Boiled yam end‐user preferences and implications for trait evaluation

Summary This study aimed to establish the quality characteristics of raw and boiled yam by involving stakeholders along the food chain using a methodology that includes a state of knowledge review, focus group discussion and individual interviews, participatory processing diagnosis with processors and consumer testing. Predictive characteristics of yam for producing a high‐ and low‐quality boiled yam were related to morphological or physicochemical characteristics: peeled yam discoloration and mucilage content being negatively appreciated while the ease of peeling, viscous state of cooking water and the ease of breaking yam into pieces positively valued. High‐quality boiled yam should be white or yellowish, sticky to the fingers, nonfibrous, easy to chew, crumbly/friable, with a sweet taste and a good smell. The overall liking of boiled yam is greatly penalised by a too dark colour, hard to the touch, no sweet taste and no friability while eating.


Introduction
Yam (Dioscorea spp) is a major staple food in the tropics and subtropics of Africa, Southeast Asia and the Caribbean regions. It is a source of dietary calories and nutrients and contributes to household income (Egesi et al., 2003;Liu et al., 2007). The most important species of yam in West Africa are white yam (Dioscorea rotundata) and the African complex D. cayenensis-D. rotundata, also referred to as Guinea yam (Loko et al., 2015). The bulk of edible yam is produced in the yam belt of West and Central Africa, which accounts for approximately 97.2% world production with 73.02 million tons produced in 2017 (FAOSTAT, 2019). Benin is the fourth largest producer of yam in Africa behind Nigeria, Côte d'Ivoire and Ghana. Benin leads consumption in West Africa at 395 kcal per capita per day, followed by Côte d'Ivoire (331 kcal), Ghana (314 kcal), Togo (234 kcal) and Nigeria (204 kcal) (Laly et al., 2019).
Yam is consumed in various forms, including products prepared from raw tubers (pounded, boiled, fried or stewed) and processed in dried flour (yam chips). Boiled yam is considered an important food product that can be consumed at all meals and as a snack. It is consumed in the rural and urban areas of Benin as a street food. Boiled yam pieces are prepared by peeling, washing, slicing and cooking, which is performed by immersing the yam pieces in boiling water or by steaming. In the 'yam belt', users select yam varieties depending on their end-use suitability: for pounding, boiling, drying (Zannou et al., 2007;Loko et al., 2015).
Several varieties of yam are grown and used to make different products. Laboko is recognised as the most preferred yam variety for pounded yam (Akissoe et al., 2011), and Kokoro for dried yam flour (Akissoe et al., 2003). For boiled yam, all varieties can, by tentative effort, be used, but all of them are not really acceptable for boiling and consuming (Loko et al., 2015). Some characteristics required for raw yam (tuber size, appearance, freshness, ease to use) ( Barlagne et al., 2017) and boiled yam (white colour, crumbly and sweet taste) have been reported in literature (Naziha, 2002;Egesi et al., 2003;Otoo & Asiedu, 2009;Ranaivosoa et al., 2010), that have conducted individual interviews or focus group discussions, involving either farmers, processors or consumers.
Quality characteristics' demand of crop and product are specific to each type of stakeholders in a food chain. Previous research has developed methodologies that capture information from all food chain stakeholders on many food crops. Dedehouanou et al. (2017) argued that participatory evaluation of maize varieties along the food chain was required to ensure objective information and suggested that spatial and gender representativeness have to be considered in the choice of different users during consultation. A new approach was developed by Fliedel et al. (2016) for better assessing the adoption of new cassava genotypes, in a view of providing information to breeders early in varietal improvement programmes. It involved several successive steps such as qualitative surveys all along the food chain to identify quality criteria of a good cassava crop and product, effective participation of processors to identify the ability of genotypes to make a good product, and a 'all-in-one' method coupling hedonic test, JAR 'Just All Right' test, CATA 'Check All That Apply' questions to assess the acceptability and preferences of products by a large number of consumers. Furthermore, Forsythe et al. (2021) reported an interdisciplinary and participatory methodology to assess user acceptability of root, tuber and banana varieties, with the ultimate aim of informing breeding programmes in trait selection. This study aimed to identify the quality characteristics (morphological, postharvest, physicochemical, technological and sensory characteristics) of raw and boiled yam, required by producers, traders, processors and consumers in Benin, and how they influence varietal preferences.

