Food Quality and Preference 17 (2006) 116125

www.elsevier.com/locate/foodqual

Consumer expectations for sensory properties in virgin olive oils

Gabriella Caporale

a,*

, Sonia Policastro b, Angela Carlucci a, Erminio Monteleone

Smell and Taste Association, Via dellEdilizia, 85100 Potenza, Italy

Dipartimento di Biologia Difesa e Biotecnologie Agro Forestali, Universita` degli Studi della Basilicata, Via dellAteneo Lucano 10,
Campus Macchia Romana, 85100 Potenza, Italy
Received 1 November 2004; received in revised form 6 July 2005; accepted 8 July 2005
Available online 1 September 2005

Abstract
Information has been shown to create expectations concerning sensory properties and acceptability of food products, and to
inuence their evaluations. Studying the impact of information is particularly relevant for traditional products which communicate
about typicality. Extra virgin olive oil is a typical Mediterranean production whose typicality is strongly aected by the origin of its
raw material and the manufacturing technology. The present study aims (1) to explore the appropriateness of several sensory
descriptors in evaluating the typicality of certain extra virgin olive oils, (2) to assess the impact of information about the origin
of the product on the sensory prole perception, (3) to study how the eect of sensory expectations can inuence liking and typicality responses for the experimental oils obtained from a dened cultivar. Working with a panel of consumers familiar with several typical extra virgin olive oils produced in Lucania, a set of monovarietal extra virgin olive oils were evaluated. Results show that
there are well dened expectations for some of the sensory properties which characterize the typical olive oils presented. The sensory
disconrmations leading to complete assimilation in sensory perception are associated to higher typicality ratings. Our results also
revealed that bitterness and pungency proved to be the most appropriate sensory descriptors of certain typical olive oils.
2005 Elsevier Ltd. All rights reserved.
Keywords: Olive oil; Appropriateness test; Typicality; Expectations; Origin

1. Introduction
The beliefs and attitudes of individuals with regards
to food products are strongly dependent on their cultural traditions and on their education and culinary habits and they may vary with information (Issanchou,
1996). Recent research on the eects of dierent types
of information about nutritional qualities, or food
ingredients and use, or the origin of raw material, or
the manufacturing process, etc. suggest the possibility that these messages may also generate hedonic
expectations and inuence food acceptability. Most of
the empirical studies conrm that the inuence of
*

Corresponding author.
E-mail address: caporale@unibas.it (G. Caporale).

0950-3293/$ - see front matter 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.foodqual.2005.07.011

information on liking after tasting could be explained

by an assimilation model as a possible consequence of
the occurrence of a disconrmation phenomenon
(Caporale & Monteleone, 2001; Caporale & Monteleone, 2004; Johansson, Haglund, Berglund, Lea, & Risvik, 1999; Kahkonen & Tuorila, 1996; Lange,
Rousseau, & Issanchou, 1999; Siret & Issanchou, 2000).
Disconrmation, i.e. the dierence between the expected acceptability of a product and the blind acceptability of the same product before being aected by
expectations, aects food perception and acceptance.
When the stimuli/product is better than expected, the
disconrmation is positive. Conversely when the stimuli/product is worse than expected the disconrmation
is negative. A number of predictive models (e.g. assimilation and contrast) have been suggested in order to

G. Caporale et al. / Food Quality and Preference 17 (2006) 116125

explain the eect of discrepancies between expectations

and actual product performance on acceptance (Anderson, 1973; Cardello & Sawyer, 1992).
Assimilation model occurs when liking of the product
moves in the expectation direction. Thus, when the differences between liking without (blind) and with information (actual) prove to be signicantly dierent from
zero it means that there was a signicant eect of written
information on hedonic scores and an assimilation occurs if liking after exposure to information is inuenced
in the direction of expected liking. Furthermore, in the
case of assimilation, when expected minus actual liking
results signicantly dierent from zero it means that
the score in the presence of information was located between the blind score and the expected score, i.e. consumers do not completely assimilate towards their
expectations and assimilation is not total (Siret &
Issanchou, 2000). Several authors suggested that if consumers do not completely assimilate the discrepancy between their expectations and perceived quality it can be
assumed that they will revise their expectations after repeated exposures (Deliza & MacFie, 1996; Goering,
1985; Lange et al., 1999).
Information about country of origin has a tremendous inuence on the acceptance and success of products (Dichter, 1962). Other authors showed that
country of origin has a strong inuence on product evaluation (Peterson & Jolibert, 1995). In particular countryof-origin information may be studied as a cognitive cue,
an informational stimulus about or relating to a product
that is used by the consumers to infer beliefs regarding
product attributes such as quality (Bilkey & Nes, 1982;
Steenkamp, 1990). In addition to its role as a quality
cue, country of origin has symbolic and emotional meaning for consumers. So there are cognitive, aective and
normative mechanisms which govern country-of origin
eects (Verlegh & Steenkamp, 1999). Some authors report that the region-of-origin cue has a direct eect,
not only indirect via perceived quality, on regional product preference for some consumer segments, especially
those resident in the products region of origin. Products
place of origin may elicit emotions based on consumers
experience with the region able to directly aect regional
product preference. When consumers are aware of the region, the region-of-origin cue triggers consumers
associations with the region, which are then used to evaluate the product. The stronger and more favourable consumers associations with the region are, the greater the
level of success of regional products (Van der Lans,
van Ittersum, De Cicco, & Loseby, 2001).
Typical foods are characterized by a complex quality
prole dened by several conditions such as geographic
origin, culture, tradition, environmental factors, genetic
properties and the manufacturing process know-how.
These factors may aect the sensory prole of products
determining their specicity. The typicality of a product

