Developing messages to encourage healthy, sustainable dietary substitutions: A qualitative study with US emerging adults

Abstract

Prior research shows that adopting simple dietary substitutions (e.g., replacing beef with poultry or plant-based entrees) can improve dietary quality and reduce the negative environmental consequences of food production, but little is known about how to encourage people to adopt these substitutions. This study aimed to examine reactions to messages encouraging healthy, sustainable dietary substitutions among emerging adults ages 18–25. We conducted four online focus groups with a diverse sample of US emerging adults (n=28; 61% female). Focus groups explored emerging adults’ reactions to messages encouraging them to adopt three target dietary substitutions: replacing beef and pork with poultry and plant-based entrees; replacing juice with whole fruit; and replacing dairy milk with non-dairy milk. We transcribed discussions verbatim and adopted a thematic approach to analyzing the transcripts. Results showed that participants perceived messages to be most effective at encouraging the target dietary substitutions when the messages: encouraged specific, achievable dietary changes; linked these dietary changes to clear consequences; included personally relevant content; included statistics; were succinct; and used a positive tone. Across the target dietary substitutions, two message topics (small changes, big benefits, which emphasized how small dietary changes can have large positive health and environmental impacts, and warning, which discussed the negative health and environmental impacts of dietary choices) were generally perceived to be most effective. A few participants expressed doubt that the target dietary substitutions would have meaningful environmental impacts. Results suggest that campaign messages to encourage healthy, sustainable dietary substitutions may be more effective if the messages make the target dietary substitutions seem achievable and use statistics to clearly describe the positive impacts of making these changes.

1. Introduction

Every year, more than 500,000 Americans die from diseases caused in part by unhealthy diet (Mokdad et al., 2018). At the same time, up to one-third of all anthropogenic greenhouse gas emissions are caused by food production, making food production a major contributor to climate change (Crippa et al., 2021). Encouraging people to adopt healthier and more sustainable diets is therefore an important priority for addressing diet-related diseases and reducing the negative environmental consequences of food production (Willett et al., 2019).

Eating a diet high in fruits, vegetables, and legumes and low in red and processed meat reduces both risk for diet-related diseases and the negative environmental consequences of food production (Clark et al., 2019; Conrad et al., 2023; M. Nelson et al., 2016; O’Malley et al., 2023; Pollock et al., 2022; Poore & Nemecek, 2018; Rose et al., 2019; Scarborough et al., 2023; Van Horn et al., 2016; Willett et al., 2019; Willits-Smith et al., 2020). Adopting new dietary patterns, however, is challenging (Hill et al., 2003; Roberto et al., 2015). One promising strategy for overcoming this challenge is to encourage people to make simple dietary substitutions in which they replace just one food with another, without making other changes to their diets. This “small changes” approach may be easier for consumers to implement than more substantial changes to the diet, such as becoming vegetarian (Hill, 2009; Hill et al., 2003; Hills et al., 2013; Stroebele et al., 2009; Wehbe et al., 2022). Moreover, encouraging simple dietary substitutions could appeal to a wider audience (including those uninterested in making substantial changes) and could boost self-efficacy by helping people to make gradual, achievable changes (Bandura, 1998; Hill, 2009; Hill et al., 2003).

Prior research has identified three simple dietary substitutions that would improve health and reduce the negative environmental consequences of food production. Specifically, studies show that replacing beef with poultry or plant-based proteins in mixed dishes and entrees, replacing juice with whole fruit, and replacing dairy milk with soy or almond milk would improve individuals’ dietary quality by 3–10% and (if universally adopted) reduce the national dietary carbon footprint by 2–37% (Grummon et al., 2023; Rose et al., 2022). However, it is unclear how to encourage people to adopt these dietary substitutions. Prior research indicates that messages discussing health or environmental aspects of food choices are a promising strategy for prompting dietary change (Dijkstra & Rotelli, 2022; Grummon et al., 2022; Herchenroeder et al., 2022; Silva Souza & O’Dwyer, 2022; Taillie et al., 2021, 2022; Wolstenholme et al., 2020), but it remains unknown what topics such messages should address. Messages could, for example, warn consumers about the negative health and environmental consequences of food choices (Grummon et al., 2021; Taillie et al., 2021, 2022), or alternatively, emphasize descriptive social norms about others making these dietary changes (Salmivaara et al., 2021; Sparkman et al., 2020; Sparkman & Walton, 2017). Identifying the most promising topics for messages about dietary substitutions is important because even small improvements in message effectiveness can yield meaningful improvements in health outcomes when multiplied across a population (Grummon et al., 2019).

