Collection of Data

COMPREHENSIVE CHAPTER NOTES: COLLECTION OF DATA

INTRODUCTION AND MEANING OF DATA COLLECTION

**Data collection** is the systematic process of gathering information to provide evidence for solving economic problems and reaching sound conclusions.

**Key Concepts:**

**Variable**: A characteristic that changes or varies. Represented by letters X, Y, or Z

Example: Food grain production in India varies from 108 million tonnes (1970-71) to 272 million tonnes (2016-17)

**Observation**: Each value of a variable

Example: 108 million tonnes is one observation of food grain production

**Data**: A collection of observations that serves as a tool for understanding economic problems

Provides factual information for decision-making

Forms the foundation for economic analysis and policy-making

**Purpose of Data Collection:**

To provide evidence for sound problem-solving

To understand trends and fluctuations (e.g., food grain production variations)

To support economic policy decisions

To enable comparative analysis across time and regions

---

SOURCES OF DATA: PRIMARY VS SECONDARY

Data can originate from two main sources, each with distinct characteristics and applications.

PRIMARY DATA

**Definition**: Data collected directly by the researcher for the first time through first-hand information.

**Characteristics:**

Collected through surveys, enquiries, and interviews

Original and specific to the researcher's needs

More time-consuming and expensive

High accuracy for specific research questions

Researcher maintains full control over methodology

**Example**: A researcher interviewing school students to determine the popularity of a filmstar among them collects primary data.

**Advantages:**

Tailored to specific research objectives

High reliability and validity

Data confidentiality can be maintained

Allows for control over quality of data

**Disadvantages:**

High cost (requires trained personnel, travel, incentives)

Time-consuming process

Requires significant resources

Cannot be collected retroactively for past periods

SECONDARY DATA

**Definition**: Data already collected, processed (scrutinised and tabulated), and published by another agency, which is then used by other researchers.

**Characteristics:**

Obtained from published sources (government reports, documents, newspapers, books, websites)

Primary to the original collector but secondary to subsequent users

Readily available and cost-effective

May not perfectly match researcher's specific needs

**Sources of Secondary Data:**

Government publications and statistical departments

RBI reports and monetary policy documents

Census reports and demographic data

Academic journals and research papers

Newspapers and magazines

International organisations (World Bank, IMF, UN)

Online databases and websites

NGO publications

**Example**: When a researcher uses data previously collected on filmstar popularity in a similar study, that data becomes secondary data.

**Advantages:**

Saves time and cost

Allows access to data from past years/decades

Enables comparative analysis across regions

Useful for longitudinal studies

**Disadvantages:**

May not exactly match research requirements

Accuracy depends on original collector's methodology

Possible outdated information

Limited control over data quality and collection methods

---

METHODS OF DATA COLLECTION: SURVEY DESIGN AND INSTRUMENTS

**Survey**: A systematic method of gathering information from individuals to describe specific characteristics (price, quality, usefulness, popularity, etc.).

QUESTIONNAIRE/INTERVIEW SCHEDULE PREPARATION

A questionnaire is the primary instrument used in surveys. It may be self-administered by respondents or administered by trained enumerators.

