Unit 3: Data and Methods of Data Collection | Research Methodology

Unit 3 Overview

Data are the basic raw materials of research. Every scientific investigation depends on the collection, organization, analysis and interpretation of appropriate data. Without reliable data, even a well-designed research study may produce inaccurate or misleading conclusions.

Data may be collected from people, biological specimens, environmental samples, laboratory experiments, documents, records, observations, questionnaires, interviews and many other sources. The method selected for collecting data depends on the research question, objectives, nature of the study, available resources and ethical requirements.

Key idea: Data collection is the systematic process of obtaining relevant information from appropriate sources using suitable methods and instruments to answer research questions and achieve research objectives.

This unit deals with the nature and forms of data, measurement scales, methods of collecting data in social sciences, laboratory-oriented methods, non-response and the construction of appropriate research questions.

3.1 Needs and Nature of Data

Introduction

Data are observations, measurements, facts, responses or information collected for a particular research purpose. Data provide the empirical basis for scientific investigation. Researchers collect data to describe phenomena, compare groups, test hypotheses, identify relationships and make conclusions.

Definition: Data are recorded observations, measurements or information collected systematically for the purpose of research, analysis and interpretation.

Need for Data in Research

Data are required at almost every stage of research. The major needs are:

Understanding a problem: Data help the researcher understand the magnitude and characteristics of a research problem.
Describing a phenomenon: Data allow researchers to describe the occurrence or distribution of a phenomenon.
Comparison: Data permit comparison between groups, places, treatments or time periods.
Testing hypotheses: Quantitative and qualitative evidence can be used to determine whether a proposed hypothesis is supported.
Finding relationships: Data help identify associations between variables.
Decision making: Research findings based on reliable data can support scientific and practical decisions.
Prediction: Historical and current data may help researchers identify trends and make reasonable predictions.
Evaluation: Data are used to evaluate programmes, interventions, treatments and policies.

Nature of Data in Natural Sciences

Natural sciences include disciplines such as biology, microbiology, chemistry, physics, environmental science and related fields. Data in these sciences are commonly obtained through observation, measurement, experimentation and laboratory analysis.

Natural science data are often characterized by greater emphasis on measurable variables, controlled conditions, reproducibility and objective observation. However, qualitative observations are also important, especially in biological and environmental research.

Examples of Natural Science Data

Temperature of a culture medium
pH of a water sample
Bacterial colony count
Concentration of glucose in blood
Diameter of inhibition zone
Body weight of experimental animals
Number of microorganisms per millilitre
Presence or absence of a pathogen

Nature of Data in Social Sciences

Social sciences deal with human beings, social groups, behaviour, attitudes, beliefs, opinions, institutions and social relationships. Therefore, data may be obtained through questionnaires, interviews, observations, schedules, surveys, records and documents.

Social science data may be both quantitative and qualitative. For example, age, income and number of family members are quantitative, whereas opinions, attitudes and experiences may be qualitative.

Different Forms of Information

1. Primary Data

Primary data are original data collected directly by the researcher for the specific purpose of the current research study.

Examples include laboratory measurements, questionnaire responses, interview responses, direct observations and experimental results.

Example: If a researcher collects water samples from different points of the Bagmati River and determines their coliform counts, those measurements are primary data for that study.

Advantages of Primary Data

Collected specifically for the research objective.
Researcher has greater control over the collection process.
Data can be collected using standardized procedures.
Current information can be obtained.

Limitations of Primary Data

Time-consuming.
May be expensive.
Requires trained investigators.
Errors may occur during collection.

2. Secondary Data

Secondary data are data that have already been collected, recorded or published by another person, organization or institution for a different purpose or sometimes for a related purpose.

Examples include census reports, government publications, hospital records, research articles, books, institutional reports, databases and previous research studies.

Advantages of Secondary Data

Usually less expensive.
Saves time.
Useful for background information and literature review.
Can provide historical information.
Useful for comparing current findings with previous studies.

Limitations of Secondary Data

May not exactly match the current research objectives.
Data may be outdated.
Quality may be difficult to verify.
Methods used for original collection may be unknown.
Definitions and measurement methods may differ.