117

depends therefore upon the detection of certain quality

allowing it to be distinguished from other products of
the same commercial category, and for this reason typical products are identied with well dened sensory
expectations by consumers familiar with them. It may
be supposed that the quality of typical food products is related to the conformance of these products to a characteristic sensory prole, and thus we can assume that when
these products are experienced and respond in terms of
expected and actual sensory proles, they are accepted
by consumers and present a high quality level.
Amongst typical food products extra virgin olive oil
plays an important role as a typical example of Mediterranean production whose typicality is strongly aected
by the origin of its raw material and the manufacturing
technology. In recent years the European Union (EU)
has recognised many protected extra virgin olive oils
(protected designation of origin or protected geographical indication). These typical oils, which are mainly
famous for their sensory properties, present a complex
and specic qualitative prole including both intrinsic
and extrinsic factors.
For many years market researchers have addressed
the question of how failure to satisfy expectation aects
consumer satisfaction of products (Cardello, 1994).
However, it is only recently that sensory scientists have
begun to address the question of how expectation eects
may inuence consumer perceptions of sensory attributes (Deliza & MacFie, 1996; Deliza, MacFie, & Hedderley, 2003; Kahkonen & Tuorila, 1998; Lange et al.,
1999; Prescott & Young, 2002). Kahkonen and Tuorila
(1998) found that when Bologna sausages were labelled as reduced fat, they were rated as lower in saltiness and fattiness than a regular sausage, dierences
that were not apparent when no information was given.
Most of the cited studies on the impact of information on expected and actual sensory properties deal with
products often unfamiliar to consumers or at least anonymous or commercial (Deliza et al., 2003; Kahkonen &
Tuorila, 1998; Tuorila, Cardello, & Lesher, 1994). It
would be very useful when marketing typical products
to investigate how information about the products origin can aect the perception of sensory descriptors and
how the eect of sensory expectations can inuence liking and the response of typicality according to consumers familiar with the product.
The present study was designed to examine the impact of information about the origin of the product
(production area and cultivar) on the expected and perceived hedonic and sensory properties of extra virgin olive oils.
The work explored a possible approach for dealing
with studies related to the measurement of the sensory
expectations of consumers towards typical products
using two separate phases: (1) the selection of the
most appropriate attributes for describing the sensory

118

G. Caporale et al. / Food Quality and Preference 17 (2006) 116125

properties of certain typical oils; (2) the analysis of

expectations, generated by information on product origin, towards the most appropriate sensory attributes.
With this approach it was necessary to select consumers who were familiar with the typical oils being tested
thus justifying the use of a specically chosen panel.

2. Materials and methods

2.1. Products
Six samples representing three typical extra virgin olive oils produced in Lucania were chosen for the study.
These samples were made exclusively from specic olive
cultivars growing in specic areas of Lucania. Table 1
presents the name and the origin of the six experimental
oils. The oils were produced in November 2003 and
tested from January to the end of February 2004 in
order to keep their sensory properties stable. Before and
during the experiment the oils were kept out of the light
and were stored at a controlled temperature (12 C) and
humidity (65%). They were presented to the consumers
into 60-ml dark glasses and served at 25 C. The presentation order of the samples followed Williams Latin
Squares balanced for rst-order and carry-over eects
(MacFie, Bratchell, Greenho, & Vallis, 1989).
2.2. Consumers
Consumers were recruited at specialized virgin olive
oil shops in Potenza (main town in the region of Lucania,
Southern Italy). They were selected using predetermined
screening criteria based on purchasing and consumption
frequency as well as familiarity with the typical extra
virgin olive oil of Lucania used in the study.
One hundred and eight subjects were interviewed and
were asked to spontaneously give as many names of
Lucania oils as they could. For each quoted typical
oil, consumers were asked to report how frequently they
purchased it (1 = never; 2 = less than once a year;
3 = once or twice a year; 4 = 35 times a year; 5 = less
than once a month; 6 = once a month; 7 = 23 times a
month). Then, they were asked to indicate how frequently they consumed it either at home or in a restau-

Table 1
Features of the six experimental oils
Product

Cultivar

Production area

A1
A2
B1
B2
C1
C2

Coratina

Vulture

Ogliarola

Bradano

Maiatica

Ferrandina

rant (1 = one a year or less; 2 = less than once a month;

