Preparation of the basil syrups

The syrups were made using fresh basil leaves (purchased in Sekolah Seniman Pangan, Javara Indonesia), sugar, and water as the main ingredients. The main formulation had followed the previous study with modification [Pratama et al., 2013]. The first step was to make a sugar solution by adding caster sugar (100 g) and water in a 1:1 (w/v) ratio. The solution was then mixed and boiled until the sugar dissolved. The second step was to make the basil extract by blending the chopped basil leaves (100 g) using an HR2223 blender (Philips, Jakarta, Indonesia) with water in a 1:1 (w/v) ratio for 2 min and filtering the mixture using a stainless-steel filter mesh (20 mesh, BZ Wire Mesh Products Co., Anping, China) to separate the liquid (extract) and the sediment. The basil extract was then heated until it reached 100°C for around 2 min and blended with the syrup solution in various proportions. The basil syrups were then placed in sterilized glass jars for the cooling process. Afterwards, they were sealed and packaged in closed 10-mL disposable cups to be evaluated by consumers. Before evaluation, the samples were stored at ambient temperature.

This study evaluated the different addition levels of basil leaf extract to the syrup solution, i.e., 5%, 10%, 15% and 20% (v/v) (samples F1, F2, F3 and F4, respectively). In addition, three alternative sugar versions were analyzed, including: 10% (v/v) of basil extract with 50% sugar reduction (the sugar portion (50 g) was replaced with 4 g of steviol glycosides (Stevia, Tropicana Sweetener, Nutrifood, Cikarang, Indonesia) (sample F5), 10% (v/v) of basil extract with no added sugar (100% sweeteners; the sugar portion (100 g) was replaced with 8 g of steviol glycosides) (sample F6), and 10% (v/v) of basil extract with the addition of 50 g of honey (Madurasa, Natural Honey, Banten, Indonesia) instead of sugar (sample F7). All formulations were presented in Table 1. Three batches were prepared for each of the formulations.

Determination of CATA method attributes

All the attributes were generated and discussed during the focus group discussion (FGD) session by eight trained panels (5 women and 3 men, mean age: 28.6 years) from the Food Technology Department, Bina Nusantara University, Jakarta, Indonesia. Each of the panelists was served all the samples and asked to write all the sensations that they perceived [Ervina et al., 2020]. In addition, raw basil leaves, sugar, lemongrass, mint, honey, fresh cut lemon, fresh cut grass, and raw leafy vegetables were also presented as the references to confirm the perceived sensation from the samples. The similar attributes generated by panelists were then grouped together (i.e., thin, watery, water-like and water merged as thin). The CATA questions for the non-sensory attributes were adapted from the previous study [Kim et al., 2013].

Procedures in consumer tests

In total, 60 untrained panelists participated in this study (50% men and 50% women, mean age: 24.5 years). Prior to the tests, an introduction regarding the study was explained and the digital consent forms were collected by asking the respondents to thick the “agreed” box in the online questionnaire if they were willing to participate. In addition, a verbal consent was also asked at the beginning of the evaluation, all the participations was voluntary. The list of the ingredients used in the samples were provided to the consumers at the beginning of the test to avoid recruiting someone with food allergies or dietary restrictions with one of the ingredients. The participants were also informed that all the data and information collected will be processed and stored as anonymous. The study has been approved by the ethical committee of the Research and Technology Transfer Office (RTTO), Bina Nusantara University, Jakarta, Indonesia (referral code: 60/VR.RTT/IV/year 2022) and has followed the World Medical Association (WMA) declaration of Helsinki [WMA, 2013].

