Social Isolation in a Population-Based Sample of Emerging Adults: Who Is on Their Own?

Abstract

Purpose:

This study aimed to identify the associations between social isolation, social contextual factors, and behavioral and psychological health during emerging adulthood.

Methods:

Participants (n = 1,568) were drawn from EAT 2018 (mean age = 22.1 ± 2.0). Logistic regressions tested the associations between social isolation, contextual factors, and behavioral and psychological health.

Results:

Approximately 16% of the sample reported social isolation. Emerging adults with socially marginalized identities (e.g., racial/ethnic, socioeconomic), who were unemployed, and/or lived alone had higher odds of experiencing social isolation. Social isolation was associated with more risky health behaviors and lower self-esteem.

Discussion:

Emerging adults who were socially disadvantaged (e.g., unemployed) had greater risk of social isolation. Findings suggest that social isolation is more strongly linked with behavioral health than psychological health. Generally, findings suggest that access to community resources and areas wherein social connections can be formed may to be important for behavioral health in emerging adulthood.

Emerging adulthood is the distinct period between adolescence and young adulthood (Arnett, 2000) wherein young people transition from high school into college or the workforce resulting in major changes to an individual’s social roles (e.g., student to full-time employee) and contexts (Branje et al., 2014). Arnett’s theory of emerging adulthood suggests that this developmental period is characterized by increased autonomy, identity exploration, instability, a focus on the self, and feeling “in-between” (Arnett, 2000, 2007). The major transitions in social context during emerging adulthood have been linked with frequent demographic changes including changes to young peoples’ place of residency and occupation (Branje et al., 2014) as young people explore new roles, worldviews, and life directions (Arnett, 2000). This instability often disrupts young people’s social networks (Lapierre & Poulin, 2020), leading them to forge relationships in their new contexts and roles (Arnett, 2000). A failure to develop and maintain relationships can negatively impact behavioral and psychological health (Child & Lawton, 2019; Lapierre & Poulin, 2020). This brings cause for concern, as emerging adulthood is a vulnerable period associated with heightened behavioral and psychological health risk (Lapierre & Poulin, 2020; Sussman & Arnett, 2014).

The development of close friendships during emerging adulthood is crucial as social relationships provide young people with a sense of a belonging and purpose, which may help to reduce behavioral and psychological health risk (Umberson & Montez, 2010). Yet, some young people may struggle to form meaningful social connections in their new settings (Lapierre & Poulin, 2020), which may leave them socially isolated. Importantly, social isolation is categorically distinct from loneliness. Social isolation refers to the physical absence of close social relationships in an individual’s life (Child & Lawton, 2019). In contrast, loneliness is characterized as perceived isolation, wherein an individual may feel alone whether or not they are actively maintaining close interpersonal relationships (Child & Lawton, 2019). Social isolation has been linked to a number of negative behavioral health outcomes (e.g., substance use; Sussman & Arnett, 2014) and psychological health concerns, such as depression (Santini et al., 2015) and suicide attempts (Calati et al., 2019). A majority of social isolation research has focused on older adults (>60 years old; Nicholson, 2012; Steptoe et al., 2013), leaving many unanswered questions about the demographic and contextual factors linked with social isolation as well as the associations between social isolation and psychological and behavioral health during emerging adulthood.

