Consumers’ acceptance and valuation of healthier rice: implications for promoting healthy diets in the Philippines

The role of nutrition in transforming food systems is crucial. However, despite its importance, very few countries are on track to meet the global nutrition targets, including diet-related non-communicable diseases (NCDs) (Development Initiatives, 2022). The prevalence of the double burden of malnutrition, defined as the coexistence of undernutrition with overweight, obesity and diet-related NCDs, is prominent in low- and middle-income countries (LMICs), particularly in Asia and Africa (Escher et al., 2024). Modifiable lifestyle-related factors, such as food choice, play a crucial role in preventing malnutrition and mitigating the risk of developing NCDs (Development Initiatives, 2022). Understanding the factors that influence consumers’ food choice, including physiological and nutritional needs, socio-cultural context, hedonic motivations and attitudes and beliefs toward food, is essential for the development of context-specific nutrition interventions to promote healthy diets (e.g. Karanja et al., 2022). Maintaining a healthy diet helps prevent malnutrition. However, the cost of a healthy diet in LMICs is typically unaffordable (Schneider et al., 2023), as it requires larger quantities of more expensive food groups (FGs) than the starchy staples (Hirvonen et al., 2020).

Rice is one of the leading staple foods and the main source of starch for most of the population in Asia. Rice is mainly consumed in milled form (i.e. white rice), which poses a nutritional challenge. Low dietary diversity and heavy reliance on white rice can lead to the development of micronutrient deficiencies among consumers. For these reasons, among others, international and national research centers continuously explore multiple pathways to enhance the nutritional quality of rice through biofortification as a complementary food-based solution to existing nutritional interventions (Tiozon et al., 2021). This current study focuses on the ongoing efforts to develop and deploy conventionally bred healthier rice types, particularly brown, colored and low glycemic index (GI) rice (IRRI, 2024; Kasote et al., 2022). Unpolished rice, or commonly called brown rice, is produced by dehulling the paddy, thus retaining the bran layer in the grain. The bran layer is abundant in fiber and micronutrients, while the remaining grain is mainly composed of starch (Ravichanthiran et al., 2018). Colored rice comprises rice varieties with pigmented bran and has antioxidant properties that are accumulated in the bran layer. Hence, the removal of the bran layer through milling to produce white rice diminishes the nutritional property of rice. Furthermore, most white rice varieties have a high GI (Jukanti et al., 2020). The consumption of high GI foods contributes to the expansion of NCDs in Asia (Wee and Henry, 2020).

Promoting the consumption of brown rice has been used as a nutritional intervention in various contexts due to its nutritional benefits (e.g. Chiang, 2019; PHILRICE, 2013; Zhang et al., 2010). However, apart from its lower storability, some empirical studies have shown the inferior sensory properties of brown rice, which has a negative influence on consumer acceptability (e.g. Saleh et al., 2017). For colored rice, studies that explore the nutritional diversity have been conducted to predict the phenolic and mineral content of rice based on its color and biochemical indicators (Buenafe et al., 2022). These efforts aim to help breeders to develop rice varieties with enhanced nutritional properties, as the presence of phenolic compounds in colored rice has been associated with numerous nutritional benefits, such as antioxidant, anticancer and antidiabetic properties (Tiozon et al., 2021). Mainstreaming of GI and resistant starch in breeding programs is also being done. Genetics for low GI have been defined and led to the development of pre-breeding lines in a high-yielding background (IRRI, 2023; Jukanti et al., 2020).

Many empirical studies have evaluated the intrinsic quality traits of rice (e.g. physical, eating and cooking quality) to help breeders incorporate market-oriented traits in varietal improvement programs and to help policymakers formulate value chain-upgrading strategies to increase smallholders’ market access. However, through a systematic literature review of 106 studies, Custodio et al. (2023) observed that few studies evaluate nutrition-related attributes, which are inherent characteristics of rice but cannot be verified after consumption (i.e. these are credence attributes). This study aims to fill this knowledge gap by (1) investigating consumers’ perception of healthier rice types, (2) measuring and comparing acceptance and willingness to pay (WTP) for three healthier rice types and (3) identifying the factors that influence WTP. The empirical results contribute both to nutrition education strategies and breeding programs. On the one hand, they provide insights into how healthier rice products should be promoted to nudge consumers toward increasing uptake of these products. On the other hand, they provide “market intelligence” that can support the identification of future market segments for breeding pipelines and seed systems developing and delivering nutritious rice to address malnutrition (CGIAR, 2024).

