Effect of three types of oils and their level of incorporation on sensory quality of sorghum cookies

Lipid (oil and fat) is one of the basic and important components used in the production of cookies. It plays several roles in their composition and a judicious choice of this ingredient makes it possible to have cheap, delicious and very nutritious cookies. This study aimed to evaluate the effect of refined cottonseeds oil (RCO), refined palm oil (RPO) and red (or crude) palm oil (CPO) on sorghum (Sorghum bicolour var. Gampela) cookies. Oils were incorporated at levels of 16%, 20% and 24%. Cookies quality were evaluated based on quantitative sensory profile, sensory acceptability, preference classification and paired-comparison test. The results of sensory profile tests showed that the sensory descriptors ranged from very low to medium intensity (1.33 to 3.83 on a scale of 5). Both acceptability and classification tests showed that cookies produced with 20% of refined oil (RCO and RPO) and those with 16% of CPO were the most preferred and RCO and RPO with 20% of the oil were appreciated as control cookies produced with 20% of margarine. The colour, texture and fat impression had most influenced the consumers’ preferences. RCO, RPO and CPO exhibit the potential to be used as substitutes to margarine to produce sorghum cookies.


Introduction
Undernutrition remains a major issue in the world, particularly in developing countries, being an underlying cause of 45% of child mortality (Abubakar et al., 2012;WHO, 2020). In developing countries, nutritional deficiencies are still major causes of death and disease, especially among vulnerable and socially disadvantaged people like women and children less than five years of age (WHO, 2020). It is estimated that up to 219 million children are deficient in vitamin A (Mason et al., 2005) and over 1 billion people are susceptible to zinc deficiency in developing countries (Black et al., 2008). Burkina Faso is one of the developing countries showing a high prevalence of undernutrition. In this country, it is estimated that 360,048 children aged 6-59 months will suffer from acute malnutrition during 2020 (Ministère de la santé, 2020).
Cookies are dry foods made from cereal flour, sugar and fat with other minor ingredients such as milk, salt, flavouring agent and eggs (Manley, 2011;Mancebo et al., 2015;Devi and Khatkar, 2016). These are highenergy-density foods with a long shelf life that vary from three months to several years (Mancebo et al., 2015). Oil is known to be one of the crucial cookies' components affecting its shelf life (Manohar et al., 1999;Mancebo et al., 2015). The oil quality including its origin and concentration is so important for young children (Manohar et al., 1999;Moreno et al., 2014). The quality of lipids in cookies processing determines the quality of the final product including the taste, texture, shelf life and cost (Zoulias et al., 2002;Jacob and Leelavathi, 2007;Devi and Khatkar, 2016).
In Burkina Faso, various oilseeds are used in lipids processing, including cottonseeds, peanuts, shea nuts, sesame seeds, soybeans and cashew nuts (FEWS, 2017;Sawadogo et al., 2020). Indeed, cottonseeds oil is the main vegetable oil produced and the most available. The annual production, in Burkina Faso, was estimated to 125 000 t (Ouédraogo, 2014;Sourabié et al., 2019). In Africa, Burkina Faso with Mali is the first cottonproducing countries. This makes cottonseeds oil a stable and available source of vegetable oil (FEWS, 2017;Sourabié et al., 2019). Alongside cottonseed oils, other oilseeds include palm oil (refined palm oil and red palm oil) largely used in Burkina Faso. Red palm oil was promoted as a source of vitamin A for women and children in central-north Burkina Faso (Zagré et al., 2002). Among children, the risk of inadequate vitamin A intake dropped from 87 to 60%, based on dietary intakes of vitamin A-rich foods, and the prevalence of low serum retinol concentrations (≤0.70 µmol/L) decreased from 84.5 to 66.9% (Zagré et al., 2002). Also, palm oil is the most widely used vegetable oil in the world because of its competitive price, naturally excellent oxidative stability, unique solid content profile, high nutritional value, free of trans fatty acids and cholesterols, antioxidant properties (May 2013;Noor-Lida et al., 2017). Although there are various oils in Burkina Faso including palm oil and cottonseed oil, margarine remains the most commonly used oil in the production of cookies (Songre-Ouattara et al., 2016;Songre-Ouattara et al., 2017). The use of this lipid is associated with its properties adapted for cookies production such as good texture, plasticity, high melting point and oxidation stability (Tarancón et al., 2013;Louiza and Hafsa, 2017;Costa et al., 2019). However, margarine is imported into most parts of Africa including Burkina Faso and is consequently not always accessible to the general population (FEWS, 2017;Sourabié et al., 2019). In addition, some components of margarine including trans fatty acids can be an underlying cause of human health disorders (Dhaka et al., 2011;Costa et al., 2016;Louiza and Hafsa, 2017;Costa et al., 2019). Due to the characteristics of industrial trans fats (texture, plasticity and high melting point) (Tarancón et al., 2013;Costa et al., 2016), it is a challenge to develop cookies using other lipid sources (Costa et al., 2019).
In line with dietary diversification to promote the consumption of nutrient-rich foods, there is a need to investigate other processing cookies using local raw material (Yeh et al., 1998). However, the production of cookie with cottonseeds oil and palm oil, as the oil used locally, might impact the sensory quality of products, such as colour, odour, aroma, texture, flavour, fat impression, bitter aftertaste and overall acceptability.
This study aimed to evaluate the effect of RCO, RPO and CPO on the sensory properties of sorghum cookies. RCO, RPO and CPO were selected because of their availability, acceptability, inexpensive cost and their essential fatty acids and carotenoid content. These fats could thus replace margarine in the production of cookies.

