UGC NET Questions (Paper – 1)

Randomisation is the process of assigning participants to experimental and control groups using random mechanisms, such as random number tables or computer-generated sequences. This practice helps ensure that groups are equivalent on both known and unknown variables at the start of the experiment. By reducing systematic differences between groups, randomisation strengthens internal validity and makes it more likely that observed effects are due to the treatment. Hence, the procedure described in the stem is correctly called randomisation.

Option A:
Stratification involves dividing a population into homogeneous subgroups and can be used before sampling or randomisation, but it does not by itself eliminate selection bias in group assignment. It is more related to sampling than to within-experiment group allocation, so it is not the best answer.

Option B:
Matching pairs participants on certain characteristics such as age or ability before assigning them to groups, which can improve equivalence on those specific variables, but it is not purely random. It cannot control for all unmeasured variables, so it differs from the chance-based process emphasised in the stem.

Option C:
Option C, randomisation, uses chance to distribute both measured and unmeasured characteristics across groups, reducing the risk that pre-existing differences, rather than the treatment, explain outcomes. This aligns exactly with the purpose stated in the question.

Option D:
Standardisation refers to keeping procedures, instructions and conditions uniform across participants to reduce procedural variation. While important for internal validity, it does not address selection bias in the same way randomisation does.

Non-probability sampling refers to techniques in which elements of the population are selected using non-random procedures such as convenience, purposive or quota sampling. In these methods, the exact probability of any given element being included in the sample is unknown. This limits the ability to generalise findings using statistical inference. Because the stem explicitly states that selection probabilities are unknown, non-probability sampling is the correct answer.

Option A:
Probability sampling, by contrast, ensures that each element of the population has a known, non-zero chance of selection, often through random processes. This property enables estimation of sampling error and supports strong generalisation to the population. Since the question mentions unknown probabilities, probability sampling cannot complete the stem correctly.

Option B:
Non-probability sampling is often used in exploratory or qualitative studies, or when constructing sampling frames is difficult or impossible. Although it may introduce bias, it can be practical and informative in certain contexts. These characteristics fit the description provided in the stem, confirming non-probability as the appropriate term.

Option C:
Systematic sampling, a form of probability sampling, selects every kth element from an ordered list after a random start, so individual selection probabilities are known. Thus, systematic sampling does not align with the condition of unknown probabilities.

Option D:
Stratified sampling, another probability method, divides the population into homogeneous groups and draws random samples from each stratum. Here again the chance of selection is known and can be calculated, which contradicts the stem. Therefore, stratified sampling is not the right completion.

Random sampling gives each unit in the population a known, often equal, chance of being selected, which makes the selection process impartial. This greatly reduces systematic bias in the composition of the sample because personal judgment and convenience are minimised. While it does not completely eliminate all error, it helps ensure that differences between the sample and population are mainly due to chance. Therefore, the correct effect on selection bias is that random sampling tends to reduce it.

Option A:
If random sampling were to increase bias, it would defeat its main purpose in scientific research. The entire rationale for using random methods is to minimise systematic distortions in the sample. Thus, the option suggesting an increase in bias is conceptually incorrect.

Option B:
Random sampling reduces the likelihood that any particular type of individual is overrepresented or underrepresented for systematic reasons. It supports the use of probability theory to estimate sampling error and build confidence intervals. These features correspond directly to the idea of reducing selection bias as stated in the question.

Option C:
Creating bias would mean deliberately skewing the sample in favour of certain subgroups, which is characteristic of convenience or judgement sampling, not random sampling. Therefore, the suggestion that random procedures create bias is misleading.

Option D:
Ignoring bias does nothing to protect the quality of inferences and would be a sign of poor methodological practice. Random sampling is chosen precisely because it addresses concerns about bias rather than ignoring them, so this option does not fit the role described in the stem.