PreprintPDF Available

Belonging Affirmation Reduces Employee Burnout and Resignations in Front Line Workers

Preprints and early-stage research may not have been peer reviewed yet.

Abstract and Figures

Burnout, an increasingly important challenge in modern workplaces, has been associated with negative health outcomes, high turnover, and poor organization performance. Yet, we have limited causal evidence about how to mitigate its risk. We posit that workers are more susceptible to burnout if they feel undervalued in their roles. Drawing on evidence from educational psychology, we hypothesize that increasing social belonging and support within the undervalued group may reduce workplace identity threat and subsequently affect burnout and turnover. Consistent with these hypotheses, we first find that worker burnout is significantly correlated with feeling undervalued. Second, in a multi-site field experiment (n=536), we find that a six-week intervention that primes social belonging within the low-status group significantly reduces burnout by about 8 points (0.4 SD) and resignations by more than half (3.4 percentage points). Our findings suggest that light-touch belonging-affirmation exercises can improve worker and organizational well-being. Introduction:
Content may be subject to copyright.
Belonging Affirmation Reduces Employee Burnout
and Resignations in Front Line Workers
Elizabeth Linos, PhD1, assistant professor; Krista Ruffini, MPP, MA1 ; Stephanie
Wilcoxen, MPA2
1 Goldman School of Public Policy, University of California, Berkeley
2 Behavioral Insights Team, Brooklyn, NY
Burnout, an increasingly important challenge in modern workplaces, has been associated with
negative health outcomes, high turnover, and poor organization performance. Yet, we have
limited causal evidence about how to mitigate its risk. We posit that workers are more susceptible
to burnout if they feel undervalued in their roles. Drawing on evidence from educational
psychology, we hypothesize that increasing social belonging and support within the undervalued
group may reduce workplace identity threat and subsequently affect burnout and turnover.
Consistent with these hypotheses, we first find that worker burnout is significantly correlated with
feeling undervalued. Second, in a multi-site field experiment (n=536), we find that a six-week
intervention that primes social belonging within the low-status group significantly reduces burnout
by about 8 points (0.4 SD) and resignations by more than half (3.4 percentage points). Our
findings suggest that light-touch belonging-affirmation exercises can improve worker and
organizational well-being.
Keywords: employee burnout, social belonging, turnover, field experiment
Employee burnout -- a workplace syndrome characterized by emotional exhaustion, cynicism,
and reduced personal accomplishment is associated with poor worker health, high turnover,
and poor organizational performance (Borritz, Rugulies, Christensen, Villadsen, & Kristensen,
2006; Salvagioni et al., 2017; Dai, Milkman, Hofmann, & Staats, 2015). Recent evidence suggests
that 1 in 5 employees are burnt out (Wigert & Agrawal, 2018), costing the U.S. health-care system
$125 to $190 billion per year (Goh, Pfeffer, & Zenios, 2016). Among frontline workers like doctors,
social workers, or law enforcement officers, burnout rates can be significantly higher (Garman,
Corrigan, & Morris, 2002; Lloyd, King, & Chenoweth, 2002; Shanafelt et al., 2010; Wallace,
Lemaire, & Ghali, 2009).
While employee burnout appears to generally be on the rise, there is ample evidence that
workplace conditions and job resources correlate with levels of reported burnout. In particular, the
Job Demands-Resources (JD-R) model posits that having high levels of social support and strong
relationships can help buffer workers against burnout in environments with high job demands
(Bakker, Demerouti, & Euwema, 2005; Broadhead et al., 1983; Hakanen, Bakker, & Demerouti,
2005). This suggests that social connectedness may be a protective resource against burnout. In
parallel, a separate literature has shown that when people are uncertain about whether they
belong -- when their social connectedness is threatened -- they may view adversity as a threat
to their identity, and subsequently perform poorly (Walton & Cohen, 2007).
A fundamental question then is how to build social connectedness and strengthen workers’ sense
of belonging, particularly for individuals who may face adversity at work. The existing literature
suggests self-affirmation exercises may be one strategy to protect against identity threat. Based
on the theory that people are motivated to protect their sense of self-worth (Steele, 1988), these
techniques ask people to reflect on their values and positive attributes. Affirmation exercises have
been shown to alleviate stress (Creswell et al., 2005), improve academic performance of
stigmatized groups (G. L. Cohen, Garcia, Purdie-Vaughns, Apfel, & Brzustoski, 2009), and reduce
avoidance of potentially threatening health information (Howell & Shepperd, 2012). Recent
evidence also indicates that self-affirmation exercises are most effective when the respondent
reflects on his or her social bonds. For example, when students are asked to affirm their values,
the focus of their reflections often lands spontaneously on valued social relationships (Creswell
et al., 2007). This may be because our sense of identity is often developed as a reflection of social
relationships (Mead 1934). Therefore, explicitly affirming social bonds may shield individuals
against identity threat and reduce the negative impact of adversity on self-worth; social belonging
may be the “key ingredient” of why affirmation exercises are so effective (Shnabel, Purdie-
Vaughns, Cook, Garcia, & Cohen, 2013). Put differently, the education psychology literature
suggests that self-affirmation exercises may be most effective when they are in fact belonging-
affirmation exercises. We build upon this literature by examining whether belonging-affirmation
could mitigate the specific type of identity threat and adversity that people face at work.
We posit that workplace identity threat depends on perceptions of how others view a specific
workplace role. If employees feel undervalued, because they believe others do not understand or
appreciate the role that they play in a work environment, they may be particularly likely to
experience workplace identity threat when facing adversity. If people derive self-worth from their
professional roles, they can also face collective threat when the image of their profession as a
whole is threatened (Cohen & Garcia, 2005). For example, nurses may feel undervalued relative
to physicians in a health setting; administrative staff may feel undervalued relative to tenured
professors in an academic environment; and (in the case we study here) 911 dispatchers may
feel undervalued relative to police officers in law enforcement. This prediction is consistent with
studies that conceptualize devaluation as a form of social rejection that threatens one’s sense of
belonging (Richman & Leary, 2009).
On the other hand, a strong sense of social belonging and support within the undervalued group
may serve as a resource to help workers cope with workplace identity threat. For example, we
know that increasing identification with their own gender group helps women cope with a “chilly
climate” in STEM fields (Walton, Logel, Peach, Spencer, & Zanna, 2015). We hypothesize that
through a similar mechanism, belonging-affirmation exercises will strengthen in-group social
relationships and subsequently reduce worker burnout, especially for workers who face workplace
identity threat.
