Download as pdf or txt
Download as pdf or txt
You are on page 1of 17

Food Quality and Preference 87 (2021) 104043

Contents lists available at ScienceDirect

Food Quality and Preference


journal homepage: www.elsevier.com/locate/foodqual

Effects of “craft” vs. “traditional” labels to beer consumers with different T


flavor preferences: A comprehensive multi-response approach

Sara R. Jaegera, , Thierry Worchb, Tracey Phelpsa, David Jina, Armand V. Cardelloc
a
The New Zealand Institute for Plant and Food Research Limited, Mt Albert Research Centre, Private Bag 92169, Victoria Street West, Auckland 1142, New Zealand
b
Friesland-Campina, Wageningen, the Netherlands
c
A.V. Cardello Consulting and Editing Services, Framingham, MA, USA

A R T I C LE I N FO A B S T R A C T

Keywords: Common definitions of craft-style beers emphasize that they are brewed by small, local breweries and that they
Consumer research have complex, novel and, often, bold flavors. In recent research conducted with declared craft beer drinkers who
Product labels tasted beers in a blind condition, Jaeger et al. (2020) [FQAP, 103884] identified the existence of two preference
Information effects segments. One segment (Craft-style Likers) preferred the characteristic complex, novel and bold flavors of many
Beer
craft-style beers, while the other segment (Traditional-style Likers) preferred the less complex and milder flavors
associated with traditional-style beers. In the present research, a multivariate response approach that included
the evaluation of liking, sensory, holistic/conceptual, and cognitive/emotional/situational responses, as well as
beer-related attitudes and behaviors, was used to assess the same set of six beers in a similar sample of declared
craft beer drinkers (n = 122) as used in our previous study. However, beers were now labeled to reflect the style
of the beer (craft vs. traditional) and the size (small vs. large) and regional origin (local vs. international) of the
brewer. Results confirmed the existence of the same two preference segments. Although the effects of labeling
the beers in accordance with their craft- vs. traditional-style natures did not influence sensory, holistic/con-
ceptual or cognitive responses, they did influence emotional and situational responses. For craft-style beer likers,
the information increased their positive emotions and the situational uses for craft-style beers, while for the
traditional-style beer likers the labeling had a corresponding positive effect on their judgments of the traditional
style beers. The results confirm previous findings on beer and product labeling, which show that sensory
properties are less susceptible to extrinsic information effects than are affective, emotional and other, non-
sensory product judgments. Implications for craft and traditional brewers are discussed and suggestions for
future research presented.

1. Introduction that are typical of most mainstream beers. Craft beer drinkers have also
been shown to place a high degree of importance on the quality of the
1.1. Flavor preferences and segments among craft beer drinkers beer (Aquilani, Laureti, Poponi, & Secondi, 2015) and on its authenti-
city (Gómez-Corona et al., 2016). In addition, they experience more
Both beer experts and trained profile panelists characterize craft positive emotions from craft-style beers (Jaeger et al., 2020) and at-
beers as being of high flavor complexity (Medoro et al., 2016). Con- tribute greater appropriateness of their use in a variety of drinking si-
sumer research also shows that craft beer drinkers seek a wide variety tuations, compared to lager-style beers (Jaeger et al., 2020).
of beer styles and flavors. These often include highly innovative beers, In recent research with declared craft beer drinkers, Jaeger et al.
some of which have high levels of hops, strong flavor profiles and/or (2020) identified two distinct flavor preference segments. One segment,
high alcohol content. Work by Gómez-Corona and collaborators designated as a strong flavor liking segment, enjoyed the character-
(Gómez-Corona, Escalona-Buendía, García, Chollet, & Valentin, 2016; istically novel and high flavor impact profiles of craft-style (CS) beers
Gómez-Corona, Escalona-Buendía et al., 2017) found that the unique- (CS Likers), whereas another segment preferred the lower flavor impact
ness of the flavors of craft-style beers is an important driver of their of more traditional-style (TS) beers (TS Likers). Although the consumers
consumption by this group of consumers. In fact, many craft beer differed in their flavor preferences, they did not differ greatly in either
drinkers have come to eschew traditional beers with low flavor impact their sensory or holistic/conceptual characterizations of craft or


Corresponding author.
E-mail address: sara.jaeger@plantandfood.co.nz (S.R. Jaeger).

https://doi.org/10.1016/j.foodqual.2020.104043
Received 29 May 2020; Received in revised form 28 July 2020; Accepted 28 July 2020
Available online 31 July 2020
0950-3293/ © 2020 Elsevier Ltd. All rights reserved.
S.R. Jaeger, et al. Food Quality and Preference 87 (2021) 104043

mainstream beers. Both preference clusters characterized the craft-style 1.2. Evidence of information and labels affecting beer perceptions and liking
beers used in the study as ‘bitter,’ ‘hoppy,’ and having ‘a strong aroma’
and ‘a long aftertaste,’ while characterizing the sensory properties of Although relatively little data exist showing the effects of extrinsic
the traditional-style beers used in the study as ‘watery,’ ‘weak’ and product factors on craft beer perception, acceptance, and consumption,
having a ‘short aftertaste.’ Thus, even though each cluster exhibited in other food and beverage categories the extrinsic factor that has been
different flavor preferences, both characterized the sensory properties shown to be most influential on these response variables is the in-
of the beers similarly. With regard to holistic and conceptual responses, formation provided to consumers about the product, especially at point
these too were similar between preference clusters, with both clusters of purchase through packaging and labeling (e.g., Cardello & Sawyer,
characterizing the high flavor profile craft beers as ‘full-bodied,’ ‘sub- 1992; Cardello, 1994, 2007; Deliza & MacFie, 1996; Tuorila,
stantial,’ ‘masculine,’ ‘complex,’ ‘powerful’ and ‘memorable’ and the Meiselman, Cardello, & Lesher, 1998; Schifferstein, Kole, & Mojet,
traditional-style beers as ‘familiar,’ ‘smooth,’ ‘balanced,’ ‘comforting,’ 1999; Jaeger & MacFie, 2001; Cardello, 2003; Caporale & Monteleone,
‘easy to drink,’ and ‘thirst-quenching.’ Although some differences were 2004; Stefani, Romano, & Cavicchi, 2006; Mielby & Frøst, 2010;
observed between the preference clusters for specific holistic/con- Piqueras-Fiszman & Spence, 2015). The elements of product informa-
ceptual judgments and for specific beers, overall, the holistic/con- tion that can affect product perceptions and liking are varied and can
ceptual responses to the beers by cluster were highly similar. include the brand name, the manufacturer, nutrition information,
In contrast to the above, the cognitive/emotional/situational use quality information, processing information, geographic origin of the
variables collected in Jaeger et al. (2020) showed greater divergence product, and visual and other sensory aspects of the label and its
between the preference clusters. Although the beer drinkers in both packaging (Cardello, 2007).
clusters characterized the craft-style beers similarly, i.e., “I think “beer Specifically with regard to beer, one of the first studies to examine
aficionados” would enjoy this beer,” “this beer has an authentic taste,” label information effects was that of Allison and Uhl (1964), who de-
“this is a unique beer,” and “in my opinion, this tastes like a craft beer,” monstrated that consumers who could not discriminate or recognize
this segment diverged from the low flavor profile likers on almost all of their preferred beer by taste alone, were able to do so in the presence of
the other variables, characterizing the high flavor profile beers as those their labels. Similarly, Guinard et al. (2000, 2001) found that the pre-
“they typically drink,” “that like-minded drinkers drink,” for “when ferences and liking of consumers for a variety of domestic, imported
they want more than one” and that evoke “positive emotions,” all and specialty beers differed significantly in blind vs. brand- and price-
characteristics that the traditional-style liker segment ascribed to the labeled conditions. They found that the effect of information was par-
more traditional-style beers. Thus, while low flavor profile beer likers ticularly pronounced for consumers in their twenties, an age group that
perceive craft-style beers to be authentic, unique, good quality craft comprises the largest segment of the craft beer market. In an interesting
beers that would be consumed by beer aficionados, they accurately study, Lee, Frederick, and Ariely (2006) had patrons of a pub evaluate
perceive traditional-style beers to be those that they themselves typi- their preference for a traditional mainstream beer vs. the same beer
cally drink and from which they derive positive emotions. They also with a few added drops of balsamic vinegar either a) blind (no in-
view these latter beers as the ones they desire when they want to drink formation about the added ingredient), b) informed about the vinegar
more than one beer and when they want to savor a beer by themselves. before they tasted the beer or c) after they tasted the beer. Results
Lastly, they view the latter beers as the ones that they would consume showed that preference for the vinegar-added beer was most affected
in many different social situations, and very importantly, they accu- when the information was presented prior to tasting the beer, sug-
rately perceive these beers to be “mainstream,” a characteristic also gesting that information influences responses through an anticipatory
ascribed to these beers by the craft-style flavor liker segment. mechanism, such as expectations (Cardello, 1994, 2007). In other re-
The study by Jaeger et al. (2020) also revealed differences in atti- cent research, Barnett, Velasco, and Spence (2016) showed that beer
tudes and behaviors toward beer between the two preference segments. poured from a bottle was rated as tasting better than those poured from
Although the two clusters did not differ in their general liking of beer, a can, while Barnett and Spence (2016) found that beer accompanied by
their involvement with beer or their need for uniqueness in foods/ a green label emphasizing the beer’s citrus/fruity notes produced sig-
beverages, the high flavor likers agreed significantly more with such nificantly higher ratings of its perceived quality, taste and the magni-
statements as “I seek out beers made by small, local, independent tude of its fruity/citrus notes than did an unlabeled beer or one ac-
breweries,” “I like beers that are of high quality and authenticity,” “I companied by a brown label. Lastly, Silva et al. (2017) had consumers
purchase and refill flagons of beer at local breweries” and “I have a beer drink either regular beer or non-alcoholic beer labeled correctly or in-
tap in my house,” all of which reflect attitudes that lie at the heart of the correctly (as regular beer or non-alcoholic). When non-alcoholic beer
definition of craft beer (Acitelli & Magee, 2017; Elzinga, Tremblay, & was labeled incorrectly as regular beer, liking of the beer increased.
Tremblay, 2015) and craft beer drinkers (Aquilani et al., 2015; Gómez- When regular beer was labeled as non-alcoholic, liking did not change,
Corona et al., 2016) and behaviors that reflect regular purchase and use but the emotional responses to the beer were more negative.
of craft beers. Jaeger et al. (2020) concluded that their results showed
that the differences in these two flavor segments likely represent 1.3. Types of label information affecting judgments of beer
changes in flavor preferences among declared craft beer drinkers that
result from greater frequency and/or length of exposure to the novel With regard to the nature of the information that can have an effect
flavors and varieties of craft beers. on the liking of beer, Caporale and Monteleone (2004) showed a strong
In light of the two flavor preference segments found by Jaeger et al. effect of information about the manufacturing process of the beer.
(2020), the question arises as to whether these two segments respond Sester, Dacremont, Deroy, and Valentin (2013) showed that a wide
differently to information that identifies a beer as being more char- variety of other forms of information can affect perception of beer.
acteristic of a true craft-style beer or of a more mainstream beer. It These investigators had consumers give word associations to 14 dif-
might well be hypothesized that information that leads the beer drinker ferent brands of beer by either tasting them in a blind condition or by
to assume that any given beer is a true craft-style beer may create more holding and examining bottles containing their brand labels. Their re-
appeal for that beer among a high flavor preference segment, because sults showed that the simple tasting of beers elicited mental re-
such information would lead to the expectation that the beer will have presentations that were primarily related to the intrinsic characteristics
the bolder, more complex craft-style flavor profile that these consumers of the beer, i.e., its sensory properties. However, in the brand label
enjoy. In contrast, any information that evokes expectations of a more condition, these representations were much broader and related to the
traditional-style beer would create greater appeal among the low flavor semantic and visual information on the bottle. For example, mental
preference segment. representations having to do with “nature” were evoked by bottle labels

