We conducted a cross-sectional study among Midwest farmers to examine whether perceived stress and symptoms of depression differ by co-op membership, and how these outcomes differ in relation to the frequency of use and satisfaction of co-op and non-co-op programs and social support. The Midwest is a north central region in the United States comprising North Dakota, Minnesota, Wisconsin, Michigan, South Dakota, Nebraska, Iowa, Kansas, Missouri, Illinois, Indiana, and Ohio (Figure 1). A survey with online and phone participation options was administered from January to May 2021. Online participants were recruited from farmer organizations in which the responding staff emailed the study invitation to their farmer-members. Farmer organizations we reached out to include farmer-to-consumer groups (e.g., farmers’ markets, online buying group), member organizations (e.g., Practical Farmers of Iowa, Farmers Union), and extension offices. Phone participants were recruited from a subscription list purchased from an agricultural data marketer. Study participant eligibility criteria were farm operators who were 18 years or older AND managed a farm during the last 12 months. Participants completing the survey received a $15 check. The study was approved by the University of Iowa Institutional Review Board.

The primary exposures were the frequency of programs used and program satisfaction. Programs encompassed marketing and sales, technical support, financial services, education, policy engagement, and farmer support networks from co-ops and non-co-ops. Frequency of programs used was measured by a 5-point scale (0 = never to 4 = very often). Participants who indicated using any programs were asked to rank program satisfaction using a 4-point scale (0 = dissatisfied to 3 = satisfied). The secondary exposure was social support, measuring by a 4-item scale that was identified in studies examining stress following the 1980s farm crisis.^{16, 17} Participants ranked their emotional support, financial support, tangible support, and informational support from disagree (0) to agree (3).

The outcome measures, perceived stress and symptoms of depression, were assessed by the 10-item Perceived Stress Scale (PSS) and the 10-item Center for Epidemiologic Studies Depression Scale (CES-D), respectively.^18,19 The PSS instrument assesses stress on a 5-point scale (0 = never to 4 = very often) related to unpredictable, uncontrollable, and overloading life events occurring in the last month.²⁰ The maximum total PSS score is 40 and high scores suggest increased stress levels. The CES-D instrument assesses feelings in the last week on a 4-point scale (0 = none of the time to 3 = all of the time).²¹ The maximum total CES-D score is 30 and high scores indicate increased symptoms of depression. Both tools have been used to examine stress and depression among farmers.^22–25

The use of each program was aggregated into never and ever due to its left-skewed distribution. Never included participants who never used a program from co-ops and non-co-ops. Ever included participants who used a program at least once from at least one source. Program satisfaction was collapsed into not completely satisfied and completely satisfied due to its right-skewed distribution. Not completely satisfied included participants who were dissatisfied, slightly dissatisfied, and slightly satisfied with programs from co-ops and non-co-ops and completely satisfied included participants who were satisfied with programs from at least one source. “Slightly satisfied” was grouped with not completely satisfied because of small cell sizes.

The total scores of PSS and CES-D were calculated for each participant by first reversing scores of positively phrased symptoms on each instrument and then by summing up their respective 10-item scores.^{18, 19} Residuals between program use, social support, and CES-D scores were not distributed normally, even after log-transforming the scores. We converted these CES-D scores into a five-level ordinal categorical variable.²⁶

We investigated whether participants differed across demographics and farm operation characteristics using Chi-square tests and whether outcome variables differed by co-op membership and exposure variables using two sample t-tests and Kruskal-Wallis tests. Associations between exposure and outcome variables were examined using multivariable linear and ordinal logistic regressions. The regression models first examined whether each exposure variable was associated with the outcome variables. All exposure variables with statistical significance were then included in subsequent models to investigate their effects on the outcome variables. All analyses were conducted using SAS version 9.4 (SAS Institute Inc., Cary, North Carolina).

The total sample encompassed 307 participants (197 co-op, 110 non-co-op), among them 112 participated by phone and 195 through online. The phone response rate was 9.9%. The online response rate was undetermined because the number of individuals that received the study invitation from farm organizations was unknown. Co-op participants had a significantly higher mean age than non-co-op participants [co-op = 56 years (SD = 14), non-co-op = 47 years (SD = 13), p < 0.0001]. Co-op and non-co-op participants were also significantly (p < 0.05, Table 1) different in demographics and farming characteristics but similar in mental health stressors.

