Site and participants

The study was conducted in seven communities surrounding the Manombo Special Reserve (MSR) in southeastern Madagascar: Sahamahitsy, Madiorano, Andrafia, Maropanahy, Manombo, Morafeno, and Anivorano. These sites were selected based on logistical convenience (road access) and community willingness to participate.

Stratified random sampling was implemented at both the village and household levels. Stratification occurred first at the village level and then at the household level within villages, ensuring geographic representation while maintaining random selection among eligible households. Within each village, ten households were randomly selected from a list meeting eligibility criterion: permanent residence of at least five individuals, including one adult male and one adult female, and at least one child aged 1 to 10 years. The minimum household size criterion ensured the presence of multiple adults and at least one child, allowing simultaneous assessment of adult perceptions (the primary focus), intra-household sharing, and children’s reactions as a secondary group of interest. This design was essential for understanding adoption dynamics. Exclusion criteria included refusal of consent or reported allergies/intolerances to insect- or crustacean-based products, as determined during face-to-face screening. The survey was administered over several weeks, with participants sampling a different product each week and providing ratings.

A total of 71 and 73 households participated in the pre- and post-consumption surveys, respectively. Overall, 127 adult participants were recruited for surveys and FGDs (Table 1); all participants who completed the pre- and post-surveys were invited to join the FGD without exclusion criteria in place. Data were recorded from 67 children across all villages, with adult respondents reporting children’s preferences in surveys. Adult participant self-reported ages ranged from 16 to 101 years, of whom 55% were women. The complete participant flow, including recruitment, attrition, waiting-list replacements, and analytical sample sizes, is presented in Figure 1.

During the study sensitization phase, one pre-selected village was replaced by a demographically comparable village after all households declined to provide consent, citing traditional beliefs concerning insect consumption. This replacement happened before the study was initiated.

Tested products

Participants evaluated three novel products: dehydrated whole BSF larvae (whole larvae), BSF larvae powder (BSF powder), and crackers containing BSF larvae powder (BSF crackers), representing non-processed, semi-processed, and processed food items, respectively. A white bean soup (WB) served as a control; WB is familiar to communities, and is a protein-containing, non-insect food with comparable serving style, in combination with rice or cassava, allowing contrast between novelty and habitual diet. BSF crackers were instructed to be eaten as a stand-alone snack. All products were sourced from a local producer adhering to experimental standards and underwent microbiological analysis by an accredited laboratory prior to use.

Data collection

Five structured questionnaires were administered face-to-face to adult participants at their households. A pre-consumption questionnaire collected demographic information, neophobia levels (Pliner and Hobden, 1992), protein source consumption habits, food beliefs and attitudes (adapted from Ruby et al., 2015; with Likert scale), and visual rankings of edible insects. Four post-consumption questionnaires assessed product acceptance (with 5-point hedonic scale – ‘dislike extremely’ to ‘like extremely’), symptomatic experiences, and shifts in food beliefs and attitudes after consuming the three test foods (adapted from Barton et al., 2020; with 5-point Likert scale – ‘strongly disagree’ to ‘strongly agree’). Children’s product preference (with Yes/No Liking response) was assessed based on caregiver observation of one child per household after tasting all products, including the control. A binary response format was used for children because multi-point hedonic scales are not developmentally appropriate for the age range studied (1-10 years), and caregivers reported observed behavioural indicators (consumption, refusal, or requests for more) rather than subjective ratings.

Questionnaires were developed in English, translated into the local dialect, and administered by trained interviewers who had previously engaged with community members. Surveys were conducted in July and August 2024, each lasting approximately 45 minutes. Data collection utilised an application developed in CommCare HQ (2024, Dimagi).