2
S.R. Jaeger, et al. Food Quality and Preference 87 (2021) 104043

containing images of animals; religious representations were evoked by whether the responses of these two segments are differentially affected
information on the labels that indicated manufacture by religious-af- by this information.
filiated (abbey) brewers; and even geographic representations were
evoked by information about the locale in which the beer was pro- 2. Research aim
duced. The latter type of representations can be important elements in
the perception of craft-style beers, because key characteristics of these A study was conducted to explore the potential effects of “craft” vs.
beers that make them popular among craft beer drinkers are the “mainstream” labeling information on responses to both craft- and
proximity of the locale in which they are brewed (as opposed to being mainstream-style beers and among craft beer drinkers who either prefer
brewed in another country or distant geographic locale) and the small high flavor impact beers or low flavor impact beers. The specific aim
size of the brewery (vs. a large mainstream brewery). was to investigate how labels that impart information about whether a
The above studies show that various elements of information about test beer is “craft” or “mainstream” and whether it is produced in a
a beer can have a profound effect on its sensory attributes, liking, and small, local brewery vs. a large, international, mainstream brewery,
other consumer-centric evaluative properties of the beer. Of additional influence consumers’ sensory, liking, holistic/conceptual, and cogni-
importance for the purpose of the present research are: 1) that these tive/emotional/situational use responses to the beers, and whether
effects can be more influential in younger consumers, who are typical of these label effects are dependent on the beer preferences of the con-
craft beer drinkers (Guinard et al., 2000, 2001) 2) that the manu- sumer.
facturing process of the beer (important to craft-style beer drinkers) can We hypothesize that, among a high flavor segment that prefers
be an important element in these effects (Caporale & Monteleone, novel and/or bolder flavors consistent with craft-style beers, labeling
2004), and 3) that the geographic location in which the beer is brewed craft-style beers as being from a small, local craft brewer will increase
(also important to craft beer drinkers) can evoke important pre-trial their liking of these beers and other positively valenced responses to
mental representations and expectations about the beer (Sester et al., them, while lowering any negatively valenced responses. In contrast,
2013). for the low flavor preference segment, such labeling may well decrease
liking and other positively valanced evaluations of these beers, while
1.4. Label information likely to differentially affect beer judgments among increasing any negatively valanced evaluations. Labeling the tradi-
declared craft beer drinkers tional-style beers as being brewed by a large mainstream brewery
outside of New Zealand (NZ) is predicted to decrease positive affect
As suggested by the research cited above, two elements of in- (increase negative affect) toward these beers among the high flavor,
formation that are important to many beer drinkers are the manu- craft-style flavor likers, but increase positive affect (decrease negative
facturing process of the beer and the geographic location in which the affect) among the low flavor, traditional-style flavor likers. In general,
beer is brewed. These two elements, along with the size of the brewery, we hypothesize that the label conditions will enhance most differences
are especially critical to craft beer drinkers and to the very definition of between the two flavor segments, insofar as the labels reinforce the
craft beers, because they serve to distinguish craft beers from more craft vs. traditional differences among the beers, essentially adding
traditional, mainstream beers. For example, Elzinga et al. (2015) has extrinsic information that confirms the intrinsic (sensory) character-
identified the two most important characteristics of craft beer as 1) the istics of the beers.
traditional nature of the beer production method, and 2) the small size
of the production facility in which it is brewed. Acitelli and Magee 3. Materials and methods
(2017) have defined craft beer as beer brewed in “any small, in-
dependently owned brewery that adheres to traditional brewing prac- 3.1. Experimental overview and data sources
tices and ingredients.” Among Italian beer drinkers, Donadini and
Porretta (2017) have also reported that craft beers are those brewed on The experimental approach combined tasting of beer samples with
a small scale, using premium local and unconventional ingredients; that questionnaires about beer-related attitudes and behaviors. The tasting
they are unpasteurized, unfiltered, additive-free, natural and genuine; component comprised a multi-response measurement approach, where
that they use heirloom grains and are spontaneously and naturally acceptability, sensory characterization, holistic/conceptual percep-
fermented; that they are sold in specific outlets only and are brewed for tions, and cognitive / emotional / situational responses were obtained
a niche market. Lastly, Gómez-Corona et al. (2016) have reported that for beers with “high” and “low” flavor intensity. Consumers were al-
beer enthusiasts describe craft beer as being “produced on a small located to one of two information conditions in which the beer samples
scale,” “that has a limited production,” and “that is carefully produced” were presented either without information (B: blind) (n = 120) or with
or produced by “a more complicated process.” Taking the above find- information (IN: Information) (n = 122) (Table 1). Further, in the IN
ings into consideration, it is clear that craft beers stand in stark contrast condition, the label information given to participants was linked to the
to mainstream beers that are produced by large, international brewers flavor intensity of the beer sample such that beers with “low” flavor
in large batches, using standard ingredients, and being subjected to intensity (LF: L1, L2, L3) were paired with the information: “Brewed by
high levels of quality control, filtering, processing and, often, con- a large mainstream brewery outside of New Zealand,” and beers with
taining additives (Elzinga et al., 2015). “high” flavor intensity (HF: H1, H2, H3) were paired with the in-
Based on the above, it would seem that the informational elements formation: “Brewed by a small craft brewery in New Zealand.” Ques-
that would be most effective in characterizing a beer as being a craft- tionnaires (completed after beer tasting) were used to gain information
style vs. a traditional-style beer are the size of the brewery, the man- about participants’ beer-related attitudes and beer-specific behaviors.
ufacturing process of the beer, and whether the beer is brewed locally. The study was a direct extension of Jaeger et al. (2020) with the
Moreover, since it has been shown that most beer drinkers consider the data from consumers in that study (n = 120) representing the B con-
characterizing element of craft beers to be that they possess “rich, in- dition of the present research. The data from consumers in the IN
tense and composite” flavor and “unusual sensory notes” (Donadini & condition (n = 122) have not previously been reported but were col-
Porretta, 2017), it can be predicted that declared craft beer drinkers lected at the same time as those for the B condition and under identical
who prefer these latter flavors would respond more favorably to in- conditions, allowing a direct comparison of the two information con-
formation that a beer is craft-style vs. traditional-style. In the present ditions. Thus, when we refer to the B condition in the present paper, we
research we examine whether the two preference clusters found in are re-using the data from Jaeger et al. (2020) for the purpose of un-
Jaeger et al. (2020) still emerge in declared craft beer drinkers when derstanding how label information (B vs. IN) modulates consumer re-
the beers are labeled according to their defining characteristics and sponses to craft-style and traditional-style beers.

3
S.R. Jaeger, et al. Food Quality and Preference 87 (2021) 104043

Table 1
Overview of the study which implemented a between-subjects design with two information conditions (B and IN) where consumers either evaluated the samples blind
(B condition, n = 120) or with information (IN condition, n = 122). In the IN condition, the information given to participants was linked to the flavor intensity of the
beer sample such that beers with “low” flavor intensity (LF: L1, L2, L3) were paired with the information: Brewed by a large mainstream brewery outside of New
Zealand. Beers with “high” flavor intensity (HF: H1, H2, H3) were paired with the information: Brewed by a small craft brewery in New Zealand. The B condition data
were from Jaeger et al. (2020).
Beer Information in B condition Information in IN condition: Brewed by a… Beer name

L1 None large mainstream brewery outside of New Zealand Mac’s Gold


L2 None large mainstream brewery outside of New Zealand Budweiser
L3 None large mainstream brewery outside of New Zealand Carling Black Label
H1 None small craft brewery in New Zealand Mac’s Sassy Red
H2 None small craft brewery in New Zealand Emerson 1820 IPA
H3 None small craft brewery in New Zealand Suffolk Vintage Ale

Note. IPA – India Pale Ale.

3.2. Participants Serving size (55–60 mL) was determined to ensure that the total
intake of alcohol during each research session was < 1.5 New Zealand
The research drew on responses from 242 New Zealand male con- standard drinks (< 15 g pure alcohol). This requirement was imposed
sumers between 20 and 66 years old with diverse socio-economic by the study’s ethics approval.
backgrounds (Supplementary Table 1 has full details). The total sample The beers were stored at 1–2 °C and served within 5 min of being
included the 120 participants from Jaeger et al. (2020) and an addi- removed from refrigeration (~5 °C). Clear glasses (tumbler shape,
tional 122 participants who were recruited according to the same cri- 150 mL) labelled with 3-digit random codes were used (see Jaeger et al.
teria used by Jaeger et al. (2020). (2020) for an image of the samples as presented to consumers).
Male consumers were chosen because, demographically, they con-
stitute the largest proportion of craft beer drinkers. Selection criteria
included declared liking for beer, declared consumption at least weekly, 3.4. Product labels
and a declared preference for beers from smaller craft-style breweries
whose beers have less traditional, typically bolder and stronger flavor, As noted in the Introduction (Section 1.1), two of the key char-
e.g. hoppy pale ales, fruit beers or chocolate stouts. acteristics of craft-style beers are that they are manufactured using
The participants attended one 2¼ hour research session at the Plant traditional brewing techniques and that they are brewed in small, local
& Food Research (PFR) Consumer Insights Facility in Auckland. Prior to breweries. These characteristics of craft-style beers have, in fact, been
each session, participants gave written informed consent and took a shown to drive craft beer drinkers’ perceptions and conceptualizations
breathalyzer test for alcohol content (to ensure no measurable blood of these beers, with Gómez-Corona et al. (2016) showing that craft beer
alcohol concentration prior to start of the session). Compensation in drinkers are driven to seek out authentic beers, Caporale and
cash was given at the end of the session, and breathalyzer testing before Monteleone (2004) showing that information about the beer manu-
leaving the research facility was compulsory for all participants. The facturing process affects liking of beers, and Sester et al. (2013)
study was approved by the PFR Human Research Ethics Committee. showing that the geographic location in which the beer is brewed is an
important element in generating pre-trial mental representations and
expectations about the beer.
3.3. Samples With this in mind, we developed two distinct information labels that
addressed these critical elements of information about the beers:
3.3.1. Beer selection “Brewed by a small craft brewery in New Zealand” and “Brewed by a
Table 1 lists the 6 beers used in the research, which were identical to large mainstream brewery outside of New Zealand.” These two labels
Jaeger et al. (2020). Beers in the “high flavor” (HF) intensity group were then used to describe either the HF beers “Brewed by a small craft
(beers H1, H2 and H3), besides being stronger in flavor intensity brewery in New Zealand,” or the LF beers “Brewed by a large main-
compared to beers in the ”low flavor” (LF) group, were selected to re- stream brewery outside of New Zealand” in the label information (IN)
present two types of strong flavor expression common to craft beers – condition of the study.
hoppy (bitter) and malty (sweet with hints of caramel, toffee and nuts). The labels were pilot tested, first with 22 people who were asked to
H2 was selected to represent an imperial India pale ale style with high name breweries/beer brands that they associated with each of the two
hop bitterness. H3, in comparison, was not hoppy or bitter, but malty labels. In response to “brewed by a small craft brewery in New Zealand”
and sweet. H1 was selected to be an intermediary between H2 and H3. participants correctly listed Garage Project, Tuatara, Emersons, Epic
In the “low flavor” (LF) intensity group (beers L1, L2 and L3), beers and ~ 50 other smaller craft breweries. The most common responses to
were selected to be lager style, with variation in flavor expression ef- “brewed by a large mainstream brewery outside of New Zealand” were
fected by including: i) a sample with expected very low flavor while not Heineken, Budweiser, Corona and Carlsberg. These results strongly
being low in alcohol content (L2), ii) a sample with higher alcohol indicated that the two information labels were appropriately inter-
content than standard for stronger lager flavor expression (L3), and iii) preted by consumers. The second step was to confirm that the two la-
a sample with slight sweet characteristics (L1). bels were associated with expected product characteristics. Next, an-
All samples were commercially available in New Zealand and pur- other 88 people (self-declared beer drinkers) were asked to imagine
chased at local supermarkets and liquor stores, and the alcohol content beers produced as described by the different labels and rate them using
of beers in the LF and HF groups were chosen to be as similar as possible 7-pt semantic differential scales (low quality – high quality, undesirable
(see Jaeger et al. (2020) for additional details regarding the samples). – desirable, weak flavour – high flavour, liked by me – disliked by me,
popular – unpopular). Paired t-tests confirmed that the beers linked to
3.3.2. Sample presentation the label “brewed by a small craft brewery in New Zealand” were
Sample presentation was identical for the B and IN conditions, since perceived as significantly higher in quality, desirable and strong flavor
label information was presented on the ballots that accompanied the (p ≤ 0.002). Expected liking and popularity were not significantly
individual samples. different, which fits with the preference segmentation for craft-style

