Basic vs. Applied Research

Basic research (also called pure or fundamental research) aims to develop or expand knowledge and theory for its own sake, without any immediate practical use in view. The researcher asks "what is true?" rather than "what will this solve?" Its findings usually feed the general body of knowledge and may only be applied years later.

Applied research aims to solve a specific, practical, real-world problem. It uses existing theories and knowledge to answer a concrete question for a client, organisation or society.

Examples

• Basic: Studying how memory encoding works in the human brain; deriving the theory of relativity; investigating the general causes of inflation in a closed economy.
• Applied: Testing whether a new fertiliser raises paddy yield in Chitwan; Nepal Rastra Bank studying why remittance inflows are not translating into credit growth; a bank measuring customer satisfaction with mobile banking.

The two are not watertight compartments — basic research often supplies the theory that applied research later uses (e.g. learning theory → classroom teaching method).

Quantitative vs. Qualitative Research

In practice the two are often mixed (mixed-methods design) — e.g. a survey first, then interviews to explain the numbers.

Role of Research in Management Decision Making

Management is essentially a series of decisions under uncertainty, and research reduces that uncertainty by supplying evidence instead of guesswork.

1. Problem identification — Research defines the real problem. A bank may think its problem is low deposits, while research shows it is poor branch service quality.
2. Information for planning — Market research, feasibility studies and demand forecasts give the factual base for objectives, targets and budgets.
3. Evaluation of alternatives — Research compares options (new branch vs. digital channel) on cost, risk and return, so the best alternative is chosen rationally.
4. Reducing risk and uncertainty — Pilot studies and test marketing allow decisions to be tried on a small scale before full commitment.
5. Organising and staffing — Job analysis, workload studies and employee surveys guide structure, staffing levels and training needs.
6. Motivation and control — Attitude surveys, performance data and feedback research show whether plans are on track and where corrective action is needed.
7. Monitoring the environment — Continuous research tracks competitors, customers, technology and policy changes, keeping strategy current.
8. Accountability and justification — Documented research findings justify decisions to the board, shareholders and regulators.

Conclusion: Research converts intuition into informed judgement. It does not replace managerial experience, but it supplies the reliable data on which sound, timely and defensible decisions rest — which is why an institution like Nepal Rastra Bank maintains a full research department for policy and management decisions.

Literature Review: Phases

A literature review is a systematic survey of existing knowledge on a topic. It generally passes through five phases:

1. Identification of the topic and scope — The researcher defines the research question and decides the boundaries (time period, geography, sector) of the literature to be covered.
2. Searching the literature — Relevant sources (journals, books, reports, theses, NRB publications, government statistics) are located using databases, libraries and reference lists.
3. Screening and evaluating sources — Material is skimmed and judged for relevance, credibility, recency and methodological soundness; irrelevant or weak sources are discarded.
4. Analysis and synthesis — The retained literature is read closely, findings are compared and grouped into themes, agreements and contradictions are identified, and gaps in existing knowledge are noted.
5. Writing the review — The synthesis is presented in a structured, critical narrative that establishes the theoretical base and justifies the present study.

Linkage among Problem Statement, Objectives and Hypothesis

These three elements form a logical chain in a research project:

• Problem statement identifies the gap or difficulty — what is not known or not working — and states why it is worth studying. It sets the boundary of the inquiry.
• Research objectives convert the problem into specific, measurable statements of what the study intends to achieve. Each objective flows directly from the problem statement, so that achieving the objectives amounts to solving the stated problem.
• Research hypothesis is a tentative, testable proposition about the relationship among variables. It is derived from the objectives and prior literature, and it guides the choice of design, data and statistical tools. Testing it either supports or rejects the proposed explanation, which in turn answers the problem statement.

Thus the sequence is: problem → objectives → hypothesis → design and data → analysis → conclusion, with each step justified by the preceding one. A weak problem statement produces vague objectives and untestable hypotheses; a well-stated problem yields focused objectives and testable hypotheses.

Research Designs: Descriptive, Exploratory and Analytical

Descriptive design describes the characteristics of a population, situation or phenomenon — for example, the pattern of remittance inflows into Nepal. It answers what, who, where, when and how much, using surveys, observation and secondary data. It does not explain causation.

• Merits: simple, inexpensive, useful for policy description, findings are generalisable if sampling is sound.
• Demerits: cannot establish cause and effect, risk of bias in self-reported data, may merely restate known facts.

Exploratory design is used when the problem is new or little understood. It explores ideas, formulates hypotheses and clarifies concepts through literature study, expert interviews, case studies and pilot surveys.

• Merits: flexible, open to new insights, useful for building hypotheses and refining research questions.
• Demerits: conclusions are tentative and not generalisable, highly dependent on researcher judgement, may yield inconclusive results.

Analytical (explanatory/causal) design goes beyond description to test hypotheses about relationships and causes — for example, the effect of interest rate on inflation. It uses experiments, causal-comparative or correlational designs and statistical techniques such as regression.

• Merits: identifies cause and effect, supports prediction and theory testing, provides a strong basis for policy decisions.
• Demerits: costly and time-consuming, requires rigorous control of variables, difficult to apply in social settings where many factors cannot be controlled.

In practice, the three designs are often combined: exploratory work precedes descriptive measurement, which in turn supplies the data for analytical testing.

Classification of Primary and Secondary Data

Data are the raw facts and figures collected for a research purpose. On the basis of their source and the stage of use, data are classified into two broad types:

1. Primary Data

Primary data are those which are collected for the first time by the researcher himself for the specific study at hand. They are original in character, fresh, and have not undergone any previous statistical treatment. Because the researcher controls the collection, primary data are usually more accurate, reliable and directly relevant to the objective — but they are costly, time-consuming and require trained investigators.

Sources / methods of collection:

• Direct personal interview
• Telephone interview
• Mailed questionnaire
• Observation
• Experimentation (laboratory or field)

Examples:

• A Nepal Rastra Bank team conducting a household survey in Kathmandu to measure inflation expectations for the current monetary policy review.
• A student distributing questionnaires to 200 depositors of a commercial bank to study e-banking satisfaction.
• A researcher recording daily prices of vegetables in the Kalimati market for three months.

2. Secondary Data

Secondary data are those which have already been collected and compiled by some other agency and are simply reused by the researcher for his own purpose. They are second-hand in nature — the researcher is not the original collector. They are cheap, quick and easily available, but they may be outdated, incomplete, or collected for a purpose different from the present study, so they must be carefully scrutinised for reliability, suitability and adequacy before use.

Sources:

• Published: NRB publications (Economic Review, Monetary Policy, Annual Report), Nepal Rastra Bank Act, CBS Statistical Yearbook of Nepal, Economic Survey (MoF), National Accounts Statistics, IMF/World Bank reports, journals, newspapers.
• Unpublished: records of banks and financial institutions, company annual reports, internal office files, previous research theses.

Examples:

• Using the Economic Survey figures of GDP growth for the last five years to analyse the trend of economic growth.
• Using NRB's Balance of Payments statistics to study the trend of remittances.
• Using the census report of 2021 to obtain the population of a district.

Key difference: primary data are original and study-specific; secondary data are already available and reused. In practice, most research uses secondary data first to build the background, and primary data to fill the specific gap.

---

Questionnaire: Content, Structure and Administration Process

A questionnaire is a formal set of questions prepared in a logical order and used to collect primary data from respondents, especially when the number of respondents is large and widely scattered. It is the most common tool of survey research.

(a) Content of a Questionnaire

The content means what is asked. It should be decided strictly on the basis of the research objectives. A good questionnaire generally contains:

1. Identifying information — name (if needed), address, age, sex, occupation, income group, education, etc., to classify respondents.
2. Subject-matter questions — the core questions on the topic under study, e.g., saving habits, loan use, satisfaction with banking services, awareness of monetary policy.
3. Cross-checking questions — repeated or differently worded questions to test the consistency of replies.
4. Open-ended questions — no fixed alternatives; the respondent answers in his own words. Useful for opinions and suggestions.
5. Closed / pre-coded questions — fixed alternatives such as yes–no, multiple choice, ranking or rating scales. Easy to tabulate and analyse.
6. Instructions — brief directions to the respondent or enumerator on how to fill the form.

