Pesquisa Agropecuária Brasileira
Print version ISSN 0100-204X
CARGNELUTTI FILHO, Alberto et al. Sample size for estimating the Pearson correlation coefficient among corn characters. Pesq. agropec. bras. [online]. 2010, vol.45, n.12, pp. 1363-1371. ISSN 0100-204X. http://dx.doi.org/10.1590/S0100-204X2010001200005.
The objective of this work was to determine the sample size for estimating the Pearson correlation coefficient among characters from three corn hybrids. For the analyses, 361, 373 and 416 plants were taken randomly from the hybrids single, triple and double, respectively. For each plant, the following characters were measured: major and minor stem diameters, plant and, heights, weight, length and diameter of ear, number of rows per ear, weight and diameter of cob, a hundred-grain weight, number of grain per ear, grain length and grain yield. For each of the 91 pairs of characters and hybrids, the sample size, based on bootstrap with replacement of 1,000 samples, of each simulated sample size, was determined. In the estimation of Pearson correlation coefficient with the same precision, the sample size (number of plants) increases in the direction of character pairs with lower intensity of the linear relationship, independently of the hybrid cross type. For the 91 pairs of characters studied, 252 plants are sufficient to estimate the Pearson correlation coefficient, in the bootstrap confidence interval 95%, equal to 0.30
Keywords : Zea mays; experimental design; bootstrap resampling; linear relationship.