Psychology says the person whose eyes keep drifting back to one option on the menu or one face on a dating app has basically already decided; a 2003 study found how your gaze locks onto your choice

Think about the last time you swiped through a dating app or stood at a diner counter trying to choose between the burger and the salad. You probably found yourself looking back at one of them before you knew you'd made a choice. That’s no coincidence, according to the study, ‘Gaze bias both reflects and influences preference,’ published in the journal Nature Neuroscience by Caltech scientist Shinsuke Shimojo and colleagues. The 2003 study, with eye-tracking equipment, showed that gaze reliably shifts towards whatever option a person is about to select in the final seconds before the decision.

What the researchers actually did
In the study, researchers recorded eye movements while small groups of participants viewed pairs of human faces and selected which one they found more attractive, taking as long as they wanted before pressing a key to record their choice. No one spoke, and no response was required until the decision was final; the only thing that was measured was where each person's eyes lingered.

At the beginning of a trial, attention was distributed approximately equally between the two faces. But the study found that as the seconds passed, gaze was increasingly biased in the direction of whichever face the person ultimately chose. And in the toughest comparisons, when the two faces were rated equally attractive, that gaze bias was as high as 83% in the moments preceding the decision. They called this the gaze cascade effect because the bias grew over time, like a snowball rolling downhill, rather than staying flat.

Why this isn't just "looking at what you like"
The obvious objection is that people just look longer at what they already prefer, which would not be surprising. The study says the researchers tested this with two control tasks: choosing the rounder face and choosing the less attractive face. In both, gaze bias was significantly weaker and plateaued earlier than in the attractiveness task, where it kept growing until the moment of choice. To rule out simple memory, the same comparisons were repeated a day later with the same participants. Roughly a quarter of the trials resulted in a different choice the second time, and the gaze cascade still occurred in those reversed decisions, suggesting the pattern reflects live decision-making and not simply remembering a previous answer, the study said.

The researchers also tested abstract, computer-generated shapes with no obvious “right” answer to lean on. They found the gaze cascade there, too, and in fact, it was the strongest effect seen anywhere in the experiment, suggesting the pattern is not limited to human faces or social cues.

Then they tried to manipulate it
But the researchers didn’t stop at observing the pattern. They attempted to control it directly. In a follow-up experiment, participants were forced to stare at one of the faces without choosing to, by repeatedly alternating between one face being shown on screen for 900 milliseconds and the other for only 300 milliseconds. Participants were 50 percent more likely to prefer the longer-shown face when the pair was repeated six or more times, even though the faces were carefully matched for attractiveness in advance, the study found. The effect was not statistically significant with only two repetitions. Longer gaze didn’t just reflect a pre-existing preference; it appeared to help create one. This only happened when gaze itself was manipulated: in a control version of the experiment where participants stared at the center of the screen rather than following the alternating faces, the bias disappeared completely.

Does this hold up today?
Good science gets retested, and the gaze cascade effect has largely stood up. A 2023 replication study published in Behavior Research Methods by researchers at the University of Amsterdam and Hebrew University of Jerusalem found the same core pattern in 134 participants using ordinary webcams as eye-trackers, though the effect size shrank somewhat online compared with the lab.

Separately, according to a Nature Neuroscience study, ‘Visual fixations and the computation and comparison of value in simple choice,’ by Caltech economist Ian Krajbich and colleagues, the way people look directly shapes how they compare and choose between options in everyday decisions, such as picking a snack. Together, these results suggest that visual attention and visual preference are more closely linked than most people think.

The model they built was a binary-choice computation in which fixations actively guide how values are compared, rather than simply recording what people already want. Tested on humans with eye-tracking, it could quantitatively account for both complex fixation-choice relationships and several fixation-driven decision biases.

What this actually means for your next decision
None of this means an app or a menu can read your mind, and it would be an exaggeration to say every choice is dictated by gaze alone. What the research does suggest, without overclaiming, is that where your eyes go isn’t just a side effect of a decision made somewhere else in the brain; it seems to be part of how that decision gets constructed in the first place.

The next time you find yourself repeatedly visiting the same dating profile, the same entrée, or the same pair of sneakers online, you may want to take note. That lingering gaze may be your brain in the middle of making up its mind, a few seconds before you're consciously in on it.