Our study focused on Asia, where rice is mainly grown and consumed and where the prevalence of malnutrition is high. We focus on the case of the Philippines in Southeast Asia, where (1) consumer awareness of brown rice is high, but consumption is low (Chiang, 2019), (2) studies on consumers’ evaluation of colored and low-GI rice are lacking (Custodio et al., 2023), and (3) the prevalence of diet-related NCDs is high (Schneider et al., 2023). We targeted the capital city Manila, where most of the urban population are from middle-income class households, categorized into upper-middle, middle-middle and lower-middle income tiers. The focus on the middle class in this current study is within the premise of the country’s long-term development goal to be a predominantly middle-class society. Albert et al. (2018) argued that investigating the middle-class consumers is crucial as they provide valuable insights on the potential effective demand for novel products since they have more disposable income to try these products.

This study embeds consumer acceptance and valuation for healthier rice types in a conceptual framework based on existing literature on behavioral intentions to purchase, substitute or consume rice and on food choice for improved health and nutrition outcomes (Figure 1). Many of these studies are guided by the food systems framework for diets and nutrition (HLPE, 2017) and the socio-ecological model to explain the factors that influence behavioral intentions (Downs et al., 2020).

Socio-ecological models in food choice research emphasize the structures and processes shaping dietary influences. These layers of influence are often examined to illustrate how individual and environmental factors explain differences in food choice. The layers closest to diets are the “individual factors” and the food environment (Downs et al., 2020; Herforth and Ahmed, 2015). In our conceptual framework, these variables refer to the socio-economic characteristics for the former and the past consumption and purchase of rice as proxy indicators of the availability and physical access to different rice types, which may also be considered as a proxy for consumption and purchase behavior in general. Many studies that evaluated preferences for rice attributes include socio-economic characteristics of potential buyers, with mixed results in different contexts. For example, income was found to significantly influence WTP for sustainably produced rice in Vietnam (My et al., 2018), whereas Sriwaranun et al. (2015) reported non-significant association between income and WTP for organic rice in Thailand. Education was found to negatively affect WTP for local rice in Nigeria (Babatunde et al., 2019). Sriwaranun et al. (2015), on the other hand, did not find a significant relationship between education and WTP for organic rice in Thailand. In terms of gender, Bairagi et al. (2021a) found that women in the Philippines are more likely to purchase heirloom rice than men, while some studies did not find a significant association between gender and purchase intention for imported rice brands in Nigeria (e.g. Obih and Baiyegunhi, 2017). Some studies also took into account household size. Bairagi et al. (2021b) and Babatunde et al. (2019) reported a negative influence of household size on purchase intention and valuation of colored and local rice, respectively. To represent consumer access to different rice types, Bairagi et al. (2021a) included the consumption of brown rice as a predictor in their econometric model and found a positive influence on purchase intentions for colored rice in the Philippines. However, in some of these studies, many socio-demographic variables (e.g. marital status and age) were not statistically significant in influencing behavioral intentions.

Measures of attitude are often included as a predictor for understanding behavioral intentions. My et al. (2018) found a positive association between the perceived importance of healthy eating behavior and attitude toward high-quality rice in Vietnam. For a broader context, we included the food-based dietary guidelines (FBDG) into our measurement of attitude toward healthy eating behavior. FBDG is a context-specific set of food-based recommendations to promote overall health (FAO, 2024). As such, adherence to health-promoting guidelines may be considered as a construct of healthy eating behavior and may have a possible influence toward behavioral intention to purchase and/or consume healthier food products in general and healthier rice in particular. Although there are studies that have examined the association between attitudes and consumption of specific FGs (reviewed in Philippi et al., 2016), we have not yet found a study that has examined attitude toward FBDGs as a possible predictor of behavioral intentions to purchase or consume specific rice types.