Fourteen gender-segregated Focus Group Discussions (two per community) were conducted in August 2024 in private settings. Gender segregation was implemented to create comfortable discussion environments and reduce potential social inhibition, particularly when discussing food taboos and household decision-making. Discussions, lasting 35-55 minutes, were audio-recorded with participant consent. Each FGD included 4 to 11 participants (average = 6.9) who were also survey respondents. The FGDs were designed to complement the questionnaires, hypothesising that factors such as cultural taboos or neighbourhood influence might contribute to product rejection.

Focus group discussions (FGDs) were conducted in room condition and using a semi-structured format guided by a pre-designed topic guide, after being piloted with local technicians and trained enumerators fluent in the local dialect who would assist with data collection. Each session lasts between 60 and 90 minutes. The main topics covered included: the criteria participants used to make BSF larvae-based products enjoyable; community perception of this new food compared to conventional protein sources and other insects; social influence on participants after consumption; and willingness to farm BSF in the future. Moderators emphasised open participation. The first author, a Malagasy female post-doctorate researcher in food science specialising in food behaviour, supervised data collection, established rapport with participants, and subsequently moderated the FGDs, assisted by a native technician for translation and note-taking. A third researcher from MBC observed for project continuity. Only data from participants who completed all pre- and post-consumption surveys were included in attitude and belief comparisons.

Data analysis

Survey data analysis began with frequency counts for each variable. Enumerators were trained to standardise interpretation. ‘Frequent’ referred to consumption at least once per week, while ‘moderate’ indicated occasional or about once every couple of weeks. ‘For occasional periods only’ referred to festive period; after receipt of salary (once a month). Pearson’s chi-square test assessed significant differences across the tested products. Means and standard deviations were calculated for Liking and Likert scale data. One-way ANOVAs were applied to Liking and Likert scale data to examine variances between groups, followed by the Least Significant Difference (LSD) post hoc test to identify specific significant differences among treatment means. Student’s t-test compared pre- and post-consumption data. Symptomatic experience data were analysed using Factorial Correspondence Analysis to characterise the symptomatic profile of each product. All visual charts and Factorial Component Analysis were generated with R 4.4.1; while statistical analyses were performed using XLSTAT 2023 (Addinsoft) with significance set at $p\mathrm{<}0.05$. For simplicity, percentage results are rounded up or down to the nearest whole number.

Audio-recorded FGDs were transcribed verbatim in the original language and translated into English. Independent thematic analysis was conducted by the first author and a trained research assistant involved in data collection. Both analysts coded transcripts separately using iterative coding of transcriptions and field notes. Emergent themes were compared, discrepancies were discussed, and final themes were agreed upon through consensus between analysts. FGD results are presented using verbatim quotes translated into English alongside thematic analysis.

Ethical approval was obtained from the Malagasy Academy Ethical Board (No. 2024-65/AM/CMEST-P/). Participants provided written and oral informed consent regarding the objectives and motivations of the research. Confidentiality was maintained through the anonymisation of data.

Protein food consumption practices

As for protein food consumption patterns, leafy vegetables were the most frequently consumed protein-rich foods, at least once a week (98%), with saltwater fish being frequently consumed (80%). Freshwater fish and legumes were moderately consumed (54 and 47%, respectively). Animal-source proteins such as pork, beef, and insects were rarely consumed (1, 2 and 6%, respectively) (Figure 2). In terms of origins, while beef and pork are usually purchased at markets (open-air markets), leafy vegetables, eggs, and chickens are home-produced by households (Figure 3).

Beliefs and attitudes towards insects as food: pre-consumption results

Before consuming BSF products, 92% of participants reported having eaten insects, while only 2% had consumed insect powder. Of those who had eaten insects, 90% reported liking the taste, and 88% did not consider insect food disgusting.

Of the 8% of participants who had never eaten insects, 6% were willing to try, while 2% reported that the thought of eating insects made them feel sick. Among all participants, 88% considered insects (in any form: plain, powder, chopped) edible. 97% of participants reported that their neighbours and friends ate insects. 95% would try a dish made with insects, while 60% confirmed they would try insect larvae.

