CHAPTER III
RESEARCH METHODOLOGY
3.0 Introduction
Research methodology, as defined by Warren and Jansen (2020), is concerned with the
feasibility of any particular research. This chapter emphasized the methodology techniques
discussed in Chapter 1 and the body of literature review discussed in Chapter 2. It discussed
the study's research philosophy, research approach, research design, and research strategy. It
also defined the target population, sample frame, sample size determination, and sampling
methods and techniques. Furthermore, it described the data sources, research instruments, and
data collection methodologies and processes used by the researcher in the approach to
determine the impact of digitalization of tax systems on revenue collection. The researcher
also gave the study's validity and reliability tests. The chapter summary articulated the data
analysis and display techniques.
3.1 Research Onion
The research onion, proposed by Saunders et al. (2016) is a tool, which helps to organize the
research and develop research design following the layers of the research onion
systematically. It underpins specific ontological and epistemological assumptions, which lead
to choice of strategies, techniques and methods different from ones used in business studies.
Figure 3.1 shows the research onion.
Source: Saunders et al. 2019
Figure 3. 1 Research Onion
3.2 Research Philosophy
Research philosophy is concerned with the development of knowledge and its nature as given
in the works of Saunders et al., (2019). It is concerned with beliefs or assumptions made
about the data collected concerning a phenomenon, analysed, and applied to a study
(Zukauskas, 2018; Bell et al., 2022; Saunders et al., 2019). According to Burrell and Morgan
(2016), these assumptions focus on reality factors encountered during the research
(ontological), human knowledge (epistemological), and how the researcher's values influence
the research process (axiological). The researcher relied on a well-thought-out set of
assumptions to come up with credible research philosophy, to underpin the research
methodology, research design, research strategy, and data collection and analysis techniques
(Zukauskas, 2018).
Studies reveal that there is a manifold of philosophies that can guide data collection and
analysis; positivism, interpretivism, critical realism, postmodernism, and pragmatism
(Ibanga, 2020; Zukauskas, 2018; Saunders et al 2019). Critical realism is concerned with
reality which it regards as the most important philosophical consideration, however, critical
realism was not applied as it mildly relies on subjectivist knowledge. Interpretivism is
focused on the study of human meanings’ and assumes that humans are different from
physical phenomena due to creating meanings (Ibanga, 2020; Bell et al., 2022; Saunders et
al., 2019). Interpretivism was not applicable in this study as it was subjective in nature and is
by origin a critique philosophy to quantitative research. Pragmatism strives to combine
objectivism and subjectivism, facts and values underpinning a study (Saunders et al., 2019;
Ansari et al., 2016). Therefore, pragmatism was not applied as it still contained more
elements of subjectivism which the researcher found to be less applicable in the mixed
research.
3.2.1 Positivism research philosophy
Positivist philosophical thinking suggests that knowledge can be gathered objectively, using
facts and or statistical information that can be proved scientifically (Saunders et al., 2019;
Bryman and Bell, 2016). Therefore, for this study, the research was conducted based on the
positivism philosophy. In adopting the positivist philosophical stance, the researcher focused
on the research problem in which the research purpose was guided by the four research
objectives. Considering the nature of the research purpose where the research objectives are
deterministic, and in quest of establishing causal relationships between the independent
variable, digitalization and revenue collection the dependent variables positivism is deemed
most relevant and applicable (Ibanga, 2020). Furthermore, considering the above positive
attributes, the use of the positivist philosophical stance allowed the researcher to eliminate
subjective thinking or opinions that would lead to biased information.
3.3 Research Approach
The research approach following Pettey (2017) entails techniques and tactics for research that
can shift judgments from broad hypotheses to comprehensive techniques for information
gathering and analysis. According to reviewed literature, there are two types of research
approaches, the deductive approach, and the inductive approach (Saunders et al., 2019;
Creswell, 2018; Ibanga, 2020). The researcher adopted the deductive approach as the study
was guided by a positivist philosophy anchoring the study, thus, objectivism of the research
was paramount whereas an inductive approach would have negatively affected the credibility
of the quantitative research as it relies on subjectivism (Bryman and Bell, 2016). In addition,
various studies suggest that a deductive approach is essential as it allows the researcher to
develop a research design and a research strategy to test the hypothesis and theory(s) for the
study.
