Bad Astronomy

So, About That Huge Discovery Last Year About the Young Universe …

galactic dust
Planck map of the effect of dust on our galaxy’s magnetic field. It was observations like this used by BICEP2 scientists to calibrate their own of the early Universe. 

Photo by ESA/Planck Collaboration. Acknowledgment: M.-A. Miville-Deschênes, CNRS – Institut d’Astrophysique Spatiale, Université Paris-XI, Orsay, France

Well, this is disappointing. A team of astronomers announced last year that they thought they had strong and direct evidence for cosmic inflation in the extremely early Universe. But, after follow-up analysis, it turns out the evidence is far weaker, and it appears inconclusive one way or the other.

Bummer.

OK, so what’s the scoop? First, here’s the announcement from last year when it was thought that the evidence was really good. The details are there.

But here’s a summary: Right after the Universe formed, it’s thought that it underwent an extremely short (to us) period of rapid growth, way faster than the usual cosmic expansion we see today. This event is called “inflation”, and it explains a variety of conditions in the Universe today that would be pretty hard to understand without it.

The problem is, the evidence for inflation is indirect; circumstantial. It explains stuff we already see pretty well, but it also makes predictions about the Universe, too, but we haven’t seen that evidence yet. For example, inflation would have created ripples in the fabric of spacetime, called gravitational waves. These ripples would affect matter in the Universe and that would in turn affect the light we see coming from the early Universe.

Until last year, that effect hadn’t been seen. The big announcement was that, using a telescope in Antarctica called BICEP2 (Background Imaging of Cosmic Extragalactic Polarization), that evidence had finally been seen. It made a big splash.

The problem comes in when the astronomers looked at things that might mimic the signal they were looking for. For example, dust (long, complex carbon-molecules that are much like fireplace soot) floating in space can look very much like the signal BICEP2 was seeking. The astronomers knew this, and used data from the ESA mission Planck to investigate it. Planck measured the amount of dust lying along the direction BICEP2 was looking, and the astronomers concluded the amount of dust in their line-of-sight was low. The signal they saw, therefore, must be from inflation.

And here’s the bummer part: They were using preliminary Planck data. When better data from Planck were released, the astronomers used that, and found that the amount of galactic dust in their view was much higher than they previously thought. That weakens their case considerably.

The bottom line is that they do still see some evidence for gravitational waves affecting the light from the early Universe, but it doesn’t look like it rises to being statistically significant, and it’s certainly not as strong as they first thought.

In other words, they cannot say if they have direct evidence for inflation or not. The new results neither prove nor disprove inflation. They’re inconclusive. The observations may all be the result of dust in our galaxy, or it may be that there is some excess signal caused by inflation. Given the current data, we just can’t say.

On a personal note, I reported on this last year as straight news. I assumed that such an announcement had been vetted and the results peer-reviewed. They hadn’t been at that time. As such, I think it’s up to the scientists making the claim to make that clear, and to be more circumspect in their announcement … just as it’s up to those of us reporting on big news to be skeptical and make sure that the process of peer review has been fully respected. That’s on me, and I blew it.

Mind you, I don’t necessarily have a problem with big announcements that are made when the scientists themselves aren’t sure what they mean. Specifically, I’m remembering the faster-than-light neutrino announcement, when the scientists said, basically, “Look, we’ve investigated this as much as we can, and we know it sounds crazy, but our results seem to say that FTL particles are possible. What did we miss?” They were very skeptical, and were asking for others to pitch in and see what they found. It turns out there was a loose cable in the equipment (yes, seriously), that messed up their timing experiments.

There were missteps in that whole mess, but overall I rather like the way it went. They had weird results, they couldn’t explain them, they asked for help, and a culprit was found. Next!

In the case of BICEP2 it was tougher because the results were very esoteric. Doubt was cast relatively quickly on the results, and I watched as rumors flew. I wanted to write a follow-up and say things may not be as they seem, but it was hard to separate fact from speculation. I decided to wait until there was better evidence one way or another. And here we are.

And there’s a final irony here: This new announcement was made by the BICEP2 team, but their results aren’t yet published. They’ve been submitted to a physics journal, and the Planck data they used will be made public next week. So even these results aren’t peer-reviewed yet!

The difference, though, is that this is not a paradigm-shifting announcement, but a retraction. The bar is set lower for such things, so I feel safe enough reporting on it. If, however, someone else comes along and says the retraction needs to be retracted, well, we’ll deal with that if it comes up.

Science is messy sometimes, and it’s made messier by the need and pressure to announce results … and the need and pressure on some of us to write about them. We all need to be more careful in the future.