TAMPA -- Despite growing awareness of fake online reviews, a new research article finds that consumers still overwhelmingly trust what they read — even when they shouldn’t.
The article, “,” was recently published online in the premier journal Information Systems Research.
The research investigates a key question: Are consumers naturally skeptical of online reviews, or do they tend to believe them? The answer lies in what psychologists call a “truth bias” — the tendency to assume information is truthful unless there’s strong evidence otherwise.
“Our research is among the first to examine how consumers make real or fake judgments of online reviews,” said Dezhi “Denny” Yin, an associate professor at the ±«Óătv Muma College of Business and one of the article’s co-authors. “A better understanding of the consumer perspective is critical, as it is consumers who are the ultimate target of review manipulation.”
Other co-authors include Samuel D. Bond of the Georgia Institute of Technology and Han Zhang of Georgia Tech and Hong Kong Baptist University.
The research conclusions are derived from five experimental studies conducted between 2018 and 2023, in which Yin and his co-authors gave study participants a collection of reviews and asked them to classify each review as “real” or “fake.”
Even when told in advance that half the reviews were fabricated, participants consistently classified the majority of reviews as real.
In one example, participants were shown 20 restaurant reviews and told that only 10 of the reviews were authentic. All the reviews were presented on a single screen, making it easy for participants to go “back and forth” to calibrate their judgments. Nonetheless, they still classified an average of 11.38 reviews as authentic.
“This illustrates the power of truth bias in this context,” Yin said.
The researchers also explored how the tone of reviews — positive or negative — affects perceptions of authenticity. Real-world data from a variety of online platforms shows that negative reviews are more likely to be fake than positive reviews.
However, participants in the studies were substantially more likely to trust negative reviews than positive reviews. “Our findings suggest a striking contrast between reality and perception,” Yin said.
The research has significant implications for platforms and marketplaces that rely heavily on consumer reviews. The researchers argue that relying on users to “report” suspicious content is largely ineffective. Instead, platforms should prioritize identifying and mitigating fake negative reviews and labeling potentially fraudulent content.
They also suggest that interface design can play a role in reducing deception, for example, by grouping positive and negative reviews separately or providing rating-based sorting tools.
Yin and his co-authors also hope to inspire more research that builds on theories of deception and consumer psychology to combat misinformation in the digital marketplace.