The problem of normalization of data is frequently discussed in the literature but few studies have been made to determine the actual effect produced when normalized and non-normalized data are compared. In this study two types of distributions were studied in which correlations are made between certain data with and without correction for skewness and kurtosis. The correlations, plots and curves of the material, and the changes created will be presented and discussed. It would seem that in most cases the benefits of normalizing data are hardly worth while since curvilinearity and breaks in data induce sources of error which compensate for any benefits derived.
Proceedings of the Iowa Academy of Science
© Copyright 1952 by the Iowa Academy of Science, Inc.
Collins, C. C. and Lauer, A. R.
"An Empirical Study of the Effect of Normalizing Data [Abstract],"
Proceedings of the Iowa Academy of Science: Vol. 59:
, Article 47.
Available at: https://scholarworks.uni.edu/pias/vol59/iss1/47