Furthermore, the deductive approach allows for the generalisation of statistical findings on
regularities in human social behaviour and structures the methodology to facilitate
replication, which enhances reliability. Therefore, the researcher adopted this approach as it
ensured the credibility of findings as the study relied on both quantitative and qualitative data
collection methods.(Saunders et al. 2019). The diagram in Figure 3.1 depicts the research
onion from which the researcher sourced the appropriate research philosophy, research
approach, research design, research strategy, and data collection methods and analysis
techniques.
3.4 Research design
Research design is referred to as, the set of methods and procedures used in collecting and
analysing measures of the variables specified in the research problem (Kirumbi, 2018).
According to Saunders et al., (2019) researchers can adopt an exploratory research design,
causal-explanatory research design, and a descriptive research design depending on the
research purpose and philosophy underpinning the study. Lissa (2018) suggests that scholars
should appreciate that the selection of the research design is guided by the research
philosophy and the research purpose.
3.4.1 Causal-Explanatory research
The causal-explanatory research design is concerned with establishing the causality
relationship between the variables underpinning a study (Saunders et al., 2019). It is focused
on studying a situation, or a problem to express the relationship between the independent
variable and the dependent variable. In this regard, the researcher adopted a causal-
explanatory research design, which was guided by the positivist philosophy and the research
purpose aimed to determine the impact of digitalisation of tax system on revenue collection.
The causal research design was employed, as it was relevant to address the extent to which
digitalization of tax can enhance tax revenue collection at ZIMRA. In addition, causal
(explanatory) research design enabled the researcher to collect and analyse data using
quantitative methods, which reduce subjectivity associated with qualitative methods (Bell et
al., 2022; Saunders et al., 2019). It is also important to note that the cross-sectional method
guided data collection as the study was academic, thus, meant to fulfil the requirements of the
degree program under study.
3.5 Research Methodology
Research methodology is a very creative endeavour that investigates the type and way in
which data is acquired, assessed, and understood. It specifies the research activities,
objectives, methodologies, measures, and applications (Vargas-Hernandez, 2016). Three
research approaches were presented to the researcher: quantitative research, qualitative
research, and mixed-method research (Ansari et al., 2016; Saunders et al., 2019).
Qualitative Approach
There are different definitions for qualitative research. In general, these methods aim to
address societies’ scientific and practical issues and involve naturalistic and interpretative
approaches to different subject matters. These methods utilize various empirical materials
such as case studies, life experiences, and stories that show the routines and problems that
individuals are struggling with in their lives through focusing on their in-depth meaning and
motivations which cannot be defined by numbers. Qualitative research aims to collect
primary, first-hand, textual data and analyse it using specific interpretive methods. It is a
useful method in studying a phenomenon with limited accessible information as its nature is
exploratory.
Quantitative Approach
Quantitative research is the method of employing numerical values derived from observations
to explain and describe the phenomena that the observations can reflect on them. This method
employs both empirical statements, as descriptive statements about the meaning of the cases
in real words not about the ought of the cases, and methods. It also applies the empirical
evaluations intending to determine to which degree a norm or standard is fulfilled in a
particular policy or program. Finally, the collected numerical data is analysed using
mathematical methods.
Mixed-method Approach
Mixed-method methods simply employ a combination of both qualitative and quantitative
approaches based on the purpose of the study and the nature of the research question aiming
to provide a better understanding of the subject. However, the focus can be on both methods
equally, or on one of the methods considering the selected integration process. Utilizing the
integration of both methods can help researchers to address complex research circumstances
in different research fields such as social and health research. As these methods cover the
advantages of both qualitative and quantitative methods, they can be useful in case that
employing one of the approaches is not adequate in a study.
