Comparação de grupos “versus" comparações emparelhadas
Keywords:
comparação emparelhada, diferença significativa, teste t, variância combinada, coeficiente correlaçãoAbstract
In this paper the author considered group and paired comparisons and has shown by theoretical considerations when one method is preferable to the other. The subject was put in quantitative reasoning and it was proved that when the data of the two groups are positively correlated, pairing is better. Also it was considered the number of degrees of freedom lost in pairing. Finally, with a practical example, it was shown the advantage of pairing in relation to group comparison. This last method was completely fallacious when applied to the given example.
References
(1) Cochran, W. G. and Cox, G. M. (1950) Experimental Designs, p. 29. John Wiley & Sons, Inc., New York
(2) Fisher, R.A. (1946) Statistical Methods for Research Workers, pp. 124-125, 10th. edition. Oliver and Boyd. London
(3) Fisher, R. A. (1947) The Design of Experiments, pp. 234-236, 4th. edition. Hafner-Publishing Co., Inc., New York
(4) Love, H. H. (1943) Experimental Methods in Agricultural Research, pp. 13-16. University of Puerto Rico, Rio Piedras
(5) Snedecor, G. W. (1946) Statistical Methods, p. 84, 4th. edition. Iowa State College Press, Ames
