Inequities in Social Media Platform Access and Use for Migrant and Seasonal Farmworkers,North Carolina,2023: Implications for Health Promotion

Abstract

Migrant and seasonal farmworkers (“farmworkers”) are a unique population with structural barriers to being reached in health promotion efforts. Social media platforms and communication campaigns are increasingly used in health promotion efforts with farmworkers. However, digital exclusion, sometimes referred to as the digital divide, can limit the reach of social media interventions and campaigns. It is critically important for program planners to know the potential reach of different communications channels. We conducted a secondary analysis of a time-venue convenience sample survey of farmworkers that was fielded in North Carolina (NC) in May–December 2023 at housing and clinic locations. We assessed social media use for seven platforms (Facebook, GroupMe, Instagram, Snapchat, TikTok, WhatsApp, YouTube), randomizing each participant with internet access to be asked about one platform. We calculated what percentage of farmworkers in NC would likely be missed via each platform based on use of the platform and inconsistent or no internet access. Several social media platforms (WhatsApp, Facebook, YouTube) were widely used by farmworkers in NC. However, use and access were not universal, with even the most used platforms missing approximately 1 in 10 farmworkers who had consistent internet in their housing locations. Considering use of platforms in combination with having no or inconsistent internet access in housing, platforms missed a minimum of approximately 1 in 3 farmworkers. Campaigns and interventions should consider using multiple social media platforms and other channels to maximize reach. Future research is needed to disentangle positive and negative effects of using social media to reach farmworkers.

In North Carolina (NC), there are approximately 80,000 migrant and seasonal farmworkers (“farmworkers”) who work in the state’s agricultural economy each agricultural season (Lambar & Thomas, 2019). The NC Farmworker Health Program has reported that 95% of farmworkers in NC were Latine, 75% identified as male, and 87% had incomes below the federal poverty line (Lambar & Thomas, 2019). Farmworkers are faced with many hazards linked with agricultural work, including heat and pesticide exposure, respiratory illnesses, and repetitive stress injuries (Lambar & Thomas, 2019). In addition, issues like poor housing quality, language barriers, limited political power, and lack of transportation hinder efforts to promote farmworker health (Lambar & Thomas, 2019; Wiggins, 2020). Farmworkers in NC frequently live in isolated, employer–provided housing (Summers et al., 2015), and outreach to this population has often relied on community health workers and the connections they form (Harwell et al., 2022).

The COVID-19 pandemic showed the importance of digital accessibility in outreach and dissemination of health information. For example, a report by Ramos and colleagues documented the use of the social media platform Facebook to provide health information to cattle feed yard workers during the COVID-19 pandemic (Ramos et al., 2020). Community health workers in NC and other states have reported benefits from use of social media for communicating with farmworkers (LePrevost et al., 2024; Ramos et al., 2020; Saldanha, 2023). More broadly, social media platforms can be part of effective behavior change interventions (Vereen et al., 2023), but it is important to consider the equity of their reach (Welch et al., 2016). Other research has shown differences in use of social media platforms in ways that can impact the equitable reach of health promotion interventions, such as for lesbian, gay, bisexual, and transgender people (Seidenberg et al., 2017). However, a mapping review and more recent literature searches have not revealed any research examining the reach of social media platforms to farmworkers in NC or nationally (Bloss et al., 2022). Without knowing how many farmworkers use social media and the type of social media platforms they use, social media interventions, communication campaigns, and outreach efforts may miss many of the people program planners are trying to reach.

Theoretical Framework

Our research was informed by the knowledge gap hypothesis (Beacom & Newman, 2010; Lind & Boomgaarden, 2019), which posits, in part, that social and economic resources facilitate access to and use of information, and The inequality paradox (Frohlich & Potvin, 2008), which suggests that interventions with differential reach or impact can widen disparities while also improving population health. These frameworks suggest that communication and health promotion interventions can be effective from the perspective of the overall population yet can simultaneously leave behind people with fewer resources. Implementation is often patterned by available resources or social power (Woodward et al., 2021). Attention is thus needed to factors that can make implementation of health promotion programs more equitable. These theoretical frameworks push health promotion programs to think about where interventions may fail to fully reach a population of interest and who cannot access an intervention. Additional investment (i.e., greater dose) of campaigns and alternative strategies may be warranted when interventions fail to be equally effective or fail to reach all members of the population. We drew on these frameworks to inform this research project looking at the potential reach of social media interventions to farmworkers in NC.

