Fans of tomato sauce and mozzarella cheese were scratching their heads last week at TripAdvisor’s announcement of the best pizza cities in the United States. Why? The travel web site announced that the top two cities were San Diego and Las Vegas.
Really? San Diego and Las Vegas? As opposed to, say, New York and Chicago?
At first glance, many commentators were perplexed by the ratings. But the mystery can be solved by reviewing the methodology that TripAdvisor utilized to compute the summary data. Apparently, according to the web service, the “top spots were determined based on the highest average rating by city for all restaurants that serve pizza.”
So a “walking city” like New York, with its modest sidewalk pizzerias selling single slices for as little as $1 per paper plate, will inevitably earn a lower average rating in comparison to other cities. After all, it’s difficult for a $1 pizza stand in Hell’s Kitchen to compete against the Wolfgang Puck Pizzeria and Cucina in the luxurious Crystals at CityCenter development in Las Vegas.
Is it appropriate for TripAdvisor to compute its ratings in this manner? From a purely statistical perspective, probably not. And yet the readers of any survey should always remember to understand its statistical methodology before deciding whether to place any confidence in its findings.