Analisis Kepuasan Konsumen Berdasarkan Nilai Ekstrinsik dan Intrinsik.

ANALYSIS OF CUSTOMER SATISFACTION BASED ON
EXTRINSIC AND INTRINSIC VALUE

DEWI CAHYANI

DEPARTMENT OF STATISTICS
FACULTY OF MATHEMATICS AND NATURAL SCIENCE
BOGOR AGRICULTURAL UNIVERSITY
BOGOR
2015

PERNYATAAN MENGENAI SKRIPSI DAN
SUMBER INFORMASI SERTA PELIMPAHAN HAK CIPTA*
Dengan ini saya menyatakan bahwa skripsi berjudul Analysis of Customer
Satisfaction Based on Extrinsic and Intrinsic Value adalah benar karya saya dengan
arahan dari komisi pembimbing dan belum diajukan dalam bentuk apa pun kepada
perguruan tinggi mana pun. Sumber informasi yang berasal atau dikutip dari karya
yang diterbitkan maupun tidak diterbitkan dari penulis lain telah disebutkan dalam
teks dan dicantumkan dalam Daftar Pustaka di bagian akhir skripsi ini.
Dengan ini saya melimpahkan hak cipta dari karya tulis saya kepada Institut
Pertanian Bogor.

Bogor, August 2015
Dewi Cahyani
NIM G14110054

ABSTRACT
DEWI CAHYANI. Analysis of Customer Satisfaction Based on Extrinsic and
Intrinsic Value. Supervised by ASEP SAEFUDDIN and DIAN
KUSUMANINGRUM.
I Love Indonesia is one of the solutions given by the Indonesian Government
to improve the awareness of local products. The development of local product
quality does not only build the nation's economy, but also brings a good image of
Indonesia in international trade. To retain customers and survive in the industry,
every company should understand the customers by fulfilling their needs, both of
which can be assessed before purchase or after purchase so that customer
satisfaction can be maintained. This study analyzes the effects of extrinsic and
intrinsic value on customer satisfaction and loyalty of undergraduate students of
IPB on an Indonesian local bag company, XYZ with Structural Equation Modeling
(SEM). A total of 300 respondents was collected and analyzed. The results shows
the goodness fit indices are adequate. It concluded that both extrinsic and intrinsic
value influence customer satisfaction and they are not statistically different with a

coefficient 0.42 and 0.46, respectively. It is also known that customer satisfaction
influences customer loyalty with a coefficient 0.70. Beside that the most important
things for customer are the bag performance in extrinsic value and perceived quality
of product in intrinsic value.
Keywords: customer loyalty, customer satisfaction, extrinsic value, intrinsic value,
product quality, structural equation modeling

ABSTRAK
DEWI CAHYANI. Analisis Kepuasan Konsumen Berdasarkan Nilai Ekstrinsik dan
Intrinsik. Dibimbing oleh ASEP SAEFUDDIN dan DIAN KUSUMANINGRUM.
Aku Cinta Produk Indonesia adalah salah satu solusi yang diberikan
Pemerintah Indonesia dalam meningkatkan kesadaran akan produk dalam negeri.
Peningkatan kualitas produk dalam negeri tidak hanya membangun ekonomi
bangsa, namun juga membawa nama baik Indonesia di pasar internasional. Setiap
perusahaan sepatutnya memahami konsumen dengan memenuhi segala
kebutuhannya, baik kebutuhan yang dapat dinilai sebelum pembelian atau setelah
pembelian agar kepuasan konsumen dapat terjaga. Hal tersebut dilakukan untuk
mempertahankan konsumen dan tetap berada dalam industri. Penelitian ini
menganalisis pengaruh dari nilai ektrinsik dan intrinsik terhadap kepuasan dan
loyalitas mahasiswa S1 IPB pada perusahaan tas lokal Indonesia, XYZ dengan

menggunakan Pemodelan Persamaan Struktural (PPS). Total 300 responden telah
diambil dan dianalisis. Hasil menunjukkan bahwa indeks kebaikannya dapat
dikatakan cukup. Kesimpulan dari penelitian ini adalah nilai ektrinsik dan intrinsik
mempengaruhi kepuasan konsumen dan keduanya tidak berbeda nyata dengan
masing-masing koefisien sebesar 0.42 dan 0.46. Diketahui pula bahwa kepuasan
konsumen mempengaruhi loyalitas konsumen dengan koefisien sebesar 0.70.
Disamping itu, hal terpenting bagi konsumen adalah performa tas yang termasuk

dalam nilai ektrinsik dan persepsi kualitas produk yang termasuk dalam nilai
intrinsik.
Kata kunci: kepuasan konsumen, kualitas produk, loyalitas konsumen, nilai
ektrinsik, nilai intrinsik, pemodelan persamaan struktural

ANALYSIS OF CUSTOMER SATISFACTION BASED ON
EXTRINSIC AND INTRINSIC VALUE

Scientific Paper
to complete the requirement for graduation of
Bachelor Degree in Statistics
at

Department of Statistics

DEPARTMENT OF STATISTICS
FACULTY OF MATHEMATICS AND NATURAL SCIENCE
BOGOR AGRICULTURAL UNIVERSITY
BOGOR
2015

ACKNOWLEDGEMENTS
Alhamdulillah, many thanks to Allah Subhanahu Wata’ala for giving me the
strength so I can finish my research and bachelor thesis. This research theme is
about Structural Equation Modeling (SEM) with the title Analysis of Customer
Satisfaction with Extrinsic and Intrinsic Value.
I would like to express my sincere gratitude to my supervisors, Prof Dr Asep
Saefuddin MSc and Dian Kusumaningrum SSi MSi for all suggestions, comments,
and kindness of their heart to guide me to do my research. Besides I also want to
thanks to SURE GSB IPB and all of my respondents, without all of you, the data
would not be collected and this research could never be done. Also thanks to all my
friends who have continuously supported me during this research and paper writing
that I cannot mention one by one. Especially for my family who always give support

to keep on fighting to finish this research.
I hope this research can be useful for many people.