3 = 13 times a month; 4 = once a week; 5 = 24 times a
week; 6 = every day; 7 = more than once a day).
A consumer was included in the panel if he/she: (1)
listed at least four of the ve typical oils produced in
Lucania and included in the list the three oils used in this
study; (2) reported buying each of the three selected typical oils at least 3 times a year; (3) reported consuming
each of the three selected typical oils at least 13 times
a month. Because of the restrictive criteria used in the
selection of consumers, only 34 consumers satised all
the conditions and agreed to participate. Thus the consumer panel consisted of 15 females and 19 males aged
between 25 and 50.
2.3. Information
Three forms of information each indicating one of the
three dierent origins of the six oils (olive cultivars and
their relative production area) were used. One form was
An extra virgin olive oil made exclusively from Coratina olives grown in the Vulture area. The Ogliarola
and Maiatica oils from the Bradano and Ferrandina
areas respectively were described in the same way.
2.4. Procedures
Consumers participated in seven sessions, at threeday intervals. Each session lasted about 30 min. All
the tests were conducted in individual sensory booths
and the data were collected with the FIZZ software
ver. 1.31 (Biosystemes, Couternon, France). Table 2
summarises the experimental design.
2.4.1. Session 1. Blind-condition-liking of olive oils
and hedonic expectations created by the information
about the origin of the experimental olive oils
During the rst session consumers were asked to rate
their liking of the six experimental oils presented in a blind
condition. After a few minutes they had to rate their liking
expectations for (a) oils produced exclusively from Coratina olives grown in the Vulture area; (b) oils produced
exclusively from Ogliarola olives grown in the Bradano
area; (c) oils produced exclusively from Maiatica olives
grown in the Ferrandina area. During this evaluation
the information about the origin of the oils was shown
on the computer screen in each booth and the assessors
did not have a real product in front of them.
2.4.2. Session 2. Eect of information about the origin
on perceived liking of the experimental olive oils
In this session, consumers were asked to rate their
perceived liking of the six oils presented with the same
information used in session one. In this session the three
forms of information really corresponded to the tasted
oils (see Table 2).

G. Caporale et al. / Food Quality and Preference 17 (2006) 116125

119

Table 2
Summary of the experimental design for the consumer study
Session

Stimulus presentation

Type of evaluation

Type of rating

Aim

Six typical oils

Tasting

Liking

Study on the oils liking in the blind

condition and hedonic expectations
created by the information about the
origin of the experimental olive oils

Three forms

Expectation of liking

Six typical oils with the form

corresponding to their origin

Tasting

Liking

Study on the eect of information

about the origin on actual liking of
the experimental olive oils

Six typical oils with the form

corresponding to their origin

Tasting

Typicality

Study of the typicality of the

experimental oils

Nine sensory descriptors

Appropriateness

Study of the appropriateness of

sensory descriptors of typical extra
virgin olive oils

Six typical oils

Tasting

Bitterness and pungency intensity

Study on the oils bitterness and

pungency intensity both in blind
condition and as expected in relation
to the information about the origin

Three forms

Expected bitterness and pungency intensity

Coratina oils

Tasting

Bitterness and pungency intensity

Coratina oils with the form

corresponding to their origin

Tasting

Bitterness and pungency intensity

After tasting the products without (session 1) or with

information (session 2), consumers rated the hedonic
responses (blind, expected and actual) on a 100 mm linear scale ranging from Extremely unpleasant at the
left end to Extremely pleasant at the right end.
2.4.3. Session 3. Study of the typicality of the
experimental oils
In the third session consumers were asked, after tasting, to rate the typicality of the six experimental oils
presented with the same information that was used in
sessions 1 and 2. This response was given on a
100 mm linear scale ranging from Not at all typical
at the left end to Very typical at the right end.
2.4.4. Session 4. Study of the appropriateness of
sensory descriptors of typical extra virgin olive oils
In this session consumers evaluated the appropriateness of certain sensory attributes in describing the typicality of extra virgin olive oils of Coratina, Ogliarola
and Maiatica. Respondents rated the appropriateness
of nine attributes selected from those generally used in
the sensory evaluation of extra virgin olive oils (Mojet
& de Jong, 1994): apple, cut grass, almond, artichoke,
green olive, tomato, pungency, astringency and bitterness. The appropriateness response was rated on a 7point scale ranging from 1 = not at all appropriate
to 7 = extremely appropriate, according to the

Study on the eect of the

information about the origin on the
perceived bitterness and pungency
intensity in Coratina oils

following instructions: Please read carefully the list of

sensory attributes shown in the video and rate the
appropriateness of each one in describing the typicality of Coratina, Ogliarola and Maiatica oils.
2.4.5. Session 5. Study of the bitterness and
pungency intensity of oils, both in a blind condition
and as expected in relation to the information
about the origin
In the fth session, performed after the analysis of results collected in session 3 and the selection of the most
typical oils, consumers were asked rst to evaluate the bitterness and pungency intensity of the six experimental oils
presented in a blind condition and then to rate the
expected intensity of bitterness and pungency for the three
information forms. These sensory descriptors were selected as the most appropriate in describing these typical
extra virgin olive oils. The study of sensory expectations
was carried out using the same procedures as in the study
of liking expectations. The bitterness and pungency intensity ratings obtained by consumers in a blind condition
were related to those obtained by a trained panel.
2.4.6. Session 6. Blind-condition perception of
bitterness and pungency intensity in Coratina oils
In this session consumers rated the bitterness and
pungency intensity of Coratina oils evaluated in a blind
condition.