The population we study, 911 dispatchers, have inherently stressful jobs. These workers answer
approximately 2,400 calls a year (NENA, 2019; BLS, 2018), and often have to make life-or-death
decisions in minutes, while also managing the emotions of call-makers and communicating
quickly and accurately with officers they dispatch. Call dispatchers have been dubbed the
“forgotten victim” when it comes to stress in policing (Sewell & Crew, 1984) because although
they face much of the same trauma as emergency responders or law enforcement officers, they
are considered clerical workers, and so do not get as much formal support, benefits or recognition
for their work. Recent studies report high levels of alcohol abuse, PTSD, and depression among
this population (Lilly, London, & Mercer, 2016). In qualitative interviews, 911 dispatchers and their
supervisors noted that burnout was commonly discussed as “part of the job,” but was also a
common reason why people leave the job.
We conduct two studies to examine the relationships among social belonging, workplace identity
threat and burnout among this population. Study 1 establishes the hypothesized relationship
between burnout and workplace identity threat through a survey of 911 dispatchers in one major
US city (n=129). Study 2 experimentally tests a six-week belonging-affirmation intervention in a
multi-site randomized control trial (n=536), and measures the impact on reported burnout and
turnover over a six month period.
In order to quantify the extent of burnout among 911 dispatchers and to examine the association
between perceived social support and burnout, we sent a survey to all 911 dispatchers and call-
taker in the Los Angeles Police Department, the third largest police force in the United States
(n=532). The response rate on the survey was 30 percent.
Burnout is measured using the Maslach Burnout Inventory (MBI) Human Services Survey. The
MBI is a validated questionnaire of 22 items designed to measure burnout on three dimensions:
emotional exhaustion, depersonalization, and personal accomplishment (Maslach and Jackson,
1981). Respondents receive a score based on a Likert scale for each question. For the purposes
of this analysis, we used a composite score aggregating across the three dimensions, with
“personal accomplishment” reversed so that higher scores are associated with negative
outcomes. This composite measure ranges from 0 to 125 points.
The survey also included questions on social belonging and social support, drawing from previous
studies on social belonging and social support ((Walton & Cohen, 2007); (Wright, Burt, &
Strongman, 2006) (Richer & Vallerand, 1998); (Chen, Gully, & Eden, 2001)). The full
questionnaire can be found in the supplementary information.
Survey responses were also linked to a separate administrative data set of employee
demographics. As such, as we can control for employee gender, race and ethnicity, as well as
tenure on the job. We conduct multivariate regression analysis to examine the association
between perceived social belonging and burnout in this population, controlling for employee
demographic characteristics.
Table 1 below shows the survey population is overwhelmingly (83 percent) female and identifies
as a racial or ethnic minority (83 percent). On average, dispatchers are 40 years old and have
been working in their current role for 13 years. Few are college-educated: 62 percent have no
more than a high school education. The demographics of the population who completed the
survey match the demographics of the overall LAPD dispatcher workforce with no statistically
significant differences in race, gender, education, age, or tenure.
Burnout, particularly on the emotional exhaustion and depersonalization scales, is high among
our study population, both in absolute and relative terms. Almost 43 percent exhibit high levels of
burnout across all measures, and these rates are high relative to those found among other front-
line occupations, including nurses (Greenglass, Burke, & Fiksenbaum, 2001; Poncet et al., 2007),
physicians ((Embriaco, Papazian, Kentish-Barnes, Pochard, & Azoulay, 2007), and teachers
(Hakanen, Bakker, & Schaufeli, 2006).
Table 1: Demographic characteristics
(Study #1)
Years employed
<= HS education
MBI score, emotional exhaustion
MBI score, depersonalization
MBI score, personal accomplishment
MBI score, composite
Table 2 shows that a greater sense of belonging and social support is significantly correlated with
less burnout. In particular, respondents who agreed with the statement “When something bad
happens, I feel that maybe I don’t belong in law enforcement” scored 13 points (0.56 SD) higher
on the MBI scale. In contrast, those who felt that they had a colleague they could speak to scored
13 points (0.57 SD) lower on the MBI scale. This suggests a strong association between
perceptions of burnout and broader social belonging and social support.
Table 2: Correlation between Social Belonging and Burnout
(Study #1)
MBI score
When something bad happens,
I feel that maybe I don’t belong in law enforcement.
There is someone at work I can talk to
about my day to day problems if I need to.
Average MBI score
% Agree
Notes: Table shows the association between perceived social belonging and
support, measured as agreeing with each statement, and burnout, measured as
the composite MBI score among 129 dispatchers in LA county (Study #1). Robust
In order to test the hypothesis that burnout is associated with whether workers believe that others
value their role, we also asked dispatchers whether different groups understand the role that
dispatchers play, and whether other groups value the work they do. Fig.1 shows that while nearly
two-thirds of respondents (65 percent) strongly agreed that other dispatchers understand the role
of dispatchers in public safety, the majority (54 percent) strongly disagreed that the general LA
community understood their role. Perceptions of police officers’ beliefs took a middle ground, with
about half of the sample either moderately or strongly agreeing that police officers understood
PSRs’ roles. This is an important category because police officers are LAPD employees with
whom they interact regularly. In contrast, a much greater share (66 percent) felt valued by other
Figure 1: Perceptions of Understanding Dispatchers’ Role, by Group
Notes: Figure shows the fraction of respondents in Study #1 disagreeing or agreeing that
other dispatchers (PSR), police officers (PO), or the LA community (LA) understands the
role of dispatchers in law enforcement (N = 129).
In line with our hypothesis, Table 3 shows that a dispatcher’s perception of whether they are
valued and understood is significantly negatively correlated with burnout. Dispatchers who report
feeling understood by police officers and the LA community also report significantly lower levels
of burnout, measured by the MBI scale. It is noteworthy that perceptions of how police officers
view them is a stronger predictor of dispatcher burnout than perceptions of how peers view them.
This suggests that being in a low-status role may be crucial to the link between workplace identity
threat and burnout.