The present study aimed to fill two critical gaps in the literature regarding social isolation during emerging adulthood. During emerging adulthood, both close friendships and romantic relationships are important forms of social connectivity and the importance of each type of relationship is fluid during this period based on romantic relationship status (Umemura et al., 2017). For instance, emerging adults may spend a period of time focusing on the development of a close romantic relationship, which may weaken intimacy or closeness with friends (Pitman & Scharfe, 2010; Umemura et al., 2017). Because it is developmentally normative for the relative importance of friendships and romantic relationships to differ as emerging adults enter and exit romantic relationships (Pitman & Scharfe, 2010; Umemura et al., 2017), social isolation in the present study was conceptualized as the absence of both close friendships and a close romantic partner. The first aim was to identify contextual and sociodemographic factors linked with social isolation during emerging adulthood in a racially/ethnically and socioeconomically diverse sample. Specifically, several social contexts that are likely to shift during emerging adulthood were examined as predictors of social isolation (Arnett, 2000, 2007), including: parental status, job status, living situation, and student status. Key markers of emerging adulthood (e.g., changes in job or student status) and connectedness during emerging adulthood have been shown to differ by demographic factors, such as socioeconomic status (SES), race/ethnicity, and age (Schwartz, 2016; Syed & Mitchell, 2013). Thus, the associations of race/ethnicity, SES, and age with social isolation were also examined. Given the exploratory nature of this aim, there were no a priori hypotheses about the directions of these associations. The second aim of the study was to identify the associations between social isolation, psychological health, and behavioral health in a population-based sample of young adults. It was hypothesized that social isolation during emerging adulthood would be linked with higher rates of health risk behaviors and poorer psychological health.

Method

Study Design and Participants

Data were derived from EAT 2018 (Eating and Activity over Time), the second wave of a population-based study examining eating and weight behaviors and their associated factors in young people. During the first wave of data collection (EAT 2010), 2,793 middle and senior high school students from 20 urban public schools in Minneapolis-St. Paul, Minnesota, in the United States completed classroom surveys and anthropometric measures (Neumark-Sztainer et al., 2012). Specific schools were chosen on the basis of students’ demographic characteristics, as an important goal of the study was to learn about the weight-related health of ethnically/racially and socioeconomically diverse adolescents. As such, the sample includes a higher prevalence of Black or African American, Asian, Latinx, and Mixed or other race emerging adults (81%) relative to estimates of the demographic distribution of emerging adults of color (35%) in the Twin Cities (Minnesota Compass, n.d.).

For EAT 2018, original participants were mailed letters inviting them to complete the survey. The response rate for EAT 2018 was 65.8% of the original participants who could be contacted at follow-up. All study protocols were approved by the University of Minnesota’s Institutional Review Board Human Subjects Committee.

Attrition from EAT 2010 to EAT 2018 did not occur completely at random, wherein nonresponders at EAT 2018 were more likely to be male, non-white, and have parents with low educational attainment in 2010 than those who participated in EAT 2018. Consistent with other analyses using this cohort (e.g., Hazzard et al., 2020), inverse probability weighting (IPW) was used to manage missing data (Little, 1986). IPW minimized response bias and allows for extrapolation back to the EAT 2010 school-based sample. Weights for IPW were derived as the inverse of the estimated probability that an individual responded at the two time points based on characteristics reported in 2010. After weighting, there were no significant differences between the analytic sample and the full EAT 2010 sample (p > 0.9).

EAT Survey Development

The EAT 2018 survey was modified from baseline to improve the relevance of items for young adults (Larson et al., 2013). Scale psychometric properties were examined in the EAT 2018 sample. Focus groups (n = 29) were conducted to pretest the EAT 2018 survey and the test-retest reliability of measures was examined using data from a subgroup of 112 participants who completed the EAT 2018 survey twice within a period of 3 weeks. Survey development and test-retest reliability for baseline measures have been described elsewhere (Neumark-Sztainer et al., 2012).

Measures

Social isolation.

Participants were asked if they had: (1) a significant other; (2) a closest friend; (3) a second closest friend; and (4) a third closest friend. Participants who indicated that they did not have any close friends, nor did they have a significant other were categorized as experiencing social isolation.

Demographic characteristics.

Sex was assessed at EAT 2018 and included options: Male, Female, or different identity [test-retest = agreement 100%]. Due to a lack of power (n = 11), participants with a different identity were excluded from analyses. Age was calculated with participants’ date of birth and EAT 2018 survey completion date [test-retest agreement = 100%]. Race/ethnicity was assessed at EAT 2010 and included categories: White; Black/African American; Hispanic/Latinx; Asian; or Mixed or Other (test-retest agreement: 98%–100%). Socioeconomic status (SES) was assessed at EAT 2010 and included categories: low, low-middle, middle, upper-middle, and upper (test-retest r = 0.90; Sherwood et al., 2009). SES was based parent education attainment and employment, eligibility for free/reduced-price school meals, and public assistance.

