4.

RESEARCH METHODOLOGY

4.1. Introduction

Appropriate research assessment and development are needed for appropriate research
methodology. It will lead to good data analysis and more valid and accurate output.

This chapter provides explanations of the method and procedures for conducting this
research. It includes the research design, population, sampling procedures, data
collecting instrument, and statistical techniques used.

4.2. Research Design

A quantitative research approach has been designed and implemented, as this research is
formed “according to the meanings taken from numbers, results collected in a numerical
standardized form and analysis directed by diagrams and statistics” (Saunders, et al.,
2009).

A survey was used to collect the primary data, as it helps to collect data from a large
sample of individuals who are questioned about their opinions and views. The survey
participants were requested to respond to the variables-related questions that helps to test
the research hypotheses, an explanation about the research topic, goals, and research
population, and a targeted sample with some guidelines was provided before the survey
questions. The survey was approved by the Istanbul Aydin University ethical committee.
The research started by finding the main idea and reviewing the previous relevant studies and
articles to improve and develop the idea, till reaching the point of formatting the main research
variables, and forming the research questions and hypotheses and related research conceptual
framework. Then research was designed and the needed data was collected, to get appropriate
answers for the research questions and perform the test of the research hypotheses to generate
results and form the final conclusion.

4.3. Population

This study is discussing the impact of online advertisements, brand awareness, and
demographical factors on consumers’ buying behavior. The targeted population was consumers
living in Istanbul - Turkey.

4.4. Sampling Method

To select the target respondents in this research, a purposive sampling method was used. such a
method is done without using any probability technique, according to the researcher's judgment
based on the criteria that were set to target respondents (Etikan & Bala, 2017).

The targeted sample was 250 or more as according to Varoquaux, (2018), from 200 to 300
observation will lead to 7% errors which is generally accepted in quantitative studies, the total
accepted is 313 responses to the survey questions.

The following graph explained in Reyes & Ghosh, (2013) shows the relation between the sample
size and the margin of error, and will lead to concluding that the more sample size the less err.
 Figure 4.1: Sample size & Margin of Error (Reyes & Ghosh, 2013)

4.5. Data collecting Instrument

This research is considered quantitative research, so the collection of the data was done by a
survey that was adapted from two articles:

- Online advertisements scale was adopted from --------------------(write the reference from
where you bring the scale)
- Brand awareness scale was adopted from --------------------
- Demographical factors scale was adopted from --------------------
- Consumers’ buying behavior scale was adopted from --------------------

For the demographical factors a direct question was asked to the respondents as below:

- Occupation has an impact on consumer buying behavior.

- Age has an impact on consumer buying behavior.
- Education has an impact on consumer buying behavior.
- Material status has an impact on consumer buying behavior.
A 5 point Likert scale (1 = Strongly disagree, 2 = Disagree, 3 = Neutral, 4 = Agree, 5 = Strongly
agree) such close-ended questions type was selected as it needs the minimum writing activity
which makes it easy and less time consuming for the respondent.

The survey starts with an introduction about the researcher, the research main idea and goals, and
an explanation about the targeted respondents, the survey questions started with the demographic
part to get an idea about the target respondent, the second part of the survey includes the main
variables measuring questions, the survey was in English and no translation was done.

4.6. Statistical Techniques

The statistical techniques used for this research, Factor Analysis, and multiple regression.

 Factor Analysis: which finds the factor loading of the scale’s items used, it measures the
relationship between the observed variables themselves and their relationship with their
observed variable. (Byrne, 2012).

 Multiple regression analysis which is a strong statistical analysis that examines the

relationship between two or more variables one dependent and others independent it tests the
effect of one or more independent variables on a dependent variable.

Analysis software used for this research is IBM SPSS version 23 and IBM SPSS AMOS version
22.

4.7. Ethical consideration of the research

Any research needs to be done and implemented with ethics. Ethical considerations were
followed in all the processes of the research, the data collection process of this research was
carried out by a high privacy and confidentiality procedure as the data was only used for the
analysis. The data was not used for personal interest at any stage of the research implementation,
the participant privacy was considered as no name or identity was requested in the survey.
5. ANALYSIS AND DISCUSSIONS

5.1. Introduction
This chapter includes the analysis done for the data collected, it includes the Simple percentage
analysis, factor analysis, and regression analysis with an explanation about the results.

