Professional Documents
Culture Documents
369391
doi:10.1093/eurrag/jbi011
Abstract
Research on consumer quality perception is reviewed using the Total Food Quality
Model as a structuring device. The relationship between food safety and quality is
addressed, and is discussed in the context of research on consumer risk perception.
Quality and safety perception is linked to food choice and consumer demand,
addressing questions of price perception and the validity of willingness-to-pay
measurements. It is concluded that food quality and safety are central issues in
todays food economics, though many research questions remain to be addressed.
Keywords: consumer, quality, risk, branding, meansend approach
JEL classification: Q13, Q18, M31, D11
370
Klaus G. Grunert
and low production costs. But with increasing global competition on food
markets, new competitors have been entering the competitive arena, and old
competitors have been catching up on their competencies in efficient production and quality control. In such a situation, more fragmented, heterogeneous and dynamic consumer demand creates opportunities for those
producers and value chains that are willing to take the risk to differentiate
their products, aim at serving specific target markets, and adapt to local conditions even under the wings of a global marketing approach. As a result,
many sectors in agribusiness these days compete not only on efficiency and
quality control, but on adding value. Adding value is a customer-oriented
conceptwe only add value to food products to the extent that those consumers at whom the final product is targeted actually perceive these products
as betterperceive them as having more quality.
The prominence of the concepts of quality and safety in agribusiness is
thus driven by all actors in the marketplace. And it is not surprising that
research has followed.
371
leading to shadow prices for certain quality attributes. For products not
(yet) on the market, or when actual demand data is not available, consumer
willingness to pay can be measured using methods such as contingent valuation or experimental auctions.
Whereas the first stream thus deals with the demand side, the second
deals with the corresponding supply side. Providing added safety and/or
differentiated quality may require changes in the organisation of agricultural
and food production, mainly with regard to governance structures of value
chains. Whenever issues of quality and safety cannot be addressed at the
final processing level, but have to pervade the whole value chain, relationships among value chain members may have to change. Food safety, for
example, is closely linked to the traceability issue and often leads to closer
links among value chain members. When quality is already differentiated at
the farm level, new forms of contractual arrangements between farmers and
processors may be called for. The difficulties of making such arrangements,
especially in co-operative organisations, have been discussed widely in the
literature.
These two streams of research together constitute the traditional economic approach to dealing with quality and safety issues. More recently, a
third stream has been added. It deals with the question of how quality and
safety is perceived by consumers, and how these perceptions influence
consumer decision-making. Consumer preferences are thus not only
regarded as being revealed in their demand, but their formation in interaction with the supply of goods becomes a separate area of inquiry. Thus,
this stream of research can be seen as mediating between supply and
demand, as it is the perception of the supply of goods that leads to the
demand for these goods, as illustrated in Figure 1.
In this paper, I will concentrate on consumer perception of and demand
for quality and safety, i.e. the right-hand part of Figure 1. Issues concerning
the provision of quality and safety will be mentioned only briefly at the end
of the paper.
There is an abundance of ways in which the term quality, both in food and
otherwise, has been defined (see the 1995 special issue of Food Quality and
Preference for a broad range of proposals). There is general agreement that
quality has an objective and a subjective dimension. Objective quality refers
to the physical characteristics built into the product and is typically dealt
with by engineers and food technologists. Subjective quality is the quality
as perceived by consumers. The relationship between the two is at the core
of the economic importance of quality: only when producers can translate
consumer wishes into physical product characteristics, and only when
372
Klaus G. Grunert
consumers can then infer desired qualities from the way the product has
been built, will quality be a competitive parameter for food producers.
In the subjective realm we can, as a gross simplification, distinguish
between two schools of thought about quality. The first one, which we can
call the holistic approach, equates quality with all the desirable properties a
product is perceived to have. The second, which we can call the excellence
approach, suggests that products can have desirable properties that consumers, in their own language, may not view as part of quality. In food, convenience is sometimes named as an example: consumers may say that
convenience goods are generally of low quality, even though they regard
convenience as a desirable property of food products (see, e.g. Zeithaml,
1998; Olsen, 2002). In the following, we will use the holistic approach.1
It follows from the holistic approach that food safety is part of food
quality, at least to the extent that consumers believe food safety to be a desirable property. We can usually assume that this is the case, at least up to a
point. Safety may, however, be different from other quality aspects in the way
in which it affects consumer decisions, a question to which we will return.