Primary Data vs Secondary Data

Basis	Primary Data	Secondary Data
Meaning	Collected directly for the present study.	Already collected by someone else.
Source	Participants, experiments, observations, specimens.	Books, reports, records, databases and publications.
Cost	Usually higher.	Usually lower.
Time	More time required.	Usually quicker to obtain.
Control	Greater researcher control.	Limited control over original collection.

Cross-sectional Data

Cross-sectional data are collected from different individuals, groups, locations or units at approximately one point or a particular period of time.

The main purpose is to describe the characteristics of a population or compare different groups at a particular time.

Example: Measuring the prevalence of intestinal parasites among 500 students during a particular academic year is a cross-sectional study.

Categorical Data

Categorical data classify observations into distinct categories or groups. The categories may represent characteristics such as sex, blood group, occupation, disease status or type of microorganism.

Categorical data may be nominal or ordinal depending on whether the categories have a meaningful order.

Categorical data → Categories → Classification → Frequency/Proportion

Time-Series Data

Time-series data consist of observations collected over successive points in time. The purpose is usually to study changes, trends, patterns or seasonal variation.

For example, monthly bacterial counts in a river over two years constitute time-series data.

Spatial Data

Spatial data describe observations according to their geographical location or physical position.

Spatial information is particularly important in environmental, epidemiological and geographical research.

Example: Coliform counts measured at Sundarijal, Gokarna, Guheshwori, Pashupati and Tilganga represent spatially distributed data.

Ordered Data

Ordered data are data in which observations or categories can be arranged according to a meaningful order or rank.

Examples include disease severity classified as mild, moderate and severe, or satisfaction classified as poor, fair, good and excellent.

3.2 Types of Data According to Scales of Measurement

Introduction

A measurement scale describes the way observations are classified, ordered or numerically represented. The choice of statistical analysis depends partly on the scale of measurement used.

The four classical scales of measurement are:

Nominal scale
Ordinal scale
Interval scale
Ratio scale

Qualitative Data

Qualitative data describe qualities, characteristics, categories, experiences or meanings rather than primarily expressing quantities.

Examples include colour, type of microorganism, gender category, occupation, symptoms and opinions.

Quantitative Data

Quantitative data are numerical data representing quantities that can generally be counted or measured.

Examples include age, weight, temperature, blood pressure, bacterial count and concentration.

Qualitative Data	Quantitative Data
Usually describes qualities or categories.	Represents numerical quantities.
Example: blood group.	Example: haemoglobin concentration.
May be nominal or ordinal.	May be discrete or continuous.

1. Nominal Scale

The nominal scale is the simplest level of measurement. It classifies observations into mutually distinct categories without any meaningful order or ranking.

Numbers may sometimes be assigned to categories, but these numbers are only labels and do not indicate quantity.

Examples

Blood group: A, B, AB, O
Sex category
Type of microorganism
Marital status
Presence or absence of disease

Key point: Nominal data have categories but no natural ranking.

2. Ordinal Scale

The ordinal scale classifies observations into categories that have a meaningful order or rank. However, the exact difference between adjacent categories is not necessarily equal.

Examples

Mild, moderate and severe disease
Poor, fair, good and excellent quality
First, second and third rank
Low, medium and high risk

Key point: Ordinal data tell us which observation is higher or lower in rank, but do not necessarily tell us the exact numerical difference between ranks.

3. Interval Scale

The interval scale has ordered values with equal intervals between measurements. However, it does not possess a true or absolute zero.

Because the intervals are equal, addition and subtraction are meaningful. However, statements based on ratios are not generally meaningful.

Example

Temperature measured in degrees Celsius or Fahrenheit is commonly used as an example of interval measurement because zero does not represent complete absence of temperature.

4. Ratio Scale

The ratio scale possesses all the characteristics of the interval scale and also has a meaningful absolute zero.

Therefore, all major arithmetic operations, including ratios, can be meaningfully interpreted.

Examples

Weight
Height
Age
Length
Concentration
Number of bacterial colonies
Time duration

Comparison of Measurement Scales

Scale	Classification	Order	Equal Interval	True Zero	Example
Nominal	Yes	No	No	No	Blood group
Ordinal	Yes	Yes	No	No	Disease severity
Interval	Yes	Yes	Yes	No	Temperature °C
Ratio	Yes	Yes	Yes	Yes	Weight, height

3.3 Data Collection in Social Sciences

Social science research requires information about people, their behaviour, opinions, attitudes, experiences, relationships and social conditions. Several methods are used to obtain such information.