120

G. Caporale et al. / Food Quality and Preference 17 (2006) 116125

2.4.7. Session 7. Study of the eect of the information

about the origin on the perceived bitterness and pungency
intensity of Coratina oils
In the last session consumers rated the bitterness and
pungency intensity of the experimental oils presented
with the information reported in session 1.
The sensory ratings were collected on a 100 mm linear scale ranging from Not at all at the left end to
Extremely at the right end.
2.5. Sensory analysis by trained panel
In order to test the consistency of the consumer data
with the data obtained from a trained panel, the bitterness and pungency intensity of the six experimental oils
was evaluated by a panel of 10 trained judges. The oils
were evaluated in triplicate in evenly balanced sequences. A 9-point scale ranging from Not at all at
the left end to Extremely at the right end was used
for this test.

of actual liking, typicality response, as well as blind

bitterness and pungency intensity, assuming subjects
and oils as factors, and in terms of expected liking and
expected bitterness and pungency intensity, assuming
subjects and information as factors. A two-way analysis
of variance, assuming subjects and attributes as factors,
was also performed on the appropriateness ratings in
order to reveal signicant dierences between proposed
sensory attributes.
Finally, Pearson correlation coecients were calculated between the mean intensity scores obtained from
consumer and trained panels for both bitterness and
pungency attributes in order to test the agreement of
the consumer panel data with the trained panel.

3. Results
3.1. Blind evaluation and hedonic expectation
created by the information about the origin of the
experimental olive oils

2.6. Data analysis

The study on the occurrence of both hedonic and sensory disconrmations was carried out by paired t-tests
performed on the dierences between the expected and
blind mean hedonic ratings and bitterness and pungency
mean intensity ratings. In the same way a paired t-test
was carried out between actual and blind mean ratings
as well as between actual and expected mean ratings in
order to study if an assimilation eect occurred and
what its extent might be.
Two-way ANOVA and Fisher least signicant dierence multiple comparison test were performed to reveal
signicant dierences amongst the oil samples in terms

Blind and expected mean liking scores and their standard errors are reported in Table 3. The two-way analysis of variance (ANOVA), carried out on the blind
hedonic ratings of the six experimental oils, showed that
the oils were not signicantly dierent in terms of blind
liking (p > 0.05). Similarly the two-way analysis of
variance revealed that the hedonic expectation ratings
obtained after examination of the three information
forms about the origin of the oils were not signicantly
dierent (p  0.05). These results were not unexpected
since the consumers recruited in this study were selected
on the basis of their familiarity and use of these typical
olive oils.

Table 3
Expectation eect on the oils acceptability
Product

Ratings

A
p

A1

48.12
(2.74)

58.76
(2.06)

55.26
(2.51)

10.65
0.004
Disconrmation

7.15
0.02
Assimilation

A2

55.62
(3.01)

58.76
(2.06)

58.17
(2.36)

3.15

n.s.

2.55

n.s.

B1

51.85
(3.49)

57.70
(2.47)

52.70
(3.04)

5.85

n.s.

0.85

n.s.

B2

47.18
(3.95)

57.70
(2.47)

57.41
(2.82)

10.53
0.03
Disconrmation

10.23
0.02
Assimilation

C1

52.38
(3.03)

61.17
(2.76)

55.20
(2.81)

8.79
0.02
Disconrmation

2.82

C2

57.73
(3.28)

61.17
(2.76)

55.08
(2.85)

3.44

n.s.

3.61

3.50

n.s.

0.29

n.s.

n.s.
n.s.

E = Expected acceptability mean scores; B = blind acceptability mean scores (baseline); A = actual acceptability mean scores (with information);
M = mean; standard error in parentheses.

G. Caporale et al. / Food Quality and Preference 17 (2006) 116125

3.2. Eect of information about the origin on actual

liking of the experimental olive oils
Table 3 presents expected and actual liking (R) mean
ratings of the six experimental oils and the results from
the paired t-test. A signicant dierence (p 6 0.05)
shows that disconrmation occurred. In particular results from the oil A1 carrying the information oil from
Coratina of Vulture area reveal that the expected mean
ratings were signicantly higher (t33 = 3.09, p = 0.004)
than the blind score. This signicant dierence reveals
that a negative disconrmation occurred. On the contrary no disconrmation was observed for the oil A2
carrying the same information and for the oil B1 presented as oil from Ogliarola of the Bradano area
whereas for the oil B2, carrying the same information
as B1, a negative disconrmation (t33 = 2.28, p = 0.03)
occurred. Finally for the two oils presented as oil from
Maiatica of the Ferrandina area a negative disconrmation (t33 = 2.34, p = 0.02) was generated for oil C1
while no disconrmation was found for oil C2.
The signicance of the dierences between the actual
liking of the product (R) and the baseline (B) calculated
by a paired t-test allowed us to evaluate the eect of the
resulting disconrmation (Siret & Issanchou, 2000). The
results are shown in Table 3. A signicant dierence
(p 6 0.05) reveals an eect of information on hedonic
scores. Amongst the conditions that generated hedonic
disconrmation, oil A1 presented an assimilation eect
(t33 = 2.35; p = 0.02) associated to the negative disconrmation. Furthermore the study of the signicance of
the dierences between the actual (R) and the expected
(E) liking of the product allowed us to evaluate if the
assimilation was complete or not (Siret & Issanchou,
2000). Thus, as reported in Table 3, for the oil A1 the
occurring assimilation eect was total (p > 0.05). The
oil B2 also showed an assimilation eect (t33 = 2.53,
p = 0.02) associated to a negative disconrmation and
this assimilation resulted complete too (p  0.05).
For the oil C1 assimilation was not signicant
(p = 0.38). Similar results, following both positive and
negative disconrmation, have been already reported
in previous works (Caporale & Monteleone, 2004;
Lange et al., 1999; Siret & Issanchou, 2000). Although
no clear interpretation of the phenomenon has been
given, this result could be partly explained by a smaller
and less signicant disconrmation (see Table 3). It is
possible that this explanation is not the only possible
one. In any case, the outcome of these observations does
not allow the formulation of specic hypotheses.
Results from this study showed that when these typical oils, which present the same hedonic expectations,
were tasted a signicant disconrmation occurred for
three of the six products and these disconrmations
proved to be negative. The dierent origin specied in
the information did not aect the disconrmation occur-