Although most Americans fall short of dietary recommendations (Liu et al., 2020, 2021; Rehm et al., 2016), emerging adults (ages 18–25 years) are an especially important group to target with interventions to encourage healthy, sustainable dietary substitutions for three reasons. First, emerging adults have lower dietary quality than other adults (Krebs-Smith et al., 2010; M. C. Nelson et al., 2008; Rehm et al., 2016). Second, emerging adulthood is a distinct developmental period during which people gain substantial autonomy and form new identities (Arnett, 2018). Implementing interventions among emerging adults therefore has disproportionate potential to shape behavioral and health trajectories in lasting ways (Arnett, 2007, 2018; M. C. Nelson et al., 2008). Indeed, longitudinal cohort studies find diet quality in emerging adulthood predicts both diet quality and likelihood of weight gain, obesity, and type 2 diabetes in middle and older adulthood (Christoph et al., 2019; Dunn et al., 2000; Murthy et al., 2016). Third, emerging adults have a strong interest in both health and environmental sustainability (YouGov, 2020), suggesting that messages that educate or remind emerging adults about the links between their dietary choices and health and environmental consequences could be a powerful way to encourage them to change their diets (Macdiarmid et al., 2016; McInnes et al., 2023).

The objective of this study was to examine emerging adults’ perceptions of messages encouraging healthy, sustainable dietary substitutions and to identify message topics that most resonate with them. Although prior research has examined attitudes, beliefs, and experiences with adopting healthy, sustainable dietary choices (de Boer et al., 2017; Faber et al., 2020; Ford et al., 2023; McInnes et al., 2023; Mollaei et al., 2022, 2023; Pelletier et al., 2013; Pieniak et al., 2016; Poobalan et al., 2014; Slotnick et al., 2023; Whittall et al., 2023; Williams et al., 2023), limited research has examined emerging adults’ reactions to messages about specific dietary substitutions. We therefore adopted a qualitative approach to allow in-depth exploration.

2. Methods

2.1. Participants and Recruitment

In May 2023, we recruited participants for online focus groups. We used purposive sampling strategies to obtain diversity in terms of gender and current college enrollment. Eligible participants were emerging adults (ages 18–25) residing in the US who were able to read and speak English and who had access to a computer with Internet. We recruited participants through social media advertisements on Facebook and Instagram and through ResearchMatch, a national registry that connects people interested in research studies with researchers. Recruitment materials described the study in general terms, indicating that participants would share their thoughts on food choices in a group discussion (Supplemental Figure 1 shows a sample recruitment ad).

Potential participants completed an eligibility screener programmed in Qualtrics. A member of the research team contacted those who were eligible per the screener to complete a brief call on the videoconferencing platform Zoom (Woolfall, 2023). During the call, the research team confirmed participants’ eligibility (i.e., that they were age 18–25, lived in the US, spoke English, and had a computer with Internet) and obtained verbal informed consent. Participants who were deemed eligible and consented were scheduled for one of the focus group sessions. We aimed to conduct four focus groups with 5–9 participants per group (20–36 people overall). After conducting the four focus groups, we determined that we had reached saturation (i.e., no new themes or perspectives related to our research questions emerged) and did not conduct additional focus groups. We did not conduct follow-up discussions. Supplemental Figure 2 shows the flow of participants through the study, including the number excluded at each stage of recruitment.

2.2. Procedures

Focus group sessions took place on Zoom and lasted approximately 90 minutes. We chose to conduct focus groups online (vs. in-person) to enable national recruitment. The sessions were facilitated by a three-person team, including a moderator, a note taker, and a technical support person. The moderator (AHG) was a female assistant professor with a PhD in health behavior and prior experience conducting qualitative focus groups (Grummon et al., 2014). The moderator did not have contact with participants prior to the focus groups. During the discussion, the participants, moderator, note taker, and technical support person kept their Zoom videos on and displayed their names (or, if they preferred, a pseudonym). At the conclusion of the discussions, participants provided demographic information via an anonymous online survey programmed in Qualtrics. Participants received a $50 Amazon gift card for completing the focus group. We recoded audio and video for all discussions. The Stanford University Institutional Review Board approved the study procedures (IRB #67669).

