Postmodern Mindset and Behavior of Indonesia Consumer: Local

Brand Preference, Early Adopter, Consumer Credit, and Online

Buying based on Social Class

Anna Triwijayati

, Melany

and Dian Wijayanti

Departement of Management, Universitas Ma Chung, Villa Puncak Tidar N-10, Malang, Indonesia

Department of English Letters, Universitas Ma Chung, Villa Puncak Tidar N-10, Malang, Indonesia

Department of Accounting, Universitas Ma Chung, Villa Puncak Tidar N-10, Malang, Indonesia

Keywords: Consumer, Social Class, Local Brand Preference, Early Adopter, Online Buying

Abstract: Indonesian consumers’ mindset is formed by the development of marketing based on information and

communication technology. This article aimed at analyzing the differences in postmodern the mindset of

Indonesia Consumer about Local Brand Preference, Early Adopter, Open-mindedness to credit facilities,

and Online Buying based on social classes. This research is a descriptive study by 1000 samples of

Indonesian consumers from three social classes, which are upper, middle, and lower social classes. Kruskal

Wallis, Anova Test, and Discriminant were used for data analysis, and the results showed that there were

differences in local brand preferences in the social class of Indonesian consumers, where the lower social

class had higher preferences for the local brand. As for in other results, there is no difference between social

classes based on Early Adopter characteristics, although the data description showed that the upper class

tends to be an early adopter segment. Related to Open-mindedness for credit facilities, the results of the

study showed that there is no difference in the social class of Indonesian consumers. The test results on the

Online buying and Thinking mindset variables indicate that there are differences in preferences in the

Online buying and Thinking mindset of consumption in the social class of Indonesian consumers with

groups that showed the highest average differences are the middle class and lower class. The discriminant

equation in all three social classes shows that the Thinking mindset variable has the biggest influence on the

three social classes. When it was seen the tendency in the three social classes, the variable Open-

mindedness to credit facilities is stronger in the lower social class; Online Buying is stronger in the upper

and middle classes, and conservative thinking mindset is stronger in the lower classes.

1 INTRODUCTION

Consumer behavior in today’s digital era has

experienced many changes (OECD, 2010). There are

two major conditions driving this change. The

internet has changed the moment when consumers

decide to buy a particular product (Marie and Grybś,

2013). Dynamic and continuous technological

changes produce new products and services, jobs

and business systems, new lifestyles and

interpersonal communication, and are not only a

driving force for cultural change but at the same

time are subject to change itself (Grubor and Marić,

2015). The development of communication and

information technology and the rise of digital media

brought a new wave of consumerism (OECD, 2010;

Eroğlu, 2014). Second, there is a change in socio-

economic conditions (OECD, 2010), where many

countries experience changes in business

competition, energy, financial services,

telecommunication, and transportation industries.

Consumers have migrated from old model industries

to new era industries, from mass-produced products

to customized ones and from obsolete products to

innovative products (Czarniewski, 2014).

Rapid changes in information and

communication technology also have an impact on

shifts in the whole marketing of products and

services. In the era of modern technology,

consumers obtain a new 'ammunition' as a producer

of a new online reality (Hamouda, 2012). The

impact of the internet on consumers will continue to

grow. In China, for example, 68% of the middle

Triwijayati, A., Melany, . and Wijayanti, D.

Postmodern Mindset and Behavior of Indonesia Consumer: Local Brand Preference, Early Adopter, Consumer Credit, and Online Buying based on Social Class.

DOI: 10.5220/0009959001270142

In Proceedings of the International Conference of Business, Economy, Entrepreneurship and Management (ICBEEM 2019), pages 127-142

ISBN: 978-989-758-471-8

127

class have access to the internet, compared to 57%

of the total urban population. The Indonesia Internet

Service Provider Association survey in 2018 showed

that internet users in Indonesia were 171.17 million

people or 64.8% of the total population of Indonesia.

Consumer behavior has now entered the era of

postmodern behavior. Postmodern, an era outside of

modernity (Singh, 2011), is becoming increasingly

plural in values and lifestyles (Bouagina and Triki,

2014). Many researchers argue that consumption in

this era is the most crucial representation of

postmodernism in contemporary life (Zhongqi et al.,

2016). Consumers' shopping habits have changed

over the past decade in which creativity and

productivity, as well as digital technology, are

integrated with shopping or consumption

experiences (Yasav, 2015; Firat, Kutucuoğlu, Saltik,

Tunçel, 2013).

Postmodernism offers an alternative to joining a

global consumption culture where commodities and

forms of knowledge are driven away from human

control (Singh, 2011). On the one hand, ethnic

diversity distinguishes products consumed, and on

the other hand, many societies will become part of

the globalization of consumption (Czarniewski,

2014). Changes in postmodernism consumer

behavior is an un-ending process (Deepak and

Harneet, 2017) that is adaptive, flexible, fragmented,

liberated, and symbolic (Bouagina and Triki, 2014).

Post-modern has a link with consumer behavior.

One of the characteristics of postmodern consumers

is that they no longer have traditional values of

society and are unpredictable in behavioral patterns

(Berner and Tonder, 2003). Nowadays, people tend

to be consumptive and are encouraged to continue

consuming, using, and discarding to live (Bati,

2008). Changes in paradigms in consumer behavior

increase the need to engage with consumers (Eroğlu,

2014) effectively.

Consumption is a primary social process, and

many functions, as well as motivations for

consumption, are derived from social (Schor, 2002).

Social class is also related to consumer behavior and

becomes the basis for understanding consumer

behavior (Shavitt, Duo, Hyewon, 2016). A set of

characteristics is found to be systematically different

in each social class in the psychological domain,

including norms and habits, abstract-level modes of

thought, the domain of behavior, and the domain of

physical influence (Iqbal and Ismail, 2011). Every

consumer desire in a social class is important for

marketers because the buying behavior in a social

class is the same (Durmaz, 2014) and differs

between classes (Iftikhar, Hussain, Kahn, and Iiyas,

2013).

Each social class places society according to their

values in society (Durmaz and Taşdemir, 2014).

Social class variables are useful for differentiating

consumers based on ethnocentrism values (Strehlau,

Ponchio, and Loebel, 2012). Consumer

ethnocentrism is more dominant in consumers who

like local products (Candan, Aydm, and Yamamoto,

2008). Developed country consumers, in general,

will judge their domestic products to be of higher

quality than other countries' products (Ping, Lobo

and Li, 2012).

