Life Events Effect on Employee

Organizational Withdrawal Behaviors

Anthony H. Brown
 
PhD Candidate in Psychology
 
 Walden University

November 2, 2014
 Abstract

 Bhagat’s life events model was evaluated for its ability to predict the effect of employees’ life events
on organizational withdrawal behaviors using cohort data from the ongoing National Longitudinal
Survey of Youth. The literature review revealed that research on the probable effects of life events on
OWBs is needed. The life event variables of marital, family, and health status were used as predictors
to estimate the criterion organizational withdrawal behaviors measured by current work status.
Participants with better health and economic status had a greater chance of currently working than
those with poorer health. The analysis did not show a significant association between marital status
and current working status, between residence and current working status, as well as between having
children and current working status. Additional analyses to determine whether crisis decisions and
uncertainty navigation impact organizational withdrawal behaviors were statistically insignificant.
Organizations that understand the impact of life events on organizational withdrawal behaviors likely
have more awareness of the ways to better manage employees who experience stressful situations,
and experiences that can have a negative influence on the work environment. The research is believed
to promote positive social change as its thought to help organizations see developmental opportunities
that reduces cost through training programs that allow employees to learn and apply evidence-based
interventions to manage the effects of negative life and lifestyle events, and learn how to value a
better life that could decrease imbalances between life and work.

Walden University 2
 Overview
 The study proposed that, Hanisch’s (2000) characterization of employees’
Organizational Withdrawal Behaviors (OWBs) was behaviors not necessarily
leading to turnover, but as behaviors interfering with day-to-day situational
and environmental processes. Porter and Steers (1973) suggested that to
conceptualize employees’ OWBs life events would need be framed as the IVs
rather than moderators to test the process of OWBs. Bhagat (1983) articulated
that life events lead to OWBs, and called for a new theory that would further
the research on their relationship.
 The theories CDT and UNT were used to address the call for a new theory.
Their use was as moderators predicted to enhance the relationship between the
decision processes to a negative event, and the array of strategies people use
to navigate difficult life patterns changes on OWBs (Sweeny, 2008; Sweeny
& Cavanaugh, 2010). Hanisch (2000) contended that research shedding light
on what contributes to employees’ OWBs remains sparse. The current
research sought to establish if employees’ life events contributed CDT and
UNT that ultimately impact employee OWBs.   

Walden University 3
 Literature Summary
 Hulin’s (1969) and Porter and Steers (1973) posited that researchers should
investigate life events as the independent variable.

 Bhagat (1983) suggested an alternative theory could add support for the
prediction of life events effect on employees’ performance.

 Hanisch and Hulin (1990) developed theoretical articles that posited

aggregated data could operationally identify withdrawal behaviors, and define
them as OWBs.

 Ostermann et al (1999) developed an international multidimensional

Situation, Work, and Self (SWS) survey model that used measurable variables
to, link stressors individuals experience from their life styles are brought to
work, which affects work demands, thus impact employees’ occupational
stress, health, and work behaviors.

Walden University 4
 The Problem

 Three research questions are staged to address the problem that was not
thoroughly reviewed in the past 30 years on organizational literature. Each were
assumed to support Bhagat’s and Hanisch’s proposition that the problem is more
research is needed to outline the neglected relationships of life events on
employee OWBs. They serve as three research questions, the basis for this study’s
research, design, and analytic argument: (1) Do the marital, family, health, and
life event changes of employees impact OWBs? (2) Do the marital, family, health,
and life event changes of employees influence attitudes that impact OWBs? and
(3) Do the marital, family, health, and life event changes of employees influence
expectations that impact OWBs?

Walden University 5
 Nature of the Study
 NLYS79’s original survey comprised 12,686 participants collected over 35
years on employees’ lifespan data starting at late adolescence. The data
reduction process, resulted in 5,121 respondents (40.4%), limiting the study to
an overall sample size of 7,656 participants, however this sample was
sufficient to address the study RQs and test their respective hypotheses. Only
58.9% of the original 1979 respondents participated in 2010, with greater
retention of female (62.0%) than males (57.2%) in the sample.