The questions must be simple, short, unambiguous, free from personal bias, and should not ask for unnecessary personal or confidential information.

(b) Structure of a Questionnaire

Structure means how the questions are arranged. A well-structured questionnaire follows a psychological order:

1. Introduction / covering letter — purpose of the study, assurance of secrecy, name of the sponsoring body, and request for cooperation.
2. Warm-up (easy) questions — simple, impersonal questions at the beginning to create interest and confidence.
3. Basic information — general particulars of the respondent.
4. Core questions — the main subject-matter questions, placed in the middle.
5. Sensitive / personal questions — income, age, caste, political views, etc., placed near the end when rapport is established.
6. Open-ended and suggestion questions — at the end, allowing free expression.
7. Closing / thanks — expression of gratitude and space for the enumerator's remarks (date, time, signature).

(c) Administration Process

Administration means how the questionnaire is delivered, filled and returned. The main steps are:

1. Designing the draft — prepare questions in line with objectives, keeping language simple and unambiguous.
2. Pre-testing (pilot survey) — try the draft on a small representative group to detect defects in wording, order and length.
3. Revision and finalisation — remove or modify faulty questions on the basis of the pre-test.
4. Reproduction — print or digitise the final questionnaire in sufficient copies, with clear numbering and spacing.
5. Selection and training of enumerators — select trained investigators and brief them on the objective, method of asking and recording.
6. Distribution — the questionnaire may be:
   • Mailed / e-mailed to respondents (cheap but low response rate),
   • Personally delivered by the enumerator (higher response, costlier),
   • Administered collectively to a group of respondents at one place.
7. Follow-up / reminder — send reminders or revisit non-respondents to reduce non-response bias.
8. Collection and scrutiny — collect the filled forms, check for completeness, consistency and editing of errors.
9. Tabulation and analysis — code the responses and process them for statistical analysis.

Precautions in administration: keep the questionnaire short, assure confidentiality, avoid leading questions, use a polite tone, and ensure a high response rate for the findings to be representative.

Parametric vs Non-parametric Tests

Parametric tests are statistical tests that make specific assumptions about the population from which the sample is drawn — mainly that the data follows a known probability distribution (usually the normal distribution), that the observations are independent, and that the variances are homogeneous. They work on interval or ratio scale data and are based on the estimation of population parameters such as the mean (μ) and standard deviation (σ). Examples: z-test, t-test, F-test, ANOVA, Pearson's correlation.

Non-parametric tests are distribution-free tests that do not require the data to follow any particular probability distribution. They make few or no assumptions about the population parameters and are generally used for nominal or ordinal scale data (ranks, categories, signs). Examples: Chi-square test, Mann–Whitney U test, Wilcoxon signed-rank test, Kruskal–Wallis test, Spearman's rank correlation, sign test, run test.

Key Differences

Views on the Statement

The statement that "limited inferential analysis techniques are used in many research papers, articles and research reports" is largely true, especially in the context of developing countries like Nepal. My views are as follows:

1. Dominance of descriptive analysis: A large number of research works stop at frequency tables, percentages, means and graphs. Inferential tools such as hypothesis testing, confidence intervals, regression and ANOVA are either absent or applied superficially.

2. Over-dependence on a few tests: Even where inference is used, it is usually confined to the t-test, Chi-square test and simple correlation. Advanced techniques — multiple regression, logistic regression, factor analysis, panel data models, structural equation modelling — are rarely seen.

3. Reasons for limited use:

   • Weak statistical training and lack of software skills (SPSS, STATA, R).
   • Poor quality and small size of data sets, which discourage advanced inference.
   • Time, resource and supervision constraints in academic institutions.
   • Tendency to copy conventional methods rather than select the technique appropriate to the research question.

4. Consequences: Findings remain descriptive and cannot be generalised to the population; relationships among variables remain untested; policy recommendations lose empirical strength.

5. However, the statement is not universally valid: In medical, agricultural, econometric and central-bank research, rigorous inferential methods are increasingly used. Growing access to software and open data is gradually improving the situation.

Conclusion: The criticism is fair for a large share of student theses and general reports, but not for all research. The remedy lies in strengthening statistical education, insisting on proper research methodology, and matching the choice of test — parametric or non-parametric — to the nature of the data and the objective of the study.

General Format of a Research Report

A research report is a formal, structured document that communicates the objectives, methods, findings and conclusions of a study. Although the exact format varies with the institution (academic thesis, NRB survey report, feasibility study), the general format contains the following parts:

1. Preliminary (Front) Matter

• Title page — title of the study, name and designation of the researcher, name of the institution, place and date of submission.
• Letter of transmittal / certificate — forwarding letter to the authority; in academic reports, a certificate of approval by the supervisor.
• Acknowledgement — thanks to persons and organisations who helped.
• Table of contents — chapter and section headings with page numbers.
• List of tables, figures and abbreviations — for quick reference.

2. Main Body

• Chapter I: Introduction — background of the study, statement of the problem, objectives, research questions/hypotheses, significance, scope and limitations, and definition of key terms.
• Chapter II: Literature Review — review of related theories, previous studies and reports, identifying the research gap.
• Chapter III: Research Methodology — research design, population and sample, sampling technique, sources of data (primary/secondary), instruments of data collection (questionnaire, interview, observation), data analysis tools (percentage, mean, correlation, regression, chi-square) and limitations of the method.
• Chapter IV: Data Presentation and Analysis — presentation of collected data in tables, graphs and charts, followed by analysis, interpretation and testing of hypotheses.
• Chapter V: Summary, Conclusion and Recommendations — summary of major findings, conclusion drawn from them, and practical recommendations or suggestions for further study.

3. Supplementary (Back) Matter

• Bibliography / References — list of books, articles and reports cited, in a standard style (APA, MLA, Chicago).
• Appendices — questionnaire, interview schedule, raw data, calculation sheets, maps.
• Index (if required).

Characteristics of a Good Research Report

1. Clarity — simple, unambiguous language; technical terms defined; ideas expressed so that even a non-specialist reader can follow.
2. Completeness — contains all essential parts from title page to appendices; no important information is left out.
3. Accuracy and objectivity — facts, figures and calculations are correct and free from personal bias; findings are based on evidence.
4. Logical organisation — chapters and sections arranged in a systematic sequence so that one part leads naturally to the next.
5. Conciseness — only relevant matter is included; repetition and unnecessary detail avoided, yet the report is not so brief as to be unclear.
6. Reliability and validity — data are collected and analysed by sound methods, so conclusions can be trusted and generalised within the stated scope.
7. Proper documentation — every borrowed idea or figure is acknowledged through footnotes and references; sources are traceable.
8. Attractive presentation — neat typing, proper headings, well-labelled tables and diagrams, uniform format and correct grammar and spelling.
9. Timeliness — prepared and submitted within the stipulated time, so that it remains useful for decision-making.
10. Usefulness — findings and recommendations are practical, action-oriented and relevant to the problem, policy or organisation concerned.

Conclusion: A research report is thus a systematic document with front matter, main body and back matter. Its value depends on clarity, accuracy, objectivity, logical arrangement, proper referencing and practical usefulness — qualities that make it a reliable basis for policy decisions in an institution like Nepal Rastra Bank.