**Essential Guidelines for Questionnaire Design:**

**1. Length and Conciseness**

Minimize number of questions

Avoid unnecessary or repetitive questions

Respects respondent's time

**2. Language and Clarity**

Use simple, understandable language

Avoid ambiguous or difficult terminology

Technical jargon should be minimized

Ensure questions are precise and unambiguous

**3. Logical Sequence and Flow**

Arrange questions in order of comfort for respondent

Progress from general to specific questions

**Poor Example**: Ask about justification of electricity charges before asking about regularity of supply

**Good Example**: Ask about regularity of electricity supply first, then about charges

**4. Precision and Clarity**

Questions must be clear and specific

**Poor**: "What percentage of your income do you spend on clothing in order to look presentable?"

**Good**: "What percentage of your income do you spend on clothing?"

**5. Avoiding Ambiguity**

Provide specific options for responses

Enable quick and correct answering

**Poor**: "Do you spend a lot of money on books in a month?"

**Good**: "How much do you spend on books in a month?"

(i) Less than Rs 200

(ii) Rs 200-300

(iii) Rs 300-400

(iv) More than Rs 400

**6. Avoiding Double Negatives**

Do not use "Wouldn't you" or "Don't you" constructions

May lead to biased or confused responses

**Poor**: "Don't you think smoking should be prohibited?"

**Good**: "Do you think smoking should be prohibited?"

**7. Avoiding Leading Questions**

Questions should not guide respondents toward particular answers

Maintain neutrality

**Poor**: "How do you like the flavour of this high-quality tea?"

**Good**: "How do you like the flavour of this tea?"

**8. Avoiding Questions Indicating Alternatives**

Do not suggest or limit response options within question

**Poor**: "Would you like to do a job after college or be a housewife?"

**Good**: "What would you like to do after college?"

TYPES OF QUESTIONS IN QUESTIONNAIRE

**Open-Ended (Unstructured) Questions**

Allow respondent freedom to answer in their own words

No predetermined alternatives provided

Example: "What is your view about globalisation?"

**Advantages:**

Captures detailed, personalised responses

Reveals respondent's true perspective

Helps discover unexpected viewpoints

**Disadvantages:**

Difficult to interpret and analyse

Hard to standardise and score

Variations in response styles make comparison difficult

Time-consuming to process

**Closed-Ended (Structured) Questions**

Provide predetermined response options

Respondent selects from given alternatives

**Two-Way Questions** (Binary choice):

Only two possible answers: Yes or No

Simple and quick to answer

**Multiple Choice Questions**:

More than two options available

Example: "Why did you sell your land?"

(i) To pay off debts

(ii) To finance children's education

(iii) To invest in another property

(iv) Any other (please specify)

**Advantages of Closed-Ended Questions:**

Easy to use, score, and codify for analysis

Standardised responses enable comparison

Quick to complete

Suitable for quantitative analysis

**Disadvantages of Closed-Ended Questions:**

Difficult to construct with balanced alternatives

Possibility that true response not available in options

Restricts responses by providing predetermined choices

May not capture nuanced perspectives

Respondents may select "Any Other" if options inadequate

**"Any Other" Option**:

Allows respondents to write responses not anticipated by researcher

Prevents forced selection of unsuitable option

---

MODES OF DATA COLLECTION: THREE BASIC METHODS

Three primary methods exist for collecting survey data, each with distinct advantages, disadvantages, and applications.

PERSONAL INTERVIEWS

**Definition**: Researcher (or investigator) conducts face-to-face interviews with respondents.

**When Used**: When researcher has access to all members of population.

**Advantages:**

**Highest response rate** among all methods

Personal contact enables explanation of study and answering queries

Opportunity to request expansion on important answers

Minimises misinterpretation and misunderstanding

Non-verbal cues (facial expressions, reactions) provide supplementary information

Allows use of all types of questions (open and closed)

Best method for sensitive or complex questions

Interviewer can clarify ambiguous questions immediately

**Disadvantages:**

**Most expensive method** (requires trained interviewers, travel expenses, time)

Time-consuming (lengthy duration to complete survey)

Presence of researcher may inhibit honest responses (social desirability bias)

Possibility of interviewer influence on responses

Cannot reach geographically dispersed populations easily

Requires large team of trained enumerators

**Best For**: Urban areas, small populations, questions requiring clarification.

MAILING (QUESTIONNAIRE) SURVEYS

**Definition**: Questionnaire is sent to respondents by mail with request to complete and return by specified date.

**Modern Variants**: Online surveys and SMS surveys.

**Advantages:**

**Least expensive method** among all three

Allows researcher access to remote/dispersed areas difficult to reach personally or by telephone

No influence by interviewer on responses (maintains objectivity)

Maintains anonymity of respondents (particularly important for sensitive questions)

Respondents can take sufficient time to give thoughtful answers

Suitable for geographically spread populations

No need for trained enumerators

**Disadvantages:**

**Low response rate** (many questionnaires not returned)

Less opportunity to clarify ambiguous instructions

Possibility of misunderstanding questions

Cannot monitor understanding of questions

Reactions of respondents cannot be watched

Long response time

Questionnaires may be lost in mail

Incomplete questionnaires received

Respondents may not fill it carefully

Limited use for complex or open-ended questions

Cannot be used by illiterate population

**Best For**: Educated populations, literate respondents, questions needing anonymity, geographically dispersed areas.

TELEPHONE INTERVIEWS

**Definition**: Investigator asks questions over telephone; respondent answers orally.

**Advantages:**

**Relatively low cost** (cheaper than personal interviews, no travel required)

Can be conducted in shorter time than personal interviews

Relatively high response rate

Allows researcher to clarify ambiguous questions for respondents

Helps when respondents reluctant to answer in-person (maintains some distance)

