Download as pdf or txt
Download as pdf or txt
You are on page 1of 23

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/255698040

Extending the Prevalent Consumer Loyalty Modelling: The Role of Habit


Strength

Article  in  European Journal of Marketing · July 2013


DOI: 10.1108/03090561311285565

CITATIONS READS

41 1,552

4 authors:

Svein Ottar Olsen Ana Alina Tudoran


UiT The Arctic University of Norway Aarhus University
89 PUBLICATIONS   4,166 CITATIONS    28 PUBLICATIONS   461 CITATIONS   

SEE PROFILE SEE PROFILE

Karen Brunsø Wim Verbeke


Aarhus Business School Ghent University
53 PUBLICATIONS   3,453 CITATIONS    427 PUBLICATIONS   21,698 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

SRV2701 View project

Seafoodplus View project

All content following this page was uploaded by Wim Verbeke on 06 May 2014.

The user has requested enhancement of the downloaded file.


European Journal of Marketing
Emerald Article: Extending the prevalent consumer loyalty modelling: the
role of habit strength
Svein Ottar Olsen, Ana Alina Tudoran, Karen Brunsø, Wim Verbeke

Article information:
To cite this document: Svein Ottar Olsen, Ana Alina Tudoran, Karen Brunsø, Wim Verbeke, (2013),"Extending the prevalent consumer
loyalty modelling: the role of habit strength", European Journal of Marketing, Vol. 47 Iss: 1 pp. 303 - 323
Permanent link to this document:
http://dx.doi.org/10.1108/03090561311285565
Downloaded on: 04-02-2013
References: This document contains references to 51 other documents
To copy this document: permissions@emeraldinsight.com

Access to this document was granted through an Emerald subscription provided by UNIVERSITEIT GENT

For Authors:
If you would like to write for this, or any other Emerald publication, then please use our Emerald for Authors service.
Information about how to choose which publication to write for and submission guidelines are available for all. Please visit
www.emeraldinsight.com/authors for more information.
About Emerald www.emeraldinsight.com
With over forty years' experience, Emerald Group Publishing is a leading independent publisher of global research with impact in
business, society, public policy and education. In total, Emerald publishes over 275 journals and more than 130 book series, as
well as an extensive range of online products and services. Emerald is both COUNTER 3 and TRANSFER compliant. The organization is
a partner of the Committee on Publication Ethics (COPE) and also works with Portico and the LOCKSS initiative for digital archive
preservation.
*Related content and download information correct at time of download.
The current issue and full text archive of this journal is available at
www.emeraldinsight.com/0309-0566.htm

Consumer
Extending the prevalent consumer loyalty
loyalty modelling: the role of habit modelling
strength
303
Svein Ottar Olsen
Tromsø Business School, University of Tromsø, Tromsø, Norway Received 8 June 2010
Ana Alina Tudoran and Karen Brunsø Revised 20 February 2011
Accepted 6 May 2011
Aarhus School of Business and Social Sciences, Aarhus University,
Denmark, and
Wim Verbeke
Ghent University, Ghent, Belgium
Abstract
Purpose – This study aims to address the role of habit strength in explaining loyalty behaviour.
Design/methodology/approach – The study uses 2,063 consumers’ data from a survey in
Denmark and Spain, and multigroup structural equation modelling to analyse the data. The paper
describes an approach employing the psychological meanings of the habit construct, such as
automaticity, lack of awareness or very little conscious deliberation.
Findings – The findings suggest that when habits start to develop and gain strength, less planning
is involved, and that the loyalty behaviour sequence mainly occurs guided by automaticity and inertia.
A new model with habit strength as a mediator between satisfaction and loyalty behaviour provides a
substantial increase in explained variance in loyalty behaviour over the traditional model with
intention as a mediator.
Originality/value – This study contributes to the existent literature by providing an extension of
the prevalent consumer loyalty theorizing by integrating the concept of habit strength and by
generating new knowledge concerning the conscious/strategic and unconscious/automatic nature of
consumer loyalty. The study derives managerial implications on how to facilitate habit formation and
how to influence habit-based versus intention-based loyalty behaviour. The external validity of this
study is assured by nationwide representative samples in two countries.
Keywords Intention, Habit strength, Loyalty, Structural equation modelling, Food, Consumer behaviour,
Customer loyalty
Paper type Research paper

1. Introduction
Research on marketing argues that strategies to gain consumer loyalty and to prevent
consumers from switching to a competitive product are essential business questions
today (Reinartz et al., 2005). The need to understand what drives consumer loyalty has
spawned a number of publications examining the determinants and mediators of
consumer loyalty behaviour (Ball et al., 2004; Chiou and Droge, 2006; Evanschitzky and
Wunderlich, 2006; Gustafsson et al., 2005; Yi and La, 2004).
European Journal of Marketing
This work was performed within the EU FP6 Integrated Project SEAFOODplus, contract no. Vol. 47 No. 1/2, 2013
FOOD-CT-2004-506359. The financing of the work by the European Union is gratefully pp. 303-323
q Emerald Group Publishing Limited
acknowledged. 0309-0566
The first two authors contributed equally to this study. DOI 10.1108/03090561311285565
EJM Past research on the psychological process that drives loyalty behaviour has mainly
47,1/2 focused on the deliberate or goal-directedness intentions mechanism (Oliver, 1999;
Pritchard et al., 1999). For example, Oliver’s (1999) model assumes that product loyalty
is the outcome of active planning, which starts with episodes of positive cognition and
affect towards a product or brand, and ends with a strong commitment (intention)
directed towards repeatedly purchasing the same product/brand. While plausible in
304 the case of less frequent behaviours or in the initial stages of product adoption, this
assumption may not be applicable to continued behaviours or behaviours
characterised by frequent purchases (such as food and drink purchasing and
consumption), as it ignores the fact that frequently performed behaviours in stable
contexts become habitual and thus automatic over time (Limayem et al., 2007; Ouellette
and Wood, 1998). As argued by Webb and Sheeran (2006) and others, when behaviour
is performed in stable contexts and for low-involvement and frequent purchases,
consumers’ behaviour can be initiated and executed without needing the person’s
conscious intent and guidance (Webb and Sheeran, 2006, p. 261). Consumers tend to
buy the same brands of products across different shopping episodes, the same amount
at a given retail store across repeat visits, and consume similar types of foods in a meal
across days (Wood and Neal, 2009). Habits but not intention hide the most common
form of repeat purchase or repeat consumption (Triandis, 1980). Habits develop and
gain strength by satisfactory repetition, and over time become automatic, so that
repeated behaviour can occur without awareness and self-instruction (Wood and Neal,
2009). Therefore, ignoring the habit-persistence effects may systematically
overestimate the intentional product/brand loyalty (Seetharaman, 2004).
The overall objective of this work is to investigate the key role of habit strength in
loyalty behaviour in the context of food consumption behaviour. More specifically, the
objectives are:
. to introduce the most prevalent consumer loyalty model in the field of marketing
(Oliver, 1999), highlighting its strengths and shortcomings, and to extend the
current theorizing on loyalty by integrating the notion of habit strength as a
mediator between cumulative satisfaction and loyalty behaviour; and,
respectively
.
to test this extended model empirically and compare its predictive ability with
the original model in the field of food consumption.

