The production of different types of food has varying impacts on the environment (Nijdam et al., 2012), which has important implications for contributions and adaptation to climate change (Campbell et al., 2017). While livestock production emits high levels of greenhouse gases (GHGs)—generating approximately 15% of anthropogenic GHGs (Gerber et al., 2013)—the production of pulses, which is an alternative to livestock-based foods as a plant-based protein source, contributes significantly less (Nijdam et al., 2012; Oliveira et al., 2019). A study estimating the reduction in GHGs when pulses were substituted for beef in people’s diets found that this change alone would have achieved 50–75% of the US’s then-current target reduction of GHGs (Harwatt et al., 2017). Pulses may also present a more acceptable substitute to animal-source proteins than other alternatives. An article on alternative proteins found that consumers preferred plant-based proteins—such as those found in pulses—to other alternative proteins, such as insects and lab-grown meat (Onwezen et al., 2021).

At the same time, human dietary choices have been increasingly linked to negative health impacts (Popkin et al., 2006). Low diet quality contributes to high rates of overweight and obesity (Hales et al., 2018; Hall and Guo, 2017), which, through related non-communicable diseases, is a leading cause of death in the US (Murphy et al., 2021). High intake of ultra-processed, calorie-dense foods leads to weight gain (Hall et al., 2019), which is linked to increased risk of type-2 diabetes, cancer, and heart disease (Preston et al., 2018). In contrast, higher intake of pulse foods has been linked to greater nutrient density and increases in diet quality (Mitchell et al., 2021).

Pulses are more environmentally sustainable (Nijdam et al., 2012), healthier (Foster-Powell et al., 2002; Mitchell et al., 2009), and more affordable than many other protein foods (Drewnowski and Conrad, 2024). These traits are important to many consumers (Akaichi et al., 2017; Cox et al., 2012; Glanz et al., 1998), although individuals may place different values on taste, health, sustainability, and affordability (Robinson et al., 2022). Despite offering numerous and diverse benefits, consumption of pulses in the US is significantly below recommended levels (Guenther et al., 2006; Perera et al., 2020).

Barriers to pulse consumption may include a host of factors, including lack of awareness of the benefits that pulses offer and perceived negative impacts, such as digestive issues (Henn et al., 2022; Winham et al., 2019, 2016), difficulty in identifying pulses in crowded retail environments (Melendrez-Ruiz et al., 2021), or lack of knowledge of what pulses are (Winham et al., 2020). Differences in awareness of pulses may mean that consumers are unfamiliar with the diversity of pulse types and the effects of consuming different pulses. For instance, beans are familiar to most Americans, while lentils and chickpeas may be less well known (Rehm et al., 2023; Sanjeevi and Monsivais, 2024; Winham et al., 2020). Further, research shows that a widely held barrier to consumption—that pulse foods lead to intestinal gas and bloating (Palmer et al., 2018; Winham et al., 2019)—is less likely to result from the consumption of lentils, chickpeas, and peas, potentially due to differences in alpha-galactosides in the raffinose family (Veenstra et al., 2010).

Additionally, consumers that are interested in environmentally responsible food choices have little information on which to rely during food choice. Robust literature has examined the impact of eco-labeling on consumer choice (Edenbrandt et al., 2025; Katare and Zhao, 2024; Muller et al., 2019; Rondoni and Grasso, 2021; Taufique et al., 2022; Vanclay et al., 2011); however, real-world implementation of these labels is limited in many countries. In the US, for instance, only the organic label is widely used in the marketplace (Big Room Inc., 2025). Further, it is believed to be unlikely that regulations will be enacted any time soon that mandate the use of eco-labels (Cole, 2025).

Efforts to identify ways to motivate changes in consumer behavior to improve human health and the sustainability of food production systems have engendered a significant body of literature (Cadario and Chandon, 2020; White et al., 2019). However, much of this research uses simple choice sets featuring two to four items at a time rather than the dozens to hundreds of products per category in many food retailers, which likely leads participants to pay more attention to products and information than they would in complex environments (Meißner et al., 2020). In fact, mobile eye-tracking studies documenting shoppers’ attention to food and label information in real-world supermarkets found that only a small fraction of products and information are considered (Machín et al., 2023, 2020). Indeed, a study examining gaze behavior in a virtual reality supermarket found that pulses did not systematically capture attention (Melendrez-Ruiz et al., 2021). A frequent finding in studies on the impact of nutrition information is that many consumers report not observing product information, and only a small fraction of those who do notice it report using it (Cantor et al., 2015). The impacts of labeling interventions estimated from simple experimental studies are significantly larger than when implemented in real-world supermarkets (Dubois et al., 2021).

