950 K. Hoppert et al. / Appetite 59 (2012) 949–955
such as nutritive or health-related claims may shape the final deci- Variation of intrinsic properties: Preparation of yoghurt samples
sion (Johansen, Naes, Oyaas, & Hersleth, 2010; Visschers & Siegrist,
2009; Wansink & Park, 2002). Plain yoghurt and yoghurt mixtures (0.1%, 1.5%, 3.5% or 10% fat)
Contradictory results in literature may be explained by taking served as basis for vanilla yoghurt preparation. By considering
into account that intrinsic and extrinsic attributes evoke different residual lactose, total sugar content was adjusted to 16 (regular)
responses which jointly determine the decision for a particular or 11.2% (relative reduction: 30%). A total of 16 combinations
food. For example, studies carried out with cheese revealed posi- was prepared by incorporating 4% or 7% vanilla flavouring prepara-
tive (Childs & Drake, 2009; Drake, Lopetcharat, & Drake, 2009; tion. The procedure was the following: sugar was weighed into a
Yates & Drake, 2007) or negligible (Levis, Chambers, & Johnson, beaker, dissolved in vanilla flavouring and mixed with yoghurt.
2000; Ritvanen et al., 2005) effects of an increasing fat content Depending on fat content and the amount of flavouring, 10–40 ll
on product acceptance. Because intrinsic and extrinsic attributes of 1% b-carotene in demineralised water was added per 100 g to
are processed by different senses (only vision is responsible for ensure that colour differences DE were, based on reflectance mea-
extrinsic attribute perception), different mechanisms come into surements, <3. Sensory experiments were carried out after storage
play. In this context, we expect that fat content is evaluated differ- for 24 h at 6 C. Approx. 40 g portions were filled into glasses
ently: with increasing fat content, there might be a steady increase which were closed with twist-off lids. Sample serving temperature
in sensory liking. Considering its extrinsic expression, consumer was 10 +/- 2 C.
judgments may be strongly affected by health-orientation and/or
social norms so that a higher fat content might be viewed rather Variation of extrinsic properties: Labels
negatively. The final choice is then a holistic expression of both
processes. To date, only few studies (e.g., Ares, Gimenez, & Deliza, Yoghurt glasses were served with labels, attached by the exper-
2010; Enneking, Neumann, & Henneberg, 2007; Johansen et al., imenter immediately before serving. We systematically manipu-
2010) examined the perception of intrinsic and extrinsic attributes lated nutrition declaration and other elements of the labels to
simultaneously. However, only a limited set of attributes was ex- have extrinsic attributes on the same levels as intrinsic attributes.
plored, and the same attributes were not manipulated both intrin- All other extrinsic information was identical. Labels showed a
sically and extrinsically at the same time so that it remains unclear mountain motive and a fake brand name and contained the follow-
how both processes interact in food choice. ing design elements: (1) A table with nutrition information (fat
We aim at filling this void by presenting an approach that con- content, sugar content, energy per 100 g) in a design typical for
siders intrinsic and extrinsic expression of three attributes of yo- the German market. (2) Claims for variation of fat content (‘‘fat
ghurt. Two of them (fat and sugar content) can be considered as reduced’’, ‘‘cream yoghurt’’), for lower carbohydrate content
health-related, whereas the third one (flavour) is health-unrelated. (‘‘reduced sugar’’), and for higher flavouring (‘‘more vanilla
We intend to answer the following questions: (1) Are intrinsic flavour’’). In line with the study design, 16 variants (4 fat content,
properties of real yoghurt or extrinsic information supplied with 2 sugar content, 2 flavour level) were designed and printed on
the product primarily responsible for food choice? (2) How do self-adhesive paper.
effects of intrinsic and extrinsic processes differ for the same attri-
bute? (3) Do isolated approaches, which neglect the divergent pro-
Participants
cessing of intrinsic and extrinsic attributes, bias the assessment of
consumer preferences? For that purpose, we establish a procedure
We recruited 101 students (65 female, average age: 24.1 +/-
which combines sensory experiments with conjoint analysis (CA)
3.7 years; 36 male, average age: 23.7 +/- 3.1 years) from lectures.
that is frequently used for assessing the impact of extrinsic attri-
In a questionnaire, they were asked for body height and body
butes on product evaluation (e.g., Enneking et al., 2007; Jaeger,
weight; body-mass index range was 18.1–41.3 (median: 23.1).
2000; Jervis, Lopetcharat, & Drake, 2012; Valeeva, Meuwissen,
Fifty-seven percent of the participants claimed that they do their
Oude Lansink, & Huirne, 2005). For the first time, we include both
shopping themselves, and 74% indicated to consume yoghurt
aspects of the same attributes: subjects evaluated yoghurt with
several times a week. For each single session, which lasted approx.
different intrinsic intensities which were systematically modified
1 h, the participants received an allowance of 10 €.
through yoghurt formulation, and simultaneously considered
extrinsic information given by labelling. Owing to the fact that
traditional full profile CA is restricted to only a small number of Experimental procedure
attributes and levels, we apply a computer-based, adaptive type.
The evaluation tasks of adaptive conjoint analysis (ACA; Johnson, Rationale for using adaptive conjoint analysis
1987) are optimised on the basis of previous decisions so that reli-
able results can be achieved with a small number of assessments. Instead of directly asking a subject how important an attribute/
To the best of our knowledge, ACA in this setup has not been used level combination is, we apply a decompositional approach. To
until now in combination with sensory assessment of foods. increase ecologic validity of the results, participants are asked to
evaluate full product profiles in a CA setting. Because of the large
number of attributes (three intrinsic and three extrinsic character-
Materials and participants istics, 16 combinations in each category), which allows 256 combi-
nations, we applied an adaptive, computer-based variant of CA. By
Materials combining compositional and decompositional methods, ACA is
capable to reliably estimate individual preferences for a large set
Plain stirred yoghurt with a fat content of 0.1% or 10% (residual of attributes while using a manageable number of choice tasks
lactose: 6.0% or 4.8%; protein content: 4.7% or 3.5%, respectively) (Orme & Johnson, 1996). For several reasons, we omitted the com-
was obtained from Molkerei Hainichen-Freiberg GmbH & Co positional part by setting equal importance for all attributes and
(Freiberg, Germany). A natural vanilla flavouring preparation was their levels: (1) Evaluation of single attributes in the compositional
provided by Zentis GmbH & Co KG (Aachen, Germany). Sucrose part is risky because respondents may guess the purpose of the
was purchased in a supermarket, and S-carotene was from Ru
การแปล กรุณารอสักครู่..