5.2. Simple Percentage Analysis

to understand the targeted respondents and their answers, a Simple percentage analysis was done
such analysis gives an accumulated summary of the respondent’s answers according to the
frequency distribution of the data collected, it is calculated by the following formula: Percentage
= (Number of Respondents * 100) / Total Number of Respondents.

Demographic questions include questions about the characteristics of the research survey
respondents:

Table 5.1: Gender Q1

Frequency Percent
Male 136 43.5%
Female 177 56.5%
Total 313 100.0%

More than half of the respondents are females they represent 57% of the total responses.

Table 5.2: Age Q2

Frequency Percent
18-29 years old 196 62.6
30-40 years old 73 23.3
Over than 40 years old 44 14.1
Total 313 100.0%

More than half of the respondents are less than 30 years old they represent 62% of the total
responses, also 23% are between 30 to 40 years old.
5.3. Descriptive statistics, Skewness, and Kurtosis:

Descriptive statistics explain the basic features of the data in a sample, descriptive statistics are
used. They offer quick summaries of the sample and the steps. They form the basis for nearly all
quantitative analyses of results.

Skewness and Kurtosis are part of the normality assessment which shows whether a data set is
well-modeled by a regular distribution and how likely it is to naturally spread a random variable
underlying the data set (Baghban, et al., 2013).

Skewness, happen when the data distribution is either below or above the mean presenting a non-
asymmetrical distribution around the mean, for not having a Skew the Skew analysis results must
be between 3 and -3. On the other hand, kurtosis happens when the distribution of the data has a
higher or lower peak compared to the normal distribution, the results of the analysis must be
between 8 and -8 to indicate that there is no kurtosis (Kline, 2011).

The table below shows the descriptive mean and standard deviation results, and also presents the
skewness and kurtosis results which indicate that the survey questions’ responses are not skewed
and don’t have kurtosis.

Table 5.3: Descriptive statistics, Skewness and Kurtosis

3>Sk>-3 8>Ku>-8

Mean Std. Deviation Skewness Kurtosis

Online advertisement Q1 3.18 1.367 -.462 -1.175

Online advertisement Q2 3.45 1.132 -.158 -1.423

Online advertisement Q3 3.55 1.365 -.836 -.680

Online advertisement Q4 3.46 1.288 -.697 -.705

Online advertisement Q5 3.51 1.284 -.658 -.754

Online advertisement Q6 3.50 1.321 -.771 -.663

Online advertisement Q7 3.61 1.314 -.789 -.569

Online advertisement Q8 3.66 1.352 -.730 -.752

Table 5.3: Descriptive statistics, Skewness, and Kurtosis Continued.

3>Sk>-3 8>Ku>-8

Mean Std. Deviation Skewness Kurtosis

Online advertisement Q9 3.50 1.284 -.748 -.609

Online advertisement Q10 3.67 1.268 -.894 -.295

Online advertisement Q11 3.54 1.320 -.797 -.597

Online advertisement Q12 3.63 1.284 -.923 -.293

Online advertisement Q13 3.47 1.337 -.621 -.881

Online advertisement Q14 3.57 1.333 -.781 -.660

Online advertisement Q15 3.52 1.318 -.771 -.632

Online advertisement Q16 3.67 1.231 -.946 -.086

Online advertisement Q17 3.51 1.311 -.666 -.769

Online advertisement Q18 3.57 1.292 -.809 -.514

Online advertisement Q19 3.60 1.270 -.813 -.442

Online advertisement Q20 3.56 1.262 -.819 -.438

Online advertisement Q21 3.57 1.267 -.805 -.474

Online advertisement Q22 3.63 1.213 -.929 -.088

Online advertisement Q23 3.59 1.250 -.840 -.342

Online advertisement Q24 3.45 1.344 -.727 -.779

Online advertisement Q25 3.65 1.223 -.958 -.085

Online advertisement Q26 3.56 1.239 -.825 -.375

Online advertisement Q27 3.71 1.254 -.901 -.246

Online advertisement Q28 3.71 1.300 -.957 -.234

Brand awareness Q29 3.72 1.221 -1.032 .121

Brand awareness Q30 3.65 1.252 -.914 -.225

Table 5.3: Descriptive statistics, Skewness, and Kurtosis. Continued.

