It's true that the color cannot be trusted on an uncalibrated scanner. And even if it were calibrated, my guess is that most people are not using a hardware calibrated monitor using firebox with color profile enable, plus everyone sees color differently based on so many factors...
However, scanners are general pretty stable in terms of color variations, by default, most likely it has a higher color temperature and gives out a slight blueish color cast, but it shouldn't be too bad, and it should be much more accurate than an uncalibrated monitor or printer. So given a choice of "uncalibrated devices", a scanner is a good pick
sorry I posted the diagram in a hurry before heading out this afternoon without any explanation.
brecklundin was correct about the color axis.
Lab is an absolute color space(unlike RGBs). L=luminosity(brightness, values from 0->100, black->white), "a" axis goes from Red to Green, and "b" axis goes from Blue & Yellow. The upper 2 dots represents the background of the kindles, and the lower dots represent the bold texts' color. Humans eyes have extremely wide shifting dynamic range(based on the current light condition), so the absolute brightness is not much of an issue, the contrast is much more important.
Based on the scanned images,
K2 has the luminosity value of 42(texts) & 71(background), and K1 has the value of 39 & 69. With the relative contrast difference less than 4%, it is too small to make any conclusion given the test was not done in a controlled environment.
Although
if the images were accurate representation of their difference, 4% difference
is noticeable, as human eyes can detect 400:1 to millions:1 contrast ratio.