No explicit study to the authors’ knowledge takes this approach. A few studies
asserting the role of habit in loyalty behaviour present four major shortcomings. First,
these studies lack a properly rigorous argument and substantiated theoretical base
(Limayem et al., 2007). Second, there is a lack of precision in relation to
conceptualisation and measurement of the habit notion. The habit construct is
typically conceived as past behavioural frequency or consecutive product purchase
( Jolley et al., 2006; Ouellette and Wood, 1998; Seetharaman, 2004). The majority of
studies omit the key aspects of habit, such as automaticity, lack of awareness or very
little conscious deliberation (Verplanken and Orbell, 2003). Third, most of the previous
studies do not include actual behaviour in their models to test it empirically. For
instance, Trafimow (2000) analyses the relationship between habit and intention to
behave as a proxy for actual behaviour, suggesting that future research should
examine these effects for actual behaviour. Fourth, habit is usually considered to be a
negative construct both in daily life, e.g. referred to in a context of bad habits or Consumer
addiction (Lindblad and Lyttkens, 2002), as well as in some parts of the loyalty
literature (Dick and Basu, 1994). The reason for this is the assumption that habit or
loyalty
inertia is something passive, convenient, a non-conscious form of retention associated modelling
with “spurious loyalty” and different from “true loyalty” as an active, planned and
conscious component of decision-making processes (Dick and Basu, 1994). However,
habit can be a positive trait or outcome for consumers and businesses. People develop 305
habits for convenience in order to save cognitive effort and time (Wood and Neal, 2009),
which is not negative per se. Without habits, people would be doomed to plan, guide
consciously and monitor every action (Neal et al., 2006). Buying a brand or product or
going to a particular restaurant out of habit may be a consequence of satisfied actions
over time (Triandis, 1980). On the other hand, the reason why individuals perform
active planning, compare and evaluate may be because they lack skills and knowledge
of the product or because they became less satisfied over time and continue to consider
and evaluate additional alternatives. Thus, buying, consuming or using a brand or
product out of habit may be something businesses want to encourage, even though
over time it is less driven by active planning and a decision to act.
Overall, the present study aims to address these limitations. The study contributes
to the marketing literature and practice by:
.
providing an extension of the prevailing consumer loyalty theorizing by
integrating the concept of habit strength into the traditional Oliver’s (1999)
model;
.
by generating new knowledge concerning the conscious/strategic and
unconscious/automatic nature of consumer loyalty behaviour; and
.
by proposing managerial implications or guidelines on how to enable habit
formation and how to influence habit-based versus intention-based loyalty
behaviour.

The external validity of this study is strengthened by the use of two representative
samples of consumers in two countries with different food consumption cultures
(Denmark and Spain).

2. Overview of the framework


Oliver’s (1999) hierarchical model of loyalty seeks to explain consumers’ behaviour to
repurchase or repatronize a preferred product/service consistently over time (Oliver,
1999). Oliver’s model positively relates satisfaction to loyalty behaviour through the
mediator role of intention. This perspective is in accordance with the traditional
attitude-intention-behaviour approach in social psychology (Ajzen, 1991), suggesting
that intention is the main causal mechanism behind the enactment of behaviour. In the
marketing literature, most studies rely on hierarchical mediation through intention to
predict consumer loyalty behaviour (Chiou and Droge, 2006; Evanschitzky and
Wunderlich, 2006; Han et al., 2008). However, Oliver’s model cannot account for the
possibility that long-practised behaviour may no longer be under motivational
conscious control, but rather influenced by antecedents other than intention. To
overcome this limitation and improve the explanatory power of the model further, we
incorporate the habit-strength construct as an alternative less conscious automatic
cause explaining loyalty behaviour. In developing our study, we begin with a brief
EJM discussion of the fundamental concepts. Next, we turn to discuss Oliver’s model (Model
47,1/2 1 in Figure 1) and the proposed model (Model 2 in Figure 1) in more detail.

2.1 Satisfaction
Satisfaction is the consumer’s fulfilment response and the degree to which the level of
fulfilment is pleasurable (Oliver, 1999). Satisfaction is a key to building and retaining a
306 loyal base of long-term consumers (Limayem et al., 2007). For satisfaction to affect
loyalty behaviour, consumers should experience frequent or cumulative satisfaction
episodes (Oliver, 1999). Repeated satisfaction with a product, service or brand becomes
aggregated over time and forms consumer cumulative global satisfaction with the
product, service or brand (Johnson et al., 1995).

2.2 Loyalty intentions


Loyalty intention is a deeply held commitment to repurchase a specific product, service
or brand (Oliver, 1999). Intention captures different motivational factors that influence
human behaviour (Eagly and Chaiken, 1993). According to Ajzen (1991), intention is an
indicator of how hard people are willing to try – how much effort they are planning to
exert – to perform the behaviour. Loyalty intention is often used as a substitute for
loyalty behaviour or as the ultimate dependent variable in satisfaction-loyalty studies
(exceptions include, e.g. Gustafsson et al., 2005; Seiders et al., 2005). Mittal and
Kamakura (2001) question this practice and argue that intention and behaviour are
different constructs.

2.3 Loyalty behaviour


Loyalty behaviour is the consumer’s tendency to patronize a product, revealed through
behaviour (action) that can be measured and that impacts directly on sales (Worthington
et al., 2010). Businesses pursue loyalty behaviour because such behaviour in consumers
can secure profitability and long-term sustainability (Reinartz et al., 2005). Marketing
literature defines and measures loyalty behaviour as the self-reported frequency of
purchases/consumption over time (Gustafsson et al., 2005; Pritchard et al., 1999; Yi and

Figure 1.
A conceptual framework
for examining the
mediating role of habit
strength on the
satisfaction-loyalty
relation
La, 2004) or the relative amount of purchases over time (Heskett, 2002). This is also the Consumer
common practice among market players, who consider that an action of repeated loyalty
purchase of the same product signals consumer loyalty (e.g. Ball et al., 2004).
modelling
2.4 Habit strength as routine
Habits are most easily envisaged as simple routine behaviours, such as eating a meal
between five and six o’clock every workday. Triandis (1980) defines habit as 307
situation-behaviour sequences that are, or have become, automatic, so that they occur
without self-instruction. The most recent definitions of habit associate habit with a
repeated behaviour that has gained automatic qualities and that is performed under
relatively stable conditions with a minimal focus of attention (Ouellette and Wood, 1998).
For instance, Verplanken and Aarts (1999) describe habits as learned sequences of acts
that have become automatic responses to specific cues, and are functional in obtaining
certain goals or end states. Overall, it is inferred that, although repetition is a necessary
condition for habits to develop, the defining feature of a habit is automaticity.
A major problem in earlier studies of habitual behaviour is that habit is usually
measured as behavioural frequency, such as: “How many times during the past
year/month/week did you buy/use/eat . . .?” (Ouellette and Wood, 1998). However, as
pointed out, repeated behaviour need not necessarily involve habituation (Ajzen, 2002).
For instance, how many tries does it take to form a habit? To use the frequency of past
behaviour as a proxy for habit strength, and to test its effect on future behaviour, is to
use circular logic. If habit strength is defined as the tendency of past behaviour to exert
a unique affect on subsequent behaviour, then one cannot empirically demonstrate
situations in which someone has a strong habit to act and resist (Jaccard and Blanton,
2005). As the concept of habit is strongly rooted in behaviourist approaches to learning
theory, most studies of the development of habits take a mechanistic perspective of
habitual behaviour (Verplanken and Orbell, 2003). This approach makes no references
to internal psychological states or mental processes in explaining, defining or
operationalizing the construct. Likewise, most studies show any discriminant validity
of habit measures with a scale measuring behavioural frequency as loyalty behaviour.
In effect, the two constructs overlap greatly in the marketing literature.
In an attempt to address these limitations and to improve the content validity of the
habit construct, some researchers recommend using a cognitive measure of habit
strength. Verplanken and Orbell (2003) present a generic 12-item measure, namely the
Self-Report Habit Index (SRHI). This scale, or facets of this scale, have recently been used
to assess different individuals’ habits, including eating habits (Honkanen et al., 2005).
Although the scale includes several facets, such as frequency, routine, automaticity and
identity, here we focus only on the part covering routine and automaticity. These aspects
are in accordance with the formal definitions of habit, discussed earlier in this study, and
are accepted as an important criterion for habit strength (Ajzen, 2002). In addition, we
include an item covering repetition over time in accordance with our discussion of the
meaning of habit as “repeated behaviour” (Ajzen, 2002) and with the repetition and
practising of a skill in a given context (Ouellette and Wood, 1998).