Food choice is also influenced by psychosocial factors (Herforth and Ahmed, 2015), such as consumers’ perceived diet quality. To assess diet quality, we examined dietary diversity at the household level based on the FGs consumed at different eating occasions. Dietary diversity is considered one of the components underlying a healthy diet (HLPE, 2017) and an indicator of diet quality (Schneider et al., 2023). FAO (2011) provided a guideline for calculating the household dietary diversity score (HDDS), originally designed to gauge a household’s economic ability to access a variety of foods (Verger et al., 2019). Several studies, including those by Hammond et al. (2017) and Ritzema et al. (2019), have employed the HDDS as an indicator of diet quality. In our conceptual framework, the HDDS was disaggregated by main eating occasions and snack occasions, following the analytical framework of Evans et al. (2015), in which occasions were categorized as (1) main eating occasions (breakfast, lunch and dinner) and (2) snack occasions (morning, afternoon and late night). Several studies have shown the importance of eating occasions for understanding the construct of a healthy diet (e.g. Huseinovic et al., 2016). Lastly, we included an information treatment in our framework to assess potential differences in valuation based on incremental level of information (e.g. My et al., 2018), particularly the generic health benefits of the rice type as a credence attribute. Claimed health benefits cannot be verified by consumers themselves, and they have to rely on quality cues such as labeling (or information on the package) (reviewed in Custodio et al., 2023). Having detailed information about the novel goods and their characteristics acts as an important stimulus for consumers, which would enable them to thoroughly assess the values they would attach to such products (De Groote and Kimenju, 2008).

We collected data in Manila, the capital city of the Philippines, through a two-phase approach. The first phase consisted of six online focus group discussions (FGDs). Recruitment, group moderation and analysis were outsourced to a market research (MR) company. The consumers were recruited in person by the MR company, following a set of qualifying criteria similar to the recruitment criteria in the second phase. The participants were then informed of the schedule for the online FGDs. To capture experiences of households with different family composition, participants in two sessions belong to a household (HH) with children aged 13–20 years old, participants in the other two sessions have children aged 12 years old and below and one session has participants who live in a household without any children below 20 years old. A total of 24 individuals participated in the FGDs, each group comprising four respondents. The FGDs were conducted over three days in October 2020, with two consecutive sessions held in the afternoon each day. Each FGD session lasted between 30 and 40 min and involved one moderator, one note taker, a session recorder and two research team members as observers. Webcams were turned on only for the moderator and the four participants, while non-participants kept their webcams off. The research team crafted the discussion guide used by the moderator. Before each session, the research team provided a short briefing with the moderator, followed by feedback after each session to improve facilitation. The FGD results were used to guide questionnaire development in the next phase, particularly in exploring the constructs of a healthy diet, which resulted in the inclusion of the FBDG and perceptions of diet quality in the survey questionnaire. The FGD results also provided guidance in developing the preliminary code frame of food items based on respective FGs, which were then used to derive the HDDS in the survey data analysis. Additionally, consumers’ overall impressions of the healthier rice types were explored, which was then integrated in the overall analysis. In the second phase, a survey was conducted in April 2021, with 600 respondents. Data collection was outsourced to the same MR company that moderated the FGDs. Selection criteria were set as follows: (1) 20–59 years old to capture a subset of the economically active population age group (PSA, 2022c), (2) from a household in the middle-income class tiers (i.e. income level from PhP19,000 to PhP115,000) and (3) the main decision-maker or with active involvement in both household food shopping and meal planning or preparation. A “soft” quota on income classes was set to reflect two subclasses based on population distribution as: (1) 65% for lower-middle (between PhP19,000 and Philippines Peso (PhP) 38,100) and (2) 35% for middle-middle and upper-middle (between PhP 38,100 and PhP114,200). A (+/−) 5% deviation from the set sample size in each subclass was allowed to improve the survey completion rate. No form of quota was imposed on age and gender because these are secondary criteria and would be household-specific depending on who is involved in food shopping and meal planning and/or preparation. A between-subjects design was implemented in the survey to avoid learning and order effects and hence improves internal validity of the study (Hair et al., 2010). Each respondent was randomly assigned to one of three rice types (i.e. brown, colored and low GI) and received either the baseline information or the information treatment for the assigned rice type. With this design, a total of six independent groups were analyzed. The information received was either (1) the control, which presents the product description and examples of dishes that can be prepared with the rice for specific eating occasions as the baseline information, or (2) the treatment, which presents the same product description as in (1) with additional generic information about the health benefits of the rice products that help reduce the risk of contracting NCDs (refer to the Supplementary materials). Designing the baseline information was guided by the gastronomic systems research (GSR) framework, which proposes that specific rice-based dishes and consumption occasions are possible entry points for nutritional and dietary interventions for the introduction of novel food products and ingredients (Cuevas et al., 2017; Samaddar et al., 2020; Custodio et al., 2021; De Steur et al., 2022). Adding the health benefits in the information treatment was guided by a previous study (My et al., 2018) that examines the potential differences in valuation based on incremental levels of information and, in this case, presenting the credence attribute of rice. Both the FGDs and surveys were conducted online, as face-to-face interviews were not possible at the time of the study due to the COVID-19-related lockdown. Data collection in both phases was outsourced to an MR company.