373
374
Klaus G. Grunert
What motivates consumers to buy one food product rather than another?
This issue has been dealt with in the means end approach to consumer
behaviour (Reynolds and Olson, 2001), which has been widely used in analysing consumer food choices. The basic assumption of means end theory
is that consumers are not interested in products per se, but in what the product is doing for themin the self-relevant consequences of the product, in
the way the product helps them attain their life values (in this way the
approach is related to the Lancaster approach to analysing consumer
demand). Whether a consumer finds a product attractive is supposed to
depend on the extent to which this consumer can link his perception of the
products characteristics to self-relevant consequences and values. Such
links are called means end chains, because they are chains of subjective
associations where the product is a means to achieve ends as defined by the
consumer.
When asking consumers in an open-ended interview what they regard as
food products of good quality, the answers always radiate around four central concepts: taste (and other sensory characteristics), health, convenience,
andfor some consumersprocess characteristics such as organic production, natural production, animal welfare, GMO-free, etc. (Bruns et al.,
2002). Placing these concepts in a means end chain context, food quality,
as perceived by consumers, becomes an intermediate conceptmore
abstract than concrete product attributes such as fat percentages, colour and
packaging, but more concrete than life values such as being responsible,
protecting your family, or having fun and excitement in life. In other words,
from the means end perspective quality is a bridging conceptby forming
impressions of the quality of a product, consumers form a judgement on
whether the characteristics of the product, as they have been perceived, will
help in attaining that consumers life values.
Means end researchers have mostly applied a technique called laddering
to study how consumers mentally link product characteristics to more
abstract quality dimensions and from there to life values. Numerous studies
employing this method have been reported in the food area (e.g. Grunert
and Grunert, 1995; Nielsen et al., 1998; Bredahl, 1999; Jaeger and MacFie,
2000; Valette-Florence et al., 2000; Bech-Larsen, 2001; Grunert et al.,
2001; Miles and Frewer, 2001; Fotopoulus et al., 2003; Russell et al.,
2004). Results from laddering studies are usually presented in so-called
hierarchical value maps, and Figure 3 shows an example from a study on
Figure 3. Hierarchical value map for fresh fish (from Nielsen et al., 1997).
375
376
Klaus G. Grunert
377
Among the broad range of applicable extrinsic cues, three have received
special attention in the food area as possible quality indicators: brands, cues
related to product origin, and quality labels.
The role of brands is amply documented in the marketing, management
and economics literature. The classic economic argument is that brands
influence purchase to the extent that they reduce risk and communicate the
positioning of the product to the consumer (Erdem and Swait, 1998); more
psychologically oriented approaches have supplemented this by the notion
that brands having such functions will, in the mind of the consumer, have
associations that are strong, favourable and unique (Keller, 1993; Krishnan,
1996). Both types of approach imply that brands become powerful cues for
consumers to the extent that consumers actually find them predictive of the
quality of the product. To the extent that a brand is widely used for quality
inference, it accumulates brand equity, i.e. becomes a valuable asset for the
manufacturer owning it. A special phenomenon in the food area is the
recent rise of retailer brands (also called private labels or own labels) as an
alternative to the traditional manufacturer brands (Laaksonen and Reynolds,
1994; Burt, 2000). Retailer brands can, in principle, serve as quality cues to
the consumer in the same way as manufacturer brands. However, the comparative lack of a brand history and the historical association of retailer
brands with generic products and low-price alternatives results, in many
cases, in consumers taking a retail brand as a cue indicating low rather than
high quality.