In cognisance of these three approaches, the researcher settled for a mixed research approach,
which conducts research using both quantitative and qualitative data, empirical procedures,
and empirical claims (Bell et al., 2022). The researcher used the mixed methods technique
because the study was influenced by a natural scientist's positivist philosophical attitude.
With mixed methods study the possibility of replicating the study over time due to the
standardized methods. This was the main reason for settling for this methodology.
3.5 Research Strategy
The research strategy adopted can be guided by the nature of the research purpose, objectives,
and philosophical stance, amongst other factors (Saunders et al., 2019). The main types of
research strategies include experiment, archival research, case study, grounded theory,
survey, action research, and ethnography. Thus, for this study, the research strategy adopted
was a single case study on NBS Harare. The researcher was guided by the positivist
philosophical stance which encompassed a deductive research approach that called for a need
to test the causal relationship between the independent variable (digitalization of tax system)
and the dependent variable (revenue collection).
3.5.1 Case Study
A case study is defined as a research strategy that takes in an empirical investigation of a
particularly modern phenomenon within its real-life context employing multiple sources of
evidence (Schoch, 2020; Saunders et al., 2019). The researcher focused on a single case study
of ZIMRA Kurima House as the researcher is familiar with the place. A single case study was
suitable as it allowed the researcher the opportunity to observe and analyse a phenomenon
that few scholars have considered exploring further within the banking sector.
3.6 Target Population
The target population is defined as the group of individuals or participants with specific
attributes of interest and relevance (Asiamah et al, 2017; Creswell, 2018). According to
Emory and Cooper (2016) the general population is the general pool from which we can
identify our target population. In addition, Best and Kahn (2019) define population as any
group of individuals who have one or more characteristics in common. The population for
this study comprised of all the eighty (80) ZIMRA workers at Kurima House.
In selecting a representative sample, the researcher used simple random sampling method
which is a probability sampling method. According to Elliott (2020), a simple random
sampling technique entails developing an accurate sampling frame, selecting elements from
the sampling frame, and locating the exact element that will be selected for inclusion in the
sample. The researcher used this technique to randomly select participants for the study as the
researcher has no guarantee that all the respondents forming the sample for this study are
going to be available for participation in the study upon engagement.
3.7 Sample Frame
A sample frame is defined as a specific list that closely approximates all the elements in a
population; a researcher operationalizes a population by developing this specific list known
as a sample frame (Elliott, 2020). For this study, the researcher relied on the employee
database from the (Human Resources Department). The employee database was key to
providing the number of employees from all departments as these respondents were familiar
with the digitalisation initiatives undertaken by ZIMRA and therefore, were essential to the
credibility of results.
3.7.1 Sample Frame and Procedure
The researcher collected data concerning the impact of digitalisation of tax system on
revenue collection from employees (both non-managerial, and managerial). The managerial
respondents were engaged for the purpose of obtaining credible data based on international
trends, whereas non-managerial respondents were also key in terms of years of work
experience as they provided the researcher with data based on current trends.
3.7.2 Sample Size Determination
According to Bukhari (2020) determining the appropriate sample size for the study is the
crucial stage of the research process as it affects generalizability in quantitative studies while
affecting credibility in qualitative studies. The researcher focused on the use of the sample
size determination table by Krejcie and Morgan (1970), which is used to measure quantitative
sample sizes. The table by Krejcie and Morgan (1970) states that for a target population size
of 700, the sample size of 248 was appropriate for this study. Therefore, the researcher was
guided by this information in determining the sample size appropriate for the study.
Figure 3. 2 Sample Size table
The researcher employed the use of Krejcie and Morgan's (1970) sample size table in
determining the sample size for the study. As highlighted by Krejcie and Morgan (1970), for
a target population size (N) of 80, the sample size (S) of 66 was appropriate for carrying out a
study. Therefore, for this study, the researcher adopted the sample size (S) of 66 respondents
as they were adequate for this study and emanated from a target population of respondents
that were familiar with digitalisation activities at ZIMRA. The questionnaires distributed
were targeted to 54 employees and interviews were held with 12 managers to understand their
perception of the impact of digitalisation of tax systems on revenue collection. This was in
line with scholar Maxwell (2016) who alluded that a case study researcher typically gathers
data from a small number of individuals to obtain a deeper understanding of an issue.