Table 3: Correlation between Feeling Understood and Burnout
(Study #1)
Recognize the role of PSRs in public safety
I feel valued by…
Police officer
Police officer
Average MBI score
% Agree
Strongl y
agree n or
agr e e
Strongl y
agr e e
Percent LA dispatchers responding groups
understand PSR's role
Notes: Table shows the association between feelings of being understood and valued by each
population in the column header, and burnout, measured as the composite MBI score among 129
dispatchers in LA county (Study #1). Robust standard errors in parentheses.
Although these survey results present a significant correlation between feeling undervalued,
feeling like you belong, and levels of burnout, these results do not establish a causal link between
social belonging and burnout. In order to examine whether promoting social belonging can
causally reduce burnout, Study #2 involved conducting a field experiment in nine mid-sized cities.1
Study #2 aimed to evaluate whether a low-cost, light-touch intervention aiming to affirm belonging
and strengthen the value of the dispatcher profession could reduce self-reported burnout and
increase retention. We collaborated with nine mid-sized US cities to conduct this experiment and
implemented a two-arm randomized controlled trial, stratified by city. All 911 dispatchers that were
employed at the time of randomization were included in the experiment (n=536).
To maximize statistical power, participants in each city were matched into pairs based on the
amount of sick leave they had taken over the previous six months. The research team randomized
members of each pair into treatment or control with equal probability.
A supervisor, or other leader selected by the department, sent weekly emails to the treatment
group in each city over a period of six weeks. Each email (after the first one) included two
components: a story from an existing dispatcher and a prompt nudging workers to reflect on their
common experiences as dispatchers. The emails were framed as an opportunity for participants
to give advice and support to newer dispatchers. The intervention aimed to have dispatchers
affirm social belonging by prompting respondents to consider how dispatchers could support and
provide resources for each other, therefore elevating the value of social belonging to the greater
network. On the online platform, members of the treatment group could read stories submitted by
all other dispatchers. Participants were blinded on the scope and nature of the intervention; the
control group received a simplified version of the first week’s email to inform them of the multi-city
collaboration, but did not receive additional emails. Table 1 in Supplementary Materials provide
complete email language for each week.
We measured the impact of the intervention using both survey burnout and administrative
retention data. The survey was conducted at three points: before the launching of the intervention,
immediately following the six-week intervention, and four months later. Administrative records
were collected for this full period as well. Fig. 2 outlines the timing of the intervention and follow-
up period.
1 Cities include Albuquerque, Cambridge, Glendale, Greensboro, Mesa, Portland, Salt Lake City, Tempe,
and West Palm Beach.
Figure 2: Social Belonging Intervention
(Study #2) Timing
Notes: Figure shows the timeline for the multi-city social belonging intervention. See text
for details.
Burnout is measured by the Copenhagen Burnout Inventory (CBI), a validated survey consisting
of 19 questions in personal-, work-, and client-related burnout categories.2 Participants receive a
score ranging from 0 to 100 based on a Likert scale for each question. We report on all three sub-
categories, as well as on total composite score immediately following the six-week intervention
and four months after the intervention had ended. We measured sick leave use and turnover over
the six-week treatment period and full six-month (24 weeks) period through March 2018 using
administrative data from each city. In the final survey, we also asked respondents whether they
felt their work provided a sense of personal accomplishment; whether they felt they had the skills
necessary for their position; and whether they felt connected to other dispatchers.
All statistical analyses were conducted using Ordinary Least Squares. For each individual i in city
c at time t, we measured the effect of the intervention on burnout and resignations,  as:
 =  + . ′ + + 
 is a binary variable equal to one if the individual was assigned treatment; ′ is a vector
of baseline demographic characteristics: race, gender, tenure, and pre-period leave; is a vector
2 We used the shorter CBI rather than the MBI for this population in order to minimize response burden and
cost to cities. Previous work has documented the two indices are highly correlated (Kristensen, Borritz,
Villadsen, & Christensen, 2005; Winwood & Winefield, 2004).
of city fixed effects;  is an i.i.d. error term. All inference is conducted using Eicker-Huber-White
robust standard errors.
In exploratory analysis, we also conducted subgroup analysis by gender and race. We are limited
to only exploring racial and gender differences using the survey data, given that not all cities
maintain demographic information in their administrative data.
Table 4, Panel A shows the mean baseline characteristics of all participants, separated by
treatment and control group. In cities for which demographic information is available in the
administrative data, more than three quarters of the sample is female, and nearly 80 percent is
white. Our sample is more female than the national average (54 percent), but similar on racial
composition. On average, participants had nearly 10 years of tenure and took an average of 1.5
hours of sick leave a week in the six months before the trial began. Chi-squared tests on each
variable confirm that treatment and control are balanced across these characteristics.
The average response rate across the interim and final burnout surveys was 28.6 percent. Table
4, Panels B and C show the treatment and control groups have similar demographic
characteristics in both survey waves, and these characteristics are also similar to the full
participant sample. These patterns mitigate concerns that change in reported burnout is not due
to systematic differences in who takes the survey, but rather reflects change in underlying burnout
in the treatment group.
Table 4: Demographic and baseline characteristics by treatment group
(Study #2)
Panel A: All trial participants
Years employed
Average pre-trial weekly sick hours
Panel B: Interim CBI burnout survey
Years employed
Panel C: Final CBI burnout survey
Years employed
Notes: This table reports means and standard deviations of the full (Study #2) sample in Panel A,
and interim (final) burnout survey respondents in Panel B (C). Columns (1)-(4) describe trial
participants -- the full ITT sample. Columns (5)-(8) describe those who remained employed
through the full trial. Column (1) summarizes the full sample, and columns (2) and (3) describe
the control and treatment groups, respectively. Column (4) reports p-values for the null
hypothesis of perfect randomization.
Table 5 presents the main burnout findings. Immediately after the trial ended, treated employees
reported lower levels of burnout, although differences are not statistically significant (panel A). Six
months after the first email was sent, however, the social belonging intervention appears to have
caused a significant reduction in burnout for each of the sub-categories and the composite score
(panel B). These coefficients are large in magnitude and statistically significant; the intervention
reduced personal-related burnout by about 9 points, work-related burnout by 7 points, and client-
related burnout by approximately 9 points. The composite burnout index fell more than 8 points,
or more than 0.4 standard deviations. This is similar in magnitude to the difference in average
reported burnout between social workers and administrative staff in other studies (Kristensen,
Borritz, Villadsen, & Christensen, 2005).