 Participant attrition over two decades had a greater drop among Caucasians
(49.8%) than African Americans (31.1%) and Hispanics (19.1%). Participants
under 50 years old at the time of interview in 2010 were excluded from the
study. These specific variable responses were limited the study to a set age
demographics (Personal Communication, Steve McClaskie, January 25,
2014).

Walden University 6
 Nature of the Study
cont.

 Life events were measured by employees marital status (married, never married, separated,
divorced, and widowed); family status (children or no children); residence (rural and
urban); and poverty status (poor or not poor); and health status (excellent, very good, good,
fair, and poor). OWBs was measured by the employees 'current working status as employed
or unemployed as of the 2010 interview phase. Descriptive statistics of selected NLSY79’s
data, identified that the employees with medium to high job satisfaction and the reason for
job status change would not be due to an employer’s decision (e.g., termination, layoff,
etc.). The latter of which established the basic criterion for OWBs data collection from
NLYS79’s work history response data that was subcategorized as labor status, time/tenure,
and experience items.
 Moderators included participants’ attitudes, which was an associated measure of
 CDT and their mean chronic condition and self-esteem score. In addition to CDT,
participants’ expectation was included as an associated measure of UNT and the mean SF-
12 physical and mental score respectively.
 The respective null hypotheses for each of the RQs were assessed using logistic regression,
and is what is being reported next.

Walden University 7
 Research Questions & Hypotheses
RQ1: Do marital, family, and health status life events changes impact
employee OWBs?
 H01: The marital, family, health, and life event changes of employees, as measured by
demographic item responses; do not impact OWBs, as measured by labor status, time,
tenure, and experience item responses.

RQ2: Do marital, family, and health life event changes influence employee
attitudes to impact reasons for employee OWBs?
 H02: Employee marital, family, and health life event changes, as measured demographic
item responses, do not influence employee attitudes (CDT), as measured by household item
responses, and do not impact employee OWBs, as measured by labor status, time/tenure,
and experience item responses.

RQ3: Do marital, family, and health life event changes impact employee
expectation to impact reasons for employee OWBs?
 H03: Employee marital, family, and health life event changes, as measured demographic
item responses, do not impact employee expectations (UNT), as measured by health item
responses, and do not impact the reasons for employee OWBs, as measured by labor status,
time/tenure, and experience item responses.
Walden University 8
 Purpose

 The intent of the study was to research, test, and accept the hypotheses that the
life events of employees do not predict OWBs. The research provided an
understanding of the likely origin and nature of OWBs, as operationally defined
by Hanisch (2000). It also posited that the CDT and the UNT does not modify life
events impact on employees OWBs. The modifier was meant to modify the
relationship of the IVs on the DV (Hosmer & Lemeshow, 2000). In this study, life
events (IVs) were used to predict employees’ OWBs (DV), with CDT and UNT
serving as the influencers (i.e., modifiers of the predicted impact).

Walden University 9
 Theoretical Framework
 CDT and UNT were hypothesized to modify the prediction that life events effects impact
employee OWBs. The inferred prediction was that CDT and UNT were important moderators
that could represent the employees’ psychological process when managing life events effects.
These effects may enhance the likely results employers perceive as the employees’ discontent
behaviors (i.e., OWBs). Sweeny (2008) and Sweeny and Cavanaugh’s (2010) theories furthered
the etiology of life events theoretical research, by implying that coping processes associated
with stressors, locus of control, mental and physical security, are descriptive of crisis decision-
making and often lead to uncertainty navigation hindrances (Folkman & Lazarus, 1985;
Lazarus & Folkman, 1987; Say et al, 2006; Wortman & Silver, 1989).