Raw materials used in the production of cookies
Sorghum (Sorghum bicolour Var. Gampela) grains were purchased in Ouagadougou (Burkina Faso) and processed into flour after shelling and milling at the Food Technology Department (DTA) of the Research Institute for Applied Sciences and Technologies (IRSAT) in Ouagadougou. RCO, RPO, CPO, margarine, sugar, powdered milk, eggs, salt, vanilla sugar, baking powder and maize starch were purchased from a food trade in Ouagadougou.

Process of cookies production
The cookies were produced by adapting the process developed by Songré-Ouattara et al. (2017) (Figure 1). Only, the fat and flour contents varied, but the total percentage was kept constant to 72.9% for each cookie (Table 1). For each type of fat, three levels of incorporation were used: 16%, 20% and 24%, resulting in 9 cookie formulas. A control cookie was produced with margarine at the level of 20%. The cookies formulas codification was given in Table 2. The cookies were produced at the workshop of the Food Technology Department (DTA) of the Research Institute for Applied Sciences and Technologies (IRSAT) in Ouagadougou. After production, the cookies were packaged in a plastic bag and hermetically sealed using a thermo-sealing machine. They were then stored at room temperature (around 28°C) and protected from light for further analysis.

Determination of sensory attributes of cookies
The sensory attribute includes sensory profile, Acceptability test, Ranking test and paired-comparison test. To perform these determinations, a total of 11 panels was done with 48 experienced tasters. During each panel, three or four different cookies were coded and placed on a plate using a random table (Cochran and Cox,1957).

Sensory profile
The sensory profile of cookies was determined according to the model described by Stone and Sidel (2004). A quantitative assessment of 16 descriptors was made on a 5-point scale ranging from 1 (very low intensity perceived) to 5 (very high intensity perceived). Definitions of all sensory terms are given in Table 3. Thus, each taster was provided with cookies of the same oil at different levels of incorporation, together with water to rinse the mouth after each cookie. The tasters reported their assessments on a data collection sheet.

