job seeking self-efficacy:
Item 1 17.417 *3542.147
Certainty of retirement plans:
Item 1 0.431 3525.160
Item 2 10.607 *3535.336
Item 3 4.255 3528.984
Familial and marital satisfaction:
Item 1 0.716 3525.445
Item 2 0.029 3524.759
Item 3 3.647 3528.377
Attitudes toward retirement plan:
Item 1 9.299 *3534.029
Model 2
Demographic and predictor
variables 0.213 0.145 0.119
Notes: *p,0.05; n¼2,409, df ¼2
Table III.
Significant likelihood
ratio tests for the seven
variables predicting
employment status
IJM
31,3
328
money, you would probably keep on working,” individuals were 27.30 percent more
likely to consider themselves partly retired than completely retired
(x2ð1;n¼1,982) ¼8.825, p,0:05, Exp(B) ¼1.273), 41.60 percent more likely to
consider themselves not retired at all than completely retired
(x2ð1;n¼1,982) ¼19.364, p,0:05, ExpðBÞ¼1:416), 21.40 percent less likely to
consider themselves completely retired than partly retired (x2ð1;n¼1,982) ¼8.825,
p,0:05, Exp(B) ¼0.786), and 29.40 percent less likely to consider themselves
completely retired than not retired at all (x2ð1;n¼1982) ¼19.364, p,0:05,
ExpðBÞ¼0:706). Job involvement as a significant predictor of partly retired was
expected and supported, while job involvement significantly predicting not retired at
all was unexpected (Table V). Thus, H3 was partially supported.
Referent group Employment status BWald x2 Exp(B) 95% CI for Exp(B)
Completely retired Partly retired
Item 1 20.328 5.181 *0.720 0.543-0.955
Item 2 20.097 0.538 0.907 0.699-1.177
Not retired at all
Item 1 20.157 1.230 0.855 0.648-1.128
Item 2 20.221 3.026 0.802 0.625-1.028
Partly retired Completely retired
Item 1 0.328 5.181 *1.388 1.047-1.841
Item 2 0.097 0.538 1.102 0.850-1.430
Not retired at all
Item 1 0.171 1.095 1.187 0.861-1.636
Item 2 20.124 0.664 0.884 0.656-1.190
Not retired at all Completely retired
Item 1 0.157 1.230 1.170 0.887-1.542
Item 2 0.221 3.026 1.248 0.972-1.601
Partly retired
Item 1 20.171 1.095 0.843 0.611-1.161
Item 2 0.124 .664 1.132 0.840-1.524
Notes: *p,0.05; n¼1,982; df ¼1
Table IV.
Work schedule
inflexibility variable
among three employment
status categories
Referent group Employment status BWald x2 Exp(B) 95% CI for Exp(B)
Completely retired Partly retired 0.241 8.825 *1.273 1.086-1.492
Not retired at all 0.348 19.364 *1.416 1.213-1.653
Partly retired Completely retired 20.241 8.825 *0.786 0.670-0.921
Not retired at all 0.107 1.269 1.113 0.924-1.340
Not retired at all Completely retired 20.348 19.364 *0.706 0.605-0.825
Partly retired 20.107 1.269 0.899 0.746-1.082
Notes: *p,0.05; n¼1,982, df ¼1
Table V.
Job involvement variable
across three employment
status categories
Bridge
employment
decisions
329
Job seeking self-efficacy reliably separated participants who consider themselves
partly retired, completely retired, as well as not retired at all. Specifically, for every
scale rating increase in individuals’ certainty to find an equally good job in the same
line of work within the next few months of losing their current job, individuals were
3.60 percent more likely to consider themselves partly retired than completely retired
(x2ð1;n¼1,982) ¼4.392, p,0:05, Exp(B) ¼1.036) and 6.90 percent more likely
to consider themselves not retired at all than completely retired
(x2ð1;n¼1,982) ¼16.619, p,0:05, ExpðBÞ¼1:069). Job seeking self-efficacy
as a significant predictor of partly retired was expected and supported, but job
involvement as a significant predictor of not retired at all was unexpectedly supported
(see Table VI). Thus, these results provided partial support for H4.