Important methods include schedules, questionnaires, interviews, surveys, census and observation.

Schedule

Meaning of Schedule

A schedule is a structured data collection instrument containing a set of questions or items that are completed by a trained investigator or enumerator during interaction with the respondent.

Definition: A schedule is a structured form containing questions or information items that are filled in by an investigator after obtaining information from respondents.

Purpose of Schedule

To collect standardized information.
To obtain complete responses from respondents.
To assist respondents who have difficulty reading or writing.
To maintain uniformity among different investigators.
To reduce omission of important information.
To facilitate systematic analysis.

Kinds of Schedule

Observation schedule: Used to record systematically what the researcher observes.
Interview schedule: Contains questions asked by an investigator during an interview.
Rating schedule: Used to record the degree or level of a characteristic using a rating system.
Evaluation schedule: Used to assess performance, programmes, activities or characteristics.

Content of a Schedule

A well-prepared schedule may contain the following components:

Title of the study
Identification number
Date and place of data collection
Instructions to the investigator
Respondent identification information
Main questions
Response categories
Observation or remarks section
Investigator's signature or identification

Limitations of Schedule

Requires trained investigators.
Can be expensive when large populations are studied.
Interviewer bias may occur.
Data collection may be time-consuming.
Respondents may provide socially desirable answers.
Quality may vary between investigators.

Questionnaire

Meaning

A questionnaire is a structured research instrument consisting of a series of written questions presented to respondents to obtain information relevant to the objectives of a study.

Definition: A questionnaire is a systematic set of written questions used to collect information directly from respondents.

Types of Questionnaire

1. Structured Questionnaire

Contains predetermined questions and response categories. It provides high uniformity and is convenient for quantitative analysis.

2. Unstructured Questionnaire

Provides greater freedom for respondents to express their views and may contain more open-ended questions.

3. Open-ended Questionnaire

Respondents answer in their own words without being restricted to predetermined alternatives.

4. Closed-ended Questionnaire

Provides predetermined response options such as yes/no, multiple choice or rating categories.

5. Mixed Questionnaire

Contains both open-ended and closed-ended questions.

Requisites of a Good Questionnaire

A good questionnaire should be carefully designed because poorly constructed questions can produce biased or unreliable information.

Clear language: Questions should be easy to understand.
Simple wording: Technical terms should be avoided unless respondents understand them.
Specific questions: Questions should focus on one issue at a time.
Relevant questions: Every question should contribute to the research objectives.
Logical sequence: Questions should follow a meaningful order.
Neutral wording: Questions should not lead respondents toward a particular answer.
Reasonable length: The questionnaire should not be unnecessarily long.
Appropriate response options: Options should be clear and sufficiently comprehensive.
Pilot testing: The questionnaire should be tested before final use.
Ethical considerations: Privacy and confidentiality should be respected.

Importance of Questionnaire Method

Can collect information from a large number of respondents.
Relatively inexpensive compared with many interview methods.
Provides standardized questions.
Responses can be coded and statistically analyzed.
Can be distributed physically or electronically.
Respondents may have greater privacy for sensitive questions.

Limitations of Questionnaire Method

Low response rate may occur.
Respondents may misunderstand questions.
Researcher cannot always clarify questions immediately.
Incomplete responses may occur.
Recall bias may affect answers.
Social desirability bias may occur.
Not suitable for people with serious literacy difficulties without assistance.

Difference Between Schedule and Questionnaire

Basis	Schedule	Questionnaire
Completed by	Usually investigator/enumerator.	Usually respondent.
Interaction	Usually involves direct interaction.	May not require direct interaction.
Literacy requirement	Less dependent on respondent literacy.	Usually requires sufficient literacy.
Cost	Usually higher.	Usually lower.
Clarification	Investigator can clarify questions.	Clarification may not be immediately available.
Bias	Interviewer bias may occur.	Response and interpretation bias may occur.