Table 5: Copenhagen Burnout Index, Full Sample
(Study #2)
Panel A: Interim
(after all emails sent)
Control group mean
Effect size
Panel B: Final
(4 months after all emails sent)
Control group mean
Effect size
Notes: This table reports OLS coefficient estimates robust standard errors
in parentheses). The dependent variable is the burnout score, either by
subindex (columns (1) - (3)) or overall (column (4)). Burnout score ranges
from 0 to 100. Panel (A) reports scores immediately after the final email
was sent in Study #2; Panel (B) reports scores 4 months after the final
email was sent. Treatment is a dummy for the social support treatment.
All specifications include demographic controls for race, gender, tenure,
as well as city fixed effects. * p < 0.10, ** p < 0.05, *** p < 0.01.
Table 6 presents the key resignation findings. We find that the intervention reduced resignations
among employees who remained employed throughout the intervention (i.e. those exposed to the
full treatment) by 3.4 percentage points in the four months post-intervention, relative to a control
group mean of 5.1 percent (column 1). This result is unchanged when we include controls for
demographic characteristics and fixed effects for each city, confirming that the randomization
process was successful. When we include resignations that happened during the intervention in
columns (3) and (4), point estimates suggest a smaller reduction in resignations and the effect
become statistically insignificant. As we discuss below, this apparent lagged response may be
due to requirements around providing notices of leave or the cumulative effect of receiving
additional emails. In results available upon request, we do not find any significant difference in
sick leave use.
Table 6: Resignations, by Treatment Dosage
(Study #2)
Full trial
Full trial
City FE
Control group
Notes: This table reports OLS coefficient estimates (robust standard errors in
brackets). The dependent variable is a binary variable equal to 1 if the employee
resigned his or her position in Study #2. Columns (1) and (2) includes all employees
assigned treatment or control; columns (3) and (4) limit the sample to those who
remained employed throughout the trial (week 7 post-
randomization). All
specifications include controls for pre-period leave; columns 2 and 4 additionally
include city fixed effects and demographic controls for race, gender, and tenure.
* p < 0.10, ** p < 0.05, *** p < 0.01.
Fig. 3 presents a Kaplan-Meier curve illustrating that treatment group resignations exceeded
those of the control group in the first few weeks of the trial (although not significantly), but in the
four months post-intervention resignations remained relatively flat for the treatment group, but
continued to increase among control group members. In additional results, we confirm these
findings are not driven by the experiences of any single city.
Figure 3: Kaplan-Meier Survival Plot for number of employees resigning their position by
treatment assignment
(Study #2)
If building social support reduces identity threat, we should observe particularly large reductions
among groups most susceptible to identity threat. Consistent with this hypothesis, in exploratory
analysis we find suggestive evidence that females and employees of color (employees identifying
as a racial/ethnic group other than non-Hispanic white) experienced particularly large reductions
in burnout on all scales (Table 7).
Table 7: Final Copenhagen Burnout Index Survey, by Gender and Race/Ethnicity
(Study #2)
Panel A: All workers
Control group mean
Panel B: Female employees
Control group mean
Panel C: Non-white employees
Control group mean
Notes: This table reports OLS coefficient estimates robust standard errors in
parentheses). The dependent variable is the burnout score 4 months after the final email
was sent, either by subindex (columns (1) - (3)) or overall (column (4)). Burnout score
ranges from 0 to 100. Panel (A) reports scores for the full sample in Study #2; Panel (B)
reports scores for female employees; Panel (C) reports scores for employees identifying
as non-white or Hispanic. Treatment is a dummy for the social support treatment. All
specifications include demographic controls for race, gender, tenure, as well as city fixed
effects. * p < 0.10, ** p < 0.05, *** p < 0.01.
In order to directly examine possible mechanisms for these findings, we included three key
questions enquiring about employees’ perceived sense of personal accomplishment at work, their
understanding of their own skills, and perceived connectedness to colleagues in the final survey
(Table 8). While point estimates suggest the treatment improved well-being, particularly among
women and employees of color (panels (B) and (C)), these findings are only significant for
employees of color. Although suggestive, these findings are consistent with our hypothesis: that
the intervention reduced identity threat and bolstered self-worth by strengthening workers’ sense
of social belonging and support. We note, however, that sub-group samples are small and we
lack statistical precision to make decisive statements on this point. Future research could explicitly
test for differential effects of this type of workplace intervention in samples with greater gender
and racial diversity.
Table 8: Perceived Accomplishment, Skills, and Connectiveness by Gender and
Race/Ethnicity (Study #2)
Panel A: All workers
Control group mean
Panel B: Female employees
Control group mean
Panel C: Non-white employees
Control group mean
Notes: This table reports OLS coefficient estimates robust standard errors in parentheses).
The dependent variable is self-
reported sense of accomplishment (column (1)); skills
(column (2)); and connectiveness (column (3)) 4 months after the final email was sent (Study
#2). Score ranges from 0 to 100. Panel (A) reports scores for the full sample; Panel (B)
reports scores for female employees; Panel (C) reports scores for employees identifying as
non-white or Hispanic. Treatment is a dummy for the social su
pport treatment. All
specifications include demographic controls for race, gender, tenure, as well as city fixed
effects. * p < 0.10, ** p < 0.05, *** p < 0.01.
The education psychology literature has established that the strength of one’s social relationships
and sense of belonging affects numerous outcomes. Yet, while the importance of belonging at
work and the alarming rise in burnout have both been widely recognized, there is limited evidence
that combines the psychological evidence around self-affirmation and belonging with the
organizational research on burnout. Our studies show a strong relationship between feeling
undervalued, low levels of social belonging, and high burnout at work. In an important contribution,
we conduct a field experiment which, to our knowledge, is the largest of its kind, and find that a
low-cost belonging-affirmation intervention can substantially reduce burnout and resignations.