3>Sk>-3 8>Ku>-8

Mean Std. Deviation Skewness Kurtosis

Brand awareness Q31 3.62 1.263 -.887 -.279

Brand awareness Q32 3.63 1.282 -.941 -.272

Brand awareness Q33 3.68 1.182 -1.020 .125

Brand awareness Q34 3.74 1.155 -1.040 .268

Brand awareness Q35 3.78 1.209 -1.073 .197

Brand awareness Q36 3.80 1.225 -1.115 .293

Brand awareness Q37 3.71 1.284 -.950 -.230

Brand awareness Q38 3.77 1.234 -1.077 .182

Brand awareness Q39 3.65 1.226 -.929 -.124

Brand awareness Q40 3.73 1.204 -.988 .020

Brand awareness Q41 3.62 1.250 -.874 -.309

Brand awareness Q42 3.72 1.228 -1.002 .028

Demographical factors Q43 3.58 1.293 -.819 -.477

Demographical factors Q44 3.56 1.302 -.824 -.475

Demographical factors Q45 3.79 1.183 -.981 .046

Demographical factors Q46 3.59 1.313 -.813 -.516

Consumers' buying behavior Q47 3.67 1.282 -.980 -.162

Consumers' buying behavior Q48 3.64 1.274 -.918 -.283

Consumers' buying behavior Q49 3.65 1.229 -.928 -.165

Consumers' buying behavior Q50 3.64 1.241 -.962 -.120

Consumers' buying behavior Q51 3.60 1.338 -.886 -.436

Consumers' buying behavior Q52 3.74 1.246 -1.013 .001

Table 5.3: Descriptive statistics, Skewness, and Kurtosis. Continued.

3>Sk>-3 8>Ku>-8

Mean Std. Deviation Skewness Kurtosis

Consumers' buying behavior Q53 3.64 1.219 -.932 -.131

Consumers' buying behavior Q54 3.64 1.246 -.991 -.062

5.4. Reliability Assessment (Cronbach alpha test)

Reliability is the method that is done to evaluate the quality of the measurement tools, it is
defined as “to what level the test scores are free from measurement error that occur when testing
something” (Muijs, 2004). If there is an unreliable measurement the relation between variables
will not be significant. One way of testing reliability is the Cronbach alpha test which is “an
internal consistency measure, shows how closely related a set of items are as a group” (Muijs,
2004). When the test results are more than 0.70 then the measurement is reliable, and 0.60 is
accepted (Muijs, 2004).

The reliability test results of each variable of this study are shown below:

Table 5.4: Reliability Assessment (Cronbach alpha test)

Cronbach's Alpha Results

Online advertisement 0.94 Accepted reliability
Brand awareness 0.88 Accepted reliability
Demographical factors 0.70 Accepted reliability
Consumers' buying behavior 0.82 Accepted reliability

5.5. Exploratory factor analysis EFA:

“EFA help to determine to what extent the observed variables are connected to their underlying
factors and to identify the minimal number of factors that account for the covariation among the
observed variables” (Byrne, 2010).
Questions with multiple loading between variables were removed, to get the needed factor
loading. (KMO) factor measure and Bartlett's Test of Sphericity as shown in the below table
were giving a positive indicator for a good loading:
Table 5.5: (KMO) and Bartlett's Test of Sphericity:

KMO Bartlett's Test of Sphericity

Online advertisement 0.923 .000
Brand awareness 0.898 .000
Demographical factors 0.707 .000
Consumers' buying behavior 0.856 .000

The factor loading of the survey questions is presented in table 5.6:

Table 5.6: Exploratory factor analysis EFA

Question Factor loading Question Factor loading

Online advertisement Q1 .494 Online advertisement Q18 .545

Online advertisement Q2 .617 Online advertisement Q19 .585

Online advertisement Q3 .608 Online advertisement Q20 .668

Online advertisement Q4 .607 Online advertisement Q21 .525

Online advertisement Q5 .577 Online advertisement Q22 .617

Online advertisement Q6 .569 Online advertisement Q23 .518

Online advertisement Q7 .639 Online advertisement Q24 .605

Online advertisement Q8 .552 Online advertisement Q25 .571

Online advertisement Q9 .573 Online advertisement Q26 .560

Online advertisement Q10 .584 Online advertisement Q27 .574

Online advertisement Q11 .600 Online advertisement Q28 .600

Online advertisement Q12 .620 Brand awareness Q29 .645

Online advertisement Q13 .603 Brand awareness Q30 .599

Online advertisement Q14 .644 Brand awareness Q31 .621

Online advertisement Q15 .636 Brand awareness Q32 .572

Online advertisement Q16 .645 Brand awareness Q33 .563

Online advertisement Q17 .638 Brand awareness Q34 .478

Table 5.6: Exploratory factor analysis EFA. Continued.