Method

The purpose of this study was to assess what percentage of farmworkers would be missed in health promotion interventions using social media platforms. We used data from a statewide survey of farmworkers conducted in NC. Five of the authors (JGLL, ECR, CEL, MR, and LEC) were involved in the implementation of this statewide survey. The full methods of the survey are published elsewhere (Lee et al., 2024), and the codebook is available in our public repository (Bloss et al., 2024). Briefly, there is not a reliable sampling frame of farmworkers in NC, and we used a time-venue sampling approach where the time was after work hours and the venue was farmworker housing or farmworker-specific health clinics. Most farmworker housing in NC is provided by the employer in isolated rural areas (Summers et al., 2015), and community health workers with local knowledge conduct outreach statewide to farmworker housing across the growing season (Lambar & Thomas, 2019). Community health workers visit farmworker housing locations to conduct health assessments, connect farmworkers to health services, and provide health education.

Data were collected from May 31, 2023, to December 27, 2023, by 19 trained bilingual community health workers and one research assistant. This time period reflects the agricultural season in NC—with early crops like blueberries and late crops like Christmas trees. Community health workers represented eight different farmworker health and social services organizations across the state, and data were collected in 39 of NC’s 100 counties. The survey design included input on questions from community health workers and NC Farmworker Health Program staff. The survey was professionally translated and piloted in Spanish with five participants.

Survey Fielding

A trained and institutional review board (IRB)-registered community health worker verbally administered the survey in English or Spanish at farmworker housing locations to up to five workers per housing site. Housing locations were selected based on the existing outreach schedules of farmworker-serving community health workers. At housing sites for families, community health workers interviewed one family member. The research assistant and community health workers also administered surveys at farmworker-serving clinics while people were waiting for services. Data collectors provided an incentive for participation (i.e., a soccer ball, first aid kit, or laundry detergent) to each participating housing site or each clinic participant. Data collectors approached potential study participants while they were waiting for services or after completing a health assessment, provided a study information sheet, answered any questions, and obtained verbal consent. To preserve the anonymity of participants, signed consent was not required. Data collectors conducted surveys in a private area and away from other people. For example, farmworker health services can take place next to farmworker housing under a portable canopy; in these cases, data collectors conducted the survey in a quiet area away from other people waiting for or receiving services to preserve privacy. The East Carolina University and Medical Center IRB approved the research protocol under an exempt determination (#22-001454).

Measures

For the main outcome of social media use, the survey programming randomized each participant who reported having internet access or who reported sometimes having internet access at their housing to one of seven social media platforms using the format, “Do you use Facebook?” The platforms were Facebook, GroupMe, Instagram, Snapchat, TikTok, WhatsApp, and YouTube. We selected these platforms based on input from community health workers regarding what platforms they expected farmworkers to be using. We also included GroupMe, which we expected to have low usage, as a contrast to WhatsApp. Randomization was conducted by the item randomization feature of Qualtrics. We assessed internet access, which is reported elsewhere (Lee et al., 2024), into three categories: having consistent internet access, having inconsistent internet access, and having no internet access. These classifications were based on the question: “Are you able to use the internet while living here in this housing through Wi-Fi or your cell phone?” Response options were yes, sometimes, no, and don’t know. Respondents answering don’t know were considered to not have internet access in these analyses. Full details on demographic measures are reported in the main effects paper on internet access (Lee et al., 2024). Briefly, we assessed if participants were more than 18 (participants under 18 were not eligible) and traveled to work in agriculture (i.e., a migrant farmworker) or lived in one place and worked seasonally (i.e., a seasonal farmworker). We also assessed gender, language, and country of birth.