Acceptability test
The assessment of the acceptability of cookies was determined according to the process described by Meilgaard et al. (2007) through the following sensory descriptors: colour, aroma, odour, texture, taste, fat impression, aftertaste and overall acceptability. A hedonic scale from 1 to 5 was used to assess each descriptor (Table 4) (Meilgaard et al., 2007). The three cookies (16%, 20% and 24% of oil) coded in a tasting tray were brought to each taster simultaneously. They Descriptor Definition Scale of appreciation (1-5) Colour 1 = Very low intensity, 2 = Low intensity, 3 = Medium intensity, 4 = High intensity and 5 = Very high intensity brown The intensity of brown colour perceived by tasters yellow The intensity of yellow colour perceived by tasters Odour caramel The intensity of nasal sensation of caramel smell perceived by tasters vanilla The intensity of nasal sensation of vanilla smell perceived by tasters palm oil The intensity of nasal sensation of the smell of palm oil perceived by tasters Aroma caramel The intensity of retro-nasal sensation of caramel aroma perceived by tasters vanilla The intensity of retro-nasal sensation of vanilla aroma perceived by tasters palm oil The intensity of retro-nasal sensation of Palm oil aroma perceived by tasters Texture granular The intensity of granular texture in the mouth perceived by the tasters founder The intensity of the founding texture in the mouth perceived by tasters crunchy The intensity of the crunchy texture in the mouth perceived by tasters Taste sweet The intensity of sweet taste in the mouth perceived by tasters salty The intensity of salty taste in the mouth perceived by tasters bitter The intensity of bittersweet in mouth perceived by tasters Fat impression The intensity of fat in the mouth perceived by tasters Bitter aftertaste The intensity of bitter aftertaste perceived by tasters Table 3. Definition of descriptors used to assess the sensory profile of the cookies produced Cookies produced with refined palm oil RPO16 RPO20 RPO24 Cookies produced with red palm oil CPO16 CPO20 CPO24 Control cookies produced with margarine M20 were then asked to complete the acceptability sheet.

Ranking test
The ranking test made it possible to classify cookies in order of preference according to the method described by Watts et al. (1991). Two series of ranking test were performed and the cookies were always ranked from the least pleasant to the most pleasant. For the first series of ranking, the three cookies formula of each oil were presented simultaneously to each taster for ranking. And then, the most pleasant cookies of each oil were selected with the control cookies to proceed for the secondranking test. These four cookies were presented simultaneously to each taster for ranking.

Paired-comparison tests of cookies
The paired-comparison test allowed the comparison of each of the most pleasant cookies of each oil to control cookies following the model described by Stone et al. (2012). Then, the four coded cookies were brought to each taster simultaneously. Each taster was then asked to note on a degustation sheet if there were any differences between the most pleasant cookies of each oil and the control cookies M20. A comparison scale of four levels were used, of which: 1 = no difference, 2 = slightly different, 3 = different and 4 = very different.

Statistical analysis of cookies
Data were keying using IBM SPSS Statistics 21 software and descriptive analysis were generated. The sensory profile and acceptability data were statistically treated using a one-way analysis of variance (ANOVA) with the XLSTAT software, version 2016.02.27444. The means were compared by Tukey test at a significance level of 5%. For the acceptability test, the Pearson correlation coefficient was generated with R software using the package GGplot2 (Team, 2016). This made it possible to determine the interactions between the different descriptors studied. The Liker diagram of the overall acceptability of cookies was also generated with the R software. To compare the ranking test results, the Friedman test was used, followed where appropriate, by rank-sum multiple comparison tests. Finally, the diagrams of spider were made from the results of the ranking and paired-comparison tests with the excel software, version 2016.

Presentation of cookies
After production, ten samples of cookies were obtained: three cookies for each of the free oil type (16, 20 and 24% of incorporation) and one of control cookie produced with margarine. Figure 2 shows the two faces of the different cookies produced.