121

rence. In fact oils with the same origin (i.e. Coratina,

etc.) may or may not lead to disconrmation. Results
from this study indicate that identical expectations were
to some extent conrmed or disconrmed sometimes
generating assimilation.
3.3. Study of the typicality of the experimental oils
In order to further investigate the reason for these
disconrmations, typicality response was studied on
the six typical oils.
Results showed a signicant eect (F2,27 = 8.35;
p = 0.001) of the factor oil. Typicality mean ratings
and their standard errors are reported in Table 4 and
showed that the sample A2 Coratina of the Vulture
area oil presented the highest typicality ratings. Results from the ANOVA indicate that subjects perceived
some of the experimental oils to be more typical than
others and these dierences were not aected by the
origin of the oils. These results were compared to the
results of the two-way ANOVA model conducted on
the actual hedonic ratings which showed that the factor
product had no signicant eect (p > 0.05). Thus, A1
and A2 oils diered (p = 0.02) in terms of typicality
even though they are equally pleasant. These results allowed us to hypothesize that the typicality response
diers from the hedonic response since familiar consumers rated the typicality response in relation to specic
sensory expectations.
3.4. Study of the appropriateness of sensory
descriptors of typical extra virgin olive oils
This stage was carried out in order to select the most
appropriate attributes for describing the sensory properties of the typical oils used in the study.
Results from the two-way analysis of variance
showed a signicant eect of the factor attribute
(F1,98 = 10.34, p = 0.001). Mean appropriateness ratings
and their standard errors are reported in Table 5 and
ranged from 3.03 (astringency) to 5.81 (pungency) on
the 7-point scale. Pungency and bitterness attribute
mean intensity ratings were the highest. The consumer
panel rated these attributes as appropriate sensory
Table 4
Typicality mean ratings of the six products
Products

Typicality

A1
A2
B1
B2
C1
C2

61.41b (2.66)
70.62a (2.38)
58.29b (2.25)
56.44b,c (3.09)
51.09c (2.85)
61.70b (3.29)

Means with a same letter are not signicantly dierent; Standard error
in parentheses.

122

G. Caporale et al. / Food Quality and Preference 17 (2006) 116125

Table 5
Mean appropriateness ratings of nine sensory attributes
Attribute

Appropriateness

Apple
Cut grass
Almond
Artichoke
Green olive
Tomato
Pungency
Astringency
Bitterness

3.88c (0.29)
4.97b (0.22)
4.03c (0.27)
4.34b,c (0.32)
4.88b (0.35)
4.94b (0.30)
5.81a (0.26)
3.03d (0.31)
5.76a (0.26)

Means with a same letter are not signicantly dierent; standard error
in parentheses.

descriptors of the presented typical extra virgin olive

oils. This is not surprising since, in general, consumers
nd it relatively easy to identify and to describe these
sensory properties (Prescott, 1999). Furthermore, previous studies have shown that the intensities of bitterness
and pungency descriptors characterise the oils and vary
greatly (Bertuccioli, 1994; Monteleone, Caporale,
Carlucci, & Bertuccioli, 1996). Mean intensity ratings
for aroma descriptors ranged from 3.88 to 4.97. Thus,
on the all, these attributes were only slightly appropriate. This evidence is probably due to the well known difculty of consumers in describing the sensory properties
of food products (Issanchou, 2004; Lawless & Heymann, 1998). Pungency and bitterness attributes were
selected on the basis of these results.
3.5. Bitterness and pungency: blind evaluation and
expectation created by the information about the
origin of the experimental oils
In order to study the eect of information relating to
origin on bitterness and pungency perceptions of experimental oils a blind evaluation of the intensity of these
attributes was carried out.
Mean intensity ratings of bitterness and pungency for
the six experimental oils as obtained by both consumer
and trained panels are reported in Table 6.
Results of the two-way analysis of variance (Table 6)
showed that the factor oils had a signicant eect on both