Format of a Questionnaire

Title of study
Brief introduction
Purpose of the study
Instructions to respondent
Demographic questions where appropriate
Main research questions
Response options
Open-ended questions where required
Closing statement

Problems in Questionnaire Method

Non-response
Incomplete responses
Ambiguous questions
Leading questions
Recall errors
Response bias
Sampling problems
Language problems
Questionnaire fatigue

Interviews

Meaning of Interview

An interview is a method of collecting information through direct interaction between the researcher or interviewer and the respondent. Questions are asked and responses are recorded systematically.

Definition: An interview is a purposeful communication between an interviewer and respondent for obtaining information relevant to a research study.

Direct Interview

In a direct interview, the researcher asks questions directly to the respondent and obtains information through personal interaction.

Indirect Interview

In an indirect interview, information may be obtained through a third person, intermediary or indirect source when direct questioning is difficult or inappropriate.

Ways of Obtaining an Interview

Personal contact
Telephone communication
Written or electronic communication
Appointment through an organization or institution
Community or group contact

Focus Group Discussion

A focus group discussion is a qualitative data collection method in which a small group of participants discusses a particular topic under the guidance of a trained moderator.

The interaction among participants is important because participants may stimulate each other's ideas and reveal opinions, experiences and perceptions that might not emerge during individual interviews.

Characteristics

Small group of selected participants.
Specific research topic.
Moderated discussion.
Interactive communication.
Mainly qualitative information.

Advantages

Generates detailed opinions and experiences.
Allows interaction among participants.
Useful for exploring new research topics.
Can identify diverse viewpoints.

Limitations

Dominant participants may influence discussion.
Some participants may hesitate to speak.
Results are not necessarily statistically representative.
Requires a skilled moderator.

Survey and Census

Survey

A survey is a systematic method of collecting information from a population or sample using questionnaires, interviews, schedules or other standardized procedures.

Census

A census involves collecting information from every member or unit of the defined population.

Similarity Between Survey and Census

Both collect systematic information.
Both require defined populations or study units.
Both use structured data collection procedures.
Both can provide information for decision making and planning.

Differences Between Survey and Census

Basis	Survey	Census
Coverage	Usually studies a sample of the population.	Studies every unit of the defined population.
Cost	Generally lower.	Generally higher.
Time	Usually faster.	Usually requires more time.
Resources	Requires fewer resources.	Requires greater resources.
Sampling error	May occur when a sample is used.	Sampling error is not present because the whole population is covered.

Methods of Sampling Human Population

When it is not feasible to study every member of a population, researchers select a sample. Common methods include:

Simple random sampling
Systematic sampling
Stratified sampling
Cluster sampling
Multistage sampling
Convenience sampling
Purposive sampling
Quota sampling
Snowball sampling

Observation Methods

Meaning

Observation is a method of collecting information by systematically watching, listening to and recording events, behaviour, characteristics or activities relevant to a research problem.

Participatory Observation

In participatory observation, the researcher becomes involved to some extent in the activities or environment being studied while observing the phenomenon.

Advantages

Provides detailed contextual information.
Helps understand behaviour in its natural setting.
Can reveal information that participants may not report directly.

Limitations

Researcher involvement may influence behaviour.
Time-consuming.
Objectivity may be difficult to maintain.

Non-participatory Observation

In non-participatory observation, the researcher observes the phenomenon without actively participating in the activities being studied.

This approach may provide greater separation between researcher and participants but may provide less contextual understanding than participatory observation.

PRA – Participatory Rural Appraisal

Participatory Rural Appraisal is an approach that encourages local people to participate actively in identifying, analysing and discussing their own conditions, problems, resources and priorities.

The emphasis is on participation, local knowledge and community involvement rather than treating community members simply as sources of information.

RRA – Rapid Rural Appraisal

Rapid Rural Appraisal is an approach used to obtain information about a community or rural setting relatively quickly using multiple methods, such as observation, interviews and review of available information.

Controls Used in Social Survey

Controls are procedures used to maintain consistency, quality and reliability during data collection.

Standardized questionnaire or schedule
Training of investigators
Pilot testing
Clear instructions
Supervision of field workers
Randomization where appropriate
Checking completed forms
Re-interview or verification
Proper coding and data entry procedures

Observational Documents and Their Types

Observational documents are records used to systematically document observations during research.

