Habits can be great for you, such as flossing consistently to preserve oral health, or they can rather be a problem, such as smoking, or something I struggle with from time to time: doomscrolling.

I think part of the reason I sometimes get caught in the loop of scrolling endlessly is that I actually get valuable information from it, such as a timely notice related to a funding call, or an update from a person I care about. Nevertheless, I also often end up just wasting time at the mercy of endlessly scrolling cat videos and silly dances, which may be entertaining for a moment, but honestly add very little value to my day.

Interestingly, a lot of what cognitive scientists know about habit formation comes from animal studies. This is partly because animal research allows to implement free operant tasks that, because of their long and repetitive nature, humans are usually not very fond of taking part. (Understandably so!) And even if they were open to the idea, it’s usually unfeasible to let a person continuously take part in an experiment for days or weeks.

In addition, since habits are highly dependent on contextual cues and sensitive to the effects of competing activities, a nice and controlled lab environment – as much as we love it – may not necessarily be the best place for learning about how we form habits in conditions that resemble everyday life contexts. This calls for the design of methodologies that allow studying human habits in a way that feasibly allows extended periods of participation, but that at the same time does not preclude contextual cues and competing activities being available. Does it sound like a challenge?

Rani Gera, Segev Barak, and Tom Schonberg (pictured below) designed the “real-world free-operant” paradigm to address these challenges. It allows studying human habit formation through a methodology that allows 24/7 access without precluding contextual cues and competing activities naturally present in everyday life. They introduce it in their paper “A novel free-operant framework enables experimental habit induction in humans,” recently published in Behavior Research Methods, an official publication of the Psychonomic Society.

The “real-world free-operant” paradigm

Taking advantage of the ubiquity of smartphones, the research team implemented the methodology as an app. This allowed participants to be able to interact with the task at will for extended periods.

In this way, participants could self-regulate their level of engagement with the app, while still going around with the activities of their everyday lives. Of course, to ensure a baseline degree of experimental control and participants’ exposure to the app, participants were required to demonstrate an understanding of the task and engage with the app at least a few times per day.

Lastly, the research team made participants’ compensation contingent on the behavior they displayed by using the app, which ensured a baseline level of motivation, and allowed them to manipulate contextual cues related to the payment structure.

The app implemented a gamified task where participants could control a spaceship to collect gold from a fictional planet. The gold that participants collected would be stored in a “warehouse” that would be emptied at the beginning of each day by a “cargo spaceship”, which would then be converted to real money to determine the participants’ compensation.

Crucially, after some days of the experiment have passed, the virtual warehouse would be filled to capacity, thus preventing the accumulation of more gold. This made engaging in behaviors to collect gold less valuable, and, in fact, costly for participants, since the actions to collect gold in the game had a small gold cost themselves.

Technically speaking, allowing participants to freely engage with the app was intended as a period to induce habit formation, and the announcement of the warehouse being full was a manipulation designed to signal that the gold-collecting behavior was temporarily devalued (i.e., engaging in it would not benefit participants’ compensation). Reducing the value of engaging in gold-collecting behavior allowed the research team to explore the extent to which participants behaved out of habit, rather than as a means to reach the goal of maximizing their compensation.

The task is rich with subtleties and clever details, which are too many for this post, so I invite you to explore the online demo made available by the authors. This would allow you to have a better understanding of the sequence of events and about the way that task was implemented as a smartphone app.

Testing the “real-world free-operant” paradigm

Participants engaged with the app for either 4 days or 11 days. This allowed testing whether a longer engagement with the task would result in increased habit formation, as revealed by an increase of engagement in gold-collecting behavior on days when engaging in such behavior was devalued.

An extensive set of analyses was performed, combining pre-registered and exploratory analyses, as well as the calculation of theoretically informed behavioral indices. Since the results are extensive, here we will focus only on two key insights:

• Participants were more likely to respond habitually after extensive engagement with the app (as compared with a shorter engagement).
• Participants displayed differences in baseline engagement rates and their patterns of engagement, which modulated habit expression (as indicated by engaging in gold-collecting behavior on days when that behavior was devalued)
  • Participants with higher baseline engagement rates showed reduced habit expression.
  • Participants with a higher average number of trials completed in a single gaming session, and those with a higher average number of gaming sessions per day displayed a higher capacity to avoid displaying habitual behavior.

The first insight is straightforward: engaging with the task for a longer period of time resulted in increased habit formation. It might seem intuitive, but since demonstrating these effects in humans has been challenging in human studies on habit induction, the authors considered it a key finding that supported the relevance and adequacy of their methodology to study habit formation.

In the words of the authors: “This pioneering research introduces an innovative real-world free-operant approach via a gamified mobile app, paving the way for a deeper understanding of habits in humans and offering a promising method for studying complex behaviors in ecologically valid settings.”

The second finding might seem counterintuitive at first sight, as one could consider that higher engagement in the gold-collecting behavior would increase habit formation. However, it might reveal that participants who voluntarily engaged with the task to a higher degree revealed the operation of motivational aspects. If this were the case, highly motivated participants (i.e., those with higher engagement with the task) would have been more efficient in resisting the urge to engage in a habitual behavior once they encountered a signal confirming that the behavior was no longer valuable.

While there are still many open questions, this new methodology can help researchers understand better human habit formation in naturalistic settings. The authors have made it widely available through GitHub and have implemented an extended number of configurable task features and parameters for making it easy for research teams to adapt it for exploring their research questions.

Who knows, maybe the results of future research using this methodology could also help me and fellow doomscrollers take advantage of social media without getting caught in unrewarding habitual loops?

Psychonomic Society’s article featured in this post:

Gera, R., Barak, S., & Schonberg, T. (2023). A novel free-operant framework enables experimental habit induction in humans. Behavior Research Methods. https://doi.org/10.3758/s13428-023-02263-6