Here Is a Good Reason to Avoid Speaking to Humans When Ordering Food

Photo: Paul Thomas/Bloomberg via Getty Images

If in the last several years you have visited, say, a chain restaurant like Chili’s, or an eatery at an airport like Newark or JFK, you may have noticed something about the waiter who took your order: that the waiter was not a person, but a tablet. The prevailing assumption about this switch to digital ordering has been that it will likely lead to more indulgent choices in food: Tablets, unlike humans, do not judge your taste for Crispy Cheddar Bites or Spicey Loaded Boneless Wings.

So this is why a recent paper in the Journal of Consumer Research is so interesting — because it finds exactly the opposite. Across five experiments, people who ordered from a real, live human were more likely to order junk food than people who ordered with the touch of a button. In one experiment, for example, the researchers staked out a restaurant, where a waiter-slash-experimenter told her customers that they’d been given the opportunity to try a new dessert. They had two choices: chocolate or fruit. Some of the customers ordered by speaking to their human waiter; others ordered with the press of a button. The results showed that 57 percent of those who’d spoken to the waiter picked the chocolate; in contrast, just 29 percent of those who ordered by pressing a button chose the chocolate.

Similarly, in a study done in a laboratory setting, people could choose between a Twix bar and a comparatively virtuous banana. Again, some were instructed to tell the experimenter which snack they wanted, and others indicated their choice with the press of a button. Just like the restuarant study, when people voiced their decision, they were more likely to choose the chocolate — 62 percent chose the Twix when they spoke their orders, compared to the 35 percent who chose the candy when they pressed a button to do so.

This sounds completely backward, but the researchers — led by Anne-Kathrin Klesse at Tilburg University, in the Netherlands — note that similar findings have occurred in studies using the Stroop test, which is sometimes used to measure an individual’s self-control. The names of colors flash on a screen — orange, purple, green — and, sometimes, the word matches the actual color of the text. Sometimes, it does not; for example, the word orange may be printed in green. When taking this test, study participants are to name the color, not the word. Klesse and her colleagues point out that people tend to be worse at this task when asked to name the colors out loud than when they take the test by pressing a computer key.

So this suggests, the researchers argue, that self-control may be easier to regulate when people take the time to press a key; the comparative ease at which we speak, on the other hand, may lead impulsivity to override self-control. An interesting theory, though, as usual, more research is needed before anything can be said definitively. You know what — here, I’ll start, by doing my part and creating a lunch order right now on Maple.

Another Reason to Avoid Speaking to Humans