Research Method: Data Collection

There are two main methods or approaches to collecting data: quantitative and qualitative. The quantitative method mostly relates to data collection through questionnaires, which use numerical data, while the qualitative approach refers to data collection that use non- numerical data, such as interviewing (Saunders et al., 2009: 151). The qualitative method often combines observation, analysis of text and documents, interviews, recording and transcribing (Silverman, 2001: 11). Each method has its own strengths and weaknesses which must be considered as they will affect the research findings. Distributing questionnaires personally helps to build connections with respondents, and quick results can be obtained. However, it can be expensive when respondents are distributed across a large region (Sekaran and Bougie, 2009: 217). On the other hand, interviews can supply rich data while building close connections with respondents; if there is any doubt about the meaning of a question, it can be directly explained, thus enhancing the understanding of multifaceted issues. However, conducting face-to-face interviews with large numbers of respondents might lead to bias and would be expensive (Sekaran and Bougie, 2009: 217).

In order to decide which data collection method to use, the researcher should mainly consider the research methodology framework that is being employed (Sarantakos, 1994: 152) and the method’s appropriateness to the research questions established at the beginning of the research (Bryman, 2008: 624)

Nowadays, due to the complexity of social phenomena, it is common to combine both approaches, resulting in ‘triangulation’. Triangulation is a mixed method using two or more data collection and analysis methods to support the research outcome (Saunders et

al., 2009: 154). Similarly, Tashakkori and Teddlie (2009: 285) underline the word ‘integration’ and define it as “research in which the investigator collects and analyzes data, integrates the findings and draws inferences using both qualitative approaches or methods in a single study”. Thus, the research might be categorised as mixed-method if it satisfies one or more conditions as follows: two types of research questions, two types of data collection, two types of data analysis (Tashakkori and Teddlie, 2009: 286). Generally, the goal of using this mixed-method approach is to achieve complementarity, completeness, development, expansion, confirmation, compensation and diversity.

Triangulation offers certain advantages: it provides more validity to the conclusion when the outcome provides mutual confirmation; it usually (but not always) offers generally consistent data; it tends to provide mutually reinforcing results (Bryman, 1988: 131-133). Also, the quality of data collected using this method might be more credible since the weakness inherent in using a single method might be reduced, although it requires more time and financial effort (Sekaran and Bougie, 2009: 216-217). However, it should be noted that this method is not fundamentally of better quality compared to using a mono- method approach (Bryman, 2008: 624). Hence, Bryman (2008: 264) suggests some important points to be considered when using this mixed-method approach: it must be competently designed and conducted; it must be appropriate to the research question; it must consider the need to spread limited research resources.

Due to the nature of the research questions in this research, triangulation is adopted, as they require both quantitative and qualitative data and knowing. In doing so, this research particularly applies inter-method triangulation in the sense of using a quantitative method through questionnaires and a qualitative method through semi-structured interviews. The use of triangulation is rationalised on the basis of the variability of the research questions, which requires the triangulation method to be applied in order to obtain the most reliable and valid results.

5.5.5.1. Quantitative data collection method: Questionnaires

In a survey study, the most common data collection method is the questionnaire (Saunders, et al., 2007: 355), which has become a common and often-used data collection

method. To capture the main primary data required to answer the research questions and considering the advantage they offer in terms of efficiency in time, energy and cost (Sekaran and Bougie, 2009: 185), questionnaires are considered an efficient method.

Sekaran and Bougie (2009: 197) define a questionnaire as “a preformulated written set of questions to which respondents record their answers, usually within rather closely defined alternatives”. They can be personally administered or self-administered; postal or online questionnaires have also become popular.

For this research, the questionnaire is considered an efficient method of reaching out to microenterprises that have borrowed money from BMTs and BPRS, as their population size is quite large. In addition, considering the detailed nature of the questions posed, the questionnaire is considered more appropriate.

