Access to this full-text is provided by PLOS.
Content available from PLOS One
This content is subject to copyright.
RESEARCH ARTICLE
Decreased odds of depressive symptoms and
suicidal ideation with higher education,
depending on sex and employment status
Vanessa K. TassoneID
1
, Sophie F. Duffy
1
, Sarah Dunnett
1
, Josheil K. Boparai
1
,
Valentina Zuluaga Cuartas
1
, Hyejung Jung
2
, Michelle Wu
1
, Navya Goel
1
, Wendy Lou
2
,
Venkat BhatID
1,3,4,5
*
1Interventional Psychiatry Program, St. Michael’s Hospital, Toronto, Ontario, Canada, 2Department of
Biostatistics, Dalla Lana School of Public Health, University of Toronto, Toronto, Ontario, Canada, 3Institute
of Medical Science, Medical Sciences Building, University of Toronto, Toronto, Ontario, Canada, 4Mental
Health and Addictions Services, St. Michael’s Hospital, Toronto, Ontario, Canada, 5Department of
Psychiatry, University of Toronto, Toronto, Ontario, Canada
*venkat.bhat@utoronto.ca
Abstract
Background
Higher education is associated with reduced depressive symptoms and requires investment
without guaranteed employment. It remains unclear how sex and employment status
together contribute to the association between mental health and educational attainment.
This study investigated the role of sex and employment status together in the associations
of 1) depressive symptoms and 2) suicidal ideation with education.
Methods
Using 2005–2018 National Health and Nutrition Examination Survey data, cross-sectional
analyses were conducted on individuals 20 years who completed the depression question-
naire and reported their employment status and highest level of education. Survey-weighted
multivariable logistic regression models were used to explore how depressive symptoms
and suicidal ideation are associated with educational attainment in an analysis stratified by
sex and employment status. To account for multiple testing, a significance level of a<0.01
was used.
Results
Participants (n= 23,669) had a weighted mean age of 43.25 (SD = 13.97) years and 47%
were female. Employed females (aOR = 0.47, 95% CI 0.32, 0.69), unemployed females
(aOR = 0.47, 95% CI 0.29, 0.75), and unemployed males (aOR = 0.31, 95% CI 0.17, 0.56)
with college education had reduced odds of depressive symptoms compared to those with
high school education. Employed females with college education also had reduced suicidal
ideation odds compared to those with high school education (aOR = 0.41, 95% CI 0.22,
0.76).
PLOS ONE
PLOS ONE | https://doi.org/10.1371/journal.pone.0299817 April 3, 2024 1 / 14
a1111111111
a1111111111
a1111111111
a1111111111
a1111111111
OPEN ACCESS
Citation: Tassone VK, Duffy SF, Dunnett S, Boparai
JK, Zuluaga Cuartas V, Jung H, et al. (2024)
Decreased odds of depressive symptoms and
suicidal ideation with higher education, depending
on sex and employment status. PLoS ONE 19(4):
e0299817. https://doi.org/10.1371/journal.
pone.0299817
Editor: Guglielmo Campus, University of Bern:
Universitat Bern, SWITZERLAND
Received: April 26, 2023
Accepted: February 15, 2024
Published: April 3, 2024
Copyright: ©2024 Tassone et al. This is an open
access article distributed under the terms of the
Creative Commons Attribution License, which
permits unrestricted use, distribution, and
reproduction in any medium, provided the original
author and source are credited.
Data Availability Statement: The data underlying
the results presented in the study are available
from https://wwwn.cdc.gov/nchs/nhanes/default.
aspx ("NHANES 2005-2006" to "NHANES 2017-
2018" datasets).
Funding: The authors received no specific funding
for this work.
Competing interests: I have read the journal’s
policy and the authors of this manuscript have the
following competing interests: VB is supported by
Conclusions
Females demonstrated significant associations between depressive symptoms and educa-
tion, regardless of employment status, whereas males demonstrated an association only if
unemployed. Employed females, in particular, demonstrated a significant association
between suicidal ideation and education. These findings may inform future research investi-
gating the underlying mechanisms and etiology of these sex-employment status differences
in the association between mental health and education.
Introduction
Major depressive disorder (MDD) is a leading contributor to disability worldwide [1], with
North America having one of the greatest incidences of MDD globally [2]. The lifetime preva-
lences of suicidal ideation and attempts in individuals with MDD are 40.3% and 31%, respec-
tively [3,4]. Females are more likely than males to experience depression and suicidal ideation
[5,6]; however, male suicide mortality is 3 to 4 times that of females [7].
Unemployed adults are known to have higher depression odds than their employed coun-
terparts [8,9], however, these associations vary by sex. A meta-analysis by Paul and Moser [10]
found that the negative effects of unemployment were greater for men than women. Similarly,
Andersen et al [11] found a higher prevalence of minor depression in non-employed men than
in non-employed women. Suicidal ideation, often a symptom of depression, is suggested to be
increased in unemployed males, but not unemployed females [12].
Research suggests that three pathways may explain the association between unemployment
and depression. The causal hypothesis suggests that individuals may experience poor mental
health as a result of unemployment [13,14]. Becoming unemployed may lead to a loss of self-
worth, personal identity, and perceived sense of control over one’s life, as well as financial insta-
bility and a decline in overall well-being [15–17]. Regaining employment may restore mental
health to some degree [18]. The selection hypothesis suggests that poor mental health causes
unemployment [10,13,14]. The causal directions proposed in each of these hypotheses may
also operate simultaneously, resulting in bidirectionality between poor mental health and
unemployment [10,19,20]. The last hypothesis suggests that an external factor, such as educa-
tional attainment, may confound the relationship between unemployment and depression [21].
The relationship between education and employment is well established. Individuals with
higher education are generally employed at greater rates [22], while those with basic education
are more likely to transition out of paid employment due to disability, early retirement, or ful-
fillment of domestic tasks and care responsibilities [23]. While higher education provides
opportunities in the form of employment and higher income [24], it is not without costs. Edu-
cation is considered a personal investment in human capital, as the opportunity cost of achiev-
ing higher education is greater than forgoing it and entering the labor force sooner [18,25].
Obtaining more years of education to develop the skills needed for a chosen career may also be
the product of greater personal investment and identity in one’s career [26]. While higher edu-
cation generally protects against depressive symptoms and suicidal ideation [27,28], it is possi-
ble that violated expectations for employment following higher education may result in mental
discomfort.