2.5 Oliver’s model and the proposed model


The marketing literature offers relatively few theoretical models explaining
consumers’ loyalty behaviour. The most widely cited model is undoubtedly Oliver’s
EJM (1999) model. Oliver proposes a temporal approach to loyalty behaviour formation and
47,1/2 argues that consumers go through different phases from cognitive loyalty and affective
loyalty (satisfaction) through loyalty intention (commitment) before being committed
to loyalty behaviour. Hence, loyalty intention acts as mediator between satisfaction
and loyalty behaviour. Different empirical studies on consumer loyalty are
conceptually built on Oliver’s loyalty model (Chiou and Droge, 2006; Evanschitzky
308 and Wunderlich, 2006; Han et al., 2008).
A limitation of Oliver’s model is that it does not explain consumers’ loyalty
behaviour once loyalty behaviour has been installed and for frequently performed
behaviours in stable environments. The model implicitly entails every single initiation
of behaviour being directly preceded by deliberative intentions regarding the course of
action. However, many actions are simply resumptions of activities that individuals
started some time before, and forming the underlying intention is therefore
unnecessary in most situations (Wilson et al., 2000). Consistent with modern
psychological theories (Wilson et al., 2000), the attitude – behaviour link may not
necessarily depend on planned, explicit attitudes. Rather, individuals may act based on
the automatic, implicit attitudes that are stored in their memory. Automatic behaviour
occurs when the attitudes towards the behaviour are easily accessed and retrieved
from the memory and evidence exists that attitudes are highly accessible under
habitual behaviours (Verplanken and Aarts, 1999). In an attempt to fill this gap, the
present study proposes habit strength as a mediating variable between satisfaction
and loyalty behaviour. We further present more arguments on how habits are
developed though repeated performance and positive affect (satisfaction) in stable
contexts.
Triandis’s (1980) theory of interpersonal behaviour states that behaviour is a
function of intentions, facilitating conditions and habit strength. While novel
behaviours are primarily determined by intentions, repeated behaviours are primarily
determined by habit. Habits form slowly in stable contexts where the behaviour is
frequently practised (daily to several times a week) (Ouellette and Wood, 1998).
According to Wood and Neal (2009), habit formation is originally an
intention-dependent process where goals provide the initial outcome-oriented
impetus for response repetition. During the process of habit formation, individuals
gradually learn the associations between the behaviour and the characteristics of the
context in which the behaviour is performed (Wood and Neal, 2009). Once habits are
formed, perception of contexts governs the associated behaviour without awareness of
the underlying goal-directed intentions (Verplanken and Wood, 2006). Recognizing
that habits may originate in goal pursuit (intention-dependent), some scholars argue
that habits are a form of goal-dependent automaticity which emerge when context
activates a goal (Verplanken and Aarts, 1999), whereas other researchers argue that
habit context-response associations are completely independent of goals (Wood and
Neal, 2009).
Habits also may arise when cues in the performance context are associated with
positive affect (i.e. satisfied repetitions) over time (Triandis, 1980). In this case, contexts
trigger habit formation as they signal opportunities to obtain the rewarding experience
(Wood and Neal, 2009). Research on the neurotransmitter systems in the brain has
shown that the context reliably associated with positive response outcome can promote
habit performance (Wood and Neal, 2009). Additional evidence from the marketing
literature has shown a significant correlation between affect or satisfaction and Consumer
perceived habit strength in the area of recycling behaviour (Knussen et al., 2004), fruit loyalty
consumption behaviour (De Bruijn et al., 2007) and information-systems-related
behaviour (Limayem et al., 2007). Similarly, in the present study, we position modelling
cumulative satisfaction as a driver of habit strength based on the consumer’s positive
experience with the consumption of the product.
In summary, Oliver’s model (Model 1, Figure 1) relies primarily on the deliberate or 309
goal-directedness intentions mechanism to explain customer loyalty behaviour (as
measured by the frequency of behaviour). The extended model (Model 2, Figure 1)
proposes a mediating effect of habit (habit strength) in the relationship between
satisfaction and loyalty behaviour. Consistent with previous arguments, as the habits
develop and gain strength through satisfactory execution of the behaviour, the loyalty
behaviour is less likely to depend on a rational statement (loyalty intentions) and
mainly depends on the automaticity of habit.

3. Methodology
3.1 Focus on product category loyalty
Traditionally scholars have associated consumer loyalty exclusively with brands
(Chiou and Droge, 2006; Evanschitzky and Wunderlich, 2006; Oliver, 1999). Recent
studies (Nijssen et al., 2003; Pritchard et al., 1999) have associated loyalty also with
product categories, services, stores, persons or geographic region. Evidence suggests
that consumers’ decision-making and loyalty may form at different levels of the
product hierarchy from product categories to brands. For example, when deciding
what to have for dinner, consumers choose among multiple food categories, classes, or
types of products before reaching decisions among brands. In addition, some product
categories lack strong brands (e.g. fresh fruits, unprocessed meat and fresh seafood are
mostly sold unbranded). Product category loyalty is similar to brand loyalty because
the same rules that govern brand competition may also be applied to product category
competition. For example, fish competes with meat and butter competes with
margarine as much as Coca-Cola competes with Pepsi and other soft drink brands.
This study deals with consumer loyalty towards a product category (Chad et al., 2005)
and in particular consumers’ loyalty towards the consumption of fish as the main meal
protein in an in-home context.

3.2 Sample
Two representative household samples from Denmark (DK) (n ¼ 1110) and Spain (SP)
(n ¼ 953) form the basis of the present analysis. The two-sample analysis was aimed to
provide conceptual and statistical support for the investigated approach and to verify
whether specific parameters of the models are (in)variant across the countries. The
countries were selected based on expected differences in fish consumption. Spain ranks
as one of the highest consumers of fish in Europe (approximately 42 kg/capita/year cf.
FAOSTAT, 2006). Denmark in turn presents a relatively low level of total fish
consumption in Europe (approximately 24 kg/capita/year cf. FAOSTAT, 2006).
The fieldwork and pre-testing were sub-contracted to local market research
agencies. Households were selected randomly. In each household, the person mainly
responsible for food shopping and cooking was elected as the respondent. This
procedure influenced the distribution of gender in favour of females (77 per cent of the
EJM total number of respondents). Otherwise, the samples were representative of the
47,1/2 different countries, in terms of basic socio-demographics such as age, education, town
size and region.

3.3 Measures
Pilot interviews were conducted in order to pre-test the questionnaire. Only small
310 revisions were necessary. The respondents were clearly informed that this study
focused on fish as a product category, and not on shellfish or other seafood products.
All the questions were framed with regard to consuming fish as the main meal/dinner
in their home environment, in accordance with the principle of compatibility (Ajzen,
1991) and with the notion that habit is performed in relatively stable contexts (Ouellette
and Wood, 1998).
Satisfaction was measured by three seven-point semantic-differential items: “When
eating fish for the main meal/dinner at home: I feel bad-good, unsatisfied-satisfied,
unpleasant-pleasant.” This scale is frequently used to assess consumer satisfaction
(Limayem et al., 2007; Seiders et al., 2005).
Intended loyalty was assessed with three indicators: “Please indicate how likely it is
that you expect/plan/want to eat fish at home for your main course/dinner in the near
future?” These items are frequently used to measure intention as a reflective construct
within the expectancy-value theory (Armitage and Conner, 2001), and have been
adapted to the loyalty literature in different ways (Evanschitzky and Wunderlich,
2006). The respondents were given the opportunity to mark their intention on a scale
from one “very unlikely” to seven “very likely”.
Habit strength was measured by three items from the self-report index of habit
strength scale (Verplanken and Orbell, 2003). We reduced the number of items because
some items in the original scale covered other dimensions, such as frequency and
identity. Thus, our scale of habit strength focused on measures to assess lack of
awareness/consciousness, and was measured by three seven-point
semantic-differential items: “Eating fish as a main course/dinner at home is
something: I have been doing for a long time (a routine); I have no need to think about
doing; I do without thinking too much (automatically).”
Finally, loyalty behaviour was measured by one item, “How often do you eat fish as
a main course/dinner at home?”, which was to be answered on a nine-point frequency
scale ranging from “never” to “every day or almost every day”. Self-reporting
behavioural frequency in cross-sectional surveys serves as a good proxy for actual
consumption (Pritchard et al., 1999; Yi and La, 2004).

