The Influence of Trust and Celebrity Endorser on Purchase Decisions
in Shopee E-Commerce Through Shopee Loyalty Profile
Clara Dwi Puspita Rahayu and Sri Zuliarni
Business Administration, State Polytechnic of Batam, St. Ahmad Yani, Batam, Indonesia
Keywords: Trust, Celebrity Endorser, Purchase Decision, Shopee Loyalty Profile.
Abstract: The purpose of this study was to determine the effect of trust and celebrity endorser on purchasing decisions
through Shopee loyalty profile as a moderating variable. The method used is multiple regression analysis
method, with a sample of 240 respondents. Respondents' data was collected using a questionnaire. The results
show that trust has an influence on purchasing decisions on e-commerce Shopee with a t
value
7.165 > t
table
1.970. Celebrity endorsers have an influence on purchasing decisions on e-commerce Shopee with a t
value
7.230 > t
table
1.970. Trust and celebrity endorsers have an influence on purchasing decisions on e-commerce
Shopee with a value of F
count
45.231 > F
table
3.033. Shopee's loyalty profile strengthens the influence of trust
on purchasing decisions on Shopee e-commerce, as evidenced by the results of the Moderated Regression
Analysis (MRA) test with an increase of 0.02. Shopee's loyalty profile strengthens the influence of celebrity
endorsers on purchasing decisions on Shopee e-commerce, as evidenced by the results of the Moderated
Regression Analysis (MRA) test with an increase of 0.051.
1 INTRODUCTION
The progress of progressive and modern times
encourages various changes, one of which is in the
field of technology which is increasingly showing its
progress. The internet has influenced market behavior
so that it is used to advance business and market
activities. For example, if they want to buy their daily
needs, the seller can contact customers in the
industrial market via the internet. This development
triggered the birth of the term e-commerce, namely
the direct sale of goods and services via the internet.
According to Waro and Widowati (2020), a
company's success in influencing consumer
purchasing decisions is fostering customer
communication, continuing product innovation in
accordance with trends to introduce brands, and
requiring consumer trust. In addition to gaining
public trust, Shopee is also very aggressively
promoting itself through celebrity endorsers, where
celebrities will enable the platform and inform the
public about promotions carried out by Shopee, such
as product discount events, free shipping, and
cashback. Celebrities chosen by Shopee also have
significant influence and a wide audience, such as
Arya Saloka, Amanda Manopo, Angga Yunanda,
Muhammad Rizky Billar, and Lestiani Andryani,
often known as Lesti Kejora, and a series of other top
artists, not only local Shopee artists. In addition, it
attracted International Celebrities such as Jackie
Chan, Blackpink, and Cristiano Ronaldo to attract
public interest. This is in accordance with the analysis
by Ansar, Suryaningsih, et al. (2016), which states
that celebrity endorsers influence purchasing
decisions. In addition to trust and celebrity endorsers,
Shopee has a loyalty level for all its loyal users,
including Classic, Silver, Gold, and Platinum. The
higher the user loyalty, the more benefits the user will
receive, such as free shipping bonuses, extra
cashback, birthday vouchers, and other cheaper
shopping offers according to the user's loyalty level.
Ayuning (2019) claims that the Shopee Indonesia
loyalty programs application impacts impulse buying
of products.
2 RESEARCH METHOD
The method uses quantitative analysis, namely theory
testing through data analysis with statistical
procedures. The survey method collects information
from certain locations with a questionnaire as a tool
688
Rahayu, C. and Zuliarni, S.
The Influence of Trust and Celebrity Endorser on Purchase Decisions in Shopee E-Commerce Through Shopee Loyalty Profile.
DOI: 10.5220/0011862900003575
In Proceedings of the 5th International Conference on Applied Science and Technology on Engineering Science (iCAST-ES 2022), pages 688-697
ISBN: 978-989-758-619-4; ISSN: 2975-8246
Copyright © 2023 by SCITEPRESS – Science and Technology Publications, Lda. Under CC license (CC BY-NC-ND 4.0)
for data collection. This method aims to collect data
on research subjects by asking a large number of
respondents who are considered to represent a
population. The nature of this study uses associative
explanations or research that tries to explain the
position of the variables to be studied, namely the
trust variable (X1), celebrity endorser variable (X2),
purchasing decisions (Y), and Shopee loyalty profile
(Z) and the relationship between all variables. The
sampling method uses non-probability sampling, a
sample selection strategy that does not employ equal
opportunities for each component of the population to
be sampled. Non-probability in this study uses
purposive sampling (Sugiyono, 2014). The criteria
determined by the researcher are based on age with a
minimum age of 17 years, making purchases at
Shopee at least three times, and having seen one of
the celebrities promoting Shopee e-commerce.
According to Hair et al. (2014), a good sample
size to use is 100 or higher. The minimum sample size
is at least five times the number of indicators. Still, it
is more acceptable if the sample size can be ten times
the number of indicators, and it is more advisable if
the sample size is twenty times the number of hands.
In this study, the four variables had a total of
12 indicators so that when multiplied by 20, it became
240 respondents who were considered sufficient to
represent the population.
