You are on page 1of 11

AN INVESTIGATION OF CONSTRUCT VALIDITY OF

THE IMPULSIVE BUYING SCALE

Anton Agus Setyawan, SE,MSi and Ihwan Susila, SE,MSi


Department of Management Muhammadiyah University of Surakarta
Jl A Yani Tromol Pos 1 Pabelan Kartasura Surakarta 57102
e-mail: rmb_anton@yahoo.com and ihwan_ss@yahoo.com
mobile: 08156718444 (Anton) and 08122600152 (Ihwan)

Abstracts

The issue of construct validity is an important for the development of social theory.
Impulsive buying behavior as a construct in consumer behavior has lack of construct
validity test. This research investigates the construct validity and reliability of impulsive
buying behavior. This research uses the procedure to investigate the construct validity
suggested by Abe et al (1996). We use department store consumer in Surakarta as
respondents, with 150 complete survey instruments. Sampling methods in this research is
convenience. We use exploratory factor analyses and cronbach alpha method to test the
construct validity. And then, we also use measurement model of structural equation
model (SEM) which is the same as confirmatory factor analysis procedure. The result
shows that the impulsive buying scale proposed by Rook and Fisher (1995) and Lin and
Lin (2005) have weak construct validity, although it does not have serious problem in the
internal consistency.

Keywords: construct validity, impulsive buying, exploratory factor analyses, SEM.

1
AN INVESTIGATION OF CONSTRUCT VALIDITY OF
THE IMPULSIVE BUYING SCALE

Anton Agus Setyawan, SE,MSi and Ihwan Susila, SE,MSi


Department of Management Muhammadiyah University of Surakarta
Jl A Yani Tromol Pos 1 Pabelan Kartasura Surakarta 57102
e-mail: rmb_anton@yahoo.com and ihwan_ss@yahoo.com
mobile: 08156718444 (Anton) and 08122600152 (Ihwan)

INTRODUCTION

The issue of impulsive buying behavior have been attracted many marketing
scholars. There are two reasons, impulsive buying find itself interesting, first, it shows a
dark side of consumer behavior, and second, marketer could use this behavior as a way to
exploit consumer. In the early stage, the impulsive buying behavior often related with the
negative behavior such as drug abuse (Hirschman, 1992) compulsive purchasing (O’Guin
and Faber, 1989), alcoholism (Blum and Noble, 1994), kleptomania (Marlatt, et al, 1988)
and bulimia (Krahn, 1991).
Research about impulsive buying behavior is relevant in Indonesia, since this
country has a rapid development in consumer goods industry. Data from The Center of
Statistic Bureau shows that from 2004 to 2006, the growth of consumption goods selling
is more than 30 percent. In 2006, 65 percent of Indonesia’s GDP dominated by
consumption rate. The growth of retail industries is also rapid in this country compare to
other industry. Data from Asian Development Bank shows that the growth of trading
sector is almost 7 percent , it is the fastest growing sector in the industries.
Indonesian consumers have been attached by commercials from any kind of media.
The commercials have a tendency to pursue individual to become irrational. The
exploitation have their reason, it is increasing the sales of a product. This is why, the
research of impulsive buying is very interesting.
Most of the research in behavioral science has similar problem, it is finding the
right measurement. The right measurement means that a scale should have a strong
content and construct validity (see Neuman, 2000; Cooper and Schindler, 2001). Abe et al
(1996) shows that the construct validity is very important in the development process of a
theory. They also point out that several scale in consumer behavior have weak of
construct validity. Schwab (1978) emphasizes that research on construct validity of a
scale is more important than research on correlation among construct. Many research in
organizational science too busy to measure the relationship of many constructs that it
forgot the measurement itself. Schwab (1978) said that substantive research (ie, the
research of correlation among construct) is important but the construct validity research is
more important.
A good measurement scale comes from the right conceptual and operational
definitions. If a researcher wants to measure a behavior, he or she has to define the
concept properly, and it has the right operational definition appropriate with the setting.
Swan and Martin (1994) in their article wrote that the guideline to find a good empirical
test, is to fit theory-setting and testable hypotheses.