4. Analysis and results


4.1 Analytical procedures
The analyses include an examination of the reliability and validity of the constructs by
estimating a confirmatory factor analysis (CFA) model individually for each model and
each sample. Next, we apply structural equation modelling to test the proposed Model
1 and Model 2 for each individual sample. The traditional chi-square fit test is reported;
however, because it has been recognised as an inappropriate test for a large sample size
(Byrne, 2001), six other indices, considered to be robust indicators of model fit, are
included: the Comparative Fit Index (CFI), Normed Fit Index (NFI), Relative Fit Index
(RFI), Incremental Fit Index (IFI), Tucker –Lewis Index (TLI) and Root Mean Square
Error of Approximation (RMSEA). An acceptable model fit is indicated by CFI, NFI, Consumer
RFI, IFI and TLI values exceeding 0.90 and RMSEA values below 0.08 (Byrne, 2001). loyalty
4.2 Construct reliability and validity modelling
CFA was estimated as a preliminary step to confirm the factor structure and to provide
an initial test of reliability and validity of the factors (Anderson and Gerbing, 1988). We
ran a CFA for each of the two models (Model 1 and Model 2). However, for space 311
reasons, only the results for Model 2 (including all the constructs) are shown in Table I.
On the basis of the model fit indices the CFA model appeared to fit the data adequately.
Without exception, every individual item loading on the predicted constructs was
highly significant with values ranging from 0.71 to 0.98, hereby providing evidence of
the items’ convergent validity (Anderson and Gerbing, 1988). The reliability
coefficients (Cronbach alpha, composite reliability and variance extracted) were
satisfactory, ranging from 0.72 to 0.95. The intercorrelations between the constructs
are displayed in Table II. All the correlations appeared to be significant and below 0.70,
attesting to discriminant validity. Furthermore, to confirm the discriminant validity,

Country Denmarka Spainb


Construct/item Loading a CR VE Loading a CR VE

Satisfaction 0.847 0.907 0.767 0.918 0.948 0.859


Feel bad/Feel good 0.81 0.90
Unsatisfied/Satisfied 0.94 0.90
Unpleasant/Pleasant 0.91 0.85
Loyalty intentions 0.935 0.958 0.885 0.977 0.985 0.957
Plan 0.93 0.98
Expect 0.96 0.98
Want (desire) 0.85 0.95
Habit strength 0.809 0.886 0.723 0.814 0.889 0.729
I have been doing for a long
time 0.75 0.71
I have no need to think about
doing 0.84 0.74
I do without thinking
(automatically) 0.78 0.87
Loyalty behaviour
How often did you eat fish at
home? 1.00 (fixed) 1.00 (fixed)
Model fit statistics
x2 175.24 141.81
df 30 30
p-value 0.000 0.000
CFI 0.98 0.99
NFI 0.98 0.98
TLI 0.96 0.97
RMSEA 0.06 0.06
Table I.
Notes: an ¼ 1,110; bn ¼ 953; a ¼ Alpha Cronbach; CR ¼ Composite Reliability; VE ¼ Variance Results of the CFA,
extracted country-specific analysis
EJM the procedure recommended by Bagozzi et al. (1991) was followed. A series of
47,1/2 one-factor and two-factor confirmatory models was run for each pair of constructs in
the study, and a chi-square difference test was then conducted. The results revealed
that, for all the pairs of constructs, the two-factor solution was better than the
one-factor solution (p , 0:01). Overall, the data show that the measures of the proposed
constructs achieve high reliability and sufficient convergent and discriminant validity
312 across the two individual CFA models.

4.3 Common method bias


As common method bias may have confounding effects on the observed relationships
between the predictor and the criterion variables, particularly when data are
self-reported (a single source), we estimated next the single-common-method factor
approach, devised by Podsakoff et al. (2003), to check whether a common method bias
is present. The measurement model was estimated with a single-method first-order
factor added to the indicators of the four variables. It should be noted that the
measurement properties (fit indices), under the common method factor model, slightly
improved in comparison with the basic model (CFI ¼ 0:996 vs 0.980; NFI ¼ 0:996 vs
0.979; RMSEA ¼ 0:053 vs 0.061). However, the correlation estimates between the
variables remained almost unchanged between the two models. On the basis of this
analysis, common method bias appeared not to be an issue in this research.

4.4 Estimation of the proposed models


The following step in the analysis was to estimate the two proposed structural models.
We carried out the estimation following a hierarchical process. First, consistent with
Oliver’s model, we tested Model 1 to determine the effect of satisfaction and the
mediating role of loyalty intention on loyalty behaviour. Then we introduced the
variable habit strength as a mediator in the relationship between satisfaction and
loyalty behaviour (Model 2). On the other hand, to verify the mediating effect of the
proposed variables, i.e. loyalty intention and habit strength, a number of conditions
must hold (Baron and Kenny, 1986):
.
Satisfaction should have a significant effect on loyalty intentions and habit
strength.
.
Loyalty intentions and habit strength should have a significant effect on loyalty
behaviour (these first two conditions are examined in Model A).

Denmarka Spainb
Variable/country S Li H Lb S Li H Lb

Mean 5.82 4.31 4.59 5.16 5.85 4.22 5.41 6.75


Std dev. (1.34) (2.30) (1.81) (1.82) (1.25) (2.10) (1.40) (1.29)
Satisfaction (S) – –
Table II. Loyalty intentions (Li) 0.39 – 0.39 –
Construct estimated Habit strength (H) 0.49 0.36 – 0.65 0.39 –
means, standard Loyalty behaviour (Lb) 0.44 0.46 0.54 – 0.43 0.31 0.56 –
deviations and
a b
correlations Notes: n ¼ 1,110; n ¼ 953; Correlation coefficients are given in the non-diagonal elements
.
Satisfaction should have a significant direct effect on loyalty behaviour (Model B). Consumer
.
The direct effect of satisfaction on loyalty behaviour should become smaller in loyalty
absolute value when the path between the mediators (loyalty intentions and modelling
habit strength) and loyalty behaviour is opened (Model C).

4.4.1 Loyalty intentions as a mediator towards loyalty behaviour. Table III shows that
the conditions are present for the existence of a clear mediating effect of loyalty 313
intentions (Oliver, 1999). According to Model A, the relationships’ satisfaction – loyalty
intentions and, respectively, loyalty intentions-loyalty behaviour are significant for
both samples. The direct effects of satisfaction on loyalty behaviour are significant in
Model B, while they decrease in Model C, where the relation between loyalty intentions
and loyalty behaviour is admitted. An additional Sobel test (Sobel, 1982) confirms the
significance of the mediation effect of loyalty intentions on loyalty behaviour (Model 1
confirmed).
4.4.2 Habit strength as an alternative route towards loyalty behaviour. In discussing
the habit-strength construct, we provided a theoretical basis for positioning habit
strength as a mediating variable similar to the loyalty intentions in the traditional
cognitive-based hierarchy. Table III (Model 2) reports the results of the estimation of
the three models after introducing habit strength as an alternative mediating variable
between satisfaction and loyalty behaviour. First, Model D, the model reflecting the
first condition in the mediating role of habit strength, was tested. Looking at the results
of the estimation, it can be confirmed that satisfaction is significantly correlated with
habit strength, and in the expected direction, while habit strength itself has a positive
and significant effect on the loyalty behaviour. Satisfaction acts indirectly on loyalty
behaviour through habit strength. It appears that the first condition for habit strength
to exert a mediating effect in the loyalty process is fulfilled. Model E presents a
significant chi-square statistic and goodness-of-fit indicators that are slightly worse
than those of Model D. Moreover, since the difference between the chi-square values of
the two models (Dx 2(0)DK ¼ 61.21 and Dx 2(0) SP ¼ 98.98) is significant and Model D’s
explanation is clearly better, we deduce that the mediating-effect model (Model D:
satisfaction-habit strength-loyalty behaviour) is superior to the direct-effect one (Model
E: satisfaction-loyalty behaviour). Next, Model F (combining the effect of satisfaction
with the mediating effect of loyalty intentions and habit strength) was compared with
Model D. Of the two models, Model F presents the best goodness-of-fit indicators and
the differences in the chi-square statistic are significant (Dx 2(1)DK ¼ 29.38 and
Dx 2(1)SP ¼ 4.14). By transitivity, if Model D is better than Model E and Model F is
better than Model D, than Model F is better than Model E. Hence Model F is the model
that best fits the data, and hence the one that best represents the loyalty behaviour
process.
We note that the effect of habit strength on loyalty behaviour is greater that the
effect of loyalty intentions on loyalty behaviour in each of the two samples. This
finding is especially notable given that the countries examined differ considerably on
the fish consumption and habit dimensions. On the other hand, in Spain, the inclusion
of the mediating effect of habit strength almost cancels out the direct effect of
satisfaction on loyalty behaviour (b ¼ 0:09; t ¼ 1:96). In Denmark, in turn, satisfaction
affects the loyalty behaviour through loyalty intentions and habit strength (mediation),
although it continues to have a significant direct effect as well.
EJM