Contextual factors.

All contextual factors were assessed at EAT 2018. Parental status categories included: no children; one child; two or more children. Job status categories included: working full-time; working part-time; currently unemployed, but actively seeking work; and other [test-retest agreement = 83%]. Living situation categories included: alone; with parents; with roommates; or other [test-retest agreement = 89%]. Student status categories included: not a student; high school or post secondary; part- or full-time community or technical college; part- or full-time 4-year college; and graduate or professional student (test-retest % agreement = 92%).

Psychological health.

All psychological health variables were measured during EAT 2018. Depressive symptoms were assessed with a six-item scale developed by Kandel and Davies (Kandel & Davies, 1982). Responses from 1 (not at all) to 3 (very much) were averaged, and the average was multiplied by 10. Scores ranged from 10 to 30, with higher scores indicating higher depressive symptoms (test-retest r = .71). Self-esteem was assessed with six items from the Rosenberg Self-Esteem Scale (Rosenberg, 1965). Responses from 1 (strongly disagree) to 4 (strongly agree) were summed (Range: 6–25), with lower scores indicating lower self-esteem (test-retest r = .78). Stress was measured with an item asking participants to rate their average stress in the past 30 days. Options ranged from 1 (not at all stressed) to 10 (very stressed) (test-retest r = .69; Vinkour & Schul, 2002).

Behavioral health.

All behavioral health variables were measured during EAT 2018. Extreme unhealthy weight control behavior (UWCB) used in the past year (range: 0–4) included: diet pills, self-induced vomiting, laxative use, and diuretics (test-retest r = .90; Neumark-Sztainer et al., 2002). Past-year less extreme UWCBs (range: 0–5) included: fasting, eating very little, used food substitute, skipped meals, or smoked more cigarettes (test-retest r = .63; Neumark-Sztainer et al., 2002). Substance use was assessed with the question: “How often have you used the following during the past year? Other drugs (cocaine, heroin, meth, etc.)” Responses from 1 (never) to5 (daily) were dichotomized (yes/no).

Analytic Strategy

Descriptive statistics identified the frequency of social isolation and substance use and average frequency of health risk behaviors and mean psychological health in the sample. The association between demographic and contextual factors with social isolation were examined with univariate logistic regression models. Each analysis was run with multiple reference groups, to allow for a thorough understanding of the associations between the predictors and social isolation. The present study included minimal item-level missingness on the variables of interest, thus a complete cases approach was used, wherein participants with complete responses to the analytic items were included. Next, generalized linear models (GLM) examined the association between social isolation and several outcomes including: unhealthy weight control behaviors (UWCBs), depressive symptoms, self-esteem, stress, and substance use. Poisson GLM was applied for count outcomes when the mean and the variance are approximately equal (less extreme UWCBs), negative binomial GLM was used for count data with overdispersion (extreme UWCBs), linear GLM was applied for continuous outcomes (depressive symptoms, self-esteem, stress), and logistic GLM was used for dichotomous outcomes (substance use). All GLM models included age, ethnic/racial identity, and SES as covariates. A sensitivity analysis examined associations between partial social isolation (only a close romantic partner), health behaviors, and psychological health. Sensitivity analysis revealed non-significant results for all of the dependent variables among partially isolated emerging adults and thus the results were excluded from the study. All syntax associated with the present analyses has been uploaded as an online supplementary file.

Results

The sample of emerging adults included 239 participants who met criteria for social isolation (15.8%). Moreover, approximately 55% of the socially isolated emerging adults were male, relative to around 47% of those in the full sample. With regard to race/ethnicity, around 42% of socially isolated young people identified as Black/African American relative to around 29% in the full sample. All descriptive statistics in the full sample and among the socially isolated subgroup are presented in Table 1.

Table 1.