Similar to employment, the relationship between depression and education, as well as sui-
cidal ideation and education, is sex-dependent. Although both males and females experience
PLOS ONE
Depressive symptoms, suicidal ideation, and higher education
PLOS ONE | https://doi.org/10.1371/journal.pone.0299817 April 3, 2024 2 / 14
an Academic Scholar Award from the University of
Toronto Department of Psychiatry and has
received research support from the Canadian
Institutes of Health Research, Brain & Behavior
Foundation, Ontario Ministry of Health Innovation
Funds, Royal College of Physicians and Surgeons
of Canada, Department of National Defence
(Government of Canada), New Frontiers in
Research Fund, Associated Medical Services Inc.
Healthcare, American Foundation for Suicide
Prevention, Roche Canada, Novartis, and Eisai.
This does not alter our adherence to PLOS ONE
policies on sharing data and materials. All other
authors do not have any competing interests to
declare.
benefits from higher education in reducing depressive symptoms, this relationship is generally
stronger for females [29–31]. Additionally, higher education is associated with reduced sui-
cidal ideation among females only [12]. A male-only study found no significant association
between education and suicidal ideation, however, there was a negative association with sui-
cide attempt [32]. Sex differences in education may posit interesting mental health trajectories,
as females attaining higher education have demonstrated higher degrees of self-help, suggest-
ing improved health literacy and positive impacts on psychosocial skills, compared to their
male counterparts [33]. The recent decade has shown changing sex dynamics as higher educa-
tion has shifted from male- to female-dominated in the United States (US) [5,22,30].
As sex and gender roles shift over time, research must reflect the current state of such
dynamics. Studies suggest that while males are more adversely affected by unemployment
[10–12], females might experience a stronger protective effect of higher education [30]. There-
fore, an exploration considering all three variables (i.e., sex, employment status, and educa-
tion) is warranted to elucidate whether the relationship between mental health and
educational attainment for each sex remains consistent for those with differing employment
statuses. Given the previously described sex-based associations, it may be expected that:
• Hypothesis 1: The association between education and mental health outcomes will be signifi-
cant among employed females but not males
Hypothesis 1a: Employed females with higher education will experience significantly
reduced odds of depressive symptoms and suicidal ideation
Hypothesis 1b: Employed females with less than high school education might experience sig-
nificantly increased odds of depressive symptoms and suicidal ideation
• Hypothesis 2: The association between education and mental health outcomes will be signifi-
cant among unemployed females but not males
Hypothesis 2a: Unemployed females with higher education will experience significantly
reduced odds of depressive symptoms and suicidal ideation
Hypothesis 2b: Unemployed females with less than high school education might experience
significantly increased odds of depressive symptoms and suicidal ideation
To the authors’ knowledge, these hypotheses are unexplored in the existing literature. As
such, the purpose of this study was to perform a sex and employment status stratified examina-
tion of how depressive symptoms and suicidal ideation are associated with educational
attainment.
Methods
Study population
This study used data from participants who were recruited for the 2005–2018 National Health
and Nutrition Examination Survey (NHANES). NHANES contains a series of cross-sectional
datasets that are representative of the non-institutionalized resident US population. This sur-
vey series is conducted by the National Center for Health Statistics (NCHS), a division of the
Centers for Disease Control and Prevention. The multistage probability sampling of partici-
pants is conducted by the stepwise selection of counties as the primary sampling unit, then
block segments within counties, households within block segments, and finally individuals
within households [34–37]. The NHANES protocol was approved by the Research Ethics
Review Board of the NCHS and written informed consent was obtained from participants
PLOS ONE
Depressive symptoms, suicidal ideation, and higher education
PLOS ONE | https://doi.org/10.1371/journal.pone.0299817 April 3, 2024 3 / 14
prior to data collection. The authors did not have access to information that could be used to
identify participants after data collection.
Participants were excluded from this study if they were not working due to retirement,
being a student, taking care of family, or reported “other” as their reason for not working [8].
Participants were also excluded if they were <20 years of age since some of the questions used
to collect data for the current analyses were only presented to those who were a minimum of
20 years old. Included participants were required to have responded to items 1–9 on the Men-
tal Health—Depression Screener questionnaire (DPQ), report the type of work that they did in
the last week (OCD150) on the Occupation questionnaire (OCQ), and report their highest
level of education (DMDEDUC2) from the Demographic Data (DEMO).
Exposure variables
Educational attainment was categorized using responses to DMDEDUC2 on the DEMO ques-
tionnaire. Responses were categorized as <high school, high school, some college/Associate of
Arts degree, or college or above. Employment status was dichotomized as employed or unem-
ployed. Based on the responses to OCD150 on the OCQ questionnaire, employed individuals
were classified as those who reported working at a job or business or with a job or business but
not at work in the last week. Unemployed individuals were classified as those who reported
looking for work or were not working at a job or business in the last week. In particular, unem-
ployed individuals included those who were not working due to disability or health-related
reasons or were on lay-off in the last week. We opted to characterize unemployment in this
way because–unlike retirees, students, or family carers who were not employed but were
excluded from the current study–those who were unemployed due to disability, health-related
reasons, or lay-off could be perceived as being unwillingly selected into unemployment. While
previous work has also included those who were not working due to disability or health-related
reasons in analyses [38], other research has limited unemployment to those who were looking
for work or were on lay-off [39]. As such, a sensitivity analysis was conducted wherein partici-
pants who were not working due to disability or health-related reasons were excluded from the
study population since, similar to those who were retired, students, or family carers, these indi-
viduals were not in the labour force. Sex was dichotomized as male or female based on the two
response options available in NHANES.
Outcome measures
The primary outcome of this study was depressive symptoms, which was assessed using the
nine-item Patient Health Questionnaire (PHQ-9) from the DPQ. The PHQ-9 is a self-report
measure corresponding to the Diagnostic Statistical Manual of Mental Disorders,Fourth Edi-
tion diagnostic criteria for MDD [40]. Nine questions, probing the frequencies of symptoms of
depression within the last 2 weeks, were scored on a four-point Likert scale from “0” (not
experiencing the symptom) to “3” (experiencing the symptom nearly every day). The PHQ-9
has been validated and is a reliable tool for depression diagnosis with high specificity and sen-
sitivity [40–42]. PHQ-9 total scores were dichotomized with scores 10 indicating presence of
depressive symptoms, and scores <10 indicating absence of depressive symptoms [42].