Analysis

We used SPSS v.29/PC to examine frequencies. Specifically, we first calculated what percentage of farmworkers in our survey would be missed by an intervention based on their not using the social media platform and not having consistent internet access. Second, we calculated what percentage of farmworkers would be missed by an intervention based on not using the social media platform and not having any internet access. Third, among participants with consistent or inconsistent internet access, we calculated the percentage of participants who did not use each platform by gender and migrant versus seasonal status. We treated any missing data as missing at random, and we used pairwise deletion. Because of the question randomization in this time-and-venue-based convenience sample, we did not use weights and thus have not reported confidence intervals in these analyses.

Results

After removing three participants due to missing data on key outcomes, there were 1,031 participants in the survey. Of these participants, 14% were women and 86% were men. Regarding racial/ethnic identities, of which participants could report more than one, the majority reported a Hispanic/Latine (99%) identity. Participants also reported Indigenous (1%), Black or African American (0.30%), White (0.20%), or other identities (0.40%). Almost all participants were born in Mexico (91%), followed by Guatemala (4%), Honduras (2%), and the United States (1%). Regarding language, 98% currently spoke Spanish and 7% currently spoke English.

Participants commonly reported using WhatsApp, Facebook, and YouTube. In Table 1, we report results in two ways: (1) what percentage of farmworkers would be missed because they do not use a platform or because they lack consistent internet access and (2) what percentage of farmworkers would be missed because they do not use the platform or because they lack any internet access at their housing location. Results showed that the most popular platform—WhatsApp—would not be consistently available to approximately one in every three farmworkers and would not be available at all to approximately one in 10 farmworkers. Similarly, the use of Facebook would miss approximately 38% and YouTube would miss approximately 39% of farmworkers due to their not having an account or having lack of consistent internet access.

Table 1

Percentage of Farmworkers Who Would Be Missed by Potential Social Media Platform Interventions Because of Not Using the Platform by Level of Internet Access in Housing, North Carolina, 2023 Growing Season, n = 1,031

Social media site	% Missed due to nonuse of platforms or lack of consistent internet access in housing	% Missed due to nonuse of platforms or lack of any internet access in housing
WhatsApp	33.8	9.7
Facebook	37.7	13.7
YouTube	38.8	16.1
TikTok	55.8	41.2
Instagram	81.1	73.3
Snapchat	91.2	88.7
GroupMe	99.4	93.3

Note. Due to sporadic missingness and rounding, percentages do not exactly match when crossed with other variables.

We also examined patterns of social media use among participants with internet access by gender and by migrant versus seasonal status. Figure 1 shows the percentage of participants with any internet access who would be missed because they do not use social media platforms. The results showed some variability by gender, which may be due to gendered differences in migrant versus seasonal status. Migrant and seasonal farmworkers with internet access reported comparable use of social media sites.

Figure 1

Percentage of Farmworkers With Any Internet Access Who Do Not Use Social Media Platforms by Gender (A) and Migrant Versus Seasonal Status (B), North Carolina, 2023 Growing Season

Discussion

Principal Findings

In this large, statewide survey of farmworkers in NC, we identified gaps in the reach of social media platforms for communication campaigns, interventions, and outreach involving farmworkers. Specifically, even the most popular social media platform would miss approximately one in three farmworkers—taking into account both use of platform and the number of farmworkers not having consistent or any access to the internet. Our findings confirm a long-standing concern documented in the literature about the reach of social media interventions (Welch et al., 2016) and serve as a reminder of the importance of using multiple channels and, for farmworkers in particular, personal outreach (Lambar & Thomas, 2019).

Our findings are likely the result of the unique situation present in the agricultural system. First, in NC, housing is often provided by the farm owner, is located in rural isolated areas, and may not be occupied year-round (Summers et al., 2015). Second, while studies show that a growing proportion of the U.S. public has access to broadband internet, access is not distributed equally, particularly among marginalized populations such as farmworkers (Early & Hernandez, 2021). Indeed, approximately a quarter of farmworkers in NC do not have consistent or, in some cases, any access to the internet in their housing locations (Lee et al., 2024). Third, protections, pay, and power are limited by an “agricultural exceptionalism” that treats farm work differently from many other types of work in the United States and has stymied efforts to change conditions (Wiggins, 2020).

