Tamaño de muestra requerido para estimar la media aritmética de una distribución lognormal

We present close formulae to calculate the required sample size to estimate the arithmetic mean of a lognormal distribution for censored and non-censored data. These formulae were obtained by adjusting non linear models for the exact sample sizes estimates reported by Perez (1995). The formulae presented are functions of the estimated geometric standard deviation, the proportional precision from the true arithmetic mean and using confidence levels of 90%, 95% and 99%. These new close formulae correct the underestimation problem in other formulae presented in the statistical literature.