Postmodern in this study is connected with Local

Brand Preference, Early Adopter, Open-mindedness

to credit facilities, Online Buying, and Thinking

Mindset based on Social Class. Why? Because

social class also influences where and how

consumers perceive purchases (Durmaz and

Taşdemir, 2014). Lower social class, for example,

likes local products: markets that allow face to face

interaction where they get friendly service and easy

credit, often in their neighborhood (Durmaz and

Taşdemir, 2014), so that they have high

ethnocentrism values (Strehlau, Ponchio, and

Loebel, 2012). When local products are available,

high ethnocentric consumers will show a preference

for local products (Siamagka, 2009).

Consumers migrate from obsolete products to

innovative products (Czarniewski, 2014). The

emergence of the adoption of a new product

behavior is the result of the interaction of several

psychological variables (Eroğlu, 2014). The research

of Ahmed, Khan, and Samad (2016) shows that a

prosperous Indonesian consumer class is an

attractive consumer group. They are ready to adopt

new products, look for quality, and are ready to pay

for it.

The adoption of digital-based e-commerce

encourages the fulfillment of challenges related to

technology, people, and money (Quigley, 2015).

Changes in communication and information

technology in banking services encourage

consumers to think positively about credit and view

credit as an indicator of economic growth. The level

and structure of consumer spending depend on

several factors, such as the level of savings, the

propensity to save, debt, and attitudes towards a

credit (Czarniewski, 2014). In the 2010 OECD

study, 73% of young Danish people did not and

were less concerned about the interest rates on their

loans, and 64% bought goods on credit and thought

that said loans had low-interest rates (OECD, 2010).

ICBEEM 2019 - International Conference on Business, Economy, Entrepreneurship and Management

128

Modern consumers ask for permanent access to

the internet and information at any place and at any

time (Marie and Grybś, 2013). Online consumers

now have more control and bargaining power than

physical stores consumers (Eroğlu, 2014). The net-

generation is now extremely aware of differences in

various cultures in the world and more critical of the

reality created by the media (Czarniewski, 2014).

The types or modes of thought tend to be the

impact of economic resource sharing. Modes of

thought or modern mindset are usually concentrated

in the upper social class because they describe

themselves as inquisitive and interested in new

things, seeking to broaden their minds (Iqbal and

Ismail, 2011). Global change encourages a deeper

understanding of the dynamics of cultural

differences (Deepak and Harneet, 2017).

Hypothesis:

Ha1: There are differences in Local brand preference

in the social class of Indonesian consumers

Ha2: There are differences in the early adopter

segment for new products in the Indonesian

consumer social class

Ha3: There are differences in Open-mindedness to

credit offers in Indonesian consumer social

classes

Ha4: There are differences in online buying

preferences in Indonesian consumer social

classes

Ha5: There are differences in thinking mindset of

consumption in the social class of Indonesian

consumers

Ha6: Openness to credit offers is related to

Indonesian consumer social class

Ha7: Preference for online buying is related to the

social class of Indonesian consumers

Ha8: the Thinking mindset of consumption is related

to the social class of Indonesian consumers

2 METHOD

This research is a descriptive study that explains the

relationship between social class and postmodern

mindset and behavior of Indonesia Consumer in

relation to Local Brand Preference, Early Adopter,

Open-mindedness to credit facilities, Online Buying,

and Thinking Mindset based on Social Class. The

selected respondents were Indonesian consumers

over 17 years old from 31 Provinces (out of 34

Provinces) and 42 ethnicities. Data was collected

through a survey method by distributing

questionnaires utilizing accidental sampling. The

number of respondents who gave responses was

1000 people, but there are missing values that are

not counted in the analysis.

The variables used are Postmodern mindset,

which is derived into sub-variables X1 Local Brand

Preference, X2 Early Adopter, X3 Open-mindedness

to credit facilities, X4 Online Buying and X5

Thinking Mindset and the dependent variables

namely three social classes (upper, middle, and

lower). The following table describes the operational

definition.

Table 1: Operational Definition

Var Postmodern

Sub Variables

Source Description Item in the

questionnaires

X1 Local Brand

Preference

Singh (2011);

Czarniewski (2014);

Berner and Tonder

(2003)

Strehlau, Ponchio,

and Loebel (2012)

Durmaz and

Taşdemir (2014)

Global consumption culture; No

more having traditional values of

society; social class is useful for

differentiating consumer based on

ethnocentrism value; consumers

like the local product, the

marketplace that allows face to

face with the trader/selle

How important is the

local Indonesian

brand for consumer

X2 Early Adopte

Bouagina and Triki

(2014);

Eroğlu (2014)

Czarniewski (2014)

adaptive, flexible, fragmented,

liberated and symbolic; showing

the adoption behavior of a new

product; from obsolete to

innovative produc

How important is

consumer be the first

buyer of new product

X3 Open-

mindedness to

Czarniewski (2014)

Durmaz and

Level and structure of expenditure

depend on some factors such as

How important is

ein

the consumer

Postmodern Mindset and Behavior of Indonesia Consumer: Local Brand Preference, Early Adopter, Consumer Credit, and Online Buying

based on Social Class

129

credit facilities Taşdemir (2014) level of savings, the propensity to

save, debt, and attitude toward

credit facilities credit; Marketplace

that offers easy credit, often in the

nei

hborhoo

that was having an

open mind to credit

facilities

X4 Online Buying Yasav (2015); Firat,

Kutucuoğlu, Saltik,

and Tunçel (2013)

Marie and Grybś

(2013)

Eroğlu (2014)

Digital technology;

Modern consumer ask permanent

access to internet and information;

Online consumers currently have

more control and bargaining power

than physical stores consumers

How important is

consumer believe in

purchasing via the

internet

X5 Thinking

Mindset

Iqbal and Ismail,

2011

Deepak and Harneet,

2017

Modern thought or mindset is

usually concentrated on social

class. Global change encourages a

deeper understanding of the

dynamics of cultural differences.

How deep do you

think and act

conservatively/

traditionally

Analysis of the data used is Kruskal Wallis,

Anova, and Discriminant Analysis. Kruskal Wallis

test is used to find out the X1 preference for local

brands and X2 early adopters in all three social

classes. ANOVA test is used to examine the

relationship between the three social classes (upper,

middle, and lower) with independent variables. For

the X5 Thinking Mindset variable specifically, the

questionnaire is trying to find out thinking which is

contrary to modern thinking or can be called

conservative/traditional thinking. Discriminant

analysis is used to look at grouping individuals

based on more than one independent variable. The

discriminant analysis aims to classify an individual

into mutually exclusive/disjoint and exhaustive

groups based on a number of explanatory variables.

