Kontact 

I don't know how you are managing to make statements of likelihood on zero data. That is certainly not "rational".


Let's put it this way - black cars might actually be more conspicuous on the road than gray cars, but if the drivers who select black cars are so aggressive that they cause a disproportionate number of accidents, now it makes black cars seem like they are not visible when they actually are.

A high contrast black car might be an excellent choice to make other drivers see and avoid you. But you can't tell if that is true or not because no one has actually studied visibility separate from driving style.

I-Like-To-Bike

Not really at least by not the organization that matters and knows how to evaluate crash statistics. In this case - to not use at all this set of so-called "statistics" .

Extract from the cited article about car crash by color:

njkayaker

It's not "zero data". It's incomplete data. Plus, it's not data that exists in a vacuum (that is, there are other situations where there is some information about visibility).

If a driver wants to pick a color that is likely to be more safe, there are lots of reasons why they aren't likely to pick a black or gray car.

There isn't any data to support this ("so aggressive").

Black cars might be more conspicuous than gray cars in certain situations. White cars might be more conspicuous than both in more situations.

No, this doesn't seem likely at all.

Black is likely not "high contrast" in low-visibility situations and not at all "high contrast" at night.

For the the situations where black might be "high contrast" (in bright sunlight?), the effect might be small.

Is black ever considered "high visibility" on roadways?

You have to make a choice without these studies.

Kontact 

I think many people have thought that black and white make excellent visibility markers due to their contrast with other colors. And some studies like this suggest that gray cars are "visible" since they get in fewer accidents, adding to the evidence that the study isn't actually showing us much about visibility.


If I was concerned about visibility, I would pick certain colors because of studies about the human eye and colors. I wouldn't base my selection on studies that completely fail to delineate between a variety of human factors.


That's all I'm getting at - sometimes studies like this don't actually give us data. They just add a layer of ambiguity that makes something rational become something else. I think the study doesn't tell us anything about visibility and just informs us of a correlation that has multiple causes. As a society we are constantly jumping on poor statistical evidence to point to causation, and we are wrong so often.

Pardon me, it is time to take my Statins with my fat free, high sugar yogurt. Because cholesterol must cause heart disease: There were studies!

Maelochs

Police cruisers (often called “Black and Whites”) were originally painted black and white because the high contrast was so readily visible.

Drivers in red cars seem to go faster …. Some studies seem to suggest. But in low-light situations, red and even international signal orange (‘danger orange”) can appear black (which is why some fire engines went to yellow, white, or lime green.)

Maybe it has nothing to do with how visible the car is and everything to do with driver behavior (not many people buy pink sports cars and drive them like race-driver wanabees.)

There are psychological factors at play with the colors people choose for their cars …. Not a lot of pale pink Mustangs or baby-blue Camaros …. (except real race cars painted in Gulf colors …. )

And considering that All cars are pretty well festooned with lights on every surface …. Not sure the color of a car much matters at all. A car you don’t see can be any color, and in the dark, a white can and a black car are both mostly just headlights, tail lights, and side marker lights

But the hi-viz things with bikes is unrelated. Even a flashing tail light on the side of the road ahead in the distance is easy to overlook for a car driver simply because it is not on the surface of travel (apparently.)

I have seen bikes with solid tail lights that looked very much like mailbox reflectors until the headlights caught them.

Studies I have seen show pedal reflectors to be the best identifier of a bicycle from the greatest distance—so of course almost nobody has reflectors on their pedals any more (I put 3M reflective fabric on the back of my shoes …. And I haven’t been hit since …. But that is fake data.)

Agreed. There is no correlation between Cause of accident an color

Much as I hate to agree with anyone, I have to agree with the truth as I see it.

Science isn’t data. Science is data collected from narrowly controlled tests designed to sustain or disprove a narrow hypothesis.

This is more, “Let’s measure some stuff …. At random.” No control groups, The tests are not sufficiently narrow.

There may well prove to be a statistical correlation between drivers with the sequence of letters “nce” in their names, or those letters in any order … would that “prove” that one could avoid accidents by changing one’s name?

Data Analysis is the key to understanding data. Pointing out correlations is not analysis. Understanding if and if so, how those correlations relate to the quantified result is where the meaning lies.

Basically, this studies says, “We see this stuff but we don’t know what it means.” Which is fine … but then a bunch of people come along and make up meanings.

Kontact 

That isn't what I said. There is a correlation between the cause and color, we just don't know what the cause is and what its relationship is with color. Correlation is not explanation.

Maelochs

Actually, we don't even know if there is a correlation between cause and color ... all we know is that there are certain colors of cars involved in more or fewer accidents than cars of other colors. Maybe the red cars are driven faster, the black more aggressively .... or maybe the black cars are less visible and more likely to be hit.

Whether it is car color or driver behavior, being seen by others or not ... the data doesn't say. All it says is gray cars seem to be involved in fewer accidents than black car and so on.

The military (National Guard) often uses public roads when doing its weekend warrior exercises. Theoretically, the camouflaged cars should be hard to pick out from the background ... but in reality, you never hear of collisions involving troop truck, military Humvees, and tank transporters.