Sensory profile of cookies produced
The results of the assessment of the intensity of the different descriptors are shown in Table 5. The brown colour was assessed on all cookies, while the yellow colour was assessed on cookies produced with red palm oil (CPO16, CPO20 and CPO24). For the brown colour, the cookies showed three levels of intensity, i.e., very low, low and medium. Cookies produced with 24% of oil (RPO24, RCO24 and CPO24) had consistently the highest brown colour intensities. For cookies produced with refined oils, those with 24% of the oil were significantly different (p<0.05) to those produced with 16 and 20% of oil. For cookies CPO20 and CPO24, the intensities of the brown colour were not significantly different (p = 0.073). For the yellow colour, the cookies CPO16 has a higher intensity and differs significantly from cookies CPO20 and CPO24 (p<0.0001).
For caramel odour, the cookies RPO24, RCO24 and Fat sensation preference in the mouth 1 = very fatty, 2 = fatty, 3 = neither fatty -nor dry, 4 = dry and 5 = very dry For caramel aroma, cookies RPO24, RCO24 and CPO24 had the highest intensities in each type of oil cookies. Only cookies RCO24 showed a significant difference from cookies RPO20 and CPO16 (p = 0.015 and p = 0.011 respectively). For vanilla aroma, cookies RPO20, RCO24 and CPO24 had the highest intensities. There was no significant difference (p > 0.05) in the assessment of the intensity of the vanilla aroma of cookies. For the red palm aroma, it concerned only cookies produced with red palm fat. Cookie CPO24 had the highest intensity (2.96 on a scale of 5) and are significantly different (p = 0,001) from cookie CPO16.
For the granular texture, cookies RPO16, RCO24 and CPO24 showed the highest intensities. The granular texture of cookies did not vary significantly (p > 0.05). For the melting texture, cookies RPO24, RCO24 and CPO24 showed the highest intensities. The melting textures of cookies RPO16 and RCO16 were significantly different (p < 0.05) from cookies RPO24 and RCO24. No significant difference (p > 0.05) was reported in the appreciation of the intensity of melting texture of cookies produced with red palm fat. For crunchy texture, cookies RPO16, RCO16 and CPO16 had the highest intensities. Cookies produced with refined cottonseed oil and red palm fat did not show a significant difference (p > 0.05) in the assessment of crunchiness intensity. For cookies produced with refined palm oil, cookies RPO16 showed a significant difference from cookies RPO24 (p < 0.0001).
For sweetness, cookies RPO20, RCO16 and CPO16 had the highest intensities for each type of fat. Only cookies RCO16 showed an intensity of sweetness significantly different (p = 0.012) from cookies RCO24. For salty taste, cookies RPO24, RCO24 and CPO24 had the highest intensities. No significant difference (p > 0.05) was reported for the intensity of the salty taste of cookies. The bitter taste was appreciated on cookies produced with red palm fat. Cookies CPO20 had the highest bitterness intensities. However, no significant difference (p > 0.05) was observed in the bitterness intensity of the three cookies. For the fat impression, cookies RPO24, RCO24 and CPO24 showed the highest intensities in each type of oil and significant differences (p < 0.05) from cookies RPO16, RCO16 and RPO16 respectively.
For the intensity of the bitter aftertaste, the cookies RCO24, RCO16, RPO16 and CPO20 had the highest intensities. Only cookies CPO24 differ significantly (p < 0.05) from cookies RPO20, RPO24 and RCO20.

