1. Primary and Secondary Data
In statistics, data refers to the quantitative or qualitative information collected for analysis. Depending on the source of collection, data is classified into two primary categories:
Primary Data: Data collected by the investigator personally for the first time for a specific, original purpose. It is first-hand information.
Secondary Data: Data that has already been collected, compiled, and processed by some other agency or individual for another purpose, which is now being reused.
Comparison between Primary and Secondary Data
| Basis of Comparison | Primary Data | Secondary Data |
|---|---|---|
| Originality | Highly original as it is collected from the source. | Not original; it is second-hand information. |
| Cost and Time | Requires significant time, effort, and money. | Cheaper and faster to obtain. |
| Suitability | Specifically tailored to the research objectives. | May or may not match the specific research needs. |
| Source | Surveys, experiments, direct observations, interviews. | Books, government reports, journals, websites. |
2. Methods of Collecting Primary Data
While there are many methods of collecting primary data, questionnaires and schedules are two of the most widely used tools.
Questionnaire Method
A questionnaire is a systematic list of questions sent directly to the respondents, usually via mail or digital platforms. The respondents read, interpret, and write the answers themselves.
- Advantages: Economical, covers a wide geographical area, preserves respondent anonymity.
- Disadvantages: Low response rate, cannot be used for illiterate respondents, risk of misinterpretation.
Schedule Method (Question Schedule)
A schedule is also a set of questions, but it is filled out by a trained investigator or enumerator who physically meets the respondent. The enumerator asks the questions face-to-face and records the replies on the sheet.
- Advantages: High response rate, can gather info from illiterate respondents, enumerators can clarify doubts.
- Disadvantages: Very expensive, time-consuming, requires training enumerators, potential for enumerator bias.
Comparison: Questionnaire vs. Schedule
| Basis | Questionnaire | Schedule |
|---|---|---|
| Filled by | Filled by the respondent themselves. | Filled by the enumerator or investigator. |
| Cost | Relatively low cost. | High cost due to hiring and training enumerators. |
| Coverage | Can reach wide and remote areas easily. | Limited to areas where enumerators can physically travel. |
| Non-Response | High non-response rate. | Very low non-response rate. |
3. Census vs. Sample Survey
When conducting research, data can be collected from the entire population or a subset of it.
Census Method: A systematic method where data is collected from each and every unit of the entire population. Example: National Population Census.
Sample Method: A method where data is collected only from a representative subset (sample) of the population, and inferences are drawn about the whole population.
Comparison Table: Census vs. Sample
| Feature | Census Method | Sample Method |
|---|---|---|
| Coverage | Complete enumeration of all units. | Partial coverage (selected units only). |
| Accuracy | Highly accurate and detailed. | Subject to sampling errors, but reliable if sample is representative. |
| Resource Requirements | Requires massive finance, manpower, and time. | Economical in terms of money, labor, and time. |
| Suitability | Best when the population size is small and absolute accuracy is needed. | Best when the population size is infinite or large. |
4. Random and Non-Random Sampling
Sampling designs are broadly categorized into two types depending on whether selection is based on probability theory or personal judgment.
Random (Probability) Sampling
A sampling technique where every element of the population has a known, non-zero probability of being selected. It eliminates personal bias from the selection process.
- Key Feature: Random selection process (e.g., lottery method, random number generators).
- Examples: Simple Random Sampling, Stratified Sampling, Systematic Sampling.
Non-Random (Non-Probability) Sampling
A sampling technique where the selection of units depends on the subjective judgment or convenience of the researcher, rather than mathematical chance.
- Key Feature: Non-random selection; some elements may have zero chance of selection.
- Examples: Quota Sampling, Purposive/Judgmental Sampling, Convenience Sampling.
5. Advanced Sampling Methods
Quota Sampling
Quota sampling is a non-random sampling method where the population is divided into mutually exclusive sub-groups (similar to stratified sampling), and then the researcher selects a specified proportion of units (quotas) from each sub-group based on convenience or judgment.
Methods/Steps of Quota Sampling:
- Step 1: Classification: Divide the population into relevant subgroups based on specific characteristics (e.g., age, gender, income level).
- Step 2: Assign Quotas: Determine the proportion of each subgroup in the target population and set quotas accordingly (e.g., 60% male, 40% female).
- Step 3: Sample Selection: The researcher selects individuals to fill these quotas using non-random methods (convenience or purposive selection) until the target numbers are met.
Stratified Sampling
Stratified sampling is a random sampling method where the heterogeneous population is divided into homogeneous, non-overlapping groups called strata. A simple random sample is then drawn from each stratum.
Key Principles:
- The variation within each stratum is minimized (homogeneity within strata).
- The variation between different strata is maximized (heterogeneity between strata).
- Example: If a school has students from classes 1 to 10, each class acts as a stratum. A random sample of 5 students is chosen from each class.
Multistage Sampling
Multistage sampling is a complex form of cluster sampling where the sample is selected in multiple sequential stages, moving from larger geographic or administrative units to smaller units.
Example of Multistage Process:
- Stage 1 (Primary Sampling Units): Select a random sample of States in a country.
- Stage 2 (Secondary Sampling Units): Select a random sample of Districts from the chosen States.
- Stage 3 (Ultimate Sampling Units): Select a random sample of Households from the chosen Districts to conduct interviews.