Resumen

MONTESINOS LOPEZ, Osval Antonio et al. Sample size determination for estimating animal prevalence that assure narrow confidence intervals. Rev. mex. de cienc. pecuarias [online]. 2011, vol.2, n.2, pp.229-244. ISSN 2448-6698.

Sample size calculation plays an important role in the design of optimal veterinary and agricultural experiments for estimate population proportions including disease prevalence. This research proposes a method with three steps to determine the exact sample size for a binomial distribution that ensures precision in the estimated proportion and numerically displays the degree of underestimation that occurs using the traditional formula (normal approximation) for calculating the sample size. Step 1 obtain a sample size that guarantees that the relative mean width of the CI (ϖ_r) is narrower than the desired width (re); the step 2 iteratively increase the sample size until ϖ_r is smaller than the desired width (re) with a specified degree of certainty (γ) and step 3 get the required number of clusters. Simulated data were created and tables presented showing possible scenarios. An R program is given and explained that will reproduce the results in an easy way.

Palabras llave : Sample size; Binomial distribution; Narrow confidence intervals; CIRW.