Field notes: Written records of observations made during fieldwork.
Checklists: Lists used to record whether specific characteristics or behaviours are present.
Observation schedules: Structured forms used to record predetermined observations.
Rating scales: Used to record the degree or intensity of a characteristic.
Diaries: Chronological records of activities or experiences.
Audio/video records: Used where ethically and technically appropriate.

3.4 Data Collection in Lab-oriented Sciences and Technologies

Introduction

Laboratory-oriented research relies heavily on systematic observation, measurement and experimentation. In biological and microbiological research, data may be obtained from specimens, cultures, biochemical reactions, instruments and experimental procedures.

Methods of Data Collection in Natural Sciences

Direct observation
Measurement using instruments
Laboratory experimentation
Collection and analysis of specimens
Microscopic examination
Microbiological culture and enumeration
Biochemical analysis
Molecular methods
Environmental monitoring

Data Measurement

Measurement is the systematic assignment of numerical values or categories to characteristics according to defined rules.

In laboratory research, measurement may involve physical, chemical, biological or microbiological variables.

Examples

Measuring pH using a pH meter.
Measuring temperature using a thermometer.
Measuring absorbance using a spectrophotometer.
Counting bacterial colonies on culture media.
Measuring inhibition zones in antimicrobial susceptibility testing.

Observation Method in Laboratory Research

Laboratory observation involves systematic examination and recording of experimental or biological events. Observation may be direct, such as observing colony morphology, or instrument-assisted, such as recording absorbance.

Laboratory observations should be made under standardized conditions whenever possible so that measurements are comparable and reproducible.

Observation Schedule

An observation schedule is a structured form containing specific characteristics or events that the researcher intends to observe and record.

For example, a microbiology observation schedule may include colony morphology, pigmentation, haemolysis, growth pattern and biochemical reactions.

Observation Plan

An observation plan describes what will be observed, where and when observations will be made, who will perform the observation, what instruments will be used and how observations will be recorded.

What to observe → Where → When → How → Who → How to record

Check Sheet

A check sheet is a simple structured form used to record the presence, absence or frequency of predefined events or characteristics.

It is useful because information can be recorded quickly and consistently.

Rating Scale

A rating scale is an instrument used to indicate the degree, intensity, quality or level of a characteristic.

For example, microbial growth could be recorded using categories such as absent, slight, moderate and heavy growth when an appropriate standardized rating system is defined.

Likert Scale

A Likert scale is a commonly used rating scale for measuring attitudes, opinions or levels of agreement with statements.

A typical format may include:

Strongly Disagree → Disagree → Neutral → Agree → Strongly Agree

Likert-type items are particularly common in social and behavioural research.

Controlled Observation

Controlled observation takes place under conditions in which the researcher controls or standardizes important aspects of the observation environment.

This increases consistency and may reduce the influence of extraneous factors.

Non-controlled Observation

Non-controlled observation occurs in a relatively natural setting without deliberate manipulation or strict control of the environment.

It can provide realistic information about natural behaviour but may have less control over extraneous variables.

Sample and Specimen

A sample is a portion selected from a larger population for examination or analysis. A specimen generally refers to a biological or material sample obtained for examination, identification or diagnostic purposes.

Example: A portion of river water collected for microbiological analysis is a sample, while a clinical stool specimen submitted for laboratory examination is a biological specimen.

Sampling of Objects or Beings

Sampling means selecting a portion of objects, organisms, people or materials from a larger population so that the selected units can be studied.

In laboratory research, sampling must consider representativeness, sample size, collection technique, transport, storage and possible contamination.

Ethical Issues in Research

1. Anonymity

Anonymity means that the identity of a participant cannot be linked to the information provided by that participant.

2. Confidentiality

Confidentiality means that information provided by participants is protected and is not disclosed to unauthorized persons.

3. Informed Consent

Informed consent means that participants are given adequate information about the research, including its purpose, procedures, potential risks and benefits, and voluntarily agree to participate.

Important: Ethical research requires respect for participants, appropriate confidentiality, voluntary participation and informed consent where human participants are involved.

