DOI: 10.36347/sjavs.2018.v05i06.010

Pages: 343-351

Abstract: In many animal researches mostly two way cross-tabulated data were evaluated by calculating a simple chi-squared test to see if the whole table deviated from the expected pattern. Such approaches to multiway frequency tables are mostly dissatisfying. Configural frequency analysis (CFA) is a general multipurpose tool and revolutionizes how we examine the cross-tabulation of two or more count variables [1]. CFA is a useful inferential tool used to evaluate the expected configural patterns in two-way to multiway cross tabulations of frequencies [2]. The results are possible types/antitypes depending on whether the observed cell frequencies are significantly lower/higher with respect to the base model. One of the goals of this study is to put the application of CFA technique into practice to investigate local associations for animal research in count data which was exemplified by simulated data for cause of mortality of calves. Also we compared the type and antitype cell numbers for different CFA approaches. The mortality data were not very well recorded in in Turkish cattle farming systems. Because of the difficulties of obtaining such risk data for calves’ mortality [3], we used simulated data to evaluate the implementation of configural frequency analysis in calves’ mortality. We have seen that CFA is a very powerful technique to investigate local configural associations, and CFA approaches must be evaluated for animal researches in the future to evaluate local relations.