2. Name: Tiezheng Yuan
Student No: 110562836
2
Table of Contents
Section number and title Page
1. Introduction 3
2. Theory 4
3. Application to Marketing 5
4. Method 7
5. Results 8
6. Marketing Implications 10
7. Summary /Conclusions 13
List of References 14
Appendices
Appendix 1 SPSS Output 15
List of Tables
Table 1 Rotated Factor Matrix for Importance of Store Attributes 8
List of Figures
Figure 1 Segmentation Targeting Positioning framework 10
3. Name: Tiezheng Yuan
Student No: 110562836
3
1. Introduction
The aim of this study is to apply exploratory factor analysis to the data collected
which represents student attitude to the importance of supermarket features.
Exploratory factor analysis is a technique that identifies the underlying dimensions of
metric correlated data. Besides that it also achieves data reduction so that original set
of variables are replaced by smaller set of factors. In addition, exploratory factor
analysis can also be used to confirm the dimensionality of existing scales. In this
study, applying factor analysis to the student scale will reveal the dimensions
underlying the importance of supermarket features. Therefore, improving the
understanding of student food shoppers and it will help in the effectiveness of
marketing grocery shoppers.
This study is structured as follows, section two dues with the theory of exploratory
factor analysis. Next, section three will touch on application to marketing of
exploratory factor analysis through published article. Section 4 will describe the
method used to conduct factor analysis. Section 5 will present the result of the
analysis. Section six will mention the marketing implications of results. Lastly,
section 7 will be summary and conclusion of the study.
4. Name: Tiezheng Yuan
Student No: 110562836
4
2. Theory
This section aims to explain the theory of factor analysis by describing the objectives
of factor analysis, the data requirement, the full equation of factor analysis, some key
assumptions of theory and basic features of model.
Exploratory factor analysis is a technique whose objectives are to identify underlying
dimensions of an original large set of matric correlated variables with the aim of
minimum information loss. The factors derived are uncorrelated and jointly explains
the total variance of the original variables in descending order of importance.
The two broad aims of factor analysis are to:
1. Identify the number of factors
2. Interpret the meaning of the factors.
In order to conduct factor analysis, the original data are required to be metric and
correlated.
The factor analysis full equation is:
xp = bp1f1 + bp2f2 + ⊠+ bpkfk + ep
The equation assumes that each of p original variables (xâs) are determined by a linear
combination of k non-observable common factors (fâs) and the influence of a non-
observable unique factor (e).
There are several assumptions that form the basis of the equation:
1. The xâs are standardised to have a mean of zero and a variance of unity
2. The common factors are standardised to have a mean of zero and a variance of
unity
3. The covariance between common factors are zero so that the correlations between
common factors are zero
4. The covariance between common and unique factors are zero so that they are not
correlated
5. The covariance between unique factors is zero so that pairs of unique factors are
not correlated
The basic feature of the model focuses on three sets of relationships:
1. How their values are determined
2. How their variances are determined
3. How their covariance are determined
5. Name: Tiezheng Yuan
Student No: 110562836
5
3. Application to Marketing
This section aims to explain and evaluate the practical application of factor analysis
through reviewing the study conducted by Gayatri and Chew (2013). It will explain
the aim of the study, provide description of data and measures, explain the result and
interpretation, mention the value of study and lastly critic the study.
The study conducted by Gayatri and Chew (2013) aim to investigate the service
quality on Muslim (Islamic follower) customer perception and behaviour.
The measure consists of a 42-item scale designed to measure service quality in the
context of Islamic culture. A seven-point Likert scale with anchors of 1 - strongly
disagree to 7 - strongly agree were used.
Six distinct factors were identified through the study. The interpretation of factors is
established through the strength of correlation between factor and the original scale
items. Factor 1 is most strongly associated, in descending order of importance, with
âRespect for Muslim Customersâ (.790), âReligious toleranceâ (.770), âAccommodate
the needs of Muslim customerâ (.740) and âUnderstanding Islamic rulesâ (.720). It is
therefore interpreted as Islamic values.
