Suppose we conducted an experiment that found that the color red makes people depressed and the color blue makes people happy.

Now suppose we found out that when looking at a red object, person A sees red but person B perceives blue.

Would the conclusions of the experiment still apply?

My take would be that the results would still apply, just that they apply to different definitions and it would be just a matter of semantics.

  • $\begingroup$ This sounds like a linguistics question to me. :) Language is a matter of semantics. We call something 'blue' since we agree on something we see. The way we can tell someone is colorblind is since they can not differentiate between certain colors. As long as your hypothetical 'color swap' does not impact any relations to other experiences (e.g., one blue is no longer lighter/darker than another color), there would be no way to tell one person experiences 'blue' as 'red' or vice-versa. $\endgroup$ – Steven Jeuris Apr 16 '18 at 9:01
  • 1
    $\begingroup$ Also, this has nothing to do with experiment-design, unless you would be asking how to conduct an experiment which tries to assess whether or not these colors are perceived differently by people. Possibly, this is a more interesting question. Your current question presumes you can conceive of such an experiment without any prior research, making this question eligible to be closed as not framed in psychology or neuroscience. $\endgroup$ – Steven Jeuris Apr 16 '18 at 9:05
  • 1
    $\begingroup$ This is an interesting variation on the inverted spectrum thought experiment; I asked a related question about the practicality of this experiment here: psychology.stackexchange.com/q/10466/7001 $\endgroup$ – Arnon Weinberg Apr 17 '18 at 0:14

While the sensation of color is well defined, the perception of color is indeed shrouded in ambiguity (Brainard, 2001). There are no colors, just electromagnetic radiations that oscillate in different frequencies in the frequency domain. Our brains interpret those various wavelengths as color.

Now on to your experiment: if you find a statistical difference and you later find out that a group of people perceive the color blue as red and vice versa, the first thing you want to do is control for that population and take them out from the data set. Then you can compare the 'true seers' with the 'reversed seers' and see if they are statistically different in their emotional response. If yes, you'd better split them and analyze them separately. If not, a pooled analysis is fine, perhaps you can add the normal or inverse perception as a co-factor, dependent on the exact analysis.

But then, perhaps, the most important question here is - how would you find out what a person actually perceives?

- Brainard, Int. Encyclopedia Social & Behavioral Sciences, Smelser & Baltes (eds.), Pergamon Press, Amsterdam

  • $\begingroup$ This question is hypothetical because as you say, we can never know what somebody perceives. I guess I'm asking that you imagine yourself to be an outsider who knows what people perceive, even though the experimenters and subjects do not know it. Does this in any way change the validity of the results? $\endgroup$ – S S Apr 16 '18 at 11:07
  • 1
    $\begingroup$ @SS I'm at a loss what you mean $\endgroup$ – AliceD Apr 16 '18 at 11:28
  • $\begingroup$ Never mind, I think my question is poorly worded and I am confusing a few concepts. $\endgroup$ – S S Apr 16 '18 at 11:49
  • $\begingroup$ @SS Hence my suggestion to rephrase it in the comments. As indicated in Alice's answer as well, the last sentence is likely the question you should be seeking answers to/understanding better, prior to asking this 'hypothetical' question. $\endgroup$ – Steven Jeuris Apr 16 '18 at 12:36
  • 1
    $\begingroup$ If rephrased as "how to conduct an experiment which tries to assess whether or not these colors are perceived differently by people" I would find it a very interesting question as well, ideally basing the question on the source you provided in this answer already. :) As is, though, we could just as well leave this question and the OP can create a new one when interested. That said, It might set a bad example that questions like this remain open, given our new guidelines. $\endgroup$ – Steven Jeuris Apr 16 '18 at 12:39

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.