This research makes several contributions to the literature. First, we provide causal evidence that
emphasizing social relationships within an undervalued group reduces burnout and turnover. This
helps inform how job demands and job resources interact at work and provides causal evidence
for the Job Demands-Resources (JD-R) model. We also document that facilitating greater job
resources can have persistent effects, and these benefits translate into improved organizational
performance in the medium-term. The reported rise in burnout has received significant attention
in both academic journals and practitioner press. Yet, previous reviews note that almost all studies
of burnout interventions are correlational (Maslach & Schaufeli, 2017) and that the existing RCTs
show mixed evidence (Westermann, Kozak, Harling, & Nienhaus, 2014). This may be because
many of the observational studies and RCTs involve small samples (usually fewer than 100
people), short interventions, limited follow-up, and self-reported outcomes (Awa, Plaumann, &
Walter, 2010). This study provides both methodological and substantive contributions to the
existing literature by using a much larger sample, testing a virtual light-touch intervention, and
measuring both self-reported well-being and actual employee behavior for a relatively longer
Finally, this study contributes to the broader literature on self-affirmation. Written reflections about
work have been associated with decreased stress and improved well-being (Bono, Glomb, Shen,
Kim, & Koch, 2013) and advice-giving seems to be most beneficial to the advice-giver (Eskreis-
Winkler, Milkman, Gromet, & Duckworth, 2019; Belle, 2014). Recent studies suggest that
relationships can be built through reciprocal sharing, even online (Carpenter & Greene, 2015). By
nudging workers to reflect on their relationship with peers and their professional identity and read
one another’s stories, this study tests how the tools of belonging-affirmation can be used to protect
employees against burnout.
We find reductions in burnout increase over time and the effect of our intervention on resignations
is clearest in the medium-term, post-intervention period. There are several explanations for this
finding. First, as employees are required to give at least two weeks advance notice on
resignations, it is likely that the resignations that occurred in the first weeks of the intervention
reflect decisions that happened pre-intervention, and so we expect to not observe a treatment
effect in early weeks. However, it is also possible that multiple emails (a more intensive
intervention) are required for the treatment to alter perceived self-worth and belonging, and
therefore, resignations. If there is path dependence in developing a sense of belonging, it is logical
that effects should appear over time and not immediately after an intervention ends (Yeager &
Walton, 2011). Future studies will need to further explore the relationship between burnout and
resignations to disentangle these explanations. Another limitation of the study is that we do not
capture spillovers to control group members. If improving the work environment for half of a
dispatch center benefits non-treated employees in the center as well, our estimates understate
the full effect of introducing such a social-belonging intervention organization-wide. Last, to
maintain anonymity, we do not know whether active participation (e.g., sharing stories) provides
larger benefits than passive participation (e.g., receiving emails).
The study has clear implications for policymakers and practitioners. First, in the ongoing debate
about whether burnout is best addressed at the individual- or systems-level (see for example
Wallace et al., 2009), we examine an effective behaviorally-informed intermediary solution where
systems are set up for individuals to support each other. Second, the study shows meaningful
cost savings for organizations due to reductions in turnover: scaling this nearly zero-cost
intervention to all employees could save a city with 100 dispatchers more than $170,000, or
approximately three full-time equivalent staff, per year. More broadly, given that workplace stress
is associated with more than 120,000 deaths per year and more than 5 percent of annual health
care costs in the U.S., encouraging a greater sense of social belonging may provide additional
savings that are not captured in this calculation (Goh et al., 2016). Future research should also
consider whether an intervention that successfully reduces burnout can have causal impacts on
daily decision-making. Ultimately, this study contributes to a growing literature demonstrating the
power and feasibility of running behaviorally-informed field experiments in real organizations on
real organizational outcomes (Hauser, Linos, & Rogers, 2017; Pfeffer, Pfeffer, & Sutton, 2000).
Our results suggest that even low-cost approaches to improving employees’ work experience can
meaningfully benefit both employees and their organizations.
Author Contributions
EL developed the study concept and study design, and coordinated approval with partner
agencies. KR performed data analysis. SW was responsible for data collection and trial
implementation. All authors interpreted the data, drafted the manuscript and approved the final
version of the manuscript for submission.
Declaration of Conflicting Interests
The author(s) declared that there were no conflicts of interest with respect to the authorship or
the publication of this article.
Open Practices
Study 1 is part of a larger ongoing experiment that is pre-registered on the
platform (#25812). Study 2 is registered on the Open Science Framework portal at Data for both studies belongs to government agencies and because of the
nature of employee data and the data sharing agreement that allowed for this project, any
requests for the data should be directed to the relevant government agencies. The materials used
are available in the appendix and any requests for the scripts used for the analysis can be sent
via email to the lead author at [removed for anonymity purposes].
Ethical Considerations
The study was reviewed and received IRB approval at UC Berkeley (Protocol ID # 2017-05-9957)
but was designed as a quality improvement study for the cities involved.
Bakker, A. B., Demerouti, E., & Euwema, M. C. (2005). Job resources buffer the impact of job
demands on burnout. Journal of Occupational Health Psychology, 10(2), 170180.
Borritz, M., Rugulies, R., Christensen, K. B., Villadsen, E., & Kristensen, T. S. (2006). Burnout as
a predictor of self-reported sickness absence among human service workers: prospective
findings from three year follow up of the PUMA study. Occupational and Environmental
Medicine, 63(2), 98106.
Broadhead, W. E., Kaplan, B. H., James, S. A., Wagner, E. H., Schoenbach, V. J., Grimson, R.,
… Gehlbach, S. H. (1983). The epidemiologic evidence for a relationship between social
support and health. American Journal of Epidemiology, 117(5), 521537.
Chen, G., Gully, S. M., & Eden, D. (2001). Validation of a New General Self-Efficacy Scale.
Organizational Research Methods, 4(1), 6283.
Cohen, G. L., & Garcia, J. (2005). “I Am Us”: Negative Stereotypes as Collective Threats. Journal
of Personality and Social Psychology, 89(4), 566582.
Cohen, G. L., Garcia, J., Purdie-Vaughns, V., Apfel, N., & Brzustoski, P. (2009). Recursive
Processes in Self-Affirmation: Intervening to Close the Minority Achievement Gap. Science,
Vol. 324, pp. 400–403.
Creswell, J. D., Lam, S., Stanton, A. L., Taylor, S. E., Bower, J. E., & Sherman, D. K. (2007).
Does self-affirmation, cognitive processing, or discovery of meaning explain cancer-related
health benefits of expressive writing? Personality & Social Psychology Bulletin, 33(2), 238
Creswell, J. D., Welch, W. T., Taylor, S. E., Sherman, D. K., Gruenewald, T. L., & Mann, T. (2005).
Affirmation of personal values buffers neuroendocrine and psychological stress responses.
Psychological Science, 16(11), 846851.
Dai, H., Milkman, K. L., Hofmann, D. A., & Staats, B. R. (2015). The impact of time at work and
time off from work on rule compliance: the case of hand hygiene in health care. The Journal
of Applied Psychology, 100(3), 846862.