3 RESULT AND DISCUSSION
3.1 Trust
Siagian (2014). Trust is a belief held by the first party
with the intention and behavior shown by the other
party. This concept means that e-commerce can
ensure the safety of customers who make online
purchases. According to Priansa (2017), According
to Kotler & Keller (2016), the following is a list of
trust indicators, namely:
1. Benevolence
2. Ability
3. Integrity
3.2 Celebrity Endorser
According to Savitri (2017), celebrity endorser
services make an impression on customers and
encourage a selective mindset in choosing. So it can
be concluded that celebrity endorsers choose artists to
star in advertisements using certain media to provide
information and influence listeners to buy advertised
goods or services.
According to Michael and Belch (2004),
indicators that affect celebrity endorser are:
1. Credibility
2. Attractiveness
3. Power
3.3 Shopee Loyalty Profile
Shopee loyalty profile is a program created by the
Shopee company to distinguish between loyal users
who have a high frequency of online shopping and
users who do not routinely shop at Shopee e-
commerce. In other words, a program that measures
Shopee user loyalty. Shopee loyalty profile has
several levels that users can obtain, ranging from
classic, silver, and gold, and the highest level is
platinum. The higher the level, the more bonuses you
get. According to Lucas and Britt (2012), indicators
owned by loyal customers include:
1. Attention
2. Interest
3. Action (Tindakan)
After feeling interested, consumers join the
program.
3.4 Purchase Decision
As defined Kotler & Armstrong (2014), that is the
buyer's decision-making process phase as long as the
customer purchases goods or services. Choice of
quality products, brand selection, supplier selection,
time of purchase, payment method, and some
purchases. It is a dimension of purchasing decisions
(Kotler & Armstrong 2016). Factors affecting
purchasing decisions can help customers choose
products from the various available options. Kotler
(2014) states that there are indicators that determine
the influence of purchasing decisions, namely:
1. Stability in a product
2. The habit of buying a product
3. Speed in buying a product
3.5 Validity Test
The validity test, according to Sudana & Setianto
(2018), is a research strategy to conduct experiments
on the issue of the truth of a causal or causal
relationship to the validity of a causal relationship
which is then generalized to the outside. According to
Sugiyono (2014), if > 0.3, it is accepted as valid, but
if it is less than 0.3, it is rejected.
The Influence of Trust and Celebrity Endorser on Purchase Decisions in Shopee E-Commerce Through Shopee Loyalty Profile
689
Table 1: Result the Validity of the Trust Variable (X1)2.
Statement
Correlation
Coefficient
R
critical
Desc.
trust 1 ,501 ,300 Accepted
trust 2 ,478 ,300 Accepted
trust 3 ,503 ,300 Accepted
trust 4 ,461 ,300 Accepted
trust 5 ,612 ,300 Accepted
trust 6 ,479 ,300 Accepted
trust 7 ,619 ,300 Accepted
trust 8 ,601 ,300 Accepted
trust 9 ,588 ,300 Accepted
Based on table 1, which summarizes the results of the
validity test of the confidence variable, it can be
shown that nine items show a correlation coefficient
more than the R Critical 0.3 so that it is declared
acceptable/valid.
Table 2: Result the Validity of the Celebrity Endorser
Variable (X2).
Statement
Correlation
Coefficient
R
critical
Desc.
Celebrity Endorser 1 ,886 ,300
Accepted
Celebrity Endorser 2 ,926 ,300
Accepted
Celebrity Endorser 3 ,904 ,300
Accepted
Based on table 2 which summarizes the results of the
celebrity endorser test, it can be shown that the three
statement items have a correlation coefficient more
than R Critical 0.3 so that they are declared
accepted/valid.
Table 3: Result the Validity of the Trust Purchase Decision
Variable (Y).
Statement
Correlation
Coefficient
R
critical
Desc.
Purchase Decision 1 ,577 ,300
Accepted
Purchase Decision 2 ,596 ,300 Accepted
Purchase Decision 3 ,707 ,300 Accepted
Purchase Decision 4 ,708 ,300 Accepted
Purchase Decision 5 ,697 ,300 Accepted
Purchase Decision 6 ,655 ,300 Accepted
Based on table 3 which summarizes the results of the
Purchase Decision test, it can be shown that the six
statement items have a correlation coefficient more
than R Critical 0.3 so that they are declared
accepted/valid.
Table 4: Result the Validity of the Shopee Loyalty Profile
(Z) Variable (X1).
Statement
Correlation
Coefficient
R
critical
Desc.
Shopee loyalty
profile Z
1,000 ,300 Accepted
Based on table 4 which summarizes the results of the
Purchase Decision test, it can be shown that the
statement item have a correlation coefficient more
than R Critical 0.3 so that they are declared
accepted/valid.
3.6 Reliability Test
According to Sekaran & Bougie (2017), an
instrument is declared reliable if the reliability
coefficient > 0.6 because it is considered to have a
safe point determines reliability.
Table 5: Reliable Test Results.