2
The purpose of the research is to investigate the construct validity of impulsive
buying behavior scale proposed by Rook and Fisher (1995) and Lin and Lin (2005). We
test the construct validity of impulsive buying behavior, whether it use the right
definitions or not. It will give proper contributions to the development of impulsive
buying behavior concept. We test the construct using exploratory factor analysis (EFA)
and re-test it with measurement model of Structural Equation Model (SEM). This paper
will be organized as follows, in the first part we will discuss the importance of construct
validity and also how should marketing scholars manage latent variable. The second part,
discuss the concept of impulsive buying, and the final part of this paper discuss the
analyses of construct validity of impulsive buying scale.

MEASUREMENT SCALE AND LATENT VARIABLE

Item scale is an end of a construct’s measurement. In a consumer behavior research


there are many construct, which cannot be measured directly. Latent variable is a
phenomenon behind a construct reflected by the item scale (Ghisseli et al, 1981). The
uniqueness of social science is the presence of unobserved entity known as latent variable
(Purwanto, 2001). Social and behavioral sciences have very different opinion about latent
variable. Their differences are in the relevant and meaning of this variable in an empirical
research. Many social scientists argue that latent variable is difficult to measure, so that
the observed variable (indicators) considered important. This opinion is no longer
dominant.
Purwanto (2001) point out that when explaining a theory, social scientist explaining
a systematic relationship among latent variables. Scientific theory which discusses certain
latent variable could be test in the observation level, when researcher uses appropriate
structure on it (Clogg, 1992). Social scientist more interested in construct than item.
Measurement scale is an important part of social science. We gain knowledge about
human, object, event and process by observation to measurement. Phenomenon in social
science based on social theory. Consequently, theory plays an important role in the
measurement conceptualization.
According to Ghiselli et al (1981), measurement scale has a strong relationship with
reliability and validity. Reliability is the presence of homogeneity of item in a scale. A
measurement scale said to be consistent when each item has strong correlation. There are
several methods to assess reliability of measurement scale, they are cronbach alpha,
covariance matrix and item to total correlation (Cooper and Schindler, 2001).
Validity is an assessment whether a variable is a cause of a covariate among item
(Neuman, 2000). There are three kinds of validity, they are content validity, criterion-
related validity and construct validity. Content validity is the appropriateness between
scale and its domain. Criterion-related validity is the appropriateness between scale and
it standard. Construct validity is the appropriateness between a scale and it purpose of
measurement.
Schwab (1978) define construct validity as the correlation between conceptual
definition of a construct and operational procedure to assess or manipulate it. Generally,
research on behavior only discusses correlation among construct or substantive research,
whereas the problem rely on construct validity of the latent variable.
Construct validity issue is important, so that researcher do not make a fundamental

3
mistake in their research. The potential pitfall of construct validity happens when
researcher analyzing multidimensional construct. The risk of multidimensional construct
relies on the operational definition. Figure 1 shows an ideal picture construct validity of a
variable.

Figure 1. Construct Validity of A Scale Measurement

Figure 1 shows that I is a construct which cannot be measured directly. Impulsive


buying scale is our main discussion. D is other construct which theoretically have a
relationship with impulsive buying. In reality, researcher only familiar with I’ which is
the items of construct I. It explains why a construct cannot be measured directly. A
behavioral science researcher has to face items of I. In this research, we have to deal with
the item scale of impulsive buying. When a researcher analyzes the relationship between
construct I and D, what really happen is an assessment of relationship between I’ (item
scale of I) and D’ (item scale of D). It needs appropriate operational definition of I’ and
D’ as a necessary condition.