3.2 Nonwork-related variables
Certainty of retirement plans significantly predicted whether individuals consider
themselves not retired at all or completely retired, where for each scale rating increase
in the amount of discussion about retirement individuals had with their spouse,
individuals were 25 percent less likely to consider themselves not retired at all than
completely retired (x2ð1;n¼1,982Þ¼9:774, p,0:05, ExpðBÞ¼0:750) and 24.10
percent less likely to consider themselves not retired at all than partly retired
(x2ð1;n¼1,982) ¼6.219, p,0:05, ExpðBÞ¼0:759). Certainty of retirement plans
as a significant predictor of completely retired was expected and supported, while
certainty of retirement plans as a significant predictor of partly retired was
unexpectedly supported (see Table VII). The results thus partially supported H5.
Surprisingly, none of the three-items measuring familial and marital satisfaction was a
significant predictor of one’s employment status. Thus, our results did not provide
support for H6.
Negative attitude toward retirement significantly predicted whether individuals
consider themselves not retired at all or completely retired. Specifically, when individuals
were uneasy with completely retiring with spouse, individuals were 61.30 percent more
likely to consider themselves not retired at all than completely retired
(x2ð1;n¼1,982) ¼9.326, p,0:05, ExpðBÞ¼1:613) and 38 percent less likely to
consider themselves completely retired than not retired at all (x2ð1;n¼1,982) ¼9.326,
p,0:05, ExpðBÞ¼0:620) (see Table VIII). These results did not provide support for
our proposed H7.
Going beyond H7, we examined the influence of positive attitude toward
retirement on older worker’s employment status. Positive attitude toward retirement
Referent group Employment status BWald x2 Exp(B) 95% CI for Exp(B)
Completely retired Partly retired 0.035 4.392 *1.036 1.002-1.071
Not retired at all 0.066 16.619 *1.069 1.035-1.103
Partly retired Completely retired 20.035 4.392 *0.965 0.934-.998
Not retired at all 0.031 2.539 1.031 0.993-1.071
Not retired at all Completely retired 20.066 16.619 *0.936 0.906-.966
Partly retired 20.031 2.539 0.970 0.933-1.007
Notes: *p,0.05; n¼1,982, df ¼1
Table VI.
Job seeking self-efficacy
across three employment
status categories
IJM
31,3
330
significantly predicted whether individuals consider themselves not retired at all or
completely retired. Specifically, when individuals looked forward to completely
retiring with their spouse, individuals were 38 percent less likely to consider
themselves not retired at all than completely retired (x2ð1;n¼1,982) ¼9.326,
p,0:05, ExpðBÞ¼0:620) and 61.30 percent more likely to consider themselves
completely retired than not retired at all (x2ð1;n¼1,982) ¼9.326, p,0:05,
ExpðBÞ¼1:613) (see Table IX).
Referent group Employment status BWald x2 Exp(B) 95% CI for Exp(B)
Completely retired Partly retired
Item 1 20.030 0.109 0.970 0.812-1.160
Item 2 20.012 0.016 0.988 0.817-1.194
Item 3 0.035 0.241 1.036 0.900-1.192
Not retired at all
Item 1 20.054 0.415 0.948 0.805-1.116
Item 2 20.288 9.774 *0.750 0.626-0.898
Item 3 20.128 3.135 0.880 0.764-1.014
Partly retired Completely retired
Item 1 0.030 0.109 1.030 0.862-1.231
Item 2 0.012 0.016 1.012 0.837-1.224
Item 3 20.035 0.241 0.965 0.839-1.111
Not retired at all
Item 1 20.024 0.055 0.977 0.801-1.190
Item 2 20.276 6.219 *0.759 0.611-0.943
Item 3 20.163 3.634 0.850 0.719-1.005
Not retired at all Completely retired
Item 1 0.054 0.415 1.055 0.896-1.243
Item 2 0.288 9.774 *1.334 1.113-1.598
Item 2 0.128 3.135 1.136 0.986-1.308
Partly retired
Item 1 0.024 0.055 1.024 0.840-1.248
Item 2 0.276 6.219 *1.318 1.061-1.637
Item 3 0.163 3.634 1.177 0.995-1.391
Notes: *p,0.05; n¼1,982, df ¼1
Table VII.
