Int. Journal of Business Science and Applied Management, Volume 1, Issue 1, 2006
A Structural Equation Model of Customer Satisfaction and
Future Purchase of Mail-Order Speciality Food
Li-Wei Mai, University of Westminster
Westminster Business School,
University of Westminster, 35 Marylebone Rd, London NW1 5LS, UK,
Tel: +44 (0)20 7911 5000 Ext. 3003
Fax: +44 (0)20 7911 5839
Mitchell R. Ness, Newcastle University
School of Agriculture Food and Rural Development
Newcastle University, Newcastle upon Tyne, NE1 7RU, UK,
Tel: +44 (0)191 2226928
Fax: +44 (0)191 2226720
Analyses the relationship between satisfaction with mail-order speciality food attributes, overall
satisfaction, and likelihood of future purchase using a structural equation model. The results indicate
that customer satisfaction is associated with both service and product features of mail order speciality
Keywords: customer satisfaction, re-purchase intentions, speciality mail order food, structural equation
Int. Journal of Business Science and Applied Management /
In the UK the market for mail-order speciality foods has evolved in line with the growth of direct
marketing in response to changes in modern lifestyles and customer expectations. These changes have
paved the way for the exploitation of in-home shopping systems (The Economist, 1994; Victor, 1995).
The issues faced by speciality food marketers concern the identification and satisfaction of customers'
needs in the context of key elements of direct marketing such as targeting, interaction, continuity and
control with an emphasis on customer retention (McCorkell, 1997).
In this context direct marketing emphasises the importance of building a long-term relationship
with customers through quality, value and service. In this respect there is a shift in emphasis from pure
transactions-based marketing to give at least equal attention to relationship marketing. Thus in the face
of increasing competition firms pursue objectives of survival, prosperity and competitive advantage by
building products and customers through the delivery of high value (Kotler et al., 1996).
The consumer is assumed to have an objective of maximising value; the difference between the
benefits conferred in relation to the costs of acquisition, including costs of decision making, relative to
alternative products or suppliers. According to Kotler et al. (1996), maximisation of customer value is
translated into maximisation of customer satisfaction relative to expectations and product performance.
From the firm's perspective, delivery of value may be related to the notion of the value chain
(Porter, 1985). Porter emphasises that value is not just delivered by products but through all primary
activities (in-born logistics, operations, out-bound logistics, marketing and sales, and service) in
association with support activities (infrastructure, human resources, technological development and
The problem facing mail-order speciality food companies is consistent with this particular area of
research. Although, in what constitutes a relatively new sector, these firms face the traditional
transactions-based task of recruiting new customers, they also need to engage customers in a long-term
relationship; to encourage repeat business and loyalty through the delivery of value. Thus the focus of
this paper is to identify whether and to what extent such firms deliver satisfaction across the
transactions aspects of their operations and whether this creates a positive response in terms of overall
satisfaction and intentions to repeat purchases.
The structure of the paper is as follows. It begins with a discussion of the background literature
and is followed by an explanation and discussion of the research methodology. This is followed by the
presentation of the empirical results. Finally the paper closes with some summary and concluding
According to Dick and Basu (1994), consumer loyalty plays a central role in marketing strategy,
and marketing planning in the achievement of brand loyalty, vendor loyalty in industrial marketing,
service loyalty in the service sector and store loyalty in the retail sector. The marketing literature
defines loyalty as an attitude or a behavioural intention (Hallowell, 1996). Attitudinal loyalty reflects
an individual’s overall attachment to a product, service or organisation (Fornier, 1994). Behavioural
loyalty is expressed in terms of intentions to re-purchase, to increase the scale and scope of a
relationship, brand-switching or the act of recommendation (Yi, 1990; Selnes, 1993; Biong, 1993).
The widest perspective of behavioural loyalty is set within the context of the quality-value-
satisfaction (Q-V-S) literature. In a review of this literature Cronin et al. (2000) report that research
interest in Q-V-S has proceeded from a focus on perceived quality to satisfaction and hence to value
according to national awards or paradigm shifts such as total quality management, customer
satisfaction measurement, and customer value measurement. Research studies have paid attention to the
measurement of the constructs (Dabolkar et al., 2000), the relationships between them (Cronin et al.,
2000) and how they affect behavioural intentions (Bou-Llusar et al., 2001).
The elements within the Q-V-S framework tend to be defined as constructs with multiple
measures. Typically a construct of ‘Sacrifice’ is specified to depict the sacrifice in terms of price, time
and effort to accomplish the transaction. An independent construct of ‘Perceived Quality’ along with
‘Sacrifice’ is linked to a construct of ‘Perceived Value’. ‘Perceived Value’ thus reflects the influence of
the trade-off between ‘Sacrifice’ and ‘Perceived Quality’. Three alternative forms of measurement have
been applied to the perceived quality construct. One approach has been to address perceived quality
(Very poor, Very good). Another has been to employ disconfirmation measures that are related to
expectations (Much worse than expected, Much better than expected). A third approach employs
Li-Wei Mai and Mitchell R Ness
computed disconfirmation that employs measures of perceived quality and expectations and computes
disconfirmation by subtraction. A further issue is whether disconfirmation is more suitable measured in
a cross-section study or longitudinal study. In the latter approach expectations are measured prior to
service delivery while service quality is measured afterwards.
