Sampling Theory Questions for SSC CGL

Population vs Sample

A population is the entire group you want to study (e.g., all voters in India). A sample is a smaller subset chosen from the population to estimate population properties. In exams, you will be asked to identify which is which, or to understand why sampling is done. Sampling is done because studying the entire population is expensive, time-consuming, and sometimes impossible (e.g., destructive testing). The goal is to make the sample representative so that sample statistics (like mean, variance) reliably estimate population parameters.

• Population = entire group; Sample = subset chosen for study
• Population parameter = true value (unknown); Sample statistic = estimated value (calculated from sample)
• Representativeness is key: sample must reflect population diversity
• Sampling error = difference between sample statistic and population parameter (always exists)
• Sampling is cost-effective and practical for large populations

Types of Sampling Methods

SSC CGL tests your ability to recognize and apply different sampling techniques. Each method has a specific use case and assumption. Random sampling is the gold standard because every unit has equal chance of selection. Stratified sampling divides population into homogeneous groups (strata) and samples from each. Systematic sampling picks every kth unit. Cluster sampling divides population into clusters and randomly selects entire clusters. Convenience sampling (non-random) is biased and rarely correct in exams.

• Simple Random Sampling: every unit has equal probability; unbiased; used as baseline
• Stratified Sampling: population divided into strata; samples from each stratum; reduces variance for heterogeneous populations
• Systematic Sampling: select every kth unit after random start; practical and easy to implement
• Cluster Sampling: divide into clusters; randomly select clusters; useful when population is geographically scattered
• Quota Sampling: non-random; interviewer selects units to meet quotas; prone to bias
• Convenience Sampling: easiest but most biased; avoid in exams unless explicitly asked

Sampling Distribution and Standard Error

A sampling distribution is the probability distribution of a sample statistic (e.g., sample mean) calculated from all possible samples of a given size. The standard error measures how much sample statistics vary from sample to sample. If you take many samples and calculate their means, those means form a distribution. The standard deviation of that distribution is the standard error. Smaller standard error means sample statistic is more stable and closer to the true population parameter. This is tested heavily in SSC CGL.

• Sampling distribution = distribution of sample statistic across all possible samples
• Standard error = standard deviation of the sampling distribution
• Larger sample size leads to smaller standard error (more precision)
• Central Limit Theorem: sampling distribution of mean is approximately normal for large n, regardless of population shape
• Standard error depends on population variance and sample size, not population size

Confidence Intervals

A confidence interval is a range of values that likely contains the true population parameter. For example, a 95% confidence interval for the mean means if you repeated sampling many times, 95% of the intervals you construct would contain the true population mean. SSC CGL asks you to calculate or interpret confidence intervals. The interval is constructed as: sample statistic ± (critical value × standard error). The critical value depends on the confidence level (95% uses 1.96 for normal distribution, 99% uses 2.576).

• Confidence interval = point estimate ± margin of error
• Margin of error = critical value × standard error
• 95% confidence level is most common in exams
• Wider interval = lower precision but higher confidence; narrower interval = higher precision but lower confidence
• Confidence level is NOT the probability that the true parameter lies in this specific interval (it either does or doesn't); it's the long-run success rate of the method

Sample Size Determination

Determining the right sample size is crucial. Too small a sample gives unreliable estimates; too large wastes resources. Sample size depends on desired precision (margin of error), confidence level, and population variability. SSC CGL occasionally asks how sample size affects standard error or what sample size is needed to achieve a certain margin of error. The relationship is inverse: larger sample size reduces standard error and margin of error.

• Larger sample size reduces standard error and margin of error
• Sample size needed increases with higher confidence level and greater population variance
• Sample size needed decreases with larger acceptable margin of error
• For proportions, maximum variance occurs at p = 0.5, so use p = 0.5 if true proportion is unknown
• Doubling sample size reduces standard error by factor of 1/sqrt(2) ≈ 0.707

• Confusing population parameter with sample statistic. Remember: parameters are unknown (Greek letters like mu, sigma); statistics are calculated from sample (x-bar, s).
• Thinking standard error is the same as standard deviation. Standard error is the std dev of the sampling distribution, not the spread of individual data points.
• Misinterpreting confidence level. A 95% CI does NOT mean there is 95% probability the true parameter is in this interval. It means 95% of all such intervals (from repeated sampling) would contain the true parameter.
• Forgetting that standard error decreases with sqrt(n), not n. Doubling sample size does not halve standard error; it reduces it by factor of 1/sqrt(2).
• Choosing wrong sampling method. Stratified is best for heterogeneous populations; cluster for geographically scattered; systematic for ordered lists. Convenience sampling is almost always wrong in exams.

• PSSCC: Population, Sample, Stratified, Cluster, Convenience — remember sampling methods in order of bias (first two best, last worst).
• SE shrinks with sqrt(n): Bigger sample = smaller standard error. Think of it as 'more data, more confident, less error'.
• 95% CI uses 1.96, 99% uses 2.576: Memorize these z-values. 1.96 ≈ 2 for quick mental math.
• Margin of Error = z × SE: The wider the interval you want, the more confident you are. Tighter interval = less confident.

• SSC CGL Tier-2 typically has 1-2 questions on sampling. They focus on: identifying sampling method, calculating standard error, constructing confidence intervals, or determining sample size.
• Recent papers favor questions on standard error and confidence intervals over sampling methods. Practice formula-based problems.
• Watch for trick questions: 'Which sampling method is best?' — answer depends on context (heterogeneous population → stratified; scattered population → cluster).
• Time management: sampling questions are usually quick (2-3 min) if you know formulas. Memorize z-values (1.96, 2.576) and the three SE formulas.
• Common SSC pattern: 'A sample of n=100 has mean 50 and std dev 10. Find 95% CI.' This is a direct formula application. Practice until you can do it in under 90 seconds.