Page 1 of 11
Journal for Studies in Management and Planning
Available at http://edupediapublications.org/journals/index.php/JSMaP/
e-ISSN: 2395-0463
Volume 02 Issue 7
July 2016
Available online: http://edupediapublications.org/journals/index.php/JSMaP/ P a g e | 92
Data Cleaning and Preliminary Analysis for a Study
Linking Entrepreneurial Orientation, Total Quality
Management, Organisation Learning and Competitive
Intensity to SME Performance
Ramatu Abdulkareem Abubakar ; Rosli Mahmood
School of Business Management, College of Business Universiti Utara Malaysia, 06010
UUM Sintok Kedah Darul Aman, Malaysia
E-mail: ramatuabdulkareem@gmail.com E-mail: roslimahmood853@gmail.com
Abstract
This article is aimed at exploring the data
collected regarding a study linking
entrepreneurial orientation, total quality
management, organisation learning and
competitive intensity to SME performance
in manufacturing sector. Four hundred
and eight small and medium enterprises
(SMEs) operating in Kano and Kaduna in
the in north-west geopolitical zone of
Nigeria were included in this study. The
data was analysed using Predictive
Analytics Software (PASW) version 22. An
initial data cleaning and preliminary
analyses were performed in order to
confirm that the key multivariate
assumptions are not violated before
conducting the main analyses. Specifically,
the assessment of missing values, outliers,
normality test, linearity test,
multicollinearity test, and
homoscedasticity test were performed. It
was concluded that the data was fit for
further multivariate analysis.
Keywords: Data Cleaning, entrepreneurial
orientation, competitive intensity,
organisation learning, SME performance
total quality management
1. Introduction
Before conducting the main analyses of
interest, it was necessary to screen the raw
data for missing values, multivariate
outliers, normality, linearity,
homoscedasticity, and multicollinearity.
This was done to confirm that the key
multivariate assumptions are not violated
before conducting the main analyses.
Therefore, in the following section the key
assumptions, including missing values,
multivariate outliers, normality, linearity,
homoscedasticity, and multicollinearity are
explored. Additionally, initial data
screening helps researchers to better
understand the data collected for further
analysis. To date, scientific investigations
have been largely been carried out without
initial data cleaning and preliminary
analysis probably due to the burden
attached to it (Hair, Anderson, Tatham, &
Black, 1998; Hair, Black, Babin, &
Anderson, 2010; Tabachnick & Fidell,
2007). Such neglect is unfortunate
because, overlooking of the initial data
screening would seriously inflate the
standard error estimates, which in turn
underestimate the statistical significance of
a regression coefficients (Dong & Peng,
2013; Graham, 2009; Peng, Harwell, Liou,
& Ehman, 2006; Roth, 1994; Schlomer,
Bauman, & Card, 2010).
Page 2 of 11
Journal for Studies in Management and Planning
Available at http://edupediapublications.org/journals/index.php/JSMaP/
e-ISSN: 2395-0463
Volume 02 Issue 7
July 2016
Available online: http://edupediapublications.org/journals/index.php/JSMaP/ P a g e | 93
Against this background, the remainder of
the paper is organized as follows. In
section 2, we highlighted the method used
in the present study, followed by
presentation of the results in section 3. In
the final section, conclusion was drawn
based on the results of data cleaning and
preliminary analyses.
2. Methods
2.1 Sample and procedures
The participating firms in this study were
four hundred and eight small and medium
enterprises (SMEs) operating in Kano and
Kaduna’s manufacturing sectors in the
north-west geopolitical zone of Nigeria.
Specifically, seven main industries in the
manufacturing sector of these cities
include food and beverages,
packaging/containers, metal and metal
products, printing and publishing, agro- allied, furniture, and building materials.
The manufacturing sector was covered in
this study because the sector has been a
major driver of Nigeria's economic growth
and it accounted for a substantial
proportion of total economic activities in
Nigeria.
2.2 Measures
2.2.1 Entrepreneurial orientation
Entrepreneurial orientation was assessed
using Covin and Slevin’s (1989)
entrepreneurial orientation scale.
