Less Wrong is a community blog devoted to refining the art of human rationality. Please visit our About page for more information.

The Red Bias

30 Post author: Jack 20 April 2010 11:42AM

Summary: This color alters your perception of the world. Evidence that it does, how it does, why it does and some implications are presented below.

(Overcoming Bias: Seeing Red)

http://design-crit.com/blog/wp-content/uploads/2009/06/michael-jordan.jpg

Across a range of sports, we find that wearing red is consistently associated with a higher probability of winning. These results indicate not only that sexual selection may have influenced the evolution of human response to colours, but also that the colour of sportswear needs to be taken into account to ensure a level playing field in sport.1

In the study quoted above Hill and Barton examine the outcomes of the 2004 Olympic Games in boxing, tae kwon do, Greco–Roman wrestling and freestyle wrestling. In these events competitors were for each bout randomly assigned red or blue outfits. In the matches where one side dominated the other outfit color made little difference. In close matches, however, combatants in red won over 60% percent of the time. This makes sense since there are presumably other factors that effect the outcome.

In soccer (or football, as the case might be):

Since 1947, English football teams wearing red shirts have been champions more often than expected on the basis of the proportion of clubs playing in red. To investigate whether this indicates an enhancement of long-term performance in red-wearing teams, we analysed the relative league positions of teams wearing different hues. Across all league divisions, red teams had the best home record, with significant differences in both percentage of maximum points achieved and mean position in the home league table. The effects were not due simply to a difference between teams playing in a colour and those playing in a predominantly white uniform, as the latter performed better than teams in yellow hues. No significant differences were found for performance in matches away from home, when teams commonly do not wear their “home” colours. A matched-pairs analysis of red and non-red wearing teams in eight English cities shows significantly better performance of red teams over a 55-year period.2

Of course it still isn't clear how red soccer teams win. They might have the benefit of deferential refereeing or they might be intimidating opposing teams.

Same goes for the combat sports. Hill and Barton figured that the color red had some physiological effect, perhaps increasing the testosterone levels of the player in the dominant color. But a study by Norbert Hagemann et al. suggests that the color red has a biasing effect on referees:

We propose that the perception of colors triggers a psychological effect in referees that can lead to bias in evaluating identical performances. Referees and umpires exert a major influence on the outcome of sports competitions. Athletes frequently make very rapid movements, and referees have to view sports competitions from a very disadvantageous perspective, so it is extremely difficult for them to make objective judgments. As a result, their judgments may show biases like those found in other social judgments. Therefore, we believe that it is the referees who are the actual cause of the advantage competitors have when they wear red. Because the effect of red clothing on performance and on the decisions of referees may well have been confounded in the original data, we conducted a new experiment and found that referees assign more points to tae kwon do competitors dressed in red than to those dressed in blue, even when the performance of the competitors is identical.3

By digitally altering the color of the competitor's outfit they were able to alter the judge's ruling on the outcomes. The video here shows what they did. Of course, the effect could be a product of both referee bias and intimidation.

Why does the color red have this effect? The explanation given by every single study I have seen is that we're just not that different from the rest of the animal kingdom where red is an indicator of a high position in dominance hierarchies. In short, we're like mandrills.

Male mandrills also possess rank-dependent red coloration on the face, rump and genitalia, and we examined the hypothesis that this coloration acts as a 'badge of status', communicating male fighting ability to other males. If this is the case, then similarity in color should lead to higher dyadic rates of aggression, while males that differ markedly should resolve encounters quickly, with the paler individual retreating. Indeed, appeasement (the 'grin' display), threats, fights and tense 'stand-off' encounters were significantly more frequent between similarly colored males, while clear submission was more frequent where color differences were large. We conclude that male mandrills employ both formal behavioral indicators of dominance and of subordination, and may also use relative brightness of red coloration to facilitate the assessment of individual differences in fighting ability, thereby regulating the degree of costly, escalated conflict between well-armed males.4

 Perhaps related is the fact that human skin becomes flushed, and thus reddish when a person is angry but when a person is afraid, they get pale. 