(i) Questionnaire design

The questions in the questionnaire were designed as a mixture of closed-ended and open- ended questions. In addition, four types of scales were used for close-ended questions: the dichotomous scale; category scale; Likert scale; and ratio scale (see Appendix A). The dichotomous scale is used for questions with Yes or No answers (Sekaran and Bougie, 2009: 149), e.g. question number 1 on whether or not the respondent borrowed money from BMT or BPRS, etc.

Category scale is applied for multiple items to obtain a single response (Sekaran and Bougie, 2009: 149), such as question number 3 about other sources of income, or number 4 about marital status, etc.

The Likert scale is used to capture the strength of the respondent’s agreement with a statement (Sekaran and Bougie, 2009: 152). This study applied a variety of Likert scales due to the varying nature of the questions; this scale was applied for the following:

(i) Question number 18 about business problems: never, rarely, sometimes, often, and always;

(ii) Question number 25 about the level of agreement in relation to factors in the decision to borrow money from BMTs or BPRS: strongly disagree, disagree, neutral, agree, and strongly agree;

(iii) Question number 30 about the level of understanding of terms used in Islamic finance: poor, fair, good, very good, and excellent;

(iv) Question number 42 about the impact of the financing on assets, number 45 about the impact on business expenditures per month, and number 46 about the impact on household expenditures: major negative effect, minor negative effect, no effect, minor positive effect, and major positive effect;

(v) Question number 43 about the effect of the loan on economic activities, and number 50 about the social impact: negative effect, no effect, minor positive effect, moderate positive effect, and major positive effect.

The ratio scale is used for questions that can be answered in terms of number (Sekaran and Bougie, 2009: 145), such as question number 6 about age, number 17 about the value of assets, number 41 about sales/month, etc. Hence, the respondents’ precise numerical answers are more accurate and are likely to provide more flexibility when grouping the data analysis.

The open questions are provided to capture other answers that are not covered in the provided options (e.g. question number 3 about other sources of income), or to specify the answers (e.g. question number 32 about the original intention of the loan - if it is for consumption the respondents are required to specify what type of consumption), or when the options are not provided (e.g. question number 20 to specify the name of the institution from which the respondent obtained financing).

As for the structure of the questionnaire, it is designed in four sections with 50 questions in total, including closed-ended questions and short-answer, open-ended questions especially when values are required as answers.

Section A of the questionnaire captures the information regarding the microenterprises’ profiles, comprising 18 questions. The first question is a screening question asking whether the respondent borrowed money from BMT or BPRS for productive purposes in the enterprise. Other questions cover demographic information, followed by more specific information about the household size, area of business, age of business, business field, business location, assets, and problems experienced in running the business, ranging from never to always. The questions are adopted and modified from Hulme et al.

(1996), Ahmed (2002), Ascarya and Sanrego (2007) and Adnan et al. (2003).

Section B covers the twenty-two questions specific to financing aspects, which include access to financing, factors in decision-making when choosing BMT or BPRS, times and duration of the financing, level of understanding of Islamic terms used in Islamic finance, the purpose of the loan, collateral, instalments, arrears etc. These questions are adopted and modified based on the previous research by Hulme et al. (1996) and Ascarya and Sanrego (2007).

Section C aims to capture the required information about the economic impact of borrowing, by comparing the situations before financing and after financing. It therefore includes average monthly sales and net income, impact on assets, impact on economic activities, impact on monthly business and household expenditures, and impact on employment. There are eight questions in this section which are adopted and modified from Ahmed (2002), Hulme et al. (1996), Rahman (2010), Ascarya and Sanrego (2007) and Karlan and Goldberg (2007)

Section D is specifically about the social impact of borrowing, including social services obtained from BMT and BPRS, social impact (relationships with spouse/family members and other group members), religious impact (knowledge and practical) and other impacts (children’s education, nutrition, housing stock, family empowerment and women’s decision-making). The questions in this section were adopted and modified from Ahmed (2002) and Karlan and Goldberg (2007).

