%7XYH]SJXLI)JJIGXSJ4IVGIMZIH&IIV,MWXSV]

SR6ITSVXIH4VIJIVIRGIWF]7IRWSV]4ERIPW
[MXL(MJJIVIRX0IZIPWSJ8VEMRMRK
J.E. Smythe1 and C.W. Bamforth1,2

%&786%'8 report a short study conducted using individuals from dif-

ferent countries aimed at determining the effect that per-
J. Inst. Brew. 108(1):34–36, 2002
ceived beer history has on reported preference.
A study in three countries tested how consumers respond in
paired-comparison tests to two beers that differed only in their
reported history. In one test, consumers tasted two beers that had 1%8)6-%07%2(1)8,3(7
supposedly been produced with different brewing techniques. In
the second, they compared two beers with apparently different 7YFNIGXW
sugar sources. In reality, for both tests all samples came from the Three different sets of judges participated. A set from
same bottle. Results show clear indications that naïve judges will Ireland (nine male, three female, aged 29–53) represented
be likely to report a preference, while more highly trained judges an untrained sensory panel. A set from Finland (six male,
will be more prone to express no preference, though all popula-
tions tested appear to be influenced by the concept of beer history.
six female, aged 24–60) represented a trained sensory
panel. A final set from Belgium (11 male, three female,
Key words: Perception, preference, production technique. aged 25–54) represented a rigorously trained and highly
tested sensory panel. Regional breweries with experience
-2863(9'8-32 in sensory work provided the sets of judges from their
own employees. Though judges appeared to be offering
Attitudes and beliefs about a product influence con- preference ratings, the goals of this study were to regard
sumer preference and the closely related concept of pur- the judges as analytical instruments in a deliberately mis-
chase intent2. For example, among Swedish consumers of informed setting. It is therefore important to consider the
tomatoes, information supplied to consumers indicating extent of sensory or brewing experience of the judges in
that a particular sample was grown “ecologically” increased this context.
the preference for three out of four different tomato samples,
whether or not the information had any truth to it3. This 7XMQYPM
groundbreaking study strongly implicated the perceived Judges tasted two sets of pairs of beer samples (~30 mL)
history of a product for the first time in an experimental served at room temperature in plastic cups. The identity of
setting, although the fact that prior information concerning the beer varied by location, within which each panellist
a sample impacts how a panel judges is well recognised4. was given all samples from a single bottle. Judges were
Little published data exists on the effect that the per- unaware of this and were led to believe that the beers dif-
ceived production history of a beer has on the consumer’s fered in their production aspects. The beers were labelled
perception of that product. In one such study, we showed so as to indicate their supposed provenance.
that the same beer labelled as either “lager” or “alcohol-
free lager” led to different responses for perceived authen- )\TIVMQIRXEPHIWMKRERHTVSXSGSP
ticity, flavour intensity and fullness1. A number of brewing In one test, subjects were told that one of the beers was
companies position their brands on a provenance of tradi- made using traditional brewing techniques and took about
tional values, for example purest raw materials and pains- 15 days to produce a brew, while the other beer was made
taking processes occupying lengthy time periods. Con- using exactly the same ingredients, but with a revolutionary
versely, some companies are striving to apply the latest new process using modified yeasts and temperature con-
advances in brewing science in order to shave costs trols that allowed the beer to be produced within 10 h. In
through accelerated operating procedures. In this paper we the other test set, judges were told that one beer was produced
according to the Reinheitsgebot using only water, hops,
barley malt and yeast, while the other beer was made from
1 Department
exactly the same ingredients, but with 30% replacement of
of Food Science and Technology, University of Cali-
the malt by corn sugar adjunct. Judges and groups had
fornia, Davis, CA 95616 USA.
2 Corresponding author. E-mail: cwbamforth@ucdavis.edu samples and tests counterbalanced and randomised among
them. Judges used all of their senses to answer the ques-
4YFPMGEXMSRRS+6
tion of which of the two beers they preferred, if either.
8LI-RWXMXYXI +YMPHSJ&VI[MRK
Special emphasis given individually to each judge clari-

   .3962%03*8,)-278-898)3*&6); -2+

fied that it was acceptable to not have a preference be- pared to the less well trained panels. This may be related
tween samples as long as they reported this on their re- directly to the degree of panellist training (i.e., a highly
sponse sheets. Judges recorded responses on an answer trained panellist would be less likely to have a preference
sheet. Testing times ranged from 2 to 9 min per judge. in these tests). Analysis 3 tests whether traditional or modi-
fied brewing was preferred among those who expressed a
7XEXMWXMGEPQIXLSH preference. This revealed that only the highly trained
This study uses Chi-squared to test the frequency of panel had a significant preference for traditional brewing
occurrence as related to chance levels, a method com- over modified brewing in the case of the brew time study.
monly employed for interpreting responses to questions However, the general trend suggests that in almost all
about foods with multiple response categories5. populations tested there may be a slight preference for the
concept of traditional brew times, over accelerated proc-
6)79087 essing, when a preference is expressed.