Specifically, this scale contains nine items,
of which three items were designed to
measure the innovativeness dimension of
entrepreneurial orientation, three items to
assess risk taking, and the remaining three
to measure proactiveness. Entrepreneurial
orientation was rated using seven-point
Likert scale ranged from 1 = strongly
disagree to 7 = strongly agree. Thus, a low
score indicates a more conservative
entrepreneurial orientation, while high
score suggests a more entrepreneurial
orientation. Sample item include: “In our
firm, changes in product or service lines
have usually been quite dramatic.
2.2.2 Total Quality management
A seven-item scale developed by Chenhall
(1997) was used to measure total quality
management in this study. Specifically, the
items in this scale asked the participants to
indicate the extent to which their firms
have implemented programmes over the
past three years to improve the quality of
products and processes, efficiency,
minimizing waste, as well as involving
employees in the continuous improvement.
Additionally, total quality management
was rated by the participants using seven- point Likert scale ranged from 1 = strongly
disagree to 7 = strongly agree. Sample
item include: “Our firm strongly
encourages involvement of employees in
quality improvement programs (e.g.
training, involvement in improvement
teams)”.
2.2.3 Organizational learning
Organizational learning was assessed using
4-item measure of organizational learning
developed by García‐Morales, Llorens‐
Montes, and Verdú‐Jover (2006). A
seven-point Likert scale ranging from 1 =
strongly disagree to 7 = strongly agree
was used by the participants to rate extent
of their agreement with statements
describing organizational learning
practices in their organizations. Sample
item include: “Our firm’s performance has
been influenced by new learning it has
acquired.
2.2.4 Competitive Intensity
To assess competitive intensity, Jaworski
and Kohli’s (1993) Competitive Intensity
Scale was administered. Competitive
intensity was assessed with six items, such
as “There are many "promotion wars" in
our industry”, and “our competitors are
relatively weak”. Participants were asked
to use a seven-point Likert scale ranged
Page 3 of 11
Journal for Studies in Management and Planning
Available at http://edupediapublications.org/journals/index.php/JSMaP/
e-ISSN: 2395-0463
Volume 02 Issue 7
July 2016
Available online: http://edupediapublications.org/journals/index.php/JSMaP/ P a g e | 94
from 1 = strongly disagree to 7 = strongly
agree to rate extent of their agreement
with statements describing the intensity of
competition in their industry.
2.2.5 Organizational Performance
Six-item were used to assess a broad range
of SME' performance indicators. Five
items will be adapted from the work of
Powell (1995), and the remaining item will
draw from the work of Baker and Sinkula
(1999). A Sample item is: Over the past 3
years, there has been change in market
share relative to our competitors".
Responses were on a seven-point Likert
scale ranged from 1 = strongly disagree to
7 = strongly agree.
3. Results and discussion
Prior to initial data cleaning, all the
returned and usable questionnaires were
coded and entered into the Predictive
Analytics Software (PASW) version 22. In
addition, all the negatively worded items
in the questionnaires were reversed. Prior
to data coding and entry. Subsequently, the
results of the preliminary analyses are
reported and discussed.
3.1 Assessment of Missing value
According to Tabachnick and Fidell
(2007), missing value “is one of the most
pervasive problems in data analysis... its
seriousness depends on the pattern of
missing data, how much is missing, and
why it is missing” (p.62). Researchers
have argued that overlooking cases with
missing values could have serious impact
on quantitative research, leading to biased
estimates of parameters, loss of
information, decreased statistical power,
increased standard errors, as well as
weakened generalizability of findings.
(Dong & Peng, 2013; Graham, 2009;
Peng, Harwell, Liou, & Ehman, 2006;
Roth, 1994; Schlomer, Bauman, & Card,
2010).
Although there is no universally threshold
on how much missing data can be tolerated
for a given sample size and a valid
statistical analysis (Tabachnick & Fidell,
2007), researchers, particularly Schafer
(1999) has asserted that a missing rate of
5% or less is of no importance in
multivariate analysis. While Schafer
(1999) is somehow more conservative
regarding the rate of missing value in a
dataset, Bennett (2001) contended that
when the rate of missing value is more
than 10% , results of subsequent statistical
analyses may invalid and biased. One of
the popular statistical methods for
replacing missing values is “the mean
scores of all other subjects for that
variable” (George & Mallery, 2001, p. 46).
In the present study, this statistical method
was performed using the Statistical
Package for the Social Sciences (SPSS).
The results for missing values analysis are
presented in Table 1.