Weighted Sample Descriptive Statistics and Adjusted Odds of Experiencing Social Isolation Relative to Non-Socially Isolated Emerging Adult Across Demographic and Contextual Factors.

	Full Sample n = 1,568	Social Isolation n = 239
Variables	%	%	OR	95% CIs	Odds of Social Isolation Across Reference Groups
Sociodemographic Variables
Sex
1. Male	47.0	55.4	—	—
2. Female	53.0	46.6	1.55	1.19–2.03
Ethnicity/Race
1. White	18.9	6.8	—	—	2 > 3, 4; 2 = 5 3 < 4, 5 4 = 5
2. Black/African American	29.3	42.4	5.38***	3.25–8.90
3. Hispanic/Latinx	16.9	11.6	2.22**	1.24–4.00
4. Asian	19.9	21.2	3.63***	2.14–6.16
5. Mixed or Other Race	15.0	18.0	4.22***	2.42–7.34
Socioeconomic Status
1. Upper	7.3	3.0	—	—	1 = 2 3 = 4= 5 1, 2 < 3, 4, 5
2. Upper-middle	13.2	7.2	1.39	0.59–3.25
3. Middle	18.0	16.5	2.60*	1.18–5.73
4. Low-middle	22.5	28.8	3.88***	1.81–8.31
5. Low	38.9	44.5	3.39**	1.61–7.14
Age (Mean, SD)	2.03	2.06	1.00	0.94–1.07
Contextual Variables
Parental Status
1. No children	78.5	76.1	—	—	1 = 2 = 3
2. One child	13.6	16.7	1.43	0.98–2.09
3. Two or more children	7.8	7.3	1.01	0.60–1.71
Job Status
1. Working full-time	51.4	43.1	—	—	2 <3; 2 = 4 3 > 4
2. Working part-time	27.6	25.0	1.08	0.77–1.50
3. Unemployed	11.6	20.6	2.65***	1.83–3.86
4. Other (e.g., caregiver)	9.3	11.3	1.58	1.00–2.48
Living Situation
1. With roommates	19.6	12.0	—	—	2 = 3; 2 > 4 3 > 4
2. Alone	8.1	13.6	3.59***	2.10–6.15
3. With parents	46.1	57.6	2.54***	1.69–3.83
4. Other	26.2	16.8	1.13	0.70–1.84
Student Status
1. Not a student	55.5	56.6	—	—	2 = 3, 5; 2 < 4 3 = 4, 5 4 = 5
2. HS or postsecondary	3.7	5.7	1.90*	1.03–3.53
3. PT/FT community college	19.1	18.4	0.99	0.69–1.41
4. PT/FT 4-year college	19.6	16.0	0.71	0.49–1.03
5. Graduate or professional	2.1	3.3	1.52	0.68–3.41

Social Isolation Across Sociodemographic Variables

Odds of experiencing social isolation during emerging adulthood differed by sex, ethnicity/race, and SES but not age (see Table 1). Specifically, men had higher odds of social isolation relative to women (OR = 1.55, 95% CIs; 1.19, 2.03). Relative to White young people, Black/African American (OR = 5.38, 95% CIs; 3.25, 8.90), Asian (OR = 3.63, 95% CIs; 2.14, 6.16), Hispanic/Latinx (OR = 2.22, 95% CIs; 1.24, 4.99), and Mixed or other race (OR = 4.22, 95% CIs; 2.42, 7.34) young people had higher odds of social isolation. Moreover, Black/African American emerging adults had higher odds of social isolation than did Asian (OR = 1.48, 95% CIs; 1.04, 2.11), and Hispanic/Latinx (OR = 2.42, 95% CIs; 1.56, 3.74) emerging adults. Asian young people had higher odds of social isolation relative to Hispanic/Latinx (OR = 1.63, 95% CIs; 1.03, 2.60) young people. Relative to those with an upper SES, odds of experiencing social isolation were higher in those with a low (OR = 3.39, 95% CIs; 1.61, 7.14), low-middle (OR = 3.88, 95% CIs; 1.81, 8.31), and middle (OR = 2.60, 95% CIs; 1.18, 5.73) SES. There were no significant differences in odds of social isolation in emerging adults with an upper-middle or upper SES.