Table 3.2 Sample size
Cluster Population Sample
Employees 70 54
Management 18 12
3.8 Sampling methods and techniques
According to Etika and Bala (2017), sampling methods or techniques are a process of
choosing or selecting a group of respondents from the target population. The selected group
then acts as a representative of the target population. The researcher was exposed to two
types of sampling methods; non-probability and probability sampling techniques (Saunders et
al., 2019). Non-probability sampling technique is a qualitative sampling technique in which
each unit within a target population has no specifiable probability of being selected (Elliot,
2020). It comprises convenience sampling, quota sampling, purposive sampling, and
snowball sampling (Elliot, 2020; Sartedt et al., 2017).
3.8.1 Probability sampling technique
According to Sartedt et al., (2017) probability sampling technique is one whereby each
employee within the target population has a specifiable chance of being selected and involves
randomly selecting respondents from the target population to partake in the study. Sartedt et
al., (2017) suggest that the most dominant probability sampling techniques over the years are
simple random, stratified random, systematic sampling, cluster, and area sampling
techniques. In regard to the study on the impact of digitalisation on tax systems on revenue
collection, the researcher focused on both probability and non-probability sampling
techniques because the study was mixed in nature.
Furthermore, the researcher focused on cluster sampling to divide the respondents into two
relevant clusters then over and above that, the researcher employed the use of a simple
random sampling technique within the two clusters as the researcher had no assurance upon
which respondents would be available to participate in the study.
[Link] Cluster sampling
The cluster sampling technique entails dividing a target population size into clusters or
groups (Taherdoost, 2016). The stages of cluster sampling encapsulate the following:
selecting a cluster grouping for the sample frame (e.g. type of company or geographical
region); numbering each of the clusters (e.g. cluster 1 & cluster 2) and lastly drawing a
sample using random sampling (Taherdoost, 2016; Thomas, 2022).
For this study, the researcher divided the target population into two clusters, cluster 1:
Employee respondents of ZIMRA Kurima House and cluster 2: management respondents
from the management. These two clusters formed the characteristics of the population and
generated a sample that represented the target population (Thomas, 2022). After establishing
two clusters for the study, the researcher proceeded to draw a random sample from each
cluster, for data collection purposes concerning digitalisation of tax system and revenue
collection. Cluster sampling was adopted as it is easier to implement and is a cost-effective
sampling method as compared to stratified sampling.
[Link] Simple random sampling
According to Elliott (2020), a simple random sampling technique entails developing an
accurate sampling frame, selecting elements from the sampling frame, and locating the exact
element that was selected for inclusion in the sample. After establishing two clusters for the
study, the researcher proceeded to draw a simple random sample from each cluster, cluster 1:
employee respondents, and cluster 2: management respondents.
The researcher employed this technique to randomly select participants for the study as the
researcher had no guarantee that all the respondents forming the sample for this study were
going to be available for participation in the study upon engagement. In addition, simple
random sampling makes use of randomization, thus, it results in high internal and external
validity of findings to the population (Thomas, 2022). Due to the causal-explanatory nature
of research, potential confounding variables had to be reduced to avoid blighting the internal
validity, hence, simple random sampling was appropriate for this study (Thomas, 2022;
Thomas, 2020).
3.9 Data Sources
According to Saunders et al., (2019) there are two types of data sources used for conducting
research, secondary data, and primary data. Primary data is concerned with new information,
which is first in hand, and comprises interviews, quantitative data, and transcripts, amongst
others (Steefkerk, 2022). Secondary data refers to second-hand information that has been
commented on by other researchers and encapsulates, journal articles, e-books, and
dissertations, amongst others (Steefkerk, 2022). Both data sources were used as it results in
credible research results.