4
S.R. Jaeger, et al. Food Quality and Preference 87 (2021) 104043

Table 2
Response variables and wording used during taste testing of beers and in the post-tasting questionnaire. The number of items for each type of response variable is
shown between parentheses.
Variable Wording

Responses during beer tasting


Sensory (11) ‘weak aroma’, ‘strong aroma’, ‘sweet’, ‘bitter’, ‘hoppy’, ‘malty’, ‘oaky/wood-like’, ‘fruity’, ‘watery/bland’, ‘short aftertaste, ‘long
aftertaste’
Holistic and conceptual (18) Holistic: ‘balanced’, ‘complex’, ‘easy-to-drink’, ‘full-bodied’, ‘smooth’, ‘thirst quenching’
Conceptual: ‘artisan’, ‘cheap’, ‘comforting’, ‘familiar’, ‘industrial’, ‘insipid’, ‘masculine’, ‘memorable’, ‘powerful’, ‘simple’,
‘sophisticated’, ‘substantial’
Cognitive, emotional and situational (13) Cognitive: “I consider this to be a good quality beer,” “This is a unique beer,” “In my opinion, this tastes like a craft beer,” “In my
opinion, this tastes like a mainstream beer,” “I think “beer aficionados” would enjoy this beer,” “This beer has an authentic taste,” “This
beer reminds me of those I typically drink”
Emotional: “This beer evokes a positive emotion in me,” “This beer bores me” (boring)
Situational: “When I want to savour a beer by myself, this is the kind of beer I would drink,” “This is a good beer for many different
social situations,” “If I wanted to drink more than one glass of the same beer, I’d chose this one,” “I would drink this beer in the
company of like-minded beer drinkers,”

Responses obtained in the post-tasting questionnaire


General attitudes to beer (5) “I prefer beer with a strong, full-bodied taste,” “I seek out beers made by small, local, independent breweries,” “I like beers that are of
high quality and authenticity,” “I don’t usually care for industrial beers that are mass-distributed,” “I choose my beer carefully”
Attitudes to craft beer (4) “I love to experiment with different styles of craft beer,” “I see myself drinking more craft beer in the future than I do today,”
“Seeking out and drinking craft beer has become a hobby for me,” “Craft beer has changed my perception of what constitutes high
quality beer”
Craft beer involvement (3) “To me, craft beer matters”, “I have a strong interest in craft beer”, “Craft beer is important to me”
Craft beer activities and behaviours (16) “I have visited and tasted beers at a local craft brewery,” “I have researched different types of beers and brewing styles on-line,” “I
have attended a “beer fest” or other organized beer tasting event,” “I visit on-line chat rooms that discuss new craft beers and
breweries,” “I am on a mailing list of a local brewery,” “I have purchased special beer glasses designed to hold different types of
beers,” “I purchase and refill flagons of beer at local breweries,” “I have a beer refrigerator in my home,” “I have a beer tap in my
house,” “I own shirts or hats with brewer names and logos on them,” “I purchase and/or trade specialty beers on-line,” “I will drive
several hours to go to a craft brewery I like in order to purchase their beer,” “I participate in on-line beer rating sites to express my
opinion about different beers,” “I am willing to pay a very large price for craft beers that are new to the market,” “I read beer
magazines and publications,” “I buy beer from beer specialty stores”

and traditional-style beers found in Jaeger et al. (2020). Pilot work contained written information about the test samples. This was printed
participants were not eligible for the main study. in large letters at the top of each ballot page in accordance with Table 1.
Sample assessments took place in standard sensory testing booths
(white lighting, 20–22 °C, positive air flow) during sample assessments.
3.5. Empirical procedures Between samples there was a break of 1–2 min, and during this time
water and plain crackers were available for palate cleansing.
3.5.1. Measurements obtained from consumers in response to tasted craft- The study’s human ethics approval stipulated that the beer tasting
style and traditional-style beers tasks be followed by other tasks involving food tasting in order to fa-
All response variables were identical to those used in Jaeger et al. cilitate metabolism of the consumed alcohol. Because these filler tasks
(2020) and are shown in Table 2. The ballot was structured in the order: were unrelated to the present research, they are not discussed further.
i) acceptability and sensory characteristics, ii) holistic/conceptual
perceptions, and iii) cognitive/emotional/situational responses. Ac-
ceptability was measured using a fully labelled 9-pt category scale with 3.5.2. Questionnaire responses obtained from consumers
end-point anchors 1 = ‘dislike extremely’ and 9 = ‘like extremely’ Following sample evaluations, participants completed a ques-
(Peryam & Pilgrim, 1957). Sensory product descriptions were obtained tionnaire about beer. All response variables were identical to those used
with a check-all-that-apply (CATA) ballot (Ares & Jaeger, 2015; Jaeger, in Jaeger et al. (2020) and are also shown in Table 2 with additional
Hunter et al., 2019) containing 11 terms relevant for the focal set of detail below.
samples and for discrimination between beers with “low” and “high” Participants’ general attitudes to beer were obtained through five
flavor intensity. A second CATA ballot was used to obtain responses statements. Focusing on craft beer, four statements were used, and craft
relating to holistic sensory characteristics and other product con- beer involvement was measured on three items drawing on Hollebeek,
ceptualizations. With input from, for example, Jaeger, Xia et al. (2019), Jaeger, Brodie, and Balemi (2007) and directly adapted from Jaeger,
18 terms were selected to be relevant for: i) discrimination between Xia et al. (2019). Responses were obtained on fully labelled 7-pt Likert
“low” and “high” flavor intensity beers, and ii) characterization of beers scales (1 = ‘disagree strongly’ to 7 = ‘agree strongly’). Statement order
by the two different information labels used in the study. Cognitive/ was randomized across participants (Ares et al., 2015).
emotional/situational aspects of the beers (n = 13) were measured Craft beer activities and behaviors (n = 16) expected to be pre-
based on Cardello et al. (2016); Jaeger et al. (2017); Jaeger, Xia et al. valent among craft beer drinkers and applicable to the New Zealand
(2019); Gómez-Corona, Escalona-Buendía et al. (2017); Gómez-Corona, context was constructed, based on findings from the literature (Aquilani
Chollet et al. (2017), and Giacalone et al. (2015), as well as pilot work et al., 2015; Thurnell-Read, 2016; Gómez-Corona, Escalona-Buendía
with the focal set of samples. Responses were obtained on fully labelled et al., 2017; and Gómez-Corona, Chollet et al., 2017). A forced Yes-No
7-pt Likert scales (1 = “disagree strongly”; 7 = “agree strongly”). question was used, with statement order randomized across partici-
Sample presentation was monadic in accordance with a design pants (Ares et al., 2015).
based on a Williams Latin Square. Statement order for response vari- Stated liking was recorded on a 9-pt scale from 1 = ‘dislike ex-
ables was randomized across participants (Ares et al., 2015). tremely’ to 9 = ‘like extremely,’ while an 8-pt scale from ‘daily’ to
In the blind (B) condition, ballots were as described above with no ‘never’ was used to capture frequency of beer consumption.
additional information about individual samples. In the information At the end of the research session, demographic and socio-economic
condition (IN), as dictated by the experimental design, sample ballots information was obtained.

5
S.R. Jaeger, et al. Food Quality and Preference 87 (2021) 104043

Table 3 which were combined with the responses for the 122 participants in the
Average scores for overall liking for beers included in the study, obtained from IN condition. The analyses were performed within each preference
122 declared craft beer drinkers in the information label (IN) condition. The cluster – CS Likers and TS Likers, respectively.
results are also shown within consumer segment, where CS Likers (n = 65) For responses obtained during beer tasting, two-way ANOVAs were
prefer craft-style beers with “high” flavor intensity, while TS Likers (n = 57)
performed using beer sample (BS; H1-3, L1-3), information condition
prefer traditional-style beers with “low” flavor intensity. Results from Turkey’s
(IC; B, IN) and their interaction as factors. These univariate analyses
HSD post-hoc test (5% level) are shown using letters, performed for comparison
were performed for liking scores and each of the sensory (n = 11),
within preference clusters (i.e., within columns). The responses were measured
on 9-pt scales where 1=’dislike extremely’ and 9=’like extremely.’ holistic/conceptual (n = 18), and cognitive/emotional/situational
(n = 13) responses collected during product tasting. For sensory and
Beer Aggregate CS Likers TS Likers
holistic/conceptual responses obtained as CATA responses, ANOVA was
L1 5.5 cd
4.6c 6.7 a chosen over logistic regression. While the latter is more statistically
L2 5.5 d
4.5c 6.6 a correct, ANOVA is robust and acceptable for the sample sizes used in
d
L3 5.4 4.6c 6.2 ab this research and its use has been demonstrated and justified by Worch
ab
H1 6.5 7.3 ab 5.7 bc et al. (2020). This choice also made direct comparison between re-
a
H2 6.9 7.7 a 5.9 bc
H3 6.2 bc
6.9b 5.4c
sponse measures possible. For post-taste questionnaire responses, the B
and IN conditions were compared using t-tests which were performed
on the individual beer attitudinal (n = 12) and behavioral variables
3.6. Data analysis (n = 17).
To further compare B and IN responses, Multiple Factor Analysis
All analyses were performed in the R language (Core, 2019), using (MFA) was applied to, respectively, the sensory, holistic/conceptual
5% as the level of significance. and cognitive/emotional/situational datasets, retaining two-dimen-
sional solutions.