Social Isolation by Contextual Variables

Odds of experiencing social isolation during emerging adulthood differed by job status, living situation, and student status, but not parental status. Specifically, unemployed young people had higher odds of social isolation relative to those who were employed full-time (OR = 2.65, 95% CIs; 1.83, 3.86), part-time (OR = 2.46, 95% CIs; 1.63, 3.72), and those with another employment status (OR = 1.68, 95% CIs; 1.01, 2.81). Further, emerging adults who lived alone had higher odds of experiencing social isolation relative to those who lived with roommates (OR = 3.59, 95% CIs; 2.10, 6.15) and those with another living situation (OR = 3.18, 95% CIs; 1.92, 5.26). Young people who lived with their parents also had higher odds of social isolation relative to those who lived with roommates (OR = 2.54, 95% CIs; 1.69, 3.83) and those with another living situation (OR = 2.25, 95% CIs; 1.57, 3.23). Additionally, emerging adults who were not students had higher odds of social isolation relative to high school students (OR = 1.90, 95% CIs; 1.03, 3.53), but did not differ from community college, 4-year college, or graduate school students. Finally, emerging adults at a 4-year university had higher odds of social isolation (OR = 2.68, 95% CIs; 1.37, 5.28) relative to high school students.

Social Isolation and Psychological Health Outcomes

The results from GLM models partially supported the hypothesis that social isolation is associated with negative psychological health outcomes (see Table 2). Specifically, socially isolated emerging adults reported lower self-esteem relative to their non-socially isolated peers (B = ‒0.17, SE = 0.07). In contrast, the results from the tested GLM models did not support the hypothesized relationships between social isolation, depressive symptoms, and stress.

Table 2.

Adjusted Results From Univariate GLM Models Wherein Weight Control Behaviors, Depressive Symptoms, Substance Use, Self-Esteem, and Stress Were Regressed Onto Social Isolation.

Domain Variable		Non-Socially Isolated N = 1,329 (84.2%)		Socially Isolated N = 239 (15.8%)		GLM Results
Psychological Health	Range	M	SD	M	SD	B	SE	Cohen’s d
Depressive symptoms	10–30	18.4	5.85	18.1	6.18	0.02	0.07	0.05
Self-esteem	6–25	17.5	3.57	16.9	3.36	−0.17*	0.07	0.17
Stress	1–10	1.28	1.37	1.42	1.79	−0.01	0.07	0.09
Behavioral Health	Range	M	SD	M	SD	ORa	95% CIs
Extreme UWCBs	0–4	0.15	0.45	0.30	0.82	2.31***	1.69–3.16
Less extreme UWCBs	0–5	1.09	1.29	1.30	1.45	1.21**	1.07–1.37
Substance use (N, %)	0–1	65	5.2	21	8.6	1.81*	1.06–3.08

Note. UWCBs = unhealthy weight control behaviors; M = mean; SD = standard deviation; B = standardized beta; SE = standardized standard error; OR = Odds Ratio; CI = confidence interval;

^a= non-socially isolated emerging adults are the reference group;

^*= p < .05;

^**= p < .01;

^***= p < .001;

All models control for the effect of gender, SES, and ethnic/racial identity

Social Isolation and Behavioral Health Outcomes

The results from GLM models supported the hypothesis that social isolation is positively associated with behavioral health (See Table 2). Specifically, socially isolated emerging adults reported more than two times higher count of extreme UWCBs relative to their non-socially peers (OR = 2.31, 95% CIs; 1.69, 3.16), above and beyond the effects of sex, ethnicity/race, and SES. Socially isolated young people also reported a higher count of less extreme UWCBs relative to their non-socially isolated peers (OR = 1.21, 95% CIs; 1.07, 1.37). Finally, socially isolated emerging adults had higher odds of substance use than their non-socially isolated peers, above and beyond the effect of sex, ethnicity/race, and SES (OR = 1.81, 95% CIs; 1.06, 3.08).