The need for firms to measure customer satisfaction has led to the use of instruments such as
customer satisfaction and purchase intentions surveys, analysis of complaints and suggestions, ghost
shopping and lost customer analysis. A study by Wilson (2002) researched the use of customer
satisfaction measurement within the retail sector. The research reveals a high degree of usage for
monitoring customer attitudes, the overall performance of the firm and to identify problem areas. Yet
more than two-thirds of firms indicated that satisfaction measures are most useful when combined with
complementary measures. Hausknecht (1990), in a review of methods of measuring customer
satisfaction/dissatisfaction, provides a taxonomy of measurement scales, which are classified as
evaluative or cognitive, emotional or affective and, behavioural or conative approaches. However,
Halstead (1989) makes the point that satisfaction is not desirable as an end but rather as a means to
understand future customer responses. Hence interest in satisfaction is linked to customer loyalty and
retention. However, satisfaction is regarded as a necessary but not a sufficient condition to lead to
repeat purchase behaviour (Van Looy et al., 1998; Bloemer and Kasper, 1995).
Satisfaction is typically measured as an overall feeling or as satisfaction with elements of the
transaction in terms of its ability to meet customers’ needs and expectations (Fornell, 1992; Zeithaml
and Bitner, 2000). Another approach employs a disconfirmation paradigm, which examines deviations
of performance from customer expectations and norms (Bearden et al., 1981).
A series of studies has further elaborated the satisfaction-loyalty relationship for products and
services, brands and retailers and considered the interaction between these (Bloemer and Lemmink,
1992). For example, Dabholkar and Thorpe (1994) employ multiple measures of both overall
satisfaction and loyalty. Bloemer and Kasper (1995) distinguish between spurious and true (brand)
loyalty and between manifest and latent satisfaction. They also provide explicit treatment of the
situation in which purchase takes place. In a study of customers of a car dealership Bloemer and
Lemmink (1992) distinguish between the satisfaction-loyalty relationship for both dealers and brands.
La Barbera and Mazursky (1983) employ a longitudinal study, which enables them to consider the
analysis of satisfaction over time, including brand-switching behaviour.
A further issue within the Q-V-S framework is the nature of the interactions between its
component constructs including direct and indirect links to behavioural intentions. In their review of
the applications Cronin et al. (2000) identify three broad approaches that reflect researchers’ interests
in different perspectives. The Value Model is typical of service value studies and specifies that
behavioural intentions are directly influenced by service value and where service value is influenced
independently by sacrifice, service quality and satisfaction or a subset of these constructs (Zeithaml,
1988; Cronin et al., 1997).
In the Satisfaction Model behavioural intentions are directly influenced by satisfaction and where
this construct is simultaneously influenced by service value and service quality, and where service
value is simultaneously influenced by sacrifice and service quality (Cronin and Taylor, 1992;
Hallowell, 1996; Oliver 1999).
A third model, the Indirect Model reflects a focus on the interaction between service quality, value
and satisfaction. Hence there are direct effects on behavioural intentions from value and satisfaction
constructs. Value is influenced directly by quality and in turn, has a direct effect on satisfaction. Hence
there are indirect effects on behavioural intentions by quality, via value and also via value and
satisfaction, and value, via satisfaction (Ennew and Binks, 1999).
Anderson et al. (1994) provide a framework for the estimation of the economic returns arising
from the delivery of consumer satisfaction. Evidence in support of the satisfaction-loyalty-profitability
relationship is provided by Heskett et al. (1994) and Hallowell (1996). Apart from the application to
individual firms, the concept has been extended, for example in the form of the American Consumer
Satisfaction Index (ACSI), to industrial sectors or economies (Fornell, 1992). Subsequently, the ACSI
inspired the development of the European Consumer Satisfaction Index (ECSI) in association with the
European Foundation for Quality Management (EFQM) and the European Academic Network for
Customer Oriented Quality Analysis (IFCF). In 1999 a pilot study was implemented in 12 European
countries (Kristensen et al., 2001; Cassel and Ekl
f, 2001)).
A related development extends the satisfaction-loyalty relationship to include profitability.
Loyalty enhances profitability through an increase in the scale and scope of the relationship with loyal
customers, lower customer recruitment costs, reduced customer price sensitivity and lower customer
servicing costs (Hallowell, 1996). However, Reinartz and Kumar (2002) warn against the assumption
that loyalty automatically promotes greater profitability. These authors test four assertions from the
Int. Journal of Business Science and Applied Management /
customer relationship marketing paradigm, that loyal customers: are more profitable, cost less to serve,
pay higher prices; and, act as word-of-mouth marketers. They established that the respective
associations between bivariate measures of loyalty with profit, costs, price and marketing activity were
generally ‘weak’ to ‘moderate’.