Is it driver discipline, or is it that the vehicles actually stand out for being unusual, paint scheme aside? or is it that they tend to be on the road early on Saturday morning and in the middle-to-late afternoon on Sunday, when traffic is low and the sun is high?

Are those color-collisions correlation studies controlled for time? Weather conditions? traffic density?

Maybe you swa something I didn't it could be. But I don't see where the data shows anything at all about why it happens, just that ti does happen.

Why do black cars get involved in more accidents? is it Purely that people cannot see them? In daylight, at night, at dusk>? Almost all cars nowadays have their lights on when the engine is on by default ... do the drivers of black cars show a tendency to turn off the lights? Is this based on car color or driver personality?

Researchers would need to field and track a team of AVs painted in various colors to really do this test.

(That is called, "Opening a whole 'nother can of worms." No need to thank me ... I do my work out of a sense of social obligation.)

njkayaker

So what? We are talking about single color cars. It's only "high contrast" from the side. And the paint job serves to make the white part more noticable.

Maybe, it wasn't because it was "high contrast" but to make them easily distinguishable as police cars (against normal single color cars).

It's not likely that it's "everything to do". Your cop car example discounts this hypothesis anyway.

It's not unrelated just because you said so.

???

Are cyclists better off having flashing lights and being visible or not?

Are cyclists less "easy to overlook" without flashing lights and being visible?

I think a minority of cars in the US have daytime running lights.

They aren't cars. They also tend to be big. There are also very few of them and they don't spend a lot of time driving on public roads. It's a really long stretch to suggest that the camouflage makes things easier to pick from the background (the effect we are looking for).

In reality, they crash all the time.

https://www.macon.com/news/local/article155145199.html

There isn't any data that indicates that 1) this population is statistically less associated with collisions, and 2) that camouflaged cars would be more visible. You can't argue against the lack of data making claims based on no data.

What studies? It appears that this is more a myth.

So black is less visible than white. If "appearing black" is a problem, you wouldn't actually use black.

You are arguing against the conclusions of the article by offering multiple reasons that support it!

This is a strawman. No one has ever made this claim.

njkayaker

There aren't many markers (basically, next to none) with the apparent goal of visibility that use all black. Black is often used with contrasting bright colors (but we are talking about single colors).

You don't need to use these studies.

Those "certain colors" are typically not an option to choose. And, obviously, that doesn't mean black is more generally visible than white.

If you were doing this, you'd pick white over black (and you'd never pick black)!

If this had an effect, the results of the study would be expected.

You are arguing against the conclusions of the article by offering multiple reasons that support it!

wphamilton

Black is fine during the day, given normal sunlight. In the dark of night, when you have lights and reflectors, the color of our jersey doesn't really mean much. It's not the best choice, but the fear is overblown.

That study has some flaws, but that doesn't mean that it's far from the truth either. High-viz colors have been proven effective for highway workers, to the degree that it's worthwhile for them to wear but not so much so that anyone would consider a law requiring pedestrians to wear that stuff. I don't see all that much difference in safety considerations for bikers, that you'd come to a different conclusion.

Kontact 

"Certain color of cars involved" is exactly what a "correlation" is. "Correlation" means something similar to "coincidence" - that two things tend to happen together.

Kontact 

We are talking about "these studies", and I'm pointing out what is wrong with them. If you want to talk about different studies, go ahead, but my point is that these studies don't demonstrate much of anything because, for instance, we haven't disproved that the human eye is uninterested in a huge splotch of black that contrasts with the background by being so much darker.


I don't think many people understand what a correlation actually is, or why they don't demonstrate anything about causation. Car color could be the cause, but it could just as easily be a symptom (effect) of whatever the real cause is. A scientist interested in the correlation between color and accident would launch a new study to find out WHY that correlation exists rather than just regarding the correlation as making one cause more likely than another just because.


The general public really seems to struggle with causation when it comes to this sort of study and it impacts our handling of important issues like climate, economy, gun control, etc. And then we support laws that are supposed to fix problems that we misunderstood the cause of because we don't generally understand things like correlations.

The study is interesting, but it tells us nothing about the underlying cause.

squirtdad

Maybe because all the vehicle have lights on in daytime and a whole bunch of vehicle evenly spaced with lights on catches attention? Also the silhouettes are fairly different from average vehicles....

as noted vision is hugely complicated....with people often not seeing what they don't expect to see

KD5NRH

*in Italy*

A YouTube search for "italian traffic" will provide hours of amazement that there are any surviving drivers at the end of the day. Not exactly the ideal bunch to ask about the effectiveness of any single traffic safety measure.

njkayaker

Some of what you offer as being wrong about them make no sense.

Some of what you offer as being wrong about them end up corroborating them.

Yours and Maelochs's "analysis" isn't better than what you criticize.

Kontact 

If any of that was true, you would be able to articulate what the problems with my statements are.

njkayaker

I did (see the earlier post; the thing you quoted).

Some people don't understand how these studies are done.