General approach
For the study, the multidisciplinary methodology as described by Forsythe et al. (2021) was used. This includes a state of knowledge review, gendered food mapping, participatory processing diagnosis and consumer testing, that will produce a list of prioritised quality characteristics for boiled yam. It is expected to translate these quality characteristics into physical or chemical components, for breeders to be used in their breeding programme in view of developing new varieties adopted by food chain stakeholders.

Gendered food mapping in rural communities
The survey was carried out in eight rural communities of central Benin belonging to the Districts of Dassa-Zoum e and Djidja. A survey was undertaken using the multidisciplinary methodology described by Forsythe et al. (2021). Participants were yam farmers, processors/cookers and consumers of boiled yam, who were selected randomly from different localities. In total, sixteen focus group discussions (FGDs) were held (eight men and eight women). In addition, eighty individual interviews including sixty-four community members, eight market leaders and eight community leaders were completed.
Interviews were conducted in French and local languages. The questionnaire focused on yam varieties, processing steps to make boiled yam and quality characteristics of raw and boiled yam. It identified the quality characteristics valued by different user groups, prioritised them by importance and examined differences in by gender and region.

Processing diagnosis of raw yam and quality characteristics
This study was carried out in Bohicon, Zou District, a small urban centre in southern Benin. Six skilled processors with at least 10 years of processing activity were recruited to make boiled yam using their own cooking methods (boiling or steaming). Six varieties selected for their contrasting characteristics, with five Dioscorea rotundata varieties (Laboko, Kodj ew e, Gnidou, Kpaina, and Deba) and one Dioscorea alata variety (Kp et e) ( Table 1). Tubers were harvested 9 months after planting. The six varieties were coded with a random three-digit number and presented to each processor in random order. Each processor received one to three tubers of each variety, weighing between 1.0 and 3.5 kg per variety. Key parameters such as duration, temperature and mass balance were determined at each step. Each processor was asked to describe the quality characteristics (the appearance of raw yam, peeling and cooking ability) of the yam at each step of the process (appearance, peeling and cooking ability of raw yam), and assess the final product (sensory and physical characteristics of boiled yam pieces).

Key processing parameters
The dry matter content (DMC) was determined by oven drying at 105°C to constant weight according to AOAC method (1984). Quantitative processing parameters (duration, temperature, mass balance) were measured at each step of the process. The mass balance was determined at each step according to the method reported by Badoussi et al. (2015).

Consumer testing in rural and urban areas
Consumer testing was carried out with 301 consumers interviewed randomly in different locations: 129 in eight rural communities, fifty-two in Bohicon, a small urban town, and 120 in the city of Cotonou and its neighbourhood. Consumers were 18-72 years old and included 57.8% males and 42.2% females. Five out of the varieties that were used in the previous processing diagnosis were processed into boiled yam pieces by one skilled processor at each location following the same procedure. In order to maintain a warm temperature for the boiled yam samples for the consumer testing activity, batches of boiled yam pieces were made and stored for 5 min in an insulated container (60-65°C). Each boiled yam samples (50-65 g) made from each yam variety were presented consecutively to each consumer in a plastic glass coded with a 3-digit label. Mineral water was available for cleaning palate between samples testing. Consumers were asked to evaluate overall liking of each boiled yam using a 9point hedonic scale (1 = 'Extremely dislike', 9 = 'Extremely like'). The intensity of five quality characteristics identified as important during previous surveys (colour, stickiness in fingers, hardness in hand, friability in mouth and sweet taste) was evaluated using the 3-point 'just about right' (JAR) test (1 = 'TW: too weak' 2 = 'JAR: just about right' and 3 = 'TS: too strong'). Each boiled yam samples were described with a Check All That Apply (CATA) table including fifteen sensory characteristics (slightly yellow, white, dirty white, pinkish, sticky, easy to chew, no fibres, friable/tender, hard in the mouth, easy to break with the hand, slightly bitter, tasteless, sweet taste) and five emotional terms (attractive, good smell of yam, good to eat, bad taste, unpleasant to eat). All the terms were selected from aforementioned surveys and then randomised between and within respondents (Ares & Jaeger, 2013).