The perceived stress analysis used the full sample (N = 307) while the symptoms of depression analysis excluded eight participants who skipped the CES-D scale (N = 299: 191 co-op, 108 non-co-op). Co-op participants had a significantly lower perceived stress mean score than non-co-op participants [co-op = 15 (standard deviation [SD] = 6.48), non-co-op = 19 (SD = 4.92), p < 0.0001]. The mean perceived stress scores were in the range of 14 to 26 for moderate stress. Co-op participants also had a lower mean score for symptoms of depression than non-co-op participants, but this difference was statistically nonsignificant [co-op = 7 (SD = 4.53), non-co-op = 8 (SD = 5.13), p = 0.1464]. The mean depression symptom scores were lower than the cut-off 10 for significant depression risk.

Participants who ever used a program reported lower stress scores than participants who never used the same program, regardless of co-op membership. The decreased stress scores were significant (p < 0.05) for technical support among all participants, and for financial services and education among non-co-op participants. Participants who ever used support networks reported significantly (p < 0.05) higher stress scores than participants who never used this program. Associations between depression symptom scores and each of the programs was statistically non-significant (p ≥ 0.05).

Participants who were completely satisfied with programs reported lower stress and depression scores than participants who were not completely satisfied with the same programs. The lower stress scores were significant (p < 0.05) across all program satisfaction ratings except for support networks among co-op participants and all programs among non-co-op participants. The low depression symptoms scores were significant (p < 0.05) across all program satisfaction for co-op and non-co-op participants except for policy engagement and support networks.

Co-op and non-co-op participants had significantly (p < 0.05) lower perceived stress and symptoms of depression scores when they agreed that tangible support, financial support, and emotional support were available, compared to participants who slightly agreed, slightly disagreed, or disagreed. Non-co-op participants who disagreed with having informational support reported significantly (p < 0.05) lower perceived stress scores than participants who slightly disagreed, slightly agreed, or agreed with having this support.

The final analyses only included significant programs. Being older, having a large farm size (> 1,000 acres), and using educational programs were associated with decreased perceived stress, while being a female and using antidepressant medications were associated with increased perceived stress (Table 2). Satisfaction with financial services was associated with decreased symptoms of depression, while using antidepressant medications was associated with increased symptoms of depression.

The final analyses included all social support measures with statistical significance. The results show that being older and having tangible support and financial support were associated with decreased perceived stress, while being a female and using antidepressant medications were associated with increased perceived stress (Table 3). Being older and having emotional support were associated with decreased symptoms of depression, while using antidepressant medications and experiencing financial hardships were associated with increased symptoms of depression.

Our sample encompassed 64% co-op and 36% non-co-op members, which was close to the national farmer demographics. The USDA agricultural cooperative statistics documented nearly 1.9 million (56%) co-op members in 2017.^Citation27 Co-op participants were primarily experienced male farmers of large grain operations, whereas many non-co-op participants were beginning female farmers of small-scale vegetables and other specialty crops. These characteristics were consistent with the sampled region that mainly produces grain while also cultivating diverse crops and livestock. Furthermore, female farmers tend to engage in diverse small-scale productions.^Citation28

Co-op participants reported fewer symptoms of perceived stress and depression than non-co-op participants and the difference was only statistically significant (p < 0.05) for perceived stress. Regression analyses revealed three key findings. First, the use of educational programs was only protective against perceived stress, and the use of any program did not predict symptoms of depression. This observation was not surprising. Depression is a chronic condition, and it is likely shaped by factors with long-term effects, such as financial hardships, relationship problems, and social isolation. It was also possible that participants who experienced depression symptoms were less likely to seek out programs. Second, satisfaction with programs was protective against perceived stress and symptoms of depression. As satisfaction across all programs except for education increased, perceived stress decreased. Likewise, as satisfaction with marketing and sales, technical support, financial services, and education increased, symptoms of depression decreased. These results suggest that co-ops and non-co-ops influence farmers’ stress and depression through program offerings related to market stability, financial services, and social engagement. In addition, the quality of relationships with co-ops and non-co-ops (i.e., high program satisfaction) was more indicative of farmers’ mental health than the level of interaction (i.e., high frequency of participating in a program). These findings establish evidence of the protective effects of organizational participation on farmers’ mental health.