Certainty of retirement
plans across three
employment status
categories
Referent group Employment status BWald x
2
Exp(B) 95% CI for Exp(B)
Completely retired Partly retired 0.197 1.307 1.218 0.869-1.708
Not retired at all 0.478 9.326 *1.613 1.187-2.192
Partly retired Completely retired 20.197 1.307 0.821 0.586-1.151
Not retired at all 0.281 2.241 1.324 0.917-1.913
Not retired at all Completely retired 20.478 9.326 *0.620 0.456-0.843
Partly retired 20.281 2.241 0.755 0.523-1.091
Notes: *p,0.05; n¼1,982, df ¼1
Table VIII.
Negative attitude toward
retirement variable
across three employment
status categories
Bridge
employment
decisions
331
4. Discussion
The significant findings and contributions of our proposed hypotheses allow us to
better understand the influences of various retirement related factors above and
beyond demographic factors on one’s decision to retire full-time, engage in bridge
employment, or continue full-time career employment. Such an important decision in
one’s life greatly impacts the individuals themselves and their families, organizations,
as well as the society as a whole (Beehr and Bennett, 2007; Shultz, 2003; Shultz and
Adams, 2007). Therefore, identifying and understanding the effects of various
retirement related factors on retirement related decisions may be useful as a guide to
effective management of the individuals (e.g. retirement preparation), organizations
(e.g. implementation of appropriate organizational policies and practices, promoting
supportive work environment and positive work attitudes), and society as a whole (e.g.,
implementing laws protecting older workers) (Beehr and Bennett, 2007). The
importance of one’s retirement decision among early retirees suggests that it has
strong influences on both pre and post-retirement decisions and adjustment (Feldman,
1994; Hanisch, 1994). The insight gained from this study allows us to increase the
probability of success as retirees and employees, and increases workplace and
retirement satisfaction, as well as adjustment as we learn more about the underlying
processes (Wang and Shultz, 2010).
Our study helps fill in an existing gap in the literature by comparing actual bridge
employment behaviors with full-time retirement and continued career employment and
by examining various predictors of those options using longitudinal data. In the
present study, for example, we found that adding the work and nonwork related
factors to the demographic factors significantly improved the fit of the model in
predicting future employment status, and increased the effect size estimates by 50
percent or more (see Table II). Specifically, we found that job involvement, job seeking
self-efficacy, certainty of retirement plans, and attitudes toward retirement plan all
predicted employment status in a longitudinal fashion (see Table III). These results are
consistent with previous studies of the antecedents of bridge employment (e.g. Wang
et al., 2008).
However, the nature of bridge employment still requires further research,
particularly on various retirement decisions (i.e. full-time versus part-time or
same-field versus different field), in order to fully understand the phenomenon (Bennett
et al., 2005; Wang and Shultz, 2010). In addition, the statistical analysis method of
hierarchical multinomial logistic regression is likely to continue its popularity in future
Referent group Employment status BWald x
2
Exp(B) 95% CI for Exp(B)
Completely retired Partly retired 20.197 1.307 0.821 0.586-1.151
Not retired at all 20.478 9.326 *0.620 0.456-0.843
Partly retired Completely retired 0.197 1.307 1.218 0.869-1.708
Not retired at all 20.281 2.241 0.755 0.523-1.091
Not retired at all Completely retired 0.478 9.326 *1.613 1.187-2.192
Partly retired 0.281 2.241 1.324 0.917-1.913
Notes: *p,0.05; n¼1,982, df ¼1
Table IX.
Positive attitude toward
retirement variable
across three employment
status categories
IJM
31,3
332
studies wishing to compare different employment or retirement decisions (Bennett
et al., 2005; Brody and Shultz, 2006; Gobeski and Beehr, 2009; von Bonsdorff et al., 2009;
Wang et al., 2008). This paper attempts to understand the decision to retire using actual
retirement related outcomes rather than intentions. Abraham and Houseman (2004)
suggest future research to consider an individual’s plans for retirement distinctive
from retirement outcome, as intention does not always lead to the desired outcome (see
also Henkens and Tazelaar, 1997).
