that is a truly bizarre binning, and the commenter (on strangemaps) is correct to note that dividing by county is meaningless. actual location plotting would be much more informative.

Meaningless? No, I don't think so. Not precise on a point-by-point basis? Sure.

A 1-2 county region in new mexico is red. About the same area covers the state of Massachusetts in yellows and oranges. Which has had more sightings? It's impossible to say; if the red has 90 and each yellow/orange has the max then Mass is ufo-tastic. On the other hand, there could be as many as 2900 sightings in the red area.

Similarly, is that red blotch in western florida an incredibly dense ufo area, with thousands of sightings, or is it merely accidental that 90 sightings were there, with 80 or so in the surrounding counties? My objection to this map is twofold: the county segmentation is inappropriate for something so continuous, and the color bins, with sizes of 4, 9, 14, 59, and 2,810 respectively, are so arbitrarily chosen it seems manipulative. In terms of achieving the implicit goal of showing what regions of the country have the most ufo sightings relative to other areas, I think it is close to useless, or at least misleading.

Jon: "It indicates the number of UFO reports per 100.000 people by county in the continental US."

I.e., it's per capita.

Oof. Even worse, especially when not overlaid with a population density map. So big desolate counties are overrepresented, and cities are under-represented.

Er, no. It's much much better than if it was an absolute count by arbitrary geography.

The bin sizes may not be even or the same size, but at least their color can be used for comparison. E.g., Lincoln country in Nevada may be desolate, but a higher percentage of their population has reported a UFO than the general population of NYC. That, to me, is not "meaningless."

It would depend on how many sightings are enough to change the color. For example, if one family of four in the Florida panhandle saw a UFO, is that enough to turn it red due to the sparsity of the population in the region and the multiple witnesses to the one incident?

This map requires more explanation.

That's ridiculous. It tells you what it's supposed to tell you: that a certain percentage of the population in this particular county have reported seeing UFOs. Sure, you can't get the whole story without having an idea of where populations are dense, but most of us have a pretty good idea of that information when it comes to the US. Plus, I never heard this complaint when the Red/Blue/Purple county maps came out after the 2004 election, and the same thing applies there.

You guys are being whiny.

Did you make this map?

No, but I don't like contrived cynicism.

even if we accept the counties and the densities, what about the ridiculous divisions between colors?

You really hate this map Jon. I guess I am not suprised given you love of all things maps. I am okay with the map, but would have prefered a 25,000 word essay about it.

I'd like to figure out whether the author of this map chose to deliberately mislead, or whether it was misleading to begin with. And so begins my journey...

Seriously. On fire!

Apparently I'm not the only one who calls this map into question; strangemaps' commenters (well, some of them) agree:

http://strangemaps.wordpress.com/2007/11/01/196-ufo-hotspots-map/#commen...

Indeed, the lack of statistical significance would be the biggest problem.

I wasn't trying to defend the statistical significance w/ the map, but I was trying to defend the county separation, which I still think is OK. But the fact that 1 sighting could make an entire county red is indeed a problem, but that's a problem with the sample size, not the binning.

1 sample making a county red is also a problem. But you don't think it's a problem that all rates between 90 and 2900 are the same color? Or, if that number is because of an anomaly, that the bins below are also of non-even sizes?