Joel Achenbach of the Achenblog at the Washington Post is worried about science press releases:
The latter is about our press release on a paper that came out in Science last friday.
Here is the original primary press release on EurekAlert.org.
To get some sense over the degree of "control" the scientists have over press releases, note that the press release spells my name incorrectly, and provides an incorrect institutional affiliation!
That is just for perspective. I've been involved in a number of press releases, some of which have very quitetly vanished into the big pile where most press releases go to die. Some got picked up by the specialist press, lead to short blips on on-line sites like space.com or Astronomy magazine; some got wider press; and some got front pages of major newspapers and in-depth stories on leading television shows, up to and including McNeil-Lehrer and The Daily Show.
It is very hard to predict what will take off, and what will sink. Particularly when batches of press releases come out, like at major scientific society meetings.
It helps if there are no other major news that day (though not totally, when there is big news, media sometimes seeks small "nice news" as filler); it helps to have the Amazing NASA Press Machine in support, although that seems to have been switched off recently, or at least into lower gear. For news about science anyway.
The paper the press release was hinged on is a short report in Science, Raymond and Mandell are joint first authors (Avi lost the coin toss...); and there is a "very soon to be submitted" long paper - Mandell, Raymond and Sigurdsson in preparation - which reports many more simulations in much more detail.
When Science accepted our paper, things happened very quickly. The time to press is short, and, with Science and Nature, the publication time determines the timing of the press release. - This is unlike the society journals, with the Astrophysical Journal, it is merely necessary that the paper have been accepted, which gives a broad window in time until actual publication.
NASA doesn't like this time constraint, they want their press releases spaced out nicely - one per week, even one per two weeks. Rotating between launches, planets, high energy etc. NASA HQ likes control.
It wasn't our fault that the Atlantic launch was delayed and overran our press release, but I don't think NASA was happy about it. HQ that is. To the extent they cared at all.
Now Science rejects most papers at the editorial board review, they have a meta-criterion of "interest level" to their readership, which goes beyond the "correctness of the Science". Most papers who get sent to referees also get rejected.
One doesn't usually plan on getting a paper through into Science.
Especially not theory papers, most of the papers that go through are discovery or experimental papers. Preferably "firsts" in suitably sexy fields of endevour. ie we did no publicity preparation until after we heard the paper was accepted.
Anyway, we were on. The procedure is that ~ 10 days before publication a press package has to be assembled; including text and figures. Science provides this on Eurkalert, on the monday of publication week, under embargo - journalists can read the material, contact the scientists for interviews, but not publish anything until 2 pm that thursday. This was complicated by the holiday weekend in the US.
So... there were actually five press releases: a "long draft" where the authors agreed on what to emphasise thematically and as the key result. BTW I think this is the biggest cause of failure for press releases, inconsistent simultaneous press releases which emphasise the local person's pet point without hitting a broad consistent theme. We had to get usable graphics, which included me commissioning an artists impression from Nahks, including the secret "alien squid" version (which was so totally worth it!).
There was a "short" version of the draft, where we cut as much as possible. It was still too long, because this was a theory paper and we had to set the scene and put in the caveats!
Now the University of Colorado publicity was written by their press office (hence my name...), and since the first author was from there, they lead the publicity.
Science wrote their own news blurb (which I have not actually seen yet, my copy has not arrived and I was traveling and couldn't browse their web site for it).
Goddard wrote a press release based on our material, but since they are a NASA center, they had to get it approved by HQ. Which they didn't get until after the publication deadline!
Penn State decided to conform to the Goddard press release (for various ulterior motive reasons), with just modest edits for local colour. Goddard was late, Penn State was late also.
Now, a lot of the press is just straight republishing of the press release, particularly by online sites.
Some places write their own copy, and there are some very good science journalists out there, but the total number is small. eg this press release got an independent writeup at Seed, NASAwatch, and I hear Nature will do an "in the News" short. Some of the professional magazines will also probably do a writeup.
But most of the news (currently ~ 62 items on Google - including Pravda, al-Jazeera and ABC, BBC, CBC etc, which is moderately good) were straight press release, or cut down versions (eg LA Times and CNN). The BBC did their own take.
The other way a press release takes off is if the wire services run it. This one got picked up by Reuters, but not AP which is probably the key service. They also have a very good science reporter. If AP had run it, given other news of the day, we'd probably have picked up ~ 2-300 mentions in mainstream media, as opposed to less than 100. So, this was successful, but not a "home run"
So, is there a level of exaggeration.
Hm, we like to think not, but as Joel notes we're too excited to be objective. This is our baby dammit.
For this press release, we had some work to do.
We needed to explain a negative - that previous assumptions had assumed this phenomenon did not exist; we then had to sell our models, which we think strongly suggest "it" (formation of stable habitable terrestrials in systems with a jovian in short period orbit) can occur.
We then had to emphasise that this was new and important. That bit goes right up front, because if it doesn't no one will read as far as the second paragraph.
You have to hit the key point right in the first 2-3 sentences and explain why it is news and why it is important. The caveats come later, and if you care you read them there (or in the actual paper...).
We also stick in stuff that looks funny - like "took months on many computers". That actually matters scientifically, it makes several points; one of which is that this is a little hard - it is not a few minutes on a computer - but not impossibly hard.
The core code that did this, btw, is public. The result is reproducible with modest coding effort (or you could ask us, nicely); but the code does not have an efficient parallel version to my knowledge.
You can run different model realization with trivial parallelism, but each run will take you CPU months on a top of the line current CPU.
There was also a subtler theme - the paper is a bowshot in the "Rare Earth" war, within the community. It argues that habitable terrestrials are more common that thought previously, and in a fairly extreme instance, which argues indirectly for things being more common still.
However, ultimately, the paper and hence the press release are based on a theoretical models.
Most theoretical papers are either wrong or irrelevant.
BUT, the model is reproducible, and makes falsifiable predictions on what we think is an important issue, and on an issue where major scientific decisions need to be made in the medium term - like what the target selection strategy for the Terrestrial Planet Finder should be.
So, at some level, a little speculative theory paper can leverage a percentage effort of a multi-billion publicly funded observational mission.
Er, so, maybe someone should check our result.
Ok, next one will be what we actually did, as opposed to what was involved in selling it... promise. Real Soon Now.
- Log in to post comments