The readability of scientific abstracts is declining, according to the preliminary results of a major study.
This “very worrisome” finding might hinder the ability of other scholars to access and reproduce research, according to one of the study’s authors, who suggested that the increasing number of scientists contributing to each paper and rising specialisation in science are partly to blame.
To come to these conclusions, a group of graduate students from Sweden’s Karolinska Institute, led by Pontus Plavén Sigray, downloaded more than 707,000 scientific abstracts published in 122 high-impact biomedical journals between 1881 and 2015.
They calculated the readability of each document using two separate measures: one that takes into account the number of syllables per word and the number of words per sentence, and another that measures the percentage of predefined “difficult words” in each sentence.
Their findings, published on preprint server BioRxiv prior to peer review, suggest that more than a quarter of scientific abstracts published in 2015 had a readability considered beyond graduate-level English, compared with just 16 per cent in 1960.
“[T]he readability of science is steadily decreasing,” the paper concludes.
One of the authors, William Hedley Thompson, told Times Higher Education that an increasing use of “general scientific jargon” and rising levels of co-authorship were contributing to the trend.
“Writing collaboratively can be hard when everyone has comments about the content of the article,” he said. “This may mean less focus on writing clearly.”
Mr Hedley Thompson said that the increasing specialisation of science might also be to blame. The findings were “very worrisome”, he added, because unclear writing might hinder others’ efforts to replicate experiments, which is a “fundamental part of science”. It also widens the divide between researchers and the general public.
Commenting on the findings, Jessica Ridpath, senior research communications consultant at Kaiser Permanente Washington Health Research Institute, said that complicated texts took longer to read and made it more difficult for results to be “understood, used and shared by journalists, policymakers and the public at large”.
“No matter how educated a person may be, their time is still valuable,” she said. “Who wants to wade through jargon and long, convoluted sentences if the information can be described in a clearer and more readable way?”
Ms Ridpath said that describing complicated concepts in plain language “is a skill that takes time to hone”. Scientists should consider working more closely with communications professionals, while journals and research funders could set readability standards for publication, she added.
James Hartley, emeritus professor of psychology at Keele University, said that the paper “rightly suggests” that authors needed to use clearer language. But he cautioned that the Karolinska study looked only at the abstracts of papers, which are usually harder to read than the body of the text. Biomedical texts “might be” harder to read than those from other branches of science, Professor Hartley added.
“We might expect the readability of scientific text to get harder over time – as science, scientific language and technical language develop over time,” he said.