Suitable for quick surveys

Can cover dispersed populations

**Disadvantages:**

**Limited access** (many people don't own telephones, especially in rural India)

Reactions of respondents cannot be watched

Possibility of influencing respondents through tone/voice

Cannot use complex visual questions

Limited for lengthy surveys

Connection/technical issues possible

Cannot verify respondent identity easily

Respondents may terminate call abruptly

**Best For**: Quick surveys, urban educated populations, when respondents have phone access, supplementary interviews.

COMPARISON TABLE OF THREE METHODS

| Feature | Personal Interview | Mail Survey | Telephone Interview |

|---------|-------------------|-------------|-------------------|

| Cost | High | Low | Medium |

| Response Rate | Highest | Lowest | Medium-High |

| Time | Long | Long | Short |

| Interviewer Influence | High | None | Medium |

| Reactions Observable | Yes | No | No |

| Geographic Reach | Limited | Excellent | Medium |

| Best for Open Questions | Yes | No | No |

| Suitable for Remote Areas | No | Yes | Limited |

---

PILOT SURVEY (PRE-TESTING)

**Definition**: A try-out of the questionnaire with a small sample group before conducting the actual survey.

**Purpose and Functions:**

**1. Testing Questionnaire Quality**

Identifies shortcomings and drawbacks in questions

Reveals ambiguous or confusing questions

Tests clarity of instructions

Assesses appropriateness of questions for target population

**2. Operational Assessment**

Tests performance of enumerators/data collectors

Identifies training needs for investigators

Assesses suitability of data collection methodology

Provides preliminary idea about survey feasibility

**3. Resource Planning**

Estimates cost involved in actual survey

Determines time required for field work

Helps in budgeting and scheduling

Identifies logistical challenges

**Benefits:**

Prevents wastage of resources on flawed questionnaire

Allows corrections before full-scale survey

Improves quality of final data

Ensures enumerators are adequately trained

Reduces response errors

**Process**:

Conducted on small sample representing target population

Results analysed for improvements

Questionnaire revised based on findings

Enumerators trained and evaluated

---

CENSUS AND SAMPLE SURVEYS

CENSUS (COMPLETE ENUMERATION)

**Definition**: A survey that includes every element/unit of the population.

**Also Known As**: Method of Complete Enumeration.

**Characteristics:**

Covers entire population without exception

Comprehensive data collection for all units

Conducted on fixed intervals (in India, every 10 years)

Time-specific snapshot of population

**Census in India:**

**Census of India - Historical Overview:**

Conducted every 10 years

**Census 1901**: Population = 23.83 crore

**Census 2011**: Population = 121.09 crore (last held census)

**Population growth in 110 years (1901-2011)**: Increased by more than 97 crore

**Population Growth Rates:**

**1971-81 decade**: Average annual growth rate = 2.2%

**1991-2001 decade**: Average annual growth rate = 1.97%

**2001-2011 decade**: Average annual growth rate = 1.64%

(Note: Declining growth rate shows demographic transition)

**Data Collected in Census:**

Demographic information: birth rates, death rates

Educational statistics: literacy levels, educational attainment

Employment data: occupational classification

Social indicators: life expectancy, household composition

Population size and composition

**Conducting Census:**

House-to-house enquiries

Covers all households in rural and urban India

Administered by Registrar General of India

Published through official census reports

**Advantages:**

Complete information about population

No sampling error

Highly accurate picture of entire society

Captures all variations and segments

**Disadvantages:**

**Extremely expensive** (requires thousands of enumerators)

**Time-consuming** (months to complete)

Complex to administer and supervise

Data processing takes considerable time

Cannot be conducted frequently

Creates information overload (excessive data)

POPULATION AND SAMPLE

**Population (Universe)**

**Definition**: The totality of all items/individuals under study; the entire group to which study results are intended to apply.

**Characteristics:**

Includes all units possessing specific characteristics

Defined according to research purpose

May be finite (countable) or infinite (uncountable)

Basis for selecting representative sample

**Example - Research Problem**: To study economic condition of agricultural labourers in Churachandpur district, Manipur

**Population**: All agricultural labourers in Churachandpur district

**Sample**

**Definition**: A group or section of population from which information is obtained; smaller than population but representative of it.

