the error bar

about

for:

journalists wanting an independent view on brain science stories in the media
scientists needing a quick briefing on the most popular stories
students at any stage studying the brain and behaviour
and anyone else intrigued about brain science and human behaviour

we will:

provide brain science news and opinion concisely and without much gloss
highlight uncertainty and error in brain science
ask brain scientists about their mistakes, failures and regrets
regularly fail in our mission, but hope to improve some brain science communication

because:

science stories in the media try to give clear answers to clear questions, wrapped in clear stories
and while science journalists search for good science stories, and while scientists try hard to tell them...
... science doesn't always write good stories. the truth can have poor narrative, terrible punctuation; and be full of typoos and crossings out
for relatively unglossed brain science, come to the error bar, have a drink and listen to our stories

policies

with at least 25 brain science news, & ~1000 new research articles published every fortnight (e.g., search for "brain AND neur* AND psych*" here), the error bar needs some filters. here is how articles are selected for the podcast:

open-access news articles (paywalled sources are checked, but most stories published here aren't visible: Times, Telegraph, Wall St Journal, Herald Sun, Toronto Globe & Mail)
news stories must relate to one or more new scientific papers, published in the last ~3 months
open-access science papers - the full scientific paper (or a version of it) must be freely available
pre-prints, dissertations & other non-peer-reviewed papers are included, but conference presentations or research proposals are not
persistantly unreliable sources are excluded. this currently includes: Express, Mirror, Sun
some topics are very common. these are checked, but suspiciously: dementia, sleep, brain implants, dog behaviour
stories about well-known charlatans are excluded. this list currently includes: Elon Musk
stories already covered in previous episodes will not generally be covered again
stories connected to any major public health crises (e.g., COVID) are not covered - better sources are available
finally, stories are filtered by my interest & expertise, preferring stories on how the brain does sensation & movement
here are links to all currently-checked news sources

process

we're often asked here at the error bar "just how do you create a fortnighly brain science news podcast with a few dozen regular listeners?". wonder no more:

Monday 15-17h: check for news articles published since last episode & select stories
Tuesday 08-13h: read, review, summarise; create story pages on website
Tuesday 13-14h: lunch, very often a cheese & pickle sandwich
Tuesday 14-15h: record with audacity
Tuesday 16-18h: edit & mix; adjust level with auphonic
Wednesday-Thursday: listen, edit, upload to anchor.fm
Thursday 18h: check websites, RSS feed, relax
Friday: 08-09h: create wavve trailer & Tweet
...wait 10 days...

article rating systems:

media reports

for each media report covered in our stories, we assign a 4-point rating, representing our opinion of how well the media source has reported the science. we are not saying whether the general claims are true or false (e.g., 'red wine causes cancer') - we are assessing whether the media report adequately reflects the science.

the four levels are:

fiction

stories classed as 'fiction' are those which either have no scientific basis at all, which do not relate to a specific piece of scientific research, or which completely fail to represent the science accurately or adequately. there is no scientific story here.

fudge

stories which 'fudge' are those where there is a genuine piece of scientific research being reported, but the report is highly selective, it distorts the scientific methods, results or conclusions, or it seems to have non-scientific motivations. there is a real scientific story here, but the report fails to tell it.

fair

stories judged 'fair' report a genuine piece of scientific research, and on the whole they do so fairly. there may be some selection, bias or non-scientific framing. there is a real scientific story here, and the report does a fair job of telling it.

fact

stories earning 'fact' status report a genuine piece of scientific research without distorting the scientific methods, results or conclusions. the report may have non-scientific motivations or context, but this is clear and separate from the science. there is a real scientific story here, and the report tells it well.

scientific articles

for each brain science article that we read in full, we assign a 3-point rating, representing our opinion of how well the science has been done and/or reported. this is the kind of scientific opinion that we would give during the 'peer review' process. we are not saying whether the general claims are true or false (e.g., 'red wine causes cancer') - we are assessing whether the science has been done and reported adequately.

the three levels are:

reject

articles that we 'reject' are those which, in our view, have a fundamental flaw in the design or conduct of the study. this flaw means that the article is unable to provide good evidence for the claims made. an example could be the lack of a control condition or group.

revise

articles rated 'revise' are those without obvious fundamental flaws in the design or conduct, but aspects of the analysis or reporting of the study make it difficult to assess the claims made. an example could be missing details of methods, or poor statistical analysis.

accept

'accepted' articles are those without obvious fundamental flaws in the design or conduct, and which analyse and report the data correctly, fairly, and with appropriate uncertainty.