3.6.1. Analyses performed for the IN condition


4. Results
The analyses performed for the information (IN) information con-
dition (n = 122) were identical to those performed and reported in
We begin by presenting the results from the label information
Jaeger et al. (2020) on the B information condition data. Specifically, to
condition (IN) (Section 4.1) and refer readers to Jaeger et al. (2020) for
establish consumer segments based on product liking/disliking Hier-
the same analyses and associated findings linked to the blind (B) in-
archical Cluster Analysis was performed (unstandardized scores, Eu-
formation condition. The results linked to the comparison of responses
clidian distance, Ward’s method) and two-cluster solutions retained in
obtained in the IN condition vs. the B condition follow in Section 4.2.
accordance with a scree plot. Within each preference cluster, analyses
were performed on the responses obtained during product tasting and
4.1. Responses obtained in the label information (IN) condition
from survey responses, as follows.
Ratings of product acceptability were analyzed by ANOVA using
4.1.1. IN condition: Preference-based consumer segments
beer samples as a fixed factor and treating respondents as a random
Table 3 shows the liking data for the two flavor preference segments
effect. Tukey’s method was applied for post-hoc testing. Also, within
identified by performing hierarchical cluster analysis on participants in
each cluster, CATA responses were analyzed separately for the two
the IN condition of the study. As can be seen, one group of consumers
question types (sensory and holistic/conceptual) using standard pro-
(n = 65), who were designated as CS Likers, preferred all of the novel,
cedures. Cochran’s Q test (Manoukian, 1986) was used to identify sig-
“high flavor” craft-style beers (H1–H3) labeled as being brewed by a
nificant differences among samples for the frequency of use of each
small local brewery over all of the “low flavor” impact, traditional-style
CATA term. When differences were significant, the McNemar test was
beers (L1–L3) labeled as being brewed in a large, commercial brewery
used for pairwise comparisons between product. A Correspondence
outside of New Zealand. Another group (n = 57), who were designated
Analysis (CA) on the frequency table was carried out to obtain a bi-
as TS Likers, preferred all of the traditional-style beers (L1–L3) labeled
dimensional representation of samples and terms using the chi-square
as coming from large international breweries over all of the craft-style
distance. Finally, the responses to Likert statements were analyzed
beers (H1–H3) labeled as coming from small local brewers.
using the ANOVA procedures applied to acceptability scores, and
A comparison of these liking data to the liking data reported in
Principal Components Analysis (PCA) of the correlation matrix. Con-
Table 3 of Jaeger et al. (2020) for the two similarly-identified consumer
fidence ellipses (95%) around product positions in CA and PCA spaces
segments who tasted the same beers in a blind condition reveals a
were obtained using a bootstrapping procedure (Cadoret & Husson,
highly similar pattern of responses, including the fact that the CS Likers
2013).
segments in both conditions preferred samples H1 and H2 more than
The post-tasting questionnaire responses were analyzed using the
H3. In addition, it can be seen that the CS Likers segment in the IN
same ANOVA as above. Involvement with craft beer was established as
condition liked all of the HF beers more than did the CS Likers segment
a single index for each participant, summing across the three items
in the B condition of Jaeger et al. (2020), while liking all of the LF beers
(possible range: 3–21). In all sub-group analyses this was supported by
the same or less. For the TS Likers segments, no systematic differences
Cronbach alpha ≥ 0.8, which exceeded the 0.7 threshold (Nunnaly,
between responses in the two information conditions were apparent.
1978). The summed score was subject to the same ANOVA as above.
The responses to each of the 16 statements about specific activities
4.1.2. IN condition: Sensory and holistic/conceptual characteristics of
related to beers were coded as ‘yes’ = 1 and ‘no’ = 0, and analyzed
tasted craft-style and traditional-style beers
using logistic regression, with factor specification similar to the AN-
Fig. 1 (top) shows the biplots of Correspondence Analyses (CA)
OVAs used for continuous variables.
conducted on the sensory responses for the CS Likers segment (1a) and
the TS Likers segment (1b). As can be seen, for both consumer segments,
3.6.2. Analyses performed to compare responses in the B and IN the 3 HF and 3 LF beers were highly differentiated along Dimension 1
information conditions (80% and 88% explained inertia). Furthermore, the sensory attributes
Comparisons of responses obtained in the blind (B) and label in- associated with the beers were very similar for the two segments, with
formation (IN) conditions were performed. To this end we drew on the both preference segments characterizing the LF beers as ‘watery,’
data from Jaeger et al. (2020) to represent the B condition (n = 120), having ‘weak aroma’ and a ‘short aftertaste,’ while the HF beers were

6
S.R. Jaeger, et al. Food Quality and Preference 87 (2021) 104043

Fig. 1. Results for the label information (IN) condition on six tasted beer samples: L1, L2, L3, H1, H2, and H3 (refer to Table 1 for full sample details). Shown are
biplots spanned by the first two dimensions following Correspondence Analysis. Top row: (a) and (b) sensory product perceptions, with CATA terms and sample
positions for CS Likers (who prefer craft-style beers with “high” flavor intensity) and TS Likers (who prefer traditional-style beers with “low” flavor intensity),
respectively; and bottom row: (c) and (d) conceptual / holistic product perceptions, with CATA terms and sample positions for CS Likers and TS Likers respectively.
Average samples positions shown with 95% confidence ellipses. See Table 2 for full wordings.

characterized as ‘bitter,’ ‘hoppy,’ having a ‘strong aroma’ and a ‘long attributes of the beers similarly.
aftertaste.’ Furthermore, both preference segments perceived beer
sample H3 as having ‘oaky’ and ‘malty’ flavors.
Fig. 1 (bottom) shows the CA results on the holistic/conceptual data 4.1.3. IN condition: cognitive/emotional/situational product responses to
for the CS Likers (1c) and TS Likers (1d) segments. As can be seen in the tasted craft-style and traditional-style beers
bi-plots for both preference segments, the 3 HF and 3 LF beers were Fig. 2 shows the results of Principal Components Analyses (PCA)
again highly differentiated along Dimension 1 (> 92% inertia ex- conducted on the cognitive/emotional/situational data from the CS
plained). Furthermore, the holistic/conceptual attributes associated Likers and TS Likers consumer segments for the 3 HF and 3 LF beers.
with the 3 HF and 3 LF beers were highly similar between the two Like both the sensory and holistic/conceptual data (Fig. 1), for both
consumer segments, with both segments characterizing the LF beers as consumer segments, the 3 HF and 3 LF beers were highly differentiated
‘simple,’ ‘familiar,’ ‘easy to drink,’ and ‘cheap,’ while the HF beers were along Dimension 1 (85–90% explained variance; Fig. 2a and b). How-
characterized as ‘substantial,’ ‘masculine,’ ‘complex,’ ‘full-bodied,’ ‘ar- ever, unlike the data shown in Fig. 1, the cognitive/emotional/situa-
tisan,’ ‘powerful’ and ‘memorable.’ Like the sensory data shown in tional responses (Fig. 2c and d) differed between segments.
Fig. 1a and b, the holistic/conceptual data again show that, regardless For the CS Likers segment, the LF beers were characterized by only
of flavor preferences, both segments perceived the holistic/conceptual two statements, “this beer bores me” and “in my opinion, this tastes like
a mainstream beer” (Fig. 2c). Since the consumers in this cluster

7
S.R. Jaeger, et al. Food Quality and Preference 87 (2021) 104043

Fig. 2. Results for the label information (IN) condition on six tasted beer samples: L1, L2, L3, H1, H2, and H3 (refer to Table 1 for full sample details). Shown are plots
spanned by the first two dimensions following Principal Components Analysis. Top row: (a) and (b) product spaces for CS Likers (who prefer craft-style beers with
“high” flavor intensity) and TS Likers (who prefer traditional-style beers with “low” flavor intensity), respectively; and bottom row: (c) and (d) variable spaces
showing cognitive/emotional/, situational responses for CS Likers and TS Likers, respectively. Average samples positions shown with 95% confidence ellipses. See
Table 2 for full wordings.

preferred HF beers, this characterization of the LF beers was to be ex- many of the same statements used by the CS likers, i.e. “this beer bores
pected. The HF beers, on the other hand, were characterized by the me” and “in my opinion, this tastes like a mainstream beer” (Fig. 2d).
statements “this beer reminds me of those I typically drink,” “I think However, the TS Likers segment further characterized and differentiated
beer aficionados would enjoy this beer,” “if I wanted to drink more than the LF beers along Dimension 2 with the statements “if I wanted to
one glass of the same beer, I’d chose this one,” “when I want to savour a drink more than one glass of the same beer, I’d chose this one,” “this is a
beer by myself, this is the kind of beer I would drink,” “I consider this to good beer for many different social situations,” “this beer reminds me of
be a good quality beer,” “this beer evokes a positive emotion in me,” those I typically drink,” “this beer evokes a positive emotion in me,”
“this beer has an authentic taste,” “this is a unique beer,” and “in my and “I would drink this beer in the company of like-minded beer drin-
opinion, this tastes like a craft beer” (Fig. 2c). All of these character- kers” (Fig. 2d). All of these latter statements were used by the CS Likers
izations are what would be expected from a craft beer drinker who to, instead, characterize the HF beers (see above). However, the TS
prefers more complex, bolder and more intensely flavored beers (i.e., Likers segment, like the CS Likers, did characterize the HF beers with the
beers H1-3). statements “this beer has an authentic taste,” “this is a unique beer,” “in
In contrast to the data for the CS Likers segment, the comparable my opinion, this tastes like a craft beer,” “I think “beer aficionados”
data for TS Likers, who prefer LF beers, are somewhat different. For would enjoy this beer,” “I consider this to be a good quality beer,” and
these consumers, the LF beers were characterized along Dimension 1 by “when I want to savour a beer by myself, this is the kind of beer I would

8
S.R. Jaeger, et al. Food Quality and Preference 87 (2021) 104043

Table 4
Attitudes to beer and beer-related behaviors measured in post-tasting questionnaire. Mean values based on survey responses for consumers in label information (IN)
condition, where CS Likers (n = 65) prefer craft-style beers with “high” flavor intensity, while TS Likers (n = 57) prefer traditional-style beers with “low” flavor
intensity. P-values following comparison of means for preference clusters are shown, using bold font to identify those where p < 0.05.
Questions CS Likers TS Likers p-value

1. Attitudes toward beer*


1a. Stated liking for beer 8.3 8.0 0.08
1b. I prefer beer with a strong, full-bodied taste 6.2 5.1 < 0.001
1c. I seek out beers made by small, local, independent breweries 5.8 4.7 < 0.001
1d. I like beers that are of high quality and authenticity 6.4 5.5 < 0.001
1e. I don’t usually care for industrial beers that are mass-distributed 5.3 4.0 < 0.001
1f. I choose my beer carefully 6.0 5.5 0.009
1g. Craft beer has changed my perception of what constitutes high quality beer 6.3 5.3 < 0.001
1h. I love to experiment with different styles of craft beer 6.1 5.4 0.003
1i. Seeking out and drinking craft beer has become a hobby for me 4.9 3.9 0.002
1j. I see myself drinking more craft beer in the future than I do today 5.9 5.4 0.009
1k. Involvement with craft beer 17.7 15.0 < 0.001

2. Beer-related behaviors**
2a. Frequency of drinking beer 1.9 2.1 0.09
2b. I purchase and refill flagons of beer at local breweries 0.29 0.12 0.023
2c. I have a beer tap in my house 0.05 0.09 0.36
2d. I have visited and tasted beers at a local craft brewery 0.77 0.83 0.45
2e. I read beer magazines and publications 0.15 0.18 0.75
2f. I have attended a “beer fest” or other organized beer tasting event 0.62 0.63 0.85
2g. I will drive several hours to go to a craft brewery I like in order to purchase their beer 0.23 0.21 0.79
2h. I have a beer refrigerator in my home 0.49 0.42 0.43
2i. I own shirts or hats with brewer names and logos on them 0.37 0.44 0.44
2j. I am on a mailing list of a local brewery 0.20 0.18 0.73
2k. I buy beer from beer specialty stores 0.62 0.53 0.32
2l. I am willing to pay a very large price for craft beers that are new to the market 0.45 0.29 0.07
2m. I have researched different types of beers and brewing styles on-line 0.55 0.32 0.009
2n. I purchase and/or trade specialty beers on-line 0.14 0.18 0.58
2o. I participate in on-line beer rating sites to express my opinion about different beers 0.14 0.14 0.98
2p. I have purchased special beer glasses designed to hold different types of beers 0.42 0.54 0.16
2q. I visit on-line chat rooms that discuss new craft beers and breweries 0.03 0.05 0.55

Notes. *) For (1a) responses obtained on an 9-pt scale from 1 = ’dislike extremely’ to 9 = ’like extremely.’ For (1b)–(1j) responses obtained on a 7-pt scale from
1 = ’strongly disagree’ to 7=’strongly agree.’ For (1 k) summed values across 3 statements are shown, each measured on 7-pt scales from 1 = ’strongly disagree’ to
7 = ’strongly agree.’ Higher values indicate greater involvement with craft beer. **) For (2a) responses obtained on an 8-pt scale from 1 = ’daily’ to 8 = ’never’. For
(2b)–(2q) responses obtained as ‘yes’ = 1 or ‘no’ = 0 and mean values calculated.

