The following is a summary of the stated research mentioned above. The content summarized here, including the figures and tables, all belong to the researchers (unless otherwise indicated). The summary attempts to stay as close to the original paper as much as possible with some adjustments in regards to jargon, length, or to focus on bean to bar aspects.

Introduction

Thousands of years ago, cacao was enjoyed as a beverage. It was likely a bitter, zesty, and spicy drink that was made by mixing ground cocoa beans with water and spices. The cacao beverage was then becoming popular in Europe around the 16th Century where they changed up the ingredients and added things such as sugar, vanilla, and cinnamon to make it more appealing to their palate.

Chocolate has a long history of pairings in various cuisines, rituals, and ceremonies. Many recipes exist which include chocolate as an ingredient in entrees, desserts, beverages, appetizers, pasta, and meat. In Italy, many chocolatiers include chocolates with flavoured creams or liquid sealed in a shell of chocolate including rum, espresso, champagne, grappa, and whiskey.

No systemic research to date of this publication has been conducted to examine the relationships between chocolate and beverage pairings. This study investigated the hedonic response of consumers to a set of chocolate and beverage pairs, and the consumer preference patterns were evaluated. Factors such as chocolate type, beverage type, beverage liking, and chocolate liking were analyzed to see how well they matched with one another.

Materials & Methods

Chocolates

Chocolates were purchased in Varese, Italy, and ranged from 30%, 70%, and 99% cocoa content. Details on the producer and ingredients are shown in Table 1. The chocolates were cut into 15g pieces, and served at room temperature. A 3-digit random number was assigned to each sample for tracking purposes.

beverages

Beverages to pair with chocolate were selected by two focus groups. Thirty two beverage types were presented to the group, and if a beverage was chosen by at least 50% of the participants (n=18) then it was used in this study. Beverages included alcoholic and non-alcoholic beverages with a range of sensory characteristics including ethanol content, sweetness, astringency, and bitterness. A full list of beverages is shown in Table 2. All beverages were purchased in Varese, Italy.

Water and alcoholic beverages were served in 30ml portions into stemware and covered with plastic lids. Temperatures are shown in Table 2.

Coffee was brewed in a 3-cup Moka Express espresso coffee maker. Coffee was served unsweetened at 50-60°C in a porcelain cup.

Tea bags containing about 2g of tea were steeped for 3 minutes and then removed. 120 ml of black and green tea were served at 50-60°C in porcelain cups.

determination of the sensory profile of beverages and chocolate & Lexicon Generation

Sensory profiling was carried out using descriptive analysis according to Stone and Sidel (1993). Descriptive analysis is a two-stage method comprising the following stages:

A lexicon generation process
A set of sensory tests designed to quantify on a rating scale the intensity of the sensory terms established in the lexicon generation phase

Three digit random numbers were assigned to each sample for tracking. Replicates were tasted on different days. The order of presentation of the samples was balanced and randomized across samples, panelists, and replicates.

Fourteen individuals participated in the chocolate lexicon generation, and had a minimum of 3 years and maximum of 6 years experience in sensory evaluation. Five separate tasting sessions were held with cocoa solid percentages ranging from: 30%, 65%, 70%, 85%, and 99%. Participants were asked to name as many sensory characteristics as possible that they considered important enough to describe the chocolate.

Redundant terms were discussed openly with the group and panel leader in order to reduce the list of descriptors. Descriptors mentioned by at least 30% of the panel were retained and intensively discussed among panelists until a final list of 34 descriptors were selected. They ranged from aroma, to taste, and texture as shown in Table 3. Eleven of the main sensory descriptors for differences between the whole set of beverages were used in the descriptive analysis of beverages shown in Table 4. Descriptors were selected by the authors based on their extensive experience in descriptive analysis of beverages. Sensory tests were designed to quantify on a 9-point scale the intensity of the terms selected for both chocolate and beverages.

consumer hedonic test

Consumers from Milan and the vicinity were randomly recruited, and eligibility depended on:

Being regular consumers of both chocolate and alcoholic & non-alcoholic beverages
Do not have an aversion for chocolate with high cocoa content
Do not have any known food allergies or dietary intolerance
Spanned a range of age classes and regions of origin
Willing and available to participate in the study

Eighty consumers (40 male, 40 female) were randomly selected within respondents in order to have the following demographics and preference patterns:

46.25% of the participants were aged 20-29
40% were aged 30-45
13.75% were over 45 years old
66.25% lived in Northern Italy
23.75% lived Central Italy
10% lived in Southern Italy
45% preferred chocolate with a cocoa content over 65%
21.25% preferred milk chocolate
Consumed chocolate (solid or liquid) at least once a week

tasting sessions

Panelists received mineral water and dry unsalted breadsticks for palate cleansing. Each tasting session lasted around 45 minutes. The tasting occurred in 3 stages:

Stage one - panelists were asked to taste and rate how much they liked each chocolate (Table 1) or beverage (Table 2) using the 9-point hedonic scale (9 being extremely like). The chocolate and beverages were tasted separately at different times.
Stage two - panelists were served chocolate and beverages pairs and asked to score on a 9-point scale their hedonic rating. Consumers were told to take the chocolate, let it melt on the tongue, move it around in their mouth, and then sip the beverage.
Stage three - panelists evaluated whether in each pairing it was the chocolate or the beverage that dominated. At one of end of the scale they had “chocolate dominates the pairing” at -6, “neither dominates the other” in the middle of the scale at 0, and “beverage dominates the pairing” on the other end at +6. A harmonic pair was considered when they scored close to zero.

statistical analysis

One-way analysis of variance (ANOVA) and factorial ANOVA were performed according to Meilgaard et al. (2007). ANOVA was used to determine if there was a significant difference for:

liking across chocolate
liking across beverages
liking across pairings
match level of each chocolate and beverage pair

Duncan’s post-hoc test (significance at p<0.05) was performed.

Results & Discussion

Descriptive analysis of chocolate

The lowest cocoa chocolate (30%) was described as being fattier, stickier, more soluble, creamier, and more mouth coating than other chocolate samples. It also was deemed richest in milk aroma, vanilla aroma, milk taste, vanilla taste, and the sweetest.

The textural properties of the dark chocolate were more intense. The 70% was considered to be the richest in smoky, dried fruit, fruit, roasted, cacao, and sour tastes. The 99% was the grittiest, firmest, crunchiest, most astringent, tobacco, liquorice, roasted, smoky, cacao, bitter, salty, and most persistent overall taste.

descriptive analysis of beverages

The distilled beverages such as grappa, whiskey, and rum had a higher perceived level of alcohol compared to other alcoholic beverages. The liquor, flavoured and Port wines were fuller bodied than other beverages. The Lambrusco, Port, and flavoured wines were perceived to be the most astringent while coffee and flavoured wines were perceived to be the most bitter. The liquor and Port wines were considered higher in fruitiness, and the liqueur wine was considered higher in sweetness. Balsamic vinegar was the most sour beverage, while flavoured and Port wines were highest in spiciness. Coffee was higher in aroma and flavour persistence.

Hedonic response to chocolate and beverages

Table 5 lists the average liking scores for the chocolates and beverages. The lowest percentage of chocolate had the highest rating of likeness, and vice versa for the darkest. The post-hoc test revealed that sparking wine, Lambrusco wine, coffee, liqueur wine, black tea, Port wine, and balsamic vinegar were more liked than mineral water, Marzemino wine, grappa, wheat beer, rum, Danish stout, Belgian stout, green tea, and flavoured wine.

Hedonic Response to chocolate and beverage pairs

Table 6 shows the mean liking scores for the pairings.

The 30% chocolate seemed to pair well with liqueur wine, balsamic vinegar, and black tea. It appeared to be disliked when paired with Marzemino wine, mineral water, flavoured wine, rum, whiskey, and Belgian stout.

The 70% chocolate seemed to pair well with balsamic vinegar, liqueur wine, Belgian stout, Port wine, coffee, Lambrusco wine, and grappa. The least liked pairs with the 70% was Marzemino wine, mineral water, Sauvignon wine, whiskey, Danish stout, black tea, and rum. Belgian stout appeared to be preferred here than with the 30% or 99% samples.

The 99% chocolate paired the best with balsamic vinegar.

There appeared to be no significant difference when green tea, balsamic vinegar, wheat beer, Danish stout, Marzemino wine, flavoured wine, rum, and whiskey was paired with either 30% or 70% chocolate samples.

The beverages mineral water, black tea, coffee, sparkling wine, Lambrusco wine, Sauvignon wine, liqueur wine, Port wine, and grappa were more appreciated with the 30% chocolate than the 70% and 99% chocolate samples.

Chocolate and beverage match

30% Chocolate

The chocolate and beverage match results are shown in Table 7. Whiskey, grappa, and flavoured wine moderately dominated the 30% chocolate, but whiskey was far more dominate than grappa. On the other hand, rum, Marzemino wine, Sauvignon wine, Port wine, Danish stout, balsamic vinegar, Lambrusco wine, wheat beer, Belgian stout, and black tea were slightly to moderately dominated by the 30% chocolate. Mineral water and green tea were largely dominated by the 30% chocolate, and had a lower perceived match level than other beverages. The harmonic pairings for 30% chocolate were the liqueur, Sauvignon, Port wines, Marzemino, and rum. There was a lack of agreement in the match level of the 30% with whiskey, grappa, rum, and Danish stout. Good agreement was found with green tea, black tea, sparkling wine, mineral water, coffee, and Belgian stout.

70% Chocolate

With the 70% chocolate, whiskey moderately dominated the chocolate. Harmonic parings were with grappa, rum, flavoured wine, Port wine, and Marzemino wine. The 70% was moderately dominate over the Danish stout, balsamic vinegar, coffee, Belgian stout, liqueur wine, Sauvignon wine, and sparkling wine samples. The 70% was largely dominate over black tea, green tea, and mineral water. There was a low level of agreement in preference when 70% was paired with rum, whiskey, Sauvignon wine, Danish stout, and grappa, and a higher level of agreement among participants was observed with black tea, balsamic vinegar, mineral water, coffee, Belgian stout, and Marzemino wine.