Factor 2 is most strongly associated, in descending order of importance, with âDisplay
a certificate of Halal/Haramâ (.840), âHolds a Halal certificate for product/serviceâ
(.830), âStatement of Halalâ (.630), âCategorized product/service that are Halal/Haram
separatelyâ (.600), âDeclare products/services according to Islamic rule of
Halal/Haramâ (.560) and âFollow Islamic rule of Halal/Haramâ (.540). It is therefore
interpreted as Halal/Haram.
Factor 3 is most strongly associated, in descending order of importance, with âPlace
for saying prayerâ (.920), âPurifying facilitiesâ (.910), âProvide Sajadah/Mukena
praying toolsâ (.890), âProvide direction of Meccaâ (.890), âMaintain sanctity of place
for prayingâ (.890), âProvide proper praying facilitiesâ (.850) and âImportant of
Islamic religious activitiesâ (.790). It is therefore interpreted as Attention to Islamic
religious activity.
Factor 4 is most strongly associated, in descending order of importance, with âNo
fraudulent in business dealingâ (.800), âNo uncertainty in business transactionâ (.790),
âHonest in business dealingsâ (.780), âNo tamper with measurement scalesâ (.750),
âNot only engage in profiteeringâ (.680), âDeliver service according to promiseâ (.650)
and âNo support for gamblingâ (.610). It is therefore interpreted as Honesty.
6. Name: Tiezheng Yuan
Student No: 110562836
6
Factor 5 is most strongly associated, in descending order of importance, with âModest
outfitâ (.830), âUniform that promotes modesty for its staffâ (.780) and âFollow
Islamic dress codeâ (.660). It is therefore interpreted as Modesty.
Factor 6 is most strongly associated, in descending order of importance, with
âHumane treatment for its customersâ (.780), âHumane touch during transactionâ
(.780), âService using humane standardâ (.720), âTreat employees with humanenessâ
(.720), âShow trust when serving customerâ (.580), â Develop trustworthy standardâ
(.530) and âGive high trust to its customers during service transactionâ (.530). It is
therefore interpreted as Humaneness and trustworthiness.
The result of the study contributes to the extant literature by developing a service
quality measure that is relevant to the Islamic context. Besides that, through the study,
the researchers discovered that consumers with Muslim background are heavily
influenced by the factors identified which have been omitted in the previous
literatures of service quality measurement. In addition, the measurement model used
in the study is useful for future researchers who are interested in exploring other
markets and market segments in other Muslim countries. Lastly, through this study,
several practical recommendations were given to practitioners. For example,
restaurant owners were recommended to cover the windows during the fasting month
of Ramadan as high religious awareness improves customerâs perception of service
quality and potentially long-term loyalty.
There are some limitations of the study conducted by Gayatri and Chew (2013).
Firstly, the use of convenience sampling may reduce the quality of representation
compared to probability sampling techniques (Cassady, 1945). The haphazard
selection of subjects may introduce bias as researchers were given considerable
leeway to exercise their judgement concerning selection of respondents (Zikmund and
Babin, 2010). Therefore, objective statistical inferences are difficult to make when
non-probability sampling is used (Ngulube, 2005). Secondly, the study was conducted
in Indonesia, where there is a separation of state and religion and the majority of the
population is Sunni Muslim (Seibel and Agung, 2006). Thus, the applicability of the
Islamic service quality measure may be questionable in countries where Islam is the
state religion or in countries where majority of population are Shiâa Muslims.
The study is extended by conducting confirmatory factor analysis to evaluate the
model. Reliability was also conducted.
7. Name: Tiezheng Yuan
Student No: 110562836
7
4. Method
This section aims to explain and justify the method used by providing explanation of
data and measures, confirmation of data are metric and correlated and provide
explanation of method.
The data consist of 14-item scale designed to measure the importance of supermarket
features (1 = Not at all important, 5 = Very important).
The use of ordinal scale, with explicit scores and with equal intervals of unity
between descriptors, means that the scale is assumed to have interval measurement
properties and is thus metric.