3.9.1 Secondary Data
According to Ajayi (2017), secondary data is information previously gathered by someone
other than the researcher. For this study, the researcher first made use of secondary data
collected from the Midlands State University Institutional Repository, ResearchGate, and
Academia. Then proceeded to ZIMRA annual reports, employee, and customer database.
Annual reports provided the research with overall brand performance in sales value; sales
analysis, and brand market share to construct a research problem concerning minority brands.
To determine the target population and sample frame to use for the case study, the researcher
obtained an employee database from the Human Resources department. The research also
made use of literature from Research-Gate, Academia, and Midlands State University
Institutional repository library from e-journals published articles, and dissertations. The
outlined secondary data sources were key in defining the research problem as well as in
situating the study in the context of the tax administration.
3.9.2 Primary Data
Primary data was collected by the researcher directly from the main source to carry out
research. Saunders et al., (2019) suggest that primary data entails information collected for
the purpose at hand, and may include data sources such as questionnaires, and pilot studies
amongst others. The researcher collected primary data and the research required first-hand
evidence (Steefkerk, 2022). Due to the nature of the mixed methods research, statistical data
was required, hence, the researcher made use of structured questionnaires with closed-ended
questions, and interviews to the target population at ZIMRA Kurima House.
A pilot study was adopted to conduct pretests on the structured questionnaire with colleagues
in an attempt to establish if any constant responses were generated. Sierra et al., (2016)
defined structured questionnaires as analogous to the multiple questions favoured by the
researcher and can only be answered by choosing from a limited number of options, usually
using a single-word answer, ‘yes or no’, or a rating scale (from strongly agree to strongly
disagree). Therefore, in this study, the researcher used the rating scale applying it alongside a
5-point Likert scale ranging from strongly agree to strongly disagree.
3.10 Research Instruments
According to Trigueros et al., (2017) research instruments are measurement tools subdivided
into qualitative and quantitative research instruments. Qualitative research instruments are
used to observe variables under study, and they include observations, interviews,
questionnaires, surveys, history, and focus groups. On the other hand, quantitative research
instruments are to generate statistical values about the variables under study and they include
questionnaires, surveys, case studies, and dichotomous questions, multiple choice, and rating
scales (Trigueros et al., 2017).
The research was conducted using a structured questionnaire with closed-ended statements
and interview guide. Sierra et al, (2016) define structured questionnaires as corresponding to
the multiple-choice questions preferred by the researcher. This type of questionnaire was
structured with a scale to ensure that it is only answered by selecting from a limited number
of options, using a 5 Point Likert-scale rating from strongly disagree to strongly agree.
These research instruments were selected because the researcher will be efficient in mixed
methods data collection as they allowed for a speedy response from the respondents, easier to
answer, and data is quickly coded, entered, and analysed, thus allowing the researcher to
complete credible research within the time frame. Also, the research instruments ensured
statistical data was relevant to testing the credibility of the conceptual framework developed
and the causal relationship of variables.
3.10.1 Questionnaire Design
The research was conducted by employing the use of two structured questionnaires one for
each set of respondents which made use of a 5-point Likert scale with the rating scale of
strongly agree to strongly disagree (Sierra et al, 2016). The research instruments were
categorised into the following sections of the questionnaire, Section A- focused on the
demographics of the respondents, whereas Sections: B, C, D, and section E, focused on
answering or probing specific questions relating to the study.
The structured questionnaire was adopted based on its ability to solicit fact-based responses
as well as commanding the power to reduce subjectivity associated with qualitative
instruments such as in-depth interviews and focus groups (Bryman and Bell, 2016). The 5
points Likert scale was employed as it is short, simple, straightforward, forward, and
understandable, thus, the study adopted its use to find the respondents' attitudes and opinions
on the impact of digitalization of tax systems on revenue collection.