Acceptability of cookies produced
The results of the acceptability of the various descriptors are shown in Table 6. These results represent the average assessment of forty-eight tasters. The colour of cookies RCO20 and RPO20 were more preferred comparatively to cookies RCO16, RCO24, RPO16 and  RPO24. However, cookies RPO16 and RCO16 are not significantly different from cookies RPO20 and RCO20 respectively (p = 0.080 and p = 0.469 respectively). For the cookies produced with red palm oil, the colour of cookies CPO16 was over preferred from cookies CPO20 and CPO24. Statistical analysis did not show any significant difference between the assessment of the acceptability of colour of these cookies (p > 0.05).
The odours of cookies RCO20, RPO20 and CPO24 were the most liked. For the assessment of the odour of the six cookies produced with palm oil (red and refined), the statistical analysis did not reveal any significant difference (p > 0.05). However, the assessment of odour acceptability of cookies RCO20 was significantly different from the odour of cookies RCO16 and RCO24 (p = 0.028 and p = 0.013 respectively).
For the assessment of aroma acceptability, cookies RCO20 and RPO20 were the most preferred than cookies RCO16, RCO24, RPO16 and RPO24. For those produced with red palm oil, cookies OPO24 were most preferred from cookies CPO16 and CPB20. However, for each of these three oils, there was no significant difference in the assessment of the aroma of cookies (p > 0.05).
For the assessment of texture acceptability, the cookies RCO20, RPO20 and RPO16 were the most preferred and they differ significantly from the other two cookies for each type of oil (p ≤ 0.05).
For the assessment of taste acceptability, cookies RCO20, RPO20 and CPO24 were most preferred. Cookies produced with refined palm oil (RPO16, RPO20 and RPO24) and those with red palm oil (RPO16, RPO20 and RPO24) did not show any significant difference in the assessment of their taste acceptability (p > 0.05). For cookies produced with refined cottonseed oil, the most preferred cookies RCO20 are significantly different from the acceptability of taste of cookies RCO16 (p = 0.012).
For the assessment of the acceptability of fat impression, all cookies were classified as Neither fat -Nor Dry (Rank 3 on the scale of 5); except cookies RPO16. For each type of oil, tasters reported that cookies RPO24, RCO24 and CPO24 were the fattiest compared to cookies produced with 16 and 20% of oil. Cookies RPO24 differed significantly from cookies RPO16 and RPO20 (p < 0.0001 and p = 0.006 respectively). Cookies RCO16 and CPO16 presented a fat impression significantly different from cookies RPO24, RCO24 and CPO24 (p ≤ 0.05).
For the acceptability of aftertaste, the cookies RCO20, RPO20 and CPO16 were the most preferred. No significant difference has been reported on the assessment of the aftertaste of cookies produced with each type of oils (p > 0.05). However, cookies produced with red palm oil showed an aftertaste significantly different from the majority of cookies produced with refined oils (p ≤ 0.05).

Overall assessment of the acceptability of cookies
The results of the overall assessment show that cookies RCO20, RPO20 and CPO16 were the most preferred ones respectively in each type of oil. Only cookies produced with 20 and 24% of red palm oil had overall acceptability significantly different from the other cookies. However, there was no significant difference in the assessment of the acceptability of cookies produced with red palm oil. The liker-disliker diagram (Figure 3) was used to group the overall acceptability into three (03) parts. The positive assessment (pleasant and very pleasant) is presented on the right and varied from 45.8% to 93.8%. Negative assessment (Unpleasant and very unpleasant) is presented in the middle and varied from 2.1% to 10.4% of the tasters. On the left are presented the percentages of neutral tasters who find cookies to be neither pleasantnor unpleasant. These neutral appreciations varied from 4.2 to 43.8% of tasters. However, the number of tasters who found the cookies unpleasant or very unpleasant remains marginal (less than 10.5%). For cookies produced with the two refined oils, at least 64% of the tasters reported that the cookies were pleasant or very pleasant.  Figure 4 shows the correlations between the different hedonic descriptors of the cookies produced. It appears that there is a positive correlation (r = 0.08 to 0.93) between certain hedonic descriptors such as colour, odour, aroma, flavour, texture, bitter aftertaste and overall acceptability. However, the fat impression of the cookie is negatively correlated (r = -0.04 to -0.72) with the other hedonic descriptors. These different correlations are significant at p≤0.001 between texture and bitter aftertaste; at p≤0.01 between colour and odour, between odour and aroma, between texture and flavour, between flavour and bitter aftertaste; at p ≤ 0.05 between colour and aroma, between colour and bitter aftertaste, between colour and overall acceptability, between odour and flavour, between aroma and texture, between aroma and flavour and between aroma and bitter aftertaste.