3 RESULT

Referring to Table A Data of Respondents (see

Appendix), it is stated that 57% are the lower social

class, 41% middle social class, and 2% upper social

class. The grouping of social classes is based on

three factors, namely income, employment, and

education (Birkelund and Lemel, 2012), according

to the ISP formula (Mihic and Culina, 2006). The

number of female respondents was 55.2%, and the

male was 44.8%. Based on age data, respondent data

shows that the proportion of age is quite balanced

(around 22-29%) up to the age group <45 years,

while the age group> 46 years is 20.6%. Of marital

status, there are 29.6% husbands, 27.8% wives,

39.7% children, and the rest are other family

members. Data on the education level of the

respondents indicated that most of them were high

school graduates (58.9%). Income data shows that

50.4% of respondents earn Rp. 2,400,000-7,200,000

per month and 31.2% earn ≤ 2,400,000.

Respondent's employment distribution data shows

30.5% respondents are not working (including

housewife), 15.5% are students, 14.9% are

administrative employees, and 13.2% are middle

managers, small business owners, middle-level

government officials, middle-level professionals

(doctors, lawyers, lecturers, etc.), and middle-level

police/army officers. Most respondents came from

provinces on Java island (78%), and Javanese, as

well as Madurese, dominates the number of

respondents who filled out the questionnaire

(68.8%). Table 2 indicates that all of the sub-

variables of the postmodern mindset are valid and

reliable.

Table 2. Validity and Reliability

Sub

Variables

Post-

modern

Mindset

Mean

Validity

(subtotal

to total

corr.)

Reliability

(Cronbach.

Alpha)

X1 4.0437 0.591 0.863

X2 2.5584 0.665 0.860

X3 2.4446 0.730 0.855

X4 3.0608 0.584 0.863

X5 2.9675 0.396 0.872

The following results answer the research

hypothesis whether there is a relationship between

social class and postmodern mindset and behavior of

Indonesia Consumer in relation to Local Brand

Preference, Early Adopter, Open-mindedness to

credit facilities, Online Buying, and Thinking

Mindset based on Social Class. Based on Table B

ICBEEM 2019 - International Conference on Business, Economy, Entrepreneurship and Management

130

(Appendix), the mean value of X1 is 4.0437. Middle

social class has the lowest mean of 3.9431. The

lower social class has the highest mean value of X1,

which is 4,1161, which descriptively signifies that

the lower class prefers local brands.

X2 mean value = 2.5584 (less than 3.00 or even

4.00), means obtained with a slight difference

between the middle and lower classes, which were

2.5792 and 2.5365, respectively. In addition, the

upper class gets a mean of the highest value of

2.7500 in that category. In general, Indonesian

consumers lack the characteristics of early adopters.

Although the mean value of X2 is less than 3.00 or

so, descriptively, the upper classes tend to be early

adopters compared to other social classes.

The X3 obtained a mean value of 2.4446 (less

than the middle value of 3.00 or even 4.00). The

middle class has the lowest mean value with a value

of 2.3491, while the upper class has a mean of

2.6000, followed by the lower class with a value of

2.5071. Descriptively, Indonesian consumers show

less openness to facilities or access to purchase

goods on credit. However, the upper class has more

Open-mindedness towards credit facilities compared

to other social classes.

The calculation result of X4 shows that the lower

class has the lowest mean value of 2.9626. The

middle class and upper class respectively have a

mean value of 3.1852 and 3.3000. Descriptively, the

upper social class prefers online purchases.

Meanwhile, the X5 shows that the lower class

has the highest mean value of 3.0982. Middle and

upper classes, respectively, have a mean score of

2.7896 and 2.9000. This signifies that the lower

classes still tend to have conservative/traditional

thinking in consumption. These results are in line

with the results of variable X2, which shows that the

lower class is not an early adopter of new products.

After going through descriptive analysis, the data

were analyzed using the Anova Test. Before

continuing the test, one of Anova's assumptions is

the same variance. From Table C (Appendix-Test of

Homogeneity of Variance), it can be seen that the

test results show that the five variants of the

variables are not the same (see Appendix). The score

of the p-value for variable X1 and X2 = 0,000 is

smaller than 0.05 or reject H0 = 0; the meaning is

that there are different variants in the three groups

on the variables X1 and X2 so that the Anova test is

invalid to test this relationship. Consequently, the

Kruskal Wallis test is used to find out the and X2 in

all three social classes. The score of p-value for

variable X3=0.089, X4 (p-value=0.189), X5 (p-

value=0.985) is> 0.05; which indicates that all three

variants are the same. Therefore, the Anova test is

valid for testing this relationship.

Table 3 describes the result of the Kruskal Wallis

rank for variables X1 and X2.

Table 3. Kruskal Wallis Ranks

Variables ISP N Mean Rank

Preference of

local brand

upper

19 479.13

middle

404 452.04

lower

560 521.27

X2 Early

adopter

upper

20 535.58

middle

404 506.18

lower

561 481.99

Kruskal Wallis's mean rank results show that the

highest local brand preference of Mean Rank X1 is

the lower social class. It can be interpreted that

lower social class has more preference for local

brands. In the X2 variable, the highest value is in the

upper social class. These results signify that the

upper social class is the social class included in the

early adopter segment of new products.

Furthermore, the following test statistics will show

whether the differences between social classes are

significant or not significant.

Table 4. Test Statistics

X1 Preference of

local brand

X2 Early

adopter

Chi-Square 15.753 2.285

df 2 2

Asymp. Sig. .000 .319

a. Kruskal Wallis Test

b. Grouping Variable: ISP

The results of the statistical tests in Table 4 show

a p-value <0.05, then Ha1: There are differences in

the local brand preference in the social class of

Indonesian consumers, which is accepted. Whereas

for variable X2 early adopter, the test results show

no significant difference between social classes even

though the data description showed that the upper

class tended to be an early adopter segment. Thus

Ha2: There are differences in the early adopter

segment of new products in the Indonesian

consumer social class is rejected.

Postmodern Mindset and Behavior of Indonesia Consumer: Local Brand Preference, Early Adopter, Consumer Credit, and Online Buying

based on Social Class

131

Looking back at the ANOVA analysis, there are

differences in the variables X3 Open-mindedness to

credit facilities, X4 Online Buying, and X5 Thinking

mindset of the three social class groups.

Table 5. ANOVA

Sum of

Squares df

Mean

Squar

e F Sig.