Furthermore, this paper extended retirement research beyond previous studies
conducted by Weckerle and Shultz (1999), Kim and Feldman (1998, 2000), and Feldman
and Kim (2000). Weckerle and Shultz’s (1999) cross-sectional study used the HRS 1992
(Wave I) data to examine the factors that distinguished older workers based on their
reported retirement intention options. Longitudinal data, however, helps us to better
understand the antecedents of full-time retirement and bridge employment (Bennett
et al., 2005), as well as plans for retirement and realization of those plans (Abraham and
Houseman, 2004).
Additionally, Kim and Feldman (1998, 2000) and Feldman and Kim (2000) examined
the bridge employment experience of University of California (UC) system faculty
members who were offered a series of early retirement incentive. In particular, Kim and
Feldman (1998) examined the predictors of a series of three early retirement incentive
offers acceptance, while their study in the year of 2000 explored bridge employment
participation among UC faculty members. However, our study utilized a broader and
more representative sample of the general population via the HRS, thus greatly
improving the generalizability and utility of our findings.
Additional research on the retirement decision making process is needed in order to
help address various limitations in the current study. There are a few limitations
associated with the use of archival data (Shultz et al., 2005). First, the initial design of
the HRS was planned for a different purpose, leading to limited direct measures of
constructs of interest to this study. Therefore, the representativeness of older workers’
profiles may not be complete. Future studies should include additional predictors in
order to provide more comprehensive profiles of older workers’ decision on their
employment statuses as they approach retirement.
Second, the use of single-item measures was a limitation in our study as it may not
be representative of the predictor. However, Nagy (2002) supports use of single item
measures for job satisfaction. Regardless, our study might have underestimated the
relationships between predictors and older workers’ employment status due to
measurement error of the actual retirement outcome (Shultz and Whitney, 2005).
Future studies should also test these relationships using well-established (or at least
psychometrically well defined) scales in order to provide more accurate estimates of the
relationships.
Third, the longitudinal HRS data were collected at two-year time intervals, leading
to the difficulty to examine and track any changes in work related and nonwork related
variables on older workers’ employment status. Future studies may want to collect
longitudinal data with a shorter time interval in order to assess more information about
the retirement process. This would allow us to better understand the dynamic nature of
bridge employment (Wang et al., 2009; Wang and Shultz, 2010).
In summary, our study makes a significant contribution to current retirement
literature by examining the influences of various work-related and nonwork-related
Bridge
employment
decisions
333
predictors on older workers’ employment statuses (full-time retirement, bridge
employment, and continued career employment) using a nationally representative,
longitudinal data set. It provides further theoretical and methodological foundations
for future studies attempting to better understand the actual retirement decision using
longitudinal data.
References
AARP (1998), Boomers Approaching Midlife: How Secure a Future?, AARP, Washington, DC.
Abraham, K.G. and Houseman, S.N. (2004), “Work and retirement plans among older Americans”,
working paper, Upjohn Institute, Kalamazoo, MI, available at: www.upjohninst.org/
publications/wp/04-105.pdf
Alley, D. and Crimmins, E. (2007), “The demography of aging and work”, in Shultz, K.S. and
Adams, G.A. (Eds), Aging and Work in the 21st Century, Psychology Press, New York,
NY, pp. 7-23.
Bass, S.A. (2005), “New models for post-retirement employment”, in Beatty, P.T. and
Visser, R.M.S. (Eds), Thriving on an Aging Workforce: Strategies for Organizational and
Systematic Change, Kriege, Malabar, FL, pp. 161-9.
Beehr, T.A. (1986), “The process of retirement”, Personnel Psychology, Vol. 39 No. 1, pp. 31-55.
Beehr, T.A. and Bennett, M.M. (2007), “Examining retirement from a multi-level perspective”,
in Shultz, K.S. and Adams, G.A. (Eds), Aging and Work in the 21st Century, Psychology
Press, New York, NY, pp. 277-302.
Beehr, T.A., Glazer, S., Nielson, N.L. and Farmer, S.J. (2000), “Work and nonwork predictors of
employees’ retirement ages”, Journal of Vocational Behavior, Vol. 57 No. 2, pp. 206-25.
Bennett, M.M., Beehr, T.A. and Lepisto, L.R. (2005), “Working after retirement: predictors of
bridge employment”, paper presented at the annual meeting of the Society for Industrial
and Organizational Psychology, Los Angeles, CA, April.
