This article is only available in the PDF format. Download the PDF to view the article, as well as its associated figures and tables.
—In their letter, Drs Sarff, Davis, and Hyndiuk make three points. The first is that nonparametric statistics must be used to analyze data that do not approximate a normal distribution. We are aware of this and have used these methods where appropriate, as we state in our paper. Second, the appropriateness of the statistical tests that we have used is questioned. Our published description of the statistical methods we used was incomplete, and, therefore, misleading. However, the statistical tests that we used are appropriate, and the results that were derived from the statistical analyses are accurate as published. We have again verified the methodology with our statistical consultants and have checked our results. Both are correct.Finally, Drs Sarff, Davis, and Hyndiuk make a plea for the uniform presentation of data. Their letter would lead the casual reader to believe that this is an idealistic plea made solely
Leibowitz HM. Statistical Approaches-Reply. Arch Ophthalmol. 1978;96(5):911. doi:10.1001/archopht.1978.03910050507029