99% Chocolate

The 99% chocolate was slightly dominated by the Port wine and rum, and moderately dominated by the liqueur wine, Marzemino wine, balsamic vinegar, Belgian stout, and Sauvignon wine samples. On the other hand, the 99% chocolate dominated coffee, Lambrusco wine, Danish stout, sparkling wine, wheat beer, green tea, black tea, and mineral water samples. There was a poor consistency in agreement with whiskey, rum, grappa, Danish stout, Port wine, Sauvignon wine, and Marzemino wine. There was good agreement observed with mineral water, coffee, black tea, and green tea.

Drink specific observations

Whiskey and grappa seemed to dominate the three chocolate samples, while Belgian stout was dominated by all three chocolate samples.

Port wine dominated the 30% and 70% chocolate, but was dominated by the 99% chocolate.

Flavoured wine dominated 30% and 70% chocolate.

Sauvignon wine and rum dominated 30% chocolate only.

The teas, sparkling wine, and liqueur wines were more dominated as the level of cocoa went up.

Mineral water, coffee, balsamic vinegar, wheat beer, Danish stout, Lambrusco wine, and Marzemino wine

Sauvignon and flavoured wines were dominated by the 99% chocolate.

Other Pairing Observations

It appeared that the hedonic rating for the pairing decreased when the beverage was dominated by the chocolate in nearly all cases except for whiskey. In particular, when chocolate dominated coffee, balsamic vinegar, rum, flavoured wine, liqueur wine, Lambrusco wine, and wheat beer, the pairing was strongly disliked. The liking of the pairing increased when the chocolate was paired with liqueur wine, flavoured wine, Marzemino wine, Lambrusco wine, balsamic vinegar, and rum.

There were significant differences in the appreciation of many of the beverages when consumed on their own (Table 5) versus when in combination with the chocolate samples (Table 6). Overall, pairings with the 99% chocolate were not appreciated by most participants, and less liked than pairings with the 30% and 70% chocolates (both of which were more versatile in their pairings than the 99%). This is likely due to the high astringency and bitterness of the 99% chocolate. As well, the chocolate did not melt in a way favorable to consumers, and melt/texture is an important feature of chocolate.

The reason the balsamic vinegar fit well with the 99% chocolate is the contrasting extremes in both muted the unfavorable elements of those two foods. The sour, viscous and round elements of the balsamic contrasted the extreme bitterness and astringency of the 99%, muting the sour, bitter, and astringent factors which most find not as appealing (Bastian et al., 2010). A rationale behind this is found in studies with model mixtures such as that by Keast & Breslin (2002). Bitterness and sweetness are mutually suppressive and sweetness is suppressive of other basic tastes. The balsamic vinegar here may modulate the bitterness of the 99% chocolate.

Although a high heterogeneity of all the drinks sampled, most were well paired with the 30% chocolate. The sweetness of this 30% chocolate may have suppressed the slight bitterness of the Port wine, Lambrusco, and coffee. As well, the 30% was richer in fat, and fat rich foods may modify the sensory profile of a beverage making them less astringent.

Keep in mind that a consumers individual preference also came into play. Their affinity for a particular drink had a significant impact on their affective response to the chocolate and drink combination. This was measured, and the effect of the beverage liking was larger than the effect of the chocolate liking. Their affinity for a particular beverage seemed to make the more likely to enjoy the actual pairing. This supports the notion that in pairings, the liking of the pairing depends more on the drink than on the food (Bastian et al., 2010; Donadini et al., 2008). This also suggests that the beverage and chocolate do not serve as equal partners, but the beverage takes on a greater supporting role in regards to consumer pair liking.

If chocolate dominated the pairing, a drop in acceptance was generally observed. If the drink dominated the chocolate, there can be either a drop or a gain in liking that specific pairing. Therefore, in order to achieve a higher rate of beverage and pairing acceptance, the drink has to be in harmony with the chocolate, or slightly dominate as opposed to be being dominated by the chocolate.

There was more agreement with participants when it came to chocolate pairings in regards to black tea, green tea, mineral water, coffee, and sparkling wine. There was less agreement (more medium to high variation) when these pairings occurred with beverages having medium to high alcohol content such as rum, grappa, whiskey, Sauvignon wine, and Marzemino wine. Alcohol is perceived as a combination of sweet tastes, bitter tastes, odours, and oral irritation, all of which vary as a function of ethanol concentration, consumer drinking history, and human genotype. The genetically determined individual variation of ethanol-related chemosensation accounts for some variation in perceived bitterness, astringency, and viscosity from alcoholic beverages.

As well, the high individual variation in salivation, mastication, and swallowing patterns likely influences the biotransformation process that occurred when the chocolate was mixed with the drink in the mouth. The perception of chocolate melting is reported to be highly heterogeneous in humans (Markov & Tscheuschner, 1989).

Further research is required to complement the results found here with more specific information on the flavour interactions that occur when chocolate and beverages are consumed together.

The Hedonic Response To Chocolate And Beverage Pairing: A Preliminary Study