Everything I know about the universe I did not learn from newspaper headlines

The new WMAP results have told us a lot about the universe. The basic findings are:

  • The LambdaCDM model — a universe comprised of about 4% ordinary matter, 22% dark matter, and 74% dark energy — passes yet another test. The data fit quite well, and we have some new constraints on the cosmological parameters.
  • There is some evidence that primordial perturbations, the small ripples in density that later grew into stars and galaxies, did not have precisely the same amplitude on all scales. More quantitatively, the scalar spectral index n was measured to be 0.951 +0.015/-0.019 (updated — see comment below), whereas purely scale-free behavior would be n=1. It’s not as statistically significant as we would like, but it’s something.
  • Reionization, the process in which electrons were ripped from ambient hydrogen atoms when the first stars turned on, happened a little bit later than the first-year WMAP data seemed to indicate. This is an important input to our understanding of the “dark ages” between the early universe and the bright galaxies we see today.

All of this is very exciting to professional cosmologists. But consider the perspective of a newspaper that wants to convey that excitement to a popular audience. The data on LambdaCDM are important, but verifying that a known model is still consistent might not seem like earth-shattering news. The information about reionization is new, but early stars don’t quite have the origin-of-the-universe kind of implications that really seem exciting to the reader on the street. But, intriguingly, the slight scale dependence of the density perturbations fits very well with the predictions of the inflationary universe scenario. In this story, the tiny ripples in the primordial universe have their origin in quantum-mechanical fluctuations during the period when the universe is “inflating” (expanding quasi-exponentially at ultra-high energies). Since the expansion rate during inflation does gradually change with time, the amout of such fluctuations gradually evolves from scale to scale. Inflation traces back to the very earliest times about which we can sensibly speak (and long before we have any reliable data), so that is definitely something that could get the juices flowing.

So a lot of stories focused on the support for inflation as the centerpiece of the WMAP narrative. Which is fine, as far as it goes, but needs to be treated with some caveats. First, of course, even in the most generous reading, the purported detection of scale dependence was only at a level of about 3.3 standard deviations, which is not a reliable discovery by most standards in physics. (In particle-physics lingo, it’s “evidence for,” not “discovery of,” which would require 5 standard deviations.) More importantly, even if there had been incontrovertible evidence for scale dependence, that would by no means prove that inflation was right beyond reasonable doubt; it fits well into the inflation story, but certainly doesn’t preclude the possibility of other stories. And finally, it should go without saying that the evidence being discussed is somewhat indirect; it’s not like we’re looking directly at what the universe was doing 10-30 seconds after the Big Bang. (The cosmic microwave background is a snapshot of the universe about 380,000 years after the Big Bang, quite a while later.)

But those subtleties are hard to get across in a few words, and the resulting stories in the press showed evidence of the struggle between conveying the (undeniable) excitement and getting the story precisely correct. Indeed, the tension was evident right in the press release from Goddard Space Flight Center. There’s principal investigator Chuck Bennett, choosing his words with care:

WMAP polarization data allow scientists to discriminate between competing models of inflation for the first time. This is a milestone in cosmology. “We can now distinguish between different versions of what happened within the first trillionth of a second of the universe,” said WMAP Principal Investigator Charles Bennett of the Johns Hopkins University in Baltimore. “The longer WMAP observes, the more it reveals about how our universe grew from microscopic quantum fluctuations to the vast expanses of stars and galaxies we see today.”

Actually, it’s not the first data that allow us to discriminate between different models, although it is some of the most precise data to date. But the idea of “distinguishing between different versions of what happened” is a very good one, and a nice way to tell the story. Sadly, in the next sentence the possibility that inflation is not right seems to have been abandoned, as he speaks with apparent confidence about the origin of galaxies in quantum fluctuations.

This urge to overstate the case is evident elsewhere, as well. In the New York Times we read:

The reason, Dr. Spergel explained, is that the force driving inflation is falling as it proceeds. The smaller bumps would be produced later and so a little less forcefully than the bigger ones.

