Distortion of science in the media - what can we do about it?

Petroc Sumner

What was the overall process of the research?

• A paper was published with poor timing, which correlated with riots in the UK

• This led to an interest in understanding how science is perceived in the media and why it gets distorted

what was the motivation for the study?

A study conducted in Cuba on a certain brain chemical that seemed to be correlated with impulsivity, which ended up being exposed to the media and showed how distorted information can become in the media

• The time of the research being published correlated with riots in the UK

• news and media outlets stated that this paper on impulsivity in men was related to this

• e.g., the sun states nasal spray could be used to decrease this behaviour

• found lots of research that is actually just a correlation, gets incorrectly reported as a cause and gets taken out of proportion

got them interested in how they can put science into the media in a way that does not lead to distortion, etc

why is this an interesting problem?

• in single incidences one news story does not have much impact

• but the accumulative effect of many false distorted studies could cause issues for the public (eg wrong decisions about their health)

• the snowball effect as this distorted information gets out of hand

• even in doctors it can have unfluences

What were the hypothesis?

Press releases were fine, but the exaggeration was occurring in the news or its actually the press release where the exaggeration lies

where in the chain do these issues lie?

How was this translated into an experiment/area of research?

• knew that this does happen in the media, but wanted to find out if there is evidence as to why and how these things actually go wrong

  • found that press releases were the key (used to promote research to news outlets), and are the main route that science gets into the media

  • journalists rely on universities sending them these snippets of studies to provide them with information

  • Due to the high expectations of more and more content in the media, journalists do not have the time to fact-check their information and are not as conscious of taking things out of their correct context

What was the design for the study?

• did the research with a group of undergraduates who volunteered to do a research internship

• source news stories from the last year and their press releases,

• aided by the Russell Group to get a complete set of the press releases

• only things that were related to health, psychology or human behaviour

• so were looking at the original paper, the press release and then the news articles that came from this

• then looked at how much news there was from each press release (what makes a successful press release?)

• decided to code the data into three types of exaggeration

How was the first study followed up?

wanted to run a randomised controlled trial using press offices

• asked them to send their press releases to them before they were send of

• The press releases were then randomised into two conditions

• one condition where very little was changed and another condition where any exaggeration was scrutinised or changed

• then sent back and if they were happy with it they set them out

• then observed what went on in the news

What were the results of the follow-up study?

found the same trend where the press releases influenced the news, when the exaggerations were removed, there was no exaggeration in the news, and they also got the same ammount of news

What three categories were the data coded by?

three types of exaggeration

• generalising from an animal study to humans

• telling people to change the way that they behave

• causal statements from correlational research

What are the pros and cons of this design?

retrospective study from data in the past

correlational in itself (when press releases have exaggerated news, it also tends to)

only measured one way that people receive news (didn’t look at social media, radio, TV, etc)

in the follow-up study there was a limitation of the press releases, finding out what they were doing, so resulted in them removing exaggerations and leading to the control condition also being without exaggeration

What were the results

The general trend was that the exaggerations often lie in the press releases

• more often than not the news follows the press release

• If the press release did not have exaggeration more often than not, neither did the news

• this trend was strongest in the animal and human generalisation exageration

Real world applications?

• got involved in training press offices proper conduct that does not lead to distortion in the media

What can be concluded from the results?

The press releases were being followed more often than not

This is a good outcome as it is easier to fix than if it were caused by the news outlets - motivations are good, however, they are just inadvertently making these exaggerations whilst trying to get the research some attention

An animal may be the highest trend as press offices leave out the fact that Bthey are using animal testing to prevent scrutiny

between 2014-2015 the rates of exaggeration seemed to drop (correlational)

why are these findings interesting?

They also found unexpectedly that those press releases where exaggeration had taken place did not get much more news attention, if any more

This again makes it much easier to deal with these issues

what could be done to follow up from this research?

conduct these studies again

create more training for press offices

How was this research received?

• press offices were not happy with the results and felt they were being blamed