Variable Cronbach’s
Alpha
Composite
Reliability
Status
Trust (X1) ,686 > ,60 Accepted
Celebrity
Endorser (X2)
,889 > ,60 Accepted
Purchase
Decision (Y)
,725 > ,60 Accepted
Shopee Loyalty
Profile (Z)
1,000 > ,60 Accepted
Based on table 5, all statements are reliable because
the reliability coefficient of Cronbach's Alpha shows
results > 0.60.
3.7 Classical Assumption Test
3.7.1 Normality Test
The normality test is used to determine whether the
dependent variable and the independent variable in a
regression model are normally distributed. The test in
this study was carried out using normal probability
plot analysis. If the data spreads towards the diagonal
line, the data is considered to be normally distributed.
For example, using SPSS version 26, the P Plot
normality test produces the following image:
Figure 1: Normality Test P Plot.
Figure 1 shows the results of the P-plot normality test,
and it is clear that the regression model passes
iCAST-ES 2022 - International Conference on Applied Science and Technology on Engineering Science
690
because the points converge and follow the diagonal
line.
3.7.2 Heteroscedasticity Test
It can be seen through the value of Sig to evaluate
whether there is a heteroscedasticity problem. (2-
tailed) Rank Spearman. If it shows a number smaller
than 0.05, it can be concluded that heteroscedasticity
has a problem. Here are the results of the Spearman
Rank heteroscedasticity test conducted with SPSS
version 26:
Table 6: Spearman Rank Heteroscedasticity Test.
Correlations
Trust
Celebrity
Endorser
Shopee
L
oyalt
y
Profile
U
nstandardize
d
Residual
S
p
e
ar
m
an
's
rh
o
Trust
Correlation
Coefficient
1,000 ,331
**
,059 ,034
Sig. (2-tailed)
, ,000 ,367 ,597
N
240 240 240 240
Celebrity
Endorser
Correlation
Coefficient
,331
**
1,000 -,081 -,005
Sig. (2-taiIed)
,000 , ,212 ,935
N
240 240 240 240
Shopee
Loyalty
Profile
CorreIation
Coefficient
,059 -,081 1,000 -,038
Sig. (2-taiIed)
,367 ,212 , ,555
N
240 240 240 240
Unstandar
d
ized
Residual
CorreIation
Coefficient
,034 -,005 -,038 1,000
Sig. (2-taiIed)
,597 ,935 ,555 ,
N
240 240 240 240
**. CorreIation is si
g
nificant at the 0.01 level
(
2-tailed
)
.
The significance value of Sig. (2-tailed) in table 7 the
confidence variable is 0.597 which means 0.597 >
0.05. The celebrity variable shows several 0.935
which means 0.935 > 0.05. Meanwhile, the Shopee
loyalty profile variable shows a 0.555 > 0.05. So it
can be said that there is no heteroscedasticity in the
data in this study.
3.7.3 Multicollinearity Test
If the tolerance value is more than 0.10 and the value
also shows the VIF number is smaller than <10.00.
then the findings of each variable in the
multicollinearity test can be said to be free from
multicollinear symptoms. The results of the
multicollinearity test carried out with SPSS version
26 are listed below:
Table 7: Tolerance and VIF Test Results.
Coefficients
a
ModeI
Unstandardized
Coefficients
S
tandardize
d
C
oefficient
s
t Sig.
C
ollinearity Statistic
s
B
S
td. Erro
r
Beta Tolerance VlF
(Constant) 7,625 1,527 4,993 ,000
Trust ,261 ,049 ,303 5,307 ,000 ,901 1,109
Celebrity
Endorser
,403 ,065 ,354 6,200 ,000 ,897 1,114
S
hopee Loyalt
y
Profile
,511 ,153 ,183 3,335 ,001 ,976 1,024
a. Dependent Variable: Purchase Decision
The tolerance value for the confidence variable is
0.901, which means 0.901 > 0.10. VIF value shows
the number 1.109, which means 1.109 < 10.00. In the
celebrity endorser variable, the tolerance value is
0.897, which means 0.897 > 0.10. VIF value shows
the number 1.114, which means 1.114 < 10.00. In the
Shopee loyalty profile variable, the tolerance value is
0.976, which means 0.976 > 0.10. VIF value shows
the number 1.024, which means 1.024 < 10.00.
Therefore, it can be said that multicollinearity
symptoms do not exist in this study.
3.7.4 Linearity Test
In this study, the linearity test uses curve estimation.
Namely, the model is said to be linear if Sig. < 0.05,
then it is declared linear.
Here are the results of testing the linearity of the
relationship between the trust variable (X1) and the
purchasing decision variable (Y) using SPSS Version
26.
Table 8: Test Results for Linearity Curve Estimation of
Trust Variables.
