Sampling Designs – Probability Sampling and Non Probability Design

Meaning of Sampling

• Sampling = The process of selecting a part (sample) from a population to study and generalize the findings to the whole population.
• Sample = A subset of population.
• Population (Universe) = Entire group of individuals/units under study.

Meaning of Probability Sampling

• Probability sampling = a sampling technique in which every unit of the population has a known, non-zero, and equal chance of being selected.
• Based on random selection → ensures representativeness and reduces bias.

Kerlinger (1986): Probability sampling gives each element of the population a known probability of being included in the sample.

1) Types of Probability Sampling Designs

(i) Simple Random Sampling (SRS)

• Every element has an equal chance of being selected.
• Selection can be: Lottery method (manual draw) or Random number tables/computer. Example: Selecting 100 farmers randomly from a list of 1000.
• Advantage: Simple, unbiased.
• Limitation: Needs complete list of population (sampling frame).

(ii) Systematic Sampling

• Select every kᵗʰ unit from a list after a random start. Sampling interval (k) = N/n, where N = population size, n = sample size. Example: In a list of 1000 farmers, if sample = 100 → pick every 10th farmer after random start.
• Easier than SRS.
• If list has hidden pattern, bias may occur.

(iii) Stratified Random Sampling

• Population divided into strata (homogeneous groups) based on characteristics (e.g., gender, farm size, education). Random samples drawn from each stratum → proportionate or disproportionate. Example: Farmers divided into small, medium, large → random sample from each group.
• Advantage: Ensures representation of all subgroups.
• Limitation: Complex if many strata.

(iv) Cluster Sampling

• Population divided into clusters (naturally occurring groups).
• Random selection of clusters → all elements within selected clusters are studied.
• Example: Villages taken as clusters; few villages selected randomly; all farmers in those villages studied.
• Advantage: Saves cost and time.
• Limitation: Less precise compared to SRS.

(v) Multi-Stage Sampling

• Sampling carried out in stages (combination of methods). Example: District → Block → Villages → Farmers. Common in large-scale surveys.
• Advantage: Useful for large, scattered populations.
• Limitation: Increases sampling error at each stage.

(vi) Multi-Phase Sampling

• First stage: collect broad/general information.
• Second stage: collect detailed information from sub-sample.
• Example: First list of adopters/non-adopters of a technology, then detailed interviews with adopters.

Key Advantages of Probability Sampling

• Reduces researcher bias.
• Allows use of statistical tests.
• Ensures representativeness of population.
• Provides basis for generalization.

2) Non-Probability Sampling

Meaning

• In non-probability sampling, not every unit of the population has a known or equal chance of being selected.
• Selection depends on the judgment of the researcher, ease of access, or voluntary participation.
• Less representative compared to probability sampling, but widely used in exploratory, qualitative, or pilot studies.

Types of Non-Probability Sampling

(i) Convenience Sampling (Accidental Sampling)

• Sample taken from units that are easily available to the researcher.
• Example: Interviewing farmers who are present in a village meeting.
• Advantage: Quick, inexpensive.
• Limitation: High bias, low generalizability.

(ii) Purposive Sampling (Judgmental Sampling)

• Researcher selects sample deliberately based on specific characteristics relevant to study.
• Example: Selecting progressive farmers for studying technology adoption.
• Advantage: Useful when specific cases are needed.
• Limitation: Subjective, researcher bias possible.

(iii) Quota Sampling

• Population divided into subgroups (like in stratified sampling).
• Researcher ensures a predetermined quota from each group, but selection within quota is non-random. Example: 40% male, 60% female farmers included, but chosen by convenience.
• Advantage: Ensures representation of groups.
• Limitation: Non-random → bias possible.

(iv) Snowball Sampling (Network / Chain Referral Sampling)

• Existing subjects refer/recruit further participants.
• Example: Studying women SHGs – one member identifies others. Useful for studying hidden or hard-to-reach populations (drug users, SHGs, migrants).
• Limitation: Sample may not represent entire population (network bias).

(v) Volunteer Sampling (Self-Selection)

• Individuals voluntarily participate in research.
• Example: Farmers responding to a call for participation in a survey.
• Advantage:  Easy and low-cost.
• Limitation; Attracts only interested individuals → not representative.

(vi) Judgment Sampling (Expert Sampling)

• Researcher uses experts’ opinion to identify representative samples.
• Example: Selecting model villages based on experts’ recommendations.

Advantages

• Economical and quick.
• Useful when population frame is not available.
• Good for exploratory research, qualitative studies, pilot testing.

Limitations

• High sampling bias.
• Results cannot be generalized to population.
• No basis for statistical inference.