Table I summarises the results of this study. There are '32'097-327

at least three ways of interpreting this data using Chi-
squared statistics. The first (Analysis 1) is to assume an This study indicates that the suggestion of a beer’s his-
equal chance of a judge selecting any one of the three pos- tory may well play a psychological role in the perception
sible preference responses for each test. A second way of beer. This mirrors the findings of Johansson et al.3 with
(Analysis 2) is to compare responses in which the judge the perceived ecology of cultivation of tomatoes. Merely
indicated a preference against those where the judge did suggesting a difference between beer histories causes
not indicate a preference. A final means (Analysis 3) is to some judges to perceive beers as different. However, as
compare among those who expressed a preference the panel members are more highly trained, they appear to be
number that preferred traditional methods versus those less likely to declare a preference, effectively superseding
who preferred non-traditional methods of brewing. Table the misinformation they have been given. Among panel-
II summarises these results, with NS signifying a signifi- lists who do express a preference, there is a tendency for
cance level of p>0.1. an inclination toward traditional methods of brewing over
accelerated forms. However, we clearly used only small
(-7'977-32 sample sizes from the different regions. Testing a larger
sample size would more accurately represent a target demo-
Analysis 1 reveals that there is no significant difference graphic population and can guide future consumer studies.
in the judges’ choices for the responses of no preference, Marketing studies might also make use of these approaches
modified, or traditional brewing processes in either study. to determine if there is a predisposition among consumers
However, since judges unknowingly tasted the same beer for certain product characteristics.
and evaluated them as if they were different, it is a rea-
sonable to perform a Chi-squared test to determine if they
%'/23 ;0)(+)1)28 7
had a preference at all, as is done in Analysis 2. This
analysis reveals that more highly trained panels have an The authors thank the breweries involved in this study and
increased likelihood of reporting no preferences as com- their judges for their participation, as well as Aaron Magram and

TABLE I. Number of judges and preferences in each study.

Brew Time Study Adjunct Study
Degree of training Rigorously Trained Trained Untrained Rigorously Trained Trained Untrained
Country Belgium Finland Ireland Belgium Finland Ireland
# who prefer traditional 7 5 5 6 7 6
# who prefer non-traditional 1 4 5 4 3 5
# with no preference 6 3 2 4 2 1

TABLE II. Chi-squared values, degrees of freedom and significance of results for the study.
Brew Time Study Adjunct Study
Degree of training Rigorously Trained Trained Untrained Rigorously Trained Trained Untrained
Country Belgium Finland Ireland Belgium Finland Ireland
Analysis 1
Chi-squared 4.4 0.5 1.5 0.6 3.5 3.5
Deg. free. 2 2 2 2 2 2
Significance NS NS NS NS NS NS
Analysis 2
Chi-squared 0.3 3.0 5.3 2.6 5.3 8.3
Deg. free. 1 1 1 1 1 1
Significance NS p<0.1 p<0.05 NS p<0.05 p<0.01
Analysis 3
Chi-squared 4.5 0.1 0 0.4 1.6 0.1
Deg. free. 1 1 1 1 1 1
Significance p<0.05 NS NS NS NS NS

:3023   
Lucas Milligan for their help in data acquisition. JES received 3. Johansson, L., Haglund, A., Berglund, L., Lea, P. and Risvik, E.,
funds from the Jastro-Shields Scholarship in support of travels Preference for tomatoes, affected by sensory attributes and in-
during this work. formation about growth conditions. Food Quality and Prefer-
ence, 1999, 10, 289-298.
4. Lawless, H.T. and Heymann, H., Sensory Evaluation of Food,
6)*)6)2')7 Principles and Practices. Aspen Publishers, USA, 1998.
5. O’Mahony, M., Sensory Evaluation of Food: Statistical Methods
1. Bamforth, C.W., Butcher, K.N. and Cope, R., The interrelation- and Procedures. Marcel Dekker, Inc., New York, 1986.
ships between parameters of beer quality. Ferment, 1989, 2, 54-58.
2. Fishbein, M. and Ajzen, I., Belief, attitude, intention, and behav-
ior: An introduction to theory and research. Addison-Wesley,
Reading, MA, 1975. (Manuscript accepted for publication February, 2002)

   .3962%03*8,)-278-898)3*&6); -2+