The authors attribute these results to the ‘crude’ nature of loyalty measurement that typically
employs recency-frequency-monetary value criteria. Alternatively, they propose the use of event
history modelling, which establishes the probability of purchase over future time periods and
subsequently segments customers into four categories (‘Buterflies’: short term loyalty/high
profitability; ‘Strangers’: short-term loyalty/low profitability; ‘True-friends’: long-term loyalty/high
profitability; and ‘Barnacles’: long-term loyalty/low profitability) according to profitability (high/low)
and loyalty (short-term/long-term). Hence they are able to propose management strategies for each
The research methodology employed a mail survey to investigate mail-order shoppers'
characteristics, attitudes, preferences and behaviour. It was implemented as a national (UK) survey of
(3,052) customer contacts supplied by five mail-order speciality food companies located in Cumbria,
Northumberland and the Scottish Borders region. The sample frame consisted of names and addresses
supplied by these companies. The sampling method employed a stratified random sample based upon
the relative sizes of the firms' contact lists. Subsequently the survey yielded 1,639 valid responses,
representing a response rate of 54 per cent.
It should be emphasised that the sample frame consisted of names and addresses of contacts,
comprising existing customers and potential customer contacts from various sources such as
exhibitions, trade fairs and from general enquiries. The firms did not have access to the type of
customer databases that are recognised in the direct marketing literature as a key aspect of direct
marketing and which would typically contain demographic, lifestyle and behavioural information to
provide for detailed analysis (McCorkell, 1997).
The respondents were classified as "Active" or "Non-active" mail-order customers according to
the recency of their last orders. The "Active" group comprised 1,030 respondents who had shopped for
speciality food using mail-order during the previous 12 months whilst the "Non-active" group
comprised 609 respondents who had not purchased food by mail-order during the same period. Whilst
the questionnaire design incorporated questions addressed to both groups, the research reported within
this study focuses only on the "Active" group, since this group had experience of speciality mail-order
products and were thus equipped to express their evaluations on the various scales. This approach
follows the notion that customer experience is an essential requirement in the analysis of satisfaction
assessments (Bolton and Drew, 1991; Cronin and Taylor, 1994; Parasuraman et al., 1988).
Although there are no directly comparable studies, this study is broadly consistent with existing
satisfaction literature (Cronin and Taylor, 1992; Anderson and Fornell, 1994; Hallowell, 1996; Bolton,
1998; Oliver, 1999; Garbarino and Johnson, 1999; Bolton and Lemon, 1999; Bernhardt et al., 2000).
Hence it takes a conventional approach in that re-purchase likelihood is directly influenced by a
measure of overall satisfaction that in turn is influenced by a construct to represent transaction
satisfaction associated with the attributes of mail order speciality foods. The measures of the attribute
satisfaction construct were identified from in-depth discussions with managers and proprietors of mail-
order firms and follow the approach of Biong (1993) in principle. The approach is parsimonious in that
in likelihood of purchase can be traced to actionable attributes associated with mail-order speciality
food and that are relevant to customers’ encounters with mail-order transactions within the speciality
food sector. The structural model is presented in Figure 1.
The eight effective indicators of the transaction satisfaction construct are concerned respectively
with enquiry service (satis1), product selection (satis2), product quality (satis3), price (satis4),
catalogue presentation (satis5), delivery service (satis6), ordering process (satis7) and payment terms
(satis8). The response measures are defined as overall satisfaction (ovsat) with mail-order and the
likelihood of future purchase (likbuy).
Li-Wei Mai and Mitchell R Ness
The levels of transaction satisfactions are measured as five-point scales (1 = Very satisfied, 5 =
Very dissatisfied) for each of the eight satisfaction variables. Overall satisfaction is measured as a
separate entity on the same basis. Likelihood of future purchase is measured on a 5 point scale (1 =
Definitely would buy, 5 = Definitely would not buy).
Figure 1: Structural Equation Model
Each of the measures are analysed in the first instance using univariate analysis of the 10 measures
in terms of frequencies and mean scores. Following this correlation analysis of the measures is
presented. Exploratory factor analysis is applied to the eight measures that comprise the transactions
satisfaction construct and the resulting factor structure is assessed using confirmatory factor analysis.
Finally a structural equation model to analyse the relationships between the construct and measures is
estimated. Univariate analysis, correlation analysis and exploratory factor analysis are conducted using
SPSS (2003) while confirmatory factor analysis and structural equation modelling are conducted using
AMOS (Amos, 2005).
Analysis of satisfaction attributes
The percentage distributions of responses for the eight satisfaction attributes are presented in
Table 1. From the perspective of the proportion of customers who are very satisfied mail-order
speciality food firms have been relatively more successful in delivering satisfaction with respect to
product quality, delivery service, enquiry service and selection of products available. They have been
least successful with respect to the order process, payment terms, catalogue presentation and price.
Price is the least satisfactory aspect with only 10 per cent of customers who are very satisfied.
However, this is not surprising, given the nature of these foods, which generally carry price premia, and
with respect to the high levels of satisfaction with quality that indicates that it is value for money which
customers evaluate.
