Skip to main content
ARS Home » Southeast Area » Auburn, Alabama » Aquatic Animal Health Research » Research » Publications at this Location » Publication #214579
Title: Analyzing data in aquaculture: practical significance, a new paradigm for determining the importance of results

Author
item Welker, Thomas
item WELKER, TIM - USDA,ARS,FOREST SERVICE
item Klesius, Phillip

Submitted to: Meeting Abstract
Publication Type: Abstract Only
Publication Acceptance Date: 11/10/2007
Publication Date: 2/10/2008
Citation: Welker, T.L., Welker, T.L., Klesius, P.H. 2008. Analyzing data in aquaculture: practical significance, a new paradigm for determining the importance of results. In: Aquaculture 2008. February 9-12, 2008. Orlando, FL. p. 508.

Interpretive Summary:

Technical Abstract: Analyzing data and interpreting results is often the most difficult and yet important part of the scientific research process. Currently, aquaculture researchers almost exclusively employ null hypothesis significance testing (NHST), a synthesis of the Fisher test of significance and the Neyman-Pearson hypothesis test, as the method for statistical inference of research results. Statistical significance testing is concerned with whether a difference in research results is due to chance or sampling variability, and NHST methods (e.g. analysis of variance) are often misinterpreted and do not provide researchers with what they really want or need to know: what is the practical significance of research results? Practical significance, on the other hand, is concerned with whether result are useful and applicable in the real world, i.e. are there meaningful differences between experimental groups or treatment means. While the use of statistical hypothesis testing dominates in aquaculture research, it is being deemphasized in other disciplines, and most aquaculture scientists are unaware of the important criticisms of NHST. The growing awareness of the limitations of NHST methods has led to a search for ways to supplement or replace these procedures. A variety of methods (e.g. effect size magnitude, confidence intervals, focused linear contrasts, and others) has been proposed and will be examined in the context of aquaculture research.