Ranking and paired-comparison tests of cookies
The Spider diagrams below show the results of the ranking by preference test and those of pairedcomparison test of cookies ( Figure 5). The cookies RCO20, RPO20 and CPO16 are the top in the classification (Figures 5 A, B and C) of cookies with each oil and are therefore the most preferred. For the classification with control cookies (M20), cookies RCO20 came first in the ranking. Figure 5 A, B and C show the best incorporation level of each type of oil according to the first three ranking tests. For the two refined oils (cottonseeds and palm), cookies RCO20 and RPO20 were ranked first and are significantly different to the cookies prepared with 16% of oil (p ≤ 0.05) ( Table  7). Figures 5A and B show that cookies RCO20 and RPO20 tend towards the same classification, namely the top position. The classification of the free best appreciates cookies (RCO20, RPO20 and CPO16) with the control cookie M20 (Table 7) showed no significant difference between cookies RCO20 and RPO20 to the cookies M20. However, cookies RPO16 significantly different to the control cookies and the cookies RCO20 and RPO20. For the three types of cookies produced with crude palm oil, cookies CPO16 were ranked first ( Figure   Figure 3. Liker-Disliker diagram of the overall acceptability of biscuits produced Type of cookies Rank Refined palm oil cookies RPO16 2 a RPO20 1 b RPO24 3 ab Refined cottonseeds oil cookies RCO16 3 a RCO20 1 b RCO24 2 ab Red palm oil cookies CPO16 1 a CPO20 3 b CPO24 2 ab Comparison to control cookies M20 M20 2 a RPO20 3 ac RCO20 1 ac CPO16 4 b ). The cookies produced with 20% refined cottonseeds and palm oils and those with 16% red palm oil, having the highest acceptability and ranked first in the various classification tests, were used in comparison tests to control cookies. The results of this classification are shown in Figure 5D and present the classification of the cookies according to the order: RCO20, M20 (control cookie), RPO20 and CPO16 respectively first, second, third and fourth. Figure 4E displays the results of the difference of cookies RPO20, RCO20 and CPO16 to the control cookies M20. Cookies CPO16 are distinguished from control cookies and occupy two vertices: "different" and "very different". 29.2% of tasters did not find any difference between cookies RPO20 to the control cookies M20. It was 16.7% between the cookies RCO20 to the control cookies M20. However, all the  FULL PAPER tasters found that there was a difference between cookies CPO16 and the control cookies M20. Thus 20.8%, 35.4% and 43.8% of tasters found that the cookies CPO16 were respectively slightly different, different and very different to the control cookies M20.