Model Summary and Parameter Estimates
Dependent Variable: Purchase Decision
Equ
atio
n
Model Summary Parameter Estimates
R
Squar
e
F Df I df2 Sig. Constant bI b2 b3
Lin
ea
r
,177 51,337 1 238 ,000 9,350 ,363
Lo
gar
ith
mi
c
,176 50,940 1 238 ,000 -18,035 11,264
Inv
ers
e
,174 50,092 1 238 ,000 31,829
3
45,05
4
Qu
adr
ati
c
,178 25,575 2 237 ,000 11,830 ,203 ,003
The Influence of Trust and Celebrity Endorser on Purchase Decisions in Shopee E-Commerce Through Shopee Loyalty Profile
691
Equ
atio
n
Model Summary Parameter Estimates
R
Squar
e
F Df I df2 Sig. Constant bI b2 b3
Cu
bic
,178 25,575 2 237 ,000 11,830 ,203 ,003 ,000
Co
mp
ou
n
d
,167 47,785 1 238 ,000 11,823 1,018
Po
we
r
,166 47,527 1 238 ,000 3,109 ,549
S ,165 46,865 1 238 ,000 3,566
-
16,84
3
Gr
ow
th
,167 47,785 1 238 ,000 2,470 ,018
Ex
p
o
ne
nti
al
,167 47,785 1 238 ,000 11,823 ,018
Tze independent variable is trust.
From table 9, it can be seen in the linearity test using
curve estimation that the Sig value is obtained.
Between the variables of trust and purchasing
decisions, 0.000 <0.05. Therefore there is a linear
relationship between the variable (X1) trust and the
variable (Y) purchasing decisions.
Table 9: Test Results for Linearity Curve Estimation of
Celebrity Endorser Variables.
ModeI Summary and Parameter Estimates
Dependent VariabIe: Celebrity Endorser
Equatio
n
ModeI Summary Parameter EstImates
R
S
quare
F
D
f I df2 Sig.
C
onsta
n
t
bI b2 b3
Linear ,180
5
2,274 1 238 ,000 1,496 ,373
Logarith
mic
,177
5
1,028 1 238 ,000
-
1
3,016
7,354
Inverse ,171
4
9,133 1 238 ,000
1
6,152
-
1
41,42
3
Quadrati
c
,182
2
6,422 2 237 ,000 7,437 -,224 ,015
Cubic ,183
2
6,483 2 237 ,000 5,910 ,040 ,000 ,000
Compou
n
d
,149
4
1,688 1 238 ,000 3,775 1,043
Powe
r
,147
4
1,172 1 238 ,000 ,738 ,825
S ,144
4
0,159 1 238 ,000 2,973 15,949
Growth ,149
4
1,688 1 238 ,000 1,328 ,042
Exponen
tial
,149
4
1,688 1 238 ,000 3,775 ,042
The independent variable is Purchase Decisions.
In the linearity test using curve estimation, it can be
seen that the value between the celebrity endorser
variable and the purchase decision is 0.000 <0.05.
The result is a linear relationship between the
celebrity endorser variable (X2) and the purchase
decision variable (Y).
Table 10: Test Results for Linearity Curve Estimation of
Shopee loyalty profile Variables.
ModeI Summary and Parameter Estimates
Dependent VariabIe: Purchase Decisions
Equation
Model Summary
Parameter
Estimates
R
Square
F dfI df2 Sig.
Consta
n
t
bI b2 b3
Linea
r
,029
7
,090 1 238 ,008 19,794 ,476
Logarit
hmic
,035
8
,546 1 238 ,004 20,085 1,067
Inverse ,038
9
,294 1 238 ,003 21,894 -1,930
Quadrat
ic
,037
4
,527 2 237 ,012 18,553 1,693 -,259
Cubic ,038 3,11
4
3 236 ,027 17,300 3,630
-
1,12
3
,117
Compo
un
d
,029
7
,071 1 238 ,008 19,635 1,024
Powe
r
,034
8
,431 1 238 ,004 19,927 ,053
S ,037
9
,115 1 238 ,003 3,082 -,096
Growth ,029
7
,071 1 238 ,008 2,977 ,024
Expone
ntial
,029
7
,071 1 238 ,008 19,635 ,024
The independent variable is Shopee Loyalty Profile.
From table 11, it can be seen that the linearity test
using Curve Estimation obtained the value of Sig.
Between the variables Shopee loyalty profile and
purchasing decisions 0.008 <0.05. Therefore, the
relationship between the Shopee loyalty profile
variable (Z) and the purchasing decision variable (Y)
is linear. This shows that the data used is linear and
further studies can be conducted.
3.7.5 Multiple Regression Analysis
Multiple regression analysis is used to predict the
value of the dependent variable, namely trust and
celebrity endorser, if the dependent variable, namely
purchasing decisions, increases and decreases. The
following are the results of multiple regression
analysis using SPSS version 26:
Table 11: Multiple Regression Test Results.
Coefficients
a
Model
Unstandardized
Coefficients
Standardize
d
Coefficients
t Sig. B
Std.
Erro
r
Beta
1 (Constant) 8,530 1,534 5,559 ,000
Trus
t
,279 ,050 ,324 5,612 ,000
Celebrity
Endorse
r
,374 ,066 ,329 5,689 ,000
a. Dependent Variable: Purchase Decision
iCAST-ES 2022 - International Conference on Applied Science and Technology on Engineering Science
692
It can be seen in table 12 that the value of the constant
is 8.530, the coefficient value of the trust variable is
0.279, and the celebrity endorser variable is 0.374 to
produce a multiple linear regression test equation,
namely:
Y = 8,530 + 0,279X1 + 0,374X2 + e (1)
3.7.6 Partial Test (T Test)
This test is intended to determine whether the
independent variables in the regression model,
namely trust and celebrity endorser, affect the
dependent variable on purchasing decisions. The test
was carried out with a 95% confidence level. This
means that the X variable affects the Y variable if t
count > t table, but if the value of t count < t table,
then the X variable has no effect on the Y variable.