The study obtained ethics approval from the Institutional Research Ethics Committee of the International Rice Research Institute (IRRI) (Protocol Code Number: 2020-0015-A-2015-117).

Acceptance of healthier rice types was measured using a five-point scale (where 1 = “definitely unacceptable”, 5 = “definitely acceptable and 3 “neutral” as midpoint). WTP estimates can be used to estimate market demand for novel goods or changes in the qualities of existing goods. In this study, the WTP values were elicited using an open-ended contingent valuation (CV) approach, a direct method in eliciting stated preferences. The respondents were asked the maximum amount they were willing to pay for the healthier rice they evaluated based on either the baseline information or the information treatment (i.e. with additional information about health benefits) [¹]. Lusk and Hudson (2004) argued that this approach systematically yields more conservative WTP estimates than a discrete-choice format. Additionally, a consumer’s exact WTP can be elicited through the open-ended CV approach as opposed to eliciting their discrete choices (i.e. observations on whether an individual would pay more or less than a particular price) such as in single- or double-bounded dichotomous choice questions and choice-based conjoint frameworks, which adds complexity to WTP estimation (Lusk and Hudson, 2004). While CV is prone to hypothetical bias, due to a lack of perceived consequentiality and possibly as a result of strategic survey responses, De Groote and Kimenju (2008) argued that it can be used when the consumer is well informed about the new product (in this case, an unfamiliar product accompanied with information) along with the advantage of eliciting exact WTP values. Brown and colored rice, albeit existing in the domestic market, are unfamiliar products based on low consumption in the capital city and nearby peri-urban and rural areas (Bairagi et al., 2021a, b; Cabardo and Depositario, 2018). The market for low-GI rice, on the other hand, is not yet established in the Philippines. Therefore, since we are dealing with unfamiliar products, there is a need to provide basic product description (e.g. brown and colored rice are processed and sold as unpolished, which means that the bran layer is retained) and the dishes that can be cooked with the rice to serve as baseline information (refer to the Supplementary materials). The incremental information on health benefits is the information treatment. This design allows for comparisons of valuations between products based on the additional layer of information. The variable “intention to adopt FBDG” was measured using a five-point frequency scale (where 1 = “never”, 2 = “occasionally”, 3 = “once a month to a few times a month”, 4 = “once a week to a few times a week and “5 = “almost every day”) on seven statements from the FBDG for the Philippines (FAO, 2024). The variable “self-reported diet quality” refers to the overall weekday diet of the household and was measured on a five-point scale. The HDDS was computed by summing the FGs consumed in any of the defined eating occasions (FAO, 2011). We computed the HDDS with 10 FGs instead of the FAO guideline’s 12, excluding “fat and oil” and “spices, condiments.” Studies in the past have taken this approach to exclude the FGs consistently used in the target population’s cuisine (e.g. Yin et al., 2017; Zhang et al., 2020). The HDDS were computed for the main eating occasions and snack occasions.