3.11 Data Collection Procedure and Administration
The study aimed at collecting data from both employees and management at ZIMRA Kurima
House in Harare. The researcher first obtained and printed a letter of research approval
obtained from the Midlands State University to conduct the research. The researcher then
submitted the research letter to the Human resources management at the company to receive
their consent to conduct a case study on the company.
Upon approval, the researcher applied research ethics, to politely ask the respondents
(employees) at the company and the customer respondents in the target market to participate
in the study and assured them it was based on voluntary action. The researcher proceeded to
distribute self-administered structured questionnaires with closed-ended statements in person
to the respondents upon approval, and these were structured to last a period of approximately
10-15 minutes, as tried via a pilot test. The participants who could not complete the
questionnaire on the same day of receipt were given consent by the researcher to submit upon
completion. The researcher collected the questionnaires and safely guarded them in a secure
flat file upon completion of the data collection process.
Interviews were held with the management. Thirty minutes were allocated to each interview
as detailed information was obtained using this method.
The non-usable questionnaires were excluded from the actual response rate for credibility and
integrity of academic research according to non-response questionnaire factors decreed. The
complete statistical data set was then computed into the SPSS version 26 for analysis and
presentation.
3.14 Data Analysis and Presentation tools
Data analysis is the process of making sense of data collected and discovering what the
respondents have to say. The aim of data analysis in this study was to understand the
constituent parts and identify the potential patterns that exist in the study (Dahl et al., 2016).
The researcher used the Statistical Package for Social Science (SPSS) version 26 and
Microsoft Excel (2013) to edit, code, and check the captured data (George and Malley
(2019). According to Pallant (2016) SPSS is essential for quantitative research as it makes it
easier to understand findings, test hypotheses, and present the findings through statistical
tables. Therefore, SPSS version 26 was adopted for the study.
Descriptive statistics were employed to summarise, organise, and analyse the demographic
data of the respondents sampled and were also vital in analysing the patterns of responses
concentrating on std. deviation, median, and the mean (statistical analysis). Inferential
statistics were performed to assess if the findings were generalizable to the population.
Inferential statistics were key to the hypothesis testing of the study. Pearson's Correlation
analysis was used to evaluate the strength of the relationships between independent and
dependent variables (Pallant, 2011).
Bivariate linear regression analysis was done to analyse data since the research was anchored
on the independent variable and dependent variable constructs: audience targeting, location
targeting, and weather targeting indicating their association to brand awareness. The research
used data analysis techniques that were applicable in quantitative research. The main reason
was that it was anchored in positivism, hence data analysis methods had to be quantitative
(Saunders et al., 2019). A normality test using Skewness and Kurtosis was applied for testing
the distribution of the data set (Regoniel, 2020).
The SPSS was relevant in this study as it allowed for the calculation of linear regression
which was then used to measure the relationship between the independent variable and the
dependent variable. The researcher also employed the use of the 5-point Likert scale with a
ratio scale from strongly agree to strongly disagree to code the statistical data in a short,
faster, and understandable procedure (Saunders et al., 2019). The data was predominantly
presented using statistical tables, before the discussion of findings to make the data more
understandable.
3.12 Reliability and Validity of the Research Instruments
The reliability and validity tests of scales are regarded as essential factors in ensuring the
credibility of a research study. According to the study by Surucu and Maslakci (2020) both
the validity and reliability tests are to be understood by the researcher to ensure that the scales
used in the methodology of the study are tested and proven credible.
3.12.1 Reliability
According to Surucu and Maslakci (2020) reliability in research refers to the extent to which
a measurement instrument can be dependent upon producing consistent results if repeat
measurements are conducted on a study at different times. In quantitative studies, there are
three notable ways of assessing the reliability of a study, these include Test-retest reliability,
Alternative forms, and internal consistency test. This study was anchored on internal
consistency tests which can be applied in three measurements, namely split-half, item-total
correlation, and alpha reliability coefficient (Surucu and Maslakci, 2020; Regoniel, 2020).
