Authors: SEGURA-CORREA, J.C.; SANTOS-RICALDE, R.H.;
MAGANA-MONFORTE, J.G.; AKE-LOPEZ, R |
Abstract: The objectives of this study were to propose a simple approach to calculate the sample size to
build growth curves and to show the effect of sample size on curve fit and parameter estimates.
Sample size was calculated based on the body weight, of turkeys, with the largest coefficient of
variation, from hatch to 23 weeks of age. Also, a simulation study was carried out to determine
the effect of sample size on growth curve fit and parameters precision. Curves were built by age
for data of 15, 30, 50, 100, 200 and 5000 male turkeys, randomly generated using the means and
standard deviations of the observed data, and by means of the von Bertalanffy model. A table to
estimate sample size for a continuous variable is provided. As expected, sample size was a
function of the CV. Sample size affected the fit of the curve and precision of parameter
estimates. In conclusion, the approach here used can be applied to build growth curves using the
formula to estimate the sample size for body weight. Large samples provide better fit of the
curve and are more precise. |