More than half of the surveyed participants believed that insect consumption offers significant advantages, with 65% citing health benefits, 50% recognising their nutritional value, and 58% acknowledging their contribution to food diversity. Conversely, a minority (9%) expressed concerns about potential negative effects such as microbes, poisons/toxins, and allergies, with 90% of these specifically highlighting allergies as a concern.

When comparing insects to traditional animal protein sources, a substantial majority (91%) perceived insects as being more cost-effective, and 32% considered them more abundant and readily available.

Regarding the perception of environmental impact, 61% of participants believed that insect farming is beneficial for the environment. Furthermore, there was overwhelming support for further research and programmes related to insects as a food source, with 87 and 87% of participants, respectively, endorsing these initiatives.

Participants visually assessed and ranked five different insects (grasshopper, cricket, beetle, grub, and BSF larvae) based on their visual perception before tasting any insect-based products. A weighted ranking system was used in which the first choice received five points, the second choice four points, the third choice three points, the fourth choice two points, and the fifth choice one point. The total preference score for each insect was calculated by multiplying the number of participants selecting each rank by its corresponding weight. The grasshopper received the highest preference score, followed by the cricket, beetle, and grub, with the BSF being the least preferred. Notably, BSF, being unfamiliar to participants, ranked lowest, suggesting that prior exposure and familiarity significantly influence consumer acceptance of insect-based foods (Figure 4). Higher scores indicate greater visual appeal.

When presented with images of insect powders, participants expressed a clear preference for lighter-coloured powders, with white being the most preferred, as opposed to darker shades (Figure 5). The majority of participants (96%) indicated their willingness to consume a salted snack that incorporated insect powder.

Comparison of attitudes and beliefs before and after consumption of tested products

On the Food Neophobia Scale (range 10-70), high neophobia is defined as a score of 35 or higher, and moderate neophobia as a score of 25 to 35 (Pliner and Hobden, 1992). Overall, participants had a score of 33.78 ± 1.22 (mean ± SD), indicating moderate food neophobia.

After consuming insect products, regarding positive perceptions, no significant change was found in perceptions of health benefits ($p=0.060$) or nutritional value ($p=0.099$). However, an increased awareness of insects contributing to food diversity ($p\mathrm{<}0.001$) and environmental benefits ($p=0.024$) was found (Table 2).

Negative perceptions changed significantly, with a significant reduction in perceptions of insects as disgusting ($p=0.037$) and inedible ($p=0.014$). Concerns about microbes and toxins decreased post-consumption ($p=0.033$) (Table 2).

Adult consumer acceptance and preference for tested products

Statistical analysis using ANOVAs revealed a consistent pattern for all criteria (overall, colour, visual appearance, smell, taste, aftertaste, flavour and texture liking). Post-hoc comparisons revealed three statistically distinct groups, consistent across all eight sensory criteria ($p\mathrm{<}0.001$): white bean soup and BSF crackers did not differ from each other and received the highest mean liking scores; BSF powder scored significantly lower but remained above the midpoint, indicating moderate acceptance; and whole BSF larvae scored lowest, falling below the acceptance threshold (Table 3).The results from the preference ranking test aligned with previous findings, revealing a substantial difference in the overall preference rankings of the four products, as evidenced by the Friedman test (${\chi }_{\mathrm{calculated}}^2=238.42$, ${\chi }_{\mathrm{theoretical}}^2=7.81$). Further analysis using an ad-hoc test on the sums of preference ranks indicated that all pairs of product rank sums differed significantly. Based on these results, the definitive order of product preference, from highest to lowest, was established as follows: white bean soup, crackers, powder, and larvae.

Perception and behaviour among children

The majority of children responded positively to the tested products. 95% of the children expressed a liking for the BSF crackers and the WB, while 79% liked the BSF powder. The BSF larvae received a lower approval rating, with only 59% of the children indicating they liked it. A chi-square test compared the percentages of those who liked and disliked the BSF larvae and found no significant difference, nor a significant association between gender and product liking (${\chi }^2=1.20$; $p=0.750$). Additionally, no significant link was found between the age of the children and their product preferences (${\chi }^2=7.02$; $p=0.850$).

