This paper indirectly has some evidence that *may* help you. I use it in a class. It is an esoteric look at some factors that influence the taste perception of orange juice, and color (lighter vs darker orange) is a factor. Research on wine aficionados may also help here.

In general I would argue that color is but one cue and it is the combination of color and other spatial cues together that would provide enough context to alter taste perception due to expectation/prior experience.