Bogor, August 2015
Dewi Cahyani

TABLE OF CONTENT
LIST OF TABLE

vii

LIST OF FIGURE

vii

LIST OF APPENDIX

vii

INTRODUCTION


1

Background

1

Objectives

1

STRUCTURAL EQUATION MODEL

2

METHODOLOGY

3

Data Source and Sampling Technique


3

Methods

3

RESULTS AND DISCUSSIONS

7

Validity and Reliability of Questionnaire

7

Description of Respondents

7

Parameter Estimation of Structural Equation Model


8

CONCLUSIONS AND RECOMMENDATIONS

11

Conclusions

11

Recommendations

12

REFERENCES

12

APPENDIX


13

BIBLIOGRAPHY

27

LIST OF TABLE
1
2
3
4
5
6
7

Goodness of Fit
Reliability of latent variables in the questionnaire
Construct reliability of measurement model
The fit indices of measurement model

The fit indices of structural model
Path coefficients between latent variables
Direct and indirect effects between latent variables

6
7
9
9
10
11
11

LIST OF FIGURE
1
2
3
4

Path diagram
Distribution of respondents based on type of bag

Distribution of respondents based on place of purchase
Structural model

4
7
8
10

LIST OF APPENDIX
1 List of latent variables and indicators
2 Structural and measurement model
3 Questionnaire
4 Qplot of standardized residuals
5 Validity of each statement in the questionnaire
6 Profiles of the respondent
7 Standardize loading factor of measurement model
8 Initial measurement model
9 Modified measurement model
10 Spearman’s correlation between indicators

13
14
15
18
19
20
21
22
23
24

INTRODUCTION
Background
I Love Indonesia is one of the government program launched by the President
in 2009. It aims to improve the competitiveness of domestic products hoping that
all Indonesians can afford to buy it. The government strongly supports any local
company to expand its business not only in Indonesia but also abroad. The President
said that there are 2.2 million innovative industries in Indonesia that were able to
absorb 5.4 million workers, while 700 000 handicraft industries absorbed 1.8
million workers (Setneg.go.id 2009). The development of local product quality does
not only build the nation's economy, but also brings a good image of Indonesia in
international trade.
One of the famous Indonesian local brand for bags, which took part in the
development of domestic business is XYZ. It has produced bags for more than 35
years. XYZ focuses more on backpack, it is segmented for female and male teens
between 12 and 20 years old, oriented to urban sport, and from middle to upper
level economic class (Sudarmadi 2007).
To retain customers and survive in the industry, every company should
understand the customers by fulfilling their needs, both of which can be assessed
before purchase or after purchase so that customer satisfaction can be maintained.
Product quality that can be felt before buying a product is known as extrinsic value
while the product quality that can be felt after buying a product is known as intrinsic
value (Hajjat and Hajjat 2014). If they are satisfied with the product, they will be
loyal with a brand (Fornell 1992). Furthermore, Tjiptono (2007) found that
customer satisfaction contributes to customer loyalty.
This study discusses the effect of product’s extrinsic and intrinsic value on
customer satisfaction, meanwhile customer satisfaction is also assumed to build
their loyalty. This study was applied to XYZ bag owners in undergraduate program
of IPB. The extrinsic and intrinsic value of product quality, customer satisfaction,
and loyalty are variables that cannot be measured or observed directly and needs
indicators to explain their value. Therefore, Structural Equation Model (SEM) is
used to analyze the relationship among unobserved variables. The relationship was
based on Garvin’s product quality dimensions (1984), Hajjat and Hajjat (2014), and
Fornell (1992) theory. Therefore, Confirmatory Factor Analysis (CFA) SEM
approach was used in this study to confirm their relationship. By using this approach,
unobserved variable is measured by one or more observed variables based on the
theory or the definition of an unobserved variable.
Objectives
The objectives of this study are:
1. To identify the effect of extrinsic and intrinsic factors on customer satisfaction
of XYZ bag owners.
2. To identify the direct and indirect influence of product’s extrinsic value
(performance, features, conformance, and aesthetics) and intrinsic value

2
(reliability, durability, serviceability, and perceived quality) on customer
satisfaction and loyalty of XYZ bag owners.

STRUCTURAL EQUATION MODEL
Structural Equation Modeling (SEM) is one of the multivariate analysis to
analyze the relationship between complex variables. SEM can perform a variety of
tests of theoretical models such as confirmatory factor analysis and multiple
regression. There are two types of variables in SEM, latent and indicator variables.
Latent variable is a variable that cannot be measured directly but can be measured
by one or more indicators while indicators can be measured directly (Bollen 1989).
The models built in SEM consist of structural model and measurement model.
Structural model is to explain the relationship between the cause variable
(exogenous latent variable) and the effect variable (endogenous latent variable).
While measurement model is to explain the operationalization of the theory or latent
variable into indicators which is can be used to measure latent variable value
(Kusnendi 2008).
The structural equation for latent variable model is:
= � + �� + �

where,
B : mxm vector of coefficient matrix for latent endogenous variables
� : mxn vector of coefficient matrix for latent exogenous variables
: mxl vector of latent endogenous variables
� : nx1 vector of latent exogenous variables
� : mx1 vector of latent errors in equation
m : number of latent endogenous variables
n : number of latent exogenous variables
While, the structural equation for measurement model is:
=� +
= � �+

where,
y : px1 vector of observed indicator of latent endogenous variables
x : qx1 vector of observed indicator of latent exogenous variables
Ʌy : pxm vector of coefficient relating y to latent endogenous variables
Ʌx : qxn vector of coefficient relating x to latent exogenous variables
ɛ : px1 vector of measurement errors for y
δ : qxl vector of measurement errors for x
p : number of indicators of latent endogenous variables
q : number of indicators of latent exogenous variables

3
There are two types of errors in SEM namely errors associated with structural
models and measurement models. Errors in structural models are represented by
zeta (ζ) and errors in the measurement model are represented by epsilon (ɛ) for
endogenous variables and delta (δ) for the exogenous variables.