**Characteristics:**

Subset of population

Carefully selected to represent population characteristics

Enables data collection from manageable number of units

Should be representative (not biased)

**Example (continued)**: From above research problem

**Sample**: 10% of agricultural labourers in Churachandrup district

**Representative Sample**:

Possesses characteristics of population in similar proportions

Capable of providing reasonably accurate information about population

Selected through scientific sampling techniques

**Advantages of Sample Over Census:**

**Lower cost**: Requires fewer resources

**Shorter time**: Quicker to complete

**More detailed information**: Allows intensive enquiries

**Easier supervision**: Smaller team of enumerators

**Better training**: Easier to train and monitor smaller teams

**Practical feasibility**: Possible when complete enumeration impossible

---

SAMPLING TECHNIQUES: RANDOM VS NON-RANDOM

RANDOM SAMPLING

**Definition**: Sampling method where individual units from population are selected at random; every unit has equal chance of selection.

**Key Principle**: Each member of population has equal probability of being selected.

**Methods of Random Selection:**

**1. Lottery Method**

Write names/identifiers of all population units on paper

Mix papers thoroughly

Draw required number of units one by one

Each unit selected is included in sample

**Example**:

Population: 300 households in a locality

Sample size: 30 households

Process: Write names of all 300 households, mix them, randomly draw 30 names

**2. Random Number Tables**

Use published tables of random numbers

Assign numbers to all population units

Select units corresponding to random numbers from table

Ensures unbiased selection

**3. Computer-Based Selection**

Modern approach using computer programmes

Generate random samples automatically

Reduces human error in selection

**Characteristics:**

Every sampling unit (sampling frame) has equal opportunity for inclusion

Selection is unbiased and systematic

No investigator judgment involved

Produces representative samples

**Sampling Frame**: Complete list of all units in population from which sample is drawn

**Advantages:**

Unbiased selection

Results can be generalised to population

Statistically valid inferences possible

Eliminates investigator bias

**Disadvantages:**

Requires complete list of population units (sampling frame)

May be impractical for very large populations

Time-consuming if population dispersed

**Exit Polls - Application of Random Sampling:**

During elections, television networks use random sampling to predict election results:

Random sample of voters exiting polling booths are asked whom they voted for

Sample results used to predict election outcomes

**Limitation**: Exit polls not always accurate due to various factors (non-response bias, misreporting, sample not perfectly representative)

**Causes of inaccuracy**: Voters refusing to participate, memory lapses, false responses, sample bias

NON-RANDOM SAMPLING

**Definition**: Sampling method where investigator uses judgment, convenience, or predetermined criteria to select sample; not all units have equal chance of selection.

**Key Principle**: Investigator's judgment plays important role; convenience and bias may influence selection.

**Characteristics:**

All population units do NOT have equal chance of selection

Investigator bias influences sample selection

Selection based on judgment, purpose, convenience, or quota

Non-systematic approach

**Example**:

Population: 100 households in locality

Sample: 10 households

Selection process: Investigator selects conveniently situated households or households known personally

Result: Non-random, potentially biased sample

**Methods of Non-Random Sampling:**

**1. Convenience Sampling**

Select units easily or conveniently accessible

Households near investigator's residence

Examples: Mall intercepts, street interviews

**2. Judgment Sampling**

Investigator selects units based on personal judgment

Assumes certain units are more representative

Researcher's expertise guides selection

**3. Purposive Sampling**

Specific purpose guides selection

Select units serving particular purpose

Targeted selection approach

**4. Quota Sampling**

Population divided into subgroups (quotas)

Fixed number selected from each quota

Ensures representation from all groups

Example: Selecting specific number of men, women, youth from population

**Disadvantages:**

Results cannot be reliably generalised to population

High risk of bias in sample

Statistically invalid conclusions

Sampling error cannot be measured

Non-representative of population

**Advantages:**

Less expensive and time-consuming

Practical when sampling frame unavailable

Suitable for exploratory research

Easier to implement

---

SAMPLING AND NON-SAMPLING ERRORS

SAMPLING ERROR

**Definition**: Difference between sample estimate (calculated from sample) and corresponding population parameter (actual value of population characteristic).

