Memory is so much more than a storage unit in our minds. The people involved in memories influence what we recall, and, as our study shows, the connections we make between memories.
Our memory helps us learn from experiences and develop new knowledge by integrating and updating information. This process goes beyond recalling individual events; it involves connecting elements from different experiences.
For example, reading about a local park cleanup by a political group in the newspaper and then noticing the park’s cleanliness during a visit might lead you to credit that group. If you notice other parks in your city looking cleaner, you might assume the political group had something to do with that too. Memory can forge inferred connections beyond direct experiences.
Forming these connections is an adaptive process and enhances our knowledge quickly and flexibly. However, these mental shortcuts can sometimes lead to false inferences.
Our research investigated how a preference for certain groups of people influences our ability to make these inferential connections about the world. Previous studies have indicated that information from groups we like gives us better access to our memory. These groups can include anything from a football team or political party to a choir you sing in.
However, before our study, it was unclear whether this phenomenon extends to the brain’s ability to connect information from different experiences to make inferences.
The distinction between liked and disliked groups was based on participants’ own preferences. Our 189 participants were asked to create profiles of “teammates” and “opponents” by choosing faces for them and assigning attributes such as political orientation, eating habits, favourite sports and music preferences. They also completed a questionnaire to gauge how much they liked their teammates and opponents, responding to statements like “I would like to get to know this person better”.
Participants then did a computer task involving a series of events set in various scenes, such as a park, and included everyday objects like an umbrella, presented by either a teammate or an opponent.
Following this learning phase, participants were asked to make inferences by linking the objects presented in the same scene. We observed that information presented by liked sources was more readily connected. Participants inferred the connections between the objects more accurately and with greater confidence. For example, connecting the two objects shown in the park was easier if the information was presented by a teammate.
This indicates that people may prioritise information differently based on the source’s likeability.
Our data suggests that people may flag information from a distrusted or disliked source for cautious handling later, whereas they tend to trust information from a person or group they like. When liked or trusted people present information, participants focus on what is being presented rather than who is presenting it.
Our knowledge often develops from synthesising different pieces of information. Imagine you’re in a new workplace. Even if you haven’t seen everyone together, you start connecting people. When you meet Anna and Maria, and a few days later Maria and Emilia, you might infer that Maria and Emilia also work together.
If we aren’t as good at synthesising memories involving disliked groups, it can hinder our ability to expand our knowledge base. Since information from liked sources tends to align with our beliefs, partisan divides may also shape the knowledge networks of a community. So, the cleanliness of a park is more likely to be attributed to a fundraiser by a favoured organisation rather than a disliked one. This phenomenon may extend across societal debates, including climate change, where your alignment with different groups influences the attribution of causes to events like forest fires.
The results of our study show that this tendency manifests even with neutral information. In real-world situations, where information is often contentious and elicits stronger reactions, these effects could be more pronounced. For example, deciding what new stories count as fake news.
Not only are people who trust the source of false information more likely to remember it, but they are also more able to use it to make new inferences about the world. The fake news can branch out into people’s emerging knowledge.
It’s currently unclear if raising awareness of these biases helps people integrate knowledge from different sources. Previous research suggests that simply making people aware of their biases doesn’t necessarily stop them from affecting their behaviour. Future work will need to evaluate if the same holds true for the new bias revealed in our study.
Even when political divides are strong, people still identify with other groups, such as their hometown or nation. Emphasising these shared affiliations may make it possible to temporarily activate these identities and increase their influence over our thinking. While this won’t diminish the importance of other identities, it can reframe who we consider part of our group. This reframing may enhance our ability to make less biased inferences based on new information.
Our study findings suggest that social polarisation between different groups can partly be explained in terms of basic cognitive functions. Social media posts are visible manifestations of polarisation, but the true battleground lies within people’s brains.
Ines Bramao receives funding from Vetenskapsrådet VR 2019-02455
Mikael Johansson receives funding from the Swedish Research Council and the Marcus and Amalia Wallenberg Foundation.
Marius Boeltzig does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.