The secondary outcome of this study was suicidal ideation, assessed using item nine from
the PHQ-9. Participants who responded to this item with the answer “not at all” were classified
as having no suicidal ideation. Responses of “several days,” “more than half the days,” or
“nearly every day” were classified as experiencing suicidal ideation. Item nine on the PHQ-9
has been shown to have a high negative predictive value and specificity, as well as a moderate
PLOS ONE
Depressive symptoms, suicidal ideation, and higher education
PLOS ONE | https://doi.org/10.1371/journal.pone.0299817 April 3, 2024 4 / 14
sensitivity. It is recommended as a screening tool for suicidal ideation due to a low positive
predictive value [43,44].
Covariates
Covariates were selected based on existing knowledge from the literature to adjust for potential
confounding bias [30]. Covariates included age (continuous by 1-year increases) [45], race
(non-Hispanic white, non-Hispanic black, Hispanic, other/multi-racial), marital status (mar-
ried/living with partner or not) [46], and NHANES survey cycle (2-year intervals).
Statistical analysis
Data analysis was performed using Mobile Examination Center (MEC) survey weights in R v
4.2.2 with the package “survey.” MEC survey weights were used to account for the clustered
sample design, oversampling, survey non-response, and post-stratification adjustments so that
the results could be generalized to the US population. Survey weights were divided by seven to
account for combining seven survey cycles to make up the study population. The study popu-
lation was divided into four groups based on both employment status and sex. To adjust for
multiple testing and reduce the risk of type I error, significance was established at P<0.01
using two-tailed hypothesis testing. Significant demographic differences between any of these
four groups were determined using an analysis of variance for continuous variables and a χ
2
test for categorical variables. An analysis examining the reasons for unemployment was con-
ducted between sexes and proportions were tested using a χ
2
test for categorical variables. Mul-
tivariable logistic regression models were used to estimate the association between education
and depressive symptoms, as well as suicidal ideation, adjusting for age, race, marital status,
and NHANES survey cycle among each of the four groups. Participants were excluded from
analysis if they had missing data or responded “refused” or “I don’t know” to variables of
interest.
Results
Demographic characteristics
A total of 23,669 individuals aged 20 to 85 years were included in this study (Fig 1). The
weighted prevalence of depressive symptoms and suicidal ideation was 7.96% and 3.29%,
respectively. Unemployed females had the highest prevalence of depressive symptoms and sui-
cidal ideation, followed by unemployed males, employed females, and employed males.
Employed females pursued education beyond high school at the highest rate, followed by
employed males, unemployed females, and unemployed males. Table 1 and S1 Table present
the demographic characteristics for the study population and the sensitivity analysis,
respectively.
Continuous characteristics reported with weighted mean and standard deviation while cat-
egorical characteristics reported with unweighted frequency and weighted percentage. Pvalues
reported using survey weights.
A higher proportion of females than males were unemployed (15.59% vs. 13.28%). After
exploring the reported reasons for unemployment by sex, a greater proportion of females than
males were unemployed due to health reasons and disability, while a greater proportion of
males than females were unemployed due to lay-off or were looking for work (Table 2).
S2 Table presents the reasons for unemployment for those included in the sensitivity analysis.
PLOS ONE
Depressive symptoms, suicidal ideation, and higher education
PLOS ONE | https://doi.org/10.1371/journal.pone.0299817 April 3, 2024 5 / 14
Depressive symptoms and educational attainment
In both the unadjusted (S3 Table) and adjusted (Table 3) models, employed females with less
than high school education had significantly increased odds of depressive symptoms,
Fig 1. Flow chart of participant inclusion from NHANES 2005–2018 surveys.
https://doi.org/10.1371/journal.pone.0299817.g001
Table 1. Demographic characteristics of the NHANES 2005–2018 study population (n= 23,669), stratified by sex and employment status.
Female Employed Male Employed Female Unemployed Male Unemployed Pvalue
n8,866 10,218 2,227 2,358
Depressive symptoms = Yes (%) 632 (6.58) 377 (3.63) 709 (31.45) 475 (19.78) <0.001
Suicidal ideation = Yes (%) 230 (2.35) 217 (1.98) 243 (10.20) 224 (10.01) <0.001
Age, mean (SD) 42.72 (13.65) 42.34 (13.66) 48.92 (14.81) 46.28 (15.03) <0.001
Race (%) <0.001
Non-Hispanic White 3,546 (67.40) 4,134 (67.20) 881 (61.28) 955 (59.64)
Hispanic 2,252 (12.90) 2,855 (16.16) 527 (13.44) 522 (14.86)
Non-Hispanic Black 2,043 (12.10) 1,991 (9.13) 652 (18.87) 693 (18.18)
Other / Multi-racial 1,025 (7.59) 1,238 (7.50) 167 (6.41) 188 (7.33)
Marital status = Not Married / Living with partner (%) 3,834 (38.57) 3,131 (30.14) 1,353 (55.37) 1,223 (53.08) <0.001
Education (%) <0.001
<High school 1,342 (9.47) 2,208 (13.96) 755 (25.93) 870 (28.41)
High school 1,783 (19.48) 2,395 (23.85) 537 (26.12) 691 (30.87)
Some college / Associate of Arts degree 3,146 (35.48) 2,874 (29.95) 691 (32.53) 585 (28.33)
College or above 2,595 (35.56) 2,741 (32.23) 244 (15.42) 212 (12.38)
https://doi.org/10.1371/journal.pone.0299817.t001
PLOS ONE
Depressive symptoms, suicidal ideation, and higher education
PLOS ONE | https://doi.org/10.1371/journal.pone.0299817 April 3, 2024 6 / 14
compared to those with high school education. Additionally, all groups with a college degree,
except employed males, had significantly reduced odds of depressive symptoms compared
those with high school education. The associations between depressive symptoms and educa-
tion became non-significant for unemployed females and males in the sensitivity analysis
(S4 and S5 Tables).
Suicidal ideation and educational attainment
In both the unadjusted (S6 Table) and adjusted (Table 4) models, employed females with col-
lege or above education had significantly lower odds of suicidal ideation when compared to
those with high school education. The sensitivity analysis demonstrated similar results (S7 and
S8 Tables).
