[uoaei]

> Roundabouts generally require a driver to look in a direction other than their direction of travel, to determine if it is safe to go.

True for stop signs... and any other traffic control solution in existence, frankly. Yielded merges are the most obvious form of specifically unbroken traffic flow that requires the same.

> They are also designed to reduce the amount of time any given vehicle has to actually stop if there is no reason to.

So it works like a yield. That's a good thing because it reduces congestion.

> This varies based on left/right side driving countries.

There are much bigger issues resulting from switching between left- and right-side driving standards which don't have anything to do with roundabouts, so this doesn't say anything about roundabouts so much as the difference in standards.

> Lights for pedestrians might help, but then the benefit of continual traffic flow is reduced.

Comes with the territory, and is also true of every other traffic control solution in existence. The complete solution to this is to completely separate pedestrian and vehicular traffic, which is also not limited to roundabouts.

It sounds like you don't have a problem with roundabouts so much as traffic control per se, if you believe these to be reasons not to implement roundabouts.

[dataflow]

>> Roundabouts generally require a driver to look in a direction other than their direction of travel, to determine if it is safe to go.

> True for stop signs...

Not really. Stop signs make you stop first, before needing to look around, after which you continue. There's no "direction of travel" when you've stopped; you're not traveling when you've stopped. Meaning you can focus on one thing at a time, unlike with a roundabout.

[uoaei]

This seems like an argument without a point.

[dataflow]

Huh? The point is that your attention (and vision) isn't nearly as divided when driving with a stop sign than with a roundabout. You don't have to multitask nearly as much; you do one thing at a time. Less division of attention = less car accidents.

[uoaei]

It doesn't sound like you're familiar with any of the studies concerning the safety of traffic flow if you think that that is the only factor that determines intersection safety.

[dataflow]

> It doesn't sound like you're familiar with any of the studies concerning the safety of traffic flow if you think that that is the only factor that determines intersection safety.

I'm pretty sure "the only factor that determines intersection safety is division of attention" was not a stance I was taking (when is an event ever a function of just 1 variable in the real world?), but if you'd prefer to take a swipe at me regardless, it would be nice if you could make your response constructive and actually link to some studies that show I said something contrary to reality, if you're well-studied in traffic flow.

Hopefully my reply clarified for you what my earlier point was, even if you think the point was wrong.

[uoaei]

The implied point of the above comment was that, while specific linear relationships can be hypothesized, and shown in certain limited controlled experiments, the nonlinear interactions with every single other factor makes such an assertion not only meaningless but also misleading.

"Well-studied", no, I would not claim that. I follow the zeitgeist of urban planners as they discuss topics like this on the fora in which they congregate. I read some of the studies they post and discuss, and I have picked up on some of the memes present in that community. Something that comes up again and again and again and again (and which is immediately apparent upon reading the research) is the primacy of roundabouts for intersections, because they are the safest of the popular options and less prone to congestion than stop signs, stop lights, etc. Even when they do get congested, the outcomes are better for everyone on average, since everyone waits a similar amount of time compared to the asymmetry of e.g. intersections with stoplights. One common pitfall in thinking about these things is only thinking about individual wait times in a subjective sense, and not aggregate wait times in a systemic sense. The latter perspective provides much clarity.

Here, I did your Googling for you. Where do I send the invoice?

https://www.iihs.org/topics/bibliography/ref/1248