METHODOLOGY
Data Source and Sampling Technique
The data used in this research is collected by a survey. The type of sampling
technique used are snowball and purposive sampling. The data was collected among
undergraduate students of Bogor Agricultural University (IPB). Purposive
sampling was used because the population of IPB’s undergraduate students who
owned XYZ bag is unknown. Snowball sampling was used because the ownership
of goods, in this case bags, can be known by first asking people who own this bag,
therefore respondents can connect the researchers to other respondents who also
have XYZ bag. Respondents were approached one-by-one and were asked about
the ownership of XYZ bag. This study used the undergraduate students of IPB as a
sample because it is considered to represent the segmentation of XYZ, which is
associated with female and male teens between 12 and 20 years old, oriented to
urban sport, and from middle to upper level economic class. Although the main
segmentation of XYZ is teenagers, in fact it is also suitable for college students and
young executives because its design is still suitable for them. The total sample of
this study is 300 XYZ owners because the model in this study have 4 latent variables
where one of them is underidentified constructs (latent variable with fewer than
three indicators). Therefore based on Hair et al. (2010) the minimum sample size is
300. The samples are taken from 36 Departments in IPB and from batch 2010 to
2014.
Methods
The step of analysis conducted are:
1. Data preparation and exploration which describes the product.
2. Conceptualization of structural equation model was based on theory.
Garvin (1984) suggested that companies could compete on product quality
dimensions, which include performance, features, reliability, conformance,
durability, serviceability, aesthetics, and perceived quality. According to
Hajjat and Hajjat (2014) these dimensions can also be identified as extrinsic
and intrinsic value for customers. Extrinsic value can be felt before buying a
product and intrinsic value can be felt after buying a product. The first include
product performance, features, conformance and aesthetics and the second
include product reliability, durability, serviceability, and perceived quality.
Customers will be loyal with a brand if they are satisfied with the supplier or
product (Fornell 1992). Thus, the latent variables are extrinsic and intrinsic
value, customer satisfaction, and loyalty. The list of latent variables and their
indicators can be seen in Appendix 1.
3. Making path diagram.

4
Path diagram helps to explain the causal relationship between the latent
variables or structural model and between latent variable and its indicators or
measurement model. The path diagram developed can be seen in Figure 1
while the matrix of structural and measurement model is included in
Appendix 2.

Figure 1 Path diagram
4. Identifying the model using t-Rule.
According to Bollen (1989), if a parameter is not identified then a consistent
estimator cannot be determined for this parameter. Parameter identification
test using t-Rule is based on the formula:
�≤

+

+

+

where t is the number of unknown parameters and (p + q) is the number of
indicators.
5. Collecting data from XYZ bag owner with the following stages:
a. Making the conceptual questionnaires. The questionnaire consist of seven
questions about the characteristics of respondents and 26 questions that
will be used to measure the latent variables or known as indicator. The
latent variables were developed based on the theory. Likert scale was used
with a score of 1 to 5, where 5 indicates strongly agree and 1 indicates
strongly disagree. The questionnaire can be seen in Appendix 3;
b. Pretesting questionnaire to 30 respondents in order to test the validity and
reliability of the questionnaire. Spearman correlation was used to check

5
the validity while Cronbach Alpha method was used to check the
reliability;
c. Distributing questionnaires to undergraduate students of IPB who owned
XYZ bag based on purposive and snowball sampling techniques.
6. Conducting data exploration to analyze the customer’s characteristic about
gender, faculty, batch, income per month, the place of purchase, type of bag
purchase, and duration of use. Income per month and the place of purchase
will explain the market segmentation of respondents.
7. Determining the input matrix to estimate parameters.
The input matrix used in this research was the covariance matrix because the
indicators have the same scale and units.
8. Estimating parameters.
In this study, estimation of the parameters used Unweighted Least Squares
(ULS). The ULS fitting function is:
���

= ( )� [ � − �

]

where,
S
: sample covariance matrix of observed variables
Σ(θ)
: covariance matrix of model variables
According to Morata-Ramirez and Holgado-Tello (2013), ULS is a parameter
estimation used for ordinal data and have a multivariate normal distribution.
ULS is also recommended for data that has a small sample size (250 subjects).
This study used ordinal data with Likert scale and the sample size is 300. Then
check the normality of data by using a Qplot of standardized residuals. If the
data are close to the 45 degree line, it can be say the data is normal distributed
(Kusnendi 2008). The Qplot in Appendix 4 shows that the data is normal
distributed.
9. Evaluating model fit with following step:
a. Measurement model fit;
i. Indicator Validity
Valid indicators indicate that indicators measure the latent variables
correctly. An indicator is valid when the t-value of loading factor is
above 1.96 or the standardized loading factor is above 0.50 (Wijanto
2007). Moreover, to check multicolinearity between indicators it can
be tested with a correlation test (Chatterjee and Hadi 2006). A
correlation between 0.70 - 0.90 has a strong relationship (Dancey and
Reidy's 2004). Thus, one of the indicators having high correlation
should be removed in the model.
ii. Construct Reliability
Reliability of a latent variable is calculated from the construct
reliability or variance extracted. High reliability indicates that the
indicator variable is consistent to measure a latent variable. Construct
reliability and variance extracted can be calculated by:

6
�� =

(∑ = � )

(∑ = � ) + ∑ = �



=

∑= �


where CRj is construct reliability of jth latent variable, VEj is variance
extracted of jth latent variable, k is number of indicator of jth latent
variable, λij is standardize loading factor ith indicator and jth latent
variable, and ei is the measurement error of ith indicator in jth latent
variable.
Construct reliability values greater than 0.70 or variance extracted
values greater than 0.50 show that the variable has a good reliability
(Wijanto 2007).
b. Structural model fit.
The evaluation of the structural model includes significance test of
estimated coefficient. The relationship between variables is significant
when the t-value is above 1.96 (Wijanto 2007).
c. Overall model fit;
SEM does not have the best statistical test that can explain the goodness
of fit of the model like other multivariate analysis. This causes the
development of several statistical tests to evaluate the goodness of fit of
the model. Table 1 provides guideline for using fit indices based on Hair
et al. simulation that was used in this research, N applies to number of
respondents and (p+q) is the number of observed variables. Overall model
fit is used both in measurement and structural model.
Table 1 Goodness of Fit
Goodness of Fit
Chi-square
Comparative Fit Index (CFI)
Standardized Root Mean Square
Residual (SRMR)
Goodness of Fit Index (GFI)
Source: Hair et al. (2010).