Embriaco, N., Papazian, L., Kentish-Barnes, N., Pochard, F., & Azoulay, E. (2007). Burnout
syndrome among critical care healthcare workers. Current Opinion in Critical Care, 13(5),
Garman, A. N., Corrigan, P. W., & Morris, S. (2002). Staff burnout and patient satisfaction:
evidence of relationships at the care unit level. Journal of Occupational Health Psychology,
7(3), 235241.
Goh, J., Pfeffer, J., & Zenios, S. A. (2016). The Relationship Between Workplace Stressors and
Mortality and Health Costs in the United States. Management Science, 62(2), 608628.
Greenglass, E. R., Burke, R. J., & Fiksenbaum, L. (2001). Workload and burnout in nurses.
Journal of Community & Applied Social Psychology, Vol. 11, pp. 211215.
Hakanen, J. J., Bakker, A. B., & Demerouti, E. (2005). How dentists cope with their job demands
and stay engaged: the moderating role of job resources. European Journal of Oral Sciences,
113(6), 479487.
Hakanen, J. J., Bakker, A. B., & Schaufeli, W. B. (2006). Burnout and work engagement among
teachers. Journal of School Psychology, 43(6), 495–513.
Howell, J. L., & Shepperd, J. A. (2012). Reducing information avoidance through affirmation.
Psychological Science, 23(2), 141–145.
Kristensen, T. S., Borritz, M., Villadsen, E., & Christensen, K. B. (2005). The Copenhagen Burnout
Inventory: A new tool for the assessment of burnout. Work & Stress, 19(3), 192207.
Lilly, M. M., London, M. J., & Mercer, M. C. (2016). Predictors of Obesity and Physical Health
Complaints Among 911 Telecommunicators. Safety and Health at Work, 7(1), 5562.
Lloyd, C., King, R., & Chenoweth, L. (2002). Social work, stress and burnout: A review. Journal
of Mental Health , 11(3), 255265.
Poncet, M. C., Toullic, P., Papazian, L., Kentish-Barnes, N., Timsit, J.-F., Pochard, F., … Azoulay,
E. (2007). Burnout syndrome in critical care nursing staff. American Journal of Respiratory
and Critical Care Medicine, 175(7), 698704.
Richer, S., & Vallerand, R. (1998). Construction and validation of the ESAS (The Relatedness
Feelings Scale). European Review of Applied Psychology = Revue Europeenne de
Psychologie Appliquee, 48, 138–138.
Richman, L. S., & Leary, M. R. (2009). Reactions to discrimination, stigmatization, ostracism, and
other forms of interpersonal rejection: A multimotive model. Psychological Review, Vol. 116,
pp. 365–383.
Salvagioni, D. A. J., Melanda, F. N., Mesas, A. E., González, A. D., Gabani, F. L., & Andrade, S.
M. de. (2017). Physical, psychological and occupational consequences of job burnout: A
systematic review of prospective studies. PloS One, 12(10), e0185781.
Sewell, J. D., & Crew, L. (1984). The forgotten victim: Stress and the police dispatcher. FBI L.
Enforcement Bull., 53, 7.
Shanafelt, T. D., Balch, C. M., Bechamps, G., Russell, T., Dyrbye, L., Satele, D., … Freischlag,
J. (2010). Burnout and medical errors among American surgeons. Annals of Surgery, 251(6),
Shnabel, N., Purdie-Vaughns, V., Cook, J. E., Garcia, J., & Cohen, G. L. (2013). Demystifying
values-affirmation interventions: writing about social belonging is a key to buffering against
identity threat. Personality & Social Psychology Bulletin, 39(5), 663676.
Steele, C. M. (1988). The Psychology of Self-Affirmation: Sustaining the Integrity of the Self. In L.
Berkowitz (Ed.), Advances in Experimental Social Psychology (Vol. 21, pp. 261302).
Academic Press.
Wallace, J. E., Lemaire, J. B., & Ghali, W. A. (2009). Physician wellness: a missing quality
indicator. The Lancet, 374(9702), 17141721.
Walton, G. M., & Cohen, G. L. (2007). A question of belonging: race, social fit, and achievement.
Journal of Personality and Social Psychology, 92(1), 8296.
Walton, G. M., Logel, C., Peach, J. M., Spencer, S. J., & Zanna, M. P. (2015). Two brief
interventions to mitigate a “chilly climate” transform women’s experience, relationships, and
achievement in engineering. Journal of Educational Psychology, 107(2), 468.
Winwood, P. C., & Winefield, A. H. (2004). Comparing Two Measures of Burnout Among Dentists
in Australia. International Journal of Stress Management, 11(3), 282289.
Wright, S. L., Burt, C. D. B., & Strongman, K. T. (2006). Loneliness in the workplace: Construct
definition and scale development. Retrieved from
Yeager, D. S., & Walton, G. M. (2011). Social-psychological interventions in education: They’re
not magic. Review of Educational Research, 81(2), 267301.
Wigert, B., Agrawal, S. (2018). Employee Burnout, Part 1: The 5 Main Causes. Retrieved from
... Stress in 9-1-1 dispatchers involves burnout and the overall feeling of being misunderstood and underappreciated. Researchers conducted a quantitative analysis in 2019 involving 9-1-1 personnel from multiple cities (Linos, Ruffini, & Wilcoxen, 2019). Linos, Ruffini, and Wilcoxen (2019) took this first study a step further, asking the Los Angeles participants whether or not they felt a level of social belonging and social support. ...
... Although the results of this first study method yielded proof that a relationship exists between feeling a sense of belonging and being valued with job burnout, the researchers Peer Support Programs 16 © 2020 Melissa Alterio could not identify an underlying link between social belonging and burnout (Linos et al., 2019). Therefore, the researchers moved to a second study method involving a controlled experiment in nine mid-sized cities in the United States. ...
... The intention of the emails was mainly to provide support, advice, and encourage dispatchers to ruminate on their work experiences. It also served as a motivator for the recipients to deliver that same type of assistance to newer dispatchers, essentially resulting in peer-to-peer support (Linos et al., 2019). This study supplies instrumental evidence for the 9-1-1 industry as it proves a correlation between intervention programs and a decreased amount of stress among the participants, suggesting further research is necessary to support the benefits of implementing peer-based programs to reduce stress and burnout (Linos et al., 2019). ...