drink,” again differentiating the beers along Dimension 2 (Fig. 2d). 4.2. Information effects: comparison of blind (B) and label information
Thus, despite preferring the flavor of the 3 LF beers, TS Likers char- (IN) conditions
acterized the HF beers as did the CS Likers.
4.2.1. Information effects on overall liking of tasted craft-style and
traditional-style beers
4.1.4. IN condition: Post-tasting questionnaire responses Having described the findings among participants in the IN condi-
The results of the post-tasting questionnaire responses among par- tion of the study, we progress to the comparison of the B and IN con-
ticipants in the IN condition can be seen in Table 4. In the case of beer- ditions and combine the data from the 122 people who evaluated beers
related attitude statements (Table 4, top), with the exception of their with label information present to those 120 consumers that in Jaeger
stated liking for beer, every statement showed a significant effect be- et al. (2020) evaluated the same beers under identical conditions
tween the CS Likers and TS Likers segments, and in each case, the CS without label information (i.e., B condition). The analyses are per-
Likers agreed significantly more with the statements than did the TS formed and described within each preference segment.
Likers. Given the nature of these statements, which all reflected atti- Table 5 shows the results of two-way ANOVAs examining the effects
tudes and involvement with beer that craft beer drinkers are more likely of beer samples (BS; H1-3, L1-3) and information condition (IC; B, IN)
to hold, the results are highly consistent with expectations for these two conducted on all responses obtained during beer tasting. The table
flavor preference segments. The fact that the two segments liked beer entries are the percentages of significant main and interaction effects
equally is important, as it shows that the consumers did not differ in (p < 0.05) out of all possible attribute effects. For both CS Likers and
how much they enjoy beer, regardless of their flavor preference. TS Likers, the beer sample effect was significant, and the information
With regard to beer-related behaviors (Table 4, bottom) there were condition effect was insignificant. A significant BS × IC interaction
only two statements that differentiated the flavor segments, “I purchase effect among CS Likers revealed that overall liking for beer samples was
and refill flagons of beer at local breweries” and “I have researched not identical in the two information conditions, which were attributed
different types of beers and brewing styles on-line.” In both cases, the to H3 and L2. For the former, liking was higher in the presence of label
CS Likers were more likely to report these behaviors than TS Likers, information than without (6.9 vs. 6.0), suggesting that the malty and
which, again, is consistent with the expected outcome, since the be- oaky characteristics of H3 were more appreciated in the context of the
haviors were selected to be those that craft beer drinkers would likely sample being identified as a craft beer. In the case of L2, the opposite
exhibit. effect was found, meaning that it was more liked in the B than IN
condition (5.0 vs. 4.5), which suggested that identifying the beer as
“mainstream” negatively influenced liking.
For TS Likers, while the BS × IC interaction effect was non-

9
S.R. Jaeger, et al. Food Quality and Preference 87 (2021) 104043

Table 5 Fig. 3a, the dominant separation of beers in the IN condition was along
Results from 2-way ANOVA using beer sample (BS: H1-3, L1-3), information Dimension 1 with the 3 HF beers grouping on one side of the sample
condition (IC: B, IN) and their interaction (BS × IC) as fixed effects. The per- space and the 3 LF beers grouping on the other side. These product
centage (%) of response variables that were significant at the 5% level is show characterizations were very similar to those obtained in the IN condi-
for each type of responses obtained during the tasting of craft-style and tradi-
tion, where H3 was characterized as ‘malty’ and ‘oaky,’ H2 was char-
tional-style beers. Results are shown separately for the two preference clusters
acterized as ‘hoppy’ and the LF beers were described as ‘weak/watery’
where CS Likers prefer craft-style beers with “high” flavor intensity and TS
and having a ‘short aftertaste’ (Fig. 1a and b)
Likers prefer traditional-style beers with “low” flavor intensity. The responses
included overall liking (1 variable), 11 sensory variables, 18 holistic/con- For the TS Likers segment, the effect of beer samples (BS) was the
ceptual variables and 13 cognitive/emotional/situational use variables. Data same as for the CS Likers, i.e. significant for all 11 sensory attributes
from the B condition come from Jaeger et al. (2020). (Table 5) (Supplementary Table 2 has full details). Although the effect
of information condition reached significance for five attributes
Response variables BS IC BS × IC
(‘fruity,’ ‘malty,’ ‘oaky/wood-like,’ ‘weak aroma,’ and ‘sweet’), there
CS Likers was only one significant interaction effect (‘malty’). This pointed to a
Overall liking (1) 100 0 100 slightly larger effect of label information than seen in the CS Likers data.
Sensory (11) 100 0 9 The above findings are confirmed by examining the MFA bi-plot for the
Holistic/conceptual (18) 89 17 11
Cognitive/emotional/situational (13) 100 8 100
TS Likers in Fig. 3b. Here, the overall separation of beer samples was not
different than that for the CS Likers (Fig. 3a), showing the same two
TS Likers
distinct groupings of HF and LF beers and the same better differentia-
Overall liking (1) 100 0 0
Sensory (11) 100 45 9 tion of the HF beers than the LF beers. The dominant sample char-
Holistic/conceptual (18) 94 39 11 acterizations were also unchanged, such as, for example, ‘malty’ and
Cognitive/emotional/situational (13) 77 23 77 ‘oaky/wood-like’ being associated with H3. Closer inspection of the
sample configurations suggested a somewhat closer spatial grouping of
HF and LF beers when product labels were presented (IN condition),
significant (Table 5), there was nonetheless a notable difference for L3, which could be indicative of consolidating sensory perceptions among
which was more liked in the B than the IN condition (7.0 vs. 6.2). the TS Likers. However, this effect of label information was small, and
the overall conclusion was that the both TS Likers and CS Likers were
relatively uninfluenced by the information conditions.
4.2.2. Information effects on sensory product perceptions of tasted craft-
style and traditional-style beers
With regard to sensory product perceptions, Table 5 shows that the 4.2.3. Information effects on holistic/conceptual product perceptions of
effect of beer samples (BS) was highly significant for all of the judged tasted craft-style and traditional-style beers
attributes in the CS Likers segment, while the effect of information The analyses examining the effects of beer samples and information
condition (IC) was never significant (Supplementary Table 2 has full condition among CS Likers found, with exceptions for ‘balanced’ and
details). In terms of BS × IC interactions, only ‘short aftertaste’ was ‘smooth,’ that the holistic/conceptual attributes varied significantly by
significant for this group of consumers, and, overall, this pointed to beer sample, while the effect of information condition was non-sig-
only a minor effect of label information on sensory product character- nificant for all but three attributes (‘familiar,’ ‘full-bodied’ and ‘so-
izations. This conclusion was supported by examination of the data in phisticated’). Significant interaction effects (BS × IC) were only found
Fig. 3, which shows the two-dimensional bi-plot of the MFA analyses for for ‘cheap’ and ‘powerful.’ These findings are summarized in Table 5
both the CS Likers (Fig. 3a) and TS Likers (Fig. 3b) segments in both the with full details in Supplementary Table 3.
blind (B) and label information (IN) conditions. As can be seen in In keeping with the above, this was indicative of a small effect of

Fig. 3. Results for the comparison between the blind (B) and label information (IN) conditions for sensory product perceptions of six tasted beer samples (L1, L2, L3,
H1, H2, and H3). Shown are bi-plots spanned by the first two dimensions following Multiple Factor Analysis (MFA): (a) CS Likers (who prefer craft-style beers with
“high” flavor intensity) and (b) TS Likers (who prefer traditional-style beers with “low” flavor intensity). Refer to Table 1 for full sample details. See Table 2 for full
wordings.

10
S.R. Jaeger, et al. Food Quality and Preference 87 (2021) 104043

Fig. 4. Results for the comparison between the blind (B) and label information (IN) condition for holistic product perceptions of six tasted beer samples (L1, L2, L3,
H1, H2, and H3). Shown are bi-plots spanned by the first two dimensions following Multiple Factor Analysis (MFA): (a) CS Likers (who prefer craft-style beers with
“high” flavor intensity) and (b) TS Likers (who prefer traditional-style beers with “low” flavor intensity). Refer to Table 1 for full sample details. See Table 2 for full
wordings.

label information, and the MFA bi-plot for CS Likers, shown in Fig. 4a, 4.2.4. Information effects on cognitive, emotional and situational product
supports this interpretation. As expected, based on Fig. 1c, the holistic/ perceptions of tasted craft-style and traditional-style beers
conceptual attributes in Fig. 4a clearly separated the HF and LF beers, The ANOVA results for effects of beer sample, information condition
with HF and LF beers being characterized as, for example, ‘powerful’ and their interaction are again summarized in Table 5, with full details
and ‘simple,’ respectively. To illustrate the minor effects of label in- in Supplementary Table 4. The first observation relative to the previous
formation, consider ‘full-bodied,’ which when labelled “brewed by a analyses on sensory and holistic/conceptual data is the larger number
small craft brewery in New Zealand” was used more frequently to of significant interaction effects for both preference segments, sug-
characterize HF beers than when these beers were evaluated in the gesting a larger influence of label information relative to the blind
blind condition (46% vs. 36% use). In terms of the interaction effect, condition.
labeling the HF beers as being “brewed by a small craft brewery in New Starting with the CS Likers, the effect of beer sample was highly
Zealand” decreased the frequency of these beers being described as significant on all response variables, in accordance with expectations.
‘cheap’, whereas labeling the LF beers as ”brewed by a large main- The main effect of information condition only reached significance for
stream brewery outside of New Zealand” increased the frequency of “I consider this to be a good quality beer,” and the effect was reflected
characterizations of these beers as ‘cheap.’ in an increase in mean ratings when label information was compared.
In the cluster comprising TS Likers, the pattern of significant ANOVA For all variables, there was a significant BS × IC interaction effect such
effects (Table 5) resembled that found for the sensory data, in the sense that information that the HF beers had been “brewed by a small craft
that a higher number of significant effects of information condition brewery in New Zealand” increased HF beer ratings for the 11 positive
were observed relative to the data from the CS Likers (Supplementary statements, including “I consider this to be a good quality beer,” “this
Table 3 has full details). Further to 94% of effects being significant for beer evokes a positive emotion in me,” and “this beer has an authentic
beer sample (‘insipid’ was the exception), effects of information con- taste,” while decreasing ratings for the statements “in my opinion this
dition were observed for ‘comforting,’ ‘complex,’ ‘full-bodied,’ ‘mascu- tastes like a mainstream beer” and “this beer bores me.” In contrast,
line,’ ‘simple,’ ‘smooth,’ and ‘thirst-quenching,’ and two interactions information that the LF beers were “brewed by a large commercial
were significant (‘comforting’ and ‘industrial’). This, however, did not brewery outside of New Zealand” had the general effect of decreasing
override the main separation of HF and LF beers seen in all previous the ratings for each of the LF beers for the 11 more positive statements,
analyses, which for TS Likers is shown in Fig. 4b. The sample grouping is while increasing them for “in my opinion this tastes like a mainstream
consistent with previous observations, as were the holistic/conceptual beer” and, in the cases of L1 and L2, for “this beer bores me.”
associations, and, for example, linked LF beers with “simple,’ ‘cheap,’ To reduce visual clutter, the MFA results are shown separately for
‘easy to drink’ and ‘thirst quenching,’ but also ‘familiar,’ and ‘com- beer samples and variables. For CS Likers, the B and IN sample con-
forting;’ and the HF beers to “full-bodied,’ ‘ masculine,’ ‘substantial,’ figurations (Fig. 5a) were similar to those previously seen, in the sense
‘powerful,’ and ‘artisan.’ The overall finding regarding the effects of that HF and LF beers were grouped together and spatially separated on
information condition (blind vs. label) among TS Likers was that only the dominant Dimension 1, which in this case accounted for 94% of the
small changes in the holistic/conceptual space for the beers occurred. variation in the data. The effect of information was small, as can be seen
Illustrative of these small changes were ‘masculine’ and ‘simple’ which from sample positions varying mostly on Dimension 2 (6%), although
were used less and more frequently, respectively, for LF beers when for beer H3 a larger effect was seen. In the B condition, H3 was spatially
these were labeled as ”brewed by a large mainstream brewery outside separated from H1 and H2 and located towards the mid-point of Di-
of New Zealand” (8% vs. 3% and 35% vs. 41%, respectively). mension 1. In the IN condition, the 3 HF beers were more tightly
grouped and, looking at the variables plot (Fig. 5c), this change could
largely be attributed to changes in responses on the statement “this is a
good beer for many different social situations.” When information was