240 respondents are used in the formula to find the t
table, namely:
t
table
= t (α/2;n-k-1) = t (0,05;238) = 1,970 (2)
The following is a test using SPSS Version 26:
Table 12: Results of t-test of Hypothesis 1.
Coefficients
a
ModeI
Unstandardlzed
Coefficients
Standardl
zed
Coefficie
nts
t Sig.
B
Std. Erro
r
Beta
1 (Constant) 9,350 1,625 5,753 ,000
Trus
t
,363 ,051 ,421 7,165 ,000
With a regression coefficient of 0.363. H1 is
accepted. Namely, trust influences purchasing
decisions on e-commerce Shopee by the partial
regression coefficient test results in table 13, namely
t
count
7.165 > t
table
1.970.
Table 13: Results of t-test of Hypothesis 2.
Coefficients
a
ModeI
Unstandardlzed
Coefficients
Standardlzed
Coefficients
T Sig.
B Std. Erro
r
Beta
1 (Constant) 16,454 ,637 25,816 ,000
Celebrity
Endorse
r
,483 ,067 ,424 7,230 ,000
a. Dependent Variable: Keputusan PembeIian
With a regression coefficient of 0,483. H2 is
accepted. Namely, celebrity endorser influences
purchasing decisions on e-commerce Shopee by the
partial regression coefficient test results in table 14,
namely t
count
7,230 > t
table
1.970.
3.7.7 Statistical Test (F Test)
If F
count
> F
table
, then the independent variable affects
the dependent variable. 240 respondents were used to
find the F
table
formula, namely:
F
table
= F (k; n-k) = F (2;239) = 3,033 (3)
Using SPSS Version 26, the F test was assessed as
follows:
Table 14: Results of t-test of Hypothesis 3.
ANOVA
a
ModeI Sum of Squares Df
Mean
Square
F SIg.
1
Regression 377,749 2 188,875 45,231 ,000
b
ResiduaI 989,651 237 4,176
Total 1367,400 239
a. Dependent VariabIe: Purchase Decision
b
. Predictors: (Constant), CeIebrity Endorser, trust
The calculated F
count
is 45.231 > F
table
3.033 which
shows that H3 is accepted which means that trust and
celebrity endorsers have an influence on purchasing
decisions on Shopee e-commerce according to the
results of table 15 of the F statistical test
3.7.8 Uji Moderated Regression Analysis
(MRA)
Moderated Regression Analysis (MRA) can also be
called the interaction test between the independent
variable and the moderator variable, which aims to
see whether the moderator variable can strengthen or
weaken the influence of the independent variable on
the dependent variable. Here are the results of the first
regression output of the confidence variable.
Table 15: Model Summary Regression I Trust Variable.
Model Summary
b
Model R
R
Square
Adjusted R
Square
Std. Error of
the Estimate
Durbin-
Watson
1 ,421
a
,177 ,174 2,174 1,952
a. Predictors: (Constant), Trus
t
b
. Dependent Variable: Purchase Decision
In table 16, the summary model can be seen that the
R square value shows a number of 0.177.
Table 16: Anova Regression I Trust Variable.
ANOVA
a
Model
Sum of
Squares
df
Mean
Square
F Sig.
1 Regression 242,617 1 242,617 51,337 ,000
b
Residual 1124,783 238 4,726
Total 1367,400 239
a. Dependent Variable: Purchase Decision
b
. Predictors: (Constant), Trust
The Influence of Trust and Celebrity Endorser on Purchase Decisions in Shopee E-Commerce Through Shopee Loyalty Profile
693
In table 17, ANOVA can be seen the value of F,
which shows a number of 51.337.
Table 17: Regression Coefficients II Trust Variable.
Coefficients
a
Model
Unstandardized
Coefficients
Standardized
Coefficients
t Sig.
B Std. Erro
r
Beta
1
(
Constant
)
16,454 ,637 25,816 ,000
Celebrity
Endorse
r
,483 ,067 ,424 7,230 ,000
a. Dependent Variable: Purchase Decision
In table 18, coefficients can be seen the value of Sig.
which shows a number of 0.000.
The following are the results of the regression
output of the two trust variables
Table 18: Model Summary Regression II Trust Variable.
Model Summary
b
Model R
R
Square
Adjusted R
Square
Std. Error of the
Estimate
Durbin-
Watson
1 ,444
a
,197 ,187 2,157 1,942
a. Predictors: (Constant), Trust*Shopee Loyalty Profile, Trust,
Shopee Loyalty Profile
b
. Dependent Variable: Purchase Decision
In table 19, the summary model can be seen that the
R square value shows a number of 0.197.
Table 19: Anova Regression II Trust Variable.
ANOVA
a
Model
Sum of
Squares df
Mean
Square F Sig.