Full-text available
Public safety telecommunicators, often referred to as 9-1-1 dispatchers, experience a significant compassion fatigue and secondary traumatic stress. Their job duties include listening to highly emotional callers provide specific details on current, tragic, and often horrific critical incidents in volatile working conditions. This paper focuses on a systemic review of research that identifies the working conditions and responsibilities of 9-1-1 dispatchers and the subsequent emotional effects of sustained exposure to vicarious trauma. The research recognizes a minimal amount of documentation regarding the stress 9-1-1 dispatchers' experience which supports a fundamental need for comprehensive, longitudinal studies in this job classification. This review also seeks to prove the lack of intervention programs employed to reduce or mitigate the emotional effects of vicarious trauma, compassion fatigue, and secondary traumatic stress, thereby resulting in the irrefutable need for workplace peer support programs to combat these conditions felt by public safety telecommunicators.
... These facts indicate the importance of undertaking preventive interventions of an interdisciplinary nature, both psychological, organizational and medical. Due to the objectively high risk of burnout in the ECD's professional group, programs are created around the world to counter the phenomenon by strengthening selected personal resources (e.g., Linos et al., 2019). The results obtained by us may contain suggestions for preventive measures. ...
Full-text available
International research has demonstrated that emergency call operators face unique risks to their mental health, in particular job stress, and occupational burnout syndrome. There is already wide knowledge about the relationship between stress, burnout and employee personal resources, which has practical application in preventing mental health. However, more research into the subtle relationships between variables is needed. The aim of the study was to check the moderation effect of differences in the intensity of latent variables on the relationship between perceived stress, self-efficacy and professional burnout. The participants were 546 call-takers and dispatchers from 14 public-safety answering point in Poland aged between 19 and 65 years. The Link Burnout Questionnaire, the 10-item Perceived Stress Scale, the Generalized Self-Efficacy Scale, and an independent questionnaire were used to gather information. The method of path analysis was used. The study confirmed the existence of negative relationships between perceived stress (assessment of the current situation) and self-efficacy (a personal trait). Taking into account the moderating effect of latent variable: psychological comfort revealed a hidden relationship between stress and burnout. The stress-burnout relationship occurred only among participants with low level of psychological comfort, so it was not a proportional relationship. In the case of participants with a high level of second latent variable: power-to-affect, the hypothesis that a high level of this variable should weaken the relationship between stress and burnout was not confirmed. The level of latent variables did not affect the self-efficacy relationship with occupational burnout. Taking into account the differences in the intensity of latent variables showed their moderating effect, which often turned out to be different from the assumed one and obtained in the research of other authors. This allowed to discover the relationships that might otherwise have been overlooked and not included in burnout prevention. The results showed a high level of occupational burnout in the ECD’s group during the COVID-19 pandemic: 32% of the responders reported emotional exhaustion, 53% loss of professional effectiveness.
Full-text available
Burnout is a syndrome that results from chronic stress at work, with several consequences to workers’ well-being and health. This systematic review aimed to summarize the evidence of the physical, psychological and occupational consequences of job burnout in prospective studies. The PubMed, Science Direct, PsycInfo, SciELO, LILACS and Web of Science databases were searched without language or date restrictions. The Transparent Reporting of Systematic Reviews and Meta-Analyses guidelines were followed. Prospective studies that analyzed burnout as the exposure condition were included. Among the 993 articles initially identified, 61 fulfilled the inclusion criteria, and 36 were analyzed because they met three criteria that must be followed in prospective studies. Burnout was a significant predictor of the following physical consequences: hypercholesterolemia, type 2 diabetes, coronary heart disease, hospitalization due to cardiovascular disorder, musculoskeletal pain, changes in pain experiences, prolonged fatigue, headaches, gastrointestinal issues, respiratory problems, severe injuries and mortality below the age of 45 years. The psychological effects were insomnia, depressive symptoms, use of psychotropic and antidepressant medications, hospitalization for mental disorders and psychological ill-health symptoms. Job dissatisfaction, absenteeism, new disability pension, job demands, job resources and presenteeism were identified as professional outcomes. Conflicting findings were observed. In conclusion, several prospective and high-quality studies showed physical, psychological and occupational consequences of job burnout. The individual and social impacts of burnout highlight the need for preventive interventions and early identification of this health condition in the work environment.
Full-text available
Background: This study aims to: (1) examine rates of obesity and physical health complaints among 911 telecommunicators; and (2) document the role of emotion dysregulation, psychological inflexibility, duty-related distress and dissociation, and psychopathology in predicting obesity and physical health complaints in this population. Methods: The sample consisted of 911 telecommunicators from across the country (N = 758). Participants completed an online survey assessing their mental and physical health functioning. Results: A total of 82.5% of the sample reported a body mass index that fell within the overweight or obese category and an average of 17 physical health complaints within the past month. Peritraumatic reactions (distress and dissociation), emotion dysregulation, and psychological inflexibility had effects on physical health largely through psychopathology (alcohol abuse, post-traumatic stress disorder, and depression). Conclusion: Development of adapted prevention and intervention efforts with this population is needed.
Full-text available
Even though epidemiological evidence links specific workplace stressors to health outcomes, the aggregate contribution of these factors to overall mortality and health spending in the United States is not known. In this paper, we build a model to estimate the excess mortality and incremental health expenditures associated with exposure to the following 10 workplace stressors: unemployment, lack of health insurance, exposure to shift work, long working hours, job insecurity, work–family conflict, low job control, high job demands, low social support at work, and low organizational justice. Our model uses input parameters obtained from publicly accessible data sources. We estimated health spending from the Medical Expenditure Panel Survey and joint probabilities of workplace exposures from the General Social Survey, and we conducted a meta-analysis of the epidemiological literature to estimate the relative risks of poor health outcomes associated with exposure to these stressors. The model was designed to overcome limitations with using inputs from multiple data sources. Specifically, the model separately derives optimistic and conservative estimates of the effect of multiple workplace exposures on health, and uses optimization to calculate upper and lower bounds around each estimate, which accounts for the correlation between exposures. We find that more than 120,000 deaths per year and approximately 5%–8% of annual healthcare costs are associated with and may be attributable to how U.S. companies manage their work forces. Our results suggest that more attention should be paid to management practices as important contributors to health outcomes and costs in the United States. This paper was accepted by Dimitris Bertsimas, optimization.