314

Model 2
47,1/2

Table III.
Results of Model 1 and
Model 1 Model 2
Relationships Model A Model B Model C Model D Model E Model F

Denmark a
S-Li 0.39 (12.44) 0.40 (12.82) 0.39 (12.29) 0.41 (12.89) 0.41 (12.86) 0.40 (12.81)
Li – Lb 0.47 (16.72) – 0.35 (11.88) 0.30 (11.24) 0.34 (11.42) 0.27 (9.30)
S- Lb – 0.45 (14.86) 0.30 (9.94) – 0.32 (10.49) 0.18 (5.24)
S –H – – – 0.43 (13.84) 0.51 (14.40) 0.50 (14.25)
H– Lb – – – 0.52 (14.75) – 0.33 (9.75)
Model fit:
x 2 (df) 162.53 (13) 197.26 (13) 64.03 (12) 297.98 (32) 359.19 (32) 268.60 (31)
CFI 0.97 0.97 0.99 0.96 0.96 0.97
NFI 0.97 0.97 0.99 0.96 0.96 0.97
TLI 0.94 0.93 0.98 0.94 0.93 0.94
RMSEA 0.10 0.11 0.06 0.09 0.10 0.07
R2 Li 0.15 0.16 0.15 0.17 0.16 0.16
R2 H – – – 0.27 0.26 0.25
R2 Lb 0.22 0.20 0.29 0.37 0.30 0.41
Spain b
S-Li 0.41 (12.52) 0.41 (12.66) 0.40 (12.46) 0.41 (12.79) 0.41 (12.80) 0.41 (12.79)
Li – Lb 0.30 (9.41) – 0.15 (4.44) 0.11 (3.83) 0.13 (4.04) 0.10 (3.13)
S- Lb – 0.43 (13.56) 0.36 (10.64) – 0.39 (11.42) 0.09 (1.96)
S –H – – – 0.66 (17.00) 0.66 (17.16) 0.65 (16.84)
H– Lb – – – 0.52 (14.32) – 0.46 (9.95)
Model fit:
x 2 (df) 145.13 (13) 55.17 (13) 35.82 (12) 106.11 (32) 205.09 (32) 101.97 (31)
CFI 0.98 0.99 1.00 0.99 0.98 0.99
NFI 0.98 0.99 0.99 0.99 0.98 0.99
TLI 0.96 0.99 0.99 0.98 0.96 0.98
RMSEA 0.10 0.06 0.05 0.05 0.08 0.05
R2 Li 0.16 0.17 0.16 0.17 0.17 0.17
R2 H – – – 0.43 0.44 0.43
R2 Lb 0.09 0.18 0.20 0.33 0.21 0.39
Notes: an ¼ 1,110; n ¼ 953; Satisfaction (S), Loyalty intentions (Li), Habit strength (H), Loyalty behaviour (Lb)
Finally, in line with Byrne (2001) and Steenkamp and Baumgartner (1998), we Consumer
compared the value of coefficients between countries in a multi-group cross-country loyalty
analysis[1]. A constrained model, in which the structural relationships were fixed
between country samples, was first estimated. Next, the constrained structural model modelling
was compared with an unconstrained model in which all the relationships were set free
between the two samples. The differences in the chi-square values between the models
determine whether the variable country acts as a moderating variable; that is, a 315
significant decrease in the chi-square from the constrained model to the unconstrained
model implies that the country variable has a significant influence on the structural
relationship (Byrne, 2001).
The results (Table IV) show significant differences between the two countries. For
the relationships that were moderated, the paths from loyalty intention to loyalty
behaviour and from satisfaction to loyalty behaviour were consistently higher for
Denmark than for Spain. Specifically, the coefficient from loyalty intentions to loyalty
behaviour was more than twice as strong for the Danish consumers than for the
Spanish ones (0.26, p ¼ 0:000 versus 0.10, p ¼ 0:01). As regards the path between
satisfaction and loyalty behaviour, it was significantly weaker for the Spanish sample
(0.09, p ¼ 0:05). In turn, the link from habit strength to loyalty behaviour was
consistently higher for Spain (0.48, p ¼ 0:000) than for Denmark (0.35, p ¼ 0:000).

5. Discussion
Loyalty behaviour has guided marketing research for many years. Oliver’s (1999)
model proposes a model of loyalty behaviour formation that infers that individuals’
intentions are a full mediator between satisfaction and loyalty behaviour. However, the
question of the present work is whether the course of action associated with consumer
loyalty is always planned and mediated by loyalty intentions or whether an alternative
route exists between satisfaction and loyalty.
We have introduced and tested an extended model in an attempt to delineate two
distinct phenomena explaining loyalty behaviour. Although variants of these
approaches exist in the literature, there is a lack of studies that integrate the two
phenomena and discuss the implications of this extended model (Limayem et al., 2007).
Taking Oliver’s model as a starting point, individuals’ settled intentions mediate the
satisfaction-loyalty behaviour in the initial stage of product adoption and in later
behaviour maintenance associated with unstable contexts (Ajzen, 2002). However, in

Multi-group
Relationship/country Denmarka Spainb comparisonsc

Satisfaction-Loyalty intentions (b1) 0.41 (23.40) * * 0.43 (23.40) * * DK * * ¼ SP * *