To understand the determinants of consumers’ WTP, a multiple linear regression (LR) was estimated. The use of multiple LR was deemed appropriate to analyze the dataset mainly due to the cross-sectional nature of the data where the dependent variable is a continuous measure of valuations in local currency (Hair et al., 2010) and that the WTP values were neither censored at zero nor at the market price of any rice type, although the elicited values are all positive and non-zero values. A Tobit model might have been appropriate for a panel dataset that is censored and would have many zero values, as is often the case for experimental auction data such as in Teuber et al. (2016), Waldman and Kerr (2015) and Ballesteros et al. (2023). A hierarchical approach was also used in the current study to assess the block effects of the different constructs in the conceptual framework, which is similar to past studies (e.g. Chiou et al., 2023; Tama et al., 2021). They are (1) socio-economic characteristics and past rice purchase and consumption behavior, (2) attitude and perceptions toward healthy eating, (3) indicators of diet quality and (4) information treatment.

Cook’s distance was computed to detect outliers. Observations with large values of Cook’s d were inspected through a scatterplot. Although there is no rigorous cut-off (Denis, 2020), 23 outliers (i.e. 3.8% of the total sample n = 600) were removed with values more than three standard deviations from the mean d. The final sample size is 577, which satisfies the desired ratio of observations to independent variables of 20:1 (Hair et al., 2010).

The assumptions of multiple LR are met in the final estimation. Collinearity diagnostics showed variance inflation factor values well below the threshold value of 10. The Breusch–Pagan/Cook–Weisberg test for heteroskedasticity was conducted to verify whether the assumption of homoscedasticity is satisfied. Linearity of the standardized residuals was checked through visual inspection of the locally weighted scatterplot smoothing graph. A one-sample Kolmogorov–Smirnov test and visual inspections based on a normal probability plot and histogram were conducted to check normality. Other data analysis techniques included Kruskal–Wallis H test, Wilcoxon rank-sum test and correlation analyses. The qualitative results were summarized by the MR company. As such, we used the findings to substantiate the survey results in the discussion.

The descriptive statistics of the socio-demographic variables are shown in Table 1. Most respondents are the main decision-makers in both food shopping and cooking or meal planning for the household (comprising 88% of the sample). The proportion of women in the sample was 71%, which is expected as women are centrally involved in household decisions about buying and preparing food. The average age of respondents is 35, higher than the regional average of 26 years. Nearly 50% hold a university degree, which is also above the regional average (30%) (UPPI, 2022). The average monthly household income is PhP35,690 (approximately US$637), which is nearly in line with the regional average of PhP34,820 (approximately US$622) but higher than the national average of PhP25,600 (approximately US$457) (PSA, 2022a). This is not surprising, given that the study location is the capital city. The mean household size is five persons, larger than the regional average of four persons (PSA, 2022b). More than 60% of the sample is employed. Households with young children under six years, school-age children and teenage children make up 37, 36 and 26% of the sample, respectively. About 14% of the sample reported that a household member had been diagnosed with an NCD. Almost 80% of the sample buy rice from traditional retail outlets (i.e. public markets), and 21% buy rice from modern retail outlets (i.e. supermarkets and hypermarkets). Almost half of the sample frequently consumed premium white rice (PWR) and only 5% frequently consumed brown rice (BR).

We measured consumers’ acceptance of healthier rice types based on the information they received (Table 2). A high acceptance rating was observed for the healthier rice types, with mean ratings ranging from 4.1 to 4.4 on a 5-point scale. The results of non-parametric tests for independent groups show that (1) the mean differences in acceptance ratings between subjects who received the information treatment were not significantly different from those who received the basic information, and (2) there was no significant difference in the mean acceptance ratings for the three healthier rice types.

Consumers’ WTP values ranged from PhP20 to 80 (US$ 0.17 to 1.43) (Figure 2). The mean WTP values of consumers who received the information on the basic product description and examples of dishes that can be prepared/cooked with it were PhP53 (US$0.95) for brown rice, PhP56 (US$0.99) for colored rice and PhP52 (US$0.94) for low-GI rice. The mean WTP values of consumers who received the information treatment (i.e. additional information on the health benefits) were PhP55 (US$0.99) for each of the healthier rice types. Similar to acceptance, the differences in WTP values between rice types were not significant.

Relative to PWR available in the market (with an average price of PhP56 at the time of the survey), consumers discounted the healthier rice types with price discounts ranging from 1.6 to 6.4% (Figure 3). Consumers who received the basic information discounted brown rice by 5.3%, colored rice by 0.8% and low-GI rice by 6.4%. Price discounts were lower when consumers had been informed about the health benefits, which were more pronounced for brown and low-GI rice (discounted by about 2% each).

