Thousands of students prepare for taking the SAT by using the Princeton Review’s test prep services. But some of these students do not know that the cost for taking the Princeton Review’s online SAT tutoring packages is not the same for every customer, as where a student lives will change the price substantially. The ZIP code the student types into the company’s website will determine how much the student pays, from $6,600 to as much as $8,400.
Julia Angwin, Surya Mattu, and Jeff Larson report for ProPublica that a ProPublica analysis shows the company’s geographic methodology of pricing makes Asians almost twice as likely to be charged a higher price than non-Asians. For example, a ZIP code in Flushing, New York, has a 70.5% Asian population. The US Census found that the median household income in that ZIP code is $41,884, which is lower than most, but students in this ZIP code are quoted the highest price by The Princeton Review.
Pricing for the SAT course is based on “the costs of running our business and the competitive attributes of the given market,” says The Princeton Review. It added that it charges the same price everywhere in New York City. The test prep service advertises that its service includes “24-hr Online Tutoring,” and the company says the tutoring is done by tutors who live in the same area as their students and is done in one-on-one lessons online or in person.
“The areas that experience higher prices will also have a disproportionately higher population of members of the financial services industry, people who tend to vote Democratic, journalists and any other group that is more heavily concentrated in areas like New York City,” The Princeton Review’s statement said.
This kind of pricing is perfectly legal, says the company, and any such consequences are unintentional, but the pricing system will probably become more common because of services like Uber, which arrives at its pricing based on computer algorithms. The only difference, according to The Princeton Review, is that the price of its service is determined by geographic region.
A White House report, published last year, cautioned that the “algorithmic decisions raise the specter of ‘redlining’ in the digital economy – the potential to discriminate against the most vulnerable classes of our society under the guise of neutral algorithms.”
The Wall Street Journal reported in 2012 that Staples, when selling online office products, used a variety of prices according to ZIP codes. In fact, it seemed that Staples was also adjusting prices based on the user’s distance from a rival store. But the company was inadvertently charging higher prices to people in lower-income ZIP codes. Northwestern University researchers found that web sites like Home Depot, Orbitz, and Travelocity were manipulating some customers into purchasing more expensive product options. This year, in another study, it was found that Google was sending fewer ads for high-paying jobs to females.
This kind of bias becomes illegal when a business discriminates against certain protected classes, according to Andrew Selbst, an attorney who co-authored a paper entitled “Big Data’s Disparate Impact.” However, with online activities it can be difficult to understand why pricing is different in some areas or to some people.
Christian Haigh, a Harvard undergraduate student, studied sites that had users enter their ZIP codes.
“We thought maybe if you have to put in the ZIP code, they were trying to discriminate,” Haigh said. Today, Haigh and three fellow students are publishing their findings that The Princeton Review’s higher prices correlate to areas with higher income.
Haigh found that the entire New York City area is charged the highest price, $8,400. A large portion of California, all except San Diego, is charged the second-highest price, $7,200. ZIP codes in San Diego are charged the lowest price of $6,600.
ProPublica tested The Princeton Review prices and found that living in a ZIP code with a high median income or which has a significant Asian population meant the prices for the prep service would be the highest.
However, while higher income ZIP codes are twice as likely to be charged higher prices, residents of wealthy Dallas neighborhoods are charged the lowest price of $6,600. Families that lived in areas where there were a large number of Asians were 1.8 times as likely to be asked to pay higher prices no matter thee family’s income. The Princeton Review countered by saying:
“To equate the incidental differences in impact that occur from this type of geographic based pricing that pervades all American commerce with discrimination misconstrues both the literal, legal and moral meaning of the word.”