The tests were divided into two parts. The first part was the blind test, all the samples were coded in a three-digit-random number and served monadically to the participants in a random order. The second part was the non-blind test, the participants were provided with the information regarding the samples (i.e., the ingredient list, sugar content such as 100% sugar, 50% reduced sugar, no sugar added or replaced with sweetener and honey). Noteworthy is that the consumers had no information that the first and second part was actually evaluating the same samples. The blind test evaluated all the samples presented in Table 1. This evaluation included the overall liking and all the CATA attributes, while the non-blind test evaluated only samples F2, F5, F6 and F7 and focused on evaluating the overall liking and the non-sensory attributes only. Consumer liking was measured using a 9-point hedonic scale and was always evaluated first followed by the CATA questions. The participants were asked to tick the attribute (check-all-that-apply) if it was perceived in the sample [Piochi et al., 2021; Tarancón et al., 2020]. They were also asked to rinse their mouth with water before tasting and in between sample tasting. A short break for around two minutes was provided to ensure all the sensation from the previous sample had vanished before the participants continued to the next sample. The results of CATA questions were expressed as percentage of frequency of perception of each attribute by consumers.

Statistical analysis

The CATA results were analyzed using Cochran’s Q test [Galler et al., 2020]. The attribute with a significant difference (p≤0.05) was then further analyzed using McNemar-Bonferroni test. The correspondence analysis was employed to map the associations between the attributes and the samples. In addition, the penalty analysis was conducted to evaluate which attributes significantly impacted liking. To evaluate the differences between the blind and the non-blind test setting on liking, a Student’s t-test was conducted. All the data was analyzed using XLSTAT Sensory version (Addinsoft, Paris, France, Version 2019.2.2).

Sensory characteristics of the basil syrups

Thirty-two attributes of basil syrups were determined and used in the CATA method. They consisted of 3 visual attributes, 9 aromas, 10 tastes, 2 textures and 8 non-sensory attributes. The complete attributes are listed in Table 2. CATA can incorporate both sensory and non-sensory properties such as emotion or feeling when consumers tasted the products or the perception related to price, usage occasions, and product positioning [Ruiz-Capillas et al., 2021; Varela & Ares, 2012], which makes this method very consumer oriented. CATA method has been reported as a valid method to discriminate products’ characteristics using untrained panel or consumers [Tarancón et al., 2020].

The frequency of consumer-perceived CATA sensory attributes of each basil syrup recipe in a blind test is presented in Table 3. For the texture attribute, most of the samples had thin and watery-like characteristic with F6 having the most watery texture (perceived by 100% of the participants) followed by F5 (90% of the participants). F6 is the formula of basil syrup without any sugar addition (replaced by sweetener), while F5 is the formula with 50% sugar reduction (substitute with sweetener). In beverages, sugar is not only used to sweeten taste but also to improve texture and mouthfeel [Wagoner et al., 2018], therefore reducing or replacing sugar with sweetener will decrease the viscosity of the syrups.

For the color attributes, F7 (formulation with honey) appeared to be the most transparent and did not have dark green color (Table 3). The green color of basil may be covered with the yellow-to-dark amber color from honey [Aparna & Rajalakshmi, 1999]. The addition of basil extract at 10–20%, v/v (F2, F3, and F4) significantly increased the dark green color of the syrup, this may be due to the chlorophyl content of the basil leaves [Dadan et al., 2021]. Interestingly, the dark green color was significantly less perceived when the sugar content was reduced or replaced by sweetener in F5 and F6, respectively. These phenomena can be explained due to the formation of sugar degradation product due to the heating process in the sample making [Hubbermann et al., 2006]. The more sugar added in the formula enhanced the formation of the sugar-degradation product. This condition will contribute to the darker color of the final product [García et al., 2017].