Int. Journal of Business Science and Applied Management /
Table 1: Frequencies and Means for Satisfaction and Purchase Likelihood Measures
Satisfaction (Percentage response) Satisfaction
Satisfied Somewhat
Dissatisfied Very
Enquiry service 38.7 52.2 8.4 0.7 0.0 100 1.71
Product selection 37.6 55.1 6.9 0.4 0.0 100 1.70
Product quality 51.4 42.3 5.9 0.3 0.1 100 1.55
Price 10.0 52.8 29.4 7.2 0.6 100 2.36
Catalogue presentation 23.4 58.4 16.5 1.6 0.2 100 1.97
Delivery service 43.6 44.5 9.2 2.2 0.6 100 1.72
Ordering process 37.3 55.0 7.2 0.4 0.1 100 1.71
Payment terms 27.6 60.5 10.8 0.9 0.1 100 1.85
Overall satisfaction 36.5 56.3 6.6 0.3 0.3 100 1.72
Purchase likelihood (Percentage response)
Future purchase likelihood Definitely Likely Not sure Not likely Definitely
Total Mean
Likelihood rating: 58.5 35.0 4.7 1.7 0.1 100 1.50
Notes: Sample size (N) = 1028-1030 according to the presence of missing values in the data
Consideration of the two most favourable rating categories (Satisfied and Very satisfied) indicates
that for each item the major proportion of customers are at least satisfied with all eight aspects of
satisfaction. For example, 94 per cent of respondents are either satisfied or very satisfied with product
quality and 93 per cent are at least satisfied with product selection. The least satisfactory aspect is
associated with price where 63 per cent of respondents are at least satisfied but this still represents a
majority attitude. On the basis of mean scores it is product quality, product selection and the order
process which are more highly rated.
Analysis of overall satisfaction and likelihood of purchase
Following from this, consideration of overall satisfaction reveals that whilst a little more than one
third of mail-order speciality food customers are very satisfied, 93 per cent are either satisfied or very
satisfied (Table 1).
With respect to future purchase intentions, nearly 60 per cent of mail-order food shoppers
definitely intend to repeat their purchases of mail order speciality food in the future while 35 per cent
indicate they are likely to purchase again (Table 1). Consequently, the results indicate that mail-order
appears to deliver high levels of satisfaction to match or exceed customers' expectations and that a high
proportion of customers intend to repurchase.
Correlation analysis
The ultimate aim of the analysis is to examine the rationale of assuming that future purchase
intentions can be associated with overall satisfaction and in turn, that overall satisfaction is associated
with a satisfaction (with the features of mail-order) construct that is composed of eight indicators. Thus
as a preliminary step the simple correlation coefficients for these sets of variables are examined (Table
2). Statistical tests, based upon the null hypothesis that the population correlation coefficient is equal to
zero, indicate that all correlations are significantly different from zero at the 1 per cent significance
There is a significant positive association between overall satisfaction and purchase likelihood.
Furthermore, there are significant correlations between overall satisfaction and all eight satisfaction
attribute variables but, in descending order of magnitude, it is associated with the order process,
product quality, deliver service, payment terms, product selection, enquiry service, price and catalogue
presentation. The correlations between future purchase intentions and the satisfaction attributes are
generally weaker, though significant, but it is more strongly associated with product quality, enquiry
service, order process and price.
Li-Wei Mai and Mitchell R Ness
Table 2: Correlation Matrix
satis1 satis2 satis3 satis4 satis5 satis6 satis7 satis8 ovsat likbuy
satis1 1.000
satis2 0.480** 1.000
satis3 0.390** 0.581** 1.000
satis4 0.297** 0.366** 0.412** 1.000
satis5 0.363** 0.348** 0.257** 0.363** 1.000
satis6 0.318** 0.254** 0.317** 0.304** 0.340** 1.000
satis7 0.480** 0.356** 0.353** 0.297** 0.409** 0.588** 1.000
satis8 0.377** 0.266** 0.258** 0.374** 0.396** 0.456** 0.627** 1.000
ovsat 0.465** 0.466** 0.517** 0.437** 0.425** 0.500** 0.550** 0.492** 1.000
likbuy 0.309** 0.286** 0.376** 0.299** 0.235** 0.217** 0.298** 0.208** 0.452** 1.000
1. Variables are defined as follows:
satis1 = enquiry service; satis2 = product selection; satis3 = product quality; satis4 = price;
satis5 = catalogue presentation; satis6 = delivery service; satis7 = ordering process; satis8 = payment
ovsat = overall satisfaction; likbuy = likelihood of future purchase.
2. ** Indicates correlation is significant at the 0.01 level (two-tailed test)
3. Sample size (N) = 1028Exploratory Factor Analysis of the Satisfaction Construct
Exploratory Factor Analysis of the Satisfaction Construct
Factor analysis of the eight-item satisfaction construct employed the extraction procedure of
principal components with Varimax rotation. The criterion used to determine the number of factors was
based upon the derivation of factors with an eigenvalue greater than unity.
A two-factor solution was derived (Table 3). Bartlett’s test of sphericity lead to a rejection of the
null hypothesis, that the data are not correlated (
2 (28) = 2225.463, p < .001), while the KMO index of
0.827 is, according to Kaiser’s classification, ‘meritorious’ (Kaiser 1974). The two factors account for
59% of total variance and the communalities are generally respectable although those associated with
catalogue presentation (0.410) and price (0.411) is rather low. The first factor (sat1) is associated with
ordering process (0.819), payment terms (0.801), delivery service (0.762) and catalogue presentation
(0.537) and is defined as service satisfaction. The second factor (sat2) is associated with product
selection (0.833), product quality (0.827), price (0.553) and enquiry service (0.520) and is defined as
product satisfaction.