Observation Method vs Experimental Method

Basis	Observation	Experiment
Researcher intervention	Usually observes without deliberately manipulating the variable.	Researcher deliberately manipulates one or more variables.
Control	Often less control over environmental factors.	Usually greater control over relevant conditions.
Main purpose	Describe or understand naturally occurring events.	Study effects of manipulation and relationships between variables.
Example	Observing natural bacterial contamination.	Testing effect of temperature on bacterial growth.

3.5 Non-responses

Meaning of Non-response

Non-response occurs when a selected participant does not provide the required information, refuses to participate, cannot be contacted, provides incomplete information or otherwise fails to respond to the research request.

Definition: Non-response is the failure of selected study units or respondents to provide the information required by a research investigation.

Causes of Non-response

Refusal to participate
Inability to contact the respondent
Lack of interest
Long or complicated questionnaire
Sensitive questions
Lack of trust
Language difficulties
Illness or absence
Inconvenient timing
Fear concerning confidentiality

Estimation of Non-response Rate

The non-response rate indicates the proportion of selected units that failed to provide usable responses.

General formula:

Non-response rate = (Number of non-respondents ÷ Total number of eligible selected units) × 100

Example: If 20 people do not respond out of 200 eligible selected participants, the non-response rate is 10%.

Techniques for Getting Response

Explain the purpose of the study clearly.
Assure participants about confidentiality.
Keep questionnaires reasonably short.
Use simple and understandable language.
Choose an appropriate time for contact.
Send reminders when appropriate.
Provide clear instructions.
Build trust and maintain professional behaviour.
Use follow-up visits or communication where ethically appropriate.

Random Response Technique

The random response technique is a survey technique designed to encourage respondents to provide information about sensitive questions while providing greater privacy.

In this method, a randomizing device or rule determines whether the respondent answers the sensitive question directly or answers an unrelated question. The researcher uses the known probability structure to estimate the population-level response.

Purpose: The main purpose of random response techniques is to reduce fear, embarrassment and deliberate misreporting when sensitive information is being collected.

3.6 Questions Used in Data Collection

Questions are the basic instruments through which researchers obtain information from respondents in questionnaires, schedules and interviews. The quality of the questions strongly influences the quality of the data.

Types of Questions Used

1. Open-ended Questions

Respondents are allowed to answer in their own words.

Example: "What are the major problems you experience regarding access to healthcare?"

2. Closed-ended Questions

Respondents select an answer from predetermined alternatives.

Example: "Have you received vaccination? Yes / No"

3. Dichotomous Questions

Questions having two principal response categories, commonly Yes/No.

4. Multiple-choice Questions

Respondents select one or more options from a set of alternatives.

5. Rating Questions

Respondents indicate the intensity or degree of an attitude, satisfaction or characteristic using a defined scale.

6. Ranking Questions

Respondents arrange items according to preference, importance or priority.

7. Filter Questions

Filter questions determine whether a respondent is eligible for subsequent questions.

Example: "Have you used antibiotics during the last six months?"
If "No", questions specifically concerning recent antibiotic use may be skipped.

Questions to be Avoided

Poorly constructed questions can introduce systematic errors and bias. The following types should generally be avoided:

Ambiguous questions: Questions with more than one possible interpretation.
Leading questions: Questions that encourage a particular answer.
Loaded questions: Questions containing emotionally charged assumptions.
Double-barrelled questions: Questions asking two different things at the same time.
Very long questions: Questions that are difficult to understand.
Technical questions: Complex technical terms that respondents cannot understand.
Unnecessary personal questions: Sensitive information should be requested only when relevant and ethically justified.
Hypothetical questions: Should be used carefully because respondents may not accurately predict future behaviour.
Questions based on poor recall: Questions requiring respondents to remember events from an excessively long period may produce inaccurate responses.

Sequence of Questions

Questions should be arranged logically so that respondents can follow the questionnaire or interview easily.

Recommended General Sequence

Introduction and consent information
Simple and general questions
Questions related to the main topic
More specific questions
Sensitive questions where necessary
Demographic or background questions
Closing questions

Introduction → Easy questions → General questions → Specific questions → Sensitive questions → Background information → Closing

Importance of Proper Question Sequence

Creates a logical flow.
Reduces respondent confusion.
Builds respondent confidence.
Improves completion rate.
Reduces unnecessary skipping.
Improves consistency of collected data.