2.3. Measures

The moderator used a semi-structured interview guide to facilitate the discussions, using probes as needed to clarify or elicit additional details (Supplemental Table 1). We developed the interview guide based on prior studies examining responses to health messages and dietary choices (Falbe et al., 2021; Hoek et al., 2017; Moracco et al., 2016; Wiseman et al., 2016). Prior to conducting the discussions, we pilot tested a draft guide with seven participants similar to our target sample. We made minor modifications to the guide based on the pilot test. For example, in the pilot, we showed participants messages for only one of the three dietary substitutions. In the main focus groups, we revised this to show messages for all three dietary substitutions (though the discussions remained centered on just one substitution, as described below). Results from the pilot focus group were not included in analyses.

The guide had four sections. First, as warm-up questions (Vaughn et al., 1996), the moderator elicited a general discussion of emerging adults’ food choices by asking participants to describe the last meal or snack they ate and why they chose that food. Second, the moderator asked participants to develop and deliver their own messages aimed at convincing one another to make the target dietary substitutions. The moderator then asked participants to vote using the Zoom poll function for the message that most convinced them to want to make the target dietary substitutions (i.e., perceived message effectiveness). We focused on querying perceived message effectiveness because it is predictive of actual message effectiveness (Baig et al., 2019, 2021; Bigsby et al., 2013; Noar et al., 2018). After voting, the moderator asked participants about their votes and their perceptions of and reactions to the messages.

Third, the moderator showed participants seven messages (presented in random order) that the research team had developed to encourage the target dietary substitutions. The moderator asked participants to imagine that the messages were part of a campaign targeted to people their age (e.g., that messages were displayed as a post in their social media feed or a pop-up ad that appeared when shopping for food). The moderator first showed the messages one at a time for approximately 20 seconds each, then showed all seven messages side-by-side. Each of the seven messages adopted a different topic (e.g., small changes, big benefits; warning; progress over perfection), selected to create maximal divergence between topics to avoid redundancy. Sample messages encouraging participants to eat chicken or vegetarian dishes instead of beef or pork are shown in Figure 1 and all messages are shown in Supplemental Figure 3. All seven messages included references to both health and environmental sustainability. The moderator asked participants to vote using the Zoom poll function for the messages that most and least convinced them to make the target dietary substitution and to discuss their votes and other reactions to the messages. The moderator also solicited participants’ advice for improving the messages. To reduce participant burden, each of the four discussions focused primarily on messages about one of the three target dietary substitutions, with a briefer discussion of messages about the other two target dietary substitutions. Fourth, the moderator concluded the discussion with a closing question (Vaughn et al., 1996).

Figure 1.

Sample messages shown in focus groups.

Note. Figure shows messages about the beef and pork substitution; participants also viewed similar messages about juice and dairy milk substitutions as shown in Supplemental Figure 3. Theme names were added to the figure for reference and were not shown to participants.

2.4. Analysis

We transcribed the audio recordings verbatim and imported the transcripts into Dedoose qualitative data analysis software version 9.01.107. The data analysis team comprised three coders, each of whom had participated in the focus groups as moderator, note taker, or technical support. We adopted a thematic data analysis approach to analyze the transcripts (Green & Thorogood, 2018). First, we immersed ourselves in the data by reading all transcripts in depth and writing analytic and reflective memos with our reactions to the transcripts. Second, we coded the data. Coding was iterative. We developed an initial codebook of “start list” codes based on the conceptual framework and research questions (Miles et al., 2019). We then conducted an initial coding of the four transcripts in which we applied the start list codes and developed new codes using an inductive approach. We chose to code all four transcripts with the start list to ensure the codebook could be used with the range of responses seen across the discussions. We revised the codebook, adding, removing, combining, and organizing codes based on the initial coding and discussions among the team. Using the revised codebook, two analysts independently coded each transcript (i.e., each transcript was coded twice more). The two analysts met to resolve any discrepancies, with remaining discrepancies resolved by a third analyst. Third, once coding was complete, the research team used the coded transcripts to create analytic products including matrices and conceptual models (Miles et al., 2019). Fourth, the research team discussed the findings and used the coded transcripts and analytic products to identify major themes across the four focus group discussions. The research team defined and clarified each theme by writing analytic memos. We did not discuss the transcripts or analysis with participants.