Data analysis
Qualitative data were analysed based on verbatim transcription in Excel, followed by descriptive analysis and prioritisation of characteristics weighing (Forsythe et al., 2021). The robustness of characteristics cited by the respondents was checked by revisiting the similarity/synonymy between verbatim words used by respondents or transcribed by different interviewers (variation in vocabulary). Important and robust characteristics were selected and prioritised according to the ranking order or to the citations weighting, where it is necessary. In the case of the citations weighting, if the sum is more than ten, the quality criterion is considered in the definition of the product profile. Where it is appropriate, a chi-square test or Fisher's exact test was performed on number of citations in links with gender.
For participatory processing diagnosis, key quantitative processing parameters (duration, temperature, mass balance) were submitted to descriptive statistics.
The overall liking scores were submitted to variance analysis (ANOVA) and Tukey's test (P-value < 0.05). For CATA data, Cochran's Q test was carried out on the frequency of citations of each characteristic followed by correspondence factorial analysis (CFA). JAR test data were analysed for each characteristic by counting the percentage of respondents who evaluated each boiled yam sample as JAR 'just about right' or TW 'too weak' or TS 'too strong'. Penalty analysis as defined by Gere et al. (2017) was performed to identify potential directions for consumers demand on the basis of the five selected sensory characteristics. JAR and overall liking scores of all boiled yam samples were combined to determine important mean drops in overall liking when the characteristics were cited TW or TS by at least 20% of consumers (Pareto principle) (Meullenet, Xiong et al., 2007). All statistical analyses were performed using XLSTAT version 2015.6.08 software (Addinsoft, Paris, France).

Results and discussion
Food crops prioritised by gender and region: importance of yam compared to other food crops Table 2 shows that the most important food crops grown in the communities were maize, cassava, yam, soya and groundnut. The top three crops were maize, yam and cassava, maize being the most important crop irrespective of gender and region. Starchy crops (maize, yam and cassava) were positioned as priority crops while legumes (source of proteins) such as soya and groundnuts are used mainly as cash-crops, suggesting that the rationale for growing a crop in both communities and for both men and women is above all for household consumption and sale (Tables 2 and  3). Regarding the top three crops, no spatial (regional) difference was observed in ranking. However, across regions, women ranked yam as the second most important crop and cassava as third, and vice versa for men. The first position of maize in both regions and across gender can be explained considering that maize is both a cash crop and important for household consumption. In addition, it can be stored for longer time than any other crop, so it is available for common use at any time (Nago et al., 1997). According to a women's FGD, yam is preferred to other crops because it can be harvested after 6 months from planting and easily stored for home consumption. Men's FGD argued that they prefer cassava because they can sell it all along the year to raise cash, and as reported by Saranraj et al. (2019), cassava provides security against famine mainly because of its staggered harvest.
Laboko, Kokoro and Moroko yam varieties are the most important varieties irrespective of gender and region under study (Table 2), in agreement with previous work (Dansi et al., 1997;Zannou et al., 2007). Laboko was ranked the most important variety by both men and women as it has the best quality for preparing many yam dishes based on processing ability and sensory traits. Laboko was seen as conferring social prestige since farmers stated: Planting Laboko is a mark of respect and this confers to farmers the status of resource person for decision-making that affects the development of the community. Although gender difference was not significant (P-value of 0.15 for chi-square test) as far as all varieties are concerned, we observed that high proportion of women prefer Kokoro for its  The motives/rationale of planting the top five preferred yam varieties can be grouped according to several criteria depending on varieties, region and gender (Table 3): agronomic performance (high yielding, big size, and early varieties), cash-providing, and/or processing ability (good to pound, good to boil). All varieties are used both for sale and for home consumption, in agreement with Zannou et al. (2007) who pointed out that about half of harvested yam is designated for home consumption and the other half for sale. Thus, the choice of variety is influenced by variety type and regional/spatial differences in preference.