Discussion
This study examined the association between depressive symptoms and educational attain-
ment, as well as the association between suicidal ideation and educational attainment, using a
nationally representative sample of the US population, stratified by sex and employment status.
Our results revealed that employed females with less than high school education demonstrated
increased odds of depressive symptoms compared to those with high school education.
Females with college or above education had decreased odds of depressive symptoms, regard-
less of employment status. Males also demonstrated decreased odds of depressive symptoms
with college or above education, however, this was only significant amongst the unemployed.
The associations between depressive symptoms and education became non-significant in
unemployed participants, regardless of sex, when those who were not working due to disability
or health-related reasons were excluded from analyses. Employed females were the only group
Table 2. Reason for unemployment among those unemployed in the study population, stratified by sex.
Female Unemployed Male Unemployed Pvalue
n(%) 2,227 (15.59) 2,358 (13.28)
Reason for Unemployment <0.001
Health-related (%) 677 (28.75) 509 (20.10)
Disability (%) 960 (40.81) 983 (39.45)
Lay-off (%) 162 (7.67) 228 (10.10)
Looking for work (%) 428 (22.76) 638 (30.36)
Results reported using unweighted frequency and weighted percentage. Pvalues reported using survey weights.
https://doi.org/10.1371/journal.pone.0299817.t002
Table 3. Adjusted logistic regression of depressive symptoms and educational attainment, stratified by sex and employment status (study population).
Female Employed Male Employed Female Unemployed Male Unemployed
aOR (95% CI) P value aOR (95% CI) P value aOR (95% CI) P value aOR (95% CI) P value
Education
High school 1 (Referent) 1 (Referent) 1 (Referent) 1 (Referent)
<High school 1.75 (1.22, 2.50) 0.003*1.46 (1.01, 2.11) 0.050 1.06 (0.76, 1.46) 0.74 1.20 (0.85, 1.67) 0.30
Some college / Associate of Arts degree 1.03 (0.76, 1.41) 0.84 1.23 (0.88, 1.70) 0.23 0.84 (0.63, 1.13) 0.26 0.76 (0.53, 1.09) 0.13
College or above 0.47 (0.32, 0.69) <0.001*0.66 (0.43, 1.02) 0.07 0.47 (0.29, 0.75) 0.002*0.31 (0.17, 0.56) <0.001*
*indicates statistical significance (P <0.01). aOR = adjusted odds ratio. CI = confidence interval. Covariates include age, race, marital status, and NHANES survey
cycle.
https://doi.org/10.1371/journal.pone.0299817.t003
PLOS ONE
Depressive symptoms, suicidal ideation, and higher education
PLOS ONE | https://doi.org/10.1371/journal.pone.0299817 April 3, 2024 7 / 14
to demonstrate a significant association between suicidal ideation and education, such that
those with college or above education demonstrated decreased odds of this specific depressive
symptom.
The results of this study supported our hypothesis that employed females with higher edu-
cation would have a significant reduction in depressive symptom odds, whereas employed
males would not. The results also supported our hypothesis that employed females with less
than high school education would experience significantly increased depressive symptom
odds. Higher education may provide improved health literacy and help-seeking behaviors,
resulting in positive implications on mental health. For instance, Warner et al [33] found that
women had significantly higher self-help scores with increased educational attainment, while
men did not. Higher education also appears to enhance one’s sense of control, thus yielding
improvement in mental well-being, particularly among women, as they apply the skills learned
through formal education to their lives [30]. While education has positive effects on reducing
depressive symptoms in females, wealth has been found to be a protective socioeconomic fac-
tor against depression for males [29]. These differential sex-based relationships may have his-
torical roots in the perception of stability in employment, which may contribute to reducing
depressive outcomes. Historically, males were the sole breadwinners in heterosexual house-
holds, making their level of income deterministic of the household’s wealth. This notion is
challenged in present-day US households which follow a dual-breadwinner model as egalitar-
ian views on gender roles are becoming more prevalent and traditionalist views are declining
[47,48]. Despite these changes, females continue to receive unequal pay for equal work com-
pared to their male counterparts [49]. Thus, it is possible that females find less security in
wealth and more security in higher education that can provide them with the stability and
mental wellness that often accompanies having greater employment opportunities and
income. Notably, the male-female wage gap is neutralizing over time, possibly leading females
to also seek greater security in wealth [49]. Further research on the sex- and gender-based dif-
ferences in net wealth and depressive symptoms is needed, as well as the etiology of the sex dif-
ferences in the depressive symptom and education relationship.
The current study demonstrated that both unemployed females and males had reduced
depressive symptom odds with higher education when the total study population was consid-
ered. This was contrary to our hypothesis that an association between mental health and edu-
cation would exist among unemployed females only, thus demonstrating the importance of
considering sex and employment status together when assessing this relationship. As the
required qualifications for employment increase overtime, unemployed individuals with
higher education might find security in their favourable employment prospects [50,51],
regardless of sex, which could be protective against depressive symptoms. However, the
Table 4. Adjusted logistic regression of suicidal ideation and educational attainment, stratified by sex and employment status (study population).
Female Employed Male Employed Female Unemployed Male Unemployed
aOR (95% CI) P value aOR (95% CI) P value aOR (95% CI) P value aOR (95% CI) P value
Education
High school 1 (Referent) 1 (Referent) 1 (Referent) 1 (Referent)
<High school 1.96 (1.12, 3.43) 0.02 1.11 (0.70, 1.77) 0.66 0.92 (0.62, 1.38) 0.70 1.03 (0.65, 1.65) 0.89
Some college / Associate of Arts degree 0.74 (0.45, 1.22) 0.24 1.10 (0.71, 1.70) 0.66 0.77 (0.50, 1.18) 0.23 1.01 (0.60, 1.72) 0.96
College or above 0.41 (0.22, 0.76) 0.006*0.62 (0.35, 1.09) 0.10 0.81 (0.43, 1.52) 0.51 0.52 (0.23, 1.17) 0.117
*indicates statistical significance (P <0.01). aOR = adjusted odds ratio. CI = confidence interval. Covariates include age, race, marital status, and NHANES survey
cycle.