bitterness and pungency. The F-value, associated to the

oil factor for bitterness was higher than the F-value computed for pungency, indicating that samples were better
discriminated in relation to bitterness than pungency.
The relationships between the mean intensity scores
for both attributes, which were obtained from consumers and the trained panel were investigated. Bitterness
and pungency mean intensity ratings from the consumer
panel were signicantly correlated to the trained panel
data (rbitterness = 0.89, p = 0.01; rpungency = 0.81, p =
0.04). The correlation coecient values derived from
the studied relationships indicate that consumer data
tended to be consistent with the data obtained from
the trained panel.
Results of the two-way analysis of variance independently computed on the expected bitterness and pungency ratings for the three proposed origins, showed
that information about cultivar signicantly aected bitterness and pungency expectations (bitterness: F3,14 =
29.34, p < 0.001; pungency: F3,14 = 24.98, p < 0.001).
The results from the Fishers least-signicant-dierence
test showed that the expected bitterness and pungency
mean ratings for Coratina oil are signicantly higher
(p < 0.001) than those computed for Ogliarola and
Maiatica oils. No signicant dierences (p  0.05)
were found between Ogliarola and Maiatica in the
expected mean ratings for both attributes. Results seem
to indicate that bitterness and pungency were expected
to contribute to the sensory characterization of Coratina oil. On the other hand several studies showed that
the sensory prole of Coratina oils is characterized by
a strong intensity of bitterness and pungency because of
their phenol content (Tsimidou, 1998).
Due to the lack of clearly evident sensory expectations for Ogliarola and Maiatica oils, we selected
only the Coratina oils for the purpose of studying
the eect of information about origin on sensory
expectations.
3.6. Study of the eect of information about the
origin of oil on the perceived bitterness and pungency
In order to study the eect of information regarding
origin on the actual perceived intensity of bitterness

Table 6
Blind mean intensity ratings of bitterness and pungency from consumer and trained panels
Mean ratings
A1
Consumer panel
Bitterness
47.44b
Pungency
44.23a,b
Trained panel
Bitterness
Pungency

5.57a
5.93a,b

Fsample

LSD

45.35b
43.23b

13.63
5.70

0.000
0.000

8.18
8.38

3.70c
3.70c

22.38
15.58

0.000
0.000

0.73
0.93

A2

B1

B2

C1

C2

40.23b
42.67b

45.91b
46.32a,b

64.79a
53.79a

31.70c
31.44c

4.50b
5.57b

6.10a
6.53a

2.77d
3.40c

3.83b,c
4.03c

Means with a same letter are not signicantly dierent.

n.s.
45.11
(3.39)
Pungency

E = Expected mean scores; B = blind acceptability mean scores (baseline); A = actual acceptability mean scores (with information); M = mean; standard error in parentheses.

6.44
16.00
0.002
Assimilation
22.44
0.000
Disconrmation
0.03
10.03
9.09
0.02
Assimilation
19.12
0.000
Disconrmation
57.79
(3.14)
54.20
(3.05)
64.23
(3.51)

8.62
7.97
0.025
Assimilation
16.59
0.000
Disconrmation
65.94
(3.28)
56.58
(2.72)
65.20
(3.48)
40.64
(2.83)
48.61
(3.48)
Bitterness

AA1
E
BA2

41.79
(3.63)

n.s.
0.73
25.30
0.000
Assimilation
24.56
0.000
Disconrmation
0.015

p
M

AA2

p
BA2
M

AA2

p
BA2
M

EA2

123

EA1
M

AA1
BA1

p
M

AA1

p
BA1
EA1

M
BA1

AA2
Ratings

Many researchers have studied how extrinsic factors

interact with intrinsic characteristics in the overall evaluation of a product but these studies only concerned hedonic measurements. Dierent models were considered
(Anderson, 1973) to describe how a dierence between
expected and blind evaluation inuences product evaluation and most of these studies revealed assimilation
eects (Caporale & Monteleone, 2004; Cardello & Sawyer, 1992; Deliza, 1996; Lange et al., 1999; Siret &

Attribute

4. Discussion

Table 7
Expectation eect on the perceived bitterness and pungency of the Coratina oils

and pungency in Coratina oils (A1 and A2), dierences between expected and blind mean ratings and differences between actual and blind mean ratings of both
bitterness and pungency were calculated (Table 7).
In A1 oil the signicant dierence (t33 = 4.17,
p = 0.000) between the expected and blind bitterness
mean intensity ratings revealed that a negative disconrmation occurred. Similarly the signicant dierence
between actual and blind bitterness mean ratings
(t33 = 2.35, p = 0.025) indicated that an assimilation effect was generated. The study of the signicance of the
dierence between actual and expected bitterness mean
intensity ratings showed that this assimilation was
incomplete (t33 = 2.54, p = 0.015). In the same way
the dierence between expected and blind pungency
mean intensity ratings proved to be signicant (t33 =
4.92, p = 0.000) thus revealing a negative disconrmation. Furthermore, this disconrmation generated an
assimilation eect (t33 = 2.37, p = 0.02) which was also
incomplete (t33 = 2.15, p = 0.03). This result suggests
that the Coratina Oil A1 did not seem to correspond
to expectations in terms of perceived bitterness and pungency intensity and this discrepancy induced consumers
to assimilate towards their sensory expectations albeit
not completely.
Results from the A2 oil showed that expected mean
intensity ratings for both bitterness and pungency attributes were higher (respectively t33 = 5.98, p = 0.000 and
t33 = 5.21, p = 0.000) than the baseline indicating that
negative disconrmations occurred. Furthermore the
signicance of the (RB) and (RE) bitterness and pungency mean intensity ratings (respectively t33 = 7.45,
p = 0.000; t33 = 3.42, p = 0.002) indicated that assimilations occurred and that they were complete (p > 0.05).
The eect of the information about the oils origin on
the perceived bitterness and pungency is reported in
Figs. 1 and 2.
Results obtained from the two Coratina oils (A1
and A2) are very similar, but the assimilation eects observed in the A2 oil were complete. This means that the
impact of the information about the origin on the perceived intensity of bitterness and pungency was stronger
in the A2 oil than in the A1 oil.