3. Results

Of the 28 participants in the four focus groups, 6 (21%) were aged 18–21 years, 10 (36%) were aged 22–23, and 12 (43%) were 24–25 (Table 1). A majority (n=17, 61%) identified as female, 10 (36%) identified as male, and 1 (4%) identified as non-binary or another gender. The sample had a relatively diverse racial/ethnic composition: 3 (11%) identified as Latino(a) or Hispanic (irrespective of race), while 9 (32%) identified as Asian, 2 (7%) identified as Black or African American, 15 (54%) identified as white, and 2 (7%) identified as another race or multi-racial. About half (n=15, 54%) were currently enrolled in college, and nearly all (n=27, 96%) had completed at least some college.

Table 1.

Characteristics of focus group participants, n=28

Characteristic	%	N
Age, years
18–21	21%	6
22–23	36%	10
24–25	43%	12
Gender
Female	61%	17
Male	36%	10
Non-binary or another gender	4%	1
Latino(a) or Hispanic	11%	3
Race
Asian	32%	9
Black or African American	7%	2
White	54%	15
Other or Multiracial	7%	2
Highest level of education completed to date
High school diploma	4%	1
Some college or technical school	18%	5
College graduate	61%	17
Graduate or professional degree	18%	5
Currently enrolled in college or university	54%	15
Household income, annual
$0 to $24,999	18%	5
$25,000 to $49,999	14%	4
$50,000 to $100,000	25%	7
$100,000 to $200,000	32%	9
$200,000 or more	11%	3
Financial situation (self-described)
Doesn’t meet basic expenses	0%	0
Just meets basic expenses	18%	5
Meets needs with a little lift	32%	9
Lives comfortably	50%	14
Household size
1	43%	12
2–3	29%	8
4 or more	29%	8
Diet quality (self-described)
Poor	0%	0
Fair	32%	9
Good	32%	9
Very Good	29%	8
Excellent	7%	2
Vegetarian or Vegan	14%	4

3.1. Themes

Six themes emerged from participants’ reactions to the messages encouraging specific dietary substitutions: five themes about the characteristics of effective messages (depicted in Table 2 alongside example quotes) and one theme about the potential impacts of personal dietary choices. Across the focus groups, the small changes, big benefits and warning messages were especially well-received by participants, eliciting 21 of 27 votes for the most effective message (1 participant did not vote).

Table 2.

Message characteristics that enhanced perceived message effectiveness

Characteristic	Example quotes
Encourages a specific, achievable dietary change	“I think something that, like, the ones that I liked all had in common was, like, a manageable action. Like even for [the progress over perfection message] … it says, like, ‘Just swapping chicken for beef or pork once a day.’”
Links the specific dietary change to clear consequences	“I like that they’re actually tying the statistics to specific things like heart disease or diabetes.it seems, I don’t know, like, it’s just better, I guess, to, like, know exactly what you’re trying to fix when you’re making this replacement.”
Discusses behaviors, consequences, or topics that the audience finds personally relevant	“I voted for the first pitch [as the most effective] because I felt like, like [another participant] said, there’s good information in there. I think, I think the main argument was that it was healthier and also costs less, which I think makes sense and is pretty relevant, I think, to my life and also a lot of others.” “Yeah, I, so, I picked the third [message] [as the most effective] just because I think it was the most relatable to me. Uh, even though the first one had a lot of information and that kind of stuff, the third one was something that, uh, related to me a lot because I drink a dairy milk in a daily basis and just stopping, like, to think about that, it just grabbed my attention.”
Includes a statistic, especially if the statistic is perceived as large	“I went with [the warning message] [as the most effective message]. I think that one works out really well because they use, like, actual statistics and stuff to convince someone. Not, just kinda, like, taking someone else’s opinion for it.” “I think for me the numbers just make it more realistic. [The small changes, big benefits and warning messages] are the two [messages] that have, like, the most numbers in them, and what the other ones are more like, ‘they’re possible, this will help.’ But this is like, this is concrete, and that’s what would convince me.” “I personally think [including statistics] is helpful. Like I mentioned earlier, I was pretty surprised by how high those numbers were.”
Is short	“I think [another participant] also alluded to earlier that being really succinct is important. I think oftentimes messaging can be caught up in a lot of words that people may not have access to or jargon or there just might be a lot of words making the message really overwhelming. So, I think all of these messages should, like, be evaluated for how effectively they use language and if it’s efficient and to the point.” “I went with [the small changes, big benefits message] [as the most effective message] just because it’s shorter and it’s more to the point.”
Uses a positive tone	“I think one aspect I liked about [the small changes, big benefits message] as well was a lot of the kind of dialogue around it was more positive. Like, if you take this step, you’re going to see these benefits both to the environment and your health.”