https://doi.org/10.1371/journal.pone.0299817.t004
PLOS ONE
Depressive symptoms, suicidal ideation, and higher education
PLOS ONE | https://doi.org/10.1371/journal.pone.0299817 April 3, 2024 8 / 14
association between depressive symptoms and education became non-significant in both sexes
after removal of those who were not working due to disability or health-related reasons from
analyses. This difference in findings might be explained by an early theory regarding causal
attributions of unemployment [52,53]. Feather and Davenport [52,53] reported that
depressed affect was positively associated with causal attributions which emphasized stable
and frustrating external factors, such as social, economic, and political forces, as reasons for
unemployment (e.g., in people who were needing or looking for work). In contrast, individuals
who reported internal attributions for unemployment appeared to exhibit less negative affect
[53], which was supported by subsequent research reporting that an internal locus of control
was associated with lower levels of depression [54]. As such, grouping individuals who might
attribute unemployment to internal factors, such as health-related reasons or disability, with
those who were looking for work or were on lay-off might have skewed the association
between depressive symptoms and education in the current study, regardless of sex. It is worth
noting, however, that the unemployment group drastically decreased in size upon removing
individuals who were not working due to disability or health-related reasons from analyses,
which could have also impacted the results of the sensitivity analysis.
The odds of suicidal ideation, a possible symptom of major depression [55], were signifi-
cantly decreased amongst employed females with college or above education. This is in line
with a previous study which demonstrated a female-specific protective effect of higher educa-
tion against suicidal thoughts [12]. The literature on the association between suicidal ideation
and education shows mixed results. Studies demonstrate that, in addition to sex, age may be an
important moderator. Research has shown that young adults with a college degree (2- or 4-year
degree) had reduced odds of suicidal ideation, while middle-aged adults with a 4-year college
degree had higher odds of suicidal ideation [45]. A study specifically in young adults found that
among those with more than high school education, suicidal ideation was not associated with
unemployment within the last two years, however, it was positively associated with lifetime
unemployment [56]. Within the current study, age was only considered as a covariate in analy-
ses and participants were categorized as unemployed based on whether they worked within the
last week, both of which could potentially explain the lack of association between suicidal idea-
tion and education in unemployed females. Future studies may analyze the associations investi-
gated here with an added interest for the effects of age and duration of unemployment.
Limitations
Due to the cross-sectional nature of NHANES, causation could not be established in this
study. The questionnaires used by NHANES are also limited by their self-reported nature, sub-
jecting responses to potential recall, social desirability, and non-response biases. An additional
limitation of the NHANES dataset includes responses to the sex question being restricted to
“male” or “female”, without consideration of those who are intersex. As such, potential inter-
sex individuals who felt compelled to respond to the sex question would have been inaccu-
rately dichotomized as male or female. Detection of suicidal ideation may also require an
improved method in place of item nine on the PHQ-9 due to evidence suggesting that it has
limited utility in certain demographic and clinical groups [57]. However, a systematic review
suggests that there is no gold standard instrument to achieve this and a new tool is needed
[58]. It is possible that external events, such as the Great Recession, influenced the results of
our analyses. Due to constraints of the NHANES dataset, we were limited in our ability to
account for such external factors (i.e., by adjusting for survey cycle). Similarly, due to restric-
tions of the NHANES dataset, we were unable to account for the gender pay-gap in analyses.
Finally, NHANES asks participants to report on their work situation within the last week,
PLOS ONE
Depressive symptoms, suicidal ideation, and higher education
PLOS ONE | https://doi.org/10.1371/journal.pone.0299817 April 3, 2024 9 / 14
making it impossible to account for the total length of time that participants were not working
for in analyses. As such, the “unemployment” group may be characterized by high heterogene-
ity not only in terms of length of unemployment, but also in terms of the number of times that
one has been unemployed, which could confound the results of the current study.
Conclusion and future directions
This study adds to the existing literature by further exploring the sex-based associations
between depressive symptoms and education, as well as between suicidal ideation and educa-
tion, from the perspective of employment status. Specifically, employed females with less than
high school education demonstrated increased odds of depressive symptoms. Employed and
unemployed females, as well as unemployed males, had reduced depressive symptom odds
with higher education. Employed females with higher education also demonstrated decreased
odds of suicidal ideation. These findings suggest that there are interesting differences in
depressive symptom and suicidal ideation odds with higher education based on sex and
employment status. They also highlight potential vulnerability amongst employed females
with less than high school education who might benefit from efforts to promote secondary
school completion, including early intervention at the school-age or provision of incentives to
encourage re-entry into the educational system at the adult-age, and preventative mental
health programming to promote resiliency.
Future studies should further investigate how higher education and employment status
may together influence sex-based dynamics in depressive symptoms and suicidal ideation
while also accounting for health status amongst the unemployed (i.e., not working due to
health-related reasons or disability vs. looking for work or on lay-off), length of unemploy-
ment, the type of job one would be working if they were not unemployed (i.e., blue- or white-
collar work), net wealth, ethnicity, and the gender pay-gap. Longitudinal studies may aid in
understanding the directionality of this relationship.
Supporting information
S1 Table. Demographic characteristics of the sensitivity analysis (i.e., individuals not
working due to disability or health-related reasons excluded; n = 20,540), stratified by sex
and employment status. Continuous characteristics reported with weighted mean and stan-
dard deviation while categorical characteristics reported with unweighted frequency and
weighted percentage. P values reported using survey weights.
(DOCX)
S2 Table. Reason for unemployment among those unemployed in the sensitivity analysis
(i.e., individuals not working due to disability or health-related reasons excluded), strati-
fied by sex. Results reported using unweighted frequency and weighted percentage. P values
reported using survey weights.
(DOCX)
S3 Table. Unadjusted logistic regression of depressive symptoms and educational attain-
ment, stratified by sex and employment status (study population). *indicates statistical sig-
nificance (P <0.01). OR = odds ratio. CI = confidence interval.
(DOCX)
S4 Table. Unadjusted logistic regression of depressive symptoms and educational attain-
ment, stratified by sex and employment status (sensitivity analysis). *indicates statistical
significance (P <0.01). OR = odds ratio. CI = confidence interval.
(DOCX)
PLOS ONE
Depressive symptoms, suicidal ideation, and higher education
PLOS ONE | https://doi.org/10.1371/journal.pone.0299817 April 3, 2024 10 / 14
S5 Table. Adjusted logistic regression of depressive symptoms and educational attainment,
stratified by sex and employment status (sensitivity analysis). *indicates statistical signifi-
cance (P <0.01). aOR = adjusted odds ratio. CI = confidence interval. Covariates include age,
race, marital status, and NHANES survey cycle.
