I'm a little late talking about this, but check out this study.
More than 5,000 dog experts, including breeders, trainers, groomers,
veterinarians, shelter staff, rescuers, and others completed the survey. They were shown 2 photos of a mixed-breed dog and asked to guess it's ancestry. Then the dogs were DNA tested and the results of the guesses are published next to the results of the DNA test.
Now, I want you to look at the results, but completely ignore the DNA testing. It is not an accurate science, and it is irrelevant. Look at the "Top Responses" column. Every single one of the dogs had at least 5 guesses. In other words, a group of experts couldn't agree that any one of the dogs looked like "predominantly". (In fact, a popular "Top Response" is "No Prominent Breed")
Field identification of a mixed-breed dog is impossible, and pointless. Just like breed specific legislation.