वैज्ञानिक अनुसन्धान प्रक्रियाको महत्व: नियमनकारी परिप्रेक्ष्य (नेपालका बैंक तथा वित्तीय संस्था)

१. परिचय

वैज्ञानिक अनुसन्धान प्रक्रिया भन्नाले समस्या पहिचान, उद्देश्य निर्धारण, परिकल्पना, तथ्य सङ्कलन, विश्लेषण र निष्कर्षसम्म पुग्ने क्रमबद्ध, वस्तुनिष्ठ तथा प्रमाणमा आधारित विधि हो। नेपाल राष्ट्र बैंक (NRB) ले बैंक तथा वित्तीय संस्था ऐन, २०७३ र राष्ट्र बैंक ऐन, २०५८ अन्तर्गत बैंक तथा वित्तीय संस्थाको नियमन, सुपरिवेक्षण र वित्तीय स्थायित्व कायम राख्ने जिम्मेवारी वहन गर्दछ। यस्तो जिम्मेवारी प्रमाण र तथ्यमा आधारित हुनुपर्ने हुँदा वैज्ञानिक अनुसन्धान प्रक्रियाको महत्व अत्यधिक रहन्छ।

२. नियमनकारी परिप्रेक्ष्यमा अनुसन्धान प्रक्रियाको महत्व

(क) तथ्यमा आधारित नीति निर्माण (Evidence-Based Policy Making) मौद्रिक नीति, पुँजीकोष सम्बन्धी निर्देशन, तरलता अनुपात, कर्जा–निक्षेप अनुपात जस्ता नीतिगत निर्णय वैज्ञानिक तथ्याङ्कमा आधारित हुनुपर्दछ। कच्चा तथ्याङ्कको वैज्ञानिक विश्लेषणबाट मात्र त्रुटिरहित नीति बन्न सक्छ।

(ख) जोखिम पहिचान तथा वित्तीय स्थायित्व कर्जा जोखिम, तरलता जोखिम, ब्याजदर जोखिम र सञ्चालन जोखिमको वैज्ञानिक मूल्याङ्कनबाट नै प्रणालीगत जोखिम (systemic risk) पहिचान गरी समयमै सुधारात्मक कदम चाल्न सकिन्छ। यसले वित्तीय संकट रोक्न मद्दत गर्दछ।

(ग) प्रभावकारी सुपरिवेक्षण तथा निरीक्षण स्थलगत (on-site) तथा स्थलबाहिर (off-site) सुपरिवेक्षणमा सङ्कलित तथ्याङ्कको सांख्यिकीय विश्लेषणबाट कमजोर बैंक पहिचान गरी सुपरिवेक्षणीय कारबाही गर्न सकिन्छ।

(घ) वित्तीय अपराध तथा सम्भावित जोखिम नियन्त्रण सम्पत्ति शुद्धीकरण (AML), आतङ्कवादी कार्यमा वित्तीय लगानी (CFT) तथा ठगी प्रवृत्तिको ढाँचा (pattern) पहिचान गर्न तथ्याङ्क विश्लेषण अनिवार्य हुन्छ।

(ङ) नियमनको प्रभाव मूल्याङ्कन (Regulatory Impact Assessment) लागू गरिएका नीतिको प्रभावकारिता मापन गरी आवश्यक संशोधन गर्न अनुसन्धानले आधार प्रदान गर्दछ।

(च) वित्तीय पहुँच तथा समावेशीकरण ग्रामीण तथा सीमान्तकृत समुदायको बैंकिङ पहुँच, वित्तीय साक्षरता तथा लघुवित्तको प्रभावबारे अनुसन्धानले समावेशी नीति तर्जुमा गर्न सघाउँछ।

(छ) पूर्वानुमान तथा सङ्कट व्यवस्थापन तथ्याङ्कीय मोडेल तथा प्रवृत्ति विश्लेषणबाट भविष्यमा आउन सक्ने जोखिमको पूर्वानुमान गरी सङ्कट व्यवस्थापन योजना तयार गर्न सकिन्छ।

(ज) पारदर्शिता तथा जवाफदेहिता वैज्ञानिक अनुसन्धानमा आधारित निर्णयले नियमनकारी निकायको विश्वसनीयता, पारदर्शिता र जवाफदेहिता बढाउँछ।

३. निष्कर्ष

नेपालका बैंक तथा वित्तीय संस्थाको नियमन तथा सुपरिवेक्षणमा वैज्ञानिक अनुसन्धान प्रक्रिया नीति निर्माण, जोखिम व्यवस्थापन, वित्तीय स्थायित्व र पारदर्शिताको आधारशिला हो। तसर्थ NRB ले तथ्याङ्कमा आधारित, वस्तुनिष्ठ र पूर्वानुमानमूलक अनुसन्धानलाई संस्थागत गर्नु आवश्यक छ।

अनुसन्धान प्रस्तावको आवश्यकता र साहित्य समीक्षाको भूमिका

१. अनुसन्धान प्रस्ताव (Research Proposal) को परिचय

कुनै पनि अनुसन्धान कार्य सुरु गर्नुअघि तयार गरिने लिखित योजनालाई अनुसन्धान प्रस्ताव भनिन्छ। यसले अनुसन्धानको विषय, उद्देश्य, समस्या, विधि, समयतालिका र बजेट समेट्दछ। यो अनुसन्धानको "नक्सा" (blueprint) हो।

२. अनुसन्धान प्रस्ताव किन आवश्यक छ ?

क) अनुसन्धानको दिशा निर्धारण: प्रस्तावले अनुसन्धानको विषय, उद्देश्य र सीमा स्पष्ट पारी अनुसन्धानकर्तालाई भद्रगोल हुनबाट बचाउँछ।

ख) समस्या र उद्देश्य स्पष्ट पार्न: अनुसन्धान समस्या, प्रश्न र उद्देश्य प्रस्तावमै स्पष्ट भएपछि कार्य सहज हुन्छ।

ग) स्रोत र बजेट अनुमान: आवश्यक जनशक्ति, समय, धन र सामग्रीको अनुमान गरी स्रोत व्यवस्थापन गर्न सकिन्छ।

घ) अनुसन्धान विधि छनोट: कुन विधि, नमुना, उपकरण र विश्लेषण प्रयोग गर्ने भन्ने पूर्वनिर्धारण हुन्छ।

ङ) स्वीकृति र अनुदान प्राप्ति: विश्वविद्यालय, अनुसन्धान परिषद् वा निकायबाट अनुमति र आर्थिक सहयोग पाउन प्रस्ताव अनिवार्य हुन्छ।

च) समय र लागत नियन्त्रण: समयतालिकाले कार्य समयमै सम्पन्न गर्न मद्दत गर्दछ।

छ) मूल्याङ्कनको आधार: प्रस्तावले अनुसन्धानको गुणस्तर र सम्भाव्यता मूल्याङ्कन गर्न आधार दिन्छ।

ज) दोहोरोपन रोक्न: पहिले भएका अनुसन्धानको पुनरावृत्ति हुनबाट जोगाउँछ।

३. साहित्य समीक्षाको भूमिका

साहित्य समीक्षा भन्नाले सम्बन्धित विषयमा पहिले भएका अध्ययन, शोध, पुस्तक र प्रतिवेदनको समीक्षात्मक अध्ययन हो। यसको भूमिका:

क) ज्ञानको आधार तयार: विद्यमान ज्ञानको सीमा पहिचान गरी नयाँ अनुसन्धानको आवश्यकता देखाउँछ।

ख) अनुसन्धान खाडल (Research Gap) पहिचान: कुन पक्ष अझै अध्ययन भएको छैन भन्ने पहिचान गराउँछ।

ग) सैद्धान्तिक आधार: सिद्धान्त, अवधारणा र मोडेलको आधार प्रदान गर्दछ।

घ) विधि तथा उपकरणको चयन: पूर्व अध्ययनका सफल विधि र उपकरणबाट मार्गदर्शन पाइन्छ।

ङ) दोहोरोपन रोक्न: पहिले गरिएको काम पुनः नगर्न सचेत गराउँछ।

च) तुलनात्मक विश्लेषण: आफ्नो नतिजालाई पूर्व नतिजासँग तुलना गर्न आधार बन्छ।

छ) शोध प्रश्न र परिकल्पना निर्माण: साहित्यका आधारमा शोध प्रश्न र परिकल्पना स्पष्ट हुन्छ।