 CDT infers to factors of the information processes into three stages; “(a) accessing the severity
of the event, (b) determining response options, and (c) evaluating response options” (Sweeny,
2008, p. 61).
 UNT infers to “strategies people use to navigate uncertainty when awaiting information about
their health: (1) consequence mitigation; (2) reappraisal, and (3) emotion regulation.” (Sweeny
& Cavanaugh, 2010, p. 2).

Walden University 10
 Assumptions
 Several evaluated assumptions were found related to the database,
some were tied to the underlying external validity of the CDT and
UNT usefulness in the models. Others were linked to the internal
validity use of logistic regression (LR) modeling of the archival
database.
 The selected archival data sets provide viable source of data to
answer the RQs.
 CDT and the UNT are underlying premises of the effect of life
events on OWBs.
 The database is robust and sufficient (i.e., has multiple levels) to
conduct a LR analysis.

Walden University 11
 Limitations

 Using an archived database imposed some limitations because the

researcher has no control over how the data that was collected or
from whom they were collected, no access to the respondents, and
no ability to select or influence the instrument and/or the scoring
and coding of the data. These limitations have some impact on the
ability of the researcher to conduct a systematic analysis.
 For instance, default data cleaning methods, transformation, and
participant reduction impacts imposed limitations on the viability
of the data set. Limits imposed were due to missing data and
outliers, and uncertainty in determining whether the limitations
affected the reliability of the study.

Walden University 12
 Delimitations

 In spite of this study’s assumptions, and limitations, the study

sought to offer organizations a perspective that aligns the impact of
life events on employees’ behaviors that cannot be explained by
observing discontent behaviors. In addition, the study expanded on
research by Bhagat (1983) and Hanisch (2000) by integrating a
targeted focus on life events and undesirable behaviors by
employees, useful positive social change could be identified.

Walden University 13
 Data Analysis and Results
A preliminary breakout followed the variable selection and descriptive analysis of the 2010
NLSY79 archived dataset. Life event changes were hypothesized in terms of being better
predictors (Model Fit) over the (Null or Intercept) only model (i.e., the intercept only model,
sometimes called the null model, defines the effect of the model without predictors). Akaike
Information Criterion (AIC), was used to determine the model fit of the null model with
predictors. AIC , is a penalizing method for estimating goodness of fit based on the lowest
parameter’s affect. AIC determined based on the RQ1 that all five predictor variables had an
overall better fit over the intercept (null) model. A Chi-Square test substantiated that finding (p.
< 0.001).

The tables displays predictive variable results depicting the beta coefficient (odds) the odds-ratio
(OR), and 95% confidence interval (CI) of the OR, and the model p-values, including which
variables were used as reference, interaction terms, and variable categories. Reference categories
are the variables being compared for example, if in gender variable 'female' is a reference. The
reference was compared (i.e., the 'male' category with 'female' while holding constant the
reference in interpreting the results).

Walden University 14
 Predictors vs. Null Model
Results of the LR for RQ1 showed that after
adjusting for the effects of marital status,
poverty status, residence, and have children,
participants with excellent health had more
than 4.5 times higher odds than participants
with poor health status of currently working.
Similarly, participants with very good health
status and good health status had 4.35 times
and 3.46 times greater odds, respectively,
than participants with poor health status of
currently working. Participants with fair
health status had more than 2.71 times higher
odds than participants with poor health status
of currently working. The association
between health status and currently working
status had a dose-response relationship and
were statistically significant (p < .01). In
addition, after adjusting for the effects of
health, poor participants had 75% lower odds
than not poor participants of currently
working. The association between poverty
status and current working status was
statistically significant (p < .01).