Regarding the amount of each product consumed by the children, 84% of the children consumed all of the products that their parents shared with them. However, a chi-square test revealed a significant association between the specific product and the quantity consumed (${\chi }^2=22.16$; $p\mathrm{<}0.001$). While children consumed all of the control food and crackers, only 66% of children ate all of the larvae, while 80% of children ate all portions of the powder preparations.

Following consumption, 72% of the children requested additional portions of the products, with a significant association between the willingness to consume more and the specific product (${\chi }^2=53.95$; $p\mathrm{<}0.001$). Children were more likely to ask for more of the control food (white bean soup) and crackers, while only about half of the children expressed a desire for more larvae or powder.

A chi-square test revealed a significant association between the children’s reactions and the specific product (${\chi }^2=62.07$; $p\mathrm{<}0.001$): control food and crackers were generally met with desire, while powder was less desired. The reactions to larvae were mixed, with a combination of desire, disgust, and neutrality.

Finally, the study examined whether the products were consumed on all three days of the test. A chi-square test showed a significant association between product consumption on all days and the specific product (${\chi }^2=27.97$; $p\mathrm{<}0.001$). 95% of children consumed the control food and 95.23% of them consumed crackers on all three days. However, 40% of the children did not consume the larvae on all three days of the test.

Post-consumption symptoms in adults

Among the participants, 62 and 73% reported that white beans provided a sensation of energy and satiety, respectively. In contrast, these sensations were reported by 30 and 26% of participants consuming BSF crackers, 22% for both sensations regarding BSF powder, and 19 and 13% with BSF larvae. Additionally, 15 and 14% of participants reported experiencing nausea with BSF larvae and BSF powder, respectively, while 9% reported stomach aches with BSF powder. Importantly, none of the products caused illness, allergic reactions, or diarrhoea (Figure 6).

When asked about their willingness to incorporate BSF into their diets and to produce BSF in the future, on average, participants reported that they would be willing to incorporate approximately one and half a spoonful of BSF powder (a mean of 6.7 g, assuming 5 g per spoonful) into a single dish for their own consumption. This figure represents a hypothetical intended amount reported during the post-consumption survey, not a measured intake. Furthermore, 90% expressed a willingness to produce BSF in the future.

Focus group discussions

The results of the FGDs are organised into the following themes: prior experience with insects and general attitudes, willingness to try insects in different forms, perceived benefits and concerns, product preferences and sensory acceptance, children’s reactions, sensory symptoms and reactions post-consumption, social influence and community perception, and willingness to incorporate and produce insects. The main results of the FGD are summarised in Table 4 below. To provide more substantial evidence about the mechanisms that contextualise the process through which changes took place between pre- and post-consumption, we explore longer quotes and sections of the dialogues.

Following is a deeper exploration of the dialogues, where participants are coded as: PX-00, where X refers to the Participant ID in the order they first spoke, and -00 refers to the FGD where the conversation took place.

Prior experience with insects and general attitudes: FGDs generally corroborated survey findings, indicating an absence of cultural taboos regarding insect consumption across participant groups. While some acknowledged taboos against consuming other animals within specific subgroups, these prohibitions did not extend to BSF.

1. ∙ Facilitator 1: ‘The kids weren’t afraid?’
2. ∙ P1-01: ‘There was already something similar to it they’d seen, so they weren’t scared of it anymore.’
3. ∙ P2-01: ‘I don’t think anyone here would have a taboo against eating them. They might not like eating them, but they’re not forbidden.’
4. ∙ Facilitator 2: ‘Have you ever heard of anyone who considers it taboo to eat crickets? Maybe someone from a different village?’
5. ∙ P3-01: ‘No, no one here considers it taboo to eat. We harvest it when the season comes.’