Criteria for
N > 250 and 12 < (p+q) < 30
p-value less than 0.50
Above 0.92
0.08 or less (with CFI above
0.92)
Above 0.90

10. Interpretation of the models, which include the measurement model, the
structural model, and the relationship between latent variables.

7

RESULTS AND DISCUSSIONS
Validity and Reliability of Questionnaire
Pretest was conducted to test the validity and reliability of the questionnaire.
This test was conducted on 30 respondents. The test results show that all the
indicators are valid at alpha 5% and all indicators of a latent variable is reliable with
Cronbach Alpha value greater than 0.5. Validity and reliability value are shown in
Appendix 5 and Table 2, respectively.
Table 2 Reliability of latent variables in the questionnaire
Dimension and Latent
Variable
Performance
Feature
Reliability
Conformance
Durability

Cronbach
Alpha
0.76
0.65
0.75
0.74
0.77

Dimension and Latent
Variable
Serviceability
Aesthetics
Perceived Quality
Satisfaction
Loyalty

Cronbach
Alpha
0.62
0.81
0.94
0.86
0.86

Description of Respondents
The respondents are undergraduate students of IPB who owned XYZ bag.
The detail profiles of the respondents can be seen in Appendix 6. More than threefourths of respondents (78.7%) were female. Most of them are students from batch
2013 with 109 respondents (36.3%) and from FMIPA with 126 respondents (42%).
It was easier to get respondents from FMIPA faculty because it is the researcher’s
faculty. See the bar chart at Figure 2. Almost all respondents (92%) had a backpack
(the respondents can be have more than one type). It is not surprising because XYZ
is more focuses on backpack than others and it popular with this type. In addition,
half of them (59%) have used these bags for 1-5 years.
350
300
250
200
150
100

50
0
Other Respondents
Number of Respondents

Backpack
24
276

Shoulder
Bag
246
54

Laptop Bag

Others

280
20

298
2

Figure 2 Distribution of respondents based on type of bag
The majority of respondents had an income Rp 500 000 – Rp 1 000 000 with

8
196 respondents (65.3%). It indicated that an economic class of respondents is
middle level. Figure 3 explained the distribution of respondents based on place of
purchase (the respondents can buy the bag in more than one place). Approximately
two-third of them (69.7%) have bought the bag in Department Store. The
percentages of respondents who bought the product in Book Store and Official
Outlet are 18.6% and 16.3%. Only a few of respondents (6%) bought the product
in other places such as traditional market and a bag store. This indicates that XYZ
bag oriented to urban communities because the purchase proportion in modern
market such as Department Store and Book Store more than buying in the
traditional market (Sudarmadi 2007).
350
300
250
200
150
100

50
0
Other Respondents
Number of Respondents

Outlet
251
49

Dept Store
91
209

Book Store
244
56

Others
282
18

Figure 3 Distribution of respondents based on place of purchase
Parameter Estimation of Structural Equation Model
SEM is used to test the measurement model and structural model of the path
diagram based on the theory. The result of identification using the t-Rule shows that
the model used in this analysis is an over-identified model with t equal to 56. It
means, the model created has a smaller number of estimated parameters than the
number of data used. Structural equation model testing uses a two stage approach,
first testing the measurement model and then testing the structural model.
Measurement model explains the relationship between latent variable and its
indicators. Measurement model uses first order confirmatory factor analysis (CFA).
First order CFA is a measurement model where the latent variable is only measured
by its indicators and does not involve other latent variable (Kusnendi 2008).
The measurement model test results show that there is only one indicator
having standardized loading factor less than 0.5, which is the resistance to water
(X18). Afterwards, we check the Spearman correlation between indicators too. This
is done to make sure that there are no multicollinearity between indicators.
Spearman correlation between the indicators can be seen in Appendix 10. If a pair
of indicators have a correlation above 0.7, one indicator will be removed by trial
and error. Deletion will be performed on indicators having the most decreasing chisquare value in the measurement model. Therefore, there are four indicators that
will be removed, namely design (X8), the likelihood of functionality (X20), the
likelihood of reliability (X21), and consistency (Y5). Hence, there are five indicators

9
deleted from the model based on standardized loading factor and Spearman
correlation, namely X8, X18, X20, X21, and Y5. The standardized loading factor of
measurement models can be seen in Appendix 7 which also explains visually by
path diagram in Appendix 8 and 9 for initial and modified measurement model
respectively. After modification, all indicators have standardized loading factor
above 0.5 and t-value above 1.96.
The reliability of a latent variable is calculated based on the construct
reliability or variance extracted. The results show that all the latent variables are
reliable. It can be seen from the value of reliability construct greater than 0.70 and
variance extracted greater than 0.50. The value of construct reliability and variance
extracted can be seen in Table 3.
Table 3 Construct reliability of measurement model
Latent Variable
Extrinsic Value
Intrinsic Value
Satisfaction
Loyalty

Construct Reliability
0.9
0.9
0.9
0.8

Variance extracted
0.5
0.5
0.8
0.7

Result
Reliable
Reliable
Reliable
Reliable

Fitting indices of the measurement model can be seen in Table 4. There is
significant change between the initial and modified measurement model. Before
modification, the model has chi-square value 1303.11 then after modification, chisquare model is getting smaller, which is 590.28. Chi-square value indicates that
the modified model has a better fit with significant p value. Besides that CFI and
GFI are high (1.00 and 0.99) and SRMR is low (0.051). Thus, it can be concluded
that the modified measurement model is adequate.
Table 4 The fit indices of measurement model
Goodness-of-Fit
Statistics
Chi-square

Fit guideline values*
(N = 300 and (p+q) = 26)
Smaller value
Significant p-value
CFI
Above 0.92
SRMR
Less than 0.08
GFI
Above 0.90
* Based on Hair et al. (2010).