Information on the place or region of origin of the product has a long history in the food area, but interest has increased after the introduction of EU
regulations on protected designations of origin (PDO) and protected geographical indications (PGI). In terms of consumer behaviour, such information is a special case of country-of-origin cues, which have been subject
to comprehensive research for several decades (Verlegh and Steenkamp,
1999). Research on country-of-origin effects has established that consumers
may use origin information as a quality cue (e.g. van der Lans et al., 2001;
van Ittersum et al., 2003). Two types of mechanism can be responsible for
this (Johansson, 1989). First, consumers may use the cue to link the product
to knowledge on the region of origin, which may be relevant for forming a
quality evaluation; this may include everything from beliefs about the quality consciousness of Germans, the food addiction of the French, and beliefs
about artisanal ways of producing ham in the Parma region. Second, consumers may use this cue during repeat purchases of the product to re-identify
a product, the quality of which they found satisfactorya process that may
be especially relevant when the product does not carry a strong brand. It follows from both mechanisms that origin information will have no effect on
quality evaluations when consumers have no knowledge about the region of
origin, when the quality of the product is not in fact experienced as desirable by the consumer, and/or when we are dealing with trial (as opposed to
repeat) purchases.
378
Klaus G. Grunert
As noted, the horizontal dimension in Figure 2 refers to how quality perception changes over time. The major (but not only) distinction to be made
here is quality perception before and after purchase. Most aspects of food
quality are either experience or credence characteristics, and the way quality
perception changes over time will differ between these.
Because experience qualities can, by definition, be evaluated with high
certainty after the purchase, the expectations about these qualities formed
based on cues can be confirmed or contradicted. Confirmation or disconfirmation of expectations is the major determinant of consumer satisfaction
and of consumer intent to repurchase the product or not (Oliver, 1997).
Especially for new products, which have been bought for the first time and
where the formation of expectations at the point of purchase therefore cannot be based on previous own experience, whether or not expectations are
confirmed is a crucial point for whether the product will become a success
or not.
379
Figure 4. Average evaluation of three types of meat before purchase and after consumption (7-point scale with 1 not so good, 7 very good) (from Grunert et al., 2004).
380
Klaus G. Grunert
the evaluation of the healthiness of meat rises after the purchase, because
the fat that was visible at the point of purchase (and was used as a cue in
evaluating healthiness) melts off during preparation.
The distinction of only two phasesbefore and after purchaseis of
course a simplification. Most food products are bought several times, and
many are bought continuously over extended periods. Quality perception
may change across the whole range of purchases. It is usually assumed,
though, that the biggest change occurs in connection with the first purchase,
because in the first purchasethe trial purchasequality expectations are
necessarily based on informational cues only, not on own experience, and
the first purchase will then lead to the first actual experience with the product, which may lead to fundamental changes in the perception of quality.
When purchasing the product for the second and subsequent timesthe
repeat purchasesprevious own experiences will play a role in forming the
quality expectations, which therefore will be more accurate. But changes
may still occur, for a variety of reasons. Learning may occur with regard to
how to handle the product, resulting in better quality experiences. The situation in which the product is consumed may change, which may have an
impact on the experienced quality. When consumers are variety-seeking and
like stimulation and change, a positive quality perception may wear off over
time. For credence qualities, the quality perception can always change when
new information about the quality becomes available.
The dimensions that constitute quality in the mind of the consumer, and
especially their weights, may change over time as well. Most notably, there
may be a tendency that those quality dimensions that are amenable to own
experience will, over time, acquire a greater weight. Whereas taste and
healthiness, for example, may have equal weight in the pre-purchase
phasewhere both types of quality perceptions are cue-basedtaste may
be given a higher weight in the period during and after consumption,
because it has now been experienced, whereas healthiness is still abstract
and information-based. Over time, there is hence a danger that credence
qualities lose out to experience qualities, because the latter are constantly
fed by own consumer experience, whereas the former are not. This problem
is especially severe for functional foods, where the credence characteristic
is the major selling point (Bech-Larsen and Grunert, 2003; Frewer et al.,
2003; Verbeke, 2005).
Food products are usually consumed after having undergone some sort of
preparation, i.e. they are used in the household production process. The
home production process and not the product itself determines the quality
perceived during and after consumption and, hence, consumer satisfaction.
