Agricultural Extension Education- ICAR JRF/SRF 2024

• Measures of Association for Nominal Contingency Tables:
• In a two-way contingency table involving two purely nominal variables:
• Cramer's V: A symmetric measure of association based on the Pearson Chi-Square statistic, ranging from 0 (no association) to 1 (perfect association) for tables larger than $2 \times 2$.
• Lambda ($\lambda$, Goodman-Kruskal Lambda): A proportional reduction in error (PRE) measure that assesses asymmetric or symmetric association between nominal variables.
• Gamma ($\gamma$): Used exclusively for ordinal variables.
• Conclusion: Both (1) Cramer's V and (2) Lambda can be used, making Option 4 correct.

Correct option: (4) Both (1) and (2)

• Methods for Assessing Scale Validity:
• Validity Techniques:
• Jury Opinion (Expert Panel): Establishes Content and Face validity by having a panel of expert judges evaluate item relevance and comprehensiveness.
• Known Groups Method: Establishes Criterion and Construct validity by testing whether the scale discriminates significantly between groups known to hold opposing attitudes.
• Goodenough Technique: A scalogram analysis scoring technique developed by Ward H. Goodenough (1944) used to evaluate the reproducibility and scalability of items in Guttman scaling, rather than assessing external scale validity.

Correct option: (3) Goodenough technique

• Controlling Extraneous Variables in Experimental Research:
• In experimental design, extraneous confounding variables that threaten internal validity can be controlled through:

• Eliminating the variable as a variable: Holding the variable constant across all conditions (e.g., selecting only one gender or one soil type).
• Matching subjects: Pairing subjects with identical extraneous characteristics across experimental and control groups.
• Randomization: The most powerful tool for distributing unknown variables equally.
  • Conclusion: Both (1) and (2) are established experimental control approaches, making Option 4 correct.

• Systematic Sampling:
• Systematic Sampling is a probability sampling method where sample units are selected from a comprehensive ordered sampling frame at regular fixed intervals:

• random starting point between $1$ and $k$ is chosen, and every $k$-th case thereafter is systematically drawn until the desired sample size $n$ is reached.
• Distractors: Stratified sampling divides the population into strata; quota sampling is non-random; convenience sampling selects easily accessible units.

Correct option: (2) Systematic sampling