In order to confirm that the data are correlated, Bartlettâs Test for Sphericity is
conducted based on the following hypotheses:
H0: None of the variables are correlated
H1: The Variables are correlated
Confirmation that the test variable are inter-correlated is indicated by a KMO index of
.697, categorised by Kaiser (1974) as âMiddlingâ , while Bartlettâs Test of Sphericity
results in the rejection of the null hypothesis at the five percent significance level
(Ï2
(91) = 1848.233, Sig = .000).
Factor analysis was applied to a 14-item scale designed to measure the importance of
supermarket features (1 = Not at all important, 5 = Very important). The analysis
using SPSS 21.0 (2008) employed principal components analysis with Varimax
rotation and the extraction criterion was to derive factors with eigenvalues greater
than unity. Goodness of fit was evaluated using total variance explained and
communalities. The minimum acceptable value for communalities was set at 0.5 (Hair
et al., 2006: p 149). Following Hair et al. (2006: p128) the cut-off point for the
inclusion of factor loadings consistent with a sample size of 731 was set as .30. The
analysis resulted in a solution of 5 factors. Factor scores were saved for subsequent
analysis (See Appendix 1).
8. Name: Tiezheng Yuan
Student No: 110562836
8
5. Results
This section aims to present the results in a logical structure. Firstly, table of results
will be presented. Secondly, it will define the measures of goodness of fit and report
and evaluate the goodness of fit of the results. Lastly, it will explain method of
interpretation and interpret the results.
Table 1 Rotated Factor Matrix for Importance of Store Attributes
Store Feature Factor Number h2
1 2 3 4 5
Convenient location .181 .156 .114 .069 -.769 .665
Parking facilities -.013 .123 .132 .161 .781 .668
Pleasant atmosphere .008 -.010 .594 .402 -.187 .550
Well-known brands .010 .173 .769 .020 .099 .632
Own label products .601 .044 .132 -.255 .211 .490
High quality products .047 .066 .734 .100 .037 .557
Value for money .767 .078 .098 .084 -.148 .634
Low prices .833 -.067 -.097 .077 -.123 .729
Special offers .757 .054 -.017 .095 -.090 .594
Friendly, helpful staff .103 -.026 .201 .805 -.114 .714
Check-out speed .021 .555 .043 .531 .074 .597
Methods of payment .077 .824 .162 .024 -.013 .711
Cash-back facilities .003 .846 .055 .006 -.027 .719
Other facilities -.020 .088 .087 .560 .291 .414
Eigenvalue 2.271 1.799 1.630 1.535 1.440
Variance % 16.219 12.849 11.640 10.962 10.286
Cumulative variance % 16.219 29.068 40/709 51.670 61.956
9. Name: Tiezheng Yuan
Student No: 110562836
9
Goodness of fit is evaluated from total variance explained and communalities. Total
variance explained is combined contribution to total variance of the set of all 5
derived factors. Total variance explained is 62%. This is regarded as acceptable for
social science data. Communality is the proportion of the variance of a specific
variable explained by all the derived factors. The communalities are generally
respectable apart from Own label products and Other facilities. In summary 4
communalities were strong, 4 were respectable, 4 were acceptable and 2 were weak.
In summary, goodness of fit was regarded as acceptable.
The interpretation of factors is established through the strength of correlations
between each factor and the original scale items. Factor 1 is most strongly associated,
in descending order of importance, with âLow pricesâ (.833), âValue for moneyâ
(.767) and âSpecial offersâ (.757). It is therefore interpreted as Price and Value.
Factor 2 is most strongly associated, in descending order of importance, with âCash-
back facilitiesâ (.846) and âMethods of paymentâ (.824). It is therefore interpreted as
Transaction methods.
Factor 3 is most strongly associated, in descending order of importance, with âWell-
known brandsâ (.769) and âHigh quality productsâ (.734). It is therefore interpreted as
Quality and Branding.
Factor 4 is most strongly associated with âFriendly, helpful staffâ (.805). It is therefore
interpreted as Employee attitude.