Walden University 15
 Crisis Decision on OWBs
An LR was run to evaluate RQ2, with an
interaction term between health status
and chronic condition score was included
in the regression model to test the
influence of a number of chronic
conditions on the relationship between
health status and current work status.
Similar to previous results (shown in
Table 26), this regression model also
showed a statistically significant
relationship between health status and
current working status as well as between
poverty status and current working
status. However, the interaction terms
included between health status and
chronic condition score in the regression
model to test the influence of a number
of chronic diseases on relationship
between health status and current work
status were not statistically significant

Walden University 16
 Uncertainty Navigation on OWBs
For RQ3, an LR model, with interaction
term, SF-12 physical composite score,
marital status, poverty status, current
residence, and have children as the IVs,
was conducted. The association between
SF-12 physical composite score and
current employment status was significant
(p = .01). The interaction between health
status and SF-12 physical component
score was included in the model to
evaluate their effect on current
employment status. The model’s
interaction terms showed that the
interaction between health status and the
SF-12 physical component score was not
significant, suggesting that the SF-12
physical composite score did not influence
the association between health status and
current working status.

Walden University 17
 Test of the Hypotheses
 Moderators included participants’ attitudes, which were associated with CDT (chronic condition and self-
esteem) and participants’ expectations, which were associated with UNT (physical and mental ability).

 A summary LR model has been fitted to display whether the RQ1, model without variables versus with
variables was scientifically significant to predict a better fit. A likelihood ratio test showed that the model
with all IVs was significantly different from the null model.
Comparing the Null Model With the Model With all IVs for RQ1
Model Residual df Residual dev df Deviance Pr (> Chi)
Null model 2318 2495
Model with all IVs 2306 2377 12 118.1 <0.001*
2
Note. Chi-square distribution (χ K = df < 0.001)
 McFadden’s R2 was used to check the R2 inferences for the model with and without interaction; terms
for RQ3, SF-12 mental score . McFadden’s R2 for the model without the interaction term = 0.0499, or
~5%, whereas with the interaction term = 0.0485, or ~4.9%. A comparison of the two R2 values
suggested that the model with the interaction term did not explain any additional variance in the outcome
variable. A likelihood ratio test of RQ3 showed the model with the interaction term was not a better fit (p
= .49).

Walden University 18
 Interpretation of Findings
RQ1: Do marital, family, and health status life events changes impact employee OWBs?
 The LR model resulted in the null hypothesis being rejected because it suggested
employees’ with better health and economic status had a greater chance of currently
working than those with poorer health.
RQ2: Do marital, family, and health life event changes influence employee attitudes to impact
reasons for employee OWBs?
 The LR model resulted in the null hypothesis being accepted for failing to show that
chronic condition representing employee attitude influenced by life events did not add to
the impact of OWBs.
RQ3: Do marital, family, and health life event changes impact employee expectation to impact
reasons for employee OWBs?
 The LR modeled resulted in the null hypothesis being accepted for failing to show
perceptions of uncertainty navigation have a moderator effect on the potential effect of
life events impact to employees OWBs.

Walden University 19
 Interpretation of Findings
cont.
 Hanisch (2000) contended: (1) Independent behavioral  The results showed that RQ2 an employee’s attitude
form model where employee withdrawals are toward being healthy does not affect the quality of
indicative of individual processes independent of decision making to access life events as a crisis or to
antecedents like absenteeism, whereas, OWBs are set off OWBs responses such as presenteeism, work
implied to be impacted by life events, and moderated avoidance behaviors.
by CDT or UNT, and (2) Compensatory behaviors  The results showed that RQ3 employees’ expectation
withdrawal model where absenteeism establishes a of pending effects of health status outcomes did not
temporal act to separate from organizational. OWBs carry forward any perceptions of uncertainty
operationally define a temporal effect and not a navigation that would imply a moderator effect on the
terminal effect, such as turnover. potential effect of life events impact to employees
OWBs. The analysis outcome also was not conclusive
 Interpretation of the results could imply that OWBs that employees’ expectations would result in either
exist as a negative asset, whereas life events such as physical or mental assessments that life events might
health and economic status might lead some impact the employees’ general behavior to brace for
employees’ to engage in presenteeism, temporary bad news (Sweeny & Shepperd, 2007). That is
absences from work or at work, but not fully present employee marital, family, health or life events changes
because of physical or mental hindrances (i.e., did not influence expectations that impact OWBs or set
stressors). The importance of the RQ1 results to the off presenteeism
study was that they determined whether to abandon the  Possible conclusion of RQ2 and RQ3 is that the impact
analysis or continue, a critical test, at that juncture of of the limitations hindered the ability to conduct a
the analysis, which could have directly impacted the systematic analysis of responses associated with these
research as undeterminable. interaction term.