Measurement model
Initial
Modify
1303.11
590.28
Significant
Significant
1.00
1.00
0.059
0.051
0.98
0.99

After the evaluation of the measurement model, the next step is testing the
structural model. Structural model test results are shown in Table 5. The p-value of
chi-squared (significant and having a value of 592.58), SRMR (0.051) and GFI
(1.00) show good results. It can be concluded that the model is also adequate.

10
Table 5 The fit indices of structural model
Goodness-of-Fit
Statistics
Chi-square

Fit guideline values*
(N = 300 and (p+q) = 26)
Smaller value
Significant p-value
CFI
Above 0.92
SRMR
Less than 0.08
GFI
Above 0.90
* Based on Hair et al. (2010).

Result
592.58
(p value = 0.00)
1.00
0.051
0.99

Figure 4 Structural model
Path diagram of structural model illustrated in Figure 4. Based on that figure,
the indicators of product quality that have the greatest value to the extrinsic value
is functionality (X3) with standardized loading factor 0.83. It indicates that the bag
performance is very important for the customers. The intrinsic value of likely
quality (X19) reach the highest value of standardized loading factor (0.78). It means
that the high quality product will have great appreciation from the customers. Then
the perceived quality is also important.
Table 6 explains the relationship between latent variables and their path
coefficient. It shows that all the path coefficients are significant at alpha 5%. The
relationship between extrinsic value and satisfaction, intrinsic value and satisfaction,
and satisfaction and loyalty are positive. It indicates that the dimensions of product
quality which is divided into two groups, extrinsic and intrinsic value, have an
influence to customer satisfaction. In addition, the customer satisfaction influences
the customer loyalty i.e. the customer loyalty depends on the customer satisfaction.

11
Table 6 Path coefficients between latent variables
Structural Relationship
Extrinsic Value → Satisfaction
Intrinsic Value → Satisfaction
Satisfaction → Loyalty

Path Coefficient
0.42
0.46
0.70

t-value
2.23
2.42
15.66

Result
Significant
Significant
Significant

From path coefficient of structural model, it can be used to calculate the direct
effect and indirect effect between latent variables. Direct effect is the directly effect
from one variable to a second variable while indirect effect is the effect between
two variables that mediated by one or more variables. Then total effect is sum of
direct and indirect effect (Kusnendi 2008). See Table 7, based on the total effect,
the extrinsic and intrinsic value are not statistically different at alpha 5%. Therefore
the company is suggested to emphasize both extrinsic and intrinsic performance in
order to increase customer satisfaction. At the end, it will increase the loyalty.
Table 7 Direct and indirect effects between latent variables
Structural Relationship
Extrinsic Value → Satisfaction
Intrinsic Value → Satisfaction
Satisfaction → Loyalty
Extrinsic Value → Loyalty
Intrinsic Value → Loyalty

Direct effect
0.42
0.46
0.70
-

Indirect effect Total effect
0.42
0.46
0.70
0.29
0.29
0.32
0.32

These results are relevant only for the IPB market segmentation. The research
is based on IPB students having a specific market segmentation. Based on income
and place of purchase, the respondents are urban college student that have a middle
level economic class. The result might be different for upper level economic class,
where the extrinsic and intrinsic value may have different effect.

CONCLUSIONS AND RECOMMENDATIONS
Conclusions
The study results show that both extrinsic and intrinsic value influence
customer satisfaction in urban middle level economic class and they are not
statistically different at alpha 5%. In addition, the customer satisfaction effects the
customer loyalty. In other words, the more the customers are satisfied, the more the
customers will become loyal. Besides that the most important things for customer
are the bag performance (functionality, X3) in extrinsic value and perceived quality
of product (likely quality, X19) in intrinsic value.

12
Recommendations
First, the company has to increase both extrinsic and intrinsic value to
increase customer satisfaction. The increasing of customer satisfaction will lead to
more loyal customer. Second, to have a broader conclusion, the respondent must be
extended on other level of economic class.

REFERENCES
Bollen KA. 1989. Structural Equations with Latent Variables. New York (US):
John Wiley & Sons.
Chatterjee S, Hadi AS. 2006. Regression Analysis by Example. 4th Edition. New
Jersey (US): John Wiley & Sons.
Dancey C, Reidy J. 2004. Statistics without Maths for Psychology: Using SPSS for
Windows. London (US): Prentice Hall.
Fornell. 1992. A National Customer Satisfaction Barometer: The Swedish
Experience. Journal of Marketing. 56(1): 6-21.
Garvin DA. 1987. Competing on the Eight Dimensions of Quality. Harvard
Business Review [Internet]. [download 2015 Feb 21]; November-December
1987:101-109.
Available
at:
http://cc.sjtu.edu.cn/G2S/eWebEditor/uploadfile/2013042709184994 4.pdf
Hair JF, Black WC, Babin BJ, Anderson RE. 2010. Multivariate Data Analysis. 7th
Edition. New Jersey (US): Pearson Prentice Hall.
Hajjat MM, Hajjat F. 2014. The Effect of Product Quality on Business Performance
in Some Arab Companies. Journal of Emerging Trends in Economics and
Management Sciences. 5(5): 498-508.
Kusnendi. 2008. Model-Model Persamaan Struktural (Satu dan Multigroup Sampel
dengan Lisrel). Bandung (ID): CV Alfabeta.
Morata-Ramirez MDA, Holgado-Tello FP. 2013. Construct Validity of Likert
Scales through Confirmatory Factor Analysis: A Simulation Study Comparing
Different Methods of Estimation Based on Pearson and Polychoric Correlations.
International Journal of Social Science Studies [Internet]. [download 2015 June
3]; 1(1): April 2013. Available at: http://dx.doi.org/10.11114/ijsss.vlil.27
Sudarmadi. 2007. 10 Pengusaha yang Sukses Membangun Bisnis dari 0. Jakarta
(ID): PT Gramedia Pustaka Utama.
Tjiptono F. 2007. Pemasaran Jasa. Malang (ID): Banyumedia Publishing.
Wijanto SH. 2007. Structural Equation Modeling dengan Lisrel 8.8. Jakarta (ID):
Graha Ilmu.