Loyalty intentions-Loyalty behaviour (b2) 0.26 (12.33) * * 0.10 (3.55) * DK * * . SP * *
Satisfaction-Loyalty behaviour (b3) 0.17 (7.04) * * 0.07 (1.62)ns DK * * . SPns
Satisfaction-Habit strength (b4) 0.49 (26.13) * * 0.67 (26.13) * * DK * ¼ SP * *
Habit strength-Loyalty behaviour (b5) 0.35 (11.19) * * 0.48 (10.48) * * DK * * , SP * *
Notes: a n ¼ 1110; bn ¼ 953; cComparisons across countries were made based on unstandardised Table IV.
coefficients; nsNot significant; *Significant at p ¼ 0:01; * * Significant at p , 0:001; Model fit: Results of Model 2
x2 ¼ 577:25; df ¼ 114; p ¼ 0:000; CFI ¼ 0:98; NFI ¼ 0:98; RFI ¼ 0:96; IFI ¼ 0:98; TLI ¼ 0:97; cross-country
RMSEA ¼ 0:04; Standardised regression coefficients are reported; t-values in parentheses comparisons
EJM later behaviour performed frequently in stable contexts, consumer behaviour is
47,1/2 automatic and less guided by plans and deliberations, which are represented by the
formation of intentions. Once habit starts to develop and gain strength through
satisfactory execution of the behaviour, the satisfaction-action loyalty sequence may
occur automatically, through habit strength and thus less guided by behavioural
intention.
316 The results from our study in a food consumption in-home context corroborate
previous arguments. Habit strength has proved to be a relevant mediator in the two
samples analysed, regardless of their differences in habits of fish consumption. The
only difference across countries was registered in the magnitude of the coefficient
estimators. For more habituated individuals in fish consumption (Spanish fish
consumers who are among the heaviest fish consumers in Europe), behavioural
intention was a weaker antecedent of loyalty behaviour when compared with less
habituated individuals (Danish fish consumers whose fish consumption is only half the
one as observed in Spain). That is, consistent with the conceptual argument, Spanish
consumers, as the high fish consumption country, were more likely to undertake less
planned and more intuitive or inertia-based responses with regard to the buying or
consumption of the involved product category.
The inclusion of a cognitive measure of habit strength in Oliver’s model led to a
notable increase in the explanatory power of the proposed model from 0.29 to 0.41 in
Denmark, and from 0.20 to 0.39 in Spain, respectively. Based on the Cohen f2 effect size
measure for hierarchical multiple regression, the habit strength had an effect size of
0.203 in Denmark and respectively 0.311 in Spain, which represents a medium effect in
both samples (Cohen and Cohen, 1983). The model in which habit is proposed to
mediate the link between cumulative satisfaction and loyalty behaviour possesses a
significantly higher explanatory power than the traditional model.
The overall findings of this research have shown that for the fish consumption
behaviour in an in-home context, habit strength can have a stronger effect on loyalty
behaviour than loyalty intentions. Consumers’ intentions were found to be significant
but relatively weaker in mediating consumer satisfaction in each sample. The
individual country analysis produced a similar pattern of results, hereby providing
validation of the findings. Overall the current study provides evidence of the
appropriateness of a model of loyalty behaviour that distinguishes between intentions
and an automatic process outside conscious intentions.

6. Managerial implications
In view of this discussion and with regard to studying the role of habits in
understanding and managing consumer loyalty in the context of food marketing, the
results suggest several implications. The dominant approach of consumer loyalty
behaviour in a business-to-consumer context has been to see loyalty behaviour as a
planned, conscious process based on consumer intentions, deliberation and
commitment towards the product category or brand (Oliver, 1999). While we do not
inquire into the mediating effect of intention on loyalty behaviour, we argue that in
some circumstances (such as later behaviour performed in stable environments and
behaviours characterised by frequent purchases, relatively little involvement and
intense competition due to the availability of many substitutes across product category
– such as food products), this effect is partially or almost entirely suppressed by
habituation (Limayem et al., 2007). The fact that intention is not the only antecedent of Consumer
behaviour loyalty confirms the previous findings by Mittal and Kamakura (2001) and loyalty
others showing poor correlation between intention and behaviour loyalty. Therefore,
this study advocates that scholars evaluating food consumption behaviour and modelling
practitioners in food marketing and other fast-moving consumer goods should
conceive consumers’ loyalty behaviour as an action governed by two distinctive
phenomena: intentions when referring to loyalty behaviour formation and habits when 317
evaluating loyalty behaviour persistence. Intention-loyalty consumers act from their
intention, which is a product of cognitive deliberation, planning and commitment,
while habit-loyalty consumers act from their habit in a less-planned, more inertial or
automatic mode.
In a highly competitive market with low-differentiated products of comparable
quality and satisfying similar needs or goals, businesses are keenly interested in
strategies designed to strengthen consumer loyalty for an incumbent product or
product category, or to break consumers’ loyalty to competing products and product
categories. Based on this study, and considering different scenarios, specific strategies
for managing different types of product category loyalty through marketing activities
can be built, that focus either on influencing consumer motivation through intention
loyalty or by influencing their habits.
In general and consistent with the previous literature, managing consumer loyalty
towards low frequency-of-purchase product categories (such as, various categories of
consumer durables) requires the fostering and maintaining of attribute beliefs about
the product and product category as well as satisfactory experience with the product to
intention loyalty formation and further loyalty behaviour maintenance.
Intention-based loyalty formation can be accomplished by focusing on
information-based communications that encourage consumers to evaluate and
balance the benefits and costs (exclusive value) of the incumbent product and
commit consumers to choice. In this traditional scenario, consumer goal-directedness
intention mechanisms play a mediating role in consumer loyalty formation. By
contrast, for product categories characterised by high frequency-of-purchase (as in this
study), managing consumer loyalty requires additional strategies.
First, building consumer loyalty in the initial phases of product adoption for
frequently purchased products (such as various categories of food products) initially
requires the fostering of awareness and attribute beliefs about the product and product
category and satisfactory experience with the product. In that sense, actions have
usually focused on information-based communications about the product and
satisfactory episodes with the product in order to strengthen the beliefs and affect
(satisfaction) associated with the product consumption or usage (Webb and Sheeran,
2006). For instance, linked to our empirical case, the fostering of beliefs about the
health qualities of fish and continuous satisfaction with the fish products should be
positively correlated with individuals’ intentions to adopt the fish category in their
daily consumption. In this scenario, consumer intention plays a mediating role in
consumer loyalty formation.
Second, maintaining consumer loyalty for a product in circumstances characterised
by high purchase frequency (consumables) and intense competition can be obtained by
fostering habit-based loyalty based on stable contexts, cumulative satisfaction and
frequent performance. The habits of purchase or consumption that consumers form
EJM over time (“inertia”) are a natural phenomenon that happens to a product or product
47,1/2 category. Consumers fall into such a state of inertia because the cognitive effort
required while buying the product is minimal when compared with a novel alternative
that requires self-control resources. Thus, in a stable context and for frequent
behaviours marketing strategies should focus on inciting consumers to develop
quickly the habit of using the desired product (Limayem et al., 2007). Specific actions to
318 fostering habit formation for fish consumption may include cumulative satisfaction
and repeat purchase of the fish products in stable contexts in order to create
associations in the memory between the features of the environment and the product in
question (for instance, a fish stand positioned every Saturday at the supermarket
entrance in order to create a habit in a stable buying context, ready fish dinners to
create a consumption habit and coupons and bonuses to encourage frequent
purchasing). Marketers should pay more attention to actions focused on developing
consumers’ habit of purchasing and consuming the product, as consumers are often
acting automatically and are more likely to maintain their old behaviours (habits)
given the demands of everyday life such as time pressure, cognitive load and
regulatory depletion (Wood and Neal, 2009). Additionally, by locking-in consumers
into the habit of consuming an incumbent product category, businesses may prevent
loyalty behaviour formation for a competitive product category. Hence, the habit-based
strategy complements the intention-based strategy – in which fish marketers want to
prevent consumer loyalty for another product category (e.g. turkey meat) by
communicating the exclusive nutritive value of fish when compared to the meat
category and thus fostering loyalty formation based on logical thinking and
comparisons.
Finally, businesses may be interested in strategies designed to break consumer
habit-based loyalty for an incumbent product category characterised by high purchase
frequency (e.g. meat). While it can be beneficial to promote the benefits of the targeted
product category (fish), this is not always sufficient to offset consumers’ behaviour.
Habits are deeply embedded in consumers’ minds and less sensitive to (new)
information (Verplanken and Wood, 2006). Consequently, traditional
information-based interventions communicating the benefits of the fish product
category may be less effective in breaking the old habits. In order to influence the
choice of habitual-loyal consumers for a competitive product category (such as fish),
the companies should consider changing the context that activates the existing habits
for an incumbent product (i.e. meat) and thus automatically directing habit
performance (Wood and Neal, 2009). For example, changing the physical
surroundings in which consumers purchase habitually (e.g. product placement on
the shelf, store displays, introducing a similar packaging design) may change the
consumer choice of products. Changing the context breaks the purchasing habit by
suppressing the link between the context and the incumbent product in the memory
(Wood and Neal, 2009). An in-store experiment for instance showed that consumers are
significantly less likely to purchase potato chips if they are placed on top and bottom
shelves than if they are placed on the middle shelf (Sigurdsson et al., 2009). With regard
to the physical context, it is important to note that it might be extremely difficult to
change (particularly when retailers exert considerable influence on the way the
products are located on shelves and in stores). However, a closer co-operation between
producers and retailers (Fornari et al., 2009) may break consumers’ habits for specific
products and product categories. Persuasive interventions based on information about Consumer
the benefits of an alternative product would be less effective for habit-loyalty loyalty
consumers without breaking the context or the physical surrounding link (Verplanken
and Wood, 2006). This is because, for frequent behaviours, a significant portion of modelling
purchasers or consumers fall into such a state of inertia that the cognitive activity they
engage in while buying is minimal, and consumers do not consider the expected
benefits and costs of changing the current product for a rival alternative (Limayem 319
et al., 2007; Verplanken and Aarts, 1999). In support of this idea, a meta-analysis of
different interventions designed to change individuals’ behaviour through different
information-based strategies (persuasive communication, social encouragement, social
pressure, social support and information regarding behaviour and outcome) revealed
corresponding changes in behaviour only for intentional-based behaviour but not for
habitual-based behaviour (Webb and Sheeran, 2006).
In summary, the current results convey that approaches to identifying loyal
consumers or customers should focus on both intention and habit-dependent views of
loyalty. If understanding the nature of habit and differentiating between a controlled
and a less-controlled dimension of loyalty behaviour were to be implemented
successfully in business settings, this could be of value for preventing, strengthening
or breaking the old habits and loyalty to an incumbent product category or brand.