ज) अध्ययनको औचित्य स्थापित: अनुसन्धान किन आवश्यक छ भन्ने तर्कपूर्ण आधार दिन्छ।

४. निष्कर्ष

अनुसन्धान प्रस्ताव अनुसन्धानको मार्गदर्शक नक्सा हो, जसले कार्यलाई व्यवस्थित, मितव्ययी र उद्देश्यमूलक बनाउँछ। साहित्य समीक्षा त्यसको आधारशिला हो — यसले विद्यमान ज्ञानको सर्वेक्षण गरी शोध खाडल पहिचान गराई अनुसन्धानलाई नयाँ र अर्थपूर्ण बनाउँछ। दुवैको समन्वयले अनुसन्धानको गुणस्तर र विश्वसनीयता सुनिश्चित गर्दछ।

अनुसन्धान अभिकल्पका भूमिका, तथ्याङ्क विश्लेषणको निर्भरता र तुलनात्मक अध्ययन डिजाइनहरूको विश्लेषण

१. अनुसन्धान अभिकल्पका मुख्य भूमिकाहरू

अनुसन्धान अभिकल्प (Research Design) भन्नाले अनुसन्धानको उद्देश्य प्राप्तिका लागि अपनाइने समग्र खाका, रणनीति र कार्ययोजनालाई जनाउँछ। यसका मुख्य भूमिकाहरू देखाइएका छन्:

क) अनुसन्धान समस्याको स्पष्ट परिभाषा र सीमाङ्कन अनुसन्धान समस्या के हो, कुन चरहरू समावेश हुन्छन् र कुन बाहिर रहन्छन् भन्ने स्पष्ट पार्ने काम अभिकल्पले गर्दछ। यसले अनुसन्धानको केन्द्रबिन्दु निर्धारण गर्छ।

ख) उद्देश्य र प्रश्नहरूको संरचना अनुसन्धानका उद्देश्य, अनुसन्धान प्रश्न तथा परिकल्पना (Hypothesis) लाई व्यवस्थित रूपमा प्रस्तुत गर्ने आधार अभिकल्पले तयार पार्छ।

ग) तथ्याङ्क सङ्कलन विधिको निर्धारण प्राथमिक वा द्वितीयक तथ्याङ्क, प्रश्नावली, अन्तर्वार्ता, अवलोकन, प्रयोगात्मक विधि आदिमध्ये कुन विधि उपयुक्त हुन्छ भन्ने निर्णय अभिकल्पले गर्दछ।

घ) नमुना छनोट र जनसङ्ख्याको निर्धारण कुन जनसङ्ख्याबाट, कति नमुना र कुन नमुना छनोट विधिबाट तथ्याङ्क सङ्कलन गर्ने भन्ने निर्णय अभिकल्पमा हुन्छ।

ङ) तथ्याङ्क विश्लेषण विधिको पूर्वनिर्धारण कुन तथ्याङ्कलाई कसरी विश्लेषण गर्ने, कुन सांख्यिकीय औजार प्रयोग गर्ने भन्ने कुरा अभिकल्पले पहिल्यै निर्धारण गर्छ।

च) समय, लागत र स्रोतको व्यवस्थापन अनुसन्धानको समयसीमा, बजेट र उपलब्ध स्रोतसाधनको उचित परिचालन अभिकल्पमार्फत सुनिश्चित हुन्छ।

छ) वैधता र विश्वसनीयताको सुनिश्चितता आन्तरिक तथा बाह्य वैधता (Internal and External Validity) कायम राख्ने आधार अभिकल्पले प्रदान गर्छ।

ज) निष्कर्ष र सामान्यीकरणको आधार सङ्कलित तथ्याङ्कबाट प्राप्त निष्कर्षलाई सिद्धान्त निर्माण वा नीति निर्धारणमा प्रयोग गर्न सकिने बनाउने काम अभिकल्पले गर्छ।

---

२. के तथ्याङ्कको विश्लेषण अनुसन्धान अभिकल्पमा निर्भर गर्दछ?

हो, तथ्याङ्कको विश्लेषण अनुसन्धान अभिकल्पमा प्रत्यक्ष रूपमा निर्भर गर्दछ। यसका कारणहरू:

क) तथ्याङ्कको प्रकृति अभिकल्पले निर्धारण गर्छ गुणात्मक (Qualitative) अभिकल्पमा विषयगत विश्लेषण, सामग्री विश्लेषण जस्ता विधि प्रयोग हुन्छन् भने मात्रात्मक (Quantitative) अभिकल्पमा सहसम्बन्ध, प्रतिगमन, t-test, ANOVA जस्ता सांख्यिकीय औजार प्रयोग हुन्छन्।

ख) स्केल र मापनको स्तर Nominal, Ordinal, Interval वा Ratio स्केलमा सङ्कलित तथ्याङ्कका लागि फरक–फरक सांख्यिकीय परीक्षण उपयुक्त हुन्छन्। यो स्केल अभिकल्पले नै तय गर्छ।

ग) अनुसन्धान प्रश्नको स्वरूप वर्णनात्मक (Descriptive) अभिकल्पमा केन्द्रीय प्रवृत्ति र विचलनको मापन पर्याप्त हुन्छ, तर कारणात्मक (Causal) अभिकल्पमा प्रयोगात्मक तुलना वा path analysis आवश्यक पर्छ।

घ) नमुना आकार र स्वतन्त्रता नमुना आकार र डिग्री अफ फ्रिडम अभिकल्पबाट निर्धारित हुन्छ, जसले कुन परीक्षण लागू हुन्छ भन्ने निर्णय गर्छ।

उदाहरण १: एक बैंकले "ग्राहक सन्तुष्टि" अध्ययन गर्दा वर्णनात्मक सर्वेक्षण अभिकल्प अपनाएमा Likert स्केलको औसत र मानक विचलन पर्याप्त हुन्छ। तर उही बैंकले "डिजिटल बैंकिङले निक्षेप वृद्धिमा पार्ने प्रभाव" अध्ययन गर्दा प्रतिगमन विश्लेषण आवश्यक पर्छ।

उदाहरण २: लोक सेवा तयारीको प्रभावकारिता अध्ययन गर्दा प्रयोगात्मक अभिकल्प (Control Group Design) अपनाइयो भने t-test उपयुक्त हुन्छ; तर त्यही विषयलाई गुणात्मक अभिकल्पबाट अध्ययन गर्दा अन्तर्वार्ता विश्लेषण आवश्यक पर्छ।

---

३. तुलनात्मक अध्ययन अभिकल्प (Case Control Study Design) र कोहर्ट अध्ययन अभिकल्प (Cohort Study Design) को भूमिका

क) तुलनात्मक अध्ययन अभिकल्प (Case Control Study Design)

यो अभिकल्पमा पहिले परिणाम (Outcome/Disease) पहिचान गरिन्छ, त्यसपछि पछाडि फर्केर (Retrospective) कारण वा जोखिम कारक (Risk Factor) खोजिन्छ।

विशेषताहरू:

• परिणामबाट कारणतर्फ अध्ययन (Backward direction)
• सस्तो, छिटो र सानो नमुनामा सम्भव
• दुर्लभ रोग वा घटनाको अध्ययनमा उपयुक्त
• Incidence rate निकाल्न सकिँदैन, Odds Ratio प्रयोग गरिन्छ

व्यवस्थापन तथा अर्थशास्त्रको अनुसन्धानमा भूमिका:

• कुनै कम्पनी घाटामा गएपछि (परिणाम) त्यसका कारणहरू — कमजोर व्यवस्थापन, बढी ऋण, बजार प्रतिस्पर्धा — पछाडि फर्केर खोज्न
• बैंकिङ सङ्कटमा परेका बैंकहरूको कारण विश्लेषण
• गरिबीमा परेका परिवारका कारण पहिचान

ख) कोहर्ट अध्ययन अभिकल्प (Cohort Study Design)