EA2

G. Caporale et al. / Food Quality and Preference 17 (2006) 116125

Bitterness mean intensity

124

G. Caporale et al. / Food Quality and Preference 17 (2006) 116125

80
70
60
50

blind

40

expected
actual

30
20
10
0
Oil A1

Oil A2

Fig. 1. Eect of expectancy disconrmation for the Coratina oils

perceived bitterness.

Pungency mean intensity

80
70
60
50

blind
expected

40

actual

30
20
10
0
Oil A1

Oil A2

Fig. 2. Eect of expectancy disconrmation for the Coratina oils

perceived pungency.

Issanchou, 2000). In the present study both hedonic and

typicality measurements were carried out in order to
study how their corresponding responses can be aected
by information about the origin of the extra virgin olive
oils. Authors are conscious that the restricted number of
subjects, due to the restrictive criteria used in the selection of consumers, might have caused some possible limitations to the study. Anyway, results obtained by the
hedonic measurements conrm the ndings of previous
studies (Siret & Issanchou, 2000; Tuorila, Meiselman,
Cardello, & Lesher, 1998) which revealed that the information evoking the origin aects product acceptability
and creates a favourable hedonic expectation by familiar
consumers (Caporale & Monteleone, 2001).
In our study information about the origin led in certain cases to a negative disconrmation which either did
or did not generate an assimilation eect for overall liking. In any case, when consumers had external information, no signicant dierences appeared between the oils
in terms of hedonic ratings.
On the contrary, the results obtained from the typicality response indicated that the experimental oils had
dierent levels of typicality. Information on origin did
not have the same eect on dierent consumer responses
such as liking or typicality.

The signicant dierences found in the typicality

scores for the oils under examination raise the problem
of dening the extent to which sensory properties inuence this result. If their inuence is signicant, then
information on origin could generate expectations
regarding specic sensory properties, in consumers
who are familiar with the product.
The results have highlighted the fact that, even
though consumers may be familiar with the product,
they tended to consider only those attributes which are
relatively easy to recognise (e.g. bitterness and pungency) as being appropriate for the description of the
sensory properties of the presented oils.
The reliability of the consumers sensory responses
was supported by the consistency of consumer data with
the data obtained from the trained panel.
No correlations were found between typicality scores
and bitterness, or pungency scores (rbitterness = 0.001,
p = 0.99; rpungency = 0.246, p = 0.64) obtained in the
blind condition. This nding would seem to show that
the relationship between typicality of the presented oils
and these attributes was not univocal (i.e. the greater
the intensity of bitterness perceived the greater the typicality) but rather that it was specic (i.e. the same perceived intensity of bitterness may be associated with
dierent perceptions of typicality for oils of dierent origins). On the basis of these results we cannot exclude the
possibility that there may be other attributes, besides
bitterness and pungency, which are important descriptors of other typical oils. This hypothesis could be the
focus of a study on a large number of oils.
In spite of its limitations, this study has, nevertheless,
made it possible to evaluate the eect of information
about origin on expectations in respect of specic sensory properties of typical oils. Consumers assimilated
the perception of bitterness and pungency to their expectations as a consequence of negative disconrmation.
However, complete assimilation to expectations occurred in only one of the two oils presented. As already
stated, the possible consequences of incomplete assimilation towards hedonic expectations have been widely debated by many authors. Very little has been said about
the consequences of incomplete assimilation towards
expectations concerning the sensory properties of products. The data at our disposal does not permit us to
make detailed hypotheses. However, the results of some
experiments ought to be noted. The eect of information
about origin on the perceived bitterness and pungency
of the Coratina oils indicated that sensory disconrmations leading to complete assimilation in both bitterness
and pungency perception were associated to higher
typicality ratings. No analogous correspondence between complete assimilation and hedonic rating was
found. If the typicality response has weight in the consumers evaluation of the perceived quality of a typical
product then it cannot be excluded that under repeated

G. Caporale et al. / Food Quality and Preference 17 (2006) 116125

tasting, sensory disconrmations which do not assimilate totally, can determine a loss of product reputation
through lack of reliability.

5. Conclusions
Results from this research conrmed that information evoking the origin of food product create a
favourable hedonic expectation in familiar consumers.
Interestingly, information on origin did not have the
same eect on dierent consumer responses such as liking or typicality. Although, it was shown that information about origin aects the expectations with regards
to specic sensory attributes in familiar consumers, the
methodological approaches to studies on sensory property expectations need perfecting. Furthermore, data
from the present research suggests that further studies
are needed to verify if, under repeated tasting, it is possible that sensory disconrmations which do not assimilate totally, can determine a loss of product reputation.