Brody, A. and Shultz, K.S. (2006), “Bridge employment: does occupational self-efficacy determine
which bridges are crossed?”, paper presented at the annual meeting of the Society for
Industrial and Organizational Psychology, Dallas, TX, May.
Brown, M.T., Fukunaka, C., Umemoto, D. and Wicker, L. (1996), “Annual review, 1990-1996:
Social class, work, and retirement behavior”, Journal of Vocational Behavior, Vol. 49 No. 2,
pp. 159-89.
Cahill, K.E., Giandrea, M.D. and Quinn, J.F. (2006), “Retirement patterns from career
employment”, The Gerontologist, Vol. 46 No. 4, pp. 514-23.
Davis, M.A. (2003), “Factors related to bridge employment participation among private sector
early retirees”, Journal of Vocational Behavior, Vol. 63 No. 1, pp. 55-71.
Doeringer, P.B. (1990), Bridges to Retirement, Cornell University ILR Press, Ithaca, NY.
Feldman, D.C. (1994), “The decision to retire early: a review and conceptualization”, Academy of
Management Review, Vol. 19 No. 2, pp. 285-311.
Feldman, D.C. and Kim, S. (2000), “Bridge employment during retirement: a field study of
individual and organizational experiences with post-retirement employment”, Human
Resource Planning, Vol. 23 No. 1, pp. 14-25.
Gobeski, K.T. and Beehr, T.A. (2009), “How retirees work: predictors of different types of bridge
employment”, Journal of Organizational Behavior, Vol. 30 No. 3, pp. 401-25.
Hanisch, K.A. (1994), “Reasons people retire and their relation to attitudinal and behavioral
correlates in retirement”, Journal of Vocational Behavior, Vol. 45 No. 1, pp. 1-16.
IJM
31,3
334
Henkens, K. (1999), “Retirement intentions and spousal support: a multiactor approach”, Journal
of Gerontology: Social Sciences, No. 54B, pp. S63-S73.
Henkens, K. and Tazelaar, F. (1997), “Explaining retirement decisions of civil servants in the
Netherlands: intentions, behavior, and the discrepancy between the two”, Research on
Aging, Vol. 19 No. 2, pp. 139-73.
Henretta, J.C. (2001), “Work and retirement”, in Binstock, R.H. and George, L.K. (Eds), Handbook
of Aging and the Social Sciences, 5th ed., Academic Press, San Diego, CA, pp. 255-71.
Juster, F.T. and Suzman, R. (1995), “An overview of the Health and Retirement Study”,
The Journal of Human Resources, supplement, Vol. 30 No. 4, pp. S7-S56.
Kiefer, T. and Briner, R.B. (1998), “Managing retirement – rethinking links between individual
and organization”, European Journal of Work and Organizational Psychology, Vol. 7 No. 3,
pp. 373-90.
Kim, H. and De Vaney, H.A. (2005), “The selection of partial or full retirement by older workers”,
The Journal of Family and Economic Issues, Vol. 26 No. 3, pp. 371-94.
Kim, S. and Feldman, D.C. (1998), “Healthy, wealthy, or wise: predicting actual acceptances of
early retirement incentives at three points in time”, Personnel Psychology, Vol. 51 No. 3,
pp. 623-42.
Kim, S. and Feldman, D.C. (2000), “Working in retirement: the antecedents of bridge employment
and its consequences for quality of life in retirement”, Academy of Management Journal,
Vol. 43 No. 6, pp. 1195-210.
Nagy, M.S. (2002), “Using a single-item approach to measure facet job satisfaction”, Journal of
Occupational and Organizational Psychology, Vol. 75 No. 1, pp. 77-86.
Reitzes, D.C., Mutran, E.J. and Fernandez, M.E. (1998), “The decision to retire: a career
perspective”, Social Science Quarterly, Vol. 79 No. 3, pp. 607-19.
Ruhm, C.J. (1990), “Bridge jobs and partial retirement”, Journal of Labor Economics, Vol. 8 No. 4,
pp. 482-501.
Shultz, K.S. (2003), “Bridge employment: work after retirement”, in Adams, G.A. and Beehr, T.A.