The basil aroma was perceived by more than 50% of the consumers in most of the formulations except for F7 (honey) and this trend was similar for the basil taste (Table 3). The taste and aroma of grassy were perceived high in the sample with 20% (v/v) addition of basil extract (F4). The cis-3-hexanal is abundant in green plants, including basil, and this aromatic compound contributes to the fresh-cut grass aroma or grassy odor [van Nieuwenburg et al., 2019]. Interestingly, the taste and aroma of lemongrass was also perceived in most of the samples, indicating that basil leaves contain similar odorants with lemongrass, such as linalool (contributes to floral and sweet notes), citral (lemon, citrus, sour), and myrcene (spices) [Patel et al., 2021; Skaria et al., 2006]. The bitter taste was perceived as extremely high in F6 (100% sweetener). The addition of sweetener to fully replace sugar in F6 may result in bitter note and bitter aftertaste of the basil syrups. Steviol glycosides were used as sweeteners in this study. This class of compounds has been reported to provide a strong bitterness intensity and bitter aftertaste [Hellfritsch et al., 2012; Tao & Cho, 2020]. The stevia compounds were selected in this study because they are the natural-non-caloric sweetener and well-known by consumer [Tao & Cho, 2020]. The bitterness of steviol glycosides became the main challenge of food industries to use stevia as a non-nutritive sweetener [Hellfritsch et al., 2012; Tian et al., 2022]. Sweetness was perceived in most of the basil syrup samples except for F6. The sweetness intensity in F6 may be covered by the lingering bitter taste sensation from stevia and this may reduce the sweetness sensation [Tian et al., 2022].

The sensory characteristics of each formulation were also mapped in a symmetric plot and presented in Figure 1. The two principal components of PC1 and PC2 contributed to 81% of variability. Based on the CATA mapping, the sensory profile between F1 (5% extract, v/v), F2 (10% extract, v/v), F3 (15% extract, v/v) and F4 (20% extract, v/v) lies on a similar group and in proximity. These samples were characterized with taste and aroma of sweet, lemongrass, basil, and dark green color. Moreover, samples made with 100% sugar (F1, F2, F3, and F4) lay on left quadrants and were separated with the samples made with sweetener or honey (F5, F6, and F7) which lay on right quadrant based on PC1. In addition, F5 and F6 were also separated with F7 based on PC2. Based on the mapping, F5 and F6 were associated with thin, watery, bitter aroma and bitter taste, while F7 was closely related to taste and aroma of citrus, honey, and was transparent in color. The sensory profiles of each formula presented in the mapping from Figure 1 were in line with what was described based on data in Table 3.

Figure 1

Plot of correspondence analysis of sensory attributes of basil syrups with different contents of basil extract and sugar. SWE, sweetener; SR, sugar reduction (50% sugar); HN, honey; A, aroma; T, taste. Syrup codes and their formulations are detailed in Table 1.

Blind test vs. non-blind test: Impact on consumers’ perception and liking

This study also investigated the consumer perception for non-sensory attributes and their liking in different settings of blind test and non-blind test. Based on Cochran’s Q test (Table 4), “healthy drink” was perceived as an important attribute in the non-blind test (p<0.001) but not in the blind test evaluation (p=0.110). Sugar content in the basil syrups may relate with health perceptions, because sugar is commonly associated with metabolic diseases such as diabetes, overweight and obesity [Reis et al., 2017]. The consumers were informed regarding the sugar content of the samples (reduced sugar, replaced by sweetener, honey) in the non-blind test. The disclosure of health-related information of a product could significantly increase the willingness of consumers to try and to buy a product, and this information could positively modulate consumer perception [Grasso et al., 2017]. In the case of sugar content, consumers may think that the sugar-reduced and the sugar-free version of the basil syrup is associated with healthiness, such as diabetes prevention, reduced calorie intake, and prevention of overweight and obesity, and thus able to positively impact consumers perception and their acceptability [Reis et al., 2017; Wardy et al., 2018].

Table 5 presents the consumer perception for the non-sensory attributes in the blind and non-blind test of each basil syrup formula. The “healthy drink” attribute was perceived as significantly higher in the non-blind testing compared to the blind test, especially for the samples with sugar reduction (F5), no sugar (F6) and honey (F7). Sugar reduction has been closely correlated with health status as the WHO suggests to reduce sugar intake by less than 10% of the total dietary intake with a further reduction of 5% providing additional health benefits [Deliza et al., 2021; WHO, 2015]. Moreover, previous studies have reported the association between sugar reduction and consumer perceptions [Deliza et al., 2021; de Souza et al., 2021; Wardy et al., 2018] showing that consumers will provide positive attitude and prefer sugar-reduced or sugar-free version for health reasons. However, the perception of “sugary drink” was similar in both settings. We assume this could be influenced by the texture of the basil syrups. In the blind test, consumers already perceived the sugar-free samples as less sugary even though they had no information regarding the sugar content. The sugar-free version was characterized as thin, watery and having bitter taste, thus this might have influenced consumers perception regarding the sugar levels [Wagoner et al., 2018].