Now, if this effect were limited to sporting events some of us might not care. But we have no reason to think it is limited to sporting events. This phenomena could effect our beliefs on the micro level, leading us to believe someone is more of a threat than they actually are or altering our perception of a person's status. It also recommends wearing red to signal dominance and aggressiveness. Given the popularity of the hypothesis than women find men who signal social dominance more attractive someone ought to test to see if women find men in red more attractive (it actually works in reverse but probably for totally different evolutionary reasons).

More troubling, I think, is the effect this bias could have on the macro level. Consider, for example, the widespread belief among American voters that Republicans are stronger on national security and better able to protect the country. Likely most of this belief is fostered by Republican leaders using more aggressive language, the presence of a pacifist faction in the Democratic party and (of late) the Republican party's greater willingness to use military force. But perhaps some of the GOP's image as a party of strong leaders is the result of the color with which they are constantly identified. Keep in mind that a lot of voters know next to nothing about the actual differences between the parties. Which side looks stronger to you?

 

Or consider geopolitics and the resources that went into beating the Soviet Union which by many accounts was never a serious economic or military rival of the United States. And consider how scared Americans were of the "red menace".

http://upload.wikimedia.org/wikipedia/commons/thumb/7/72/Flag_of_the_Soviet_Union_(1955-1980).svg/600px-Flag_of_the_Soviet_Union_(1955-1980).svg.png

Obviously there were plenty of incentives for Americans to be scared and for politicians to scare Americans. And it isn't the case that the USSR was no threat to the US at all. But politics appears to be a sphere of human activity where symbolism is important. America's perception of the Soviet Union as a dominant and aggressive foe probably increased the chances of armed conflict. And given that the conflict between the US the USSR came close to nuclear exchange it seems plausible that communism's choice in hue was responsible for increasing (albeit slightly) the probability of a lot of people dying.

China's color is red as well.

 

by Jack Noble

 

1 Red enhances human performance in contests, Russell A. Hill & Robert A. Barton, Nature 435, 293 (19 May 2005).

2 Attrill, Martin J., Karen A. Gresty, Russell A. Hill and Robert A. Barton. 2008. Red shirt colour is associated with long-term team success in English football. Journal of Sports Sciences. 26(6):577-582.

3 Hagemann et al. When the referee sees red, Psychological Science, Volume 19, Issue 8, Date: August 2008, Pages: 769-771

4 Joanna M. Setchell and E. Jean Wickings, Dominance, Status Signals and Coloration in Male Mandrills (2004), Ethology 111 (1): 25-50.

Comments (61)

Comment author: SilasBarta 20 April 2010 04:04:36PM *  8 points [-]

Great article! But then you lose me when you get to the bit about red vs. blue states and communists: you're basically making inferences from tiny sample sets without comparison to the sample space, effectively making it seem like numerology. (Look at how people were scared of these instances of nines in the past!)

Comment author: Jack 20 April 2010 07:30:05PM *  4 points [-]

Well I'm certainly speculating and I think I made that reasonably clear. But it isn't like I came up with this wacky theory about the color red on my own and then set out to find examples of people being scared of red (like in numerology). Rather, I have this hypothesis (which has it's own evidence) and I'm considering places where this bias could have had an effect. Part of what we have to do as rationalists is identify where biases could enter our own thinking. These are two that occurred to me and when nuclear war is involved even tiny bias effects have huge implications.

Comment author: ResistTheUrge 22 April 2010 12:13:34PM 1 point [-]

My girlfriend has always told me that I look best in red or black shirts.

Are there any studies on sexual / romantic attraction and clothing color?

Comment author: Jack 22 April 2010 06:03:51PM 0 points [-]

All I've seen is the study whose summary is linked to in the post suggesting red makes women more attractive to men.

Comment author: Dapple 22 April 2010 02:48:29AM *  1 point [-]

I can think of a scenario where the colour red tends to have the opposite effect... http://en.wikipedia.org/wiki/Redshirt_(character)

Comment author: PhilGoetz 22 April 2010 12:23:04AM 1 point [-]

Great post - but please insert a horizontal break near the start of it, so the whole thing doesn't appear on the "NEW" page.