Factor 5 is most strongly associated, in descending order of importance, with âParking
facilitiesâ (.781) and âInconvenient locationâ (.769). The negative coefficient for
âConvenient locationâ (-.769) in Table 1 is interpreted as Factor 5 is positively
associated with âInconvenient locationâ. Therefore, Factor 5 is interpreted as
Accessibility.
10. Name: Tiezheng Yuan
Student No: 110562836
10
6. Marketing Implications of Results
This section aims to apply strategic and tactical marketing theory to the results
through the use of Segmentation, Targeting and Positioning (STP) framework.
However, this section will only briefly mention the strategic (STP) theory. It will
focus more on tactical (7ps) aspect of marketing theory. The 7ps are Product, Price,
Place, Promotion, People, Process and Physical evidence.
Figure 1 Segmentation Targeting Positioning framework
Strategic (STP)
Demographic segmentation (age) was used in the study as the research was conducted
at Newcastle University with full-time undergraduates (18-25 years old).
Differentiated targeting approach was adopted to specifically target the undergraduate
segment. Product and Brand positioning strategy would be specifically designed
based on undergraduateâs characteristics.
Tactical (7ps)
Product
Supermarkets could focus on ensuring high quality of products as quality is one
important dimension identified. Retailers could communicate clearly with its suppliers
regarding quality standards. Besides that, supermarkets could encourage suppliers to
adopt a quantitative (statistic software) approach to quality control. Customer
Strategic
âą Segmentation
Startegic
âą Targeting and Positioning
Tactical
âą Marketing Mix (7ps)
11. Name: Tiezheng Yuan
Student No: 110562836
11
feedback on quality could be conducted by supermarkets as a way to monitor quality
and satisfaction. In addition, internal (employee) feedback on quality could also be
encouraged. Maintaining high quality standard could be integrated into the
organisation culture.
Price
Price and value is the most important dimension identified by student shoppers
concerning supermarket features. Supermarkets could adopt a cost-plus pricing
strategy when targeting students. Besides that, supermarkets could adopt lean
production techniques such as Just-in-time (JIT) stock management approach.
Retailers only hold stocks that it need and therefore reducing storage cost. The cost
saved could be passed on to customers thus lowering product price to attract student
shoppers.
Place
Accessibility is one dimension identified by student shoppers. There could be parking
facilities nearby supermarkets. If the supermarket is located at an inconvenient
location, the availability of parking facilities gives students the opportunities to use
their cars.
Promotion
Promotional efforts could focus on price and brand as these two dimensions are
identified important to student shoppers. Retailers could step up promotional efforts
when there are special discounts. In addition, quality brand names could also be
included in promotions. Retailers could use both above and below the line promotion
methods. For example advertise through internet, radio and magazines.
People
Staff attitude is one important factor identified by student consumers. Organisations
could place more importance in its human resource management. It needs to recruit
people with the right values and attitudes. Adequate training could be provided to
equip employees with the necessary skills to provide quality service. Occasionally,
there could be refresher or upgrader courses to keep up with raising customer demand.
Employees could also be rewarded for showing consistent good attitude. Customer
feedback could be encouraged to reflect on staff attitude.
12. Name: Tiezheng Yuan
Student No: 110562836
12
Process
Retailers could introduce convenient transaction methods to improve customer
shopping process. It could introduce cash-back service at the tiles. Besides that,
retailers could provide variety of payment methods. For example customers could pay
by credit card, direct debit or cash.
Physical Evidence
Employees could wear a smiley badge as evidence to show commitment towards
friendly and helpful service. Posters showing commitment towards quality assurance
could be displayed.
13. Name: Tiezheng Yuan
Student No: 110562836
13
7. Summary
This section will provide a brief summary of research aim, research method, key
results and value of results. It will also evaluate the study and provide suggestions for
future research.
The aim of this study is to apply exploratory factor analysis and identify the
underlying dimensions concerning studentsâ attitude to the importance of supermarket
features. Factor analysis was applied to a 14-item scale designed to measure the
importance of supermarket features (1 = Not at all important, 5 = Very important).
