adults) using the nine-point hedonic scale. The con-
sumers’ sugar and fat intakes were estimated using a food
frequency questionnaire. Increased fat caused higher
creamy, doughy, eggy and mouthcoating intensities and
lower coolness, ice crystals and melting rate intensities (p
< 0.05). Increased sugar resulted in higher vanilla,
almond, buttery, sweet, creamy, doughy and mouth-
coating intensities and lower coolness and ice crystals
intensities (p < 0.005). DOL of the samples did not differ
between males and females except for flavor, which males
liked more than females (p < 0.01). Response surface
methodology indicated that 14.3% fat and 14.8% sugar
resulted in the highest sensory acceptability. Optimum
flavor was obtained with 14.1% fat and 15.4% sugar, and
optimum texture/mouthfeel was obtained with 13.2% fat
and 14.0% sugar. Only limited relationships were found
between dietary intake and hedonic measures.
27.THElNFLUENCEOF FAT,SUGARAND
NON-FAT MILKSOLIDSON SELECTED
TASTE,FLAVOR ANDTEXTURE
PARAMETERSOFAVANILLAICECREAM
Chantal R. Stampanoni Koeferli, Patrizia Piccinali and
Susanne Sigrist,
Sensory Science and Evaluation Department of
Givaudan-Roure Flavors Ltd, 8600 Dtibendorf,
Switzerland
Vanilla flavored ice creams varying in fat, sugar and
non-fat milk solids were prepared to study the influence of
the three design variables on sweetness, flavor notes, such
as vanillin, phenolic, caramel, buttery, creamy, milky and
whey-like, as well as textural attributes such as manual
firmness, coldness, ice crystal and melting rate. Analysis of
variance, principal component analysis and partial least
squares regression indicated that fat addition increased
primarily the buttery and creamy notes as well as mouth-
coating. Coldness, ice crystal perception and the melting
rate were decreased by fat. Sugar addition mainly
increased sweetness, caramel and vanillin notes, whereas
it decreased milkiness. Non-fat milk solids decreased the
coldness, ice crystals and melting rate perception, whereas
they increased creaminess and mouth-coating.
28.APAN EUROPEAN SENSORY
PROFILING AND PREFERENCESTUDYOF
COFFEE
Einar Risvik,
European Sensory Network, TN0 Nutrition and Food
Research Institute, P.O. Box 360, 3700 AJ Zeist, The
Netherlands
The study was undertaken by the European Sensory
Network (ESN) to compare trained sensory panels across
Europe, to determine a possible European vocabulary
for describing coffees, to investigate preference of Eur-
opean consumers, and to identify the sensory attributes
driving these preferences. The profiling study was
undertaken by 11 sensory panels in eight European
countries and the consumer test by eight panels in seven
European countries. Sixteen samples of coffee were pro-
filed by 104 trained subjects. Eight of these coffees were
selected for preference evaluation by 605 consumers. For
all panels, the coffee sample space, obtained by a prin-
cipal component analysis was oriented by flavors related
to the degree of roasting and bitterness for one dimen-
sion and acidity for the other one. In order to validate
these sample structures, more sophisticated statistical
methods, based on RV coefficient and analytical per-
mutation tests of samples, were used. These techniques
allow the panel homogeneities to be quantified and an
estimation of the number of dimensions required by the
sample structures to be made. Moreover, a permutation
test of assessors enabled it to be determined whether the
assessors are exchangeable between the panels. In terms
of developing a European vocabulary, analyses of var-
iance including panel, assessor within panel, sample and
sample by panel interaction effects were undertaken.
From the whole set of the 259 different attributes used
by the 11 panels, a list of 13 key-attributes for describing
coffee was defined. As a prerequisite to determining
which sensory attributes drove preference, a clustering
procedure of the 605 European consumers found eight
groups of consumers, each group having a different pat-
tern of preference. A sensory interpretation of these pat-
terns was proposed from correlation and preference
mapping analyses. Moreover, no clear relationships were
found between these clusters and country, sex and age of
the consumers.
29.CROSS-NATIONALSEGMENTATION OF
CONCEPTS FOR COFFEE
Howard R. Moskowitz, Dorit Cohen, Alex Gofman and
Richard Guggisberg,
Moskowitz Jacobs Inc., White Plains, NY 10604, USA
Sensory analysts have long been interested in indivi-
dual differences in products. With the advent of multi-
national research efforts, this interest has expanded into
cross-national differences for products on the one hand,
and reactions to product concepts on the other. This
study reports on a world wide study with 900 consumers,
each of whom evaluated different experimentally
designed concepts for coffee. The study enables the
researcher to identify the “utility” or part-worth con-
tribution of each element to consumer purchase intent.
Individual models have been created for the 900 con-
sumers to identify the specific hot buttons for product
communication. The results show that although
