In December, Gizmodo did its “Definitive coast-to-coast 3G data test” in which it compared the mobile broadband performance of Verizon Wireless, AT&T, and Sprint in eight cities across the U.S. and then published the results, which came up with a surprise winner: Sprint.

Gizmodo used comparable USB dongles from the big three U.S. cellular carriers to measure bandwidth in eight cities:

  • New York
  • Boston
  • Chicago
  • San Francisco bay area
  • Seattle
  • Austin
  • Portland
  • Raleigh-Durham

As the chart below shows, Sprint had the best download bandwidth in four of the eight cities and was highly competitive in the other four. In terms of national average, Sprint was the clear winner, with just under 1.5 Mbps average while AT&T and Verizon averaged about 1.2 Mbps.

Chart by Gizmodo

The test also measured upload speeds and  AT&T was the winner there with an average of over 600 Kbps while Sprint and Verizon each averaged about 400 Kbps. Of course, upload speed is not nearly as important as download speed for most of the work that the average professional does on the road.

Here is the methodology that Gizmodo used for its tests:

“In lieu of jetsetting all around the country, we FedExed our testing package from one staffer’s home to the next, until we’d hit eight of the country’s biggest cities. In each city, testers were instructed to put the three cards-one each from AT&T Sprint and Verizon-through some pretty rigorous paces. (Note: In case you’re wondering why T-Mobile isn’t represented, the carrier doesn’t yet have the coverage required, nor does it yet offer 3G USB dongles or cards like the ones we used in testing. Next year, maybe.)

The testers chose three to five locations (preferably including one suburban spot). Parking themselves somewhere, they would connect each card to the laptop, running Speakeasy’s bandwidth test five times for each device, and then follow it up with an auxiliary battery of repeated pageload and file download tests, in order to verify Speakeasy’s readings.”

Interestingly, these results contradict a study that Computerworld did earlier in 2008:

Chart by Computerworld

The reason for the difference is probably related to the fact that Computerworld’s tests were not done nationwide but were limited to the northeast corrider and centered primarily around New York City (although Sprint had it best performance in New York in Gizmodo’s tests). Another factor in the disparity could be time. Computerworld’s tests were done at the beginning of 2008 while Gizmodo’s tests were in the second half of the year. And, of course, the final factor could be testing methodology.

For those who have used multiple 3G services, what has your experience been in comparing the U.S. carriers? Join the discussion and take the poll below.

Also, if you currently have 3G broadband access, please run a speed test from your 3G-powered system using or (or, better yet, both) and then post the results plus your location and your carrier in the discussion.