Table 3: Satisfaction Construct: Exploratory Factor Analysis
Factor Number Attribute
1 2
Enquiry service .442 .520 .466
Product selection .135 .833 .712
Product quality .124 .827 .699
Price .324 .553 .411
Catalogue presentation .537 .350 .410
Delivery service .762 .121 .596
Ordering process .819 .252 .734
Payment terms .801 .161 .667
Eigenvalue 2.514 2.180
Variance (%) 31.429 27.253
Cumulative variance (%) 31.249 58.682
h2 refers to communality
Sample size (N) = 1028
Confirmatory Factor Analysis of the Satisfaction Construct
The results of confirmatory factor analysis are presented in Table 4. The non-constrained
regression weights linking the sub-constructs or factors and their associated measures are all
statistically significant. The covariance between the two sub-constructs service satisfaction (sat1) and
product satisfaction (sat2) is positive and significant and is associated with a correlation of 0.674.
Int. Journal of Business Science and Applied Management /
Table 4: Satisfaction Construct: Confirmatory Factor Analysis
Weight Standard Error Critical Ratio
satis1 sat2 0.921 0.069 13.444 0.615 0.378
satis2 sat2 1.027 0.070 14.761 0.727 0.528
satis3 sat2 1.053 0.071 14.796 0.726 0.527
satis4 sat2 1.000 na na 0.554 0.307
satis5 sat1 0.786 0.053 14.924 0.529 0.279
satis6 sat1 1.113 0.058 19.034 0.673 0.453
satis7 sat1 1.146 0.052 22.164 0.855 0.730
satis8 sat1 1.000 na na 0.727 0.528
sat1 sat2
0.136 0.012 10.978 0.674 na
χ2(19) = 216.179, P = 0.000
Default 0.080 0.848 0.921 0.920
Saturated na na 1.000 1.000
Independence 0.205 0.000 0.000 0.000
1. Variables are defined as follows:
satis1 = enquiry service; satis2 = product selection; satis3 = product quality; satis4 = price;
satis5 = catalogue presentation; satis6 = delivery service; satis7 = ordering process; satis8 = payment
terms; sat1 = factor 1 (service satisfaction), sat2 = factor 2 (product satisfaction)
2. Sample size (N) = 1028.
Structural Equation Model
The structural equation model follows conventional linkages between satisfaction constructs,
overall satisfaction and likelihood of future purchase. The model employed maximum likelihood
estimation because of the presence of missing values in the data. Two versions of the model were
estimated. Version 1 revealed a significant covariance between service and product satisfaction
constructs of 0.136 so that Version 2 imposed a constraint between them. The estimated relationships
are presented in Table 5
Table 5: Estimated Structural Equation Model
Weight Standard Error Critical Ratio
ovsat sat1 1.605 0.108 14.904 0.891 0.794
ovsat sat2 1.000 na na 0.608 na
satis1 sat2 1.032 0.061 16.807 0.624 0.291
satis2 sat2 1.089 0.058 18.645 0.697 0.279
satis3 sat2 1.158 0.060 19.298 0.721 0.342
satis4 sat2 1.000 na na 0.517 0.254
satis5 sat1 0.788 0.049 15.984 0.543 0.239
satis6 sat1 1.111 0.054 20.745 0.688 0.322
satis7 sat1 1.110 0.045 24.932 0.845 0.393
satis8 sat1 1.000 na na 0.739 0.464
sat1 < -- > sat2 0.136 na na 0.719 na
likbuy ovsat 0.497 0.029 16.950 0.473 0.215
Goodness of Fit Measures
Absolute Incremental
Chi-Square Model RMSEA TLI IFI CFI
χ2 = 69.359
Def 0.069 0.877 0.922 0.922
df =17 Sat na na 1.000 1.000
P = 0.000 Indep 0.196 0.000 0.000 0.000
Li-Wei Mai and Mitchell R Ness
1. Variables are defined as follows:
satis1 = enquiry service; satis2 = product selection; satis3 = product quality; satis4 = price;
satis5 = catalogue presentation; satis6 = delivery service; satis7 = ordering process;
satis8 = payment terms; ovsat = overall satisfaction; likbuy = likelihood of future purchase;
sat1 = service satisfaction; sat2 = product satisfaction.
2. Sample size (N) = 1028
The goodness of fit measures presented in the table generally follow the recommendations of
Hoyle and Panter (1995) except for the inclusion of root mean square error of approximation
(RMSEA). The chi-square measure of discrepancy tests how much the implied and sample covariance
matrices differ under the null hypothesis that they do not. The results of the test indicate a rejection of
the null hypothesis, which does not auger well for the proposed model. However, Hu and Bentler
(1995: 78) suggest that the test may not be a good enough guide to model adequacy because of model
mis-specification, the power of the test, or violation of some technical assumptions underlying the test
(Hu and Bentler 1995: 77-8).
In the case of the model examined here it is highly likely that this is associated with the large
sample size (1028). Hair et al. (1998) report that if the sample size becomes large enough significant
differences will be found for any specified model. Hence they suggest that the test is not reliable
outside the sample range of 100-200 observations. However, it is also likely that the technical
assumptions underlying the test have been violated.
