My friend and frequent blog commenter, Mike Conn, passed on to me a link from Marginal Revolution that suggested you could determine the quality of an ethnic restaurant by tasting a single, representative dish. Much discussion ensued.

As usual, I have opinions, starting with the opinion the the entire undertaking is rather dubious. However, if we assume that it is possible to judge an ethnic restaurant by one, touchstone dish, I think:

Japanese: For tempura, I agree that one bite tells all. But, I don' t think tempura's a good test. For sushi, I think that one bit of tobiko can tell you if the seafood is fresh and properly stored. Otherwise, I'd say gyoza — surprisingly hard to get right.

Chinese: Must be regionalized. For dim sum, one commenter suggested har gow. I think shui mai is a better test as few restaurants do it well. For northern Chinese, try a noodle dish such as a soft noodle chow mein. For Szechuan, I agree with the commenter who suggested kung pao chicken. For Cantonese, see dim sum.

Thai: Tom Ka Gai

Vietnamese: If it's a pho joint, obviously, the pho. Otherwise, bun cha although I have a strong preference for the style served with whole lettuce leaves for wrapping the rest of the ingredients and dipping into the nuoc cham. I believe this to be the Northern style but I may have my compass points confused.

Korean: Seafood pancake??? Are you kidding??? One commenter nailed it — bulgogi, pure and simple.

Indian: As a commenter pointed out, Indian must also be regionalized. Agree to avoid chicken dishes as a test because it's easy to get the meat tender. Understand the logic of the poster choosing saag ghost (lamb and spinach) but too much of India is vegetarian to use a meat dish as a litmus test. For vegetarian Indian, a chana dish would be my test.

Italian: Commenter was, IMO, close. Virtually any pasta with red sauce. Noodle should be al dente, and "there should [not] be [a] pool of tomato water on the plate." Words cannot describe how much a pool of tomato water beneath a pile of pasta hacks me off. Another commenter suggested linguine with clam sauce. The problem I have with that suggestion is that the quality of clams can vary greatly from batch to batch, even if the restaurant uses a very good purveyor. While it's possible screw up the dish in other ways, too much of the overall quality on any given day is out of the restaurateur's hands for it to be a definitive test.

Pizza: As a reader suggested, crust is a big test but I disagree about toppings: toppings are NOT easy. Too many place apply too much or too little. Proper amounts and balance between cheese and sauce very important.

Deli: I don't like brisket or pastrami or anything else on rye so the test proposed by the comments fails for me. Ham and cheese on wheat would be my test unless it's a sub shop. Then, Italian sub.

Mexican: I'm torn between carnitas or mole poblano.

American diner: Commenter said, "Hamburger." Hear! Hear!

French/Belgian Bistro: Commenter said, "The mayonaise [sic] that comes with the fries for Moules Frites." Are you out of your freakin' mind??? Bistro test is steak frites or duck confit, probably the latter.

British pub: Why would anyone eat in a British pub??? If push came to shove and I had to devise a test for pub grub, I'd say the shepherd's pie. If they say it's made with beef, it's false advertising.

High-end, Michelin-star type place: Commenter said, "I usually know with the amuse-bouche whether the rest of the meal will be any good." Dude, you're clueless!!! The amuse bouche tells you nothing. I've had great amuses followed by so-so meals and vice versa.

Most Mediterranean/Middle Eastern: Kebabs. Simple dishes, nowhere to hide if the ingredients or prep are sub-par.

Salvadorian: Pupusas, especially if they contain queso.


Of course, the best test is probably to ask your server what is the house specialty and to order it!