(Eds), Retirement: Reasons, Processes, and Results, Springer Publishing Company,
New York, NY, pp. 214-41.
Shultz, K.S. and Adams, G.A. (2007), “In search of a unifying paradigm for understanding aging
and work in the 21st century”, in Shultz, K.S. and Adams, G.A. (Eds), Aging and Work in
the 21st Century, Psychology Press, New York, NY, pp. 303-19.
Shultz, K.S. and Whitney, D. (2005), Measurement Theory in Action: Case Studies and Exercises,
Sage Publications, Thousand Oaks, CA.
Shultz, K.S., Hoffman, C.C. and Reiter-Palmon, R. (2005), “Using archival data for I-O research:
advantages, pitfalls, sources, and examples”, The Industrial-Organizational Psychologist,
Vol. 42 No. 3, pp. 31-7.
Szinovacz, M.E. (2003), “Contexts and pathways: retirement as institution, process, and
experience”, in Adams, G.A. and Beehr, T.A. (Eds), Retirement: Reasons, Processes, and
Results, Springer Publishing Company, New York, NY, pp. 6-52.
von Bonsdorff, M.E., Shultz, K.S., Leskinen, E. and Tansky, J. (2009), “The choice between
retirement and bridge employment: a continuity theory and life course perspective”,
International Journal of Aging and Human Development, Vol. 69 No. 2, pp. 79-100.
Wang, M. (2007), “Profiling retirees in the retirement transition and adjustment process:
examining the longitudinal change patterns of retirees’ psychological well-being”, Journal
of Applied Psychology, Vol. 92 No. 2, pp. 455-74.
Bridge
employment
decisions
335
Wang, M. and Shultz, K.S. (2010), “Employee retirement: a review and recommendations for
future investigation”, Journal of Management, Vol. 36 No. 1, pp. 172-206.
Wang, M., Adams, G.A., Beehr, T.A. and Shultz, K.S. (2009), “Career issues at the end of one’s
career: bridge employment and retirement”, in Baugh, S.G. and Sullivan, S.E. (Eds),
Maintaining Focus, Energy, and Options through the Life Span, Information Age,
Charlotte, NC, pp. 135-62.
Wang, M., Zhan, Y., Liu, S. and Shultz, K.S. (2008), “Antecedents of bridge employment:
a longitudinal investigation”, Journal of Applied Psychology, Vol. 93 No. 4, pp. 818-20.
Weckerle, J.R. and Shultz, K.S. (1999), “Influences on the bridge employment decision among
older USA workers”, Journal of Occupational and Organizational Psychology, Vol. 72 No. 3,
pp. 317-29.
Zhan, Y., Wang, M., Liu, S. and Shultz, K.S. (2009), “Bridge employment and retirees’ health:
a longitudinal investigation”, Journal of Occupational Health Psychology, Vol. 14 No. 4,
pp. 374-89.
About the authors
Chanjira Pengcharoen (now Chanjira Luu) is currently a Human Resources Analyst at the Los
Angeles County Office of Education in Los Angeles, California. Prior to her current position, she
was the Human Resources Analyst with the Hacienda La Puente Unified School District and the
Human Resources Intern with the Los Angeles Unified School District. Ms Luu earned her
Master’s degree in Industrial and Organizational Psychology from the California State
University, San Bernardino, and bachelor’s degree in Psychology with Labor and Workplace
Studies specialization from the University of California, Los Angeles (UCLA). Chanjira
Pengcharoen is the corresponding author and can be contacted at: Luu_Chanjira@lacoe.edu
Kenneth S. Shultz is an organizational psychologist and professor in the psychology
department at California State University, San Bernardino. He teaches classes in work and
organizational psychology as well as statistics and research methodology. He has published
extensively on the topics of retirement and bridge employment. He recently co-edited (with Gary
Adams) the book Aging and Work in the 21st Century published by Psychology Press/Taylor
and Francis and is currently completing a co-authored book (with Mo Wang and Deborah Olson)
titled Mid and Late Career Issues: Psycho-social Dynamics and Perspectives to be published by
Psychology Press/Taylor and Francis.
IJM
31,3
336
To purchase reprints of this article please e-mail: reprints@emeraldinsight.com
Or visit our web site for further details: www.emeraldinsight.com/reprints