Discussion

The present study aimed to identify sociodemographic and contextual factors linked with social isolation during emerging adulthood and to examine the association between social isolation and behavioral and psychological health. Nearly 16% of the sample experienced social isolation, which suggests that social isolation is relatively common and highlights the importance of understanding the factors linked with social isolation in emerging adulthood. The present findings demonstrated that social isolation risk differed by demographic and contextual factors. Furthermore, social isolation was associated with several behavioral health concerns but was only one psychological health concern. Each of the main study findings are described in detail.

Contextual Factors Linked With Social Isolation

The current study revealed that men are more likely to be socially isolated than women, which is consistent with past research on social isolation during emerging adulthood (Adamczyk, 2016). For instance, past research suggests that men may have smaller social networks during emerging adulthood relative to women because men may be more likely to seek social support and intimacy from romantic partners, whereas women may seek social support and connection from friends, family, and partners (Adamczyk, 2016; Vandervoort, 2000). Thus, it is possible that men who are not romantically partnered are at greater risk of experiencing social isolation during emerging adulthood because of their potentially smaller social network. Similarly, recent research suggests that romantic deficits during emerging adulthood are associated with greater loneliness among men than they are among women (Hopmeyer et al., 2020), and thus it is possible that men more strongly rely on intimate partnerships at this development period relative to women.

The results from the study also revealed that emerging adults of color, with a lower SES, and/or those who are unemployed were more likely to experience social isolation relative to young people who were White, of higher SES, and/or at least part-time employed. Specifically, emerging adults of color were more likely to experience social isolation relative to their White peers, with Black/African American and Mixed or other race young people experiencing the greatest risk. While the Twin Cities metropolitan area is ethnically and racially diverse, the sample includes a higher percentage of emerging adults with socially minoritized racial and ethnic identities (approximately 81%) than is estimated for emerging adults in the region (approximately 35%; Minnesota Compass, n.d.) and thus the overrepresentation of these young people in the social isolation subgroup is of concern. Moreover, the Twin Cities metropolitan area is racially and ethnically segregated (Orfield, 2015), and thus it is possible that emerging adults of color have less access to social support, thus driving disparities in social connectivity. Indeed, some theories suggest that exposure to interpersonal racism can impact the ability to form close relationships and that social isolation is one mechanism though which racial health disparities emerge (Brondolo et al., 2012). Additionally, emerging adults with a lower SES have more time constraints than those with higher SES (Pelletier & Laska, 2012), which may be partially due to longer working hours for lower pay. Thus, young people with a lower SES may have less available time to foster friendships or to engage in identity development and career exploration activities typically associated with emerging adulthood (Arnett, 2000). Similarly, unemployed young people may not have access to new social networks after grade school, which may in turn increase their odds of experiencing social isolation.

The present findings suggest that the transition to a 4-year college, which often includes relocation to a new region, may be a particularly risky time period as students in a 4-year college were more likely to experience social isolation than high school students. Moreover, consistent with past research (Cacioppo et al., 2003), living alone or with parents was linked with higher odds of experiencing social isolation relative to those who lived with roommates or others. To this end, it has been theorized that identity formation and close social relationships are inherently intertwined (Erickson, 1968, 1982) and thus, emerging adults who live alone or with their parents during emerging adulthood, a period characterized by accelerated autonomy and identity development, may be less socially connected to peers (Arnett, 2000; Doumen et al., 2012; Koepke & Denissen, 2012).

Social Isolation and Behavioral and Psychological Health

The present findings suggest that social isolation is associated with, or indicative of engagement in risky health behaviors (e.g., UWCBs, substance use). It is possible that socially isolated young people engage in risky health behaviors to self-medicate due to a lack of social support during young adulthood (Kahle et al., 2020). Further, self-esteem was the only psychological health outcome associated with social isolation, wherein social isolation was linked with lower self-esteem. Social relationships have been shown to give young people a sense of purpose and belonging (Umberson & Montez, 2010), thus those who are not well connected may internalize their lack of friendships as a reflection of their worth. It is also possible that the failure to develop close relationships as young people enter new social contexts and roles during emerging adulthood reduces self-esteem. Given the cross-sectional nature of the analyses, it may also be that low self-esteem predicts social isolation such that young people with low self-esteem less frequently engage in settings wherein new friendships could be forged.

