Hello and welcome to the forums.
This sounds like an interesting project.
With regards to the various colors which absorbed the most heat, I think the main consideration might be how dark those colors were. I would think that blue and purple would not absorb so much; unless they were very dark blues and purples. Here is a spectrum analysis of heat lamps: http://heelspurs.com/a/led/black3.gif
which shows, as you might expect, that heat lamps emit the most energy in colors towards the red end of the spectrum, as expected. That would explain orange - did you also try red?
Why did the colored liquids absorb more heat than the paper covered ones? I hope other experts can chime in on this, but consider how the heat absorption works. With the paper, a part of it is exposed to the heat, and the paper absorbs the energy, transfers it to the glass (and may have some air between the two, also), which then transfers it to the water. In the colored water case, the heat absorption happens right in the water, which should be more efficient. Did you have a control with no coloring? How much did that one heat up?