VileR wrote on 2020-01-05, 12:32:
badmojo wrote on 2020-01-05, 02:19:
The background colour is pretty powerful but I prefer a good contrast over the all too common light grey text on a white background that has me bobbing my head to find a workable viewing angle.
Yup, the overall choice of bright-on-dark is very welcome; screens are not paper. OTOH there are objective (i.e. scientifically backed) reasons for avoiding very saturated colors as site backgrounds - some examples... not that the colors here are especially bad or anything, but toning it down slightly wouldn't hurt.
This is a long and boring post about scientific research so please feel free to skip it as it doesn’t have any meaningful bearing on your life probably, but I typed it out so I might as well send it anyway 😀
Before I start, please understand that it is not my intent to use this post to dismiss the concerns that people have. Rather, I only want to be clear about (1) how the site measures up to the standards in that article, and (2) the conclusions the article draws from the articles it cites.
From the examples given by UX Movement we can determine whether the VOGONS background meets their saturation criteria. The HSB saturation of the VOGONS background is 60%.
The “bad” background colours given have saturation values of 84%, 85%, 88%, 91%, 93%, and 100%.
The “good” non-pastel background have saturation values of 38%, 56%, 67%, and 70%.
So the VOGONS background falls within the range of “good” saturation, according to that article.
Now, about that research. The first study is on the use of colour to draw attention. The second and third studies are about arousal and valence (i.e. pleasantness) when using colours of different hues and intensities. The leading claim in the UX Movement article that “too much [saturation] in a large area overstimulates the retinas which can strain the eyes” remains an arbitrary and unsourced claim. None of the linked studies include any attempts to draw causal relationships between physiology and measured responses, with the exception of a short note in study three when discussing the paradoxical (lower) arousal scores during exposure to blue light.
How were the studies conducted, and what do they actually say?
The first study, DOI:10.1002/col.10214, looks at the effects of hue, saturation, and brightness on attention by showing participants a 9x7 grid of colour blocks on a CRT in an office-like environment using eight different coloured backgrounds, with one axis being H and the other being SB. The coloured backgrounds always have maximum saturation and brightness (SB100). They ask the participants “Which colour square attracts your attention the most on the background colour on the screen?”. The result was that colour squares at SB100 were considered the most attention-grabbing against the SB100 backgrounds 67% of the time. There was no statistically significant (P<.05) difference on hue for their main image set, but somehow they got highly significant (P=0.0001) results for 4 of 8 colours in their supplementary image set. I didn’t see any explanation for this in their analysis so this makes me slightly concerned there was some flaw in at least part of the study. Also, the raw data is unavailable for analysis (of course), so it’s not known to me what people were choosing when they weren’t choosing the SB100 colour squares.
The second study is not available without a fee so I am unable to review it. The abstract sounds very similar to the third study, though it seems to reach different conclusions. The abstract talks about self-reported arousal and valence measurements and their correlation to skin conductance response (SCR) magnitude. The abstract reports that only colour saturation, not hue or brightness, was associated with a measurable change in SCR, and that this change in SCR was correlated only with self-reported arousal ratings.
The third study, DOI:10.1007/s00426-017-0880-8, looks at the effect of colour on SCR, heart rate (HR), and self-reported feelings of arousal and valence. It tested pure red, green, blue, and grey. There were 3 levels of brightness for each group, and for the colours, 3 different levels of saturation. Participants were shown colours on a Eurolite LED light panel in a dark room. The results showed arousal was positively correlated with warm colours, high saturation, and high brightness. The results also showed valence was highest in medium saturation colour, then high saturation colour, then low saturation colour, which is inconsistent with other studies. There was no statistical significance in valence based on hue alone, but there was when combined with saturation and brightness, which is also inconsistent with other studies. HR showed no correlations between any of the chromatic stimuli and the one significant effect (greys) was <1bpm difference from baseline. There was a sex-selective correlation in the valence of greys (lower in women). The authors describe various limitations and inconsistencies in the discussion section and suggest further study is needed to determine anything beyond (1) “color has systemic effects on the emotional state of a person viewing the color” and (2) these effects are not solely due to any single property of colour (i.e. it’s not just saturation which causes arousal).
So any practical conclusion that I can see to take from these particular studies is pretty tentative. We can say (from study 1) that when you are displaying a background with maximum saturation and brightness, other colours with maximum saturation and brightness are the best default choice to draw attention away from the hideous background. We can say (from studies 2 and 3) that there is some psychological arousal response to colours, and that this arousal response is more intense with saturation—but that it is also observable with changes to hue and brightness in certain combinations.
The UX Movement article seems to oversimplify the details of the first study since hue was competing against a background which was maximally bright and saturated. Against different backgrounds, the impact of hue may become more significant, but the study doesn’t test that. The third study actually shows a higher correlation between hue and arousal than between brightness and arousal, so the UX Movement article’s claim that “brightness [has] a more significant impact” seems to actually be wrong here according to the data.
OK, that’s it from me for now. Always happy to continue to receive research from which I can draw better conclusions, or critiques if it seems like I have misinterpreted study data or results (I am not a data scientist).