Question Factor loading Question Factor loading

Brand awareness Q35 .585 Demographical factors Q46 .548

Brand awareness Q36 .558 Demographical factors Q45 .421

Brand awareness Q37 .610 Demographical factors Q46 .548

Brand awareness Q38 .580 Consumers' buying behavior Q47 .612

Brand awareness Q39 .560 Consumers' buying behavior Q48 .568

Brand awareness Q40 .574 Consumers' buying behavior Q49 .647

Brand awareness Q41 .615 Consumers' buying behavior Q50 .654

Brand awareness Q42 .537 Consumers' buying behavior Q51 .546

Demographical factors Q43 .716 Consumers' buying behavior Q52 .609

Demographical factors Q44 .744 Consumers' buying behavior Q53 .633

Demographical factors Q45 .421 Consumers' buying behavior Q54 .563

5.6. Multivariate Assumptions

Before testing the linearity and the regression relation between independent and dependent
variables some conditions must be taken into consideration these conditions are called the
Multivariate Assumptions and they are the Normality test and the Multi-collinearity.

5.5.1 Normality Assessment

Normality assessment is the main condition for many statistical tests in this study according to
the below shape and the number of collected data it is concluded that the data is normally
distributed, better distribution can be obtained when more data can be collected:
 Figure 5.1: Normality Assessment

5.5.2. Multi-Collinearity analysis

When independent variables are correlated and not independent from each other there is a Multi-
Collinearity issue, if such correlation between variables is high there will be some difficulties
when interpreting the regression results.

(VIF) and tolerance is used to test Multi-Collinearity: the VIF must be less than 3 and tolerance
must be more than 0.1 to accept that there is no Multi-Collinearity issue (O’Brien, 2007).

In this study, there are three independent variables one variable was put as dependent and the
other two as independents to see the results:
Table 5.7: Multi-Collinearity analysis results

Collinearity Statistics: Part 1

Dependent Variable: Online advertisement Tolerance VIF

Brand awareness .664 1.506

Demographical factors .664 1.506

Collinearity Statistics: Part 2

Dependent Variable: Brand awareness Tolerance VIF

Online advertisement .742 1.348

Demographical factors .742 1.348

Collinearity Statistics: Part 3

Dependent Variable: Demographical factors Tolerance VIF
Online advertisement .379 2.641

Brand awareness .379 2.641

From the table above it is concluded that there is no multi-collinearity for the independent
variables as shown in the results of Tolerance and VIF.

5.5.3 Outliers and influencers

Outliers and influencers are different from the other scores or different responses from the other
respondents, the standard deviation calculation is used to find if there are responses that are far
away from the mean of each response when the result is more than 3 an Outlier is influencing the
results (Kline, 2011).

According to the test done for this research, the only higher amount is less than 0.15 which leads
to conclude that there are no Outliers and influencers in the responses of this study
 Figure 5.2: Outliers and influencers Assessment

Regression Analysis for Hypotheses Testing

Regression is defined as a predictive analysis tool that examines the relationship between
independent and dependent variables, to fit a mathematical function describing how the value of
the response changes when the values of the predictors vary

Linear regression is the simplest form of regression, which in the case of multiple regression,
works on explaining the data by simply fitting a hyperplane minimizing the absolute error of the
fitting.

Regression Analysis for Hypotheses Testing result is the P-value that is between 0 and 1and, for
accepting or rejecting a hypothesis we look at the cutoff point 0.05 if it is less the hypothesis is
accepted if more it is rejected (Hair Jr, et al., 2014).

The Regression Analysis for Hypotheses Testing results is shown in the list of tables below
table:
Table 5.8: R Square table

Adjusted R Std. Error of the

Model R R Square Square Estimate
1 .778 a
.605 .597 .57621704
a. Predictors: (Constant), Demographical factors Q46 Material status ,
Demographical factors Q45 Education , Demographical factors Q43
Occupation, Online advertisement, Demographical factors Q44 Age,
Brand awareness

R-squared shows how well the regression model fits the observed data. in this research R-
squared is 61% reveals that 61% of the data fit the regression model.