Five factors were identified, namely, Price and Value, Transaction methods, Quality
and Branding, Employee attitude and Accessibility. The results generated contribute
to the understanding of student segment concerning supermarket features. It also
enables marketers to develop more specific marketing strategies targeting student
segment.
The study can be further improved by using probability sampling technique such as
stratified sampling to provide a better representation of the population. The use of
quota sampling (non-probability) technique in the study may not provide an unbiased
representation of population (Peterson and OâDell, 1950). As a result, objective
statistical inferences are difficult to make when non-probability sampling is used
(Ngulube, 2005).
In future research, the study could be conducted using mixed method approach. For
example, focus group could be conducted before factor analysis. Such a way, the data
gathered is triangulated and therefore improve the credibility and validity of result
(Homburg et al., 2012). In addition, more information could be gathered using mixed
method approach. Participants might be willing to provide more information in a
focus group compared to face-to-face survey as they feel more secure answering
questions in a group (Powell and Single, 1996). Therefore enable the researcher to
gather more information.
14. Name: Tiezheng Yuan
Student No: 110562836
14
List of References
Cassady Jr, R. (1945) 'Statistical sampling techniques and marketing research',
Journal of Marketing, 9(4), pp. 317-341.
Gayatri, Gita and Chew, Janet (2013) âHow do Muslim consumers perceive service
quality?â, Asia Pacific Journal of Marketing and Logistics, 25(3), pp. 472-490.
Hair, J.F., Black, W.C., Babin, B.J., Anderson, R.E. and Tatham, R.L. (2006)
Multivariate Data Analysis. Upper Saddle River, New Jersey: Pearson Education, Inc.
Homburg, C., Klarmann, M., Reimann, M. and Schilke, O. (2012) 'What drives key
informant accuracy?', Journal of Marketing Research (JMR), 49(4), pp. 594-608.
Kaiser, H. F. (1974) âAn index of factorial simplicityâ, Psychometrika, 39(1), pp. 31-
36.
Ngulube, P. (2005) 'Research procedures used by Master of Information Studies
students at the University of Natal in the period 1982â2002 with special reference to
their sampling techniques and survey response rates: A methodological discourse',
The International Information & Library Review, 37(2), pp. 127â143.
Peterson, P.G. and O'Dell, W.F. (1950) 'Selecting sampling methods in commercial
research', Journal of Marketing, 15(2), pp. 182-189.
Powell, R.A. and Single, H.M. (1996) 'Focus Groups', International Journal for
Quality in Health Care, 8(55), pp. 499-504.
Seibel, H.D. and Agung, W.D. (2006), Islamic microfinance in Indonesia, University
of Cologne Development Research Centre, Cologne.
SPSS (2008), SPSS for Windows (Version 21.0), Chicago, IL, USA: SPSS Inc.
Zikmund, W.G. and Babin, B.J. (2010) Exploring Market Research (10th
edition).
London: Cengage Learning.
15. Name: Tiezheng Yuan
Student No: 110562836
15
Appendix 1 SPSS Output
KMO and Bartlett's Test
Kaiser-Meyer-Olkin Measure of Sampling
Adequacy.
.697
Bartlett's Test of
Sphericity
Approx. Chi-Square 1848.233
df 91
Sig. .000
Communalities
Initial Extraction
Convenient Location 1.000 .665
Car-Parking Facilities 1.000 .668
Pleasant Shopping
Atmosphere
1.000 .550
Wide range of well known
brands
1.000 .632
Wide range of own-label
products
1.000 .490
High Quality Products 1.000 .557
Value for money 1.000 .634
Low Prices 1.000 .729
Special Offers 1.000 .594
Friendly, Helpful Staff 1.000 .714
Speed of Check-Out 1.000 .597
Method of Payment 1.000 .711
Cash-Back Facilities 1.000 .719
Other Facilities 1.000 .414
Extraction Method: Principal Component Analysis.