The descriptive statistics for the variables are summarized in Table 1. A hierarchical approach was used to assess the block effects of the constructs in the conceptual framework (Table 3). The variables for socio-economic characteristics and rice consumption and purchase behavior are the independent variables in Model 1. In Model 2, measures of attitude toward healthy eating (i.e. intention to adopt FBDG and self-reported diet quality) were included in the hierarchical regression analysis. The indicators of diet quality (i.e. HDDS main eating occasions and HDDS snacks) were included in Model 3. And lastly, the information treatment variable was included in Model 4.

In Model 1, the variables past consumption of rice (i.e. PWR and brown rice) and place of rice purchase had no significant effects on WTP. Among the socio-economic variables, monthly household income and employment had significant positive effects on WTP. Household size had a significant negative effect on WTP. In Model 2, these three socio-economic variables remain significant, and the additional variables intention to adopt FBDG and self-reported diet quality had significant effects on WTP. In this stage, the R² improved by 0.03, which is significant and suggests that attitude toward healthy eating explains the additional 3% variation in the model. In Model 3, the indicators of healthy eating behavior had a significant effect on WTP, with an improvement in R² by 0.02. In this stage, the measures of attitude toward healthy eating remain significant. Income and household size remain significant. Employment was no longer significant in Model 3. In Model 4, the information treatment variable was added, which did not improve the overall explanatory power of this complete model.

The complete Model 4 (Table 3) indicated that previous consumption of brown rice had no significant effect on WTP, which contrasts with findings by Bairagi et al. (2021a). Monthly household income remained the only socio-economic variable that has a significant positive effect on WTP, which is consistent with other studies (e.g. My et al., 2018; Cuevas et al., 2016). Age, gender and household composition did not have a significant influence on WTP in the complete model. However, smaller families tended to exhibit higher WTP. The non-significance of the presence of children was not expected as it was reported to have negative effects in previous studies (e.g. Bairagi et al., 2021b; Sriwaranun et al., 2015). Self-reported diet quality had a significant positive influence on WTP. Consumers’ WTP tends to increase by PhP1.4 (US$0.02) for each unit increase on the five-point scale of how they describe the quality of their households’ diets. Similarly, consumers’ WTP tends to increase by PhP2.4 (US$0.04) for each unit increase in their intended frequency of consumption of the food items in the FBDGs. The block effect of HDDS for main eating occasions and for snack occasions was positive, but only the latter was statistically significant. The information treatment did not have a significant influence on WTP. The possible reasons for this non-significant effect, particularly of the health benefits information, need further investigation because it does not corroborate with other consumer studies that have evaluated nutrition-related attributes (e.g. Herrington et al., 2023; De Steur et al., 2016).

The results of our study reveal that consumers generally accepted healthier rice, similar to previous studies that evaluated brown rice and colored rice, although these were sensory studies that measured acceptability based on specific cooking and eating quality attributes (e.g. Choi et al., 2020; Lu et al., 2018). Although consumers accepted healthier rice types, they discounted it relative to PWR available in the market, despite being exposed to generic information on the health benefits offered by these products. Among the three types, it was observed that the lowest price discount is for colored rice.

Consumers discounted brown rice by about 2–5%, and possible explanations for this may be drawn from other studies in Asia. For example, consumers in China considered brown rice inferior to white rice in terms of taste and quality (Zhang et al., 2010). Consumers in Bangladesh discounted low-milled rice, which gives the grains a distinctive light brown color (Herrington et al., 2023). Therefore, a plausible explanation for the price discount for brown rice is that it is perceived by consumers as an inferior version of white rice. The FGDs conducted prior to the survey substantiates this. Consumers described brown rice as tasting stale (i.e. walang lasa and matabang), having a “hard” texture (i.e. matigas) and taking longer to cook (Table 4). Nevertheless, they viewed brown rice as “healthy.” The FGD results also suggest high awareness of brown rice, but this needs further investigation because it contrasts with the survey findings by Bairagi et al. (2021a) and Cabardo and Depositario (2018), which indicated unpopularity of brown rice in the capital city and nearby peri-urban and rural areas. This could be due to the historical role of brown rice in nutrition intervention programs implemented by the government to address (1) vitamin B1 and protein energy deficiencies in the 1970s and (2) the prevalence of NCDs in 2011–2018 (Chiang, 2019). Furthermore, in the early 2000s, the Asia Rice Foundation promoted brown rice to tackle the increasing prevalence of hidden hunger by positioning it as a new type of healthy food. However, this campaign did not create sustained impacts. These imply that brown rice is considered inferior relative to white rice based on appearance and eating and cooking quality attributes, albeit viewed as “healthy.” Therefore, promotion strategies need to emphasize the nutritional benefits of brown rice and mainstream dishes where the different taste of brown rice pairs well with other ingredients. In other words, rather than focusing on substituting brown for white rice in regular dishes, awareness campaigns should focus on promoting the taste and nutrition benefits of dishes based on brown rice (e.g. Cuevas et al., 2017).