(ii) Administration of the questionnaires

The questionnaires were personally administered to MEs who are customers of BMTs and BPRS in East Java, Indonesia. This method of administering was employed because the characteristics of the respondents as well as the character of the questions posed are very specific and detailed, such as personal financial information of the respondents (Saunders et al., 2007: 357). Hence, this approach is considered helpful for building connections with the respondents, for providing clarification when there is difficulty in answering the questions, and to ensure that the questionnaires are completed; they can then be collected immediately (Sekaran and Bougie, 2009: 217). It should also be noted that personally administering the questionnaire proved to be expensive due to the large numbers of respondents being located in many cities/towns in East Java, Indonesia.

(iii) Pre-tests

Pre-tests of the questionnaire were conducted in a number of stages. The first stage was conducted through distribution of written questionnaires to some PhD students who were undertaking research into Islamic microfinance. Based on the suggestions collected during the first pre-test, the questionnaire was developed for the second pre-test. The second stage was conducted using online questionnaires to obtain more responses from students and non-students, both in Durham and in Indonesia. The third pre-test was distributed to actual participants in Indonesia to obtain responses from microenterprises, the targeted respondents. This provided an opportunity to develop the questionnaire and also gave initial validity to the process.

5.5.1.2. Qualitative data collection method: Semi-structured interviews

The literature generally indicates that there are two types of interview: unstructured interview and structured interview (Sekaran and Bougie, 2009). Unstructured interviews aim to capture broad preliminary issues that emerge; hence, no specific questions are set in advance beyond the issues to be discussed (Sekaran and Bougie, 2009: 186). On the other hand, when the researcher knows what information is required and accordingly prepares predetermined questions, it becomes a structured interview (Sekaran and Bougie,

2009: 188). It should be noted that researchers often implement a method in between unstructured and structured interviews, namely the semi-structured interview, in which “the content and sequence aren’t fully specified in advance” (Jankowicz, 2000: 237). The advantage of semi-structured interviews is the flexibility of the order of the questions, which helps to direct all the questions to the interviewees, allowing opportunities to raise new questions from the answers provided by the interviewees, with the objective of exploring further (Jankowicz, 2000: 243). This is commonly used in case-study and survey designs; it is also suitable for researchers who are not sure what kind of responses they may receive from the interviewees (Jankowicz, 2000: 246).

Interviews can be used to substantiate the data gathered through questionnaires, but they can also be used as a single method of data collection. In this research, in order to expand the available data for analysis, face-to-face semi-structured interviews are used to gather qualitative data from Directors/Managers of BMTs and BPRS with the objective of validating the data from the questionnaires and also directly engaging with another set of interviewees in relation to their perception of the running of their organisations (see Appendix B). Face-to-face one-to-one interviews were chosen to explore and understand the issues, although this method may be expensive when it involves many interviewees spread over a large area (Sekaran and Bougie, 2009: 217). It should be noted that the interviewees and the respondents of the questionnaires in this study are from the same institutions.

Interview questions’ design and process

The interview questions used in this study are designed as a mixture of open-ended and closed-ended questions (see: Appendix B). While the open-ended questions offer flexibility to the interviewees in responding, the closed-ended questions are those which the interviewees will probably be able to answer easily and quickly, although there is always an opportunity to provide comments. Among others, the main questions in the interview schedule used in this study relate to the research questions, including the potential and prospect of BMTs, BPRS and MEs, and the role of BMTs and BPRS in developing MEs, all of which use open-ended questions, and the financing and non-

financing services provided to MEs, which use a mixture of open-ended and closed-ended questions, such as question number 13 on the collection of repayment, etc. The detailed interview questions are presented in the Appendix B.

In regard to the interview process during the field research between mid-August and mid- November 2011, the interviews with 22 Directors/Managers of BMTs and BPRS took place in the 12 BMTs’ and 10 BPRS’ offices in various cities/towns in East Java; each interview lasted between 1.5 hours and 2 hours. The interviewees sometimes suggested that the researcher meet the BMTs’ and BPRS’ staff members, who could provide more detailed information, such as marketing staff, administration staff, etc.