PART 3New Experiments Three Factors On The Box plot, Box notch & Outlier Values

Abstract

New two experiments of the three factors, in this study were constructed to investigate the effects, of the fixed variations to the box plot on subjects' judgments of the box lengths. These two experiments were constructed as an extension to the group B experiments, the ratio experiments the experiments with two variables carried out previously by Hussin, M.M. (1989, 2006, 2007). The first experiment box notch experiment, and the second experiment outlier values experiment. Subjects were asked to judge what percentage the shorter represented of the longer length in pairs of box lengths and give an estimate of percentage, one being a standard plot and the other being of a different box lengths and also varying with respect to other elements such as, box width & box notch or box width & outlier values. When Hussin, M.M. (1989) suggested in the future research points (1, 2, 3), can take account of wider range of the variables levels and the changing length of the standard box plot effects on the subjects' perception of the box length and further investigations could be made into variations such as box notch with box length; box notch with box length; whisker length, and also outlier values with these variations. However, both experiments were used the stander box length as the middle box length levels were not used in the experiments. The results of these two experiments suggest that these variations effected the subjects' perception of box length, as a results of the interactions between box plot features by creating visual illusions as which effect the subjects ability to accurately judge box length, both experiments were run in statistics department, Baghdad University.

Keywords

box plot