Given this situation, goodness of fit is also indicated by root mean square error of association
(RMSEA) that is recommended as an alternative to the chi-square test in the case of large samples
(Hair et al. 1998), and three type 2 indices suggested by Hoyle and Panter (1995); Tucker-Lewis index
(TLI), incremental fit index (IFI), and comparative fit index (CFI). RMSEA lies below the upper
threshold value of .080 regarded as ‘reasonable’ by Brown and Cudeck (1993) while values of TLI, IFI
and CFI approximate to the lower threshold of 0.9 suggested by Hair et al. (1998).
For the measurement models, the reliability of the indicators of the satisfaction sub-constructs is
conducted using the Reliability procedure within SPSS (SPSS, 2005). This yielded Cronbach alpha
coefficients of 0.768 for service satisfaction (sat1) and 0.723 for product satisfaction’ (sat2). Both
coefficients exceed the acceptable threshold level of 0.7 suggested by Nunally (1978). The respective
construct reliabilities (CR) of these sub-constructs are 0.801 and 0.737, which are above the threshold
level of 0.7 suggested by Hair et al. (1998). The respective variances extracted (VE) are 0.507 and
0.416 compared to a threshold value 0.5 suggested by the same authors, so that the VE for product
satisfaction is disappointing.
With respect to the structural model service satisfaction (sat1) and product satisfaction (sat2) have
a strong positive effect on overall satisfaction (ovsat) and explain 79 per cent of the variance of this
measure. The regression weight of product satisfaction (sat2) was constrained to unity to achieve
identification of the model but the coefficient of service satisfaction (sat1) is highly significant. Overall
satisfaction (ovsat) has a moderately positive effect on future purchase intentions (likbuy) but only
manages to explain 22 per cent of the variance of this dependent measure. The coefficient is also highly
significant. Examination of the standard coefficients reveals that service satisfaction (sat1) has a
stronger impact on overall satisfaction (ovsat) compared to product satisfaction (sat2) and that overall
satisfaction (ovsat) has a comparatively weaker impact on future purchase intentions (likbuy).
Examination of the relations between the two satisfaction constructs and their respective measures
reveals that all non-constrained coefficients are highly significant. For service satisfaction (sat1) the
relative importance of measure’s coefficients indicates that the highest associations in descending order
of importance are product quality, product selection, enquiry service, and price. For product
satisfaction (sat2) the highest associations are ranked in descending order as ordering processes,
payment terms, delivery service and, catalogue presentation.
The study set out to analyse customer satisfaction in terms of eight satisfaction attributes, a
measure of overall satisfaction and likelihood of future purchase.
The initial analysis considered analysis of individual scale items. The implications of this analysis
for mail-order speciality food businesses depend very much on the perspective that is taken. From the
perspective of the proportion of customers who are very satisfied there is concern because only in the
Int. Journal of Business Science and Applied Management /
case of product quality are a majority of customers very satisfied. Thus it appears that mail-order
speciality food firms should pay attention to all other aspects of mail-order operations.
On the other hand, if the criterion is to consider customers who are at least satisfied, satisfied or
very satisfied, there is more cause for an optimistic stance. Analysis of the eight-item satisfaction scale
reveals high levels of satisfaction with the eight attributes, especially with respect to product quality,
delivery service and enquiry service.
Thus it is not surprising that consideration of overall satisfaction reveals that 93 per cent of mail-
order speciality food shoppers are either satisfied or very satisfied. With respect to future purchase
intentions, nearly 60 per cent of mail-order food shoppers definitely intend to repeat their purchases in
the future. Consequently, the results indicate that mail-order appears to deliver high levels of
satisfaction that matches or exceeds customers' expectations and that a high proportion of customers
intend to purchase in the future.
Simple correlation analysis reveals that there are significant correlations between overall
satisfaction and likelihood of future purchase and that each of these variables are significantly
correlated with each of the eight items of the satisfaction attributes.
The results of the SEM show that it is possible to establish credible inter-relationships between the
sub-constructs of transaction satisfaction with mail order, overall satisfaction, and re-purchase
intentions. The measures of fit are acceptable and all free parameters are strongly significant.
Satisfaction with the transactions of mail order has a strong association with overall satisfaction.
However, higher levels of overall satisfaction have a weaker association with intentions to re-purchase.
The structural relations indicate that both product and service aspects of the mail-order transaction have
strong associations with overall satisfaction, but satisfaction with the service transaction is more
important. Consequently, the message to mail-order firms is that they need to expand their vision of
their respective businesses beyond that of a food delivered by post. After all, some speciality food
products are available in speciality stores but it is the high level of customer care and service that
differentiates the mail-order product from the in-store equivalent. The most important message is that
satisfaction and hence re-purchase likelihood are dependent on integrated features of both product and
service aspects of the mail order business.
The results are consistent with current emphasis on building customer satisfaction and loyalty,
which is envisaged as a goal and a key element in the achievement of company objectives, through
service aspects of the marketing mix. They are broadly compatible with the results of other studies
applied to other sectors that identify the positive link between satisfaction and response (Hallowell,
1996; Bolton, 1998; Oliver, 1999; Garbarino and Johnson, 1999; Bolton and Lemon, 1999; Bernhardt
et al., 2000). Hence the most important aspects of customer satisfaction revealed in this study can be
associated with those service marketing elements concerned with physical evidence (catalogue
presentation, product quality, product selection), process (order process, delivery service, payment
terms) and people (enquiry service).