Most studies in complex retail or online environments examine interventions focused on labeling or product positioning (Katare and Zhao, 2024). However, there is a small, but growing, body of literature suggesting that reminder messages are effective at promoting healthy choices in crowded product environments (Arslain et al., 2020; Gustafson et al., 2018). These reminder messages are meant to activate consumers’ innate goals and direct attention to accomplishing target outcomes. Reminders about health appear to refocus attention to health attributes (van Koningsbruggen et al., 2011), even among sample of participants not actively trying to change eating patterns (Hare et al., 2011), and can lead to increased frequency of healthy behaviors, such as physical activity (Calzolari and Nardotto, 2017; Habla and Muller, 2021). While subtle messages may go unnoticed (Papies et al., 2014), resulting in small impacts on choices (Panzone et al., 2021), reminders are stronger statements directing attention to the target outcome.

Stimulus-rich environments, such as retail markets with myriad unique products, may distract from other long-term goals of consumers, such as making environmentally friendly choices. Unless individuals actively consider the health and environmental impacts of choices or have established healthy consumption patterns (Gustafson, 2022; Tuyizere and Gustafson, 2023), consideration of health or environmental attributes may be forgotten or overlooked while shopping in the supermarket. While research shows that priming consideration of sustainability increases the likelihood that consumers will choose more environmentally friendly products (Thøgersen and Alfinito, 2020; Torma et al., 2018), most of this research has not been conducted in complex choice settings, which are more cognitively demanding to navigate and reduce the number of attributes that consumers consider when making choices (Meißner et al., 2020). Indeed, primes implemented in complex environments frequently have modest impacts on choices (Panzone et al., 2021; Papies et al., 2014).

The advent of online shopping has led to new areas of research examining changes in the choice environment. There are multiple novel interventions enabled by the technological interface used in online shopping. For instance, studies on the use of subsets or filters that permit shoppers to shape the choice environment according to their preferences have found that these tools can significantly increase the choice of foods with beneficial nutrient and/or environmental characteristics (Gustafson et al., 2025, 2024). Research on filters in visually crowded environments find less use of these tools, which, upon follow-up discussion with research participants, was due to the tools going unnoticed (Godden et al., 2025). A stronger intervention signal involves exposing consumers directly to products with key attributes of interest. A study that involved interventions examining (in part) forced versus voluntary commitments to a low-carbon product set found that both reduced emissions embodied in the products (Panzone et al., 2023). Recent research on personalizing food suggestions based on individual health priorities found that exposing participants to a priority-responsive subset of all available products (with the option to view all available products instead) significantly increased the choice of healthy products and improved the nutrient profile of chosen products relative to an optional personalized product set condition (Gustafson et al., 2025). Although exposing participants to the personalized product set intervention condition was the most effective, participants in that condition reported lower levels of product satisfaction with their product choices. This finding suggests that there may be trade-offs between efficacy, intervention forcefulness, and satisfaction with choices resulting from intervention conditions.

In this research, we conduct an experiment on the choice of pulse products, which were defined as foods with pulses among the first three ingredients. Since pulses have beneficial human health and environmental profiles (Drewnowski and Conrad, 2024; Oliveira et al., 2019), promoting the choice of pulse products presents an opportunity to improve human nutrition and decrease environmental impacts of food production. In the absence of eco-labeling, consumers attempting to make climate friendly choices will have to rely on other attributes, such as general information or beliefs about product types. The use of pulses as a source of protein and other important nutrients is an effective way to promote human and environmental health (Oliveira et al., 2019). Thus, strategies that promote consumption of pulse foods present an opportunity to decrease the impacts of food production on the environment while also combatting the health crises caused by poor diet.