Third, social support was protective against perceived stress and symptoms of depression. Participants having social support reported lower stress and depression scores than participants who did not have social support. Regression analysis revealed that when multiple types of support are available, only tangible support and financial support decreased perceived stress, and only emotional support decreased symptoms of depression. While the results confirm that social support reduces psychological distress among older farmers,¹⁷ they point out that it does not reduce stress for farmers who are a female or using antidepressants. These observations were consistent with the literature that the mental health effects of social support are specific to the type of support⁷, and the heterogeneity is likely due to variability in social support measures.⁹

The overall findings of decreased perceived stress and symptoms of depression that were associated with specific programs, program satisfaction and social support are consistent with the literature that social connectedness lessens mental health risks.^29,30 Specifically, social connectedness creates psychological gains (e.g., sense of belonging, shared social identity), influences healthy behaviors, and provides coping resources.^31,32 For instance, US adults who had diverse social engagement such as attending faith-based activities and interacting with friends reported fewer symptoms of depression than adults whose social networks were limited to few family members or friends.³¹ Cruwys et al found that depression relapse reduced by 63% when increasing social groups from zero to three.³² To our knowledge, no studies have investigated the mental health impacts of farmers’ relationships with organizations.^33,34 This study fills a critical gap in the social connectedness literature.

Female farmers reported higher stress levels than male farmers, which is consistent with the limited mental health literature on this population.³⁵ The underrepresentation of female farmers in studies hinders our understanding of contextual factors impacting their wellbeing. Nonetheless, the additional roles (e.g., a caregiver, a childcare provider, and a health insurance policyholder through off-farm employment) they undertake and the lack of recognition as farmers³⁶ may worsen their mental health more than that of male farmers.

The cross-sectional study design limited our ability to determine causality between program participation and mental health. The findings however suggest that future prospective studies to establish causality are warranted. Using interviewers to administer the questionnaire may have led participants to respond to questions in ways they viewed were socially desirable. This bias may underestimate stress and depression symptom scores,³⁷ inflating the effects of program use, program satisfaction, and social support on outcomes. However, to minimize the likelihood of low statistical power from a small sample size, we decided to offer online and phone survey options to encourage study participation. We examined the effects of online and phone participation on outcomes and observed statistically nonsignificant (p ≥ 0.05) results. In addition, participants with negative affectivity may have had overstated stress and depression, program dissatisfaction, and social support.³⁸ This may have been a problem because the questionnaire was administered during the COVID-19 pandemic, but we do not expect that negative affectivity biased findings. The COVID-19 pandemic has had an impact on mental health across populations.³⁹ Nondifferential bias was possible as the pandemic may have resulted in participants increasing their reporting of exposure (program use) regardless of the outcome status (stress, depressive symptoms). Finally, our sample included farmers who participated in different co-ops and organizations, which likely introduced heterogeneity in program use and satisfaction and therefore the ability to interpret the association of programs on mental health. Future studies comparing programs within the same organization could reduce this heterogeneity.

This study makes a convincing argument for investigating the effects of social connectedness on farmers’ mental health. Organizational participation overlaps the social-ecological model that has been used to examine multilevel mental health determinants. At the organizational level, co-ops allow farmers to collectively negotiate for markets and services.^{15, 40} Through the cooperative process, farmers maintain their independence while developing a sense of belonging and control. Researchers refer to this phenomenon as social identity framing (e.g., we the co-op owners) whereby members perceive themselves as part of a group that shares values and beliefs rather than an isolated self.^{33, 34, 41} The self-concept and practices of aligning individual attitudes and behaviors with the group are powerful predictors of health. The material and psychological resources provided by organizations to address farmers’ economic and social needs should be emphasized in public health research.

From a policy perspective, supporting farmer organizations is needed to ensure equitable production and trading environments for farmers. Co-ops have encountered ongoing consolidation threats.⁴² Industry consolidations have generated concerns among small-scale (e.g., 70 cows vs the national average 175) co-op members regarding viability of their co-ops, farms, and their social identity. Policy support to strengthen the development of farmer organizations (co-ops and non-co-ops) can protect infrastructures that are critical to farmers’ livelihoods.

In public health practices, partnering with community-led organizations in the development and implementation of mental health programs may be cost-effective and culturally relevant. Farmers face rural mental health resource shortages and stigma to seek help.⁴³ This study suggests that an alternative intervention through organizations such as co-ops that address upstream risks are likely to resonate with and are welcomed by farmers.