Walden University 20
 Implication for
Future Recommendations
Social Change
 Organizations that understand the impact of life events
on OWBs likely are more aware of ways to better
manage employees who experience stressful situations  Organizational researchers should examine metrics
or other experiences that can have a negative influence
linking employees’ OWBs to performance and
on the work environment. They also see the benefit of
retention analysis. The modeling approach applied by
developmental opportunities that could reduce the cost
this research introduced a method for the scientific
of managing turnover by offering training programs that
application of a clinical theory that OWBs
allow employees to learn and apply evidence-based
interventions to manage the effects of negative life operationally defines a process toward turnover. OWBs
events, thus decreasing imbalances between life and are a response action impacted by multiple antecedents
work. of life events.

 Organizational interventions should be based on  Future, researchers will want to access research using
identifying methods of classifying and labeling archival data, secondary data, or primary data that can
problems that negatively affect employee developmental advance this study. Future researchers might choose to
processes and work. Training management personnel on address the limitations by enhancing the capturing of
the benefits of having targeted assistant programs data points to provide better classification of life events
focused on how to best identify decreases or sharp impact on OWBs. Researchers can accomplish this by
increases in employees accountability that may affect designing a survey questionnaire or use secondary
employee long-term task sustainability. Identified database built to explore participant responses.
features would be pushed to being reshaped into a
piloted program called developmental improvement
protocols.

Walden University 21
 Conclusion

 The study established that when framing life events as the IVs,
employee OWBs are not only impacted by job dissatisfaction,
organizational demands, or the lack of organizational justice but
also support the measure of multiple antecedents, narrow
identification of the problem, and aid in classification of the
problem. Overall, how organizations frame (i.e., identify or
classify) a problem will determine the approach to resolve or
reduce the effect of the problem. Bhagat (1983) and Hanisch
(1995) suggested that researchers need new technologies as well
as research designs and methods that go beyond single-item
measures to study employee OWBs.

Walden University 22
 Conclusion
cont.

 The study attempted to respond to this call and form applied

approaches and introduce new theories that extend life events
research on OWBs into a new research design not perceived
during the original hypotheses of what operationally defined
OWBs. The current study is but the first step in defining,
understanding, or forecasting the factors that might impact
employee OWBs. One final recommendation is to continue
research on the impact of employees’ life event history and
styles and their association with employee work
dysfunctions.

Walden University 23
 Summary

 Life events were introduced as a likely impact on OWBs, yet most of the
research has focused entirely on undesirable job behaviors (Hanisch, 2000).
There exist probable links among life events, the theoretical framework, and
employee OWBs. The underlying value of crisis decisions and uncertainty
navigation was believed a subliminal effect derived from the outcome of
negative life events, it was further believed to be a definitive link with OWBs
(Bryant, Moulds, & Guthrie, 2000; Cosentino, Chute, Libon, Moore, &
Grossman, 2006). This research was structured to isolate this impact so to
promote a positive social change in how employers assess causes of turnover
and to provide wider lens onto the ways life events impact the psychology of
work conditions that affect employees’ OWBs.

