Graph of the Day: Hype in Science
It's getting worse...
An analysis of 2,300,000 abstracts from scientific papers shows that scientists are increasingly hyping their findings. The graph below shows the frequency of 35 words indicative of hype from 1980 to 2020.
Curiously, the trend is more pronounced in psychology than in biology or physics.
The graph is from a new paper in Scientometrics by Moritz Edlinger, Finn Buchrieser, and Guilherme Wood, titled “Presence and Consequences of Positive Words in Scientific Abstracts.” Here’s the abstract (an abstract about abstracts):
Abstracts are the showcase of scientific studies, crafted to make an impression on the reader within a limited space and to determine the amount of attention each study receives. Systemic conditions in the sciences may change the expressive norm and incentive scientists to hype abstracts to promote their work and career. Previous studies found that terms such as “unprecedented”, “novel” and “unique” have been used increasingly in recent history, to describe one’s own research findings. The present study investigates the use of valence-loaded scientific jargon in the abstracts of scientific articles. Sentiment analysis with dictionaries specifically attuned to detect valence-loaded scientific jargon was employed to analyze more than 2,300,000 MEDLINE abstracts from the fields of psychology, biology, and physics. Results show that over the last four decades, abstracts have contained an increasing amount of valence-loaded scientific jargon, as previously observed in earlier studies. Moreover, our results reveal that the positive emotional content of abstracts is increasing in a way that cannot be accounted for by the increase in text length, which has also been observed in the same time period. There were small differences between scientific disciplines. A detailed analysis of the distribution of valence-loaded scientific jargon within abstracts reveals a strong concentration towards the end of the text. We discuss these results in light of psychological evidence relating positive emotions with the propensity to overestimate the value of information to inform judgment and the increase in the competition for attention due to a pressure to publish.
The system pushes for this shite