Strengths and Limitations

The present study contained several strengths and limitations. This study is the first known population-based study to identify demographic and contextual predictors of social isolation and the links between social isolation and psychological and behavioral health in emerging adults. The present results offer information about the subgroups of emerging adults who are at greatest risk of experiencing social isolation and offer a foundation for future research on social isolation during this period. Other strengths include the racially/ethnically diverse sample of men and women. The current sample was collected from schools in the metropolitan areas of the Twin Cities, and thus future studies should examine social isolation during emerging adulthood in a national sample. The current study did not assess homelessness as a potential living situation. Given the stigma associated with homelessness, future research should examine the association between homelessness and social isolation. Moreover, the present study did not measure social isolation with a validated psychometric tool and thus, future research should assess potential predictors of social isolation with validated psychometric assessments. Finally, the present study did not assess loneliness. While social isolation and loneliness are two overlapping constructs, the two constructs are only moderately correlated (Matthews et al., 2016) and are associated with differing psychological and behavioral health concerns (Steptoe et al., 2013). Future research should examine social isolation and loneliness in emerging adulthood.

Implications

The results from this study offer insight into the subgroups of the emerging adult population who may be at greatest risk of social isolation, as well as the association between social isolation and behavioral health. Specifically, emerging adult populations for whom social isolation risk was greatest, include: men; Black/African American, Asian, and Mixed or other race emerging adults; those transitioning to a 4-year college, experiencing unemployment, who live alone, and/or have a lower SES. Moreover, social isolation was positively associated with multiple behavioral health variables, but not psychological health variables. Thus, access to community resources, among other areas wherein social connections during emerging adulthood, may be important for behavioral health during emerging adulthood. For instance, 4-year colleges may integrate social programming for new students to reduce social isolation risk during this transition. Future longitudinal research should seek to distinguish characteristics that predict social isolation (e.g., structural barriers, racism) from those that are byproducts of experiencing social isolation.

Author Biographies

Melissa Simone, PhD is a postdoctoral research fellow in the Midwest Regional T32 Training Program in Eating Disorders Research at the University of Minnesota. Their research focuses on socioecological processes contributing to psychological health disparities, with a focus on sexual and gender minority populations, as well as eating disorders and concurrent mental health concerns.

Katie Loth, PhD, MPH, RD is an assistant professor at the University of Minnesota Department of Family Medicine and Community Health. Her research explores social and environmental influences on child and adolescent weight status and disordered eating behaviors. Loth is also a practicing clinical dietician at a broad spectrum family medicine clinic.

Carol B. Peterson, PhD is an associate professor at the University of Minnesota Medical School Department of Psychiatry and Behavioral Sciences. She is an adult psychologist whose research focuses on identifying novel approaches to treat eating disorders, particularly by targeting the psychological and neurobiological factors that maintain these conditions.

Jerica M. Berge, PhD, MPH, LMFT, CFLE is a professor and vice chair for research in the Department of Family Medicine and Community Health at the University of Minnesota. She is both a behavioral medicine clinician and researcher. She examines emotional and physical indicators of health across the life course.

Marla E. Eisenberg, ScD, MPH is a professor in the Department of Pediatrics at the University of Minnesota School of Medicine. Her research focuses on social influences on high risk health behaviors among adolescents and young adults.

Dianne Neumark-Sztainer, PhD, MPH, RD is a professor and the division head of the Division of Epidemiology and Community Health at the University of Minnesota School of Public Health. Her research focuses on the prevention and reduction of weight-related problems, including obesity, poor dietary intake, inadequate physical activity, disordered eating behaviors, and body dissatisfaction.