Understanding this context for our findings is important as multiple studies report the use of WhatsApp and other social media platforms to conduct outreach to farmworkers (LePrevost et al., 2024; Ramos et al., 2020; Saldanha, 2023). Researchers are also testing social media platforms to promote HIV prevention among seasonal farmworkers (Rodriguez et al., 2024). In older studies, farmworkers have reported a willingness to receive health messages via text messages (Price et al., 2013; Sandberg et al., 2016). Social media interventions and outreach are important and likely effective strategies given that farmworkers live in rural areas and have limited access to transportation. However, this study suggests it is likely important to leverage social media channels as well as other lower-tech channels. Our findings support research showing that adding additional channels for interventions beyond social media may increase reach and, ultimately, intervention impact (Vereen et al., 2023).

Although we cannot directly compare our results with findings among the general public, our study problematizes research showing no inequities in social media access by race or class in national surveys (Chou et al., 2021; Huo et al., 2019), suggesting it may be important to consider occupation or migratory status. Researchers should continue to examine social media interventions for equitable reach across populations.

Implications for Practice

In the broader context of the literature and following our theoretical framework, our findings show that it is critical to recognize the limits of social media reach and the role of digital inequities in limiting access to social media platforms. Interventionists, health educators, community health workers, and health communication professionals should consider the reach of different social media platforms when planning campaigns. Posting clinic updates to Facebook, for example, will not reach all farmworkers. Using a WhatsApp-based group would reach more farmworkers, but it would still miss a meaningful proportion of farmworkers who do not use WhatsApp or lack consistent internet access. Practitioners should consider who is being missed when using one or more social media platforms and how these missing farmworkers might be reached (e.g., through outreach visits to housing units where they know the internet is unreliable). Additional channels or approaches may be needed, such as radio (Young et al., 2019), in-person outreach, or other strategies.

Public health practitioners should consider a nuanced view of the effects of social media platforms in health promotion. While our focus was on the reach of benefits from social media interventions, it is also important to consider negative, unintended consequences of social media use (Jahnel et al., 2022). Although we did not examine misinformation or mistrust in social media, it is vital to note that social media can also spread misinformation or be used to socially engineer phishing or other fraud. Thus, future interventions should be attentive to the risks of social media and its reach.

Implications for Research

Beyond exposure to misinformation in social media, there are other unintended effects that may cause harm such as exposure to norms promoting body dysmorphia or exposure to tobacco product use. Reach within social media platforms may also be disrupted by social media algorithms that are written in ways that discriminate against individuals based on race, gender, or economic status (Terrasse et al., 2019). Unintended negative effects of social media use must be weighed against the many potential benefits (Terrasse et al., 2019), such as the ability of farmworkers to make connections with advocacy groups. Future research should work to disentangle the positive and negative consequences of social media reach to farmworkers. Future research should also determine how to maximize reach across social media and more traditional channels. For example, what is an optimal mix of platforms and communication channels to reach farmworkers for a particular occupational safety campaign?

Strengths and Limitations

There are important strengths to this study including its size, use of time-venue sampling, and support from community health workers in the planning process. Indeed, the idea to assess social media use was suggested by community health workers at a NC Farmworker Health Program training. However, there are also limitations. We were not able to use probability sampling in this study, which means the findings may not generalize across NC or to other states. Our estimates of reach are also dependent on the sampling used. That said, without a sampling frame of farmworkers, time-venue sampling represents a best practice for reaching a population that is isolated. We did not assess every possible social media platform, and, due to needed efficiencies in survey administration, we asked each person with internet access about a single randomly selected social media platform. Because of this, we cannot assess the reach of campaigns that leverage multiple social media platforms. Other qualitative approaches could better help understand how social media platforms are used, preferences for interventions and outreach, and potential harms. Finally, innate to this topic area, our data are from the 2023 growing season and changes in social media platform landscapes are likely to continue to be rapid.

Conclusion

Social media interventions represent an important strategy for health outreach to farmworkers, but there are important limitations to their reach because of digital exclusion. To be clear, we are not calling for an abandonment of social media interventions; rather, we hope this research highlights how, following our theoretical framework, interventions using only social media may widen inequities between those with and without access. Health promotion professionals should assess reach of social media platforms and consider additional channels for interventions, communication, and outreach to promote equity.