Improving Color Discrimination Through Set Manipulation

The Problem: To clarify the theoretical dilemma posed by the Helmholtz/Hering controversy about the nature of color discrimination, this study examined relationships between three set manipulated groups and two task types, counting and matching colors These variables were selected to investigate their effects upon learners' discrimination of object stimulus color. Experimental Design A 3 x 2 unbalanced factorial design was employed where color discrimination test scores were analyzed by a two-way analysis of variance, as well as a two-way analysis of covariance with California Achievement Test Cognitive Skill Index scores as the covariate. Procedure: A research sample of grade seven middle school students (N = 148) was assigned to three set manipulated treatment groups of two cells each. Cell A subjects performed color discrimination tasks of counting changes of stimulus color while Cell B subjects performed tasks of matching changes of stimulus color from the same visuals Subjects were blocked into three categories according to their assigned set, resulting in three analysis groups collapsed across task type levels implicit set (N = 50), explicit pictorial set (N = 48), and explicit pictorial/verbal set (N = 50) The data from the six cells were also analyzed according to task type collapsed across three levels of set: task A, color counting (N = 74), and task B, color matching (N = 74) Findings: Hoi— There would be no difference in color discrimination scores among the three Set. manipulated groups was rejected (F = 8 41, df = 2/147, p < 01) Ho2— There would be no difference in Task Type scores was rejected (F = 43 28, df = 1/147, p < 01), color counting performance being superior to color matching. Tukey's HSD post hoc comparison indicated that the treatment group receiving both visual and verbal cued instruction achieved significantly higher scores than the implicit (no) treatment group, and the pictorial cued treatment group. Ho3—There would be no interaction between levels of Set and Task Type was not rejected. The analysis of covariance on the same data supported the ANOVA findings (&&£. F = 5 91, df = 2/147, P < 01, and Task Type F = 51 31, df = 1/147, p < 01).