We developed an intervention to target three drivers of pulse choice in a complex choice environment, featuring six food product categories with 50 unique products per category. The interventions we studied were (1) varying the prevalence of pulse products in the market (10% vs. 20% pulse products), which reflects growth trends in the global market for pulse products (Choudhury, 2025); (2) increasing awareness or attention to nutritional and/or environmental benefits of pulse products during the choice process (no message, nutrition only, environment only, nutrition and environment messaging); and (3) choice environment interventions addressing the (implicit) search costs associated with identifying pulse products (pulse products intermixed with other products vs. the option to filter out non-pulse products). We examined the impact of these interventions on the choice of pulses and the content of nutrients of public health concern in the foods chosen by participants in the choice task. While there are groups attempting to provide access to environmental impact information for food products—see, e.g., (CarbonCloud, 2025)—the available data do not cover many of the products used in this experiment, so our analysis was limited to nutritional outcomes. We additionally examined outcomes capturing beliefs and subject knowledge about pulse food relationships to health and environment outcomes, active consideration of health and environmental impacts during choice, and satisfaction with product choices to corroborate impacts of the interventions on different elements of behavior.

We pre-registered the research design and analysis plan with the Open Science Framework (OSF)¹ for an experiment and survey implemented in Qualtrics, which is a widely used survey platform (Qualtrics, 2015). The survey is provided in full in the supplementary materials and is also available as a component of the OSF pre-registration. The research protocol was approved by the Institutional Review Board at the authors’ institution (protocol #20221122409EX). The research was performed in accordance with guidelines for research with human subjects, including the Declaration of Helsinki. All participants were adult (≥19 years of age) residents of the United States (US) and provided informed consent prior to the research.

Participants were recruited from Dynata’s consumer panels.² The researchers worked with Dynata to recruit a sample of participants representative of key demographic characteristics in the US, using US Census-based demographic quotas. The quotas set sample characteristics for sex, age, income, and education. We used the software program G*Power to calculate sample size (Faul et al., 2009). The aim was to be able to detect a small effect size (0.2) with 0.95 power at 0.05 alpha error probability. The targeted sample size was complete responses from 6,000 participants. We collected usable responses from 6,401 participants.

Table 3 reports the summary statistics of the demographic characteristics for the full sample and compares it to estimates of the distribution of demographic characteristics in the US population (US Census Bureau, 2023). As Table 3 shows, the sample matches the US Census demographic characteristics closely. There were no significant differences between the sample and US Census data (p-values from the Fisher’s Exact Tests are reported in the table). Fisher’s Exact Tests also revealed no statistically significant differences in the distribution of any of the demographic variables across intervention conditions (Supplementary Table S4).

In this article, we extend and integrate recent reports examining the impact of messaging and choice environment interventions on product choice. We also estimated the impact of an experimental condition that increased the number of options with the target attribute (a pulse within the first three ingredients in a product) to simulate a recent trend in the development of more pulse products. While the presence of products and attributes in the market is the result of market interactions between producers’ product offerings and consumers’ purchase decisions, the development of pulse products has been growing steadily and that growth is expected to increase over the next decade (Choudhury, 2025).

We found significant impacts of all intervention arms on the choice of pulse food products. While messaging increased the choice of pulse products, the provision of a pulse filtering option led to markedly greater increases in choice of pulse foods. The combination of messaging and filtering had the greatest impact, which aligns with previous findings showing that combined messaging and filtering has a higher impact on choices than messaging alone, which also had a significant impact relative to a control condition (Gustafson et al., 2024). Importantly, the conditions featuring pulse filters also had significant impacts on the amount of nutrients of public health concern in the US, increasing the content of underconsumed nutrients (such as dietary fiber and potassium) and decreasing the content of overconsumed nutrients (such as saturated fat and sodium). This finding aligns with the fact that pulses are a rich source of many important nutrients (Drewnowski and Conrad, 2024), but provides valuable evidence highlighting that promoting food products that contain pulses as an ingredient also improves the nutritional profile of selected foods.

The fact that the pulse filter option alone had a larger impact than the messages alone may seem counterintuitive. However, the filter itself—containing the words “pulse”—may have primed consideration of health and/or environmental concepts (Urminsky and Goswami, 2019), while also markedly reducing search costs to identify pulse products. In our exploratory analysis, we found significant positive relationships between individual characteristics related to habits, knowledge, and priorities about health and environment during food choice and use of the pulse filter. This suggests that the presence of the filter may have enabled individuals who regularly consumed pulses and had greater health and environmental preferences to focus in on pulse products.