Betwee

Groups

6.335 2 3.167 2.302 .101

Within

Groups

1348.39 980 1.376

Total

1354.72 982

X4 Betwee

Groups

12.826 2 6.413 4.971 .007

Within

Groups

1269.52 984 1.290

Total

1282.35 986

X5 Betwee

Groups

22.445 2 11.222 10.095 .000

Within

Groups

1090.51 981 1.112

Total

1112.95 983

ANOVA test results show no differences in the

X3 in the three social class groups (Not

rejecting/accepting H0), with p-value = 0.101. Thus

Ha3: There are differences in openness to credit

offers in Indonesian consumer social classes is

rejected. If the test results show Ho failed to be

rejected (there is no difference), then further tests

(Post Hoc Test) can’t be carried out.

ANOVA test results on variables X4 and X5

showed a significant difference (rejecting H0) on the

variable X4 (p-value=0.007) and the X5 (p-

value=0.000) between the three social classes.

Hence, the next test will be done to see which

groups are different through further testing (Post

Hoc Test). Thus, alternative hypotheses 4 and 5,

namely Ha4: There are differences in online buying

preferences in Indonesian consumer social classes

and Ha5: There are differences in Thinking mindset

of consumption in Indonesian consumer social

classes are accepted.

Then the further test (Post Hoc Test-Bonferroni

Test) is used to determine the difference between X4

and X5. Table 6 indicates that the groups which

showed an average difference in the X4 (marked

with an asterisk "*") were the middle class and

lower class. Likewise, for the X5, it shows that the

middle class and lower-class group have different

mean values.

Table 6. Bonferroni Test

Dep.

Var

(I)

(J)

ISP

Mean

Diff. (I-

Std.

Erro

Sig.

95%

Confidence

Interval

ower

Boun

Upper

Bound

.11481 .2608

1.000 -.5091 .7387

.33737 .2584

.576 -.2825 .9572

-.11481 .2601

1.000 -.7387 .5091

.22255

.0740

.008 .0450 .4001

-.33737 .2584

.576 -.9572 .2825

-.22255

.0740

.008 -.4001 -.0450

.11040 .2415

1.000 -.4688 .6896

-.19821 .2399

1.000 -.7736 .3772

-.11040 .2415

1.000 -.6896 .4688

-.30861

.0688

.000 -.4737 -.1436

.19821 .2399

1.000 -.3772 .7736

.30861

.0688

.000 .1436 .4737

Furthermore, to examine the role of the three

variables X3, X4, X5 in Indonesian consumer social

class, a Discriminant Analysis test was conducted.

Discriminant analysis was carried out on variables

X3, X4, and X5 because the variables X1 and X2

did not meet homogeneous criteria.

According to Table D in the Appendix, there

were 978 respondents’ analyzed cases. Twenty

respondents are upper class, 400 middle-class

respondents, and 558 respondents are lower class.

The mean value of X3 in the upper social class is

2.6, in the middle class are 2.3 and in the lower class

is 2.5. It can be descriptively interpreted that the

upper class has more Open-mindedness towards

purchases on credit, and the middle class has a more

careful thought about credit. In the X4, the mean

value of the upper class is 3.3, the middle class is

3.1, and the lower class is 2.9. These results indicate

that the upper classes have a greater interest in

internet purchases. In the X5, the upper class has a

ICBEEM 2019 - International Conference on Business, Economy, Entrepreneurship and Management

132

mean value of 2.9, the middle class is 2.78, and the

lower class is 3.09. These results indicate that the

lower classes of Indonesian consumers still think

conservatively/traditionally in their consumption

behavior. Referring to Table 7, the results of the

analysis to test the average similarity of variables are

presented. This test uses Wilks' lambda and

significance values.

Table 7. Tests of Equality of Group Means

Variables Wilks'

Lambda F df1 df2 Sig.

3 .995 2.235 2 975 .108

4 .990 4.827 2 975 .008

5 .980 9.994 2 975 .000

Based on the results of the average similarity test

variables, the significance value of X3 is 0.108>

0.05, which means there is no significant difference

in the group. Thus, Ha6: Openness to credit offers

relating to Indonesian consumer social classes is

rejected. Online Buying value (X4) of 0.008 <0.05

means there are differences in groups. Ha7:

Preference for online buying related to the social

class of Indonesian consumers is accepted. While

the value of Thinking mindset (X5) of 0.000 <0.05,

means there are differences in groups. Hence, Ha8:

the Thinking mindset of consumption related to the

social class of Indonesian consumers is accepted.

Online Buying and Thinking mindset contribute to

differences in the behavior of social class groups,

and the results of this test are consistent with the

results of the ANOVA test.

Covariance and Correlation analysis results in

Table E (see Appendix) show that the correlation

between independent variables is not greater than

0.5. This means that there is no correlation between

independent variables. The results of the calculation

of variance similarity test using Box'M yielded

results of significance values of 0.062> 0.05. This

means that the variance of the two data groups is

identical/homogeneous.

In Table H (Appendix), the calculated F value of

the X3 is 9.994, X4 is 6.918, and X5 is 6.348.

According to the results of the discriminant analysis

using the stepwise method, the statistical value of F

and its significance indicate that the three variables

are the same in the three social class groups (upper,

middle, and lower).

Table I Variables in the Analysis (see Appendix)

shows that there are three stages of variables

included in each stage of the model. In step 1, 2, and

3, all three numbers Sig. Of F to Remove value are >

0.05. The results of the calculation of Table J (see

Appendix) show that in step 0, there is no Sig.F to

enter the value of <0.05. Likewise, the calculation

results in steps 1 and 2 show the variables coming

out of the model in each stage, which until stage 2,

there is only 1, namely X3, but finally, in stage 3,

nothing is excluded.

Wilks' Lambda (see Appendix) shows the value

of the percentage of variance in variables which can

explain differences in the division of three groups. In

step 1, the lambda value = 0.980; in step 2 = 0.972

and in step 3 = 0.962. This means that >95% of the

variance of the variable is not able to explain

differences that divide the three social class groups.

Until stage 3, the Sig value remains <0.05, and then

until stage 3, the independent variables enter the

model.

In Table L Eigenvalues (see Appendix), there is a

canonical correlation value used to measure the

degree of relationship between discriminant results

or the amount of variability that can be explained by

the independent variables on the dependent variable

of social class. The calculation results in Table K

Eigenvalues show the Canonical Correlation rate of

0.193. This value is far enough from 1 so that it can

be interpreted that the relationship between

discriminant values and groups is very low.

Likewise, the second function shows a value of

0.028. From this table, the canonical correlation

function 1 value of 0.193 is obtained when it is

squared = 0.3725, meaning 37.25% of the variance

of the independent variable (social class group) can

be explained from the discriminant model that is

formed. In contrast, the value of the canonical

correlation function 2 is extremely small.