13
Appendix 1 List of latent variables and indicators
Latent Variable

Extrinsic Value �

Intrinsic Value �

Satisfaction �
Loyalty �

Indicator
Performance
Comfortable
Up to date design
Functionality
Feature
Materials used
Pocket placement
Conformance
Design and function
Price and quality
Aesthetics
Design
Color options
Durability
The main materials
Color main material
Zipper/magnetic clasp/drawstring
Bag strap
Serviceability
Ease to service
After-sales service
Reliability
Comfortable for a long time used
Stitching
Resistance to water
Perceived Quality
Likely quality
The likelihood of functionality
The likelihood of reliability
Expected quality
Overall satisfaction
Consider to loyal
Recommendation
Consistency

Indicator
Notation
X1
X2
X3
X4
X5
X6
X7
X8
X9
X10
X11
X12
X13
X14
X15
X16
X17
X18
X19
X20
X21
Y1
Y2
Y3
Y4
Y5

14
Appendix 2 Structural and measurement model
1. Structural Model

= � + �� + �

] [� ] + [


[� ] = [

2. Measurement Model

=

[

]

[ ]











[

=



[



] [ ] + [� ]


= � �+














=�


[ ]+


]

+

[


� [� ] +

[ ]
� ]

]

15
Appendix 3 Questionnaire
INSTITUT PERTANIAN BOGOR
FAKULTAS MATEMATIKA DAN ILMU PENGETAHUAN ALAM
DEPARTEMEN STATISTIKA
Jl. Meranti Wing 22, Level 4 Kampus IPB Dramaga-Bogor 16680
Kuesioner Penelitian
Terima kasih atas partisipasi Anda menjadi salah satu responden dalam penelitian
saya Dewi Cahyani (G14110054), mahasiswa Departemen Statistika IPB, dengan
judul penelitian Analysis of Customer Satisfaction Based on Extrinsic and Intrinsic
Value. Saya sangat menghargai kejujuran Anda dalam mengisi kuesioner ini.
Identitas Anda akan dijamin kerahasiaannya dan hasil survey ini semata-mata
hanya akan digunakan untuk tujuan penelitian, bukan komersial.
PENYARINGAN
Apakah Anda memiliki
tas merek XYZ?

(a) Ya
(b) Tidak

KARAKTERISTIK RESPONDEN
Nama
Departemen/Fakultas
Angkatan
Jenis Kelamin
(1) Laki- laki
(2) Perempuan
Uang saku per bulan
(a) < Rp 500.000
(b) Rp 500.000 – Rp 1.000.000
(c) Rp 1.000.001 – Rp 2.000.000
(d) Rp 2.000.001 – Rp 2.500.000
(e) > Rp 2.500.000
Jenis tas XYZ yang
(a) Tas Ransel (backpack)
dimiliki
(b) Tas Selempang (shoulder bag)
(dapat lebih dari satu)
(c) Tas Laptop (laptop bag)
(d) Lainnya, sebutkan .......................
Lokasi pembelian tas
(a) Outlet resmi XYZ
XYZ
(b) Department Store (contoh: Matahari, Ramayana)
(dapat lebih dari satu)
(c) Book Store (contoh: Gramedia, Toko Gunung
Agung)
(d) Lainnya, sebutkan .......................
Lama telah
(a) < 1 tahun
menggunakan tas XYZ
(b) 1 – 5 tahun
(c) > 5 tahun

16
Silang (X) setiap skor penilaian sesuai dengan penilaian Anda terhadap
penyataan-pernyataan berikut
Keterangan:
 1 Sangat tidak setuju
 2 Tidak setuju
 3 Netral
 4 Setuju
 5 Sangat setuju
No

Pernyataan

A. PERFORMA
1
Tas XYZ nyaman digunakan
2
Desain tas XYZ up to date
3
Tas XYZ berfungsi dengan baik
B. FITUR
1
Bahan yang digunakan tas XYZ baik
2
Penempatan kantung tas XYZ baik
C. KEANDALAN
Tetap nyaman digunakan dalam jangka
1
waktu yang lama
2
Jahitan tas XYZ kuat
3
Tas XYZ tahan terhadap air
D. KESESUAIAN
1
Desain tas XYZ sesuai dengan fungsinya
Harga tas XYZ sesuai dengan
2
kualitasnya
E. KETAHANAN
1
Bahan utama tas XYZ tidak mudah robek
Warna bahan utama tas XYZ tidak
2
mudah pudar
Resleting/katup magnet/serut pada tas
3
XYZ tidak mudah rusak
Pegangan (strap) tas XYZ tidak mudah
4
putus
F. KEMUDAHAN SERVIS
Tas XYZ mudah diperbaiki jika terdapat
1
kerusakan
2
Layanan purnajual tas XYZ baik
G. ESTETIKA
1
Desain tas XYZ menarik
2
Pilihan warna tas XYZ menarik

1

2

Penilaian
3
4

5

17
No

Pernyataan

H. PERSEPSI KUALITAS
Sangat besar kemungkinan tas XYZ akan
1
berkualitas tinggi
Sangat besar kemungkinan tas XYZ akan
2
berfungsi dengan baik
Sangat besar kemungkinan tas XYZ akan
3
dapat diandalkan
I. KEPUASAN KONSUMEN
Kualitas tas XYZ sesuai dengan yang
1
saya harapkan
2
Saya puas pada keseluruhan tas XYZ
J. LOYALITAS KONSUMEN
Saya memikirkan akan setia pada tas
1
XYZ
Saya merekomendasikan tas XYZ pada
2
orang-orang terdekat saya
Saya akan tetap membeli tas XYZ
3
walaupun banyak produk lain sejenis
yang memiliki penawaran menarik

1

2

Penilaian
3
4

5

18
Appendix 4 Qplot of standardized residuals

19
Appendix 5 Validity of each statement in the questionnaire
Dimension and
Latent Variable
Performance
Feature
Conformance
Aesthetics
Durability