Sensory profile of the cookies produced
Lipids generally influence several sensory descriptors of cookies such as colour, texture, mouthfeel, flavour and aroma of food (Maache-Rezzoug et al., 1998;Drewnowski and Almiron-Roig, 2010). The results of the assessment of the intensity of the different descriptors varied from very low intensity (ranking 1 on 5 scales) to medium intensity (ranking 3 on 5 scales). The assessment of the intensity of the colour of the cookie, in this study, agrees with those reported by Yılmaz and Öğütcü (2015) on cookies prepared with wax oleo gels. These authors found the intensity of the colour of the cookies was moderate (Ranking 3 on 5 scales). However, the other descriptors (e.g., aroma and odour) presented a lower intensity than those of Yılmaz and Öğütcü (2015). The results of the assessment of the intensity of the colour of cookies reported in this study show significant variations depending on the incorporation level of each oil used. Similarly, the intensity of several descriptors varied significantly with the level of oil incorporation. This is the case of caramel odour of all cookies, red palm odour of cookies produced with crude palm oil, caramel aroma of all cookies, the aroma of crude palm oil of cookies produced with red palm oil, melting texture of cookies produced with both refined oils, crunchy texture of cookies produced with refined palm oil, sweet taste of cookies produced with refined cottonseeds oil, fat impression of all cookies. A similar effect has been reported for biscuits with variations in the percentages of fat (Drewnowski et al., 1998;Biguzzi et al., 2014). These authors have reported that variations in the percentages of fat in biscuits are perceptible to consumers. On the other hand, the variation of refined oil type (RCO and RPO) was not perceptible to the judges about the assessment of the intensities of all the studied descriptors. However, descriptors such as odour and aroma of vanilla, granular texture, salty and bitter taste and the bitter aftertaste of the produced cookies remained low to very low intensity and without significant variation either according to the type of oil or according to the level of incorporation. Concerning the level of incorporation, the results obtained in this study showed that the cookies with the highest oil content (24%) had the highest intensities of melting texture. This could be justified by the effect of high oil content on cookies melting texture. These results corroborate those reported by Lai and Lin (2006) who showed that an increase in fat content increases the melting texture of cookies. Similarly, Pareyt et al. (2010) reported that decreasing fat content in cookies increases the crunchy texture. Taste plays a very important role in the acceptability of cookies (Chung et al., 2014). The intensity of the three tastes studied did not vary significantly. This could be explained by the constant level of sugar in all cookies. Nevertheless, the sweet taste of cookies RCO16 is significantly different from cookies RCO24. This could be explained by the interactions between refined cottonseed oil and sugar. Previous research of Abdallah et al. (1998) has reported that certain lipids enhance the sweetness in cookies and cakes. These lipids promote the release of fat-soluble aromatic molecules that increase the sweetness of food (Pangborn, 1987;Abdallah et al., 1998). As for the assessment of the intensity of fat impression, the results obtained in this study agree with those reported by (Drewnowski et al., 1989;Abdallah et al., 1998;Yackinous and Guinard, 2000) These same authors also reported that the perception of fat in food depends on the level of lipids.