The information provided to the consumers before the test could influence their perception to the product directly. For example, the information regarding coffee quality before tasting was reported to significantly increase consumers preferences [Bemfeito et al., 2021]. Moreover, another study suggests that the nutritional information, such as sugar content, was able to influence consumer perception regarding physical health and emotional aspects of wellbeing [Reis et al., 2017]. The same study also showed that consumer perception was significantly changed when they were provided information regarding the sugar content of the juice, in which consumers preferred juice without added sugar (sugar-free) [Reis et al., 2017]. The use of sweetener is associated with less calorie foods, and consumers may have expectation that this product will provide health benefits [Deliza et al., 2021; Wardy et al., 2018].

The hedonic aspect was also evaluated to compare the liking score of the basil syrups in the blind and the non-blind setting. Based on the Student’s t-test, the liking showed to be significantly higher (p=0.050) for the basil syrup with sweetener (no added sugar) in the non-blind test (mean liking score 4.72) compared to the blind test (4.03), while there were no significant differences for the other samples (Table 6). In addition, the results of the penalty analysis for the non-sensory attributes were different in the blind test and the non-blind test. The mean impact is presented in Table 7. The results show that “sugary drink” positively affected consumer liking in the blind test but not in the non-blind test setting. Moreover, “fresh”, “innovative” and “healthy drinks” were able to significantly promote liking while “herbal drink” was found to decrease liking in the non-blind test. However, the results were different in the blind testing, as only the “sugary drink” positively impacted liking. The non-nutritive sweetener used in the syrup formulation may provide a “healthy” perception for consumers. Based on the penalty analysis, the “healthy drink” attribute significantly (p=0.050) influenced liking (mean impact of 0.43). This may promote consumer acceptability towards the samples since a non-nutritive sweetener is associated with non-caloric food and sugar intake reduction [Deliza et al., 2021] as well as weight and diabetes management [Farhat et al., 2021; Lohner et al., 2017], which is closely related with healthiness aspect. The liking of basil syrups was influenced by the different attributes in the blind test and the non-blind test, suggesting that consumer perception can be modulated by the information provided prior to testing.

Our study found that the drivers of liking for non-sensory attributes were different when comparing the blind and the non-blind test evaluation. The information provided to consumers prior to tasting can influence their perception and liking. Our study corroborates previous studies [Bemfeito et al., 2021; Fibri & Frøst, 2020; Henrique et al., 2015; Mazhangara et al., 2022] that also found similar results in comparing consumers perception in the blind and the non-blind setting. For example, information regarding the label of organic foods was able to influence consumers’ perception of the products and increase their purchase interest [Asioli et al., 2018]. Further, information regarding coffee quality was shown to change consumer preferences and increase their liking significantly compared to the blind testing [Bemfeito et al., 2021]. A similar result was also reported for sausages [Mazhangara et al., 2022] and cooked ham [Henrique et al., 2015], suggesting that the information obtained regarding the samples (ingredients, processing, origin) in a non-blind setting can shift consumer perception and preferences. A previous study reported that when consumers were presented with information regarding salt-reduction in the biscuit samples their liking was higher compared to a normal-salt biscuit [Vazquez et al., 2009]. However, when the same biscuits were tested in the blind test setting the liking was similar between normal-salt and reduced-salt biscuit. Sensory and non-sensory cues may interact to build consumer perception, and the interaction between these cues depends on such factors as the product characteristics, familiarity, brand, sensory profiles, prices or any additional claim related to health and wellness [Carrillo et al., 2012], which can influence consumer preferences.