(DOCX)
S6 Table. Unadjusted logistic regression of suicidal ideation and educational attainment,
stratified by sex and employment status (study population). *indicates statistical signifi-
cance (P <0.01). OR = odds ratio. CI = confidence interval.
(DOCX)
S7 Table. Unadjusted logistic regression of suicidal ideation and educational attainment,
stratified by sex and employment status (sensitivity analysis). *indicates statistical signifi-
cance (P <0.01). OR = odds ratio. CI = confidence interval.
(DOCX)
S8 Table. Adjusted logistic regression of suicidal ideation and educational attainment,
stratified by sex and employment status (sensitivity analysis). *indicates statistical signifi-
cance (P <0.01). aOR = adjusted odds ratio. CI = confidence interval. Covariates include age,
race, marital status, and NHANES survey cycle.
(DOCX)
Acknowledgments
The authors would like to acknowledge Shakila Meshkat for assistance with proofreading and
language editing.
Author Contributions
Conceptualization: Vanessa K. Tassone, Sophie F. Duffy, Navya Goel, Venkat Bhat.
Data curation: Sophie F. Duffy.
Formal analysis: Sophie F. Duffy, Michelle Wu.
Investigation: Navya Goel.
Methodology: Sophie F. Duffy.
Supervision: Hyejung Jung, Wendy Lou, Venkat Bhat.
Validation: Michelle Wu.
Visualization: Vanessa K. Tassone, Sophie F. Duffy, Michelle Wu.
Writing – original draft: Vanessa K. Tassone, Sophie F. Duffy.
Writing – review & editing: Vanessa K. Tassone, Sophie F. Duffy, Sarah Dunnett, Josheil K.
Boparai, Valentina Zuluaga Cuartas, Hyejung Jung, Michelle Wu, Navya Goel, Venkat
Bhat.
References
1. Ferrari AJ, Charlson FJ, Norman RE, Patten SB, Freedman G, Murray CJL, et al. Burden of depressive
disorders by country, sex, age, and year: Findings from the Global Burden of Disease Study 2010.
PLOS Med. 2013; 10(11):e1001547. https://doi.org/10.1371/journal.pmed.1001547 PMID: 24223526
PLOS ONE
Depressive symptoms, suicidal ideation, and higher education
PLOS ONE | https://doi.org/10.1371/journal.pone.0299817 April 3, 2024 11 / 14
2. Liu Q, He H, Yang J, Feng X, Zhao F, Lyu J. Changes in the global burden of depression from 1990 to
2017: Findings from the Global Burden of Disease study. J Psychiatr Res. 2020 Jul1; 126:134–40.
https://doi.org/10.1016/j.jpsychires.2019.08.002 PMID: 31439359
3. Cai H, Jin Y, Liu S, Zhang Q, Zhang L, Cheung T, et al. Prevalence of suicidal ideation and planning in
patients with major depressive disorder: A meta-analysis of observation studies. J Affect Disord. 2021
Oct 1; 293:148–58. https://doi.org/10.1016/j.jad.2021.05.115 PMID: 34192629
4. Dong M, Zeng LN, Lu L, Li XH, Ungvari GS, Ng CH, et al. Prevalence of suicide attempt in individuals
with major depressive disorder: a meta-analysis of observational surveys. Psychol Med. 2019 Jul; 49
(10):1691–704. https://doi.org/10.1017/S0033291718002301 PMID: 30178722
5. Platt JM, Bates LM, Jager J, McLaughlin KA, Keyes KM. Changes in the depression gender gap from
1992 to 2014: Cohort effects and mediation by gendered social position. Soc Sci Med. 2020 Aug;
258:113088. https://doi.org/10.1016/j.socscimed.2020.113088 PMID: 32540513
6. Nock MK, Borges G, Bromet EJ, Cha CB, Kessler RC, Lee S. Suicide and Suicidal Behavior. Epidemiol
Rev. 2008 May 14; 30(1):133–54. https://doi.org/10.1093/epirev/mxn002 PMID: 18653727
7. Garnett M, Curtin S, Stone D. Suicide Mortality in the United States, 2000–2020 [Internet]. National
Center for Health Statistics (U.S.); 2022 Mar [cited 2023 Apr 11]. Available from: https://stacks.cdc.gov/
view/cdc/114217
8. Zuelke AE, Luck T, Schroeter ML, Witte AV, Hinz A, Engel C, et al. The association between unemploy-
ment and depression–Results from the population-based LIFE-adult-study. J Affect Disord. 2018 Aug;
235:399–406. https://doi.org/10.1016/j.jad.2018.04.073 PMID: 29677604
9. McGee RE, Thompson NJ. Unemployment and depression among emerging adults in 12 states,
Behavioral Risk Factor Surveillance System, 2010. Prev Chronic Dis. 2015 Mar 19; 12:E38. https://doi.
org/10.5888/pcd12.140451 PMID: 25789499
10. Paul KI, Moser K. Unemployment impairs mental health: Meta-analyses. J Vocat Behav. 2009 Jun 1; 74
(3):264–82.
11. Andersen I, Thielen K, Nygaard E, Diderichsen F. Social inequality in the prevalence of depressive dis-
orders. J Epidemiol Community Health. 2009 Jul 1; 63(7):575–81. https://doi.org/10.1136/jech.2008.
082719 PMID: 19293167
12. Kim DH, Rodrı
´guez Andre
´s A, Leigh JP. Sex-specific impact of changes in job status on suicidal idea-
tion. Crisis. 2020 Mar; 41(2):89–96. https://doi.org/10.1027/0227-5910/a000605 PMID: 31140318
13. Stauder J. Unemployment, unemployment duration, and health: selection or causation? Eur J Health
Econ. 2019 Feb; 20(1):59–73. https://doi.org/10.1007/s10198-018-0982-2 PMID: 29725787
14. Zimmerman FJ, Katon W. Socioeconomic status, depression disparities, and financial strain: what lies
behind the income-depression relationship? Health Econ. 2005 Dec; 14(12):1197–215. https://doi.org/
10.1002/hec.1011 PMID: 15945040
15. Coetzer R, Carroll E, Ruddle JA. Depression, anxiety and employment status after traumatic brain
injury. Soc Care Neurodisability. 2011 Nov 17; 2(4):200–7.