The combination of messaging and filter conditions resulted in the highest likelihood of pulse products being selected. This reflects findings from other settings. A number of articles examining messages that prime or prompt consideration of health and the environment has found impacts on blood oxygenation level dependent measures of brain function (Hare et al., 2011), purchases in supermarkets (Gustafson et al., 2018; Papies et al., 2014), the set of products considered and information used during the choice process (Arslain et al., 2021; Gustafson et al., 2021; Gustafson, 2023; Gustafson and Rose, 2023) as well as choice of healthier/sustainable products in experimental settings documenting choice process variables (Arslain et al., 2020; Gustafson et al., 2024), and valuation of products (Hosni et al., 2025).

The results related to message type—health only, environment only, health and environment combined—are mixed. While we expected that the combined message would be most impactful in promoting the choice of pulse foods, it is not true in all conditions. For example, while the combined message does lead to the highest likelihood of pulse choice among the Ten Filter messaging conditions, it has the lowest likelihood among the Twenty Filter messaging conditions. While some research suggests that combined health and environment messages are most effective at promoting pulse-related consumption (Lemken et al., 2017), our results may highlight tradeoffs associated with attempts to present multiple pieces of information. For instance, a recent study on the impact of obesity and food waste information on portion size preferences found that simultaneously presenting information yields smaller effects than sequentially presenting information (Hosni et al., 2025). Of the health and environmental messages, results suggest that the health messages had a greater impact on the selection of pulse products, which reflects findings in other studies that health concerns appear to be more relevant and impactful during food choice than environmental concerns (Henn et al., 2022).

Further, our results point toward the role of complexity in limiting the effectiveness of informational, labeling, or incentive-based interventions to promote healthier choices. In complex, real-world retail environment, the attention paid to products and relevant attributes for healthy decisions, such as warning labels or nutrition information appears is limited (Cantor et al., 2015; Dubois et al., 2021; Elbel et al., 2009; Machín et al., 2023, 2020). In fact, more complex choice environments have been found to decrease individuals’ attention to products and product attributes (Meißner et al., 2020). While reminders to consider important outcomes, such as health, have been found to shift attention to healthier products and to relevant attribute information (Arslain et al., 2021), the potential impact of reminder messages in isolation may be limited. Evidence suggests that these messages are markedly more impactful when the choice environment aligns with the message: at a rural supermarket serving a predominantly minority community, a reminder message about fruit and vegetable consumption goals led to significantly higher purchases of produce as well as healthy foods in general (Gustafson et al., 2018). However, a reminder message about general healthy food purchasing goals, with products identified by a locally designed, culturally appropriate labeling system (Gustafson and Prate, 2019), did not significantly increase purchases (Gustafson et al., 2018). Note that fruits and vegetables were contained within the set of all healthy products and thus it should have been easier for shoppers in the general message condition to find healthy products they would want to purchase than shoppers in the fruit and vegetable message condition. In online settings, pulse-focused messages combined with a pulse attribute filter significantly outperformed pulse messages alone in an environment in which pulse products were randomly intermixed with non-pulse products (Gustafson et al., 2024). Further, while messages may recruit attention to attributes that tend to be overlooked during choice in complex environments, they may also crowd out attention to other health or environmental attributes (Gitungwa et al., 2024; Lemken et al., 2024).

Our results about the relationship between demographic characteristics and pulse choice also provide important information. While previous studies on pulses in a US context have tended to focus on specific populations, such as low-income women or ethnic minority women in Midwestern communities (Palmer et al., 2018; Winham et al., 2019, 2016), we have data from a large sample that reflects the national distribution of key demographic characteristics. We find significant results for all demographic characteristics except race/ethnicity. Women were significantly less likely to select pulse products. Recent analysis of National Health and Nutrition Examination Survey (NHANES) data on legume consumption (which includes pulses) shows insignificant differences in consumption by gender (Perera et al., 2020), but these NHANES data reflect shared consumption environments among household members rather than individual food choices. We find lower levels of pulse choice in older adults than young adults, which may partially reflect differences in attention to sustainability issues (Yamane and Kaneko, 2021), and also reflects findings in previous studies (Perera et al., 2020). Pulse choice also differs by income. While pulses are affordable (Didinger and Thompson, 2020; Drewnowski and Conrad, 2024)—and, in fact, were less costly on average in this experiment, we find increased odds of pulse consumption with higher levels of income. The educational variable differentiates estimated pulse choice more than any other demographic characteristic. Respondents with education beyond high school were significantly more likely to choose pulse products than those with high school or less.