11
S.R. Jaeger, et al. Food Quality and Preference 87 (2021) 104043

Fig. 5. Results for the comparison between the blind (B) and label information (IN) condition for the cognitive / emotional /situational responses of six tasted beer
samples (L1, L2, L3, H1, H2, and H3). Shown are plots spanned by the first two dimensions following Multiple Factor Analysis (MFA). Top row: (a) and (b) product
spaces for CS Likers (who prefer craft-style beers with “high” flavor intensity) and TS Likers (who prefer traditional-style beers with “low” flavor intensity), re-
spectively; and bottom row: (c) and (d) variable spaces showing cognitive, emotional, and situational responses for CS Likers and LF Likers, respectively. Refer to
Table 1 for full sample details. See Table 2 for full wordings.

presented that H3 had been brewed by a “small craft brewery in New beer” and “this beer reminds me of those I typically drink.” The inter-
Zealand,” its mean rating increased more than those for H1 and H2. action effects were significant for 10 of 13 response variables (ex-
Another observation pertained to a somewhat closer alignment or cluding “if I wanted to drink more than one glass of the same beer, I’d
grouping of the response variables on both the left and right sides of the chose this one,” “I would drink this beer in the company of like-minded
space in the IN condition, indicating that the label information served beer drinkers” and “this beer has an authentic taste”).
to solidify the perception of the beers into a more rigid and simpler For TS Likers, the information presented about the beers had the
perceptual space. effect of dispersing the MFA space, making Dimension 1 a little less
Among TS Likers, ANOVA (Table 5) revealed significant effects of dominant (71%) and Dimension 2 relatively more important (16%)
beer samples on most response variables, with the exceptions of “I compared to CS Likers. In the sample space (Fig. 5c), this was particu-
consider this to be a good quality beer,” “when I want to savour a beer larly observed for LF beers where L1 and L3 were more clearly sepa-
by myself, this is the kind of beer I would drink” and “I would drink this rated on Dimension 2. For L3 this was linked to an increase in the rating
beer in the company of like-minded beer drinkers.” For the main effect for “this beer bores me” and a decrease in the rating many of the other
of information (which for LF beers was “brewed by a large commercial positive variables including “when I want to savour a beer by myself,
brewery outside of New Zealand”), significance was only reached on this is the kind of beer I would drink,” “I consider this to be a good
two response variables: “I think “beer aficionados” would enjoy this quality beer,” and “I would drink this beer in the company of like-

12
S.R. Jaeger, et al. Food Quality and Preference 87 (2021) 104043

minded beer drinkers” (Fig. 5d). For L1 the effect of label information of liking among the beers within clusters and between these data and
was less pronounced relative to L3, and the changes in mean scores those in Table 3 in Jaeger et al. (2020), namely that beer H2 was liked
smaller. Particularly notable were the differences between the two best and beer H3 was liked least by both segments, provides additional
variable spaces, where among TS Likers (Fig. 5d) dispersion was pro- support for the reliability and robustness of these preference clusters
nounced relative to CS Likers (Fig. 5b) and the changed position of the and the patterns of preferences within them.
vector linked to the variable “I consider this to be a good quality beer” The existence of preference segments within declared craft beer
illustrated how LF beers in the IN condition were less positively eval- drinkers had not been demonstrated prior to our earlier study. The
uated by LF likers, despite their taste preferences for traditional-style demonstration that these preference clusters exist even when beers are
beers. labeled as to their craft-style vs. traditional-style nature and source, as
Overall, the MFA results (Fig. 5) supported the ANOVA results in this study, means that these preferences outweigh information and
(Table 5) with regard to a more notable effect of information condition any corresponding cognitive associations that arise from their knowl-
on the cognitive/emotional/situational responses to beers of both CS edge of the nature and sourcing of the beers. As such, these data provide
Likers and TS Likers. For CS Likers, besides a tendency to consolidate the an important caution for craft brewers that simply producing beers with
positive and negative perceptions, respectively, of HF and LF beers, the complex and strong flavor profiles will not capture all of the existing
biggest change was for H3 which became more tightly grouped with the market of consumers who view themselves as craft beer drinkers.
other two HF beers in the IN condition. TS Likers, despite their taste Rather, a strategy of developing a range of beers with varying levels of
preferences for “low” flavor and traditional-style beers, also evaluated complexity and flavor strengths will better serve the goal of greater
HF beers more positively and LF beers more negatively when these were market share.
presented with the label information (“brewed by a small craft brewery
in New Zealand” and “brewed by a large commercial brewery outside of 5.1.2. Sensory, holistic/conceptual and cognitive/emotional/situational
New Zealand,” respectively). The largest change was observed for L3. responses to tasted beers with label information
Having confirmed that the above two preference clusters exist
4.2.5. Information effects on post-tasting questionnaire responses among the IN consumers in the present study, these two preference
Relatively few effects of information were observed on the re- groups were compared in terms of the sensory, holistic/conceptual, and
sponses to the post-test questionnaire addressing beer-related attitudes cognitive/emotional/situational responses to tasted craft-style and
(Supplementary Table 5 has full details). The only significant effect traditional-style beers.
among the CS Likers was for the statement “I see myself drinking more In terms of the consumers’ sensory characterizations of the labeled
craft beer in the future than I do today,” where agreement (7-pt Likert beers and their holistic/conceptual responses to them, the two pre-
scale) with this statement was significantly higher in the IN condition ference clusters were found to be highly similar (Fig. 1). Thus, re-
than in the B condition). For TS Likers, the only significant effects were gardless of the flavor preferences of the consumers and/or their direct
for “seeking out and drinking craft beer has become a hobby for me” knowledge of the brewing source of the beers, declared craft beer
where agreement was less in the IN condition and for the measured drinkers perceived and conceptualized craft-style and traditional-style
construct of “involvement with craft beer,” where involvement was beers in the same way. This suggests that neither consumer group dis-
again lower in the IN condition. torted their perceptions or conceptualizations of these beers based on
With regard to the effects of information on the post-tasting ques- the labels, as might be predicted by either cognitive dissonance theory
tionnaire responses that addressed beer-related behaviors, no effects (e.g. the perception of a disliked beer is altered to be more consistent
were observed in the CS Likers segment, while only two effects were with one’s negative perception of it) or by assimilation of the in-
seen within the TS Likers segment. For the latter consumers, the like- formation (in accordance with expectation theory) to enhance positive
lihood (yes = 1, no = 0) of “researching different types of beer and perception of beers they expect to like (Festinger, 1962; Olson & Dover,
brewing styles on-line” was less in the IN condition, whereas the like- 1979). Rather, these two preference segments perceive the sensory and
lihood of “having a beer tap in my house” was higher, although the conceptual differences between these styles of beer similarly and, in-
absolute values on the likelihood index were extremely low. stead, simply like one of the flavor types more than the other.
In contrast to the sensory and holistic/conceptual aspects of the HF
5. Discussion and LF beers used in this study, the cognitive/emotional/situational
responses did vary by preference cluster in the IN condition (Fig. 2).
5.1. The label information (IN) condition Although many of the cognitions about the quality of the beers and for
whom they were targeted were the same, both groups perceived the
5.1.1. Liking segments based on responses to the tasted beers presented with beers that they prefer as the ones that evoked positive emotions in them
label information and that they would drink with fellow beer drinkers in various situa-
Given the findings of Jaeger et al. (2020), who used the same beers tions. These data are consistent with the notion that both consumer
as used in this study and consumers drawn from the same overall po- groups accurately recognize and internalize the factual elements asso-
pulation of declared craft-style beer drinkers in New Zealand, it was ciated with craft-style vs. traditional-style, but that their personal
expected that the same two distinct flavor segments found in our past emotional responses to them and their self-awareness of how they
research would likely emerge in the present data, albeit that the beers personally would use or not use the beers in different situations is
in the present study were labeled with information about the beers’ aligned with the beers that they prefer (vs. do not prefer). The basic
brewery source. The hierarchical cluster analysis of the acceptability elements of this finding are consistent with the findings from Jaeger
data obtained in this study confirmed the existence of these same two et al. (2020), where these same beers were presented in a blind con-
clusters in the present data, as are reflected in the liking ratings shown dition (no labels), indicating the robustness of the findings and the
in Table 3. The existence of these two clusters, one that preferred the 3 likely ability to generalize the findings to other craft vs. traditional-style
HF craft-style beers with more complex, bolder and more intense flavor beers. Moreover, the data are consistent with the notion that consumers
(CS Likers segment) and one that preferred the 3 LF traditional-style can accurately perceive the factual and objective differences among
beers with less complex and mild flavor (TS Likers segment) provides food and beverage products in terms of both intrinsic and extrinsic
confirmation that these two distinctive preference groups exist among characteristics, but that they make their choices based on subjective
declared craft beer drinkers and extends this finding to the case where factors related to their affective preferences, their emotional responses,
the consumer is fully aware of the craft vs. mainstream nature of the their values and other unconscious elements. The latter stands in con-
beer and its brewing source. Moreover, the consistency in the patterns trast to rational choice theory (Homans, 1974; Scott, 2000), which, as it