3.1.1. Theme 1: Messages were perceived as more effective when they encouraged specific, achievable dietary changes.

Participants found messages to be more effective when the message encouraged a specific (i.e., well-defined) and achievable dietary change (Table 2). For example, many participants indicated they found the small changes, big benefits and progress over perfection messages effective because these messages encouraged replacing “just one serving per day” of the target food with the substitute food or making the target dietary substitution “just one day per week.” Participants indicated liking that these messages provided specific information about how frequently they should make the target dietary substitutions. For example, a participant in focus group 1 said:

I think something that, like, the ones that I liked all had in common was, like, a manageable action. Like even for [the progress over perfection message] …it says, like, ‘Just swapping chicken for beef or pork once a day.’ And then if you compare it to [the future generations message] like it says, ‘Start eating chicken,’ but it doesn’t really say like how much do I have to change to make this like effort worth it?

Participants also indicated they found the small changes, big benefits and progress over perfection messages to be effective because these messages emphasized making an incremental change (i.e., “just one” language), which made the target dietary changes seem “manageable,” “reasonable,” and “achievable” to participants. Participants contrasted these achievable dietary changes with more drastic dietary changes that would require completely avoiding a food or food group, despite none of the messages encouraging such changes. For example, a participant in focus group 2 said:

I know lots of people would be like, I don’t wanna, you know, cut off, like – or cut out beef or pork entirely from my diet. And so that recommendation of swapping chicken for beef just, like, one day per week, I think is a really reasonable and, like, achievable goal for a lot of people.

3.1.2. Theme 2: Participants found messages to be more effective when the message linked the specific dietary change to clear consequences.

Participants indicated they found messages to be more effective at encouraging the target dietary substitution when the message indicated that making (or failing to make) that substitution would lead to clear beneficial (or negative) consequences (Table 2). For example, a participant in focus group 3 said she found the small changes, big benefits message and the warning message to be effective because these messages linked the dietary changes to consequences that she found clear: “I like that they’re actually tying the statistics to specific things like heart disease or diabetes…it seems, I don’t know, like, it’s just better, I guess, to, like, know exactly what you’re trying to fix when you’re making this replacement.” By contrast, participants found messages to be less effective when the messages described a consequence that was not clear and understandable to them. For example, one participant indicated she found the future generations message to be the least effective message because:

I’m very data-driven and just saying ‘greener, healthier’… I don’t know how I’m being more green. I don’t know what part of my health I’m improving. It all sounds great, but I don’t actually know what kind of an impact I’m making.

Similarly, several participants indicated messages that described how the substitutions would affect dietary quality were ineffective because participants found the term “dietary quality” unclear. One participant in focus group 4, for example, said:

In [the warning message], the second number has changed to ‘leads to 5% worse dietary quality.’ I don’t think that’s very tangible. I, I don’t know what exactly is meant by that. I can understand the general sentiment. But I certainly don’t know how they’re defining dietary quality.

3.1.3. Theme 3: Participants found messages to be more effective when the message content was more personally relevant.

Participants indicated they found messages to be more effective when the message discussed behaviors, consequences, or statistics that participants found personally relevant (Table 2). A participant in focus group 2, for example, said she found one message particularly effective because she was already considering making the target dietary substitution:

I voted for the [message about replacing juice with fruit] [as the most effective message] because I am currently in the process of trying to reduce, like, juices or sugar-added drinks and try to change that into fruit, specifically blueberry or oranges, so I kinda resonated with it.

Similarly, another participant in focus group 2 said he found one of the messages most effective because the consequences in the message were relevant to him:

I voted for the first pitch [as the most effective] because…the main argument was that it was healthier and also costs less, which I think makes sense and is pretty relevant, I think, to my life and also a lot of others.

By contrast, many participants indicated they found the social norms message ineffective because they did not find information about what other people are doing to be personally relevant. For example, a participant in focus group 4 said, “I also find that [the social norms message] was not that convincing to me either… ‘Cause it’s like, I don’t care about what other people do.”