1 Regression 269,748 3 89,916 19,332 ,000
b
Residual 1097,652 236 4,651
Total 1367,400 239
a. Dependent Variable: Purchase Decision
b. Predictors: (Constant), Trust*Shopee Loyalty Profile, Trust,
Shopee Loyalty Profile
In table 20 ANOVA can be seen the value of F, which
shows a number of 19.332
.
Table 20: Second Regression Output Trust Variable.
Coefficients
a
Model
Unstandardized
Coefficients
Standardized
Coefficients
t Sig.
B Std. Erro
r
Beta
1
(Constant) 6,390 4,507 1,418 ,158
Trus
t
,427 ,141 ,495 3,017 ,003
Shopee Loyalty
Profile
1,392 1,826 ,498 ,762 ,447
Trust*Shopee
Loyalty Profile
-,032 ,057 -,378 -,554 ,580
a. Dependent Variable: Purchase Decision
In table 21, it can be seen that the constant value
shows a number of 6.390 while the value of the
confidence coefficient (X1) is 0.427. The coefficient
value for the Shopee loyalty profile is 1.392 while the
coefficient value for the trust* Shopee loyalty profile
variable is -0.032 so that the regression test results
obtain the following equation:
Y = 6,390 + 0,427 X1 + 1,392 Z – 0,032 X1Z + e (4)
The comparison of R2 in the first regression is 0.177.
At the same time, after the moderator variable, the
value of R2 becomes 0.197, so it can be concluded
that H4 is accepted. Namely, the Shopee loyalty
profile strengthens the influence of trust on
purchasing decisions on Shopee e-commerce.
In table 20 the second regression ANOVA is
obtained F
count
of 19.332 > F
table
3.033, which means
that the variables of trust, Shopee loyalty profile, and
trust*Shopee loyalty profile simultaneously affect
purchasing decisions.
Table 21: Model Summary Regression I Variable Celebrity
Endorser.
Model Summary
b
Model R
R
Square
Adjusted
R Square
Std. Error
of the
Estimate
Durbin-
Watson
1 ,424
a
,180 ,177 2,170 1,885
a. Predictors: (Constant), Celebrity Endorser
b. Dependent Variable: Purchase Decision
In table 21, the summary model shows that the R
square value shows a number of 0.180.
Table 22: Anova Regression I Variable Celebrity Endorser.
ANOVA
a
Model
Sum of
Squares
df
Mean
Square
F Sig.
1 Regression 246,250 1 246,250 52,274 ,000
b
Residual 1121,150 238 4,711
Total 1367,400 239
a. Dependent Variable: Purchase Decision
b. Predictors: (Constant), Celebrity Endorser
In table 22, ANOVA can be seen the value of F,
which shows a number of 52.274.
iCAST-ES 2022 - International Conference on Applied Science and Technology on Engineering Science
694
Table 23: Regression Coefficients II Variable Celebrity
Endorser.
Coefficients
a
Model
Unstandardized
Coefficients
Standardized
Coefficients
t Sig.
B Std. Error Beta
1 (Constant) 16,454 ,637 25,816 ,000
Celebrity
Endorse
r
,483 ,067 ,424 7,230 ,000
a. Dependent Variable: Purchase Decision
In table 23 coefficients, it can be seen that the Sig
value shows a number of 0.000.
Here are the results of the second regression
output of the celebrity endorser variables.
Table 24: Model Summary Regression II Variable
Celebrity Endorser.
Model Summary
b
Model R
R
Square
Adjusted
R Square
Std. Error of
the Estimate
Durbin-
Watson
1 ,480
a
,231 ,221 2,111 1,878
a. Predictors: (Constant), CelebrityEndorser*Shopee Loyalty
Profile, Celebrity Endorser, Shopee Loyalty Profile
b. Dependent Variable: Purchase Decision
In table 24, the summary model can be seen that the
R square value shows a number of 0.231.
Table 25: ANOVA Regression II Variable Celebrity
Endorser.
,753ANOVA
a
Model
Sum of
Squares
df
Mean
Square
F Sig.
1 Regression 315,597 3 105,199 23,604 ,000
b
Residual 1051,803 236 4,457
Total 1367,400 239
a. Dependent Variable: Purchase Decision
b. Predictors: (Constant), CelebrityEndorser*Shopee Loyalty
Profile, Celebrity Endorser, Shopee Loyalty Profile
In table 25 ANOVA, it can be seen that the F value
shows a number of 23,604.
Table 26: Celebrity Endorser Variable Second Regression
Output.
Coefficients
a
Model
Unstandardized
Coefficients
Standardized
Coefficients
t Sig.
B Std. Erro
r
Beta
1
(Constant) 12,434 2,120 5,866 ,000
Celebrity Endorser ,754 ,219 ,662 3,440 ,001
Shopee Loyalty
Profile
1,488 ,771 ,532 1,930 ,055
CelebrityEndorser
*Shopee Loyalty
Profile
-,094 ,080 -,370
-
1,172
,243
a. Dependent Variable: Purchase Decision
In table 26, it can be seen that the constant value
shows a number of 12.434 while the celebrity
endorser coefficient value (X2) is 0.754. The
coefficient value for the Shopee loyalty profile is
1.488, while the coefficient value for the celebrity
endorser*Shopee loyalty profile variable is -0.094, so
that the regression test results obtained the following
equation
Y = 12,434 + 0,754 X2 + 1,488 Z – 0,094 X2Z + e (5)
The comparison of R
2
in the first regression is 0.180,
while after the moderator variable, the value of
R
2
becomes 0.231, so it can be concluded that H5 is
accepted. Namely, the Shopee loyalty profile
strengthens the influence of celebrity endorsers on
purchasing decisions on Shopee e-commerce.