16. Infurna FJ, Gerstorf D, Ram N, Schupp J, Wagner GG, Heckhausen J. Maintaining perceived control
with unemployment facilitates future adjustment. J Vocat Behav. 2016 Apr; 93:103–19. https://doi.org/
10.1016/j.jvb.2016.01.006 PMID: 26924845
17. Crowe L, Butterworth P. The role of financial hardship, mastery and social support in the association
between employment status and depression: results from an Australian longitudinal cohort study. BMJ
Open. 2016 May; 6(5):e009834. https://doi.org/10.1136/bmjopen-2015-009834 PMID: 27235296
18. Cornacchione E, Daugherty JL. Trends in opportunity costs of U.S. postsecondary education: A national
HRD and human capital theory analysis. New Horiz Adult Educ Hum Resour Dev. 2013; 25(2):62–82.
19. Stolove CA, Galatzer-Levy IR, Bonanno GA. Emergence of depression following job loss prospectively
predicts lower rates of reemployment. Psychiatry Res. 2017 Jul; 253:79–83. https://doi.org/10.1016/j.
psychres.2017.03.036 PMID: 28359031
20. Olesen SC, Butterworth P, Leach LS, Kelaher M, Pirkis J. Mental health affects future employment as
job loss affects mental health: findings from a longitudinal population study. BMC Psychiatry. 2013 Dec;
13(1):144. https://doi.org/10.1186/1471-244X-13-144 PMID: 23705753
21. Marmot M, Wilkinson R. Social Determinants of Health. 2nd ed. OUP Oxford; 2005.
22. Irwin V, Wang K, Hein S, Zhang J, Burr R, Roberts A, et al. Report on the condition of education 2022.
Washington, DC: National Center for Education Statistics; 2022 p. 46. Report No.: 144.
23. Schuring M, Robroek SJW, Carrino L, O’Prinsen AC, Oude Hengel KM, Avendano M, et al. Does
reduced employment protection increase the employment disadvantage of workers with low education
and poorer health? J Epidemiol Community Health. 2020 Oct; 74(10):851–7. https://doi.org/10.1136/
jech-2020-213772 PMID: 32611691
PLOS ONE
Depressive symptoms, suicidal ideation, and higher education
PLOS ONE | https://doi.org/10.1371/journal.pone.0299817 April 3, 2024 12 / 14
24. McMahon WW. The external social benefits of higher education: Theory, evidence, and policy implica-
tions. J Educ Finance. 2021; 46(4):398–430.
25. Steiber N, Berghammer C, Haas B. Contextualizing the education effect on women’s employment: A
cross-national comparative analysis. J Marriage Fam. 2016; 78(1):246–61.
26. Turner JB. Economic Context and the Health Effects of Unemployment. J Health Soc Behav. 1995; 36
(3):213–29. PMID: 7594355
27. Kim JH, Park EC, Yoo KB. Effects of the gap between socioeconomic status and perceived social class
on suicidal ideation: Unique perspectives using a longitudinal analysis. Arch Gerontol Geriatr. 2015 Nov
1; 61(3):384–91. https://doi.org/10.1016/j.archger.2015.06.002 PMID: 26186964
28. Cohen AK, Nussbaum J, Weintraub MLR, Nichols CR, Yen IH. Association of adult depression with
educational attainment, aspirations, and expectations. Prev Chronic Dis. 2020 Aug 27; 17:E94. https://
doi.org/10.5888/pcd17.200098 PMID: 32857033
29. Back JH, Lee Y. Gender differences in the association between socioeconomic status (SES) and
depressive symptoms in older adults. Arch Gerontol Geriatr. 2011 May 1; 52(3):e140–4. https://doi.org/
10.1016/j.archger.2010.09.012 PMID: 20947184
30. Ross CE, Mirowsky J. Sex differences in the effect of education on depression: Resource multiplication
or resource substitution? Soc Sci Med. 2006 Sep 1; 63(5):1400–13. https://doi.org/10.1016/j.
socscimed.2006.03.013 PMID: 16644077
31. Van de Velde S, Bracke P, Levecque K. Gender differences in depression in 23 European countries.
Cross-national variation in the gender gap in depression. Soc Sci Med. 2010 Jul 1; 71(2):305–13.
https://doi.org/10.1016/j.socscimed.2010.03.035 PMID: 20483518
32. Milner AJ, Shields M, Currier D, King TL. Male-dominated occupations, employment status, and suicidal
behaviors among Australian men. Crisis. 2020 Jan; 41(1):54–63. https://doi.org/10.1027/0227-5910/
a000610 PMID: 31310165
33. Warner E, Nannarone M, Smail-Crevier R, Manuel D, Lashewicz B, Patten S, et al. The relationship
between depression risk perception and self-help behaviours in high risk Canadians: A cross-sectional
study. BMC Public Health. 2020 Dec; 20(1):1–12.
34. Curtin LR, Mohadjer LK, Dohrmann SM, Kruszon-Moran D, Mirel LB, Carroll MD, et al. National Health
and Nutrition Examination Survey: Sample design, 2007–2010. Vital Health Stat 2. 2013 Aug 1;
(160):1–23.
35. Zipf G, Chiappa M, Porter KS, Ostchega Y, Lewis BG, Dostal J. National Health and Nutrition Examina-
tion Survey: Plan and operations, 1999–2010. Vital Health Stat Ser 1 Programs Collect Proced. 2013
Aug;(56):1–37. PMID: 25078429
36. Chen TC, Clark J, Riddles MK, Mohadjer LK, Fakhouri THI. National Health and Nutrition Examination
Survey, 2015–2018: Sample design and estimation procedures. Vital Health Stat 2. 2020 Apr;(184):1–
35. PMID: 33663649
37. Johnson CL, Dohrmann SM, Burt VL, Mohadjer LK. National Health and Nutrition Examination Survey:
Sample design, 2011–2014. Vital Health Stat 2. 2014 Mar;(162):1–33. PMID: 25569458
38. Van Domelen DR, Koster A, Caserotti P, Brychta RJ, Chen KY, McClain JJ, et al. Employment and
physical activity in the U.S. Am J Prev Med. 2011 Aug; 41(2):136–45. https://doi.org/10.1016/j.amepre.