13
S.R. Jaeger, et al. Food Quality and Preference 87 (2021) 104043

relates to food choice, has been described by Köster (2009) as the influencing sensory judgments, as contrasted with affective and other
“fallacy of conscious choice.” non-sensory responses.
The present data also show limited effects of labeling on holistic/
5.1.3. Post-tasting questionnaire responses conceptual responses (Table 5), with only the judgments for ‘cheap’ and
Responses to the post-test questionnaires regarding beer-related ‘powerful’ showing significant interaction effects for CS Likers and only
attitudes and behaviors (Table 4) were consistent with expectations. ‘comforting’ and ‘industrial’ showing similar effects for TS Likers.
Since these attitudes and behaviors were selected to be those that craft However, the effects that were observed were consistent with ex-
beer drinkers would hold and/or exhibit, the fact that craft-style likers pectations based on the labeling, e.g. when labelled “brewed by a small
agreed significantly more with almost all of the attitudes reflects the craft brewery in New Zealand”) (IN condition) the term ‘full-bodied’
fact that this segment has adopted the attitudes that are common was used more frequently to characterize HF beers than in the blind (B)
among most craft beer drinkers. Similarly, with regard to involvement condition, and inversely, decreased the frequency of these beers being
in beer, the greater involvement among these CS Likers is consistent described as ‘cheap.’ Similarly, labeling the LF beers as ”brewed by a
with their preference for beers with novel and stronger flavors, which large mainstream brewery outside of New Zealand” increased the fre-
would likely create desire to seek out more information about different quency of them being characterized as ‘cheap’ and ‘simple.’
beer flavors, which brewers produce such flavors, how they are pro- In contrast to the sensory and holistic/conceptual product percep-
duced, etc.; all of which would require more involvement with craft tions, information condition had a significant effect on a large number
beer issues, materials and information. This association of greater in- of the cognitive/emotional/situational use responses, especially among
volvement with beer and a preference for more complex and stronger CS Likers (Table 5, Fig. 5). Labeling the HF beers as “brewed by a small
flavored beers is consistent with recent findings by Calvo-Porral, craft brewery in New Zealand” increased the degree to which the CS
Rivaroli, and Orosa-Gonzalez (2020) showing that highly involved Likers and TS Likers agreed with positive statements about the beers,
consumers exhibit higher preferences for unconventional or flamboyant e.g. “I consider this to be a good quality beer,” “this beer evokes a po-
beer flavors than low or medium involved consumers. sitive emotion in me,” “this is a unique beer,” “this beer has an authentic
In contrast, consumers in the TS Likers segment appeared not to taste” and “In my opinion, this tastes like a craft beer,” while decreasing
have adopted these attitudes to the same degree or to be as involved in their agreement with the statements “in my opinion this tastes like a
craft beers. mainstream beer” and “this beer bores me,” although the specific effects
Unlike the attitudinal data, only two beer-related behavior state- among LF likers were more variable. Information that the LF beers were
ments reached statistical significance between the two preference seg- “brewed by a large commercial brewery outside of New Zealand” de-
ments in the IN condition (Table 4). The behaviors were exhibited more creased the positive ratings for each of the LF beers among both CS
by the CS Likers, consistent with the preferences of these consumers Likers and TS Likers, while increasing them for “in my opinion this tastes
toward craft-style flavors that would result in a greater likelihood that like a mainstream beer” and “this beer bores me.” Thus, information
they would purchase and refill flagons of beer from a craft brewer and about the craft vs. traditional nature of the beers had the effect of
research different beer styles. Taken together, the post-test attitudinal producing higher positive perceptions of HF beers labeled as “brewed
and behavioral data are consistent with what one would expect to find by a small craft brewery in New Zealand” and producing more negative
among consumers who prefer craft-style flavors vs. traditional-style perceptions of LF beers labeled as being “brewed by a large mainstream
flavors of beer. brewery outside of New Zealand,” although the effects were less con-
sistent among LF Likers.
5.2. The effects of information: B vs. IN conditions The above effects of the information manipulations on specific
cognitive/emotional/situational use variables are consistent with pre-
5.2.1. Effects of information on responses to tasted beers vious research and existing models of the role of information in mod-
Although some studies have shown that information imparted by ulating evaluative product responses. For example, the present label
labeling can impact the discrimination and sensory ratings of beer information results are consistent with other studies of the role of in-
(Allison & Uhl, 1964; Barnett & Spence, 2016), such judgments are formation and labels on beer perception, e.g. Guinard et al. (2000) who
often more difficult to alter than other affective response variables, e.g. found that beers labeled as domestic had lower acceptance ratings and
preference and liking, which are more subjective and labile than sen- beers labeled as imported had higher acceptance ratings than the same
sory judgments (Barnett & Spence, 2016; Guinard et al., 2000, 2001; beer in a blind condition, and that of Silva et al. (2017) who found that
Lee et al., 2006). In fact, when comparing the liking responses to spe- non-alcoholic beer labeled as regular beer was rated more positively,
cific beers in the information (IN) condition vs. the blind (B) condition, while regular beer labeled as non-alcoholic elicited more negative
some differences were observed on the liking ratings obtained. These emotions. Such effects are consistent with the assimilation model of
differences are consistent with the results of Guinard et al. (2000, 2001) expectations (Oliver, 1977; Olshavsky & Miller, 1972; Sherif &
who found that the preferences and liking of consumers for a variety of Hovland, 1961), whereby the consumers’ response is altered to be
domestic, imported and specialty beers differed significantly in blind vs. consistent with the nature of the information, i.e. negative information
brand- and price-labeled conditions. Although the information pre- lowers evaluations, while positive information increases evaluations.
sented here did not involve either brand or price, the general effect of
information on liking ratings is supported by the present data. 5.2.2. Effects of information on post-tasting questionnaire responses
In contrast, the data from a comparison of the label condition (IN) of The effects of information on the beer-related attitudes and beha-
the present study with the data from the blind (B) condition (Jaeger viors obtained with the post-test questionnaire were few. The attitu-
et al., 2020) fails to show an effect on sensory characterization of the dinal data showed that the CS Likers in the information (IN) condition
beers in either preference segment, as reflected in the number of in- reported greater agreement with the statement that they would con-
teraction effects observed between beer samples and information con- sume more craft beer in the future when these beers were labeled as
dition (BS × IC), which were far fewer for the sensory variables than coming from a small, local brewer in New Zealand, suggesting that the
for either the holistic/conceptual variables or the cognitive/emotional/ label reinforced the craft-style nature of these beers and leading the CS
situational variables used in the present research (Table 5). This was Likers to have a greater propensity to purchase this preferred beer style
similarly seen in the MFA bi-plot of the sensory data (Fig. 3), where, in the future. Similarly, the TS Likers were less likely to agree with the
regardless of information condition, the sensory judgments of the beers statement that seeking out and drinking craft beer had become a hobby
were highly similar, both for CS Likers and TS Likers. As such, the for them and scored lower on the scale of involvement with craft beer
present data support previous findings of the difficulty of information when the craft beers were labeled as coming from a small craft brewer

14
S.R. Jaeger, et al. Food Quality and Preference 87 (2021) 104043

in New Zealand, which is also consistent with the label reinforcing at- likers the labeling had a similar effect on their judgments of the tra-
titudes consistent with the preferences of these consumers for tradi- ditional-style beers. These effects are consistent with the predictions of
tional-style beers over craft beers. Although the effects on reported assimilation models of the effect of expectations on product evaluation,
beer-related behaviors were also few in number, the decrease in the which predict that information that produces positive consumer ex-
probability of TS Likers having researched different types of beers and pectations about a product will result in more positive affective eva-
brewing styles in the information condition is also consistent with the luations of that product, while information that evokes negative ex-
expected effect of labeling, while the probabilities of having a beer tap pectations will result in more negative evaluation of the product
in the house were so low as to be unimportant. (Cardello, 1994; Olson & Dover, 1979).
The general consistency of the data obtained in this study and that
6. Conclusions and research contributions reported in Jaeger et al. (2020) provides evidence of the reliability and
robustness of the multi-response approach to product evaluation used
The results of the present study confirm the previous findings of in these studies. This approach employs liking, sensory, holistic/con-
Jaeger et al. (2020) that there exist two different flavor preference ceptual, and cognitive/emotional/situational use measures to construct
segments among declared craft beer drinkers. One segment can be de- a comprehensive profile of products that is amenable to detailed seg-
scribed as craft-style (CS) likers. These consumers prefer the novel and/ mentation analysis of the responses to provide important insights into
or stronger flavor characteristic of craft-style beers. The other segment the factors that drive product preferences.
can be classified as traditional-style (TS) likers. This segment prefers the Lastly, there is nothing in the present data that deters us from our
familiar and less flavorful characteristic of traditional-style beers. A previous conclusion that the differences in the two flavor segments
major contribution of the present research is the extension of this found here and in our previous research are the result of evolving flavor
finding to the situation where the beers are labeled in a manner that preferences among declared craft beer drinkers that occur as a result of
highlights and reinforces their craft- vs. traditional-styles as it relates to greater frequency and/or length of exposure to the novel flavors and
the nature of the brewer (craft vs. mainstream), the size of the brewery varieties of craft beers.
(small vs. large) and its geographic proximity (inside New Zealand vs.
outside New Zealand). As such, it allows the generalization of the
finding of specific preference clusters to situations in which beers are 7. Limitations and suggestions for future research
served with manufacturer labels that identify them as craft- vs. tradi-
tional-style and that provide additional information, such as the loca- One limitation of the present study was in the specific choice of
tion of the brewer and its size. The latter situations most frequently information provided about the beers. The choice made here was based
occur where beers are served/sold in labeled bottles or cans or where on the recognized importance of the size of the brewery, its craft vs.
this information can be inferred from the brand or name of the brewer, traditional nature and the general proximity of the brewery operation
e.g. in supermarkets, package stores and in pubs. to the definition and appreciation of craft beers by declared craft beer
The present data also confirm the fact that, although the two seg- drinkers. Had other information been provided, e.g. about the sensory
ments of declared craft beer drinkers have very different preferences for properties of the beer, the specific brand, etc., the effects may well have
craft-style vs. traditional-style beers, they both perceive the sensory and been different. Another limitation is that the labels were not crossed
holistic/conceptual attributes of craft and traditional beers similarly, between beer styles. This possibility was ruled out for the present re-
suggesting that these consumers have not distorted their perceptual or search, because we felt that labeling craft-style beer as having come
conceptual judgments to conform to their own preferences for the from a mainstream brewer and traditional-style beers as having come
beers. Rather, both groups accurately perceive the differences between from a craft brewer would not have been believable to most declared
these beers, but simply prefer one flavor profile/style over the other. craft beer drinkers. However, such a cross-labeling study could be
The fact that differences were observed between the two preference conducted with a more general population of beer drinkers and with a
clusters in their emotional reaction and in their perceived uses of the set of beers that are not so clearly of a craft-style vs. traditional-style, as
two beer styles, although their objective perceptions of them were si- were the beers chosen for study here. Finally, as it relates to the labeling
milar, is consistent with current theories of choice behavior, e.g. pro- issues, it would be useful to conduct a study in which the beers are
spect theory (Kahneman & Tversky, 1980), that emphasize subjective, presented both blind and in manufacturer-like bottles crossed with beer
non-rational elements in choice behavior and the recognized relation- style, so that the consumer is more likely to accept the “labeling” as
ships between liking, emotional responding and product use. being true and accurate. Although, we have no evidence that the con-
With regard to the role of product information on the evaluation of sumers in this study did not believe the label information, the in-
craft-style (and traditional-style) beers, the present data confirm pre- formation was presented on the ballots and not integrated with the
vious findings showing that sensory judgments are less influenced by sample/packaging. As such, it leaves open the possibility for some
product information than are affective, emotional and other more disassociation of the label from the beer, either due to experimental
subjective experiences. Although information about the small size and inattention (Oppenheimer, Meyvis, & Davidenko, 2009) or label skep-
local nature of craft brewers has been shown to be important elements ticism (Obermiller & Spangenberg, 1998).
that define craft beers (Elzinga et al., 2015), presenting this information One point of further discussion about the terminology for response
(“brewed by a small craft brewery in New Zealand”) in this study did variables used in the study is worthwhile. While we acknowledge that it
not alter the sensory judgments of the craft-style beers among either is possible to argue about specific response variable categorizations as
liker segment. Neither did the corresponding labeling of traditional- reflecting holistic, conceptual or cognitive constructs, insofar as the
style beers (“brewed by a large mainstream brewery outside of New present study was a follow-on to earlier research in our laboratory and
Zealand”) have an effect on the sensory perception of traditional-style data from the labeling condition of this study were designed to be
beers. compared with the blind condition of our previous work, we elected to
As with the failure of information to affect sensory judgments, so use the same terminology and categorization of response variables as
too, information had relatively few effects on the holistic/conceptual was used in that previous research. We acknowledge that arguments
variables measured in this study. Rather the major effects of labeling can be made for alternative categorizations, but the reader is directed to
were restricted to judgments about the emotional responses to the beers the specific response terminology in order to draw conclusions for fu-
and their situational uses with like-minded drinkers. For craft-style beer ture research.
likers, the information about the beers increased their positive emotions
and likely uses for the craft-style beers, while for the traditional-style