3.1.4. Theme 4: Including statistics in messages can enhance perceived message effectiveness.

About half of the messages that participants viewed included statistics (e.g., the warning message stated, “Producing beef creates 17 times more carbon emissions than plant-based meals and 6 times more than chicken, contributing to climate change.”). Many participants said that messages that included statistics were more effective than messages that lacked statistics (Table 2). A participant in focus group 2, for example, said, “Having the numbers just makes the message stronger,” while a participant in focus group 1 said, “I’m always more convinced by the numbers.” Statistics were especially likely to enhance perceived message effectiveness when they were perceived as large in magnitude. For example, a participant in focus group 2 said, “I feel like especially the statistics are larger numbers and so that also makes them more convincing to me because it’s just more powerful in itself.” Statistics were also more likely to enhance perceived message effectiveness when the statistics referred to consequences that were clearly defined and personally relevant (see also Theme 2 and Theme 3, above).

Participants’ responses indicated three reasons why messages with statistics were perceived as more effective. First, participants indicated that statistics made messages feel more objective and credible, whereas messages that lacked statistics were viewed as “advertisements” or “opinions” that were less trustworthy. For example, a participant in focus group 3 said she chose the warning message as the most effective because, “I think that one works out really well because they use, like, actual statistics and stuff to convince someone. Not, just kinda, like, taking someone else’s opinion for it.” Second, participants indicated that statistics made the consequences described in the messages feel “tangible,” “concrete,” and “real” to them. For example, a participant in focus group 1 said:

I think for me the numbers just make it more realistic. The [small changes, big benefits and warning messages] are the two [messages] that have, like, the most numbers in them…the other [messages] are more like, they’re possible, this will help. But this is like, this is concrete, and that’s what would convince me.

Third, participants indicated that statistics (especially statistics that were large in magnitude) caught their attention and sometimes surprised them, resulting in more engagement with the message. For example, a participant in focus group 2 said: “[The statistic in the message] felt like it was larger than I would expect…which just flagged my interest.”

Despite messages with statistics generally being perceived as effective, some participants indicated that including statistics did not enhance perceived message effectiveness when the statistics were perceived as small in magnitude or referred to consequences that participants found hard to understand or not personally relevant. For example, several participants indicated they found statistics in the messages about juice or milk to be less effective than those in the messages about beef and pork, because the former statistics were typically smaller in magnitude than the latter:

I think one thing I’ll point out is while I originally thought [the warning message] was among the more effective ones because it was focused on statistics, the statistical changes provided this time [for the juice substitution] aren’t, like, that much. Like, I feel like 1.3 times and 4% wouldn’t be that much.

One participant indicated that including uncompelling statistics in messages could make the messages less effective than those that did not include any statistics: “I love stats too, but I would prefer no stats over bad stats.”

3.1.5. Theme 5: Succinct, positive messages were well received.

Throughout the discussions, participants indicated they found messages that were succinct to be more effective than messages that were longer (Table 2). Many participants also found messages to be more effective when they used a positive tone that felt “inspirational” or “optimistic.” By contrast, participants found messages with a negative tone to be less effective. For example, many participants found the anti-industry message ineffective in part because they found its tone to be “aggressive,” “polarizing,” or “divisive.” A participant in focus group 1, for example, said she chose the anti-industry message as the least effective because, “It just felt really aggressive rather than focusing on the positives of the plant-based, it seemed to only focus on the negatives of the meat companies, and it just seemed really intense.” Despite many participants indicating they found messages with a positive tone to be effective, a participant in focus group 4 hypothesized that messages with a negative tone would better motivate people to adopt the target dietary substitution because these messages would make people “pretty much think that [diabetes and other poor health consequences] gonna come to them if they don’t change their dietary behaviors.”

3.1.6: Theme 6: Some participants doubted that their personal dietary choices would have meaningful environmental impacts.

In response to messages about the environmental consequences of dietary choices, some participants indicated that they doubted whether any individual making the target dietary substitutions would meaningfully reduce the negative environmental consequences of food production. For example, a participant in focus group 2 said, “It’s hard to imagine that a single person changing between juice or fruit, you know, makes a difference.” Others used the phrase “I’m just one person” to express their skepticism that adopting the target dietary substitution would make a meaningful impact, offering comments such as:

While I concede that there are environmental benefits to eating less beef or pork, my choice whether to eat beef or pork or not is not really significant on environmental scale. Like, I’m just one person…making that change wouldn’t really do anything on a grander scale.