**Nature of Sampling Error:**

Arises because sample represents only portion of population

Inevitable when using sampling method

Cannot be completely eliminated but can be minimised

Depends on sample size and variability in population

**Mathematical Expression:**

Sampling Error = Sample Estimate - Population Parameter

Example: If sample mean income = Rs 50,000 but actual population mean = Rs 52,000

Sampling Error = 50,000 - 52,000 = -Rs 2,000

**Causes:**

Random variation inherent in sampling process

Not all population units represented in sample

Difference in sample composition from population composition

**Reduction Methods:**

**Increase sample size**: Larger samples reduce error

**Use proper sampling technique**: Ensure randomness

**Stratified sampling**: Divide population into strata, sample from each

**Careful sample design**: Minimise variability

**Characteristics:**

Can be measured statistically

Decreases with larger sample size

Inversely related to sample size (formula: error ∝ 1/√n)

Predictable and measurable

NON-SAMPLING ERROR

**Definition**: Errors in data collection, processing, and analysis that are NOT due to sampling; occur in both census and sample surveys.

**Sources of Non-Sampling Error:**

**1. Data Collection Errors**

Ambiguous or unclear questions

Poorly trained enumerators/investigators

Misunderstanding by respondents

Respondent fatigue leading to careless answers

**2. Interviewer-Related Errors**

Interviewer bias or prejudice

Inappropriate tone or body language

Inconsistent question interpretation

Deliberate misreporting by investigator

**3. Respondent-Related Errors**

Deliberate false answers (social desirability bias)

Memory lapses and forgetting

Reluctance to answer sensitive questions

Misunderstanding of questions

Lack of cooperation

**4. Processing Errors**

Mistakes in data entry

Coding errors during categorisation

Calculation mistakes

Data transcription errors

Omission or duplication of records

**5. Survey Design Errors**

Poorly designed questionnaire

Inappropriate sampling frame

Inadequate instructions to enumerators

Faulty survey procedures

**6. Non-Response Errors**

Questionnaires not returned (mail surveys)

Refusal to participate in survey

Selected respondents not available

Results in incomplete data

**Characteristics:**

Cannot be completely eliminated

Occurs in both census and sample surveys (unlike sampling error)

Often larger and more serious than sampling error

Difficult to measure and quantify

More significant threat to data quality

**Reduction/Minimisation Methods:**

**Questionnaire design**: Careful, clear, unambiguous questions

**Enumerator training**: Comprehensive training on procedures

**Supervision**: Regular monitoring of field work

**Pilot testing**: Pre-testing questionnaire

**Clear instructions**: Written, detailed protocols

**Follow-up**: Contact non-respondents

**Data validation**: Check for consistency and accuracy

**Quality assurance**: Verification procedures

**Comparison: Sampling vs Non-Sampling Error**

| Aspect | Sampling Error | Non-Sampling Error |

|--------|---|---|

| **Definition** | Difference between sample estimate and population parameter | Errors in collection, processing, analysis (not due to sampling) |

| **Occurrence** | Only in sample surveys | In both census and sample surveys |

| **Measurement** | Can be measured statistically | Difficult to measure |

| **Relationship to Sample Size** | Decreases with larger sample | Unrelated to sample size |

| **Control** | Can be controlled by proper sampling | Difficult to control completely |

| **Source** | Inherent to sampling process | Errors in survey design and execution |

---

PRACTICAL APPLICATIONS AND EXAM-RELEVANT POINTS

**Key Takeaways for Board Examination:**

1. **Data Collection Purpose**: Provides evidence for sound economic decision-making and problem-solving

2. **Primary vs Secondary Data**: Primary (first-hand, specific) vs Secondary (processed, available from published sources)

3. **Questionnaire Design**: Must be clear, concise, unambiguous, logically sequenced, free from bias and leading questions

4. **Three Modes of Collection**: Personal interviews (highest response, most expensive), Mail surveys (cheapest, lowest response), Telephone interviews (medium cost and response)

5. **Census**: Complete enumeration every 10 years in India, providing comprehensive population data

6. **Sampling**: Representative sample more practical than census due to cost, time, and intensive inquiry advantages

7. **Random Sampling**: Scientific selection ensuring equal probability for all units

8. **Non-Random Sampling**: Uses investigator judgment; less reliable but practical

9. **Sampling Error**: Measurable difference between sample and population; decreases with larger sample size

10. **Non-Sampling Error**: More serious; occurs in all surveys; requires careful design and execution to minimise

**Indian Context Relevance:**

Census 2011 provided critical data on India's 121.09 crore population

Declining population growth rates (2.2% to 1.64%) reflect demographic transition

Sample surveys widely used in India for economic surveys (household surveys, agricultural surveys)

NSSO (National Sample Survey Office) conducts regular sample surveys for employment, consumption, economic data

Census data forms basis for allocation of Lok Sabha seats, resources, and policy planning