In the second ANOVA regression table, F
count
is
23,604 > F
table
3,033, which means that the celebrity
endorser, Shopee loyalty profile, and celebrity
endorser*Shopee loyalty profile variables
simultaneously affect purchasing decisions.
3.7.9 Coefficient of Determination Test
(Adjusted R²)
For purchase decisions with or without a moderator,
the Shopee loyalty profile is an analysis of
determination used to determine how much influence
the independent variable has on the dependent
variable. The correlation coefficient is multiplied by
100%, and the result is the coefficient of
determination.
The results of testing the coefficient of
determination of substructure 1 before moderation
using SPSS Version 26 are as follows:
Table 27: Substructure I Coefficient Determination Test
Results Before Moderation.
Model Summary
Model R R Square
Adjusted R
Square Std. Error of the Estimate
1 ,526
a
,276 ,270 2,043
a. Predictors: (Constant), Celebrity Endorser, Trust
Based on table 27, R Square shows the number 0.276,
so:
r
2
x 100% = 0,276 x 100% = 27,6% (6)
R
2
, known as the coefficient of determination, has a
value of 0.276. This shows that the purchase decision
(dependent variable) can be explained by trust and
celebrity endorser (independent variable) by 27.6%.
In contrast, the remaining 72.4 is explained by other
factors not included in this study.
The Influence of Trust and Celebrity Endorser on Purchase Decisions in Shopee E-Commerce Through Shopee Loyalty Profile
695
Table 28: Substructure Determination Coefficient 2 After
Moderation.
Model Summary
Model R R Square
Adjusted R
Square
Std. Error of the Estimate
1 .558
a
,311 .296 2,006
a. Predictors: (Constant), CelebrityEndorser*Shopee Loyalty
Profile, Trust, Celebrity Endorser, Shopee Loyalty Profile,
Trust*Shopee Loyalty Profile
Based on table 28, R Square shows the number 0.311,
so:
r
2
x 100% = 0,311 x 100% = 31,1%
(7)
This shows that after the moderating variable, namely
Shopee Loyalty Profile R Square, it increased to
31.1%, which was previously 27.6%, so it can be
concluded that the moderating variable can
strengthen the relationship between the independent
variable and the dependent variable. Likewise, with
the value of Std. An error of the Estimate, which was
previously 2,043, decreased to 2,006 since the
moderating variable.
Discussion
a) The Effect of Trust (X1) on Purchase Decisions
(Y)
Based on the results of the partial test, the results
of t
count
7.165 > t
table
1.970 as shown in table 13
so that it is stated that H1 is accepted, namely
trust (X1) affects purchasing decisions (Y).
b) The Effect of Celebrity Endorser (X2) on
Purchase Decision
Judging from the results of the partial test, the
results of t count t count 7,230 > t table 1,970 as
shown in table 14, so it is stated that H2 is
accepted, which means celebrity endorser (X2)
affects purchasing decisions (Y).
c) The Effect of Trust (X1) and Celebrity Endorser
(X2) on Purchase Decisions (Y)
The F test is used to test the hypothesis that trust
(X1) and celebrity endorser (X2) affect
purchasing decisions (Y). celebrity endorser
(X2) influences purchasing decisions (Y) on e-
commerce Shopee.
d) Shopee Loyalty Profile (Z) Strengthens the
Effect of Trust (X1) on Purchase Decisions (Y)
To test the hypothesis, which states that the
Shopee loyalty profile strengthens the influence
of trust on purchasing decisions, a Moderated
Regression Analysis (MRA) test is carrifed out.
e) Shopee Loyalty Profile (Z) Strengthens the
Effect of Celebrity Endorsers (X2) on Purchase
Decisions (Y)
To test the hypothesis, which states that the
Shopee loyalty profile strengthens the influence
of trust on purchasing decisions, a Moderated
Regression Analysis (MRA) test is carried out.
4 CONCLUSION
Through Shopee loyalty profile (Z), this study intends
to determine the impact of trust (X1) and celebrity
endorser (X2) on purchasing decisions. This study
used 240 respondents by distributing a google form
questionnaire. Data were collected and tested using
SPSS version 26 software so that the following
conclusions can be drawn:
a) Trust (X1) has an influence on purchasing
decisions (Y) on Shopee e-commerce.
b) Celebrity endorsers (X2) have an influence on
purchasing decisions (Y) on Shopee e-
commerce.
c) Celebrity endorsers (X2) have an influence on
purchasing decisions (Y) on Shopee e-
commerce.
d) Shopee loyalty profile (Z) strengthens the
influence of trust (X1) on purchasing decisions
(Y) in Shopee e-commerce.
e) Shopee loyalty profile (Z) strengthens the
influence of celebrity endorser (X2) on
purchasing decisions (Y) in Shopee e-commerce.