The study reported here is regarded as an exploratory study which could be elaborated in further
work that focuses uniquely on the issue of satisfaction and loyalty. The model is set in the context of
the satisfaction literature rather than a more general Q-V-S framework. The model excluded constructs
of sacrifice, quality, and value and so assumes that these issues are incorporated in the transactions and
overall measures of satisfaction. Further extensions of this approach could employ multiple measures
of both overall satisfaction and behavioural intentions.
There would appear to be some justification for the exclusion of the perceived quality and
perceived value constructs. Some studies have questioned whether they are synonymous or distinct,
since they tend to be highly correlated such that some studies have failed to establish discriminant
validity between them (Dabholkar, 1993; Oliver, 1993). However it should be noted that Dabhollkar et
al. (2000) regard them as distinct but highly correlated.
According to Cronin et al. (2000), the omission of the constructs of sacrifice, quality and value
excludes the possibility of exploring the impact that these constructs have on behavioural intentions,
and more complex relations that emerge from both indirect and direct effects. Hence, a more ambitious
extension in further work could employ a more comprehensive framework to include the constructs of
sacrifice, perceived risk and perceived quality in addition to satisfaction.
Li-Wei Mai and Mitchell R Ness
Amos. 2005. Amos Version 6, (PA., USA: Amos Development Corporation)
Anderson, E. W., Fornell, C. 1994. A customer satisfaction research prospectus. In Rust, R. T., and
Oliver R. L. (Eds.), Service quality: New directions in theory and practice, 241-268 (Thousand
Oaks, CA, Sage Publications.
Anderson, E. W., Fornell, C., Lehmann, D. R. 1994. Customer satisfaction, market share, and
profitability: findings from Sweden. Journal of Marketing 58 (3), 53-66.
Beardon, W.O., Crockett, M., Teel, J. E. 1981. Disconfirmation of purchase expectations and
subsequent consumer attitudes and patronage intentions. Paper delivered at the American
Marketing Association conference on Retail Patronage, May 1981.
Bernhardt, K. L., Donthu, N., Kennett, P. A. 2000. A longitudinal analysis of satisfaction and
profitability. Journal of Business Research, 47, 161-171.
Biong, H. 1993. Satisfaction and loyalty to suppliers within the grocery trade. European Journal of
Marketing 27 (7), 21-38.
Bloemer, J. M. M., Kasper, H. D. P. 1995. The complex relationship between customer satisfaction and
brand loyalty. Journal of Economic Psychology, 16, 311-29.
Bloemer, J. M. M., Lemmink, G. A. M. 1992. The importance of customer satisfaction in explaining
brand and dealer loyalty. Journal of Marketing Management, 8, 351-364.
Bolton, R. N. 1998. A dynamic model of the duration of the customer’s relationship with a continuous
service provider: the role of satisfaction. Marketing Science 17 (1), 45-65.
Bolton, R. N., Drew, J. H. 1991. A longitudinal analysis of the impact of service changes on consumer
attitudes. Journal of Marketing, 55 (1), 1-9.
Bolton, R. N., Lemon, K. N. 1999. A dynamic model of customers’ usage of services: usage as an
antecedent and consequence of satisfaction. Journal of Marketing Research, 36 (May), 171-186.
Bou-Llusar, J. C., Camisón-Zornoza, C., Escrig-Tena, A. B. 2001. Measuring the relationship between
firm perceived quality and customer satisfaction and its influence on purchase intentions. Total
Quality Management, 12 (6), 719-734.
Browne, M. W., Cudeck, R. 1993. Alternative ways of assessing model fit. In K. A. Bollen, K. A.,
Long, J. S. (Eds.), Testing structural equation models, 136-162 (Newbury Park, CA, Sage
Cassel. C., Eklöf, J. A. 2001. Modelling customer satisfaction and loyalty on aggregate levels:
Experience from the ECSI pilot study. Total Quality Management, 12 (7-8), 834-841.
Cronin, J. J., Brady, M. K., Brand, R. R., Hightower Jr., R., Shenwell, D. J. 1997. A cross-sectional test
of the effect and conceptualization of service value. The Journal of Service Marketing, 11 (6),
Cronin, J. J., Brady, M. K., Hult, G. M. 2000. Assessing the effects of quality, value and customer
satisfaction on consumer behavioural intentions in service environments. Journal of Retailing, 76
(2), 193-218.
Cronin, J. J., Taylor, S. A. 1992. Measuring service quality: a re-examination and extension. Journal of
Marketing, 56 (3), 55-68.
Int. Journal of Business Science and Applied Management /
Cronin, J. J., Taylor, S. A. 1994. SERVPERF versus SERVQUAL: reconciling performance based and
perception based - minus - expectation measurements of service quality. Journal of Marketing, 58
(1), 125-31.
Dabholkar, P. A. 1993. Customer Satisfaction and service quality: Two constructs or one? In Cravens,
D. W. and Dickson, P. R. (Eds.), Enhancing knowledge development in marketing, 10-18
(Chicago, IL, American Marketing Association).
Dabolkar, P. A., Shepherd, C. D., Thorpe, D. I. 2000. A comprehensive framework for service quality:
an investigation of critical, conceptual and measurement issues though a longitudinal study.
Journal of Retailing, 76 (2), 139-173.
Dabholkar, P. A., Thorpe, D.I. 1994. Does customer satisfaction predict shopper intentions? Journal of
Consumer Satisfaction, 7, 161-71.