Full-text available
To deliver high-quality, reliable, and consistent services safely, organizations develop professional standards. Despite the communication and reinforcement of these standards, they are often not followed consistently. Although previous research suggests that high job demands are associated with declines in compliance over lengthy intervals, we hypothesized-drawing on theoretical arguments focused on fatigue and depletion-that the impact of job demands on routine compliance with professional standards might accumulate much more quickly. To test this hypothesis, we studied a problem that represents one of the most significant compliance challenges in health care today: hand hygiene. Using longitudinal field observations of over 4,157 caregivers working in 35 different hospitals and experiencing more than 13.7 million hand hygiene opportunities, we found that hand hygiene compliance rates dropped by a regression-estimated 8.7 percentage points on average from the beginning to the end of a typical 12-hr work shift. This decline in compliance was magnified by increased work intensity. Further, longer breaks between work shifts increased subsequent compliance rates, and such benefits were greater for individuals when they had ended their preceding shift with a lower compliance rate. In addition, (a) the decline in compliance over the course of a work shift and (b) the improvement in compliance following a longer break increased as individuals accumulated more total work hours the preceding week. The implications of these findings for patient safety and job design are discussed. (PsycINFO Database Record (c) 2014 APA, all rights reserved).
Full-text available
The authors report results from a study of stress and burnout in dentists in which a new measure of burnout, the Copenhagen Burnout Inventory (CBI), was compared with the well-established Maslach Burnout Inventory (MBI; C. Maslach & S. E. Jackson, 1981). The results suggest that the CBI possesses excellent psychometric properties and seems to be an appropriate measure of burnout in populations of health professionals. By comparison, difficulties were encountered in demonstrating the assumed nexus between the 3 (subscale) components of the MBI. These results call into question a long-held definition of burnout that appears to have been widely accepted. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Full-text available
Recent randomized experiments have found that seemingly “small” social-psychological interventions in education—that is, brief exercises that target students’ thoughts, feelings, and beliefs in and about school—can lead to large gains in student achievement and sharply reduce achievement gaps even months and years later. These interventions do not teach students academic content but instead target students’ psychology, such as their beliefs that they have the potential to improve their intelligence or that they belong and are valued in school. When social-psychological interventions have lasting effects, it can seem surprising and even “magical,” leading people either to think of them as quick fixes to complicated problems or to consider them unworthy of serious consideration. The present article discourages both responses. It reviews the theoretical basis of several prominent social-psychological interventions and emphasizes that they have lasting effects because they target students’ subjective experiences in school, because they use persuasive yet stealthy methods for conveying psychological ideas, and because they tap into recursive processes present in educational environments. By understanding psychological interventions as powerful but context-dependent tools, educational researchers will be better equipped to take them to scale. This review concludes by discussing challenges to scaling psychological interventions and how these challenges may be overcome.
Full-text available
So far, the large majority of studies on burnout in the international literature have employed the Maslach Burnout Inventory (MBI). In this paper we criticize the MBI on a number of points and present a new tool for the measurement of burnout: the Copenhagen Burnout Inventory (CBI). The CBI consists of three scales measuring personal burnout, work-related burnout, and client-related burnout, for use in different domains. On the basis of an ongoing prospective study of burnout in employees in the human service sector, the PUMA study (Project on Burnout, Motivation and Job Satisfaction; N=1914 at baseline), we analysed the validity and reliability of the CBI. All three scales were found to have very high internal reliability, and non-response rates were small. The scales differentiated well between occupational groups in the human service sector, and the expected pattern with regard to correlations with other measures of fatigue and psychological well-being was found. Furthermore, the three scales predicted future sickness absence, sleep problems, use of pain-killers, and intention to quit. Analyses of changes over time showed that substantial proportions of the employees changed with regard to burnout levels. It is concluded that the analyses indicate very satisfactory reliability and validity for the CBI instrument. The CBI is being used in a number of countries and translations into eight languages are available.
Full-text available
Two experiments examined for the first time whether the specific content of participant-generated affirmation essays-in particular, writing about social belonging-facilitated an affirmation intervention's ability to reduce identity threat among negatively stereotyped students. Study 1, a field experiment, revealed that seventh graders assigned to a values-affirmation condition wrote about social belonging more than those assigned to a control condition. Writing about belonging, in turn, improved the grade point average (GPA) of Black, but not White students. In Study 2, using a modified "belonging-affirmation" intervention, we directly manipulated writing about social belonging before a math test described as diagnostic of math ability. The more female participants wrote about belonging, the better they performed, while there was no effect of writing about belonging for males. Writing about social belonging improved performance only for members of negatively stereotyped groups. Implications for self-affirmation theory and practice are discussed.
In a randomized-controlled trial, we tested 2 brief interventions designed to mitigate the effects of a "chilly climate" women may experience in engineering, especially in male-dominated fields. Participants were students entering a selective university engineering program. The social-belonging intervention aimed to protect students' sense of belonging in engineering by providing a nonthreatening narrative with which to interpret instances of adversity. The affirmation-training intervention aimed to help students manage stress that can arise from social marginalization by incorporating diverse aspects of their self-identity in their daily academic lives. As expected, gender differences and intervention effects were concentrated in male-dominated majors (<20% women). In these majors, compared with control conditions, both interventions raised women's school-reported engineering grade-point-average (GPA) over the full academic year, eliminating gender differences. Both also led women to view daily adversities as more manageable and improved women's academic attitudes. However, the 2 interventions had divergent effects on women's social experiences. The social-belonging intervention helped women integrate into engineering, for instance, increasing friendships with male engineers. Affirmation-training helped women develop external resources, deepening their identification with their gender group. The results highlight how social marginalization contributes to gender inequality in quantitative fields and 2 potential remedies.
Self-affirmation processes are being activated by information that threatens the perceived adequacy or integrity of the self and as running their course until this perception is restored through explanation, rationalization, and/or action. The purpose of these constant explanations (and rationalizations) is to maintain a phenomenal experience of the self-self-conceptions and images as adaptively and morally adequate—that is, as competent, good, coherent, unitary, stable, capable of free choice, capable of controlling important outcomes, and so on. The research reported in this chapter focuses on the way people cope with the implications of threat to their self-regard rather than on the way they cope with the threat itself. This chapter analyzes the way coping processes restore self-regard rather than the way they address the provoking threat itself.