2011.03.019 PMID: 21767720
39. Frith E, Ramulu PY, Ashar B, Loprinzi PD. Association of Single and Multiple Medical Conditions with
Work Status among Adults in the United States. J Lifestyle Med. 2019 Jan; 9(1):15–26. https://doi.org/
10.15280/jlm.2019.9.1.15 PMID: 30918830
40. Levis B, Benedetti A, Thombs BD. Accuracy of Patient Health Questionnaire-9 (PHQ-9) for screening to
detect major depression: Individual participant data meta-analysis. BMJ. 2019 Apr 9; 365:l1476. https://
doi.org/10.1136/bmj.l1476 PMID: 30967483
41. Manea L, Gilbody S, McMillan D. A diagnostic meta-analysis of the Patient Health Questionnaire-9
(PHQ-9) algorithm scoring method as a screen for depression. Gen Hosp Psychiatry. 2015 Jan 1; 37
(1):67–75. https://doi.org/10.1016/j.genhosppsych.2014.09.009 PMID: 25439733
42. Kroenke K, Spitzer RL, Williams JBW. The PHQ-9. J Gen Intern Med. 2001; 16(9):606–13.
43. Litster B, Bernstein CN, Graff LA, Walker JR, Fisk JD, Patten SB, et al. Validation of the PHQ-9 for sui-
cidal ideation in persons with inflammatory bowel disease. Inflamm Bowel Dis. 2018 Jul 12; 24
(8):1641–8. https://doi.org/10.1093/ibd/izy032 PMID: 29522100
44. Altura KC, Patten SB, Fiest KM, Atta C, Bulloch AG, Jette
´N. Suicidal ideation in persons with neurologi-
cal conditions: Prevalence, associations and validation of the PHQ-9 for suicidal ideation. Gen Hosp
Psychiatry. 2016 Sep 1; 42:22–6. https://doi.org/10.1016/j.genhosppsych.2016.06.006 PMID:
27638967
PLOS ONE
Depressive symptoms, suicidal ideation, and higher education
PLOS ONE | https://doi.org/10.1371/journal.pone.0299817 April 3, 2024 13 / 14
45. Kim J, Yoon SY. Association between socioeconomic attainments and suicidal ideation by age groups
in Korea. Int J Soc Psychiatry. 2018 Nov 1; 64(7):628–36. https://doi.org/10.1177/0020764018792592
PMID: 30084278
46. Gutie
´rrez-Lobos K, Wo
¨lfl G, Scherer M, Anderer P, Schmidl-Mohl B. The gender gap in depression
reconsidered: The influence of marital and employment status on the female/male ratio of treated inci-
dence rates. Soc Psychiatry Psychiatr Epidemiol. 2000 Jun 1; 35(5):202–10. https://doi.org/10.1007/
s001270050229 PMID: 10941995
47. Aydiner-Avsar N, Piovani C. The gender impact of unemployment on mental health: A micro analysis
for the United States. Forum Soc Econ. 2021 Oct 2; 50(4):505–29.
48. Scarborough WJ, Sin R, Risman B. Attitudes and the stalled gender revolution: Egalitarianism, tradition-
alism, and ambivalence from 1977 through 2016. Gend Soc. 2019 Apr 1; 33(2):173–200.
49. Platt J, Prins S, Bates L, Keyes K. Unequal depression for equal work? How the wage gap explains gen-
dered disparities in mood disorders. Soc Sci Med. 2016 Jan 1; 149:1–8. https://doi.org/10.1016/j.
socscimed.2015.11.056 PMID: 26689629
50. Lo
¨vestad S, Lo
¨ve J, Vaez M, Waern M, Hensing G, Krantz G. Suicidal ideation and attempts in popula-
tion-based samples of women: temporal changes between 1989 and 2015. BMC Public Health. 2019
Dec; 19(1):351. https://doi.org/10.1186/s12889-019-6685-5 PMID: 30922272
51. Alghamdi H, Szira Z, Cseh-Papp I, Varga E. Analyzing Unemployment by Educational Attainment. In
2018. p. 54.
52. Feather NT, Barber JG. Depressive reactions and unemployment. J Abnorm Psychol. 1983; 92(2):185.
https://doi.org/10.1037//0021-843x.92.2.185 PMID: 6863733
53. Feather NT, Davenport PR. Unemployment and depressive affect: A motivational and attributional anal-
ysis. J Pers Soc Psychol. 1981; 41(3):422. https://doi.org/10.1037//0022-3514.41.3.422 PMID:
7288562
54. Cvetanovski J, Jex SM. Locus of control of unemployed people and its relationship to psychological and
physical well-being. Work Stress. 1994 Jan; 8(1):60–7.
55. Williams JMG, Crane C, Barnhofer T, Van der Does AJW, Segal ZV. Recurrence of suicidal ideation
across depressive episodes. J Affect Disord. 2006 Apr 1; 91(2):189–94. https://doi.org/10.1016/j.jad.
2006.01.002 PMID: 16488023
56. Dalglish SL, Melchior M, Younes N, Surkan PJ. Work characteristics and suicidal ideation in young
adults in France. Soc Psychiatry Psychiatr Epidemiol. 2015 Apr; 50(4):613–20. https://doi.org/10.1007/
s00127-014-0969-y PMID: 25308058
57. Na PJ, Yaramala SR, Kim JA, Kim H, Goes FS, Zandi PP, et al. The PHQ-9 Item 9 based screening for
suicide risk: a validation study of the Patient Health Questionnaire (PHQ)−9 Item 9 with the Columbia
Suicide Severity Rating Scale (C-SSRS). J Affect Disord. 2018 May; 232:34–40. https://doi.org/10.
1016/j.jad.2018.02.045 PMID: 29477096
58. Andreotti ET, Ipuchima JR, Cazella SC, Beria P, Bortoncello CF, Silveira RC, et al. Instruments to
assess suicide risk: A systematic review. Trends Psychiatry Psychother. 42(3):276–83. https://doi.org/
10.1590/2237-6089-2019-0092 PMID: 32997043
PLOS ONE
Depressive symptoms, suicidal ideation, and higher education
PLOS ONE | https://doi.org/10.1371/journal.pone.0299817 April 3, 2024 14 / 14
Available via license: CC BY 4.0
Content may be subject to copyright.