Dick, A. S., Basu, K. 1994. Customer loyalty: toward an integrated conceptual framework. Journal of
the Academy of Marketing Science, 22 (2), 99-113.
The Economist 1994. A survey of television-feeling for the future. February, 12-18.
Ennew, C. T., Binks, M. R. 1999. Impact of participative service relationships on quality, satisfaction
and retention: an exploratory study. Journal of Business Research, 46, 121-123.
Food from Britain 1987. Specialist foods study: market overview (Watford, UK, Institute of Grocery
Fornell, C. 1992. A national customer satisfaction barometer: the Swedish experience. Journal of
Marketing, 56 (1), 1-21.
Fornier, S. 1994. A consumer-based relationship framework for strategic brand management,
Unpublished PhD dissertation, University of Florida.
Garbarino, E., Johnson, M. S. 1999. The different roles of satisfaction, trust and commitment in
customer relationships. Journal of Marketing, 63 (April), 70-87.
Hair, J. R., Anderson, R. E., Tatham, R. L., Black, W. C. (1998). Multivariate data analysis (5th
edition) (New Jersey, Prentice Hall International Inc).
Hallowell, R. 1996. The relationships of customer satisfaction, customer loyalty, and profitability: an
empirical study. International Journal of Service Industry Management, 7 (1), 27-42.
Halstead, D. 1989. Expectations and disconfirmation beliefs as predictors of consumer satisfaction,
repurchase intention and complaining behaviour: an empirical study. Journal of Consumer
Satisfaction, Dissatisfaction and Complaining Behaviour, 5, 1-11.
Hausknecht, D. R. 1990. Measurement scales in consumer satisfaction/dissatisfaction. Journal of
Consumer Satisfaction, Dissatisfaction and Complaining Behaviour, 3, 1-11.
Heskett, J. I., Jones, T. O., Loveman, G. W., Sasser, W. E. Jr., Schlesinger, L. A. 1994. Putting the
service profit chain to work. Harvard Business Review, March-April, 105-11.
Hoyle, R. H., Panter, A. T. (1995). Writing about structural equation models. In Hoyle, R. H. (Ed.),
Structural equation modeling: Concepts issues and applications, 158-176 (London, Sage
Hu, Li-Tze., Bentler, P. M. 1995. Evaluating model fit. In Hoyle, R. H. (Ed.), Structural equation
modeling: Concepts issues and applications, 76-90 (London: Sage Publications).
Li-Wei Mai and Mitchell R Ness
Kaiser, H. F. 1974. An index of factorial simplicity. Psychometrika, 39, 31-36.
Kondo, Y. 2001. Customer satisfaction: how can I measure it? Total Quality Management, 12 (7-8),
Kotler, P., Armstrong, G., Saunders, S., Wong, V. 1996. Principles of marketing (European Edition)
(Hemel Hempstead: Prentice Hall).
Kristensen, K., Juhl, H. J., Østergaard, P. 2001. Customer satisfaction: Some results for European
retailing. Total Quality Management, 12 (7-8), 890-897.
La Barbara, P. A., Mazursky, D. 1983. A longitudinal assessment of consumer
satisfaction/dissatisfaction: the dynamic aspect of the cognitive process. Journal of Marketing
Research, November, 393-404.
McCorkell, G. 1997. Direct and data base marketing (London: Kogan Page).
Nunally, J. C. 1978. Psychometric Theory (2nd Edition) (New York: McGraw Hill).
Oliver, R. L. 1999. Whence consumer loyalty? Journal of Marketing, 63 Special Issue, 33-44.
Parasuraman, A. Zeithaml, V. A., Berry L. L. 1988. SERVQUAL: a multiple item scale for measuring
consumer perceptions of service quality. Journal of Retailing, 64 (1), 12-40.
Porter, M. 1985. Competitive advantage: creating and sustaining superior performance (New York:
Free Press).
Reinartz, W., Kumar, V. 2002. Mis-customer loyalty. Harvard Business Review, 80 (7) July, 86-94.
Selnes, F. 1993. An examination of the effect of product performance on brand reputation, satisfaction
and loyalty. European Journal of Marketing, 27 (9), 19-23
SPSS. 2003, SPSS for Windows Version 12, (Chicago, IL: SPSS Inc)
Van Looy, B., Gemmel, P., Desmet S., Van Dierdonck. R, Serneels S. 1998. Dealing with productivity
and quality indicators in a service environment: some field experiences. International journal of
service industry management 9 (4): 359-376.
Victor, P. 1995. Tele-shops to take 20 per cent of high street sales. Independent on Sunday, June 4th.
Wilson, A. 2002. Attitudes towards customer satisfaction measurement in the retail sector. International
Journal of Marketing Research, 44 (2), 213-222.
Yi, Y. 1990. A critical review of consumer satisfaction. In Zeithaml, V. (Ed.), Review of Marketing
1990, 213-222 (Chicago, IL: American Marketing Association).
Zeithaml, V. A. 1988. Consumer perceptions of price, quality and value: a means-end model and
synthesis of evidence. Journal of Marketing, 52 (July), 2-22.
Zeithaml, V. and Bitner, M. J. 2000. Services marketing (New York: McGraw-Hill).