That, in fact, is exactly what the Wilkinson probe has measured. Dr. Spergel said, “It’s very consistent with simplest inflation models, just what inflation models say we should see.”

Michael Turner, a cosmologist at the University of Chicago, called the results, “the first smoking gun evidence for inflation.”

Here, David Spergel is being very careful to stress that the data are consistent with simple models, which is quite different from saying that it verifies those models are correct. Michael Turner is much less cautious, as “smoking gun evidence” would lead you to believe that the case was closed, which it definitely is not. It’s just very difficult to simultaneously be a cautious scientist and convey an accurate sense of the very real excitement that cosmologists have when examining these data.

If the quotes are ambiguous, the headlines are worse. Let’s face it, “Satellite Gathers Useful Data” wouldn’t sell a lot of newspapers. So many places went for the idea that we had actually observed the extremely early universe, rather than made some observations that constrained theories of the extremely early universe. So we get:

Really, WMAP did not see the origin of the cosmos, any more than seeing an infant is the same as seeing someone being born. But it’s not hard to figure out where they got the idea — the NASA press release is titled “NASA Satellite Glimpses Universe’s First Trillionth of a Second.”

Interestingly, some of the headlines were misleading in the opposite sense, by being less exciting than the truth:

We already have plenty of evidence for the Big Bang! Some more of that would be anticlimactic indeed. And, needless to say, the fact that the universe is expanding is not exactly hot news. I know what they’re all trying to say, but can’t but feeling that if people had a better general idea about what we already know about cosmology, they wouldn’t be tempted to write headlines like this.

I have great sympathy for everyone involved in the process of bringing a story like this to the public — from the scientists working on the project, to the outside scientists who help interpret the results for reporters, to the journalists themselves, to the headline-writers with the unenviable task of squeezing some subtle thoughts into just a few words. But the readers need to take some of these overly enthusiastic declarations with a grain of salt. If you want the real scoop, you have to go beyond the newspaper headlines. For example, by reading blogs.

This entry was posted in Science and the Media. Bookmark the permalink.

44 Responses to Everything I know about the universe I did not learn from newspaper headlines

  1. Sam Gralla says:

    good post, Sean. I like the way you broke everything apart.

  2. pixelator says:

    Off Topic completely, but I thought you folks would like to see this:


    Check out this piece:
    “Spatial Orientation Using Quantum Telepathy”

  3. PK says:

    Quantum Telepathy: Imagine the head lines on that one.

    After just a glance, I would say the science is correct. However, it is very unfortunate that the complete misnomer “telepathy” is being adopted by the quantum information community. It is nothing more than exploiting the non-classical correlations commonly referred to as entanglement.

  4. JustAnotherInfidel says:

    Dr. Carroll–

    This is an excellent post. I think that the reason some sense a growing public animosity towards science is here demonstrated–the problem is that we have to justify public funding of these projects to a public that expects earth-shattering results. It is difficult to convey things like the non-triviality of second generation WMAP data while being careful not to overstate its conclusions.

    Those whose job it is to inform the public have to walk a pretty fine line between keeping laypersons interested and sensationalism. I believe that the same situation happened before, with the debacle over a press release claiming Dr. Hewitt had come up with a way to test string theory–a non-trivial result, no doubt, but certainly not to the extent with which it was initially publicized.

  5. ed hessler says:

    Thanks for this typically clear analysis. While I’ve tended to worry about what I read regarding these findings, I’ve not known enough to “know” much.

    If you have time will you explain what all those white lines mean (I do sort of) but they seem oriented in so many directions that I strubble with them.

    I always hope for more accurate reporting and interpretations from media that have full time science beat reporters but this isn’t always the case.

    Again, I’m grateful.

  6. spaceman says:

    Like Dumb biologist, I am also interested in the age-old question: is the universe finite or infinite? In my opinion, this is one of the most important questions ever asked. I know this question can only be answered definitively if the universe is smaller than the horizon. Unfortunately, I have a feeling that those in favor of the small universe idea will never accept any data which concludes that non-trivial topology, if it exists, must be on a super-horizon scale.