Similar to brown rice, the results of the study showed that consumers discounted low-GI rice by 2–6% compared to PWR. The plausible explanations for the price discounts could be that (1) the concept of GI is unclear despite the information treatment and (2) low-GI rice is mainly viewed as suitable for diabetics or people with a family history of diabetes. The pre-survey FGDs indicated that most respondents were unclear about low GI but clearly associated it with diabetes prevention (Table 4). It may be possible that consumers have formed an optimistic bias toward the need to consume low-GI rice, which suggests that they have underestimated their own risk of contracting diabetes compared to others (Klein and Helweg-Larsen, 2002). This is worth further exploring in future studies on low-GI rice.

Consumers discounted colored rice, albeit to a lesser extent than brown and low-GI rice. Among the three, WTP for colored rice was not significantly different from the market price of PWR, and the discount for colored rice was the lowest (i.e. 1%). The distinct visible characteristic of colored rice could be a possible reason for the smaller discounts (compared to brown and low-GI rice), as consumers may have considered it as a “different” type of rice (i.e. not necessarily inferior). Other possible explanations for the price discounts could also be drawn from the pre-survey FGDs, which revealed that respondents had (1) low awareness and (2) mixed perceptions about the cooking and eating quality of colored rice. For example, different classifications of texture were described by the respondents. Some described it as “having a loose texture” (buhaghag), while others described it as “sticky” (malagkit). Some respondents were generally hesitant about its taste and texture. This could indicate an ambivalent attitude or mixed feelings toward colored rice, which has shown to result in a less clear relationship between attitude and behavior (Norris et al., 2019). Strategies to increase product awareness and encourage product trial may be explored to improve consumers’ familiarity with the eating quality of colored rice.

Our regression estimates provide two key insights into the potential target consumers and strategies to promote consumption of healthier rice. First, consumers who have a predisposition toward healthy eating may be more likely to consume healthier rice types, as WTP is influenced by the constructs of healthy diets. Their attitude toward healthy eating and their perception of the quality of their own diet positively influence WTP. This suggests that consumer acceptance and valuation of healthier rice need to be understood and studied in the context of healthy rice-based diets. The current quality of consumers’ diets in the context of diversity in FGs consumed also has a significant influence on WTP. Incorporating healthier rice types within the broader context of healthy eating habits in nutrition education programs could be explored to promote the consumption of healthier rice types (De Steur et al., 2022; Mattei and Alfonso, 2020). And secondly, the significant block effect of HDDS, in general, and the positive effect of HDDS snacks, in particular, also suggests an opportunity for promotion strategies for healthier rice that focus on the eating occasions, in addition to the dishes, following the GSR framework proposed by Cuevas et al. (2017). The different snack occasions could serve as entry points rather than targeting the substitution of white rice in regular dishes for main eating occasions (i.e. lunch and dinner). In other words, a differentiation strategy in terms of dishes and occasions that focuses on the health benefits of their current diet, which is mostly composed of starchy staples, is most effective. Such a strategy may not increase dietary diversity since only a third of the country’s population eats five different food groups (i.e. starchy staple, vegetable, fruit, pulse/nut/seed and animal-source food), which is recommended for daily consumption (Schneider et al., 2023), but rather enhance the health benefits of the current diet.

