Jon: I am not worthy of you.
Woman: You got that right.
Jon: So much for humility.
Garfield: You got that right.
The author writes:
This is a repetition of an experiment first done by Comments on a Postcard. Each of the three authors independently coloured a black-and-white version of the Garfield strip from 2006-01-27. The three versions were combined into this strip, of which each panel contains one colour channel from each author, in three different combinations.
Manyhills provided the red channel of the first panel, the blue channel of the middle panel, and the green channel of the last panel. He writes: I've gone for the nursery-esque "cutting out bits of translucent paper and pasting them over the picture" look. I think it works pretty well.
Henning Makholm provided the green channel of the first panel, the red channel of the middle panel, and the blue channel of the last panel. He writes: I came this close to reusing the colour scheme from #141, but decided it probably wouldn't combine well with the other submissions.
Taneb provided the blue channel of the first panel, the green channel of the middle panel, and the red channel of the last panel. He writes: I wasn't sure if I should modify the strip further, so I didn't, I hope this is okay.
Through our results seem to validate the basic concept of collaborative colouring, they do not quite reach the levels reported in the original Postcard experiment. One possible explanation is that Postcard has a well-known and distinctive colouring style that the subjects could emulate, whereas a Garfield strip implied fewer conventions for our participants to follow. This suggests that the choice of line art for this kind of experiment has a greater influence on the outcome than previously assumed. Clearly, further research in this direction is warranted.
Another factor is that the Postcard work did not disclose how to choose which colours to use from which participants in each panel. Calculations show that there are about 1.2×101 possible choices here, and it is obviously not feasible to exhaustively survey a combinatorial space of that size. Apparently the Postcard author was able to enlist a number of volunteers for an Internet-based distributed search, which raises questions of selection bias and how to maintain consistent standards. In our experiment, the Principal Investigator simply generated a small number of samples and heuristically selected the one that best showcased the airbrushed highlights in his own colouring. We do not, however, suggest this as an ideal strategy to follow in future work; further research is needed to identify appropriate protocols.
Original strip: 2006-01-27.