I am in the course of designing an application that allows users to ask survey questions by swiping along a coloured scale. For preliminary testing, I just took a coloured gradient from the internet that transitions from red green with yellow in the middle, which has worked fined, in principle. However, since the app will now be subject to scientific testing and perhaps result in a scientific paper, I was wondering about two things. First of all, is there any literature or reference as to which shades of green / red are recommended for scales that represent agreement / disagreement or any other scientific literature related to this topic? Secondly, I asked myself if there is a scientific method for obtaining a "perfectly fair" gradient between two colours, and if there is a way to obtain and especially formalise the process maybe by using mathematical formulas, so that it can be replicated by anyone. Since I'm really new to the design aspect, please excuse my limited knowledge, any hint or thought is highly appreciated.
Color theory is not a hard science, its more traditional level stuff. Mixed with older hard science studies. Now color science, which is a cognitive science subfield, is a hard science, but also really really complicated. .
There are some standard scientific scales. Like for example on better figures org. Color science gives you methods to calculate color distances, but not to my knowledge how sequences should be made. We can not really objectively answer the question how to correctly go from red to green for example because there are no straight lines in color just several different straight and curved lines.
However you might want to avoid obsessing about this too much since once you do you will find that human color vision is affected by surrounding color ambiance so if you a want to be scientifically truly accurate you would need to calibrate the tablet several times an hour. Even so you need to calibrate your tablets to show same color or a count for this in some way. No screen is exactly the same as another without this step.