The finding is uncomfortable in a specific way. A German research team pooled 15 datasets covering more than 700 people and concluded that when participants repeatedly picked an option, they later not only kept picking it but rated it as better than alternatives it had never been directly compared against.
The rating is the part worth pausing on. Repetition did not simply build inertia. It appears to have shifted how good the option seemed.
The work, published in Communications Psychology in late November, came from a group at TU Dresden led by Stefan Kiebel, professor of cognitive computational neuroscience, with lead author Ben Wagner and co-author H. Benedikt Wolf. A follow-up from the same laboratory, published in May, has since begun to fill in what the first paper left open.
Nine New Experiments and Six Old Ones
The team ran nine value-based decision-making tasks with 351 participants and reanalyzed six previously published datasets, covering another 350 participants. Participants first learned the value of options within clearly defined contexts, then faced those options recombined into new contexts.
The prediction from standard models is that people carry forward the value of an option. What the data showed instead was that people carried forward what they had done.
Options an individual had repeated more often within a context were chosen more often afterward, valued more highly, and treated with less uncertainty. That held across tasks involving gains and losses, probabilistic and continuous rewards, and comparison sets ranging from two to 66 options.
"This pure repetition can later lead to a particular option still being preferred" in new contexts even when equivalent or better alternatives exist, Wagner said in a statement from the university.
A Cheaper Explanation for Behavior Long Called Irrational
Researchers have cataloged these context effects for years. The usual explanations involve relative value learning or range normalization, models in which the brain stores an option's worth in relation to its competitors and requires fairly detailed knowledge of the environment to do so.
Kiebel's group tested a leaner idea: that the bias comes mostly from a tendency to repeat context-specific actions. They built a hierarchical Bayesian model based on two principles—learning from reward and repeating actions—and then ran it against every dataset using formal model comparison.
That two-principle model outperformed the alternatives they tested, including implementations of divisive normalization, variants of range normalization, and a model weighting internal against external reward signals. In stable environments, reward learning plus repetition was sufficient to reproduce the biased choices. The full paper is open access for readers who want the model specifications.
The interpretation is that repeated actions serve as a mental shortcut. Recalling what you did last time is computationally cheaper than reassessing a decision from scratch, and the brain appears to lean on that shortcut more heavily than models had assumed. A companion study from the same group had already found a comparable repetition bias in a sequential decision-making task.
The Newer Result Is About Whether You Can Override It
The obvious follow-up question, whether a repetition-driven preference can be suppressed when it becomes inconvenient, got an answer in May.
In a second Communications Psychology paper, the Dresden team trained 60 adults on a repeating twelve-item keyboard sequence under a 600-millisecond deadline, then occasionally forced a choice where the habitual response led to the low-reward option instead of the high-reward one.
Computational modeling indicated that models allowing habit strength to grow with the number of repetitions beat models assuming constant habit strength. More usefully, roughly half to 60 percent of participants adaptively reduced the habit's influence when it conflicted with the goal, and exploited it when the two aligned. Reaction times told the same story: fastest when habit and goal agreed, slowest when they clashed.
So the shortcut is not a trap. It is a default that a substantial share of people can override under pressure, though a substantial share apparently could not, which the authors flag as worth investigating in relation to disorders of maladaptive habit.
The Everyday Version, and Where It Stops
Anyone who has ordered the same dish at the same restaurant for years, or bought the same coffee while a comparable brand sat discounted on the next shelf, has run this experiment informally. The researchers point to shopping decisions, habits, and recurring routines as the natural applications.
There is a commercial reading, too. If repetition rather than assessed quality does much of the work in maintaining a preference, then getting someone to choose something a handful of times may matter more than the product's competitiveness afterward.
That said, the effect was studied in constrained laboratory tasks with defined options and rewards. The authors of the follow-up are explicit that a habit built over minutes to hours is not the same object as one built over years, and that their paradigm captures a mechanistic component of habit formation rather than real-world routines. Extending any of this to marriage, careers or where to live goes well past what was measured.
The clinical adjacency is worth flagging without overstating it. Habit-like repetition sits near the mechanisms researchers study in compulsive behavior and addiction, and both papers gesture at that connection. Neither tested any clinical population.
For readers looking for a practical takeaway, the honest one is small: the sense that a familiar option is better may sometimes be a reflection of how often you have chosen it rather than of the option itself. That is worth knowing. It is not a diagnosis, and it does not mean familiar choices are wrong.
Key Questions Answered
What did the researchers find?
That repeating a choice made people more likely to select it later in new contexts and to rate it as better, even when equal or superior alternatives existed.
How big was the study?
Nine new experiments with 351 participants plus reanalysis of six published datasets covering 350 more, totaling over 700 people.
Why is this different from what was already known?
Existing models explained these biases through relative value storage. This work suggests that reward learning plus simple action repetition accounts for them more efficiently.
Can the bias be overridden?
Partly. A follow-up study of 60 adults found that roughly half to 60 percent suppressed the habitual response when it conflicted with their goal and used it when it aligned.
Does this explain major life decisions?
No. The experiments involved structured laboratory choices, and the authors explicitly declined to extend the findings to long-term real-world routines.
Was the brain scanned?
No. Both studies were behavioral and computational, without neuroimaging.
Does this say anything about addiction?
Not directly. The researchers note the mechanism is adjacent to maladaptive habit, but they tested healthy volunteers making low-stakes choices.
