research methodology

CHAPTER THREE

METHODOLOGY

3.1       Research Design

The study adopted a correlation research design. Correlational research is a non-experimental research design technique that helps researchers establish a relationship between two closely connected variables. Correlational studies display the relationships among variables by such techniques as cross-tabulation and correlations (Marilyn & Jim, 2011). The research also used both quantitative and qualitative data, where this research method dealt with quantifying and analysis variables in order to get results.  It  involved  the  utilization  and  analysis  of  numerical  data  using  specific  statistical techniques to  answer  questions like  who, how much, what, where, when,  how many,  and how. The quantitative approach was employed in order to capture statistical evidence of motivation style levels and tax compliance levels as well as the relationships to obtain the correlation and regressions.

Qualitative data on the other hand was used in this study because the topical issues of compliance require social reality that should be obtained in the real life among the tax payers.

3.2 Study Population

Lira Municipality is located in Lira District in Northern Uganda. It is geographically located at latitude 20’ 17′ north of the equator and longitude 32’ 56′ east of the principal meridian. It became a municipal council in 1985. The Council has four (4) Divisions, twenty two wards (22) and sixty four cells (64).

The target populations was the SMES operating within Lira municipality and are legally registered by URA as taxpayers on the category of SMEs. Lira municipality has a total of 1,643 SMEs.(URA, 2019). For the purpose of the study, the population were proportionately spread across the four divisions. Therefore the study population was  328 SMEs in the central division, where our focus was.

3.3       Sample Size

The sample size for the quantitative data was determined using Yamane (1967) formula where;

s = X ²NP(1− P) ÷ d²(N −1) + X ²P(1− P).

s = required sample size.

X² = the table value of chi-square for 1 degree of freedom at the desired confidence level (3.841).

N = the population size.

P = the population proportion (assumed to be .50 since this would provide the maximum sample size).

d = the degree of accuracy expressed as a proportion (.05).

n= 176. Therefore, the sample size was 176 SME operators in Lira municipality

3.4       Sampling Procedure

The sample was obtained using simple random sampling. Simple random sampling was used until the desired sample is obtained. Simple random sampling (SRS) occurs when every sample of size n (from a population of size N) has an equal chance of being selected (Fricker, 2005).The choice of simple random sampling was dictated by the desire to minimize possible bias and ensure that all SMEs and their owners or managers have equal chances of being selected to participate in the study.

3.5       Sources of data

The study used primary data. An advantage of using primary data is that researchers are collecting information for the specific purposes of their study. In essence, the questions the researchers asked were tailored to elicit the data that would help them with their study. Researcher collect the data himself, using surveys, interviews and direct observations (Sajjad, 2018). The primary data was collected using questionnaires from the SMEs owners or managers in Lira Municipality. The URA staff were given interview questions to respond against.

3.6  Data Collection Methods and Instruments

Data collection is the process of gathering and measuring information on variables of interest, in an established systematic fashion that enables one to answer stated research questions, test  hypotheses, and evaluate outcomes (Kabir, 2016). A self-administered questionnaire was used to collect primary data from the SMEs operators. It contained closed ended questions. A self-administered questionnaire (SAQ) refers to a questionnaire that has been designed specifically to be completed by a respondent without intervention of the researchers (e.g. an interviewer) collecting the data (Lavrakas, 2008). Structured questionnaires was used because they are easy to administer, cost effective and appropriate for collecting quantitative data in a short time period.  A five point likert scale was used to determine the level of agreement with the questions in the Questionnaire relating the variables described above where 1= strongly Disagree, 2 =Disagree, 3=uncertain, 4= agree and 5= strongly agree.

An unstructured questionnaire was also used to collects qualitative data.

Table 3.7 represents the variables under study and their components of measure to be assessed in order to achieve the objectives of the study.

3.7       Measurement of Variables

Table 1: Measurement of Variables

Source: Author, adopted from Literature review

3.8  Validity and Reliability

3.8.1 Validity

For quality control, a pre-test of the research instruments to establish their validity was done. The instrument was given to the researcher’s two supervisors to give their opinions on the relevance of the questions using a 5-point Likert scale of strongly disagree, disagree, not sure, agree and strongly agree. The data was then considered valid for the study .

Table 2: Validity statistics

Source: Primary data (2020)

This shows that 90% of the questions were accepted as valid making the tool acceptable for data collection.

3.8.2    Reliability

The reliability of the research instrument was pretested by administering it to selected respondents and was examined for their reliability by using Cronbach’s Alpha value. Although the standards for what makes a “good” α coefficient are entirely arbitrary and depend on your theoretical knowledge of the scale in question, many methodologists recommend a minimum α coefficient between 0.65 and 0.8 (or higher in many cases); α coefficients that are less than 0.5 are usually unacceptable, especially for scales purporting to be unidimensional (Goforth, 2018). The study thus considered 0.65 or greater to be the perfect Alpha score of reliability, processed using SPSS software ver.23.   

Table 3: Reliability Statistics

Source: Primary data (2020)

The detailed analysis showing the reliability coefficients of the consolidated subsections on the questionnaire tool were as shown in Table 3. It can be concluded that the reliability of the tool used was excellent and acceptable going by the scale given by George and Mallery (2003) that an alpha score of  .9 and above is considered to be excellent; that ≥ .8 as good;  ≥ .7 as acceptable;  ≥ .6 as fair; ≥ .5 as poor; and  ≥ .5 as unacceptable. Therefore, basing on the above Cronbach’s Alpha value ( 0.928) , the reliability of the instrument  used  was excellent.

3.9  Data Processing and Analysis

Analysis is the application of reasoning to understand and interpret the data that have been collected (Kothari, 2004). By definition, qualitative data analysis is the range of processes and procedures whereby one moves from the qualitative data that have been collected into some form of explanation, understanding or interpretation of the people and situations being investigated. Qualitative data analysis is usually based on an interpretative philosophy.

Analyzing qualitative data is essentially a complex process which consists of noticing, collecting and thinking; and the purpose of this model is to show that there is a simple foundation to the complex and rigorous practice of qualitative data analysis. This process is interactive and progressive. In this study, the researcher used the judgmental practice which was a suitable method of analyzing qualitative data and the ethnographic representation of tax compliance realities.

The primary data collected was edited, coded and analyzed to identify the relationship between the tax administration systems and tax compliance. Data derived from the questionnaires was analyzed using SPSS ver 23 statistical package. Descriptive statistics and inferential was produced in form of tables.

Since the study ran a correlational analysis. Inferential statistics including Pearson correlation and regression was processed. Correlation helped to establish the relationships between the study variables while multiple regression was used to establish the effect of tax administration systems on tax compliance.  Also, sample characteristics were processed to generate frequencies and percentages. 

3.10     Ethical Considerations

The study was conducted after attaining an introductory letter from Kyambogo University to be presented to the respondents. The researcher also ensured that participation was voluntary without any forceful tendencies. The privacy of respondents was sheltered by only using the study for academic purposes and not asking for personal details. The researcher also acknowledged all the sources used. The entire report considered facts as found and did not subject them to bias and prejudgments.