5 SUGGESTION
5.1 Theoretical Suggestions
In this study, researchers used only two independent
variables, trust and celebrity endorser, one
moderating variable, namely Shopee loyalty profile,
and one dependent variable, purchasing decisions. So
that companies and readers get more mixed results
and can inform what influences buying decisions on
Shopee e-commerce.
5.2 Practical Suggestions
To improve purchasing decisions in the community,
Shopee must maintain good trust, especially in
assessing the statement with the lowest value:
"Shopee provides a 100% original guarantee feature
to consumers when buying products. -Receive the
original item. Shopee must also make the buyer's
warranty claim easier if the item received is proven
not to be original.
iCAST-ES 2022 - International Conference on Applied Science and Technology on Engineering Science
696
REFERENCES
Anshar, S. B., Suryaningsih, I. B., & Sumani. (2016).
"Pengaruh SeIebriti Endorser Dan Harga Terhadap
Keputusan Pembelian Online Melalui Moderasi
Gambar Produk Di Instagram Pada Mahasiswi
Universitas Jember." Artikel Ilmiah Mahasiswa.
Ayuning, S. S. (2019). "Pengaruh Promosi Penjualan
Terhadap lmpulse Buying Produk Pada ApIikasi
Shopee Indonesia." Jurnal Ilmu Komunikasi , 79.
Hair, J. F. (2014). "Multivariate Data Analysis 7th
edition)." New Jersey: Pearson.
Karmila, A. (2020). "Pengaruh Harga Dan Kepercayaan
Konsumen Terhadap Keputusan Pembelian MeIaIui
Marketplace Shopee." Jurnal Manajemen.
Kotler, P., & Amstrong, G. (2014). "PrincipIes of Marketing,
12th Edition, Jilid 1." Jakarta : Erlangga.
Kotler, P., & Amstrong, G. (2014). "PrincipIes of Marketing,
15th Edition". England: Pearson Education Limited.
Kotler, P., & Keller, K. L. (2016). "Marketing Management
15th Edition." England : Pearson.
Kurniawati, E. (2022). "Pengaruh Kepercayaan Dan
CeIebrity Endorser Terhadap Keputusan Pembelian
(Studi Pada Konsumen Lazada.co.id Di Kota
Semarang)." Jurnal Ilmiah Bidang Ilmu Ekono.
Lucas, D. B., & Britt, S. H. (2012). Measuring Advertising
Effectiveness: "McGraw HiII Series ln Marketing And
Advertising Paperback." New York: Literary
Licensing, LLC.
Michael, B., & Belch, G. E. (2004). "An ltegrated
Marketing Communication Perspective Edisi 6." New
York: McGraw-Hill .
Priansa, D. J. (2017). "PeriIaku Organisasi Bisnis."
Bandung : Alfabeta.
PT Shopee Indonesia. (2020, Maret 03). "Apa itu Shopee
LoyaIty." Retrieved from "help.shopee. co.id:
https://help.shopee.co.id/s/article/Apa-itu-Shopee-
Loyalty"
Savitri, N. W. (2017). "Pengaruh CeIebrity Endorser dan
Iklan Melalui Media TeIevisi Terhadap Minat Beli pada
Tokopedia di Denpasar." E-Jurnal Manajemen Unud,
Vol. 6, 3.
Shimp, T. A. (2014). "Komunikasi Pemasaran Terpadu
Edisi 8." Jakarta : SaIemba Empat.
Siagian, H., & Cahyono, E. (2014). "Analisis Website
QuaIity, Trust, Dan LoyaIty Pelanggan Online Shop."
Jurnal Manajemen Pemasaran, Vol, 8 No,2.
Sudana, M., & Setianto, R. H. (2018). "Metode PeneIitian
Bisnis & Analisis Data Dengan SPSS." Jakarta :
Erlangga.
Sugiyono. (2007). "Statistika untuk peneIitian." Jawa
Barat: Alfabeta.
Sugiyono. (2015). "Statistik Nonparametris Untuk
PeneIitian." Bandung : Alfabeta.
Sugiyono. (2018). "Metode PeneIitian Kombinasi (Mixed
Methods)." Bandung : Penerbit Alfabet.
Sugiyono. (2018). "Metode PeneIitian Kuantitatif."
Bandung : Alfabeta.
Waro, & Widowati. (2020). "Pengaruh Kepercayaan,
CeIebrity Endorser, dan Persepsi Risiko terhadap
Keputusan Pembelian Pengguna E-Commerce di Kota
Semarang." Prosiding Seminar Nasional Unimus.
Waro, S. S., & Widowati, M. (2020). "Pengaruh
Kepercayaan, CeIebrity Endorser, dan Persepsi Risiko
terhadap keputusan pembelian." Jurnal Prosiding
Seminar Nasional Unimus Vol. 3 , 2.
The Influence of Trust and Celebrity Endorser on Purchase Decisions in Shopee E-Commerce Through Shopee Loyalty Profile
697
