UGC NET Questions (Paper – 1)

Non-sampling errors are all sources of survey inaccuracy that do not stem from the act of sampling but from the way data are collected, recorded or processed. They include interviewer bias, faulty questionnaire design, respondent misunderstanding and coding or tabulation mistakes. Such errors can be large and may not decrease simply by increasing sample size. Therefore, errors arising from inaccurate responses or defective tools are correctly termed non-sampling errors.

Option A:
Non-sampling errors encompass diverse problems such as ambiguous questions that confuse respondents, poorly trained interviewers and data entry errors. These issues affect the quality of information gathered from each selected unit and can seriously distort findings. Since the stem emphasises inaccurate responses and processing mistakes, this option is appropriate.

Option B:
Random error refers to unpredictable fluctuations that occur by chance and may be part of both sampling and measurement processes; it is not restricted to the specific systematic problems listed.

Option C:
Proportional error is not a standard category in survey methodology and does not describe the cluster of issues like tool defects and processing mistakes.

Option D:
Experimental error is associated with variability in experimental outcomes and does not specifically refer to survey response or instrument problems, so it is not correct here.

Sampling error is the random difference between values obtained from a sample and the true values in the population. It occurs because the sample, even if selected properly, represents only a subset of the population. As a result, estimates like means or proportions fluctuate from sample to sample. Larger and more representative samples reduce sampling error but cannot eliminate it entirely.

Option A:
Option A describes errors arising from mistakes in coding or tabulation, which are classified as non sampling errors. These can occur even in a census where the whole population is measured.

Option B:
Option B correctly explains sampling error as the natural variation due to observing only a part of the population. It captures the essence of random fluctuation inherent in the sampling process and is consistent with statistical theory.

Option C:
Option C refers to deliberate response distortion or bias, which again falls under non sampling errors, specifically response bias. While serious, it is conceptually distinct from sampling error.

Option D:
Option D deals with faulty instruments, another form of non sampling error because it affects measurement irrespective of whether a sample or census is used.

Statements A, B, C and D distinguish sampling error from non-sampling error, and F highlights a way to reduce some non-sampling problems. A is true because sampling error results from studying only a subset, while B correctly notes that poor instruments create non-sampling error. C is accurate in classifying non-response bias as a non-sampling error, and D correctly observes that larger samples usually reduce sampling error. F is also true since properly trained field staff can lower interviewer and recording errors, whereas E is false because random sampling addresses sampling error, not all forms of non-sampling error.

Option A:
Option A ignores statement F, which draws attention to practical steps for reducing non-sampling error through training. While A, B, C and D are correct, the omission of F makes the description less complete in terms of remedies. Thus, this option is not fully correct.

Option B:
Option B adds statement E, falsely claiming that non-sampling errors can be entirely removed by random sampling. Random sampling influences only sampling error; measurement, non-response and recording errors can still occur. Because it includes a false statement, this option is invalid.

Option C:
Option C is correct because it lists A, B, C, D and F, capturing both definitions and some methods for error reduction. It properly excludes E, which overstates the power of random sampling. Therefore, this combination contains all and only the true statements.

Option D:
Option D omits statement A and includes E, thereby losing the basic definition of sampling error and adding an incorrect assertion. This combination confuses the distinction between error types and cannot be accepted.

Statements A, B and C correctly describe population, sample, representativeness and sampling error, whereas D and E are wrong. Using a sample instead of a census does not automatically remove non-sampling error such as measurement or non-response error. Similarly, simply increasing sample size cannot eliminate bias caused by an incomplete or faulty sampling frame. Therefore, the wrong statements are D and E together.

Option A:
Option A identifies only D as wrong and overlooks E. Although D is indeed incorrect, E also contains a misconception by claiming that large samples remove frame-induced bias. Because it fails to include all wrong statements, this option is incomplete.

Option B:
Option B isolates only E as wrong and ignores D. While E is clearly incorrect, D also misleads by implying that sampling automatically eliminates non-sampling error. Hence, Option B does not capture the full set of wrong statements.

Option C:
Option C is correct because it groups D and E, the two statements that misrepresent what sampling can achieve. It recognises that non-sampling errors and frame bias require other remedies such as better instruments, careful procedures and accurate frames.

Option D:
Option D incorrectly includes C among the wrong statements. Statement C is actually a standard definition of sampling error as the difference between a statistic and the corresponding parameter. Misclassifying this correct definition makes the combination invalid.

Option E wrongly presents A, which correctly defines a sample, as a wrong statement alongside D and E. By treating a basic concept as incorrect, this option cannot be accepted.