No participants expressed disagreement with these doubts, but one participant indicated that he still found the messages effective because he believed that individual action could spur collective action:

I think all we can ultimately control are our individual actions, so I feel like I’m still appealed to by messages that encourage me to take steps to reduce my carbon emissions, even if individual consumers aren’t the biggest culprit. Because I think social change always starts with the self.

For some participants, their general skepticism that their personal dietary choices would have meaningful environmental impacts was alleviated by learning that others were also making dietary choices to reduce the negative environmental consequences of food production. For example, a participant in focus group 4 said he found the social norms message to be effective because it included information about other people also adopting the target dietary substitution: “I think for me, [the social norms message] does make a bit more of a difference because for some things I think to myself, what’s the point if I don’t feel like a lot of people are doing it?” Similarly, a participant in focus group 3 tried to convince another participant to adopt the beef and pork substitution by saying, “…while it may seem like you are just one person, there are other people in the world that are making this change, um, whether you know it or not, so you’re not alone.”

4. Discussion

Prior studies have found that simple dietary substitutions – such as replacing beef with chicken or plant-based entrees – could reduce the negative environmental consequences of food production and improve dietary quality, including among emerging adults (Grummon et al., 2023; Rose et al., 2022). Our study suggests that well-designed messages may hold promise for encouraging emerging adults to adopt these substitutions. Specifically, emerging adults indicated they found messages to be effective when messages discussed a specific, achievable dietary change; linked that dietary change with a clear consequence; discussed personally relevant information; included statistics (especially when large in magnitude); were short; and used positive language. These findings provide guidance for campaign developers seeking to encourage emerging adults to make dietary changes.

Emerging adults in our study found messages to be more effective when the message contained direct, actionable advice about how to change their diets. Although all the messages we tested focused on specific dietary substitutions (rather than broader changes to the diet, such as becoming vegetarian), participants found messages to be more effective when the messages additionally indicated how often they should make these substitutions. These results suggest that messages encouraging dietary changes should be as specific as possible about how the audience should change their behavior. This finding is consistent with the Theory of Planned Behavior’s “principle of compatibility,” which posits that people are most likely to adopt a recommended behavior change when that change is a well-specified, concrete, and observable behavior (Ajzen & Fishbein, 1977; Booth-Butterfield & Reger, 2004). Additionally, our results suggested that messages may be more effective when they emphasize that the suggested dietary change is achievable, for example by using language like “just one serving” or “just once per week.” This finding is consistent with prior studies suggesting the benefits of a “small changes” approach to dietary improvement (Hill, 2009; Hill et al., 2003; Stroebele et al., 2009; Wehbe et al., 2022).

Emerging adults also found messages to be more effective when the message linked the target dietary substitution to clear and understandable health and environmental consequences such as heart disease, cancer, or greenhouse gas emissions. This finding aligns with prior studies indicating that messages are more persuasive when they focus on specific consequences compared to general descriptions of consequences (Eckhardt et al., 2015; Roberson et al., 2005; Webb & Eves, 2007). Prior research also indicates that people perceive specific consequences as more personally relevant than general consequences (Roberson et al., 2005), which could make them more likely to process messages through the central route of persuasion (Petty & Cacioppo, 1986).

Emerging adults in our study found messages that included statistics to be more effective than messages without statistics. This finding aligns with experiments finding that messages with statistical evidence have a stronger influence on beliefs, attitudes, and (in some studies) intentions than messages with narrative evidence (Greene & Brinn, 2003; McKinley et al., 2017; Zebregs et al., 2015). Statistical evidence provides more information about the prevalence, likelihood, and severity of a particular consequence, which may make people more likely to believe they would experience that consequence (Zebregs et al., 2015), in turn leading to stronger motivation for behavior change (Ajzen, 1991). Consistent with that hypothesis, emerging adults in our study reported that statistics made messages feel more “real” and “tangible” to them. Emerging adults in our study also indicated that statistics enhanced message credibility, which increases message effectiveness (Petty & Cacioppo, 1986; Pornpitakpan, 2004).