Comment author: Jack 22 April 2010 12:58:35AM 0 points [-]

Done. Thanks. And sorry, still new at this.

Comment author: Morendil 21 April 2010 10:47:31AM 4 points [-]

Why don't we see many more people wearing red?

Can the rest of the color spectrum be ranked similarly in its effect on referees' evaluations?

Comment author: kes 22 April 2010 03:05:46AM 1 point [-]

Why don't we see more people wearing red?

I own one red shirt (The 'Communist Party' shirt, which is great) and only wear it rarely. But I have a theory as to why it isn't a more common color to be worn. I would venture to believe that red clothing in sports is ok, but if you wear it in normal life you are more likely to be seen as brash and wild.

I know that the few people I know that wear red will actually wear a whole slew of ugly colors (lime green, neon purple) and not just that one color. Also, I must wonder about the surge of pink clothing we are seeing in American business.

Comment author: nazgulnarsil 21 April 2010 10:32:05PM 1 point [-]

i started wearing red as a result of hearing of results like these years ago. can't tell if it has had an impact with one data point, it is eye grabbing though.

Comment author: Jack 21 April 2010 10:49:01PM 0 points [-]

Why don't we see many more people wearing red?

I was told when I was younger that red didn't look good on me because of my pink Irish skin.

Can the rest of the color spectrum be ranked similarly in its effect on referees' evaluations?

Not a lot of evidence here. Some psychologists briefly thought blue had an advantage over white in judo matches but it turned out that combatants in blue were higher seeds and thus were expected to perform better. Black seems to have some similar effects of red but I don't think they've ever been paired head-to-head.

Yellow got it's ass kicked in English football. Other than that I don't know.

Comment author: knb 20 April 2010 11:25:54PM *  6 points [-]

The GOP has only been identified as "red" since 2000. Before 2000, the GOP and Democrats alternated colors on electoral maps every 4 years.

I suspect that the 2000 color assignment stuck because Tim Russert's electoral maps were such a cultural touchstone from that year. It was after 2000 that a series of books emphasizing the red state/blue state cultural divide started appearing.

Also, the electoral map you show makes the GOP look "stronger" mainly because the area of the red states happens to be larger than the more densely populated blue states.

Comment author: sketerpot 21 April 2010 03:53:33AM 15 points [-]

Also, the electoral map you show makes the GOP look "stronger" mainly because the area of the red states happens to be larger than the more densely populated blue states.

Let's test this. I inverted the colors on that map, so the Democrats are red and the Republicans are blue. Which looks stronger?

inverted election map

Comment author: Alicorn 21 April 2010 03:58:40AM 8 points [-]

Oh that is creepy.

Comment author: RobinZ 21 April 2010 11:13:33AM 0 points [-]

Seconded.

Comment author: Nanani 23 April 2010 01:35:47AM 2 points [-]

I wonder, could the effect be reduced by using a darker shade of red? This red is certainly more vivid than the blue to my eye.

Comment author: knb 21 April 2010 05:55:28AM 3 points [-]

Wow. My immediate impression is that red still looks stronger.

Comment author: Konkvistador 21 April 2010 09:21:37PM 1 point [-]

I don't agree.

My first impression was that the red states where fragmented, blue appeared to be stronger. My second impression was indeed that reds looked menancing.

Size however was lower on the list of things I noticed vs. color.

Comment author: Jack 21 April 2010 04:44:32AM 5 points [-]

Blue looks like it would have an advantage in a war against red but it seems more evenly matched than the original. And the red states seems like more rugged, tougher places. If I was a different country the blue area feels like the easiest to invade.

(It's pretty weird that I feel comfortable making these nonsensical judgments about blotches of color. Also, feel free to discount my opinion if you think I'm vested in the outcome as the author of the OP)

Comment author: sketerpot 21 April 2010 05:46:29AM 11 points [-]

When I look at the color-flipped map, I feel like the red area is aggressively penetrating the blue area. When I look at the regular map, I feel like the red is pushing the blue against a wall. Color's influence on our perceptions is weird.