However, the researcher adopted the Alpha reliability coefficient test and item-total
correlation as the literature reviewed regards the tests to be more relevant. Thus, Cronbach’s
alpha coefficient reliability test was then selected over Kuder-Richardson 20 & 21 alpha
coefficient test as studies by Surucu and Maslakci (2020) suggest that Cronbach’s alpha test
is the most preferred and widely used method of measuring internal consistency. Cronbach's
alpha is a technique that measures the average of all possible split-half coefficients, resulting
in different ways of splitting the scale items with values ranging from 0 to 1 (Surucu and
Maslakci, 2020). Hence, to ensure the reliability of the study results, the researcher employed
the use of the Cronbach’s Alpha test to measure the reliability of the study results by
quantifying the amount of random measurement error that exists in a sum score or average
(Hayes and Coutts, 2020; Regoniel, 2020).
3.12.2 Validity
According to Hayes and Coutts (2020) validity refers to the degree of significance of
something in giving desired results. Akamatsu et al., (2016) suggest that validity in research
refers to whether a study’s findings are true, and definite and whether the evidence they yield
supports the research findings. To accurately measure the validity of the study results, the
researcher was exposed to the use of construct, predictive, concurrent, and content validity as
measures of validity.
However, the researcher adopted predictive validity to measure how the independent variable
(digitalisation of tax system) has a causal relationship to the dependent variable (revenue
collection) underpinning the study (Surucu and Maslakci, 2020). The researcher also adopted
content validity to validate the literature reviewed, research methodology, and data analysis
and presentation tools used following authenticated past scholars to this study.
Predictive validity was applied in this study as the researcher wanted to predict the extent to
which digitalization of tax system (independent variable) can influence revenue collection
(dependent variable) of NBS brand based on the regression analysis (Surucu and Maslakci,
2020).
Saunders et al., (2019), Pallant, (2016), and Creswell, (2018) famously known for their
academic contribution in research methodology, were the source of guidance for the
researcher. Krejcie and Morgan, (1970) and Saunders et al, (2019) applied for sample size
determination for the study. Bryman and Bell, (2016) were employed in a structured
questionnaire design, hence, this was sufficient to guarantee content validity. Furthermore, a
pilot study was conducted to ensure that the validity of the study results was accurately
measured, through pretesting the structured questionnaire with colleagues.
3.13 Ethical Considerations
This study was informed by ethics which govern how the research should be conducted.
Ethics speak of moral principles and values that influence the conduct or behaviour of an
individual or group (Creswell, 2018). The researcher, therefore, was guided by the following
ethical issues that were relevant for this study.
Voluntary Participation
The researcher also assured the respondents from the company that participated in the study
were voluntary, hence it was optional to participate in the study. This allowed for an ethical
study to be conducted in researching the influence of digitalization of tax system on revenue
collection.
Anonymity
The researcher assured the respondents that participated in the study were confidential, by
ensuring that the personal data of the respondents is processed fairly and lawfully and used
only to fulfil the study. The researcher also assured the respondents that any personal
responses from participants will not be ascribed to any individual, as all the data would be
computed in an aggregate and would not be linked to any respondent.
Furthermore, the researcher went on to develop and employ the use of a closed-ended
questionnaire that did not contain the names of the respondents, as observing the anonymity
of respondents was relevant throughout the study. Lastly, the researcher demonstrated
professionalism by working tirelessly to ensure that the above ethics were observed in their
entirety throughout the data collection.
3.15 Chapter Summary
It was explained in this chapter in detail, the research quantitative research method adopted,
and the causal-explanatory research design that was used to address the research hypotheses
generated for the study. Justification for the use of a quantitative research method embedded
in a positivist philosophy was articulated, and the target population, sampling methods, and
research instruments used for the study were also outlined and discussed. In addition, the data
analysis tools were outlined, and the procedure undertaken was discussed. Furthermore,
ethical considerations for the study were also discussed. The next chapter, Chapter four, seeks
to report the findings from quantitative data, and discuss the themes that emerged during the
analysis of the research findings from the quantitative research method based on structured
questionnaires with closed-ended question and interviews.