    Having said that, I have several questions related to the finite or infinite issue which I am hoping a cosmologist could help answer.

    1). The low CMB quadrupole is in sharp contradiction with the infinite universe prediction for the quadrupole. Wouldn’t any infinite universe model which tries to accommodate this observation be considered an unnatural stretch?

    2). Luminet et al (2004) and Aurich et al (2005) and others have written highly critical papers regarding the topology conclusion reached by Spergel et al (2004). A lot of this criticism is two-pronged: they basically say that (i) the 1st year sky-maps have too much noise in them for Spergel et al to have reached the conclusion they did, and (ii), the methodology itself is in some way flawed. Who is correct? Do the WMAP 3-year sky-maps have a high enough signal-to-noise ratio for one to look for a topological signiture in them, or, will it take another satellite (i.e. the Planck Surveyor) to resolve this issue?

    3). Do Spergel et al have plans to write a paper to counter the recent criticisms that have been leveled against their “circles in the sky” analysis?

  7. Sean says:

    Ed– The white lines indicate the direction of polarization of the CMB. To date, the polarization that we’ve observed is just what you’d expect, given the observed temperature fluctuations; we hope someday to also detect the signature of relic gravitational waves from inflation. That really would be strong evidence in favor of the model.

    spaceman– I think the best we can say right now is that there isn’t any compelling evidence in favor of a finite universe. The low-multipole measurements are interesting, but by no means sharply inconsistent with an infinite universe. Most cosmologists are disposed against a finite universe with a characteristic size very close to the Hubble radius, as that seems like an unnatural fine-tuning. But we should keep an open mind.

  8. Steve Esser says:

    This was very helpful, thank you.
    As for the balance between excitement and sober reflection on the importance of this, the last line from your previous post on this has stuck in my head as adding a wallop of perspective:

    “What caused inflation, and what are the dark matter and dark energy?”

  9. Rob Knop says:

    News stories are one thing, but when it comes to keeping the public informed of what’s going on in science, I think it usually best to avoid over-emphasizing the latest results.

    Were I to give a public talk about the latest WMAP results, I probably wouldn’t really talk about those results for most of the talk. I’d rather spend the time talking about the Big Bang model, about how we look back in time when we look far away, about what the light of the CMB really is, and so forth. I’d show images from the latest results, and I’d try to do some of the technical stuff at a level that the public could understand, but mostly it would be a Big Bang talk rather than a CMB talk.

    I think it worthwhile to regularly remind the public that the Big Bang is still viable, and that indeed we’re still actively working on understanding it. People who aren’t physicists or astronomers don’t think about it that much, and can be easily distracted by headlines that either are just wrong (based on bad science), or try to over-emphasize and over-dramaticize things which are different from what was expected.

    I had a student in my class who, after seeing some stories about physics beyond relativity during the World Year of Physics, questioned the notion of a constant speed of light. I mentioned it, and he asked, “Wait, isn’t that part of Einstein’s theory? I thought that physicists were now questioning if it was right?” I explained that, well, yeah, we know that there’s something beyond GR, but there are also huge numbers of experiments that verify GR in the regimes where we’ve been able to test it, so any new theory will have to make the same predictions in those regimes.


  10. Rob Knop says:

    My own suspicion is that the low-multipole moments in the CMB, given their orientation, are telling us something that we don’t know about the solar system, but that’s just a very weak off-the-cuff suspicion.


  11. spaceman says:

    Sean, thanks for the reply. So, are you saying it comes down to two probabilities? The probability of living in an infinite universe in which in our Hubble volume the quadrupole just happens to be lower than average, and the probability that we live in a finite universe with a characteristic size very close to the Hubble radius. Both of these cases, if I am correct, seem to require, as you put it, fine-tuning. Is there anyway to quantify which of these two probabilities is higher? In other words, which case requires more fine-tuning or seems more unnatural in light of other astronomical observations and theory?

    Fine-tuning issues aside, there are cosmologists (some of them members of the WMAP team) who claim to have ruled out compact universes, and there other cosmologists who have raised serious doubts the Spergel et al (2004) circles analysis. Is the current WMAP data of high enough quality to test these small universe models?