People look at the data they can get and observe a correlation. They propose a reasonable explanation of the correlation (the article I linked to proposed that part of it was visibility and possibly "driver behavior"). They publish that. That, hopefully, will lead people to launch a follow-up study to investigate the actual causation. The data for the followup study might not even exist.

It doesn't seem likely that "driver behavior" hypothesis explains most of the difference.

The "visibility" hypothesis is corroborated by some of the things you and Maelochs list (yet, you both oddly think they don't).

Much of the problem isn't with the studies at all. It's with the media reporting of the studies.

Kontact 

Alright, so let's go through them one by one.

Of course not, but as black and white police cars show, black is a high contrast color. The study isn't about vehicles that are intentionally colored for visibility, but common colors that have variable visibility. Given how common gray cars are, black may well be relatively high contrast compared to them.

This thread is to discuss these studies.



I don't know that white is any more remarkable to the eye than black - both stand out because they contrast with more muted colors. But since cars come in many medium tone colors that are far less contrasting than either white or black, white or black may both be good choices compared to gray or tan.
The result of the study is a correlation that isn't explained by the study. It is possible that none of the common car colors have ANY impact on visibility. The study isn't able to say because they were not able to rule out the correlation between car color and behavior.


No, I'm not. And if you'd like to point to why you think I'm doing that, I will explain that as well.

njkayaker

So, you didn't read it.

I addressed that.

No, the black serves as a background to the white (that is, the purpose of the black is to enhance the visibility of the white area). And the purpose might not be to increase visibility but to make them distinguishable as police cars.

And we are talking about cars of one color. Multicolor cop cars doesn't mean that black alone is more visible than white alone.

And, this completely ignores the typical colors used in undercover police vehicles!

White reflects light; black absorbs light. This is basic stuff.

White works like those "certain colors" much, much more than black does.

If you'd pick "certain colors", it wouldn't make sense to pick black over white of those "certain colors" were not available.

You do so with the "I would pick certain colors because of studies about the human eye and colors".

Maelochs

I demur. What the study shows is a correlation (and yes, I am native English speaker with many decades of experience but thanks for the rather vague definition) Between NUMBER of Accidents and Color of Cars.

Not anything about the CAUSES of the accidents. We don’t know if people didn;’t see certin colors or the drivers were texting or what.

“There have been various studies over the years into the correlation between car colour and accident rates … ” https://www.moneysupermarket.com/car...ash-by-colour/ Notice ... Not car colors and Causes of accidents.

And this: “The visibility of grey, silver, red and blue cars can also be poor.” Followed by this: “A study in New Zealand also looked at the link between car colour and accident rates.

It too concluded that black cars were the most accident prone. But it found that silver was the safest colour – contradicting the other research.”

Since there were contradictory results, one cannot conclude Based on These Studies that visibility was the sole issue or a factor at all.

And finally: "Of course, colour is only one of many crash factors.

The age and experience of the driver, speed of the car, and the vehicle’s safety features are all much more important in determining the likelihood of an accident, not to mention the influence of drink and drugs."

You can contradict me, but the very article you cite contradicts you. The correlation is between Color and Number of collisions, not Causes at all.

atbman

Judging by some of the comments, there seems to be some confusion between clothing which is highly coloured (i.e. fluorescent - yellow/orange, etc.) and reflective. In the UK, the phrase hi-viz refers to the latter. Bright colours at night are of little or no use, whereas the reflective stuff is.

Also, in the UK, the greatest number of collisions with motor vehicles take place on rural roads and in daylight, rather that in the city, so reflective hi-viz is of no extra benefit, tho' such clothing is also usually fluorescent and it is the latter quality which is of most use in daylight.

Kontact 

I don't know what you think I didn't read since I addressed your points one by one. Again, you are going to have to articulate your objections and not just refer to them generally.


In terms of human vision, your brain does not really care if something is "really light" or "really dark". Vision has all sorts of processes built in designed to make you pay attention to things that are important in your environment. No military uses black for daytime camouflage because it contrasts to much with background colors. Crows are as easy to see as doves.

This may seem like "basic stuff" to you, but I actually studied the neuroscience of vision in college and you are making a bunch of unwarranted assumptions about how the eye works. Look up "edge detection in the eye" - contrast is vastly important for what comes to our attention.

njkayaker

Useful to make this clear. (I was aware of it.)

Reflective stuff that moves quickly (ankle straps, for example) appears to work better that doesn't.

njkayaker

Again, you have to read them. They are specific.

This is an oversimplification.

It wouldn't make sense to choose black over white "because of studies about the human eye and colors".

One reason black "contrasts too much" is because the military objects are large.

In any case, the point of choosing a color isn't likely for bright daytime conditions (where it might not matter much).

The military uses black in many other contexts but not really with the intent of making stuff "contrast too much" or "more visible".

Maelochs

i think i mentioned this above, but i cannot link to a specific study ... don't feel the need ... but pedal reflectors or reflective ankle straps are supposedly the best bicycle identifiers ... the things drivers will notice and identify as being associated with a bike more than solid red, flashing red, amber, whatever tail light.

I cannot imagine that someone could not design, build, and successfully market a device which uses pedal motion to light a tiny LED on a pedal.