15
S.R. Jaeger, et al. Food Quality and Preference 87 (2021) 104043

Funding study focusing on “uniqueness”. Food Quality and Preference, 54, 23–38.
Cardello, A. V., & Sawyer, F. M. (1992). Effects of disconfirmed consumer expectations on
food acceptability. Journal of Sensory Studies, 7(4), 253–277.
Financial support was received from two sources: 1) The New Deliza, R., & MacFie, H. J. (1996). The generation of sensory expectation by external cues
Zealand Ministry for Business, Innovation & Employment (contract and its effect on sensory perception and hedonic ratings: A review. Journal of Sensory
C11X1306), and 2) The New Zealand Institute for Plant and Food Studies, 11(2), 103–128.
Donadini, G., & Porretta, S. (2017). Uncovering patterns of consumers' interest for beer: A
Research Limited (PFR). case study with craft beers. Food Research International, 91, 183–198.
Elzinga, K. G., Tremblay, C. H., & Tremblay, V. J. (2015). Craft beer in the United States:
CRediT authorship contribution statement History, numbers, and geography. Journal of Wine Economics, 10(3), 242–274.
Festinger, L. (1962). A theory of cognitive dissonance, Vol. 2. Stanford University Press.
Giacalone, D., Frøst, M. B., Bredie, W. L., Pineau, B., Hunter, D. C., Paisley, A. G., ...
Sara R. Jaeger: Conceptualization, Methodology, Writing - original Jaeger, S. R. (2015). Situational appropriateness of beer is influenced by product
draft, Writing - review & editing. Thierry Worch: Formal analysis. familiarity. Food Quality and Preference, 39, 16–27.
Gómez-Corona, C., Escalona-Buendía, H. B., García, M., Chollet, S., & Valentin, D. (2016).
Tracey Phelps: Resources, Investigation. David Jin: Resources,
Craft vs. industrial: Habits, attitudes and motivations towards beer consumption in
Investigation, Visualization. Armand V. Cardello: Conceptualization, Mexico. Appetite, 96, 358–367.
Methodology, Writing - original draft, Writing - review & editing. Gómez-Corona, C., Escalona-Buendía, H. B., Chollet, S., & Valentin, D. (2017). The
building blocks of drinking experience across men and women: A case study with
craft and industrial beers. Appetite, 116, 345–356.
Declaration of Competing Interest Gómez-Corona, C., Chollet, S., Escalona-Buendía, H. B., & Valentin, D. (2017). Measuring
the drinking experience of beer in real context situations. The impact of affects,
senses, and cognition. Food Quality and Preference, 60, 113–122.
The authors declare that they have no known competing financial Guinard, J. X., Uotani, B., Mazzucchelli, R., Taguchi, A., Masuoka, S., & Fujino, S. (2000).
interests or personal relationships that could have appeared to influ- Consumer testing of commercial lager beers in blind versus informed conditions:
ence the work reported in this paper. Relation with descriptive analysis and expert quality ratings. Journal of the Institute of
Brewing, 106(1), 11–20.
Guinard, J. X., Uotani, B., & Schlich, P. (2001). Internal and external mapping of pre-
Acknowledgement ferences for commercial lager beers: Comparison of hedonic ratings by consumers
blind versus with knowledge of brand and price. Food Quality and Preference, 12(4),
243–255.
Sensory and Consumer Science staff from PFR are thanked for help
Hollebeek, L. D., Jaeger, S. R., Brodie, R. J., & Balemi, A. (2007). The influence of in-
in collection of consumer data. volvement on purchase intention for new world wine. Food Quality and Preference,
18(8), 1033–1049.
Homans, G. C. (1974). Social behavior: Its elementary forms. New York: Harcourt, Brace and
Appendix A. Supplementary data
World.
Jaeger, S. R., Cardello, A. V., Chheang, S. L., Beresford, M. K., Hedderley, D. I., & Pineau,
Supplementary information to this article can be found online at B. (2017). Holistic and consumer-centric assessment of beer: A multi-measurement
https://doi.org/10.1016/j.foodqual.2020.104043. It contains the fol- approach. Food Research International, 99, 287–297.
Jaeger, S. R., Hunter, D. C., Vidal, L., Chheang, S. L., Ares, G., & Harker, F. R. (2019).
lowing: 1) Summary of participant characteristics; 2) B vs. IN com- Sensory product characterization by consumers using check-all-that-apply questions:
parison for sensory responses; 3) B vs. IN comparison for holistic / Investigations linked to term development using kiwifruit as a case study. Journal of
conceptual responses; 4) B vs. IN comparison for cognitive / emotional Sensory Studies, 34(3), Article e12490.
Jaeger, S. R., & MacFie, H. J. (2001). The effect of advertising format and means-end
/ situational responses; and 5) B vs. IN comparison for post-tasting information on consumer expectations for apples. Food Quality and Preference, 12(3),
questionnaire. 189–205.
Jaeger, S. R., Xia, Y., Le Blond, M., Beresford, M. K., Hedderley, D. I., & Cardello, A. V.
(2019). Supplementing hedonic and sensory consumer research on beer with cogni-
References tive and emotional measures, and additional insights via consumer segmentation.
Food Quality and Preference, 73, 117–134.
Acitelli, T., & Magee, T. (2017). Audacity of hops: The history of America's craft beer re- Jaeger, S. R., Worch, T., Phelps, T., Jin, D., & Cardello, A. V. (2020). Preference segments
volution. Chicago Review Press. among declared craft beer drinkers: Perceptual, attitudinal and behavioral responses
Allison, R. I., & Uhl, K. P. (1964). Influence of beer brand identification on taste per- underlying craft-style vs. traditional-style flavor preferences. Food Quality and
ception. Journal of Marketing Research, 1(3), 36–39. Preference, 103884.
Ares, G., & Jaeger, S. R. (2015). Check-all-that-apply (CATA) questions with consumers in Kahneman, D., & Tversky, A. (1980). Prospect theory. Econometrica, 12.
practice: Experimental considerations and impact on outcome. In J. Delarue, J. B. Köster, E. P. (2009). Diversity in the determinants of food choice: A psychological per-
Lawlor, & M. Rogeaux (Eds.). Rapid Sensory Profiling Techniques (pp. 227–245). spective. Food Quality and Preference, 20(2), 70–82.
Cambridge, UK: Woodhead Publishing. Lee, L., Frederick, S., & Ariely, D. (2006). Try it, you'll like it: The influence of expecta-
Ares, G., Reis, F., Oliveira, D., Antúnez, L., Vidal, L., Giménez, A., Chheang, S. L., Hunter, tion, consumption, and revelation on preferences for beer. Psychological science,
D. C., Kam, K., Roigard, C. M., Paisley, A. G., Beresford, M. K., Jin, D., & Jaeger, S. R. 17(12), 1054–1058.
(2015). Recommendations for use of balanced presentation order of terms in CATA Manoukian, E. B. (1986). Mathematical nonparametric statistics. New York, NY: Gordon &
questions for sensory product characterization. Food Quality and Preference, 46, Breach.
137–141. Medoro, C., Cianciabella, M., Camilli, F., Magli, M., Gatti, E., & Predieri, S. (2016).
Aquilani, B., Laureti, T., Poponi, S., & Secondi, L. (2015). Beer choice and consumption Sensory profile of italian craft beers, beer taster expert versus sensory methods: A
determinants when craft beers are tasted: An exploratory study of consumer pre- comparative study. Food and Nutrition Sciences, 7(06), 454.
ferences. Food Quality and Preference, 41, 214–224. Mielby, L. H., & Frøst, M. B. (2010). Expectations and surprise in a molecular gastronomic
Barnett, A., & Spence, C. (2016). Assessing the effect of changing a bottled beer label on meal. Food Quality and Preference, 21(2), 213–224.
taste ratings. Nutrition and Food Technology: Open Access, 2(4). Nunnaly, J. (1978). Psychometric theory. New York: McGraw-Hill.
Barnett, A., Velasco, C., & Spence, C. (2016). Bottled vs. canned beer: Do they really taste Obermiller, C., & Spangenberg, E. R. (1998). Development of a scale to measure consumer
different? Beverages, 2(4), 25. skepticism toward advertising. Journal of Consumer Psychology, 7(2), 159–186.
Cadoret, M., & Husson, F. (2013). Construction and evaluation of confidence ellipses Oliver, R. L. (1977). Effect of expectation and disconfirmation on postexposure product
applied at sensory data. Food Quality and Preference, 28(1), 106–115. evaluations: An alternative interpretation. Journal of Applied Psychology, 62(4), 480.
Calvo-Porral, C., Rivaroli, S., & Orosa-Gonzalez, J. (2020). How consumer involvement Olshavsky, R. W., & Miller, J. A. (1972). Consumer expectations, product performance,
influences beer flavour preferences. International Journal of Wine Business Research. and perceived product quality. Journal of Marketing Research, 9(1), 19–21.
https://doi.org/10.1108/IJWBR-10-2019-0054. Olson, J. C., & Dover, P. A. (1979). Disconfirmation of consumer expectations through
Caporale, G., & Monteleone, E. (2004). Influence of information about manufacturing product trial. Journal of Applied Psychology, 64(2), 179.
process on beer acceptability. Food Quality and Preference, 15(3), 271–278. Oppenheimer, D. M., Meyvis, T., & Davidenko, N. (2009). Instructional manipulation
Cardello, A. V. (1994). Consumer expectations and their role in food acceptance. checks: Detecting satisficing to increase statistical power. Journal of Experimental
Measurement of Food Preferences (pp. 253–297). Boston, MA: Springer. Social Psychology, 45(4), 867–872.
Cardello, A. V. (2003). Consumer concerns and expectations about novel food processing Peryam, D. R., & Pilgrim, F. J. (1957). Hedonic scale method of measuring food pre-
technologies: Effects on product liking. Appetite, 40(3), 217–233. ferences. Food Technology, 11(9), 9–14.
Cardello, A. V. (2007). Measuring consumer expectations to improve food product de- Piqueras-Fiszman, B., & Spence, C. (2015). Sensory expectations based on product-ex-
velopment. Consumer-led Food Product Development (pp. 223–261). Woodhead trinsic food cues: An interdisciplinary review of the empirical evidence and theore-
Publishing. tical accounts. Food Quality and Preference, 40, 165–179.
Cardello, A. V., Pineau, B., Paisley, A. G., Roigard, C. M., Chheang, S. L., Guo, L. F., ... R Core Team (2019). R: A language and environment for statistical computing. R
Jaeger, S. R. (2016). Cognitive and emotional differentiators for beer: An exploratory Foundation for Statistical Computing, Vienna, Austria. http://www.r-project.org/

16
S.R. Jaeger, et al. Food Quality and Preference 87 (2021) 104043

index.html. liking and emotions when consuming beer or non-alcoholic beer in a bar. Food Quality
Schifferstein, H. N. J., Kole, A. P. W., & Mojet, J. (1999). Asymmetry in the dis- and Preference, 55, 58–66.
confirmation of expectations for natural yogurt. Appetite, 32(3), 307–329. Stefani, G., Romano, D., & Cavicchi, A. (2006). Consumer expectations, liking and will-
Scott, J. (2000). Rational choice theory. In G. Browning, A. Halcli, & F. Webster (Eds.). ingness to pay for specialty foods: Do sensory characteristics tell the whole story?
Understanding Contemporary Society: Theories of the Present (pp. 126–138). London: Food Quality and Preference, 17(1–2), 53–62.
Sage Publications. Thurnell-Read, T. (2016). ‘Real ale enthusiasts, serious leisure and the costs of getting ‘too
Sester, C., Dacremont, C., Deroy, O., & Valentin, D. (2013). Investigating consumers’ serious’ about beer. Leisure Sciences, 38(1), 68–84.
representations of beers through a free association task: A comparison between Tuorila, H. M., Meiselman, H. L., Cardello, A. V., & Lesher, L. L. (1998). Effect of ex-
packaging and blind conditions. Food Quality and Preference, 28(2), 475–483. pectations and the definition of product category on the acceptance of unfamiliar
Sherif, M., & Hovland, C. I. (1961). Social judgment: Assimilation and contrast effects in foods. Food Quality and Preference, 9(6), 421–430.
communication and attitude change. New Haven, CT: Yale University Press. Worch, T., Sinesio, F., Moneta, E., Abbà, S., Dreyfuss, L., McEwan, J. A., & Porcherot-
Silva, A. P., Jager, G., Voss, H. P., van Zyl, H., Hogg, T., Pintado, M., & de Graaf, C. Lassallette, C. (2020). Influence of different test conditions on the emotional re-
(2017). What’s in a name? The effect of congruent and incongruent product names on sponses elicited by beers. Food Quality and Preference, 83, Article 103895.

17

You might also like