Acceptability of cookies produced
The results of the assessment of the acceptability of the different descriptors stayed middle and varied from rank 2 to 4 on a 5-point hedonic scale. These results are agreed with those of Wani et al. (2015), Kuchtová et al. (2016) and Roofigari Haghighat et al. (2017) on cookies incorporated respectively with whey protein concentrate, white grape and green tea. These authors showed acceptability ranged from 2 to 4 on a scale of 5. In another study by Hama-Ba et al. (2019) in Burkina Faso on pancakes and gruel have also obtained hedonic appreciation varying from 3.1 to 4.6 on a scale of 5 (Hama-Ba et al., 2019). The assessment of the acceptability of the cookies produced with the two refined oils is distinct from that of cookies produced with crude palm oil. Cookies RCO20 and RPO20 were the most liked among the cookies produced with the refined oils. Thus, the type of refined oil and the levels 16, 20 and 24% of incorporation did not lead to a significant difference (p>0.05) in the assessment of the acceptability of cookies. These results are similar to those of Costa et al. (2019) who work on cookies formulated with goat cream enriched with conjugated linoleic acid. Which mean the changes occurred by the type and the levels 16, 20 and 24% of refined cottonseeds and palm oils were not noticeably perceptible to the judges. For these two refined oils, the best-liked cookies according to the different studied descriptors are in line with the overall assessment of the same cookies. The assessment of the acceptability of cookies in this study shows a link between the different studied descriptors. Several studies eISSN: 2550-2166 © 2021 The Authors. Published by Rynnye Lyan Resources FULL PAPER also have reported an interaction between the different sensory descriptors (Benoit, 2004;Drewnowski and Almiron-Roig, 2010). Thus, Benoit (2004) reported that flavour is the combination of the perception of taste, odour and mouthfeel of food. This flavour plays a very decisive role in the overall acceptability of cookies (Chung et al., 2014).
Cookies CPO16 produced with red palm oil were better liked than cookies CPO20 and CPO24. This could be explained by the smallest amount of crude palm oil in these cookies. Red palm oil has a strong smell and aroma of palm, which would not be appreciated by the tasters (Cassiday, 2017). In fact, contrary to cookies produced with the two refined oils, the assessment of the different sensory descriptors of cookies produced with red palm oil was different to that of the overall acceptability of these cookies. This could be justified by the difficulty to appreciate the different descriptors because of the strong smell and aroma of the red palm oil (Cassiday, 2017). This is in line with the results of the sensory profile for several descriptors which presented significant variations with the refined oils, as opposed to cookies produced with red palm oil (Cassiday, 2017). Thus, the assessment of the acceptability of smell, flavour and fat impression of cookies CPO24 was the most preferred. As for colour, texture and aftertaste, cookies CPO16 were the most preferred. These are in line with the overall appreciation of cookies produced with red palm oil. This also shows the importance of certain descriptors in the acceptability of cookies. Indeed, several studies established the major role of certain descriptors in the acceptability of cookies (Zucco et al., 2011;van der Sman and Renzetti, 2019). These descriptors include colour and texture. The results obtained in this study are in line with the study of Zucco et al. (2011).
The Person correlation diagram showed significant correlations between the different studied sensory descriptors ( Figure 4). Thus, colour is significantly correlated with five other descriptors (odour, aroma, fat impression, aftertaste and overall acceptability). This shows the importance of colour in the assessment of the acceptability of other sensory descriptors. These results corroborate those of Zucco et al. (2011) on wheat cookies enriched with legumes. Indeed, Zucco et al. (2011) reported that colour, texture and flavour significantly influence the overall acceptability of baked goods. However, the results of flavour assessment compared to the overall acceptability assessment of cookies produced with red palm oil differed from those reported by Zucco et al. (2011). This could be explained by the interactions between the smell and the aroma of red palm oil on the flavour of these cookies. Indeed, previous work has reported that the flavour of cookies was significantly influenced by other sensory parameters such as smell, aroma and mouthfeel (Pangborn, 1987;Abdallah et al., 1998;Benoit, 2004). After colour, come successively the aftertaste (four significant correlations), flavour and aroma (three significant correlations each), smell and texture (two significant correlations each) and finally the fat impression with one significant correlation. Fat impression showed negative correlations with all other descriptors, but it was only significant with colour (P≤0.05).

Classification and paired-comparison test of cookies produced
The result of classifications of cookie agreed with those of overall acceptability assessment. The ranking test showed that the assessment of cookies RCO20 and RPO20 is similar to that of the control cookies M20. The results of cookies with refined cottonseeds and palm oil are similar to those of Costa et al. (2019) who indicate that variation of fat type, in cookie formulation, was not perceptible to the judges. The difference test showed that cookies RPO20 were not different to control cookies M20. Cookies RCO20 were slightly different to control cookies M20. The two tests show a significant difference between cookies CPO16 and cookies M20. Cookies RPO16 occupy separate classifications from cookies RCO20, RPO20 and M20.

Conclusion
This study showed that refined cottonseeds oil, refined palm oils and red palm oil can be used to produce cookies of sorghum. The two refined oils can be used in the same proportions as margarine. For red palm oil, the level of incorporation must be less than 20% to take into account the acceptability of cookies. These three oils are alternatives to margarine, which is more expensive, in the production of cheap sorghum cookies. Among the studied descriptors, colour, texture and fat impression had most influenced the consumers' preferences. The acceptability of fat impression had a significant and negative correlation with the acceptability of colour (r = 0.72); while colour, odour, aroma, taste, bitter after taste and overall acceptability had a significant and positive correlation. eISSN: 2550-2166 © 2021 The Authors. Published by Rynnye Lyan Resources FULL PAPER through the Biscuit Production Support project. We also thank Mr Issouf Paré (Ministry of water and sanitation, Burkina Faso) for his contribution in the statistical analyses.