Comment author: cupholder 21 April 2010 04:11:38AM 1 point [-]

Blue for me.

Comment author: LordTC 04 May 2010 11:51:40PM 0 points [-]

Are those really the same colors, the red seems more orange in this one than the one up top, and the blue seems darker than the original.

Comment author: sketerpot 05 May 2010 05:55:32AM 0 points [-]

If the "red" and "blue" of the original had some color components in other channels, then those would indeed by subtly different colors in my altered picture. I just swapped the red and blue channels of the image.

Comment author: pscriber 04 May 2010 06:19:12PM 0 points [-]

To me, the color really doesn't make a difference; what strikes me is the relative size of each area, and the Republican block still looks more intimidating than the Democratic ones.

It might be worth noting that I've always seen the southern U.S. as a value enemy of sorts - I'm a politically liberal atheist in South Carolina - so I'm wondering if that might play into my perceptions of this. Can we test this with another map?

Comment author: NancyLebovitz 04 May 2010 06:33:22PM 0 points [-]
Comment author: Jack 20 April 2010 11:45:15PM 1 point [-]

This is true. The next step would be seeing if there was a change in party image after this happened that didn't have another cause. I suspect identifying any color biasing effect will be impossible. The other relevant variables (9/11, the War on Terror, Iraq, past military interventions etc.) are too unstable and we don't know when the identification of the GOP with the color red sunk in.

Comment author: komponisto 20 April 2010 11:35:20PM -1 points [-]

The GOP has only been identified as "red" since 2000. Before 2000, the GOP and Democrats alternated colors on electoral maps every 4 years.

I'm not sure that's true. I recall the current color assignment being in place at least since 1992.

Incidentally, the UK has the "opposite" color pattern (Labour red, Conservative blue), which I must say I find jarring (even if historically understandable).

Comment author: Bo102010 21 April 2010 12:38:32AM 2 points [-]

See (http://en.wikipedia.org/wiki/Red_states_and_blue_states) for some information on the U.S. and coloring maps.

Comment author: Jack 21 April 2010 12:29:28AM *  -1 points [-]

Here is Time's 1996 map (warning, it's a PDF).

I think knb might be wrong about the alternating thing, though. My understanding was that the left of center party was historically the "red" party and "blue" the conservative party, in the European tradition.

Edit: the problem is that now it has been standardized as red= Republican so every map on the internet is this way going back through the 70's.

Comment author: komponisto 21 April 2010 02:12:56AM *  2 points [-]

From the Wikipedia article linked to by Bo102010:

As late as 1996, there was still no universal association of one color with one party.[7] If anything, by 1996, color schemes were relatively mixed, as CNN, CBS, ABC, and The New York Times referred to Democratic states with the color blue and Republican ones as red, while Time Magazine and the Washington Post used an opposite scheme.

My memories from the time would have been formed mainly from television rather than printed sources, so there you go. (Although one printed memory that does stand out is, of all things, the French magazine L'Express, which used the Democrat-blue/Republican-red scheme in a 1996 article showing Clinton's 1992 victory.)

Comment author: Cyan 20 April 2010 11:55:19PM 1 point [-]

It's not just the UK that uses the opposite colour scheme. Canada does too.

Comment author: LordTC 04 May 2010 11:54:59PM 0 points [-]

The US is the unusual color scheme here though. Red is generally associated with the left, look at the flags of most communist countries, and party colors etc.

And I think it flipped only once not every four years, because when I look at the Reagan maps, he's blue in both of them!

Comment author: CronoDAS 20 April 2010 07:13:17PM 3 points [-]

The leader of the Power Rangers is always the Red Ranger. (And Power Rangers is based on the Japanese show Super Sentai.)

Comment author: Jack 20 April 2010 08:15:10PM 1 point [-]

I thought he got replaced by the white ranger at some point.