There is reason to believe that there may be many instances in which the
home production process is more important for the overall quality experience than the quality of the product itself. The best piece of meat can be
ruined by frying it too long, and a good cook can prepare a delicious meal
even from a less attractive piece of meat. We know relatively little on
381
the relative influence of these factors, and especially about the role of cooking skills.
382
Klaus G. Grunert
of irradiation and a considerable delay of the adoption of GMOs on European markets. Consumer attitudes to GMOs in food production have been
widely studied (e.g. Bredahl, 2001; Burton et al., 2001; Grunert et al.,
2003; Scholderer and Frewer, 2003; Frewer et al., 2004; Hu et al., 2004).
Consumer risk perception in more general terms has been widely studied
as well, starting with the well-known work by Slovic and colleagues
(Fischhoff et al., 1978; Slovic, 1987), and a number of regularities have
been observed (Frewer et al., 2005). Three phenomena that seem to be
rather robust will be mentioned here. First, self-imposed risk is more acceptable to consumers than technology-based risk. Thus, although meal preparation at home is, by objective standards, much riskier than meal
production in a factory, consumers tend to perceive ready-made meals as
more dangerous than meals they have cooked themselves, and the perceived
risk is amplified when new and unknown technologies are used. Second,
although consumers can usually appreciate the risk associated with their
own handling of food in general terms, they believe that the probability of
being hit themselves by a problem is lower than the probability of the average consumer being hit by the same problem, a phenomenon also known as
optimistic bias. Finally, the importance of the dimensions of dread and
familiarity in risk perception has been amply demonstrated, implying that
familiar risks are perceived as less severe than unfamiliar onesa phenomenon relevant to the perception of GMOs and other forms of novel foods.
Food safety is a major topic for public policy. Regulatory responses have
been, roughly speaking, in two categories (Caswell and Mojduszka, 1999;
Ritson and Mai, 1998). The first refers to the enforcement of common standards for food safety, which has no immediate impact on consumer food
choice, but is debatable in terms of economic efficiency when consumer
preferences for safety are assumed to be heterogeneous. The second refers
to attempts to provide transparency and encourage consumers to form their
own judgements on food safety, supported by mechanisms of public participation, consumer education, and consumer information instruments such as
labelling. This latter policy, which has been central to the European
Commissions white paper on food safety (Commission of the European
Communities, 1999), may, however, be based on unrealistic assumptions
about consumers willingness and ability to process information when making daily purchases (Grunert, 2000).
383
(e.g. Baltzer, 2002; Steiner, 2004) or, more recently, mixed multinomial
logit approaches (that take into account respondent heterogeneity, e.g.
Bonnet and Simioni, 2001) have been used to estimate WTP in cases where
market transaction data (usually scanner data) are available. In those cases
where market transaction data are either not accessible or non-existent,
because we are dealing with hypothetical improvements of quality or safety,
contingent valuation (e.g. Boccaletti and Nardella, 2000; Gil et al., 2000;
Maruyama and Kikuchi, 2004), experimental auctions (Hoffman et al.,
1993; Lusk et al., 2001, 2004; Rozan et al., 2004) and conjoint analysis
(e.g. Bech-Larsen and Grunert, 2003; Enneking, 2004; Grunert et al., 2004)
have been the most popular methods. WTP estimates that are not based on
market data are often made for quality and safety improvements that have
not yet been launched on the market. Such WTP estimates are valuable
from the perspective of both public policy makers and food manufacturers.
Results from these various approaches often differ a good deal, with
WTP as derived from market transactions often being a good deal lower
than those derived from hypothetical methods (see, e.g. Shogren et al.,
1999) In the eyes of many food practitioners, this not only has discredited
non-market based WTP methods, but it has also led to a widespread notion
that price is the most important parameter in determining consumer food
choice and that interest in quality and safety improvements is often not
widespread. However, one should be cautious about such conclusions for
two reasons.