Willingness to try insects in different forms: FGD comments consistently demonstrated participants’ willingness to try dishes incorporating insects, showing general enthusiasm, particularly when BSF products were combined with familiar staples like rice or cassava. Even individuals with initial reservations about whole larvae or powder expressed openness to consumption under certain conditions, such as understanding the product’s origin or when presented in a more culturally familiar format.

1. ∙ P4-14: ‘I liked it with rice, though. The powder, I don’t like eating it raw! But when it’s crushed like this, people seem to like it. The kids in my household really enjoyed it. The powder was very delicious.’
2. ∙ P5-14: ‘In our household, I think only three kids liked the larvae. Most of the adults didn’t eat it, or rather, they tried it, but they didn’t like it. It gave them goosebumps. We all liked the powder, though, and we ate it with rice!’

Perceived benefits and concerns: The benefits identified in survey responses, such as health and nutritional value, were frequently echoed by focus group participants. The ‘fatness’ of the larvae was often cited as an indicator of nutritional richness. Concerns expressed in the survey, including potential allergens or microbes, were also reflected in the FGDs through reports of mild adverse symptoms (e.g. itchiness or stomach ache from the powder) and sensory aversions (e.g. undesirable smell or visual resemblance to maggots or latrine larvae). These concerns were predominantly associated with the product’s appearance and unfamiliarity, supporting the quantitative finding that prior exposure significantly influences acceptance.

1. ∙ Perception of health and ‘fatness’:
   1. – P6-08: ‘I mixed it with cassava, and I ate it like I would eat spice. When I ate rice, I poured it into a small spoon. We tried pouring it onto our rice in the beginning, but we felt it was very strong. The day after it was distributed, we used our hands and ate it with rice. I felt it was good for my body. Some people said it was fatty. And full of nutrients. Some people even said that we became more in shape after eating it.’
   2. – P7-08: ‘We liked it in general because our body remains healthy after eating it, so that’s why we like it.’
   3. – P8-08: ‘Same as what they said, it tastes fatty when you try it.’
2. ∙ Sensory aversion:
   1. – P4-03: ‘We just loved it [crackers], the taste, the texture, everything. But the recent one, we didn’t really like. Also, the smell was quite strong.’
3. ∙ Unfamiliarity:
   1. – P8-10: ‘There will be changes if we see the source of it. If we see the source of these things, we might like it.’
   2. – P9-10: ‘I think these things must taste better than the things you’re asking us to compare them to, but the problem is that our minds associate them with bad things. If we see the origin of these things, we might like them more than anything else.’

Product preferences and sensory acceptance: FGD feedback strongly supported the quantitative ranking of product preferences, which identified BSF crackers as the most preferred and whole larvae as the least liked. Crackers were consistently praised for their taste and crunchiness, indicating high acceptance across all demographics. BSF powder received variable acceptance, with some participants enjoying it while others reported rejection due to its odour or dark colour. Whole larvae were frequently rejected in FGDs, primarily due to visual disgust, unpleasant smell, and perceived difficulty in swallowing, consistent with the low acceptance scores observed in the quantitative analysis.

1. ∙ Crackers:
   1. – P5-03: ‘We really enjoyed the crackers. My kids couldn’t stop eating them; they really, really loved them and kept asking for more.’
2. ∙ Powder:
   1. – P3-04: ‘The powder was our favourite and then the cracker but we did not really like the larvae! We finished all three of them.’
3. ∙ Whole larvae:
   1. – P6-07: ‘All of it. My stomach can’t handle it. I only tried the crackers. Everyone, including my wife and kids, liked it! They also did not like the larvae.’
   2. – P7-07: ‘Personally, me and my kids liked it! But the larvae, we didn’t like even seeing it!’
   3. – P4-02: ‘The only small problem was that it passes through the space between our teeth, for those of us who are missing teeth.’