Table 5.9: ANOVA table

ANOVAa
Model Sum of Squares df Mean Square F Sig.
Regression 155.594 6 25.932 78.103 .000b
1 Residual 101.600 306 .332
Total 257.194 312

a. Dependent Variable: Consumers' buying behavior

b. Predictors: (Constant), Demographical factors Q46 Material status , Demographical factors

Q45 Education , Demographical factors Q43 Occupation, Online advertisement, Demographical
factors Q44 Age, Brand awareness

Analysis of Variance (ANOVA) consists of calculations that provide information about levels of
variability within a regression model and form a basis for tests of significance, getting a P-value
of 0.000 which is less than 0.05 shows that there is variability within the research regression
model.
 Coefficientsa

Standardized
Unstandardized Coefficients Coefficients

Model B Std. Error Beta t Sig.

(Constant) -.836 .175 -4.764 .000

1
Online advertisement .241 .055 .257 4.350 .000

Brand awareness .372 .061 .384 6.092 .000

Demographical factors Q43

.044 .032 .062 1.350 .178
Occupation

Demographical factors Q44

.085 .031 .122 2.699 .007
Age

Demographical factors Q45

.037 .032 .048 1.152 .250
Education

Demographical factors Q46

.067 .029 .096 2.285 .023
Material status

a. Dependent Variable: Consumers' buying behavior

Table 5.10: The Regression Analysis for Hypotheses Testing results

The P-value, which is less than 0.05 results lead to conclude that there is an effect of the Online
advertisement, Brand awareness, Age, and Material status on consumer buying behavior.

5.5. Hypothesis Results Summary

Table 5.12: Hypothesis Results Summary

Relationships P-value Interpretation

H1 Online advertisement  Consumers' buying behavior .000 Accepted

H2 Brand awareness  Consumers' buying behavior .000 Accepted

H3. Rejected
Occupation  Consumers' buying behavior .178
1
H3. Accepted
Age  Consumers' buying behavior .007
2
H3. Rejected
Education  Consumers' buying behavior .250
3
H3. Accepted
Material status  Consumers' buying behavior .023
4
The table above shows the summary of the hypothesis results it shows that:
a. There is significant impact from online advertisement on consumers’ buying behavior.
b. There is significant impact from Brand awareness on consumers’ buying behavior.
c. There is no significant impact from Occupation on consumers’ buying behavior.
d. There is significant impact from age on consumers’ buying behavior.
e. There is no significant impact from Education on consumers’ buying behavior.
f. There is significant impact from Material status on consumers’ buying behavior.
References

Journals:

Baghban, A. A., Younespour, S., Jambarsang, S., Yousef, M., Zayeri, F., & Jalilian, F. A. (2013). How to
test normality distribution for a variable: a real example and a simulation study. Journal of
Paramedical Sciences (JPS), 4(1), 73-77.
Etikan, I., & Bala, K. (2017). Sampling and Sampling Methods. Biometrics & Biostatistics International
Journal, 5(6), 1-3.
O’Brien, R. M., 2007. A Caution Regarding Rules of Thumb for Variance Inflation Factors. Quality &
Quantity, Volume 41, p. 673–690.
Reyes, E. M., & Ghosh, S. K. (2013). Bayesian Average Error-Based Approach to Sample Size
Calculations for Hypothesis Testing. Journal of Biopharmaceutical Statistics, 23(3), 569-588.
Varoquaux, G. (2018). Cross-validation failure: Small sample sizes lead to large error bars. NeuroImage,
180, 68–77.

Books:

Byrne, B. M. (2010). Structural Equation Modeling with AMOS Basic Concepts, Applications, and
Programming (2nd ed.). New York: Taylor and Francis Group, LLC.
Byrne, B. M. (2012). Structural Equation Modeling with Mplus Basic Concepts, Applications, and
Programming (2nd ed.). New York: Taylor & Francis Group, LLC.
Hair Jr, J. F., Black, W. C., Babin, B. J., & Anderson, R. E. (2014). Multivariate Data Analysis (7th ed.).
England: Pearson Education Limited.
Kline, R. B. (2011). Principles and Practice of Structural Equation Modeling (3ed ed.). New York: THE
GUILFORD PRESS.
Muijs, D. (2004). Doing Quantitative Research in Education with SPSS (1st ed.). London: Sage
Publications.
Saunders, M., Lewis, P., & Thornhill, A. (2009). Research methods for business students (5th ed.).
Harlow: Pearson Education Limited.