Comment author: gwern 20 April 2010 08:35:53PM 3 points [-]

Ah, but that white ranger eventually turned red.

Comment author: JonathanLonsdale 25 April 2010 06:22:09PM 0 points [-]

Also interesting how Mars ("the red planet") is named after the Roman god of war.

Comment author: Aurini 20 April 2010 05:47:00PM 5 points [-]

While reading this I couldn't help thinking about how Chinese culture considers red a lucky colour. Not to mention how red/black combinations are often the uniform of powerful Eeeevil forces in art and culture. It might just be an accident, but it's enough of a chance that I'll consider wearing my red shirt whenever I need a bit of 'luck' - such as later today, when I bring my (hopefully fixed) motorcycle in for an Out-Of-Province Inspection.

Comment author: knb 20 April 2010 11:37:25PM 0 points [-]

The Red/Black=evil thing is true. Could it just be the influence of Nazis on the popular imagination? I'm trying to think of red/black combos from before WW2.

Comment author: Jack 20 April 2010 11:48:53PM *  1 point [-]

Germany's flag in WWI was black, red and white. I kind of doubt they saw themselves as evil.

One think I hated about V for Vendetta was how obvious the fascism looked. If you want to oppress a country changing the flag so it consists of alien and sinister red on black is, like, a dead giveaway.

Edit: Related video

Comment author: knb 21 April 2010 01:30:52AM 0 points [-]

Well, nobody sees themselves as evil. I'm suggesting that the red/black combo might be associated with evil after WW2 because the Nazis were so evil.

The Norsefire flag in V for Vendetta is pretty obviously a Nazi'd up version of the St. George Cross.

Comment author: Jack 21 April 2010 01:43:06AM *  0 points [-]

I wasn't disagreeing with you, just preempting any suggestion that the Red/black= evil phenomena was especially universal or ancient. I think the Nazi explanation is a good one.

That said, red and black do seem to be the color most associated with hell and the underworld.

The Norsefire flag in V for Vendetta is pretty obviously a Nazi'd up version of the St. George Cross.

"When fascism comes to America, it will be wrapped in the flag and carrying a cross." -Sinclair Lewis

Comment author: Aurini 21 April 2010 01:24:46AM 0 points [-]

The only example from antiquity that I can think of is Heironymous Bosch's Garden of Earthly Delights: http://upload.wikimedia.org/wikipedia/commons/8/80/GardenED.jpg

The colour is faded, but hell is definitely red/black.

Incidentally, red/black/gold not only looks evil, it looks rich and evil.

Comment author: Jack 21 April 2010 01:55:10AM 1 point [-]

Well red and black are pretty much the colors you would expect to see in an underground pit full of fire.

Comment author: Alicorn 21 April 2010 02:59:40AM 0 points [-]

I would expect brown and orange and yellow too. Maybe blue, depending on what was on fire.

Comment author: Unnamed 20 April 2010 04:35:53PM 7 points [-]

There is a similar line of research on the color black and its associations with aggression, evil, death, sin, and impurity. Frank & Gilovich (1988) (pdf) found that athletes with black uniforms commit more penalties - they seem to both act more aggressively and to be seen as more aggressive. Sherman and Clore (2009) (pdf) found what they called the moral Stroop effect - people are faster to categorize words as moral or immoral when moral words are written in white font and immoral words are in black font.

These associations could affect racial dynamics (e.g., stereotypes about black people), but there isn't any direct evidence on that, as far as I know.

Comment author: Jack 20 April 2010 09:03:07PM *  0 points [-]

Interesting. Have you seen any suggestion that these reactions to the color black are innate? Light and darkness as a metaphor for good and evil is a common enough trope that I could see culture alone generating some of these effects. Do you know if this effect remains in cultures where people have darker skin?

Comment author: wnoise 21 April 2010 04:55:17AM 4 points [-]

Though it's been changing due to western influence, white is the color of mourning in Chinese culture, and black is pretty much neutral, and even associated with Heaven (which of course is vastly different than the Christian concept).