Children’s reactions: FGD data reinforced quantitative findings regarding children’s product acceptance, confirming high approval for crackers and the control soup, with lower acceptance for the powder and larvae. Positive reactions from children to the crackers were emphasised, while larvae elicited mixed or negative responses, often related to appearance and texture. Parents frequently noted that their children’s trust in the food’s perceived health benefits increased their willingness to try it.

1. ∙ P6-05: ‘In our case, all three types of it were appreciated by the kids and others; they couldn’t even wait for the turn to distribute it. I even realised it brought changes to the kids’ bodies ever since they ate it.’

Sensory symptoms and reactions post-consumption: The FGDs mirrored survey findings that a small number of participants reported minor symptoms associated with BSF powder and larvae. Isolated reports of stomach ache and itchiness were primarily linked to the consumption of these two forms. Crucially, no participants in the focus groups reported significant reactions or illness.

1. ∙ P8-03: ‘It feels like you’re adding caterpillars to your meals when you eat it… and we were afraid to eat it. I was scared to see it. When I took it in my hand, I got goosebumps, and I never thought of it as something edible. I really liked the crackers, but I felt stomach pain after eating them. As for the powder, no one in my household liked it, including me.’

Social influence and community perception: Participants reported that neighbours displayed curiosity and sometimes envy regarding their participation in the study. FGDs elaborated on this social context, indicating that despite some negative rumours, trust in the project and its collaborators fostered continued consumption. Participants also shared products with guests or neighbours, who reportedly reacted positively.

1. ∙ P8-05: ‘It’s quite difficult to talk about food with someone nowadays! You want to motivate people when you share it with them. They’re also jealous of us receiving it, so they didn’t get it. Surely we’ve heard it [negative rumours], but we still continue to eat it. They said we’re eating nonsense and that it will kill us one day.’
2. ∙ P9-05: ‘Yeah, they say a lot of things, but we don’t care.’
3. ∙ P10-05: ‘Yeah, they say we’ll die from our curiosity.’
4. ∙ P11-05: ‘But we trust that you won’t do that to us.’
5. ∙ P12-05: ‘Yes, I heard that [Some people say it causes infertility] from people in Ankazotsararavina. They say that this is family planning. I told them that even if they wanted to do that to us, I already have many kids.’
6. ∙ P13-05: ‘Some friends come and ask! And we told them that it’s good!’
7. ∙ P14-05: ‘They envy us; they also want to have it! They want to eat it also!’
8. ∙ P15-05: ‘Yes, they’re right; people are curious to try it too!’
9. ∙ P16-05: ‘I don’t need to go to their home, but they come here asking, and I answer them! Because it is an edible thing, and I give whatever is available to me to my guests! What they do is actually envy us because they also want to benefit from it.’

Willingness to incorporate and produce insects: FGDs reinforced the quantitative findings of high willingness to incorporate BSF powder into meals and to engage in BSF production. Several participants emphasised that their openness would increase if they had a clearer understanding of the production process for the powder or larvae, suggesting that enhanced education and transparency could further boost acceptance.

1. ∙ P17-05: ‘I’m ready to farm it.’
2. ∙ P18-05: ‘As long as there’s training on how to farm it, I’m ready.’
3. ∙ P19-05: ‘I told my friend from the beginning of this that I’m ready, as long as you won’t kill me, I’m ready!’
4. ∙ P20-05: ‘Just go ahead and send us the technician because we’re ready.’
5. ∙ P21-05: ‘I will make sure to eat first and then sell later. It will be easy to sell if people see the evidence on me and my family.’

While overall patterns of acceptance were similar between men and women, women more frequently emphasised preparation practices and children’s reactions, whereas men more often discussed production and economic opportunities. Similarly, despite the large variability in participant numbers across FGDs, with some groups exceeding the optimum size (a maximum of 8 participants), thematic saturation across groups remained consistent.