  12. Jeff says:

    You’re missing a ‘0’ in your spectral index error bars that turn the measurement from promisingly intriguing into laughably insignificant. I prefer the promisingly intriguing, personally.

  13. Sean says:

    Jeff– good catch, fixed.

    spaceman– I’m not an expert on these papers. There is a judgment call here, and most researchers don’t think a universe just the size of our current Hubble radius is very natural, or a likely explanation for the data.

  14. Plato says:

    Why it’s good to read Wayne Hu’s work

  15. NL says:

    Very dumb cosmology question- what is the photonic contribution to the percentage distribution listed? Is it too small to list, or totally separately thought of due to photon number nonconservation?

  16. Sean says:

    Too small to list; about 0.01%.

  17. NL says:

    Anothe lazy question: is that 0.01% mostly from CMB, starlight, or BBR from matter?

  18. Sean says:

    The 0.01% is the CMB. I vaguely recall hearing somewhere that post-CMB photons (from stars and whatnot) contributed a comparable amount, but don’t quote me.

  19. That’s why when I wrote about this, I picked the idea that WMAP showed the first stars formed 400 million years post-BB. I think that’s pretty cool too, would not be discussed by other media, and allowed me to wax poetic a bit about science. 🙂

  20. Rob Knop says:

    I would guess that the number of photons in the CMB is greater than all of the other photons, but I wouldn’t guess that is the case for the energy in photons. But I’d have to dig up a “global background plot” and integrate-by-eye to really make a commitment one way or the other.

    Photons once were very important. The reason they are such a small contribution to the Universe now is that whereas the energy density normal matter goes does as 1/size^3 as the Universe gets bigger (it’s just the volume going up, with the same amount of matter), the energy density in photons goes down as 1/size^4, due to the redshift of the photons.

    (And Dark Energy… well, if it’s Vacuum Energy, the density stays constant. But I’m no expert on what it might be, and Sean is.)


  21. KC Cole says:

    I’ve just assigned this to my science journalism students. We already decided last week that the coverage was awful, but your piece is a wonderful analysis. Thanks! p.s. Turner uses “smoking gun” almost as often as science writers use “holy grail.”

  22. Sean says:

    KC, thanks, I’m glad to hear you enjoyed it. Better public understanding is the Holy Grail of science blogging.

  23. Eugene says:

    We should have a contest to come up with the best, most interesting, and accurate headline for the WMAP2 results.

  24. spaceman says:

    I agree that the press coverage tended to exaggerate the nature of the discovery. I also agree that a flashy headline is more likely to capture the public’s interest.

    Below are some of my thoughts on the new data. This is the title I would have picked for the press release.

    Special Telescope Yields New Insights into Cosmic Evolution

    From what I can tell based on critical reading skills and a trust of the WMAP team, the superbly accurate 3-year results are the product of an exhaustive and painstakingly detailed search for systematic errors and foreground contamination. A number of new techniques were employed to see if the data is of high enough quality to be used for a cosmological analysis. So, the combination of longer integration time and a more thorough analysis assures us the new results are giving us a solid picture/understanding of cosmic evolution. I certainly don’t think cosmology is solved, as there are still mysteries and cosmophenomena that need to be explained, but at least we now have a rough outline of cosmic evolution. I have a feeling that the standard model of cosmology is basically correct even though it may take decades before we fill in all of the details. Think about it like this: we knew the size and shape of the earth before we knew what it was made out of and had it all mapped; similarly, we now almost surely know the size, expansion rate, and shape (i.e, flatness) of the universe even though we do not yet know what is the dark energy and the dark matter. Humanity has little to be proud of these days on Earth, as neo-liberalism allows billionaire tourists to fly into space while billions remain without the basics. Nevertheless, we should be proud of the fact that we’ve come as far as we have in recent years in terms of being able to read the “universe story” in the sky.

  25. LambchopofGod says:

    On top of all that, before we all rush out and declare that 3rd year WMAP supports vanilla LambdaCDM models, maybe we should look at [just to take an example]