First, there is a good deal of research showing that consumers routinely
buy lots of products without knowing their prices (e.g. Chernatony and
Knox, 1992; Urbany et al., 2000; Vanhuele and Dre`ze, 2002). This price
perception research is usually based on asking consumers the price of a product right after they have placed it in their shopping basket. The percentage
of consumers not knowing the price varies widely among studies and also
across countries, but has often been found to be very lowfor example, in
the study by Vanhuele and Dre`ze (2002) in France, only 10 per cent could
recall the price of an item when asked directly at the point of sale, and only
30 per cent could guess it correctly within a 5 per cent margin or error.
Price perception is a complex phenomenon. Because, however, markets do
function and most consumers can manage their income and expenditures
flows, it appears that most consumers do form an opinion on whether the
qualities of the product outweigh the costs. This opinion may not be based
on any price information at all, only on a general assumption that in this
product category, prices are usually so low that it does not matter. It may
be based on a general assessment of the price level in the store where the
product is found, or it may be based on detailed price knowledge of this
and a range of competing product varieties. How much consumers are interested in prices, their price involvement (Lichtenstein et al., 1993; Urbany
et al., 1996), is related to a host of factors, including perceived budget
constraints, lowest price or value for money as basic purchasing motives,
or price mavenism, i.e. a desire to appear as a market expert when talking
384
Klaus G. Grunert
385
386
Klaus G. Grunert
and understand the voice of the customer (Biemans and Harmsen, 1995;
Benner et al., 2003).
Much food product differentiation has traditionally been dealt with at the
processing level. However, there has also been a trend towards increasing
differentiation already at the farm level. There are various reasons for this.
Consumers demand some kinds of product differentiation that by their
nature have to be dealt with at the farm level, such as increased animal welfare or organic production. Advances in biotechnology open up new possibilities for differentiation of both animal and plant production. And product
differentiation at the processing level involves imitation lags for competitors
that are usually short, whereas differentiation that goes back to primary production gives better protection against competitive moves.
When differentiation pervades the food chain, it will often have implications for the governance of the food chain. Differentiation starting at the
farm level leads to demands for segregation and traceability, and it requires
good communication between the members of the value chain, as intelligence on consumer demands has to travel all the way back to the farm
level. This typically requires closer forms of co-operation between value
chain members, and especially between farmers and processors, than when
the farm level output is judged by price and homogeneity only (Grunert
et al., 2005). Finding the right contractual forms for these types of relationship is a research area in its own right (e.g. Mahoney, 1992; Giraud-Heraud
et al., 1999; Boger, 2001; Bogetoft and Olesen, 2002) that is beyond the
scope of this paper. The development of relationships between farmers and
processors adds another dimension to the complexities resulting from
increased attention to food quality and safety.
References
Aarts, H. and Dijksterhuis, A. (2000). Habits as knowledge structures: automaticity in
goal-directed behavior. Journal of Personality and Social Psychology 78: 53 63.
Adaval, R. and Monroe, K. B. (2002). Automatic construction and use of contextual
information for product and price evaluation. Journal of Consumer Research 28:
572 588.
Ajzen, I. and Fishbein, M. (2000). Attitudes and the attitude-behavior relation: reasoned
and automatic processes. European Review of Social Psychology 11: 1 33.
Baltzer, K. (2002). Eftersprgslen efter fdevarekvalitet og -sikkerhed: Et pilotstudie af
danske forbrugeres eftersprgsel efter g. Working paper 11/2002. Copenhagen: FI.
Bech-Larsen, T. (2001). Model-based development and testing of advertising messages:
a comparative study of two campaign proposals based on the MECCAS model and a
conventional approach. International Journal of Advertising: The Quarterly Review
of Marketing Communications 20: 499 519.
Bech-Larsen, T. and Grunert, K. G. (2003). The perceived healthiness of functional
foods: a conjoint study of Danish. Finnish and American consumers perception of
functional foods. Appetite 40: 9 14.
387
388
Klaus G. Grunert
389
390
Klaus G. Grunert
391
All in-text references underlined in blue are linked to publications on ResearchGate, letting you access and rea