There are significant differences in the likelihood of pulse choice by product category. When choosing foods from the Frozen Patties and Burgers and the Snacks categories participants were significantly less likely to choose a pulse product than when choosing from the Frozen Dinners and Entrees category. This result is interesting because of the frequency with which pulses are suggested as a way to substitute away from animal-source foods in the burger category (Oliveira et al., 2019; Onwezen et al., 2021; Tarrega et al., 2020), and growth in pulse offerings in the snack category (Choudhury, 2025). Our findings may suggest that efforts to increase pulse consumption should focus on other food categories.

Finally, there are some important limitations to this research that should be noted. First, the decisions made in the experiment are hypothetical, which was necessitated by a desire to recruit a large sample of respondents with diverse demographic characteristics from throughout the US. While we employed techniques that have been found to mitigate the impact of hypothetical bias, we cannot rule out the possibility that choices would differ if participants were exchanging real money for real products. Second, the filter condition featured a programming error that prevented participants from removing the filter once it was in place. This resulted in increases in participants in those conditions saying that they would not choose any product. Whereas 9.4 percent of choices made by participants were the “none of these” response in the 10 and 20% pulse conditions, this increased by approximately 5.5 percentage points to 15 percent in the 10 and 20% subset conditions. The increase in the percent of “none of these” choices between conditions with and without the filter in place was consistent across messaging conditions as well. Further, as noted earlier, the pulse filter may have primed thoughts about health and/or the environment (Urminsky and Goswami, 2019), which would have exaggerated the effect of the filter intervention relative to a real-world scenario in which multiple filter options would be present. In future work on the feasibility of filters to promote desirable food choices, we will ensure that the filter is removable and that other filter options are available to avoid focally priming pulse benefits.

However, we also find evidence that the pulse filter eased the difficulty of finding a pulse food among the other products, making the filter particularly useful for a product category that some people have a hard time identifying (Melendrez-Ruiz et al., 2021; Winham et al., 2020). Finally, much of the recent research on consumer choice of sustainable food products has focused on labels or measures, such as carbon emissions or food miles, that attempt to represent a direct environmental impact. However, these labels have not been widely implemented, and even finding estimates for these values is currently difficult. We chose to focus on information that is universally available in the study area—the presence (and—in real-world product environments—the ranking in the list of ingredients) of pulses in food products. While this does not provide the summary measure of environmental impact that labels do, it does represent the information upon which many people have to make decisions in the real-world.

This research shows that interventions targeting a key ingredient—having pulses as one of the top three ingredients in a food product—can significantly improve the nutritional profiles of foods selected, which is true even though participants in all conditions had access to nutrient information. While we were unable to find comprehensive information about environmental impacts of enough food products to analyze the impact of promoting pulse foods on emissions or other important environmental outcomes, the findings about the impact on the nutritional profiles of chosen foods may very well translate to sustainability attributes as well. Given that it is unlikely that many countries, such as the US, will implement mandatory eco-labeling laws, interventions promoting information already available, such as pulse content in food products, may be the next best option.

We find that interventions that combine messaging with filtering, which allows easy identification of pulse products, resulted in the greatest increase in choice of pulse products and improved the nutritional profiles of products chosen by participants. Our results also suggest that the trend toward more pulse product offerings may lead to higher purchases of pulse products. However, the development of pulse food products may need additional evidence about consumer receptiveness to pulse foods by food category, because consumers were less likely to choose pulse food products in food categories—snacks and frozen patties and burgers—that have been a focus of increasing pulse products. Messaging that highlights affordability and convenience may help promote pulse consumption within demographic categories that have lower consumption levels, helping expand the reach of nutritional benefits of pulse consumption.