DEVELOPING BETTER MEASURES OF IMPULSIVE BUYING

Churchill (1979) wrote that a critical element in the evolution of a fundamental


body of knowledge in marketing, as well as for improved marketing practice, is the
development of better measures of the variables with which marketers work. Churchill
(1979) suggested procedure for developing better measure, they are specify domain the
construct, generate sample of items, purify the measure, assess the reliability with new
data and developing norms.
First, specify domain the construct. In this step, the researcher must be exacting in
delineating what is included in the definition and what is excluded. The researcher should

4
make sure that the process of specifying the domain of a construct is thorough review of
literature in which the variable is used and should present a detailed statement of the
reasons and evidence as to why the new measure is better.
Second, generate sample of items. In this stage researcher use exploratory
techniques, including literature searches, experience survey and insight stimulating
examples. Critical incidents and focus group also can be used to advantage at the item
generation sale.
Third, purify the measure. The calculations one performs in purifying a measure
depend somewhat on the measurement model one embraces. The most logically
defensible model is the domain sampling model with holds that the purpose of any
particular measurement is to estimate the score that would be obtained if all the domains
were used (Nunnaly, 1967 in Churcill, 1979). In this stage, researcher uses statistical
methods such as Coefficient Alpha, factor analysis and iteration.
Fourth, assess reliability with new data. The major source of error within a test or
measure is the sampling of items. If the sample is appropriate and the items seems right,
then it said that the measurement have a face or content validity. Sometimes researcher
does not consider other thing such as transient personal factors or ambiguous questions
which probably have the effects to measurement error.
Fifth, asses construct validity. Construct validity is related to the question of what
the instrument in fact measuring or else what construct, trait or concept underlies score
on a measure. A useful way of assessing the convergent and discriminant validity of a
measure is through the multitrait-multimethod matrix. This matrix is of a zero order
correlations between different traits when each of the traits is measured by different
methods (Neuman, 2000).
Sixth, developing norms. Norm quality is a function of both the number of cases on
which the average is based and their representative ness. The larger the number cases,
the more stable will be the norms and the more definitive will be the conclusions that can
be drawn.
Armstrong (1974) suggests other way to develop better measurement in consumer
behavior research. He suggests a method called eclectic research. In this research,
consumer behavior researcher uses a different approach for a construct. For example, he
uses four different approaches in analyzing the income-sales relationship. They are
subjective data, house-hold survey data, international cross section and longitudinal data
over countries.

THE CONCEPT OF IMPULSIVE BUYING

Based upon the research literature (e.g., Beatty & Ferrell, 1998; Purl, 1996; Rook &
Fisher, 1995; Rook & Gardner, 1993), impulsive buying tendency can be defined as the
degree to which an individual is likely to make unintended, immediate, and unreflective
purchases. The theory of the impulsive buying lies in the self control concept. Self
control or self regulation is self’s capacity to alter its own states or responses
(Baumeister,2002). Thus, selfcontrol overrides one incipient pattern of response and
replaces it with another. These responses may include thoughts (e.g., by suppressing
unwanted thoughts or forcing oneself to concentrate), changing emotions (getting into,
getting outof, or unnaturally preserving some emotion or mood), regulating impulses

5
(e.g., resisting temptation), and altering performances (e.g., persisting).
Baumeister (2002) states that impulses refer to incipient behavioral responses that
normally result from the encounter between a motivation and some activating stimulus,
such as when the hungry person sees food and feels an impulse to eat it. Impulsive
behavior is understood as behavior that is not regulated and that results from an
unplanned, spontaneous impulse. In particular, impulsive purchasing involves getting a
sudden urge to buy something, without advance intention or plan, and then acting on that
impulse without carefully or thoroughly considering whether the purchase is consistent
with one's long-range goals, ideals, resolves, and plans.
Furthermore, several scales have been developed to measure this tendency (e.g.,
Puri, 1996; Rook & Fisher, 1995; Rook & Gardner, 1993). Lin and Lin (2005) in their
research use this scales. Impulsive buying scale consists of 9 items. Table 1 shows 9
items of impulsive buying scale.