Across focus groups and dietary substitutions, the small changes, big benefits and warning messages were well-received, eliciting a majority of participants’ votes for being the most effective message. These two messages featured several of the characteristics that participants indicated they found effective. For example, both messages included statistics and linked the dietary substitutions to clearly defined consequences, and the small changes, big benefits message was also succinct and used a positive tone. Prior research also suggests other reasons these messages were perceived as particularly effective. The small changes, big benefits message might have resonated with participants because it directly addressed the psychological heuristic called the “peanuts effect,” which describes people’s tendency to discount small individual amounts (Roberto & Kawachi, 2015). Similarly, the warning message may have been perceived as effective because warning messages have been shown to evoke emotions and thinking about harms, two reactions that can lead to behavior change (Brewer et al., 2019; Grummon & Brewer, 2020; Hall et al., 2017; Noar et al., 2015).

A few emerging adults indicated that they doubted that making the target dietary substitutions would lead to meaningful environmental benefits. These responses exemplify the concept of low climate change efficacy (i.e., low belief that individual behaviors are impactful on the environment (Salomon et al., 2017)) as well as its more general cousin, low participative efficacy (i.e., low belief that one’s own actions will “make a difference” to collective efforts aimed at achieving group goals (van Zomeren et al., 2013)). These responses merit attention given that when individuals have low climate change or participative efficacy, they have lower intentions to reduce energy use (Salomon et al., 2017) and participate in collective action related to environmental issues (van Zomeren et al., 2019). Messages might be more effective if they address low climate change or participative efficacy, for example, by emphasizing that individual actions are easy to implement and have a strong impact on personal carbon emissions (Aberman & Plaks, 2022; Salomon et al., 2017). Pairing messages with changes to the physical environment (e.g., increasing the prominence or availability of healthier, more sustainable foods) could also boost uptake of healthier and more sustainable dietary choices (Bianchi et al., 2018; Garnett et al., 2019; Koutoukidis et al., 2019).

We note six limitations of this study. First, we used a convenience sample, though the final sample was diverse along gender, race/ethnicity, and income lines. As with other qualitative studies (Neale et al., 2014), no inferences can be drawn beyond the sample. Second, we conducted the focus groups using Zoom. Using Zoom meant that we excluded potential participants without computer or Internet access, participants’ environments may have had distractions, and body language was less easily observed compared to in-person discussions (Flayelle et al., 2022). However, using Zoom also enabled us to recruit a national sample. Third, most participants had completed at least some college, all lived in the US, and all were ages 18–25; future studies will need to establish whether our findings generalize to people with less educational attainment, other countries, and other age groups. Fourth, given the small sample size and qualitative nature of the study, we are unable to determine whether responses to health messages differed from responses to environmental messages or to examine whether participants’ characteristics (e.g., nutrition knowledge, gender, race/ethnicity, education) modified their responses to the messages. Fifth, we did not collect information on where participants lived within the US, though both dietary patterns and views on environmental sustainability vary across regions (Allcott et al., 2019; Marlon et al., 2023). Sixth, our discussions focused mostly on message content; future research will be needed to explore other potentially influential aspects of message design such as source, imagery, and medium (Demeshko et al., 2022; Keller & Lehmann, 2008; Lee et al., 2023; Pornpitakpan, 2004).

Our study also brings three key strengths. First, our sample of emerging adults was diverse in terms of gender, race, ethnicity, current college enrollment, and household income. Second, we developed an interactive discussion guide that included asking emerging adults to develop messages in their own words. Third, we focused on gathering in-depth information about reactions to messages from emerging adults, who merit special attention in nutrition research given that they have poor dietary quality (Krebs-Smith et al., 2010; Rehm et al., 2016) and that dietary patterns established during emerging adulthood track into later adulthood (Murthy et al., 2016; Reis et al., 2013).

CONCLUSIONS

In this qualitative study with emerging adults, participants reported that a variety of message topics would convince them to make healthy, sustainable dietary substitutions such as replacing beef and pork with poultry or plant-based entrees. Campaign developers aiming to encourage healthy, sustainable dietary substitutions should consider designing messages that make the target dietary changes seem achievable (e.g., by using language such as “just one”) and use statistics to clearly describe the benefits of making these changes, especially if statistics are large in magnitude. Messages should also address doubts about the potential for individual dietary choices to affect environmental outcomes.

Data Availability

The lead author has full access to the data reported in the manuscript. All de-identified original data will be made available upon request and pending human subjects’ approval from the requester’s home institution.