Walden University 24
 References
Bhagat, R. S. (1983). Effects of stressful life events on individual performance effectiveness and work adjustment
processes within organizational settings: A research model. Academy of Management Review, 8(4), 660-671.
doi:10.5465/ AMR.1983.4284672
Bryant, R. A., Moulds, M. L., & Guthrie, R. M. (2000). Acute Stress Disorder Scale: A self-report measure of acute
stress disorder. Psychological Assessment, 12(1), 61-68. doi:10.1037/1040-3590.12.1.61
Cosentino, S., Chute, D., Libon, D., Moore, P., & Grossman, M. (2006). How does the brain support script
comprehension? A study of executive processes and semantic knowledge in dementia. Neuropsychology, 20(3), 307-
318. doi:10.1037/0894-4105.20.3.307
Creswell, J. W. (2003). Research design: Qualitative, quantitative, and mixed methods approaches. Social Sciences
Research Methodology, (2nd Ed.). University of Nebraska, Lincoln, Sage.
Creswell, J. W. (2009). Research design: Qualitative, quantitative, and mixed methods approaches (3rd ed.). Thousand
Oaks, CA US: Sage Publications, Inc.
El-Habil, A. M. (2012). An application on multinomial logistic regression model. [Article]. Pakistan Journal of
Statistics & Operation Research, 8(2), 271-291. Retrieved from the Academic Search Complete database.
Folkman, S., & Lazarus, R. S. (1985). If it changes, it must be a process: Study of emotion and coping during three
stages of a college examination. Journal of Personality and Social Psychology, 48(1), 150-170. doi:10.1037/0022-
3514.48.1.150
Hanisch, K. A., & Hulin, C. L. (1990). Job attitudes and organizational withdrawal: An examination of retirement and
other voluntary withdrawal behaviors. Journal of Vocational Behavior, 37(1), 60-78.

Walden University 25
 References
Hanisch, K. A. (2000). The impact of organizational interventions on behaviors: An examination of models of withdrawal. In D. R. Ilgen
& C. L. Hulin (Eds.), Computational modeling of behavior in organizations: The third scientific discipline (pp. 33-68). Washington,
DC: American Psychological Association. doi:10.1037/10375-003
Hosmer, D. W., & Lemeshow, S. (2000). Applied logistic regression (2nd ed.). New York, NY: Wiley
Hulin, C. L. (1969). Sources of variation in job and life satisfaction: The role of community and job-related variables. Journal of Applied
Psychology, 53(4), 279-291. Retrieved from PsycARTICLES database.
Lazarus, R. S., & Folkman, S. (1987). Transactional theory and research on emotions and coping. European Journal of Personality, 1(3),
141-169. Retrieved from Academic Search Complete database.
Ostermann, R. F., Gutierrez, R. E., Brody, E. R., Chell, R., Ellis, G., Rashid, T., . . . Spielberger, C. D. (1999). The SWS-international
survey of work stress and mental health: Symposium. American Psychological Association (APA), Public Interest Directorate, 5-6.
doi:10.1037/e359012004-005
Porter, L. W., & Steers, R. M. (1973). Organizational, work, and personal factors in employee turnover and absenteeism. Psychological
Bulletin, 80(2), 151-176. doi:10.1037/h0034829
Say, R., Murtagh, M., & Thomson, R. (2006). Patients' preference for involvement in medical decision making: A narrative review. Patient
Education & Counseling, 60(2), 102-114. doi:10.1016/j.pec.2005.02.003
Sweeny, K. (2008). Crisis decision theory: Decisions in the face of negative events. Psychological Bulletin, 134(1), 61-76.
doi:10.1037/0033-2909.134.1.61
Sweeny, K., & Cavanaugh, A. G. (2010). Waiting is the hardest part: A model of uncertainty navigation in the context of health news.
Health Psychology Review, 1-18. doi:10.1080/17437199.2010.520112
U.S. Department of Labor. (2010a). National Longitudinal Survey of Youth 1979 cohort, 1979-2008 (Rounds 1-24). Retrieved from
http://www.bls.gov/
Wildemuth, B. M. (1993). Post-positivist research: Two examples of methodological pluralism. Library Quarterly, 63(4), 450-468.
doi:10.2307/4308866
Wortman, C. B., & Silver, R. C. (1989). The myths of coping with loss. Journal of Consulting and Clinical Psychology, 57(3), 349-357.
doi:10.1037/0022-006x.57.3.349
Walden University 26