Table 1 Impulsive Buying Scale


No Item
1. I often buy things spontaneously.
2. "Just do it" describes the way I buy things
3. I often buy things without thinking
4. "I see it, I buy it" describes me
5. "Buy now, think about it later" describes me.
6. Sometimes I feel like buying things on the spur of the moment.
7. I buy things according to how I feel at the moment.
8. I carefully plan most of my purchases (reverse coded)
9. Sometimes I am a bit reckless about what I buy

METHODS

The quantitative part of the study involves a survey of department store consumers
in Surakarta. The survey uses convenience sampling methods. There are 200 respondents
involve in this study. 150 completed questionnaire returned, for a response rate 75
percent. This research uses exploratory factor analysis suggested by Hair et al (1998). In
order to confirm the validity construct, we use confirmatory factor analysis (CFA). The
survey instrument in this article developed from impulsive buying scale by Rook and
Fisher (1995) and Lin and Lin (2005). The reliability or internal consistency in this
research uses Cronbach Alpha, suggested by Nunnally (1978).
CFA in this research is the part of measurement model in Structural Equation
Model. The goodness of fit of the model is GFI,AGFI,CFI, RMSEA and CMIN/DF. The
measurement model in this research, could be states in the mathematical model:
Х= Λxξ + δ
Notes:
Х = Exogen latent variable
Λx = Structural coefficient matrix which related observed exogen variable with its
latent varibable.

6
δ = error term

FINDINGS

We analyze the construct validity of impulsive buying scale with exploratory factor
analyses. Table 2 shows the result of EFA.

Table 2 Factor Loading Impulsive Buying Scale


No Item 1 2 3
1. I often buy things spontaneously. 0,536 -0,491
2. "Just do it" describes the way I buy things 0,310
3. I often buy things without thinking -0,800
4. "I see it, I buy it" describes me 0,928
5. "Buy now, think about it later" describes -0,642
me.
6. Sometimes I feel like buying things on 0,633
the spur of the moment.
7. I buy things according to how I feel at the 0,831 0,331
moment.
8. I carefully plan most of my purchases -0,508
(reverse coded)
9. Sometimes I am a bit reckless about what 0,963
I buy
Source: Data analyses.

Table 2 shows that, by using EFA the impulsive buying scale represent three
dimensions. Ideally, EFA will shows only single dimension which represent a single
latent variable, it is impulsive buying scale. Table 2 shows that impulsive buying scale
proposed by Rook and Fisher (1995) and Lin and Lin (2005) is not well defined. The
operational definition is not match with the concept, so that it defines other construct.
In order to find the internal reliability we test the impulsive buying scale by using
cronbach alpha.

Scale Scale Corrected


Mean Variance Item- Alpha
if Item if Item Total if Item
Deleted Deleted Correlation Deleted

IMP1 21.1067 23.3845 .3545 .6449


IMP2 21.2000 24.0000 .3107 .6537
IMP3 21.2000 22.9664 .3781 .6397
IMP4 21.2533 21.6401 .5226 .6081
IMP5 21.2733 21.8778 .4392 .6250
IMP6 21.1933 23.8080 .2881 .6588
IMP7 19.9333 25.1633 .1315 .6941

7
IMP8 21.2533 23.3448 .2830 .6616
IMP9 21.2933 22.6919 .4717 .6225

Reliability Coefficients

N of Cases = 150.0 N of Items = 9

Alpha = .6728

Nunnally (1978) states that a scale have internal consistency or reliability when
they have cronbach alpha above 0,6. The impulsive buying scale have internal
consistency since it achieve cronbach alpha 0,6728.
We continue our analysis to find the good construct validity of the impulsive buying
scale. Then, we use confirmatory factor analysis to find the best result. Table 3 shows the
best result of CFA.