यो अभिकल्पमा एउटै विशेषता भएको समूह (Cohort) लाई समयको अन्तरालमा अध्ययन गरी परिणाम हेरिन्छ।

विशेषताहरू:

• कारणबाट परिणामतर्फ अध्ययन (Forward direction)
• लामो समय र बढी लागत आवश्यक
• Incidence rate र Relative Risk निकाल्न सकिन्छ
• सम्भाव्य (Prospective) वा पूर्वव्यापी (Retrospective) दुवै हुन सक्छ

व्यवस्थापन तथा अर्थशास्त्रको अनुसन्धानमा भूमिका:

• नयाँ कर्मचारी भर्ना गरेपछि तालिम दिएका र नदिएका समूहको उत्पादकत्व तुलना
• सूक्ष्म ऋण लिएका र नलिएका परिवारको आर्थिक स्थिति वर्षौंसम्म अध्ययन
• लगानी गरेका र नगरेका कम्पनीहरूको नाफा वृद्धि दर तुलना

ग) तुलनात्मक विश्लेषण (Comparative Analysis)

---

४. निष्कर्ष

अनुसन्धान अभिकल्प अनुसन्धानको मेरुदण्ड हो। यसले अनुसन्धान समस्या, उद्देश्य, नमुना, तथ्याङ्क सङ्कलन र विश्लेषण विधिलाई एकीकृत रूपमा निर्देशन गर्छ। तथ्याङ्कको विश्लेषण अभिकल्पमा निर्भर हुन्छ किनभने अभिकल्पले नै तथ्याङ्कको प्रकृति र उपयुक्त सांख्यिकीय औजार तय गर्छ। व्यवस्थापन तथा अर्थशास्त्रीय अनुसन्धानमा Case Control अभिकल्प छिटो र सस्तो भए पनि कमजोर प्रमाण दिन्छ, जबकि Cohort अभिकल्प महँगो र समयसापेक्ष भए पनि बलियो कारण–परिणाम सम्बन्ध स्थापित गर्न सक्षम हुन्छ। अनुसन्धानको उद्देश्य, उपलब्ध स्रोत र समयसीमा हेरी उपयुक्त अभिकल्पको छनोट गर्नु नै सफल अनुसन्धानको कुञ्जी हो।

अध्ययन प्रतिवेदन: नेपालमा सहकारी संस्थाको सान्दर्भिकता

(A Detailed Sample Design for Selecting Sample Cooperative Institutions in Nepal)

---

१. अध्ययनको पृष्ठभूमि तथा औचित्य (Background and Rationale)

नेपालमा सहकारी आन्दोलनको इतिहास करिब ७० वर्ष पुरानो छ। विसं २०१३ मा चितवनको बखुण्डेबेसीमा स्थापित "बखुण्डेबेसी कृषि सहकारी संस्था" लाई नेपालको पहिलो सहकारी संस्था मानिन्छ। संविधानको धारा ५१ (ख) (४) ले सहकारीलाई राष्ट्रिय विकासको तीन स्तम्भमध्ये एकको रूपमा स्थापित गरेको छ। सहकारी ऐन, २०७४ र सहकारी नियमावली, २०७५ ले सहकारीको कानुनी संरचना निर्धारण गरेको छ।

नेपाल राष्ट्र बैंकको तथ्याङ्क अनुसार हाल देशभर करिब ३४,०००+ सहकारी संस्था दर्ता भई करिब ६५ लाख सदस्य आबद्ध छन्। यी संस्थाको कुल कारोबार रकम खर्बौं रुपैयाँ पुगेको छ। यसै सन्दर्भमा सहकारी संस्थाको वर्तमान अवस्था, सान्दर्भिकता, कार्यक्षमता र समस्याहरूको वैज्ञानिक अध्ययन आवश्यक छ।

---

२. अध्ययनको उद्देश्य (Objectives of the Study)

२.१ मुख्य उद्देश्य (General Objective)

नेपालमा सहकारी संस्थाको वर्तमान अवस्था तथा सामाजिक–आर्थिक सान्दर्भिकताको मूल्याङ्कन गर्नु।

२.२ विशिष्ट उद्देश्यहरू (Specific Objectives)

1. सहकारी संस्थाको संस्थागत संरचना र सञ्चालन अवस्था पहिचान गर्नु।
2. सहकारीका सदस्यहरूको आर्थिक–सामाजिक उपलब्धि विश्लेषण गर्नु।
3. सहकारी संस्थाका समस्या तथा चुनौतीहरू पहिचान गर्नु।
4. सहकारीको दिगो सान्दर्भिकताका लागि सुझाव प्रस्तुत गर्नु।

---

३. अध्ययनको क्षेत्र तथा जनसंख्या (Study Area and Population)

३.१ भौगोलिक क्षेत्र (Geographical Area)

नेपालका सातै प्रदेशको प्रतिनिधित्व हुने गरी अध्ययन गरिनेछ।

३.२ अध्ययनको एकाइ (Unit of Study)

• प्राथमिक एकाइ (Primary Unit): सहकारी संस्था
• द्वितीयक एकाइ (Secondary Unit): सहकारीका सदस्यहरू

३.३ अध्ययन जनसंख्या (Study Population)

नेपालभर दर्ता भएका सम्पूर्ण सहकारी संस्थाहरू (करिब ३४,०००) अध्ययन जनसंख्या हुन्।

---

४. नमुना ढाँचा (Sample Design)

४.१ नमुना विधि (Sampling Method)

बहुस्तरीय नमुना विधि (Multi-stage Sampling) अवलम्बन गरिनेछ:

४.२ नमुना आकार निर्धारण (Sample Size Determination)

Cochran's Formula प्रयोग गरी नमुना आकार निकालिनेछ:

$n_0 = \frac{Z^2 \cdot p \cdot q}{e^2}$

जहाँ,

• Z = 1.96 (95% confidence level)
• p = 0.5 (estimated proportion)
• q = 1 − p = 0.5
• e = 0.05 (margin of error)

$n_0 = \frac{(1.96)^2 \times 0.5 \times 0.5}{(0.05)^2} = \frac{3.8416 \times 0.25}{0.0025} = 384.16 \approx 385$

जनसंख्या सीमित (N ≈ ३४,०००) भएकाले Finite Population Correction (FPC) लागू गरिनेछ:

$n = \frac{n_0}{1 + \frac{n_0 - 1}{N}} = \frac{385}{1 + \frac{384}{34000}} = \frac{385}{1 + 0.011294} = \frac{385}{1.011294} \approx 381$

गणनात्मक नमुना आकार: ३८१ सहकारी संस्था। यसलाई सातै प्रदेशमा समानुपातिक रूपमा बाँड्दा प्रदेशगत नमुना तलको तालिमा देखाइएको छ (जम्मा ३८१) र प्रत्येक संस्थाबाट ५ जना सदस्य छनोट गरिनेछ (जम्मा १,९०५ सदस्य)।

४.३ नमुना वितरण (Sample Distribution)

(नमुना आकारलाई प्रदेशको सहकारी घनत्व अनुसार समायोजन गरिनेछ।)

४.४ नमुना छनोटका मापदण्ड (Selection Criteria)

1. भौगोलिक सन्तुलन: हिमाल, पहाड, तराई–मधेश तीनै भू–क्षेत्रको प्रतिनिधित्व।
2. संस्थागत विविधता: बचत तथा ऋण, बहुउद्देश्यीय, कृषि, दुग्ध, उपभोक्ता, महिला सहकारी समेटिने।
3. सञ्चालन अवधि: कम्तीमा ३ वर्ष सञ्चालनमा रहेका संस्था।
4. सदस्य संख्या: न्यूनतम ५० सदस्य भएका संस्था।
5. कानुनी वैधता: सहकारी ऐन, २०७४ अनुसार दर्ता भई वार्षिक लेखापरीक्षण गरेका संस्था।

---

५. नमुना छनोट प्रक्रिया (Sampling Procedure)