Serviceability
Reliability
Perceived Quality
Satisfaction
Loyalty

Indicator
X1
X2
X3
X4
X5
X6
X7
X8
X9
X10
X11
X12
X13
X14
X15
X16
X17
X18
X19
X20
X21
Y1
Y2
Y3
Y4
Y5

Coefficient
Correlation
0.88
0.74
0.87
0.91
0.82
0.87
0.90
0.91
0.90
0.75
0.77
0.87
0.76
0.78
0.85
0.81
0.74
0.83
0.91
0.85
0.94
0.92
0.93
0.94
0.84
0.90

20
Appendix 6 Profiles of the respondent
Freq
Gender
Male
Female

%

64
236
300

Freq

21.3
78.7
100

Batch
2010

2

0.7

2011

38

12.7

2012

87

29.0

2013

109

36.3

2014

64
300

21.3
100

Type of bag
Backpack
Shoulder
bag
Laptop bag
Others
Faculty
FAPERTA
FKH
FPIK
FAPET
FAHUTAN
FATETA
FMIPA
FEM
FEMA

Duration of used
Under 1 year
1 - 5 years
Over 5 years
Income per month
Less than Rp 500
000
Rp 500 000 - Rp 1
000 000
Rp 1 000 000 - Rp 2
000 000
Rp 2 000 000 - Rp 2
500 000
Over Rp 2 500 000

276

92.0

Place of purchase
Official outlet

54

18.0

Department Store

20
2

6.6
0.7

27
2
22
9
23
44
126
34
13
300

9.0
0.7
7.3
3.0
7.7
14.7
42.0
11.3
4.3
100

Book Store
Others

%

36
177
87
300

12.0
59.0
29.0
100

16

5.3

196

65.3

85

28.3

1

0.3

2
300

0.7
100

4

16.3

209

69.6

56
18

18.6
6.0

21
Appendix 7 Standardize loading factor of measurement model
Latent
Variable

Indicator

Std. loading
factor of
initial model

Performance
X1
X2
X3
Feature
X4
Extrinsic
X5
Value
Conformance
X6
X7
Aesthetics
X8
X9
Durability
X10
X11
X12
X13
Serviceability
X14
X15
Intrinsic
Value
Reliability
X16
X17
X18
Perceived Quality
X19
X20
X21
Y1
Satisfaction
Y2
Y3
Loyalty
Y4
Y5

t-value

Std. loading
factor of
modify
model

t-value

0.81
0.63
0.83

32.72
26.66
32.34

0.81
0.60
0.83

29.81
23.22
29.24

0.80
0.70

31.49
29.31

0.81
0.71

28.74
26.77

0.76
0.81

29.55
32.56

0.75
0.81

26.70
29.59

0.64
0.63

26.93
26.79

0.60

23.62

0.68
0.60
0.59
0.71

31.16
29.07
30.74
30.80

0.69
0.62
0.60
0.72

28.17
26.72
27.79
27.77

0.59
0.57

25.42
25.51

0.59
0.55

22.75
22.36

0.73
0.70
0.45

31.23
34.62
23.13

0.74
0.70
-

28.30
30.86
-

0.80
0.85
0.82
0.88
0.87
0.88
0.81
0.82

33.63
34.54
33.57
16.65
16.77
27.10
26.58
26.62

0.78
0.89
0.87
0.87
0.82
-

29.25
16.34
16.50
19.27
19.33
-

22
Appendix 8 Initial measurement model

23
Appendix 9 Modified measurement model

24
Appendix 10 Spearman’s correlation between indicators
X1
X2
X3
X4
X5
X6
X7
X8
X9
X10
X11
X12
X13
X14
X15
X16
X17
X18
X19
X20
X21
Y1
Y2
Y3
Y4
Y5

X1
1.00
0.50
0.63
0.57
0.53
0.55
0.58
0.42
0.41
0.47
0.47
0.38
0.47
0.37
0.37
0.57
0.52
0.30
0.51
0.55
0.54
0.58
0.52
0.40
0.39
0.40

X2
0.50
1.00
0.44
0.38
0.39
0.43
0.38
0.65
0.58
0.22
0.28
0.21
0.30
0.33
0.36
0.31
0.29
0.29
0.41
0.43
0.42
0.47
0.52
0.35
0.32
0.39

X3
0.63
0.44
1.00
0.62
0.54
0.54
0.57
0.40
0.37
0.47
0.40
0.39
0.50
0.41
0.43
0.52
0.51
0.29
0.54
0.61
0.58
0.57
0.56
0.34
0.38
0.40

X4
0.57
0.38
0.62
1.00
0.59
0.49
0.58
0.38
0.36
0.53
0.48
0.40
0.47
0.41
0.38
0.55
0.61
0.34
0.57
0.56
0.50
0.57
0.50
0.38
0.37
0.39

X5
0.53
0.39
0.54
0.59
1.00
0.47
0.50
0.45
0.43
0.38
0.44
0.42
0.48
0.43
0.34
0.49
0.46
0.21
0.47
0.50
0.51
0.53
0.45
0.30
0.29
0.26

X6
0.55
0.43
0.54
0.49
0.47
1.00
0.64
0.41
0.42
0.54
0.47
0.37
0.48
0.42
0.43
0.50
0.45
0.36
0.53
0.57
0.55
0.55
0.53
0.39
0.37
0.39

X7
0.58
0.38
0.57
0.58
0.50
0.64
1.00
0.39
0.34
0.56
0.44
0.41
0.50
0.40
0.34
0.56
0.55
0.34
0.56
0.57
0.56
0.53
0.49
0.38
0.41
0.38

X8
0.42
0.65
0.40
0.38
0.45
0.41
0.39
1.00
0.82
0.21
0.34
0.28
0.38
0.33
0.39
0.30
0.32
0.23
0.40
0.44
0.42
0.48
0.48
0.33
0.34
0.33

X9
0.41
0.58
0.37
0.36
0.43
0.42
0.34
0.82
1.00
0.23
0.37
0.30
0.38
0.29
0.38
0.32
0.31
0.23
0.40
0.39
0.37
0.47
0.45
0.31
0.27
0.29