7. Limitations and further research


The present study faces some limitations. The study employs cross-sectional data;
therefore, causal effects can only be inferred. Although the authors based their
arguments on Jaccard and Blanton (2005), according to whom, for stable behaviours,
cross-sectional data can be as informative as longitudinal data, prospective
longitudinal studies are recommended to verify and validate the present findings.
Second, this study did not directly assess real action (factual behaviour); rather, the
authors built their hypotheses on individuals’ self-reported opinions about their past
behaviour. Optimistic biases might have affected participants’ responses, because of
the prevailing recommendations to increase fish consumption for its nutritional and
health benefits. Therefore, a future research agenda requires replications involving
new marketing objects (brands, stores, services), other product categories that have
more negative connotations and higher loyalty ratings (such as snacks, confectionery,
fast foods) or various categories of consumer durables. New studies also require a
combination of different methodologies (quantitative and qualitative) and more
objective methods of data collection (experiments and neuroimaging techniques) to
understand and capture consumers’ habits and loyalty behaviour in different contexts
(in-home vs out-of-home, own country vs abroad). Qualitative designs, such as focus
groups, in-depth interviews or observations at the point of purchase, or a combination
of these, could be an approachable challenge for future research. For example, it would
be beneficial to combine verbal protocols with direct observation in order to study
consumers’ decision-making process in stable and unstable buying and consumption
contexts, or for relative new and current incumbent products with the same level of
satisfaction. These qualitative techniques combined with more advanced
neuroimaging techniques on consumer sub-conscious action (Jonides, 2004) could
help to determine the real drivers of consumer loyalty behaviour.
EJM Note
47,1/2 1. The invariance of the measurement model (i.e. factor loadings) was initially assessed.
Although the chi-square difference between the unconstrained model and the constrained
model (fixed loadings across groups) was statistically significant (delta chi-square ¼ 74:7;
p , 0:01), the TLI indicated a negligible change in fit (delta TLI , 0:001). The metric
invariance between the two groups was therefore accepted (Byrne, 2001). The constrained
320 model was used in the subsequent structural invariance analysis (Steenkamp and
Baumgartner, 1998).