References
Anderson, R. E. (1973). Consumer dissatisfaction: the eect of
disconrmed expectancy on perceived product performance.
Journal of Marketing Research, 10, 3844.
Bertuccioli, M. (1994). A study of sensory and nutritional quality of
virgin olive oil in relation to variety, ripeness and extraction
technology. Overview of three year study and conclusion. Grasas y
Aceites, 45(12), 5559.
Bilkey, W. J., & Nes, E. (1982). Country-of-origin eects on product
evaluations. Journal of International Business Studies, 13, 8999.
Caporale, G., & Monteleone, E. (2001). Eect of expectations induced
by information on origin and its guarantee on the acceptability of a
traditional food: olive oil. Sciences des Aliments, 21, 243254.
Caporale, G., & Monteleone, E. (2004). Inuence of information
about manufacturing process on beer acceptability. Food Quality
and Preference, 15, 271278.
Cardello, A. V. (1994). Consumer expectations and their role in food
acceptance. In H. J. H. MacFie & D. M. H. Thompson (Eds.),
Measurement of food preferences (pp. 253297). London: Blackie
Academic Press.
Cardello, A. V., & Sawyer, F. M. (1992). Eects of disconrmed
consumer expectations on food acceptability. Journal of Sensory
Studies, 7, 253277.
Deliza, R. (1996). The eects of expectation on sensory perception and
acceptance. Unpublished PhD thesis, University of Reading, UK.
Deliza, R., & MacFie, H. J. H. (1996). The generation of sensory
expectation by external cues and its eect on sensory perception
and hedonic ratings: a review. Journal of Sensory Studies, 7,
253277.
Deliza, R., MacFie, H., & Hedderley, D. (2003). Use of computer
generated images and conjoint analysis to investigate sensory
expectations. Journal of Sensory Studies, 18(6), 465486.
Dichter, E. (1962). The world customer. Harvard Business Review, 40,
113122.

125

Goering, P. A. (1985). Eects of product trial on consumer expectations, demand, and prices. Journal of Consumer Research, 12,
7482.
Issanchou, S. (1996). Consumer expectations and perceptions of meat
and meat product quality. Meat Science, 43, S5S19.
Issanchou, S. (2004). Sensory and hedonic consumer expectations
towards typical food products. In European Conference on Sensory
Science of Food and Beverages, Florence, Italy, 2629 September.
., Berglund, L., Lea, P., & Risvik, E. (1999).
Johansson, L., Haglund, A
Preference for tomatoes, aect by sensory attributes and information about growth conditions. Food Quality and Preference, 10,
289298.
Kahkonen, P., & Tuorila, H. (1996). How information enhances
acceptability of a low-fat spread. Food Quality and Preference, 7,
8794.
Kahkonen, P., & Tuorila, H. (1998). Eect of reduced-fat information
on expected and actual hedonic and sensory rating of sausage.
Appetite, 30, 1323.
Lange, C., Rousseau, F., & Issanchou, S. (1999). Expectation, liking
and purchase the eect of order behaviour under economical
constraint. Food Quality and Preference, 10, 3139.
Lawless, H. T., & Heymann, H. (1998). The Central DogmaAnalytic
vs. Hedonic Tests. In Sensory evaluation of food principles and
practices (pp. 1316). Kluwer Academic Publishers Group.
MacFie, H. J. H., Bratchell, N., Greenho, K., & Vallis, L. V. (1989).
Designs to balance the eect of order of presentation and rst order
carry-over eects in hall tests. Journal of Sensory Studies, 4,
129148.
Mojet, J., & de Jong, S. (1994). The sensory wheel of virgin olive oil.
Grasas y Aceites, 45, 4247.
Monteleone, E., Caporale, G., Carlucci, A., & Bertuccioli, M. (1996).
Prediction of virgin olive oil sensory prole. Industrie Alimentari,
35, 10661072.
Peterson, R. A., & Jolibert, A. J. P. (1995). A meta-analysis of countryof-origin eects. Journal of International Business Studies, 26,
883900.
Prescott, J. (1999). Flavour as a psychological construct: implications
for perceiving and measuring the sensory qualities of foods. Food
Quality and Preference, 10, 349356.
Prescott, J., & Young, A. (2002). Does information about MSG
(monosodium glutamate) content inuence consumer rating of
soups with and without added MSG? Appetite, 39, 2533.
Siret, F., & Issanchou, S. (2000). Traditional process: inuence on
sensory properties and on consumers expectation and liking.
Application to pate de campagne. Food Quality and Preference,
11, 217228.
Steenkamp, J. B. E. M. (1990). Conceptual model of the quality
perception process. Journal of Business Research, 21, 309333.
Tuorila, H., Cardello, A. V., & Lesher, L. L. (1994). Antecedents and
consequences of expectations related to fat-free and regular-fat
foods. Appetite, 23, 247263.
Tuorila, H., Meiselman, H. L., Cardello, A. V., & Lesher, L. L. (1998).
Eect of expectations and the denition of product category on the
acceptance of unfamiliar foods. Food Quality and Preference, 9,
421430.
Tsimidou, M. (1998). Polyphenols and quality of virgin olive oil in
retrospect. Italian Journal of Food Science, 10(2), 99116.
Van der Lans, I. A., van Ittersum, K., De Cicco, A., & Loseby, M.
(2001). The role of the region of origin and EU certicates of origin
in consumer evaluation of food products. European Review of
Agricultural Economics, 28(4), 451477.
Verlegh, P. W. J., & Steenkamp, J. B. E. M. (1999). A review and
meta-analysis of country-of-origin research. Journal of Economic
Psychology, 20, 521546.