Table 8. Wilks' Lambda

Test of

Function(s)

Wilks'

Lambda

Chi-

square df Sig.

1 through 2

.962

37.757 6 .000

.999

.770 2 .680

Table 8 shows that the Chi-square sig= 37,757

(<0.05) indicates that there is a clear difference

between the three social classes, and there are

significant differences between the three groups of

respondents based on the three independent

variables. According to Table M of the Standardized

Canonical Discriminate Function (see Appendix), it

Postmodern Mindset and Behavior of Indonesia Consumer: Local Brand Preference, Early Adopter, Consumer Credit, and Online Buying

based on Social Class

133

can be concluded that the equation or discriminant

function is almost the same as the multiple

regression function. The equations are as follows:

Zscore= 0.610 X3 (Open-mindedness to credit

facilities)

(3)

Zscore = -0.746 X4 (Online Buying) (4)

Zscore = 0.602 X5 (Thinking mindset) (5)

Table M. Structure Matrix (see Appendix) shows

the order of characteristics that most distinguish

social class (Y). In Function 1, Variable X3 is the

most distinguished, then X4, followed by X5. The

table shows the correlation between the independent

variables and the discriminant function formed.

Variable X3 has the highest correlation with a

correlation value of 0.728.

Table 9. Canonical Discriminant Function Coefficients

Variables Function

1 2

X3 .521 .548

X4 -.658 .450

X5 .572 .060

(Constant) -.956 -2.890

Unstandardized coefficients

Equation:

D1 = -0.956+0.521X3-0.658X4+0.572 X5

(1)

D2 =-2.890+0.548X3+0.450 X4+0.060 X5 (2)

The contribution of each variable can be seen in

Table L (see Appendix).

Table O, Functions at Group Centroids (see

Appendix) shows there are three different groups:

the upper-class group with negative and positive

centroids, the middle-class group with negative

centroids, and the lower-class group with positive

and negative centroids.

The accuracy of the Discriminant Test Function

is seen from the changes in social class members.

From Table P Classification Results, at the original

limit, it can be seen that 4 respondents (20%) remain

in the upper class, whereas those who were

originally classified as the upper class, move to the

middle class (middle) after using the discriminant

model are as much as 10 respondents (50%), while 6

respondents move to the lower class (30%). In the

middle class, it is seen that 179 respondents (45%)

remain in the middle class, while 100 respondents

(25%) move to the upper class, and 121 respondents

(30%) move to the lower class (lower). In the lower

class, 246 respondents (44%) remain in the lower

class, 169 (30%) move to the middle class, while the

remaining 143 (26%) move to the upper class.

Following the previous process, the following

prediction accuracy model will be established:

Prediction accuracy= (4+179+246)/978=0.438

or 44%

(6)

The accuracy of the prediction figure is 44%.

This figure is considered moderate, tends to be low,

by looking at the cause, because the possibility of

movement between social classes is very high.

Table 10. Classification Function Coefficients

Variables

ISP

upper middle lower

X3 .401 .229 .441

X4 2.545 2.532 2.270

X5 2.761 2.681 2.911

(Constant) -9.821 -9.130 -9.515

Fisher's linear discriminant functions

Based on the results of Table 10 Classification

Function Coefficients, a regression equation can be

made as follows:

Classified as upper class:

Z_score=(-9.821+0.401X3)+

2.545X4+2.761X5

(7)

Classified as middle class:

Z_score=(-9.130+0.229X3)+

2.532X4+2.681X5

(8)

Classified as lower class:

Z_score=(-9.515+0.441X3)+

2.270X4+2.911X5

(9)

From the three social class equations, the X5

variable Thinking mindset has the greatest influence

on those three social classes. However, if seen from

trends in the three social classes, the X3 is stronger

in lower social classes. The X4 is stronger in the

upper and middle classes. X5 is stronger in the lower

classes.

4 DISCUSSION

The results of the study descriptively showed that

the lower classes preferred local brands. Kruskal

Wallis rank test results also showed that the highest

local brand preference is contributed by the lower

ICBEEM 2019 - International Conference on Business, Economy, Entrepreneurship and Management

134

social class so that it can be interpreted that the

lower social class has a preference over the local

brand. From the ANOVA test, it is evident that there

are differences in local brand preferences in

Indonesian consumer social classes. This result is in

line with the research of Candan, Aydm, and

Yamamoto (2008), which states that the upper class

has less consumer ethnocentrism. In developing

countries, consumers generally perceive foreign

products to be of higher quality than local products

(Ping, Lobo, and Li, 2012).

In general, Indonesian consumers lack the

characteristics of Early Adopters. However, from the

Kruskal Wallis test, the upper classes tend to be

Early Adopters compared to other social classes.

These results mean that the upper social class is the

social class included in the Early Adopter segment

of new products. The prosperous Indonesian

consumer class is an attractive consumer group and

is ready to experience new products, look for

quality, and are ready to pay for it (Durmaz and

Taşdemir, 2014). Meanwhile, from the different test

results, there was no significant difference between

social classes, although the description of the data

showed that the upper class tended to be an Early

Adopter segment. The results of this study are in

accordance with the research of Iqbal and Ismail

(2011) that upper-class subjects described

themselves as people who are curious and interested

in new things, so they try to expand their minds.

Consumers of middle to upper income in Indonesia

buy new product categories, and they will bring

changes to the consumption mix between basic and

luxurious products (Ahmed, Khan and Samad,

2016).

The Open-mindedness to Credit Facilities

variable has the highest correlation with social class.

Descriptively, Indonesian consumers are less open to

expressing their relationship with credit facilities.

Consumers will answer briefly when asked about

financial problems, but overconfident in their ability

to manage loans (OECD, 2010). The lower-income

class often reacts shamefully when called poor (Al-

Modaf, 2002). For consumers, purchasing power or

the ability to pay for goods and services is a

determinant of the material prosperity of one's

lifestyle (Iqbal and Ismail, 2011). However, the

results of Anova and discriminant tests rejected the

hypothesis that differences exist between Open-

mindedness to credit facilities and social classes, or

in other words, there was no correlation between

Open-mindedness to credit facilities and social

classes. The results show that credit facilities and

systems are closely related to Indonesian consumers

in all social classes, although descriptively, the

upper classes have more Open-mindedness about

credit facilities than other social classes. Open-

mindedness to Credit Facilities is significantly

stronger in lower social classes.