Table 3 Factor Loading Impulsive Buying Scale


No Item 1 2 3
1 "I see it, I buy it" describes me 0,972
2. Sometimes I am a bit reckless about what 0,972
I buy
Source: Data analyses.

Structural equation model provides solutions for confirmatory factor analysis by


using it measurement model. We test the impulsive buying scale with the measurement
model. Table 4 shows the result of the measurement model of impulsive buying scale.

Table 4 Regression Weight of The Measurement Model


Estimate S.E C.R
Imp1 Impulsive 1,000
buying
Imp2 Impulsive 2.627 2.212 1.188
buying
Imp3 Impulsive 2.680 2.270 1.180
buying
Imp4 Impulsive 10.066 8.018 1.255
buying
Imp5 Impulsive 3.382 2.814 1.202
buying
Imp6 Impulsive 1.209 1.267 0.954
buying
Imp7 Impulsive 0.100 0.902 0.111
buying
Imp8 Impulsive 2.884 2.453 1.176
buying

8
Imp9 Impulsive 8.115 6.449 1.258
buying

Table 5 shows the goodness of fit of the measurement model. It adopt the criteria
proposed by Byrne (2001) dan Bagozzi and Yi (1989).

Table 5 Goodness of Fit of Measurement Model

GOODNESS OF CUT OFF VALUE ESTIMATION DECISION


FIT RESULT
Chi Square Small value 84,226 Good
Probability ≥ 0,05 0,000 Good
GFI ≥ 0,90 0,873 Marginal
AGFI ≥ 0,90 0,789 Marginal
CFI ≥ 0,95 0,829 Marginal
RMSEA ≤ 0,08 0,119 Marginal
CMIN/DF ≤ 2,00 3,119 Marginal

Impulsive buying scale have serious problem in their construct validity. It shows
from the goodness of fit of the measurement model. The problem lies in the concept
definition of impulsive buying which in-appropriate with the operational definition. It
will have a serious effect to the substantive research of this issue.

DISCUSSIONS AND IMPLICATIONS

This research shows the importance of construct validity in the consumer behavior
research. Schwab (1978) shows that research on construct validity in consumer behavior
will give major contributions to the development of consumer behavior theory. However,
little concern about this effort, which result low correlation among consumer behavior
constructs.
Abe et al (1996) shows that in an investigation of construct validity will also give
the answer of reliability, validity and generalizability of a construct. In this research we
analyze the construct validity and reliability of impulsive buying scale. The result shows
that impulsive buying scale proposed by Rook and Fisher (1995) and Lin and Lin (2005)
does not have proper construct validity. It will have a serious effect in using the scale to
measure the antecedents and consequences of impulsive buying behavior.
There are two implications of this research to the development of impulsive buying
concept. First, marketing scholars should pay attention more on construct validity of a
scale before moving further into the test of substantive research. Second, a researcher
should have an anticipation in order to prevent the weakness of construct validity. Swan
and Martin (1994) proposed three steps to overcome the problem. It called Theory-
Setting-Testable Hypotheses (TST). The steps are: theoretical statement and prediction,
determine setting and preposition and determine testable hypotheses.

