Social media users value consistency over factual accuracy

When social media users come across an unfamiliar topic, they form an opinion very quickly, drawing mainly on information that strikes them as consistent. This is one of the key conclusions of a study conducted by two researchers, including NEOMA’s Venu Puthineedi. The research also reveals that these early impressions remain stable over time. 

During the pandemic in 2020, around 800 people worldwide either injected or ingested bleach in an attempt to treat Covid‑19, despite repeated warnings from health authorities. Numerous posts on social media claimed that this potent disinfectant had therapeutic properties. 

Endless stream of information and limited attention span

This behaviour, which may seem irrational at first, becomes far more understandable in light of the researchers' recent findings. Their study explored the cognitive processes involved when users discover a new subject while scrolling through social media platforms such as Facebook or Instagram. The researchers’ verdict is clear: a cluster of posts pointing to the same conclusion is enough for us to form an opinion, even when the information is false. In other words, we are more receptive to consistency than factual accuracy. 

At the same time, the two researchers draw attention to the particular circumstances in which users navigate these platforms. The stream of content is never-ending. We consume it quickly and casually, with little cognitive effort and scattered attention. And when we jump from one subject to another, we operate on instinct and feeling. It bears no resemblance to deliberate, methodical, rational navigation.

Minds are made up after only five posts

The researchers replicated this casual browsing pattern in three experiments involving 500 adult participants, who were asked to view Instagram-style posts consisting of images and short texts. Some of the content was true, some of it false, but every post focused on two obscure parts of the body: the atlas bone (the first cervical vertebra) and the mental nerve. The aim was to expose participants to a novel subject and clearly demarcate the difference between accuracy and error since anatomy leaves no room for interpretation.

The researchers’ key finding: on social media, the earliest exposure to new content leaves a lasting imprint on our initial opinion. This opinion is based on quick, impressionistic snap judgment rather than any structured research. In the first experiment, exposure to a mere five posts with consistent – but not necessarily true – content was all it took for participants to settle on a stable opinion! This illustrates a trade‑off between our need to make sense of what we read and the limited cognitive effort we are willing to put in.

Once those first five posts have set the tone, any additional content has less and less influence on our initial opinion, even when it contradicts it. It is as though our capacity to absorb new information has reached saturation point.

Even if this initial opinion has been formed in haste, it persists over time

The “five posts” figure should not be treated as a universal rule. Rather, it captures a threshold effect: a boundary between the early information, which shapes our initial opinion, and subsequent content, which has only a marginal impact. As noted above, this mechanism operates regardless of the accuracy of the information. For example, some of the posts in the experiment claimed that the atlas was the sole pivot for head rotation, whereas this pivot actually involves three bones. Yet, these deliberate inaccuracies did not stop participants from forming their opinion after only a few clicks.

The scientists also observed that these opinions stand the test of time. Three days after the first experiment, when asked again, the internet users stuck to their initial position. Likewise, when they encounter posts that confirm or contradict their view, they are more likely to like, share and trust content that aligns with their perspective. Here again, the veracity of the information barely matters; the participants held onto their initial opinion, whether it was well-founded or not. 

To put it another way, the first opinion acts as a default filter: it influences how every subsequent piece of content on the subject is evaluated.

Confirming requires less cognitive effort than contradicting

Why do we place so much trust in superficial clues – a handful of posts saying the same thing – instead of seeking solid facts? Why do we give more credibility to information that is (seemingly) consistent and plausible than to what is genuinely true? 

The researchers believe that the explanation lies in the reduced attention span of internet users as they scroll through social networks. This dynamic encourages us to automatically gravitate towards what is familiar, what matches our existing beliefs and what is easy to interpret because it reinforces what we already know. Contradiction, on the other hand, would require a cognitive effort that we are less inclined to make. 

This limited attention leads to another outcome: when we decide whether to trust a social‑media influencer, we rely more on their profile details (such as their job, background or way of presenting themselves) than on the content of their posts. The same content, regardless of its factual accuracy, is perceived as more credible if it is shared by an influencer whose profile seems trustworthy.

These findings provide practical guidance for putting a stop to online disinformation. The suggestions made by the authors include the following: requiring platforms to check the expertise influencers claim to have (such as being a doctor, teacher or specialist); prioritising qualified sources when a story breaks in the news; and attaching warnings or credibility labels to the first posts published on the subject. But if social networks actually worked this way, would they still enjoy the same success?

Find out more

Venu Puthineedi and Ashish Kumar Jha, Where the Ball Starts Rolling? An Empirical Investigation into Initial Opinion Formation on Social Media Platforms, Information Systems Research, April 2026. https://doi.org/10.1287/isre.2024.1589