चरण १ — प्रदेश छनोट: सातै प्रदेशलाई समेटिनेछ (Census of provinces)।

चरण २ — जिल्ला छनोट: प्रत्येक प्रदेशबाट सहकारी संस्थाको संख्या र भौगोलिक वितरणका आधारमा २–३ जिल्ला छनोट गरिनेछ।

चरण ३ — सहकारी संस्था छनोट: छानिएका जिल्लाभित्र सहकारीको प्रकार (बचत तथा ऋण, बहुउद्देश्यीय, कृषि आदि) अनुसार स्तरीकृत गरी Random Number Table प्रयोग गरी नमुना छनोट गरिनेछ।

चरण ४ — सदस्य छनोट: छानिएका प्रत्येक सहकारीबाट सदस्य दर्ता अभिलेखका आधारमा Systematic Random Sampling विधिबाट ५ जना सदस्य छनोट गरिनेछ।

---

६. तथ्याङ्क सङ्कलनका विधि तथा साधन (Data Collection Methods and Tools)

---

७. तथ्याङ्क विश्लेषण (Data Analysis)

• परिमाणात्मक तथ्याङ्क: SPSS/MS Excel प्रयोग गरी वर्णनात्मक तथ्याङ्क (mean, median, mode, standard deviation), correlation र regression विश्लेषण।
• गुणात्मक तथ्याङ्क: Thematic Analysis विधि।
• तुलनात्मक विश्लेषण: प्रदेश, प्रकार र सञ्चालन अवधिका आधारमा।

---

८. नैतिकता तथा गोपनीयता (Ethics and Confidentiality)

• सहभागीको सूचित सहमति (Informed Consent) अनिवार्य।
• व्यक्तिगत तथा संस्थागत तथ्याङ्कको गोपनीयता कायम।
• निष्पक्ष तथा पूर्वाग्रहमुक्त विश्लेषण।
• नेपाल राष्ट्र बैंक तथा सम्बन्धित निकायको अनुमति।

---

९. निष्कर्ष (Conclusion)

यस नमुना ढाँचाले नेपालका सहकारी संस्थाको वर्तमान अवस्था, सान्दर्भिकता र समस्याहरूको वैज्ञानिक तथा प्रतिनिधित्वपूर्ण मूल्याङ्कन गर्न सक्ने आधार तयार पार्दछ। बहुस्तरीय नमुना विधि, Cochran's Formula र स्पष्ट मापदण्डको प्रयोगले अध्ययनको निष्कर्षलाई विश्वसनीय, तथ्यपरक र नीति–निर्माणमा उपयोगी बनाउनेछ।

प्रश्न: राम्रोसँग तयार पारिएको प्रश्नावलीका मुख्य विशेषताहरू के के हुन्? यी विशेषताहरूले कसरी विश्वसनीय र वैध तथ्याङ्क सङ्कलनमा योगदान पुर्‍याउँछन्? व्याख्या गर्नुहोस्।

---

भूमिका

प्रश्नावली (Questionnaire) भन्नाले तथ्याङ्क सङ्कलनका लागि तयार पारिएको प्रश्नहरूको व्यवस्थित सूची हो। सर्वेक्षण अनुसन्धानको मुख्य औजारका रूपमा प्रयोग हुने यो औजार राम्रोसँग तयार नभएमा सङ्कलित तथ्याङ्क अविश्वसनीय र अवैध हुन्छ। त्यसैले प्रश्नावली निर्माण गर्दा यसका मुख्य विशेषताहरूलाई ध्यान दिनु अनिवार्य हुन्छ।

राम्रोसँग तयार पारिएको प्रश्नावलीका मुख्य विशेषताहरू

१. स्पष्टता र सरलता (Clarity and Simplicity)

प्रश्नहरू छोटो, सरल र बुझ्न सजिलो हुनुपर्छ। दुई अर्थ लाग्ने, जटिल वा प्राविधिक शब्दको प्रयोग गर्नु हुँदैन। उत्तरदाताको शिक्षा तह र भाषिक क्षमताअनुसार प्रश्न तयार गर्नुपर्छ।

२. सान्दर्भिकता (Relevance)

प्रश्नहरू अनुसन्धानको उद्देश्यसँग प्रत्यक्ष सम्बन्धित हुनुपर्छ। उद्देश्यसँग सम्बन्धित नभएका प्रश्नले अनुसन्धानको केन्द्रबिन्दु भद्रगोल बनाउँछन्।

३. संक्षिप्तता (Brevity)

प्रश्नावली अनावश्यक रूपमा लामो हुनु हुँदैन। आवश्यक विषयवस्तु मात्र समेटिएको संक्षिप्त प्रश्नावलीले उत्तरदाताको समय र धैर्यको सम्मान गर्छ।

४. वस्तुनिष्ठता (Objectivity)

प्रश्नहरू तटस्थ, पूर्वाग्रहरहित र निर्देशात्मक (Leading) नभएका हुनुपर्छ। प्रश्नले उत्तरदातालाई कुनै निश्चित उत्तरतर्फ डोर्‍याउनु हुँदैन।

५. क्रमबद्धता र तार्किक प्रवाह (Logical Sequence)

सामान्य प्रश्नबाट सुरु गरी क्रमशः विशिष्ट र संवेदनशील प्रश्नतर्फ जानुपर्छ। यसले उत्तरदातालाई सहज महसुस गराउँछ।

६. उपयुक्त प्रश्न संरचना (Appropriate Question Types)

बन्द प्रश्न (Closed-ended), खुला प्रश्न (Open-ended), बहुविकल्पीय, लिकर्ट स्केल आदि आवश्यकताअनुसार प्रयोग गर्नुपर्छ। तथ्यात्मक जानकारीका लागि बन्द प्रश्न र गहिरो अनुभवका लागि खुला प्रश्न उपयुक्त हुन्छ।

७. पूर्वपरीक्षण (Pre-testing / Pilot Testing)

वास्तविक सर्वेक्षणअघि सानो समूहमा प्रश्नावली परीक्षण गरी दोष सुधार गर्नुपर्छ।

८. गोपनीयता र नैतिकता (Confidentiality and Ethics)

उत्तरदाताको पहिचान गोप्य राखिने आश्वासन दिनुपर्छ र व्यक्तिगत स्वतन्त्रताको सम्मान गर्नुपर्छ।

९. तटस्थ भाषा र सांस्कृतिक संवेदनशीलता

भाषा पक्षपातरहित र स्थानीय संस्कृतिप्रति संवेदनशील हुनुपर्छ।

१०. व्यवस्थित ढाँचा (Proper Format)

निर्देशन, परिचय, प्रश्न खण्ड र धन्यवाद जस्ता भागहरू समावेश भएको स्पष्ट ढाँचा हुनुपर्छ।

यी विशेषताहरूले विश्वसनीय र वैध तथ्याङ्क सङ्कलनमा पुर्‍याउने योगदान

विश्वसनीयता (Reliability) मा योगदान:

• स्पष्ट र सरल प्रश्नले सबै उत्तरदाताले एउटै अर्थमा बुझ्ने भएकाले पुनः परीक्षण गर्दा समान नतिजा आउँछ।
• एकरूप प्रश्न संरचना र क्रमले निरन्तरता कायम राख्छ।
• पूर्वपरीक्षणले अस्पष्ट प्रश्न पहिचान गरी त्रुटि घटाउँछ।
• तालिमप्राप्त सर्वेक्षक र मानक निर्देशनले पूर्वाग्रह (Bias) न्यूनीकरण गर्छ।

वैधता (Validity) मा योगदान:

• सान्दर्भिक प्रश्नले अनुसन्धान उद्देश्यअनुसारकै तथ्याङ्क सङ्कलन गर्ने सुनिश्चित गर्छ।
• वस्तुनिष्ठ र तटस्थ भाषाले उत्तरदाताको वास्तविक धारणा प्रतिबिम्बित गर्छ।
• उपयुक्त प्रश्न संरचनाले मापन गर्न खोजिएको विषयवस्तु नै मापन हुन्छ।
• गोपनीयताको आश्वासनले उत्तरदातालाई सत्य उत्तर दिन प्रोत्साहित गर्छ।