9
REFFERENCES

Abe, Shuzo, Richard P Bagozzi and Pradip Sadarangani (1996), An Investigation of


Construct Validity and Generalizability of The Self Concept: Self Consciousness in
Japan and the United States in Global Perspectives in Consumer Research, The
Haworth Press Inc pp 97-123..
Armstrong, J Scott (1974), Eclectic Research and Construct Validation in Jagdish N Seth
Models of Buyer Behavior, Conceptual, Quantitative and Empirical, New York,
Harper and Row.
Bagozzi, Richard dan Youjae Yi (1989), On The Use of Structural Equation Models in
Experimental Design, Journal of Marketing Research, Vol XXVI, h 271-184.
Baumeister , Roy F. (2002), Yielding to Temptation: Self-Control Failure, Impulsive
Purchasing, and Consumer Behavior, Journal Of Consumer Research, Vol 28, 670-
676.
Beatty, S. E., & Ferrell, M. E. (1998). Impulse buying: Modeling its precursors. Journal
of Retailing, 74(2), 169-191.
Blum, K. and E.P. Noble (1994). The sobering D2 story. Science 265:1346-1347.
Byrne, Barbara M (2001). Structural Equation Modeling with AMOS, Basic Concepts,
Applications and Programming, Lawrence Erlbaum Associates Publisher, London.
Churcill, Jr Gilbert A (1979), A Paradigm for Developing Better Measures of Marketing
Constructs, Journal of Marketing Research ,Vol XVI pp 64-73.
Clogg, Clifford C. (1992), “The Impact of Sociological Methodology on Statistical
Methodology,” Statistical Science, 7, 183-207.
Cooper, Donald P dan P. S. Schindler (2001). Business Research Methods. 7 th Edition
Boston. McGraw Hill.
Ghiselli Edwin E, John P Campbell dan Sheldon Zedeck (1981), Measurement Theory for
The Behavioral Sciences, New York, W.H Freeman and Company.
Hair, Joseph F., Jr., Rolph E. Anderson, Ronald L. Tatham, and William C. Black (1998),
Multivariate Data Analysis, 5th ed., Upper Saddle River, NJ: Prentice-Hall
International, Inc.
Hirschman, Elizabeth C. (1992), 'The Consciousness of Addiction: Toward a General
Theory of Compulsive Consumption'. Jottrna/ of Consumer Research, 19
(September), 155-179.
Krahn, D. (1991). The Relationship of Eating Disorders and Substance Abuse. Journal of
Studies on Alcohol 3:239-253.
Lin, Chien-Huang and Hung-Ming Lin (2005), An Exploration of Taiwan Adolescents’
Impulsive Buying Tendency, Adolescence in Health & Fitness provided free by
LookSmart Find Articles., www.findarticles.com/p/articles/mi_m2248/
Marlatt, G. Alan, John S. Baer, Dennis M. Donovan, and Daniel R. Kivlahan (1988),
Addictive Behaviors: Etiology and Treatment'. Annual Review of Psychology,
39,223-252.
Neuman, W Lawrence (2000), Social Research Methods, Qualitative and Quantitative
Methods 4th ed, Allyn and Bacon, Boston.
O'Guinn, Thomas C and Ronald J. Faber (1989), "Compulsive Buying:
APhenomenological Exploration." Journal of Consumer Research, 16 (September),

10
147 157.
Puri, R. (1996). Measuring and modifying consumer impulsiveness: A cost-benefit
accessibility framework. Journal of Consumer Psychology, 5(2), 87-113.
Purwanto, BM (2001) Pelatihan Pengukuran dan Teknik Statistik untuk Riset
Keperilakuan, Modul, QUE Project Management dan Program Magister Sains
Ilmu-Ilmu Ekonomi UGM.
Rook, D. W., & Fisher, R. J. (1995). Normative influences on impulsive buying behavior.
Journal of Consumer Research, 22(3), 305-313.
Rook, D. W., & Gardner, M. (1993). In the mood: Impulse buying's affective antecedents.
Research in consumer research (Vol. 6, pp. 1-28). Greenwich, CT: JAI Press.
Schwab, D.P (1978), Construct Validity in Organizational Behavior, in B Staw dan L
Cummings (Eds), Research in Organizational Behavior, 3-43, Greenwich, CT, JAI
Press.
Swan John E dan Warren S Martin (1994), The Theory-Setting-Testable Hypothesis
Model: A Framework to Assist Doctoral Students in Linking Theory and Empirical
Research, Marketing Education Review, Vol 4.

11