Boiled yam processing diagnosis: key processing parameters
Boiled yam is obtained through a very simple process, including peeling, washing, slicing and boiling or steaming the yam pieces until the tuber cores become soft (Fig. 1). Processors prefer boiling yam pieces when the quantities are small for home consumption, and steaming when they have large quantities to sell, mainly because steam cooking better preserves the integrity of yam tubers all day long during selling.
The water-to-yam ratio during boiling was lower than that during steaming, ranging between 0.41 to 0.59 vs. 0.77 to 2.33, respectively. Steaming time was longer than that of boiling, with values in the range of 27.4-38.0 min vs. 17.7-24.1 min, respectively. Dry mater content of raw yam pieces was similar to that of boiled yam pieces, ranging from 25.8 to 43.2 g per 100 g, db. These results are in accordance with previous work that reported a positive relation between dry matter content of raw and cooked yam pieces (Akissoe et al., 2011). These authors pointed out that dry matter content of cooked yam decreased slightly for about 4 g per 100 g (wb) during cooking and can be predicted by the uncooked value.

Raw yam quality characteristics
Quality characteristics of raw yam were collected during surveys and processing diagnosis and then disaggregated by gender (Table 4). They were morphological characteristics (appearance of skin, size and form of tuber, appearance of inner peel) and technological characteristics (skin colour during peeling, viscous exudate/mucilage). In general, women cited more characteristics than men, with more variation within those characteristics. Characteristics associated with high-quality raw yam were cited about twice compared to those indicating poor-quality raw yam. Ten out of twelve characteristics referring to high quality of raw yam are common for women and men, even if they were cited at a different frequency. Surveys and processing diagnosis data revealed that most of the characteristics were exclusively and separately grouped as high-or low-quality types (Table 4), except for rough/shaggy/humped skin, which crosses groups. Test-peeling with fingernails before buying tubers in the market provides reliable information on the raw yam quality, such as colour and water content. Significant differences were evidenced between men and women citations (chi-square P-value < 0.05) for some characteristics such as rough/shaggy/humped peel, Figure 1 Flow diagram for making boiled yam and key processing parameters. brown skin, big tuber and presence of insects (Table 4). Some of the characteristics (big tuber, smooth, presence of rootlets, skin discolouration) matched up with previous work (Naziha, 2002;Otoo & Asiedu, 2009;Barlagne et al., 2017) while others related to the form of tuber (pointed tip, cylindrical tuber, large/convex/flat head) have not been previously reported as quality characteristics of raw yam for making boiled yam. Naziha (2002) reported that appearance of tuber and postharvest storability had high importance in the choice of yam varieties while Otoo & Asiedu (2009) stated that the enzymatic oxidation appearance and colour attractiveness of peeled tubers are other main quality characteristics. Barlagne et al. (2017) reported that the main important characteristics when buying raw yam were lack of external damage, size, freshness and variety type.

Quality characteristics of yam during processing into boiled yam
The quality characteristics of six yam varieties were evaluated during processing. The main characteristics of the preferred varieties for making boiled yam were as follows: raw tuber appearance (absence of rootlets), ease of peeling, stability of white or yellow colour during peeling and cooking, viscosity of cooking water and the ease of breaking the yam piece with a fork during cooking. The main driver of quality cited by processors is the change of colour into red, purple or dark during peeling, washing/slicing and cooking. Our participatory approach revealed that poor quality characteristics of raw yam for boiling were a sticky substance/mucilage observed during peeling, or the lack of viscous aspect of boiling water. Processors grouped yam varieties into high (Laboko and Kodjewe), intermediate (Gnidou, Kpaina, and D eba) and poor quality (Kp et e).

Quality characteristics of boiled yam identified during surveys
Quality characteristics of boiled yam cited by respondents during surveys were classified into sensory (appearance, texture) and emotional (e.g. attractive, unpleasant odour, good aroma/odour, good taste) characteristics (Tables 5). The quality of boiled yams depended on the variety. Poor-quality boiled yams was mainly referred to as a discolouration (red/purple/ black colour) after cooking, poor texture (hard in the mouth/difficult to chew) and/or bad taste/bitter. Highquality characteristics of boiled yam include appearance, texture when touching, texture in mouth and taste. No significant gender association (chi-square Pvalue > 0.05) was evidenced among these characteristics. Prior to consumption, the most important characteristics were a white or yellow boiled yam while easy to break with fork/friable in fingers and sticky were the most important texture characteristics (Table 5). Boiled yam also must have a good aroma and good taste (little sweet/sugary). The ease to chew/ friable in mouth is also considered as a high-quality characteristics for boiled yam. These observations were in agreement with previous work that largely reported the same characteristics of texture, colour, taste, and smell as the major drivers of the consumers' preference for boiled yam (Naziha, 2002;Zannou et al., 2007;Ranaivosoa et al., 2010).

