Many corresponding color datasets have been collected under different illumination conditions over these years, providing adequate data for developing chromatic adaptation transforms (CAT). Nevertheless, these datasets exhibit notable, systematical discrepancies in visual data, probably stemming from their differences in the experiment setup and methodology. This necessitates a comprehensive examination of how the experiment-related factors influence the outcomes, accompanied by thorough discussions to establish theoretical references for the effective classification of datasets. Horizontal comparisons among these datasets indicate the influence of two crucial factors-viewing medium and experimental methods-on chromatic adaptation, albeit without systematic investigations. Additionally, the underlying mechanism contributing to the disparities observed among different media remains unclear. To address these limitations, this study selected three different media - surface colors, self-luminous colors, and independently illuminated surface colors, and two experimental methods - the achromatic matching and neutrality rating method. The results confirm the significant influence of the viewing medium on the adaptation state. Two crucial factors emerge as significant contributors: the color correlation with global illumination and the nature of the surface medium (reflective or self-luminous).