Sample size for estimation of direct effects in path analysis of corn

Area of Science:

• Agricultural Science
• Genetics
• Biometrics

Background:

• Estimating direct effects of variables on corn grain yield is crucial for agricultural research.
• Previous studies often require large sample sizes, impacting efficiency and cost.
• Understanding optimal sample size is key for robust statistical analysis in crop breeding.

Purpose of the Study:

• To determine the minimum sample size needed for accurately estimating direct effects on corn grain yield.
• To compare sample size requirements across different corn hybrids, harvest seasons, and analytical scenarios.
• To evaluate the efficiency of traditional versus ridge path analyses for yield component studies.

Main Methods:

• Evaluated 6340 corn plants across two harvests and three hybrid types.
• Measured eleven yield-related variables, including plant height, ear characteristics, and grain yield.
• Employed resampling techniques with replacement at four accuracy levels (95% confidence interval ranges) across nine explanatory variable scenarios and two path analysis types.

Main Results:

• The ninth scenario, excluding most explanatory variables, significantly reduced required sample size.
• Estimating direct effects with a maximum 95% confidence interval of 0.25 requires 10 to 530 plants, varying by experimental conditions.
• The ninth scenario, with a maximum 95% confidence interval of 0.25, requires only 120 plants, irrespective of hybrid, harvest, or analysis type.

Conclusions:

• The ninth path analysis scenario is recommended for efficient sample size determination in corn yield studies.
• Optimized sample size estimation enhances the practicality and cost-effectiveness of agricultural research.
• Accurate direct effect estimation is achievable with a reduced sample size under specific analytical conditions.