Every day, we are bombarded with a deluge of information, especially online, where news spreads faster than ever. But what kind of information do we choose to share, and why? A new study published in Scientific Reports explores this, revealing that people are more likely to share information that aligns with their beliefs, even if it isn’t particularly surprising. This challenges the common notion that novelty and surprise are primary drivers of information sharing.
Information theory posits that the value of information is largely derived from its novelty or the surprise it generates. The more unexpected the information, the theory suggests, the more likely it is to be shared. Previous research has often found a link between the novelty of information — such as fresh insights on health or political issues — and its rapid dissemination.
However, this study introduces a twist: the alignment of information with personal beliefs might play a more critical role than previously thought, prompting the researchers to investigate how belief-consistency influences sharing.
“Some prominent research suggested that false information spreads farther and faster than true information because the false information is more novel/surprising,” explained study author Jacob T. Goebel of Ohio State University.
“Yet, this work did not seem to account for the extent to which information might fit or not with the beliefs of people deciding to spread the information. We wanted to account for both belief-consistency and novelty/surprise to examine in a more controlled way whether surprise was playing the key role attributed to it.”
The researchers conducted two primary methods of investigation: an analysis of Twitter data and two controlled experiments. For the Twitter analysis, they gathered data from 223 tweets from politically neutral news sources and tracked how frequently each tweet was retweeted. They examined the timing of the tweets relative to when the events they reported happened, using this as a measure of novelty. They also analyzed the ideological alignment of the tweets with the beliefs of the users who shared them.
The first experiment included 226 undergraduate students from an introductory psychology course at Ohio State University. They were tasked with the role of an editor’s assistant at a news organization, where their job was to decide how much information to forward to the editor for a comprehensive understanding of an issue. Initially, participants were provided with an interview transcript discussing the effectiveness of risk-taking in firefighting. This transcript was designed to manipulate the participants’ beliefs about the topic, with the fire chief’s statements and supporting evidence either favoring or opposing risk-taking.
After establishing these initial beliefs, the participants received an “update” from a reporter that presented additional evidence either supporting or contradicting the initial information (belief consistency manipulation). This update also included information that was either previously seen in the transcript or new, thus manipulating the novelty of the information presented. Participants then assessed the novelty and surprise of the update, and decided which pieces of information were critical enough to pass along to the editor.
The second experiment was similar in design but involved 301 participants recruited from Amazon’s Mechanical Turk platform and focused on a different topic. In this experiment, the subject matter concerned whether a hypothetical country should be allowed to join the European Union. Like the first experiment, participants were initially exposed to information designed to sway their beliefs in one direction or the other. They then received an update containing information that either aligned with or contradicted these newly formed beliefs, and like before, the novelty of the information was manipulated.
Participants in Experiment 2 were also asked to rate the value of the information in the update before making decisions about what to share. This addition aimed to gather more detailed data on how perceived value influenced the decision to share information.
The researchers found that tweets posted sooner after an event — thus more novel — tended to receive more retweets. However, the ideological consistency of the tweets with the users’ beliefs emerged as a strong predictor of sharing behavior. Tweets that aligned with users’ political ideologies were shared more frequently than those that did not, regardless of their novelty. This suggests that while novelty attracts attention, congruence with pre-existing beliefs substantially influences whether information is considered share-worthy.
“Encountering a sea of information each day, people must decide what is worth sharing with others,” Goebel told PsyPost. “People don’t always share information that is “fresh” or “new”; they also react to how it fits with what they already know. Often, this leads them to share information that is consistent with their prior beliefs (even though that information is not very surprising to them). It is not safe to assume that information will be shared simply because it is novel and/or surprising.”
The two experiments reinforced these results, showing that information consistent with participants’ beliefs was not only perceived as less surprising but was also more likely to be shared. Interestingly, the novelty of the information did not significantly affect its likelihood of being shared, contrary to what traditional information theory would predict.
“When we manipulated whether potential information to be spread had or had not been seen previously, participants were not sensitive to that manipulation when judging how novel (or surprising) the information was,” Goebel explained. “In other words, people didn’t seem very adept at discerning how objectively novel the information was.”
As with any study, the research includes some caveats. For instance, the experiments were conducted in controlled settings with hypothetical scenarios, which might not fully capture the complexities of real-world information sharing.
“The experimental work created new beliefs in order to have strict control over whether the information to be spread had or had not been previously seen (in addition to whether it was belief-consistent or not),” Goebel noted. “That control is useful for making clear claims about these particular studies, but it leaves open the question of whether similar effects occur when new information is related to long-held existing beliefs.”
“Also, because the sharing occurred in the context of a scenario in which information was to be shared with an editor of a publication, different roles ascribed to the participant or to the audience might shape what kinds of information would be viewed as valuable.”
Future research could explore how these dynamics play out with different types of information, such as health advisories or political news, where individuals might have stronger pre-existing beliefs. Additionally, examining the impact of real consequences associated with sharing misinformation could provide deeper insights into how beliefs influence sharing behavior.
“We would like to examine how group and network contexts might shift sharing tendencies, perhaps enhancing or changing the kinds of effects we observed for individual decisions,” Goebel said. “For example, a network of friends might call for sharing of different information than a network of scholars or of political affiliates (or foes). These network/audience contexts might also shift the roles of novelty, surprise, and/or belief-consistency.”
“There are other contexts in which surprise draws interest or prompts processing of information one encounters,” he added. “Thus, it seems likely that there might be contexts in which surprise also makes incoming information seem worth spreading to others. It could be that, in many contexts, however, people use consistency with existing beliefs as a key factor in determining whether information is worth sharing.”
The study, “Belief-consistent information is most shared despite being the least surprising,” was authored by Jacob T. Goebel, Mark W. Susmann, Srinivasan Parthasarathy, Hesham El Gamal, R. Kelly Garrett, and Duane T. Wegener.