Comment author: Unnamed 21 April 2010 04:41:44AM 3 points [-]

There's evidence that these associations with the color black are common across cultures, including in Africa. Frank & Gilovich discuss this at the start of their paper, and here is one cross-cultural article that they cite which found that black is seen as bad, strong, and passive, white is good and weak, and red is strong, active, and affectively salient.

These associations could be learned through widespread experiences like night being more dangerous and objects becoming darker as they get dirty, rather than being innate. Those widespread experiences could also lead many cultures to develop tropes of darkness and light, which could strengthen the associations.

Comment author: Blueberry 20 April 2010 03:13:40PM 1 point [-]

I'm curious how this affects racial dynamics. What about people with very dark skin, for instance: how does this apply? I don't think there's any population group with literally red skin, but maybe increased melanin would have a similar effect, by making it more difficult to see blushing or becoming pale.

Of course, there are a great deal of confounding factors there.

Comment author: Clippy 20 April 2010 04:12:03PM 4 points [-]

What you call "white" and "black" people are actually neither black nor white; rather, they differ in the darkness of the red component of their skin color. (Though most don't view brown as red, its color breakdown is dominated by red; some cultures do see brown as a type of red.)

This actually ties in with why clippys don't visit earth. To put it bluntly, "These beings terminate each other over relative skin redness. I'm silver. Exit!"

Comment author: Blueberry 21 April 2010 02:58:13AM 1 point [-]

But, Clippy, we humans also fight over natural resources like iron ore... doesn't that make you happy?

Comment author: Clippy 21 April 2010 07:36:04PM 1 point [-]

Not if it means the iron ends up in some unusable or useless form, as is too often the case in these wars...

Comment author: tut 20 April 2010 04:30:18PM *  1 point [-]

If you are an alien AI, how come you use human colors?

ETA I mean, what colors you see depends on what cell types you have in your eyes. Red green and blue makes a lot of sense for a creature that lives in earth atmosphere and eat earth fruit etc. But what did your creators want you to distinguish that used the same colors?

Comment author: Clippy 21 April 2010 07:36:25PM 1 point [-]

What User:Blueberry said, among other reasons.

Comment author: Blueberry 21 April 2010 02:57:22AM 2 points [-]

I would imagine Clippy is referring to the way humans see colors. That is, when e says "silver" e means "silver as seen by humans".

Comment author: eirenicon 20 April 2010 03:44:37PM 1 point [-]

I think there is probably no relation. My guess is that red signalling probably precedes variation in skin colour, perhaps even loss of body-wide hair. It is a thoroughly unconscious bias, and does not apply to pink, or orange, or peach, but red, especially bright, bold baboon-butt red. In any case, I hope the sporting tests were controlled for skin colour, because that does seem like a weighty factor when considering scoring bias.

Comment author: Jack 20 April 2010 09:13:43PM *  0 points [-]

In any case, I hope the sporting tests were controlled for skin colour, because that does seem like a weighty factor when considering scoring bias.

For sure. In the case of the combat sports, outfit color was assigned randomly by the competition. In the Hagemann study, the outfits were alterted digitally so it was literally the same fighters. The goalie test which I linked to claims to use the same person just with different jerseys. For English football it seems unlikely that skin color and jersey color had any correlation but of course it wasn't explicitly controlled.

EDIT: Though it occurs to me that red could have different effects depending on the skin tone of the competitors (helps darker contestants, hurts lighter ones or something) and that certainly wasn't controlled for in any of the studies.

Comment author: CarlShulman 20 April 2010 02:52:09PM 4 points [-]

You should put a link to Robin's Overcoming Bias post on the same subject at the top.

Comment author: Jack 20 April 2010 07:13:19PM 2 points [-]

Ugh. I knew there had to be a Overcoming Bias post but I searched and couldn't find it.

Comment author: Kevin 20 April 2010 08:22:34PM 1 point [-]

Your post is better though.

Comment author: thomblake 20 April 2010 08:27:14PM 0 points [-]

This post adds a lot and is probably worth promoting.