The magic word is swarm intelligence. We rely on them like when we have to decide on the internet. Reviews from users are therefore used in many places, in books, news or hotels. However, the sentences with “likes” and stars are apparently a significant distortion that a study of economists now precisely quantified.
researchers from Israel and the U.S. for their experiment used a web forum platform similar to Reddit. What site they studied, researchers do not share the “legal reasons” with. Users can discuss it via news articles and review their comments to each other positively or negatively. Approximately 3600 people were active on the site during the five-month study, whose results were published in the journal Science.
basis of the study were more than 100,000 contributions to the debate. The researchers manipulated the first to review randomly and observed how other users have responded with their own judgments. And they did – 300,000 ratings included in the analysis, a
. The pull positive reviews
Here, the benevolence put clearly by: Positive sentences were awarded more than four times as often as negative, particularly if a comment earlier was already positive. Then the probability of a further increase by more than 30 percent higher than in the comparison group. “The little manipulation of a single positive initial result of social influence resulted in a significantly higher overall,” the researchers write.
negative initial assessments did the user sat quietly balancing: you placed for these comments more often a plus, probably to mitigate the bad first sentence. The authors summarize their findings as follows: “While accumulating a positive influence and has a tendency to vote bubbles, negative influence on the amount of users is corrected.”
However, these effects were not uniform in all subjects on: it could be evidence for contributions from the fields of politics, culture, society and business, but not as business items. The tendency to positive reinforcement could also at the polls, stock prices and product recommendations in the network play a role, scientists suspect. Here more research is needed.
track of the news
Help Get hold of free news services:
- everything from the category Science
- Twitter | RSS
- everything from the category People
- everything about behavioral research