The upper social class prefers online purchases,

and the test results show that there are differences in

online buying preferences between Indonesian

consumer social classes. Online buying preferences

are related to the social class of Indonesian

consumers and are significantly more likely to occur

in upper and middle classes. These results support

previous results, which stated that online purchases

are also in line with the tendency to become Early

Adopter consumers. Czarniewski (2014) states that

internet access affects the lifestyles of modern

consumers because consumers are able to easily

access information, products, services, and people in

the same interest.

This research also results that there are

differences in the Thinking Mindset of Consumption

in the social class of Indonesian consumers.

Thinking Mindset of Consumption is related to the

social class of Indonesian consumers and has the

biggest influence on those three social classes.

Contrary to the upper social class, the lower class

descriptively still tends to have

conservative/traditional thinking in consumption and

is not an Early Adopter of new products. A

significantly more conservative thinking mindset

also occurs more strongly in the lower classes.

Research by Iqbal and Ismail (2011) shows that the

lower classes tend not to plan for the future. In

particular, several studies indicate that young

consumers from upper social classes may have

stronger brand preferences and are more likely to

seek information before making decisions compared

to lower classes (Durmaz and Taşdemir, 2014).

Based on the accuracy of the prediction figures,

the discriminant equation shows 44% and is

considered moderate, tends to be low by looking at

the cause, which infers the possibility of movement

or the mobility of upper, middle and lower social

classes is significantly high. Although the Online

Buying and Thinking Mindset variables provide

differences and interrelationships in the behavior of

social class groups, the results of discriminant

analysis of the statistical value of F and their

significance indicate that the three variables are the

same in the three social class groups (upper, middle

and lower). The gap between social classes in

contemporary society is associated with shifting

norms, and social structures consequently will have

Postmodern Mindset and Behavior of Indonesia Consumer: Local Brand Preference, Early Adopter, Consumer Credit, and Online Buying

based on Social Class

135

an impact on changes in consumer behavior based

on social class (Shavitt, Duo, Hyewon, 2016).

5 CLOSING

The results of this study are part of a large study that

is concerned about the relationship between social

class, consumption, decision-making style, and

Indonesian national culture. Specifically, this paper

looks at the relationship between social class and

postmodern mindset and behavior of Indonesian

Consumers based on Social Class. The results of this

study are useful for practitioners and academics who

observe Indonesian consumers based on social class

and specifically their relationship with Local Brand

Preferences, Early Adopters, Open-mindedness to

credit facilities, Online Buying, and Thinking

Mindset. Even though it has been optimally pursued,

the limited time of the study has an impact on the

tendency of the participating respondents to come

from Java and Javanese from middle and lower

social classes. However, at the very least, this

research contributes to the characteristics of social

class in Indonesia and the relationship between

Local Brand Preference, Early Adopter, Open-

mindedness to credit facilities, Online Buying, and

Thinking Mindset. Future studies are expected to

continue on the variables of ethnocentrism and

postmodern characteristics.

ACKNOWLEDGMENT

This paper is based on the results of the 2018-2020

Higher Education Basic Research scheme funded by

the Ministry of Research, Technology, and Higher

Education of Indonesia.

REFERENCES

Ahmed, M. E., Khan, M. M. & Samad, N., 2016. Income,

social class and consumer behavior: Focus on

developing nations. International Journal of Applied

Business Economic Research, 14(10), pp. 6679-6702.

Al-Modaf, O. A. (2002). Class and Consumption: a

Comparative Analysis of Consumption patterns a

Cross Different Social Classes. Retrieved June 5,

2017, from

www.kau.edu.sa/Files/0003309/Files/69972_Consump

tion%20and%20Class.pdf

Bati, U., 2008. Postmodern impacts on the consumption

patterns, activities and theories. Iletisim, pp. 216-231.

Berner, A. & Van Tonder, C. L., 2003. The postmodern

consumer: Implication of changing customer

expectation for organisation development in service

organisations. Journal of Industrial Psychology, 29(3),

pp. 1-10.

Birkelund, G. E. & Lemel, Y., 2013. Lifestyle and Social

Stratification: An explorative study of France and

Norway. FMSH-WP-2013-48, October.Volume

GeWoP-3.

Bouagina, D. J. & Triki, A., 2014. From postmodernism to

postmodern consumer: The impact on the

consumption theory. Postmodern Openings, 5(2), pp.

99-117.

Candan, B., Aydm, K. & Yamamoto, G. T., 2008. A

research on measuring consumer ethnocentrism of

young turkish customer purchasing behaviors. Serbian

Journal of Management, 3(1), pp. 39-60.

Czarniewski , S., 2014. Changes in consumer behavior in

the market and the value of companies. European

Journal or Research and Reflection in Management

Sciences, 2(2).

Deepak & Harneet, 2017. Growing and changing trends in

consumer behavior. Biz and Bytes, 8(1).

Durmaz, Y., 2014. The influence of cultural factors on

consumer buying. Global Journal of Management and

Business Research: E-Marketing, 14(1).

Durmaz, Y. & Taşdemir, A., 2014. A theoretical approach

to the influence of social class on cunsumer behavior.

American International Journal of Social Science,

3(3), pp. 187-191.

Eroğlu, E., 2014. The changing shopping culture: Internet

consumer behavior. Review of Business Information

System, 18(1).

Firat, A., Kutucuoğlu , K. Y., Tunçel , Ö. & Saltik, A.,

2013. Consumption, consumer culture and consumer

society. Journal of Community Positive P, XIII(1), pp.

182-203.

Hamouda, M., 2012. Postmodernism and consumer

psychology: Transformation or break?. International

Journal of Academic Research in Business and Social

Sciences, 2(1).

Iftikhar, M., Hussain, M. F., Kahn, A. & Ilyas, S., 2013.

Social Class in a myth or reality in buying behavior.

African Journal of Business Management, March,

7(9), pp. 713-718.

Indonesia Internet Service Provider Association. 2018.

Penetrasi dan Profil Perilaku Pengguna Internet

Indonesia. Jakarta: Indonesia Internet Service Provider

Association.

Iqbal, S. & Ismail, Z., 2011. Buying behavior: Gender and

socioeconomic class differences on interpersonal

influence susceptibility. International Journal of

Business and Social Science, 2(4), pp. 55-66.

Marić, D. & Grubor, A., 2015. Contemporary consumer in

the global environment. Prague,

WWW.CBUNI.CZ,OJS.JOURNALS.CZ.

Marie, A. & Grybś, M., 2013. Modern trends in consumer

behavior in era of e-communication. Poland,

University of Economics in Katowice, pp. 263-273.