X10
0.47
0.22
0.47
0.53
0.38
0.54
0.56
0.21
0.23
1.00
0.49
0.49
0.51
0.35
0.39
0.49
0.57
0.35
0.52
0.53
0.49
0.46
0.43
0.37
0.33
0.31

25
Appendix 10 (Continue) Spearman’s correlation between indicators
X1
X2
X3
X4
X5
X6
X7
X8
X9
X10
X11
X12
X13
X14
X15
X16
X17
X18
X19
X20
X21
Y1
Y2
Y3
Y4
Y5

X11
0.47
0.28
0.40
0.48
0.44
0.47
0.44
0.34
0.37
0.49
1.00
0.54
0.43
0.43
0.34
0.38
0.40
0.28
0.35
0.38
0.39
0.43
0.44
0.29
0.26
0.25

X12
0.38
0.21
0.39
0.40
0.42
0.37
0.41
0.28
0.30
0.49
0.54
1.00
0.62
0.39
0.29
0.37
0.41
0.24
0.42
0.45
0.43
0.48
0.40
0.31
0.23
0.23

X13
0.47
0.30
0.50
0.47
0.48
0.48
0.50
0.38
0.38
0.51
0.43
0.62
1.00
0.38
0.38
0.45
0.48
0.26
0.47
0.51
0.51
0.58
0.44
0.38
0.35
0.38

X14
0.37
0.33
0.41
0.41
0.43
0.42
0.40
0.33
0.29
0.35
0.43
0.39
0.38
1.00
0.54
0.40
0.36
0.34
0.40
0.41
0.41
0.39
0.44
0.35
0.28
0.33

X15
0.37
0.36
0.43
0.38
0.34
0.43
0.34
0.39
0.38
0.39
0.34
0.29
0.38
0.54
1.00
0.36
0.32
0.29
0.45
0.49
0.44
0.36
0.44
0.30
0.29
0.36

X16
0.57
0.31
0.52
0.55
0.49
0.50
0.56
0.30
0.32
0.49
0.38
0.37
0.45
0.40
0.36
1.00
0.61
0.33
0.53
0.52
0.48
0.50
0.47
0.34
0.35
0.35

X17
0.52
0.29
0.51
0.61
0.46
0.45
0.55
0.32
0.31
0.57
0.40
0.41
0.48
0.36
0.32
0.61
1.00
0.38
0.54
0.53
0.47
0.51
0.48
0.38
0.39
0.39

X18
0.30
0.29
0.29
0.34
0.21
0.36
0.34
0.23
0.23
0.35
0.28
0.24
0.26
0.34
0.29
0.33
0.38
1.00
0.33
0.35
0.28
0.30
0.32
0.32
0.26
0.38

X19
0.51
0.41
0.54
0.57
0.47
0.53
0.56
0.40
0.40
0.52
0.35
0.42
0.47
0.40
0.45
0.53
0.54
0.33
1.00
0.83
0.77
0.59
0.53
0.41
0.39
0.36

X20
0.55
0.43
0.61
0.56
0.50
0.57
0.57
0.44
0.39
0.53
0.38
0.45
0.51
0.41
0.49
0.52
0.53
0.35
0.83
1.00
0.85
0.63
0.61
0.43
0.41
0.43

26
Appendix 10 (Continue) Spearman’s correlation between indicators
X1
X2
X3
X4
X5
X6
X7
X8
X9
X10
X11
X12
X13
X14
X15
X16
X17
X18
X19
X20
X21
Y1
Y2
Y3
Y4
Y5

X21
0.54
0.42
0.58
0.50
0.51
0.55
0.56
0.42
0.37
0.49
0.39
0.43
0.51
0.41
0.44
0.48
0.47
0.28
0.77
0.85
1.00
0.65
0.61
0.43
0.41
0.41

Y1
0.58
0.47
0.57
0.57
0.53
0.55
0.53
0.48
0.47
0.46
0.43
0.48
0.58
0.39
0.36
0.50
0.51
0.30
0.59
0.63
0.65
1.00
0.74
0.49
0.46
0.45

Y2
0.52
0.52
0.56
0.50
0.45
0.53
0.49
0.48
0.45
0.43
0.44
0.40
0.44
0.44
0.44
0.47
0.48
0.32
0.53
0.61
0.61
0.74
1.00
0.54
0.49
0.51

Y3
0.40
0.35
0.34
0.38
0.30
0.39
0.38
0.33
0.31
0.37
0.29
0.31
0.38
0.35
0.30
0.34
0.38
0.32
0.41
0.43
0.43
0.49
0.54
1.00
0.70
0.75

Y4
0.39
0.32
0.38
0.37
0.29
0.37
0.41
0.34
0.27
0.33
0.26
0.23
0.35
0.28
0.29
0.35
0.39
0.26
0.39
0.41
0.41
0.46
0.49
0.70
1.00
0.65

Y5
0.40
0.39
0.40
0.39
0.26
0.39
0.38
0.33
0.29
0.31
0.25
0.23
0.38
0.33
0.36
0.35
0.39
0.38
0.36
0.43
0.41
0.45
0.51
0.75
0.65
1.00

27

BIBLIOGRAPHY
Dewi Cahyani was born in Kebumen (Central Java) as the eldest daughter
of Dwi Eko Hartoyo and Chalimah on July 3rd, 1993. She lived and grew up in
Bekasi (West Java). She was graduated from SMAN 31 Jakarta and SMPN 2 Bekasi
and then continue her study in bachelor degree of IPB on 2011. She accepted in IPB
by SNMPTN-Undangan.
During her college time, she joined as a committee in Statistika Ria, Pesta
Sains Nasional, IPB’s Dedicated to Education, and several events in IPB. She also
experienced organization in Century IPB as the Secretary of Production and
Operation Division in 2012/2013 and Gamma Sigma Beta as the Treasurer of
Database Center Department in 2013/2014. On July 2014, she had an internship in
Research and Development Center Resources and Equipment of Post and
Information Technology (Puslitbang SDPPI), Ministry of Communication and
Informatics of The Republic of Indonesia.