Poder estatístico, tamanho amostral e variação em estudos com ácidos graxos na carne ovina

Abstract

There is a growing increase of sheep research to improve the quality of sheep meat, and, through different search diets to improve the functional fatty acids profile, with improved properties in consumer health. Currently there is a pressure from the Committee of ethics in research with animals to reduce the sample size and the number of repetitions, but this has been done empirically and without worrying about the credibility of the results to be obtained. Scientific studies related to the improvement of the quality of meat sheep need to have credibility and trust from results investigated. A way to predict the quality of the results of an experiment is the use of estimation of power of statistical tests. Another tool that researchers have employed to assess and measure the quality of their experiments is the use and classification of the coefficient of variation (CV). Therefore, the objective of estimating the statistical power, sample size and predict the sort of variation coefficient in analysis of variance under different experimental designs, with the profile of fatty acids in sheep meat. Collected articles published in the years 2015 and 2016, Web of Science databases, Scielo, Scopus and Science Direct. Collected a total of 30 articles in the respective years. To evaluate the test power of the variables under study, as low power test <80% and high power of ≥ 80% test. For the sample medium size were considered only those comments with size of effect (f) estimated in the range of 0.35 to 3.12. To the power of these test showed an average of 77% for low power test and 23% for high power test. For the sampling size is suggested between 8 to 11 repetitions per treatment. For CV ratings notes that the same differ, and these presented low precision of CV and greater prevalence and high precision of CV with lower prevalence. It is concluded that it is necessary to the planning of experiments before the conduct of research for credible results. With the increase in the number of repetitions per treatment prevents publications with empirical results.