निष्कर्ष

प्रश्नावली अनुसन्धानको मेरुदण्ड हो। स्पष्टता, सान्दर्भिकता, वस्तुनिष्ठता, तार्किक क्रम, पूर्वपरीक्षण र नैतिकता जस्ता विशेषताले युक्त प्रश्नावलीले विश्वसनीयता र वैधता दुवै सुनिश्चित गरी अनुसन्धानको निष्कर्षलाई विश्वसनीय बनाउँछ।

Importance of Defining the Research Problem Appropriately

Defining the research problem properly is the foundation of the entire research process. Its importance:

1. Gives direction and focus — A clearly defined problem tells the researcher exactly what to study, preventing aimless wandering and wasted effort.
2. Determines research design — The nature of the problem decides whether the study will be descriptive, exploratory, explanatory or experimental.
3. Guides data collection — It specifies what data are needed, from whom, and by what methods, thus avoiding irrelevant or excessive data.
4. Saves time, cost and effort — A precise problem statement prevents duplication of work and reduces the risk of collecting useless information.
5. Helps in formulating objectives and hypotheses — Clear objectives and testable hypotheses can only emerge from a well-defined problem.
6. Facilitates meaningful analysis and interpretation — Findings can be interpreted correctly only against a well-stated problem.
7. Prevents ambiguity in communication — Sponsors, supervisors and readers understand what the study intends to achieve.
8. Increases the value of the research — A well-defined problem yields conclusions that are useful for policy, practice and theory.

Techniques Involved in Defining the Research Problem

Defining a research problem is a systematic process. The main techniques are:

(a) Statement of the problem in general terms The problem is first stated broadly — for example, "low saving mobilisation in rural Nepal" — so that its dimensions become visible.

(b) Understanding the nature and origin of the problem The researcher examines how the problem arose, its background, and the socio-economic, institutional and legal setting in which it exists. This helps separate the real problem from its symptoms.

(c) Surveying the available literature Relevant books, journals, reports, government publications and past theses are reviewed to know what is already established, what methods were used, and where gaps exist. This prevents duplication and sharpens the problem.

(d) Discussion with knowledgeable persons / experts Consultation with supervisors, subject experts, bank officials, researchers and practitioners brings out practical insights, hidden dimensions and feasibility considerations not found in literature.

(e) Preliminary field observation / pilot study A small exploratory visit or pilot study in the actual field (e.g., a branch office or a sample of respondents) helps the researcher test assumptions, understand ground realities and refine the problem.

(f) Narrowing down to a specific, researchable problem From the broad area, the researcher selects one specific, feasible and manageable problem, keeping in view time, money, data availability and competence.

(g) Rephrasing the problem as a working proposition The problem is finally converted into a clear statement — objectives, research questions and hypotheses — in operational and measurable terms.

Conclusion: Appropriate definition of the research problem ensures that the study is purposeful, feasible and useful. The techniques above — general statement, study of origin, literature survey, expert consultation, pilot study, narrowing down and rephrasing — together make the problem precise, testable and researchable.

Sample Design

Sample design is the definite plan or scheme that is prepared before collecting data, which specifies how the sample is to be selected from the population. It is the blueprint that determines the target population, the sampling frame, the sampling unit, the sample size, the sampling technique and the method of estimation. In short, sample design is the complete plan of drawing a sample from a given population for the purpose of study.

Components of a Sample Design

1. Target population (universe): The entire group of elements about which the researcher wants to draw conclusions.
2. Sampling frame: The list or source from which the sample units are actually drawn (e.g., voter list, list of registered firms).
3. Sampling unit: The individual element or group of elements that is selected in the sample.
4. Sample size: The number of units to be selected — decided on the basis of precision, confidence level, variability and cost.
5. Sampling technique/method: The procedure used to select units — probability (random, systematic, stratified, cluster) or non-probability (convenience, judgment, quota, snowball).
6. Sampling error and precision: The acceptable margin of error and the confidence level desired.
7. Budget and time: The financial and time resources available, which constrain the design.
8. Method of estimation: The statistical technique used to estimate population parameters from sample data.

Snowball Sampling Method

Snowball sampling (also called chain-referral sampling) is a non-probability sampling technique in which the researcher first identifies a few respondents who meet the criteria, and these initial respondents then refer other persons from their network, who in turn refer further persons. The sample thus grows like a rolling snowball.

Procedure: Select a few initial subjects (seeds) → collect data from them → ask them to name others → contact those referred → repeat until the required sample size is reached.

Advantages

1. Useful for studying hidden, rare or hard-to-reach populations (drug users, sex workers, migrant workers, illegal traders).
2. Low cost and less time compared to locating such respondents independently.
3. No need for a sampling frame, which usually does not exist for such groups.
4. Builds trust and cooperation, as referrals come through known contacts.
5. Useful in exploratory and qualitative research where the population is unknown.

Limitations

1. It is a non-probability method, so findings cannot be generalized to the whole population.
2. Selection bias — the sample depends on the network of the initial seeds.
3. Referral bias / gatekeeper problem — respondents tend to refer people similar to themselves.
4. Privacy and confidentiality issues, since respondents must reveal their contacts.
5. The sample may not be representative, and sample size cannot be controlled precisely.
6. Initial seeds may be difficult to find, and the chain may stop early if respondents refuse to refer.

Preparing a Questionnaire on the Loan Status of Microfinance Institutions

A questionnaire is a structured set of questions designed to collect primary data from respondents. For a study on the loan status of microfinance institutions (MFIs), the questionnaire must be prepared systematically so that reliable and comparable data on loan disbursement, use, repayment and default can be collected. The following steps describe how I would prepare it.

1. Determine the Objectives and Information Needs

First, I would clearly define what the study seeks to measure. For "loan status", the key information needs are:

• Who borrows and why (borrower profile and purpose of loan)
• How much is borrowed and on what terms (amount, interest, tenure, instalment)
• How the loan is used (productive or consumptive)
• How it is repaid (repayment status, timeliness, overdue, default)
• What problems borrowers face (causes of delay or default)

2. Decide the Respondent and Unit of Study

The respondent may be the borrower (loan client), the loan officer, or both. For loan status, the borrower is the primary respondent, while records of the branch may be used for verification. The unit of study is one loan account.

3. Choose the Type of Questions

I would use a mix of:

• Closed-ended questions (yes/no, multiple choice) for easy coding and analysis.
• Open-ended questions for reasons and suggestions.
• Rating/scale questions (e.g., Likert scale) for satisfaction and severity of problems.

4. Design the Structure of the Questionnaire

A standard structure would be:

Section A – Respondent Profile: name (optional), age, sex, caste/ethnicity, education, occupation, family size, income, address.

Section B – Loan Details: source of loan, number of loans, amount, purpose, interest rate, tenure, instalment, collateral/group guarantee.

Section C – Loan Utilization: actual use of loan, proportion used for the stated purpose, diversion of funds.

Section D – Repayment Status: regular/irregular, amount repaid, outstanding balance, overdue amount, default, reasons for delay.

Section E – Problems and Suggestions: difficulties faced, support received, suggestions for improvement.

5. Draft Clear, Simple and Unbiased Questions

Questions should be short, specific and free from leading or double-barrelled wording. Technical terms should be avoided or explained. Sensitive questions should be placed later in the questionnaire.

6. Pre-test and Finalize

The draft would be pre-tested on a small sample (5–10 borrowers) to check clarity, sequence and length. Necessary corrections would then be made before final printing.