References
Ajzen, I. (1991), “The theory of planned behaviour”, Organizational Behaviour and Human
Decision Processes, Vol. 50 No. 2, pp. 179-211.
Ajzen, I. (2002), “Perceived behavioural control, self-efficacy, locus of control, and the theory of
planned behaviour”, Journal of Applied Social Psychology, Vol. 32 No. 4, pp. 665-83.
Anderson, J.C. and Gerbing, D.W. (1988), “Structural equation modeling in practice: a review and
recommended two-step approach”, Psychological Bulletin, Vol. 103 No. 3, pp. 411-23.
Armitage, C.J. and Conner, M. (2001), “Efficacy of the theory of planned behaviour:
a meta-analytic review”, British Journal of Social Psychology, Vol. 40 No. 4, pp. 471-99.
Bagozzi, R.P., Yi, Y. and Phillips, L.W. (1991), “Assessing construct validity in organizational
research”, Administrative Science Quarterly, Vol. 36 No. 3, pp. 421-58.
Ball, D., Coelho, P.S. and Machas, A. (2004), “The role of communication and trust in explaining
customer loyalty: an extension to the ECSI model”, European Journal of Marketing, Vol. 38
Nos 9/10, pp. 1272-93.
Baron, R.M. and Kenny, D.A. (1986), “The moderator-mediator variable distinction in social
psychological research: conceptual, strategic and statistical considerations”, Journal of
Personality and Social Psychology, Vol. 51 No. 6, pp. 1173-82.
Byrne, B.M. (2001), Structural Equation Modeling with AMOS – Basic Concepts, Applications,
and Programming, Lawrence Erlbaum Associates, Mahwah, NJ.
Chad, W., MacDonald, S., Lingras, P. and Adams, A. (2005), “Relationship between product based
loyalty and clustering based on supermarket visits and spending patterns”, International
Journal of Computer Science & Applications, Vol. 2 No. 2, pp. 85-100.
Chiou, J.-S. and Droge, C. (2006), “Service quality, trust, specific asset investment and expertise:
direct versus indirect effects in a satisfaction-loyalty framework”, Journal of the Academy
of Marketing Science, Vol. 34 No. 4, pp. 1-15.
Cohen, J. and Cohen, P. (1983), Applied Multiple Regression/Correlation Analysis for the
Behavioural Sciences, Lawrence Erlbaum Associates, Hillsdale, NJ.
De Bruijn, G.-J., Kremers, S.P.J., De Vet, E., De Nooijer, J., Van Mechelen, V. and Brug, J. (2007),
“Does habit strength moderate the intention-behaviour relationship in the theory of
planned behaviour? The case of fruit consumption”, Psychology and Health, Vol. 22 No. 8,
pp. 899-916.
Dick, A.S. and Basu, K. (1994), “Customer loyalty: toward an integrated conceptual framework”,
Journal of Academy of Marketing Science, Vol. 22 No. 2, pp. 99-113.
Eagly, A. and Chaiken, S. (1993), The Psychology of Attitudes, Harcourt Brace Jovanovich, Fort
Worth, TX.
Evanschitzky, H. and Wunderlich, M. (2006), “An examination of moderator effects in the
four-stage loyalty model”, Journal of Service Research, Vol. 8 No. 4, pp. 1-16.
FAOSTAT (2006), Food and Agriculture Organization statistical database, United Nations, Consumer
available at: http://faostat.fao.org/default.aspx (accessed 5 November 2009).
loyalty
Fornari, D., Grandi, S. and Fornari, E. (2009), “The role of management of product innovation in
retailer assortment: evidence from the Italian FMCG market”, The International Review of modelling
Retail, Distribution and Consumer Research, Vol. 19 No. 1, pp. 29-43.
Gustafsson, A., Johnson, M.D. and Roos, I. (2005), “The effects of customer satisfaction,
relationship commitment dimensions and triggers on customer retention”, Journal of 321
Marketing, Vol. 69 No. 4, pp. 210-8.
Han, X., Kwortnik, R.J. Jr and Wang, C. (2008), “Service loyalty: an integrative model and
examination across service contexts”, Journal of Service Research, Vol. 11 No. 1, pp. 22-42.
Heskett, J.L. (2002), “Beyond customer loyalty”, Managing Service Quality, Vol. 12 No. 6, pp. 355-7.
Honkanen, P., Olsen, S.O. and Verplanken, B. (2005), “Intention to consume seafood: the
importance of habit strength”, Appetite, Vol. 45 No. 2, pp. 161-8.
Jaccard, J. and Blanton, H. (2005), “The origins and structure of behaviour: conceptualizing
behaviour in attitude research”, in Albarracı́n, D., Johnson, B.T. and Zanna, M.P. (Eds),
Handbook of Attitudes and Attitude Change, Lawrence Erlbaum Associates, Mahwah, NJ,
pp. 125-72.
Johnson, M.D., Anderson, E.W. and Fornell, C. (1995), “Rational and adaptive performance
expectations in a customer satisfaction framework”, Journal of Consumer Research, Vol. 21
No. 4, pp. 128-40.
Jolley, B., Mizerski, R. and Olaru, D. (2006), “How habit and satisfaction affects player retention
for online gambling”, Journal of Business Research, Vol. 59 No. 6, pp. 770-7.
Jonides, J. (2004), “How does practice makes perfect?”, Nature Neuroscience, Vol. 7, pp. 75-9.
Knussen, C., Yule, F., MacKenzie, J. and Wells, M. (2004), “An analysis of intentions to recycle
household waste: the roles of past behaviour, perceived habit, and perceived lack of
facilities”, Journal of Environmental Psychology, Vol. 24 No. 2, pp. 237-46.
Limayem, M., Hirt, S.G. and Cheung, C.M.K. (2007), “How habit limits the predictive power of
intention: the case of information systems continuance”, MIS Quarterly, Vol. 31 No. 4,
pp. 705-37.
Lindblad, E. and Lyttkens, C.H. (2002), “Habits versus choice: the process of decision-making in
health related behaviour”, Social Science & Medicine, Vol. 55, pp. 451-65.
Mittal, V. and Kamakura, W.A. (2001), “Satisfaction, repurchase intent, and repurchase behavior:
investigating the moderating effect of customer characteristics”, Journal of Marketing
Research, Vol. 38 No. 2, pp. 131-42.
Neal, D.T., Wood, W. and Quinn, J.M. (2006), “Habits: a repeat performance”, Current Directions
in Psychological Science, Vol. 15 No. 4, pp. 198-202.
Nijssen, E., Singh, J., Sirdeshmukh, D. and Holzmueller, H. (2003), “Investigating industry context
effects in consumer-firm relationships: preliminary results from a dispositional approach”,
Journal of the Academy of Marketing Science, Vol. 31 No. 1, pp. 46-60.
Oliver, R.L. (1999), “Whence customer loyalty?”, Journal of Marketing, Vol. 63, pp. 33-44.
Ouellette, J. and Wood, W. (1998), “Habit and intention in everyday life: the multiple processes by
which past behaviour predicts future behaviour”, Psychological Bulletin, Vol. 124 No. 1,
pp. 54-74.
Podsakoff, P.M., MacKenzie, S.B., Lee, J.-Y.Y. and Podsakoff, N.P. (2003), “Common method
biases in behavioural research: a critical review of the literature and recommended
remedies”, Journal of Applied Psychology, Vol. 88 No. 5, pp. 879-903.
EJM Pritchard, M.P., Havitz, M.E. and Howard, D.R. (1999), “Analyzing the commitment-loyalty link
in service context”, Journal of Academy of Marketing Science, Vol. 27 No. 3, pp. 333-48.
47,1/2
Reinartz, W., Thomas, J.S. and Kumar, V. (2005), “Balancing acquisition and retention resources
to maximize customer profitability”, Journal of Marketing, Vol. 69 No. 1, pp. 63-79.
Seetharaman, P.B. (2004), “Modeling multiple sources of state dependence in random utility
models: a distributed lag approach”, Marketing Science, Vol. 23 No. 2, pp. 263-71.
322 Seiders, K., Voss, G.B., Grewal, D. and Godfrey, A.L. (2005), “Do satisfied customers buy more?
Examining moderating influences in a retailing context”, Journal of Marketing, Vol. 69
No. 4, pp. 26-43.
Sigurdsson, V., Saevarsson, H. and Foxall, G. (2009), “Brand placement and consumer choice:
an in-store experiment”, Journal of Applied Behavior Analysis, Vol. 42, pp. 741-5.
Sobel, M.E. (1982), “Asymptotic confidence intervals for indirect effects in structural equations
models”, in Leinhart, S. (Ed.), Sociological Methodology, San Francisco, CA, pp. 290-312.
Steenkamp, J.-B.E.M. and Baumgartner, H. (1998), “Assessing measurement invariance in
cross-national consumer research”, Journal of Consumer Research, Vol. 25 No. 1, pp. 78-90.
Trafimow, D. (2000), “Habit as both a direct cause of intention to use a condom and as a
moderator of the attitude-intention and subjective norm-intention relations”, Psychology
and Health, No. 15, pp. 383-93.
Triandis, H.C. (1980), “Values, attitudes and interpersonal behaviour”, in Howe, H.E. and Page,
M.M. (Eds), Nebraska Symposium on Motivation: Beliefs, Attitudes and Values, Vol. 27,
University of Nebraska Press, Lincoln, NE, pp. 195-259.
Verplanken, B. and Aarts, H. (1999), “Habit, attitude, and planned behaviour: is habit an empty
construct or an interesting case of automaticity?”, European Review of Social Psychology,
Vol. 10, pp. 101-34.
Verplanken, B. and Orbell, S. (2003), “Reflections on past behaviour: a self-report index of habit
strength”, Journal of Applied Social Psychology, Vol. 33 No. 6, pp. 1313-30.
Verplanken, B. and Wood, W. (2006), “Interventions to break and create consumer habits”,
Journal of Public Policy and Marketing, Vol. 25 No. 1, pp. 90-103.
Webb, T.L. and Sheeran, P. (2006), “Does changing behavioral intentions engender behavior
change? A meta-analysis of the experimental evidence”, Psychological Bulletin, Vol. 132
No. 2, pp. 249-68.
Wilson, T.D., Lindsey, S. and Schooler, T.Y. (2000), “A model of dual attitudes”, Psychological
Review, Vol. 107 No. 1, pp. 101-26.
Wood, W. and Neal, D.T. (2009), “The habitual consumer”, Journal of Consumer Psychology,
Vol. 19, pp. 579-92.
Worthington, S., Russell-Bennett, R. and Härtel, C. (2010), “A tri-dimensional approach for
auditing brand loyalty”, Journal of Brand Management, Vol. 17, pp. 243-53.
Yi, Y. and La, S. (2004), “What influences the relationship between customer satisfaction and
repurchase intention? Investigating the effects of adjusted expectations and customer
loyalty”, Psychology and Marketing, Vol. 21 No. 5, pp. 351-73.

About the authors


Professor Svein Ottar Olsen has carried out research in consumer behaviour and marketing
research with focus on consumer psychology and consumption behaviour. Professor Olsen has
published articles in leading marketing and international business journals in the area of
perceived quality, consumer satisfaction and loyalty, survey methodology in international
research, and use of marketing information. Svein Ottar Olsen is the corresponding author and Consumer
can be contacted at: sveinoo@nfh.uit.no
Dr Ana Alina Tudoran has carried out research and published articles in international and loyalty
national business journals in the area of brand equity, perceived quality, consumer satisfaction modelling
and the role of information on consumer behaviour. Dr Tudoran’s current research interests are
in the area of behavioural economics and quantitative methods.
Dr Karen Brunsø has carried out research on consumer behaviour and she is author of
numerous scientific publications on marketing and consumer behaviour in relation to food. In 323
particular Dr Brunsø has researched food-related lifestyles across Europe, and has worked with
the implementation of results in food companies.
Dr Wim Verbeke has carried out research on consumer attitudes, perceptions and acceptance
of agricultural and food production technologies and products. In particular, Dr Verbeke is
author of numerous scientific publications on the impact of information, food labelling and the
role of individual characteristics and individual difference variables on food consumption
decisions.

To purchase reprints of this article please e-mail: reprints@emeraldinsight.com


Or visit our web site for further details: www.emeraldinsight.com/reprints

View publication stats

You might also like