Overall liking of boiled yam during consumer testing
To support the high-and low-quality characteristics cited during the survey, boiled yam made from five of the six varieties used during the processing diagnosis were submitted for consumers testing in eight rural communities, one small town and one city. The overall acceptability of boiled yam significantly differed between the five varieties and the locations (rural and urban) (P < 0.05) ( Table 6). Laboko is the most preferred variety by consumers (score of 7.6, like very much) while Kp et e was scored the lowest (3.7, dislike). Consumers from rural communities gave higher scores for Laboko and Kodj ew e and lower scores for Kp et e boiled yams compared to consumers from Bohicon and Cotonou. The score obtained for Kp et e in the city seemed higher than in the other localities. This indicated that citizens from small town and city seemed to be less able to discriminate the quality differences between boiled yam samples, probably because these consumers are less demanding of specific quality characteristics and are used to consume any types of boiled yam. The Q Cochran's test revealed significant differences in the twenty characteristics used by consumers to describe each boiled yam sample (P-value < 0.05). The most frequent characteristics used to describe the five samples were positive descriptors including attractive, good smell of yam, easy to break with hand, not sticky in mouth, easy to chew, friable/tender, no fibres, sweet and good to eat. Figure 2 depicts a factorial correspondence analysis (FCA) of the five yam varieties and frequency of citations of CATA characteristics selected during surveys and processing diagnosis. Three groups of varieties were established. Boiled yam pieces made from Laboko and Kodj ew e varieties were qualified as high-quality products and were specifically described as attractive (slightly yellow colour for   (Fig. 3) was performed to point out how many scores of overall liking were significantly lost because the characteristic was not evaluated JAR by at least 20% of the consumers. The penalty values of boiled yam ranged from 0.2 to 1.6. The descriptors TS-colour (too strong colour i.e. dark), TS-hardness in hand, TW-sweet taste (too weak sweet) and TW-friability in mouth received the highest mean decreases in the overall liking score (penalty higher than 1.2 and for more than 68% of consumers). The CATA test revealed that the dirty white colour is the main negative characteristic cited by all the consumers to describe Kpaina and Gnidou varieties, but the JAR test showed that this characteristic did not penalise greatly the overall liking. However, both varieties were highly penalised by the TW-sweet taste. This might be explained by the fact that the consumers grant great importance to sweet taste in comparison with dirty colour. This observation was confirmed by the TWsweet taste (not sweet) cited by 22% of consumers for Kodj ew e, but with the lowest penalty value (0.2). The variety Laboko was considered as the best by processors and consumers and was not penalised by any descriptor evaluated in JAR test. In addition, stickiness in fingers was evaluated JAR by at least 80% of the consumers. Stickiness in fingers did not penalise any product and cannot be considered as a driver of liking.

Limitations and suggestions
The results obtained are likely to reflect the survey areas, particularly the consumption habit and the yam varieties cultivated. The number and the quantity of yam samples tested during the process diagnosis are likely insufficient in capturing enough detail on quality characteristics, since skilled women processors were asked to process very different varieties, which will invite them to generate a large number of good or bad characteristics throughout the process. It would be interesting to take into account the variance in sensory quality along the yam tuber (proximal, middle and distal) when distributing samples to consumers.

Conclusions
The quality characteristics of boiled yam motivate the varietal preferences and decisions of producers, processors and consumers. Gender and spatial considerations should be taken into account for yam variety choice/ preferences. Field surveys and the processing diagnosis provided reliable descriptors that were mapped with overall liking scores to better understand the consumers' expectations. These consumer expectations can be met by understanding and paying attention to both of the high-and low-quality characteristics of a variety. Three classes of quality characteristics were identified among the six yam varieties studied. Apart from Laboko, and to a lesser extent Kodjewe, all other yam varieties, in particular Kp et e which obtained the lowest score of overall liking, tested by a large number of consumers were penalised for some quality attributes that need to be addressed to biochemists and yam breeders/geneticists.