ICBEEM 2019 - International Conference on Business, Economy, Entrepreneurship and Management

136

Mihic, M., & Culina, G. 2006. Buying behavior

consumption: Social class versus income.

Management, 11(2), pp.77-92.

OECD, 2010. The changing consumer and market

landscape. 3e 2009/1 ed. PARIS: OECD.

Ping, Q., Lobo, A. & Li, C., 2012. The impact of lifestyle

and ethnocentrism on consumers' purchase intentions

of fresh fruit in China. Journal of Consumer

Marketing, January.

Quigley, J. F., 2015. Consumer behavior in digital

markets, Southern Illinois: s.n.

Shavitt, S., Duo, J., & Hyewon, C. 2016. Stratification and

segmentation: Social class in consumer behavior.

Journal of Consumer Psychology, 26(4), pp.583-593.

Schor, J., 2002. Understanding the new consumerism

inequality, emulation and the erosion of well-being,

s.l.: s.n.

Siamagka, N. T., 2009. Extending consumer

ethnocentrism: development and validation of the

ceescale, Birminghamn: University of Birmingham.

Singh, P. R., 2014. Consumer culture and postmodernism

in postmodern openings. American International

Journal of Science, 3(3).

Stranford, K. & Cowling, A., 2014. Chiense household

income, consumption and saving, s.l.: Economic

Group.

Strehlau, V. I., Ponchio, M. C. & Loebel, E., 2012. An

assessment of the consumer ethnocentric scale

(CETSCALE). Brazilian Business Review, 9(4).

Yasav, S., 2015. The impact of digital technology on

consumer purchase behavior. The Journal of Financial

Perspectives: Fintech, 3(3).

Zhongqi, J. et al., 2016. The relationship between

consumer ethnocentrism, cosmopolitanism and

product country image among younger generation

consumers: The moderating role of country

development status. International Business Review,

Volume 24, pp. 380-393.

APPENDIX

Table A. Data of Respondents

Description Up (%) Mid (%) Low (%) Tot (%)

Social Class 2.0 41 57 100

Gender Male 1.2 16 27.6 44.8

Female 0.8 25 29.4 55.2

Age 17-25 0.4 14.2 13.7 28.4

26-35 0.5 11.2 17.4 29.1

36-45 0.5 7.3 14.2 22.0

46-55 0.6 5.7 10.6 16.9

>55 0.1 2.5 1.1 3.7

Marital Status Married 1.5 21.5 35.4 58.4

Unmarried 0.6 19.4 21.6 41.6

Family status Husband 1.0 9.1 19.6 29.6

Wife 0.4 11.4 16.0 27.8

Child 0.5 19.0 20.3 39.7

Others 0.1 1.3 1.4 2.8

Education Doctoral 0 0 0.1 0.1

Specialist 0 0.2 0.2 0.4

Master 0 0.8 8.8 9.6

Bachelor 0.1 0.4 0.9 1.4

Diploma 0.3 5.9 11.1 17.3

SHC 0.2 2.9 1.2 4.3

YHC 0.9 23.8 29.9 54.6

ES 0.6 5.0 3.8 9.5

Not grad. ES 0 0.1 0.2 0.3

No education 0 1.7 0.8 2.5

Income > 38.400.000 0.1 1.0 0.9 2.0

31.200.001 –

38.400.000

0.2 0.4 0.7 1.3

24.000.001 –

31.200.000

0.1 1.0 0.8 1.9

19.200.001 –

24.000.000

0.1 1.2 0.4 4.4

Postmodern Mindset and Behavior of Indonesia Consumer: Local Brand Preference, Early Adopter, Consumer Credit, and Online Buying

based on Social Class

137

Description Up (%) Mid (%) Low (%) Tot (%)

14.400.001 –

19.200.000

0 1.6 0.9 2.5

9.600.001 –

14.400.000

0.2 3.0 1.2 4.4

7.200.001 –

9.600.000

0 2.9 1.6 4.5

4.800.001 –

7.200.000

0.2 7.8 7.0 15.1

2.400.001 –

4.800.000

0.4 11.9 23 35.3

≤ 2.400.000 0.7 10.1 20.4 31.2

Occupation Unemployed 0.3 5.5 24.6 30.5

Students 0.2 7.0 8.3 15.5

Retired 0 0.6 0.3 0.9

Machine operator,

unskilled labors

0 0.9 5.9 6.9

Skilled technician –

factory workers,

shop attendants

0.1 1.3 1.4 2.9

Administration staff 0.2 9.2 5.4 14.9

Teachers,

Engineers,

freelancer

0.5 7.1 4.2 11.8

Middle managers,

small business

owners, mid-level

government

officials,

professionals,

police/army officers

0.5 8.0 4.7 13.2

Upper level

executive

managers, medium

business owners

0.2 0.9 0.8 1.0

Upper level

government

officials,

executives, big

business owners,

professionals

0 0.6 0.9 1.5

Prov Sumatera 0 1 1 2

Banten 0 1 1 2

DKI 0 3 2 4

Jawa Barat 0 2 1 3

Jawa Tengah 0 1 1 2

DIY 0 1 1 2

Jawa Timur 1 22 42 65

Bali 0 1 1 2

NTB 0 1 1 1

NTT 0 0 0 1

Kalimantan 0 2 2 4

Sulawesi 0 4 4 9

Maluku dan Maluku

Utara 0 0 0 0

ICBEEM 2019 - International Conference on Business, Economy, Entrepreneurship and Management

138

Description Up (%) Mid (%) Low (%) Tot (%)

Papua 0 2 1 3

Ethnic Freq Tot (%)

Javanese 553 58,1

Madurese 102 10,7

Chinese 54 5,67

Bugis 37 3,89

Kaili 31 3,26

Balinese 23 2,42

Mixed 23 2,42

Bataknese 16 1,68

Sundanese 14 1,47

Banjar 12 1,26

Minahasa 10 1,05

Dayak 9 0,95

Malay 6 0,63

Makasar 6 0,63

Betawi 5 0,53

Manado 5 0,53

Sasak 5 0,53

Kei 4 0,42

Minang 3 0,32

Aceh 3 0,32

Others 31 3,26

Postmodern Mindset and Behavior of Indonesia Consumer: Local Brand Preference, Early Adopter, Consumer Credit, and Online Buying

based on Social Class

139

Table B. Descriptive data for variables

Variables Social

Class Total Mean

X1 Preference of

Local brand

upper 19 4.0526

middle 404 3.9431

lower 560 4.1161

Total 983 4.0437

X2 Early Adopter upper 20 2.7500

middle 404 2.5792

lower 561 2.5365

Total 985 2.5584

X3 Open-

mindedness to

credit facilities

upper 20 2.6000

middle 401 2.3491

lower 562 2.5071

Total 983 2.4446

X4 Online

Buying

upper 20 3.3000

middle 405 3.1852

lower 562 2.9626

Total 987 3.0608

X5 Thinking

Mindset

upper 20 2.9000

middle 404 2.7896

lower 560 3.0982

Total 984 2.9675

Table C. Test of Homogenity of

Variances

Variables Levene

Statistic df1 df2 Sig.

