Q.Define sampling error and non-sampling error. Give one example of each, and state one method commonly used to reduce each type of error.

Every survey carries some error, but the two kinds — sampling and non-sampling — arise from different causes and need different fixes.

Sampling Error is the error that arises solely because a sample, rather than the entire population, is studied — it is the difference between a sample statistic (such as $\bar{x}$) and the true population parameter (such as $\mu$) that a full census would have revealed. Example: if the true average monthly sales of 400 outlets is $\mu = ₹50{,}000$, but a sample of 40 outlets gives $\bar{x} = ₹47{,}000$, the sampling error is $|47{,}000 - 50{,}000| = ₹3{,}000$. It exists only in the sample method (a census has none), and it can be reduced by increasing the sample size or by using a more efficient design such as stratified sampling. …

Assuming non-sampling error occurs only in sample surveys — it occurs in a census too, and can in fact be larger there, simply because …