Examples of Questions

Section A – Respondent Profile

1. Sex: (a) Male (b) Female (c) Other
2. Age: ……… years
3. Occupation: (a) Agriculture (b) Business (c) Service (d) Labour (e) Other …………
4. Monthly household income: Rs. …………

Section B – Loan Details

5. From which institution have you taken the loan? (a) Bank (b) Microfinance (c) Co-operative (d) Money lender (e) Other …………
6. For what purpose did you take the loan? (a) Agriculture (b) Livestock (c) Small business (d) Education (e) Household consumption (f) Other …………
7. How much loan did you receive? Rs. …………
8. What is the annual interest rate on your loan? ………%
9. What is the repayment tenure? ……… months
10. What is the amount of your periodic instalment? Rs. …………

Section C – Loan Utilization

11. Did you use the whole loan amount for the stated purpose? (a) Yes (b) No
12. If no, what proportion was used for other purposes? ………%

Section D – Repayment Status

13. What is your present repayment status? (a) Fully repaid (b) Repaying regularly (c) Repaying irregularly (d) Overdue (e) Defaulted
14. How much amount has been repaid so far? Rs. …………
15. How much amount is still outstanding? Rs. …………
16. If repayment is delayed, what are the reasons? (Tick more than one) (a) Low income (b) Crop failure (c) High interest (d) Illness (e) Diversion of loan (f) Other …………

Section E – Problems and Suggestions

17. What problems do you face in repaying the loan? …………
18. What suggestions do you have to improve the loan system? …………

7. Points to Remember

• Use simple, unambiguous language.
• Avoid personal or sensitive questions unless necessary.
• Keep the sequence logical (profile → loan → use → repayment → problems).
• Provide "Others (please specify)" options.
• Ensure confidentiality of respondents' information.
• Keep the questionnaire as short as possible to avoid respondent fatigue.

Conclusion

Thus, a good questionnaire for studying loan status is prepared by defining objectives, identifying respondents, selecting appropriate question types, structuring it into logical sections, drafting clear questions, and pre-testing before final use. This ensures valid, reliable and analysable data on the loan status of microfinance institutions.

Relationship between Reliability and Validity of Data

Reliability refers to the consistency, stability and dependability of a measurement instrument — that is, repeated measurements under the same conditions give the same result. Validity refers to the accuracy of a measurement instrument — that is, it measures what it is actually intended to measure.

The relationship between the two can be stated as follows:

1. Reliability is a necessary but not sufficient condition for validity. A scale must be reliable before it can be valid, but a reliable scale may still be invalid. For example, a weighing machine that consistently shows 2 kg extra for every object is reliable (consistent) but not valid (inaccurate).
2. Validity implies reliability. If a scale is valid, it is automatically reliable, because a measure cannot consistently capture the true concept without being stable.
3. A scale can be reliable without being valid, but it cannot be valid without being reliable. Thus reliability is the floor, and validity is the ceiling, of measurement quality.
4. Both are matters of degree, not absolutes. A scale is reported as "highly reliable" or "moderately valid," not simply reliable or valid.

In short, reliability concerns consistency of measurement, while validity concerns correctness of measurement. Reliability is about "how well" the instrument measures; validity is about "what" it measures.

Assessing the Validity of a Multi-Item Scale

A multi-item scale (e.g., a Likert scale with several items measuring one construct) is assessed for validity through the following steps:

1. Content (Face) Validity

• Prepare the items strictly on the basis of the theoretical definition of the construct and the existing literature.
• Submit the draft scale to a panel of subject experts and get their judgement on whether each item represents the intended domain.
• Compute the Content Validity Index (CVI) or Content Validity Ratio (CVR) from expert ratings; items with low values are revised or dropped.
• This ensures the scale covers all dimensions of the construct adequately.

2. Construct Validity Assessed statistically after data collection, mainly through:

(a) Convergent Validity

• Run Exploratory Factor Analysis (EFA) or Confirmatory Factor Analysis (CFA).
• Check that all items load on their intended factor with factor loading ≥ 0.50 (ideally ≥ 0.70).
• Check Average Variance Extracted (AVE) ≥ 0.50 and Composite Reliability (CR) ≥ 0.70.
• The scale should correlate highly with other measures of the same construct.

(b) Discriminant Validity

• Verify that the √AVE of a construct is greater than its correlation with any other construct (Fornell–Larcker criterion).
• Alternatively, use the Heterotrait–Monotrait (HTMT) ratio < 0.85.
• This confirms the scale measures a distinct construct and not another one.

(c) Nomological / Criterion Validity

• Correlate the scale score with an established gold-standard measure of the same construct (concurrent validity) or with a future outcome (predictive validity).
• A high correlation confirms that the scale behaves as theory predicts.

3. Dimensionality Check

• Use exploratory factor analysis to verify that the items group into the theoretically expected number of dimensions, with eigenvalue > 1 and adequate total variance explained (usually ≥ 60%).

4. Pilot Testing

• Administer the scale to a small sample (30–50 respondents) before the main study to detect ambiguous, double-barrelled or irrelevant items and to refine wording.

5. Reliability Check as a Precondition

• Compute Cronbach's alpha (≥ 0.70) and item-total correlation (≥ 0.30); delete items that reduce reliability. Only a reliable scale can be further tested for validity.

Conclusion: Validity of a multi-item scale is established by first ensuring content adequacy through expert judgement, then confirming the factor structure, convergent and discriminant validity through EFA/CFA and AVE/CR/HTMT statistics, and finally verifying criterion-related validity against an external standard.

Components of a Research Proposal

A research proposal is a structured document that presents a plan for a study and seeks approval or funding. Its main components are:

1. Title — A clear, specific and concise statement reflecting the subject and scope of the study.
2. Introduction / Background — Background of the problem, its context, and why the study is relevant.
3. Statement of the Problem — The gap between the existing situation and the desired situation, expressed in specific terms.
4. Objectives of the Study — General and specific objectives stated in measurable, action-oriented terms.
5. Research Questions / Hypotheses — Questions the study seeks to answer, or testable statements to be verified.
6. Rationale / Significance — Why the study matters to policy, practice, and knowledge.
7. Literature Review — Review of related theories and previous studies, identifying the research gap.
8. Conceptual / Theoretical Framework — The variables and their expected relationships, based on theory.
9. Research Methodology — Research design, population and sample, data collection tools, data analysis techniques, and ethical considerations.
10. Scope and Limitations — Boundaries of the study and constraints expected.
11. Time Schedule (Work Plan) — Activities and the time allotted to each, often in a Gantt chart.
12. Budget / Cost Estimate — Expected expenditure on personnel, travel, data collection, and report writing.
13. References — Sources cited, in a standard format (APA, etc.).
14. Annexes — Questionnaire, interview schedule, and other supporting documents.

Defining the Research Methods (if I were preparing the proposal)

While preparing the proposal, I would define the research methods under the Research Methodology section, in the following sequence:

1. Research Design — State whether the study is descriptive, exploratory, explanatory, or causal, and justify the choice in light of the objectives. For a descriptive study of banking practices, for example, a descriptive and analytical design would be used.
2. Nature of Data — Specify whether quantitative, qualitative, or mixed data are required, and whether they are primary or secondary.
3. Population and Sampling — Define the population (e.g., all employees/customers of commercial banks in Nepal), the sampling frame, sample size, and sampling technique (probability or non-probability), with justification.
4. Data Collection Tools and Techniques — Name the instruments (structured questionnaire, interview schedule, observation checklist, focus group discussion) and the sources of secondary data (NRB reports, annual reports, CBS publications). Include a pilot test to check reliability and validity.
5. Variables and Measurement — Operationalise the variables and state the measurement scales (nominal, ordinal, interval, ratio) and the Likert scale used.
6. Data Analysis Techniques — Specify descriptive tools (frequency, percentage, mean, standard deviation) and inferential tools (correlation, regression, chi-square, t-test, ANOVA), along with the software (SPSS/Excel).
7. Ethical Considerations — Mention informed consent, confidentiality, anonymity, and avoidance of plagiarism.
8. Limitations of the Methods — State the constraints of the chosen design, sample, and tools so that the reader can judge the validity of the findings.

In this way, the research methods are defined in a logical sequence — design, population and sample, tools, analysis, ethics, and limitations — so that the study is systematic, replicable, and verifiable.