Rice is the main staple food in Asia, where NCDs are prevalent. It is typically consumed as white rice (i.e. milled and polished). Consumption of white rice can lead to micronutrient deficiencies, especially among consumers with low dietary diversity and whose main source of calories is rice, due to the removal of the bran layer, which contains dietary fiber and essential micronutrients. White rice varieties also tend to have a high GI.

Our study in the Philippines focused on consumers’ acceptance and WTP for conventionally bred healthier rice types (i.e. brown, colored and low-GI rice). Using a contingent valuation approach with a between-subjects design, we found that consumers accepted these healthier rice types but discounted them relative to PWR. Consumers discounted brown and low-GI rice by 2–6%, while the discount was lowest for colored rice (i.e. 0.77–0.90%). Additional information about the health benefits of these products, as a credence attribute of the rice type, did not significantly affect consumers’ WTP for them. Household income, attitude toward healthy eating and household dietary diversity had significant positive effects on WTP. Our findings provide insights for the design of nutrition education programs by promoting healthier rice consumption in the context of a healthy eating habit that should be communicated not only to the consumers who have a predisposition toward healthy eating but also to consumers who are not as health conscious. Our findings also suggest an opportunity for dish and occasion-based interventions to enhance the health benefits of their current rice-based diets (e.g. promoting the nutrition benefits of rice-based dishes and appropriate occasions rather than attempting to substitute white rice in regular lunch or dinner dishes).

Future studies should test nutrition communication strategies that move from a narrow product focus to a broader emphasis on dietary diversity by promoting healthier dishes based on healthier rice products. Subsequent research may also take on the results of the current study to explore different types of nutritional information nudges that can be later considered in the experimental design with the aim of encouraging consumers to eat healthier rice-based dishes ^[2].

Our findings also suggest further insights to increase awareness and acceptance of the specific rice types. Strategies to differentiate brown rice from white rice by focusing on the dish and the nutritional benefits are worth exploring. For colored rice, the results indicate that consumers need to know more about its eating quality. For low-GI rice, further research is needed to identify the information that needs to be conveyed to raise consumer awareness and acceptance (e.g. whether it should be clearly associated with diabetes or with health benefits that reduce the risk of having NCDs that appeal to the general population). As GI is a credence attribute, information and labeling should be carefully considered. Our study provides detailed market intelligence that supports rice breeding programs to incorporate nutrition attributes in rice target product profiles for future market segments for nutritious rice in the Philippines. This proof of concept may be further expanded to other rice-consuming countries in Asia.

This study has limitations that must be taken into account when interpreting the results. The data were collected through an online survey of middle-income class households in the capital city, Manila, limiting the representativeness of our sample for all Filipino consumers and the elderly population, who may have an important influence on food purchase decision-making in the household. Inclusion of secondary urban consumption zones and of broader age populations (e.g. with younger and/or older household members) may be explored in future studies. The design of the study is hypothetical, which may lead to overestimation of WTP values. Future studies could use revealed preference elicitation techniques to compare and validate our findings. Lastly, the current study also relied on perceived grain quality attributes of the healthier rice types. As such, sensory evaluation could be considered in future studies because eating and cooking quality attributes play an important role in consumers’ evaluation of novel food.

We would like to thank all funders who supported this research through their contributions to the CGIAR Trust Fund: http://www.cgiar.org/funders. Funding from the Biotechnology and Biological Sciences Research Council (BBSRC) SuperNutrientRice Project (No: BB/T008873/1), Ghent University Special Research Fund (UGent-BOF, No: 01W06120), the CGIAR Research Program on Rice, the CGIAR Initiative on Market Intelligence and the CGIAR Better Diets and Nutrition Science Program are gratefully acknowledged. The authors would like to thank the editor and four anonymous reviewers for their constructive comments that greatly improved this article. The authors attest originality of the results and assume full responsibility for any remaining shortcomings.

CRediT authorship contribution statement: Marie Claire Custodio: methodology, formal analysis, writing - original draft, writing - review & editing and project administration. Jhoanne Ynion: methodology, formal analysis and project administration. Matty Demont: conceptualization; formal analysis, writing - review and editing, supervision and funding acquisition. Hans De Steur: conceptualization, formal analysis, writing - review and editing, supervision and funding acquisition.

The supplementary material for this article can be found online at: https://doi.org/10.7910/DVN/RMA3NV