Preference

to Local

brand

9.222 2 980 .000

X2 Early

Adopter

17.626 2 982 .000

X3 Open-

mindedness

to credit

facilities

2.428 2 980 .089

X4 Online

Buying

1.667 2 984 .189

Thinking

mindset

.079 2 981 .925

Table D. Group Statistics

ISP Mean

Std.

Deviation

Valid N

(listwise)

nweig

hted Weighted

upper X3 2.6000 1.18766 20 20.000

X4 3.3000 1.12858 20 20.000

X5 2.9000 1.16529 20 20.000

middle X3 2.3450 1.12434 400 400.000

X4 3.1800 1.05374 400 400.000

X5 2.7875 1.00741 400 400.000

lower X3 2.5000 1.20256 558 558.000

X4 2.9606 1.18877 558 558.000

X5 3.0950 1.07857 558 558.000

Total X3 2.4387 1.17236 978 978.000

X4 3.0573 1.13869 978 978.000

X5 2.9652 1.06148 978 978.000

Table E. Pooled Within-Groups Matrices

Open-

mindedne

ss to

credit

facilities

Onlin

Buyin

Thinkin

mindset

Covariance

X3 1,360 ,643 ,147

,643 1,283 -,078

,147 -,078 1,094

Correlation

X3 1,000 ,487 ,121

,487 1,000 -,066

,121 -,066 1,000

Table F. Covariance Matrices

ISP

Open-

mindedness

to credit

facilities

Online

Buyin

Thinkin

mindset

upper X3

1.411 .916 .747

.916 1.274 .611

.747 .611 1.358

middl

1.264 .514 .159

.514 1.110 -.019

.159 -.019 1.015

lower X3

1.446 .731 .119

.731 1.413 -.154

.119 -.154 1.163

ICBEEM 2019 - International Conference on Business, Economy, Entrepreneurship and Management

140

Table G. Test Results

Box's M 20.922

F Approx.

1.689

df1

df2

1.001E4

Sig.

.062

Tests null hypothesis of equal population

covariance matrices.

Table H. Variables Entered/ Removed

a,b,c,d

Step

Enter

Wilks' Lambda

Stat df1 df2 df3

Exact F

Stat df1 df2 Sig.

1 X5

.980 1 2

975.

9.99

5.0

.00

2 X4

.972 2 2

975.

6.91

1.9

.00

3 X3

.962 3 2

975.

6.34

1.9

.00

At each step, the variable that minimizes

the overall Wilks' Lambda is entered.

a. Maximum number of steps is 6.

b. Minimum partial F to enter is 3.84.

c. Maximum partial F to remove is 2.71.

d. F level, tolerance, or

. VIN insufficient for further computation

Table I. Variables in the Analysis

tep Tolerance

F to

Remove

Wilks'

Lambda

1 X5 Thinking

mindset

1.000 9.994

X5 Thinking

mindset

.995 9.023 .990

X4 Online

Buying

.995 3.871 .980

X5 Thinking

mindset

.964 6.414 .975

X4 Online

Buying

.747 7.735 .977

X3 Open-

mindedness

to credit

facilities

.740 5.161 .972

Table J. Variables Not in the Analysis

Step

Toler

ance

Min.

Tolerance

F to

Enter

Wilks'

Lambda

0 X3 1.00

1.000 2.235 .995

X4 1.00

1.000 4.827 .990

X5 1.00

1.000 9.994 .980

1 X3

.986 .986 1.315 .977

X4 .995 .995 3.871 .972

2 X3 .740 .740 5.161 .962

Table K. Wilks' Lambda

Step

Numbe

r of Var

Lam

bda df1

2 df3

Exact F

Sta

1 df2

Sig

1 .980 1 2

9.9 2

975.0

.00

2 .972 2 2

6.9 4

1.94E

.00

3 .962 3 2

6.3 6

1.94E

.00

Table L. Eigenvalues

Function

Eigenvalu

% of

Varianc

Cumula

tive %

Canonical

Correlation

1 .039

98.0 98.0 .193

2 .001

2.0 100.0 .028

First 2 canonical discriminant functions were used

in the analysis.

Table M. Standardized Canonical Discriminant

Function Coefficients

Function

1 2

X3 Open-mindedness to credit

facilities

.610

.64

X4 Online Buying

-.746

.51

X5 Thinking mindset

.602

.06

Postmodern Mindset and Behavior of Indonesia Consumer: Local Brand Preference, Early Adopter, Consumer Credit, and Online Buying

based on Social Class

141

Table N. Structure Matrix

Function

1 2

X3 Open-mindedness to credit

facilities

.728

.104

X4 Online Buying .319 .898

X5 Thinking mindset -.492 .818

Pooled within-groups correlations between

discriminating variables and standardized canonical

discriminant functions

Variables ordered by absolute size of correlation within

function.

*. Largest absolute correlation between each variable

and any discriminant function

Table O. Functions at Group Centroids

ISP

Function

1 2

upper -.113 .194

middle -.231 -.007

lower .170 -.002

Unstandardized canonical discriminant functions

evaluated at group means

Table P. Classification Results

b,c

ISP

Predicted Group

Membership

Total

upper

le lower

Original Count upper 4 10 6 20

middl

100

121 400

lower

143

246 558

Ungr

oupe

cases

3 1 0 4

% upper

20.0

30.0 100

middl

25.0

30.2 100

lower

25.6

44.1 100

Ungr

oupe

cases

75.0

.0 100

Cross- Count upper 1 10 9 20

validated

middl

100

121 400

lower

143

246 558

% upper

5.0

45.0 100

middl

25.0

30.2 100

lower

25.6

44.1 100

a. Cross validation is done only for those cases in the

analysis. In cross validation, each case is classified by

the functions derived from all cases other than that case.

. 43,9% of original grouped cases correctly

classified.

c. 43,6% of cross-validated grouped cases correctly

classified.

ICBEEM 2019 - International Conference on Business, Economy, Entrepreneurship and Management

142