Articles claiming that the Higgs may have been seen were almost certainly wide of the mark. Martin Griffiths explains how the story came about
Popularly known as the God particle, the Higgs boson is the most sought-after particle in physics. It is believed to endow all other particles with mass and is the last undiscovered component of particle physicists’ great theoretical framework – the Standard Model. It is not hard then to see why New Scientist magazine should have devoted a long news story and a leader article in its issue of 2 March to a possible sighting of the particle. And why other publications, including The Economist, should then have carried enthusiastic articles of their own on the subject.
The story was given extra spice because the supposed discovery took place at the aging Tevatron accelerator at Fermilab near Chicago, which is pulling out all the stops to find the Higgs before the much more powerful Large Hadron Collider (LHC) switches on at CERN in Geneva either this year or next (see p7, print version only). But many at Fermilab are unhappy. Although the putative sighting was officially reported at a scientific meeting by CDF, the 600-strong collaboration carrying out the Higgs search at the US facility, it only became known to the media when it was discussed in Web logs, or blogs, by individual members of the group. Judged against the generally accepted standards within particle physics, the events seen at the Tevatron constituted very poor evidence for the Higgs, but that doubt was not made clear in the newspaper and magazine articles. So should physicists be more cautious about discussing science in blogs?
Bump hunting
The CDF detector records the collisions between huge numbers of protons and antiprotons accelerated in the Tevatron. A new particle can be revealed by looking for a particular bump in the energy spectrum of the events recorded by CDF. The specific signature that John Conway of the University of California Davis and colleagues had been looking for was the decay of one of five “supersymmetric” Higgs bosons into two tau leptons, which are very heavy cousins of the electron. These five Higgs particles are postulated by the most popular version of the theory of supersymmetry, which is designed to overcome theoretical inconsistencies within the Standard Model.
Like any search for new particles, however, the process involved the screening out of huge numbers of “background” processes recorded by the detector. For example, the Z boson, discovered in the early 1980s at CERN, can also decay to give a pair of tau leptons, and this process occurs far more frequently than the Higgs to tau decay. But because the mass of the Z boson is well known, it is possible to predict the spectrum of tau-pair masses that would be expected from Z decays. An excess of tau pairs above this prediction could indicate the presence of a new particle, and it was exactly such a bump in the spectrum that Conway saw and reported to a conference in Aspen, Colorado, on 9 January. The position of the bump corresponded to a possible Higgs boson with a mass of about 160 GeV, a value favoured by previous constraints.
Tommaso Dorigo, a member of the CDF collaboration from the University of Padova, Italy, then reported the sighting on his blog A Quantum Diaries Survivor on 19 January. This was added to in a post by Conway himself at the popular group blog Cosmic Variance, in which he described how “the hair literally rose up on the back of my neck” when he saw the bump for the first time. Stoking the fire further, Dorigo alluded to an analysis he was working on of a different decay signature – a bottom quark and antiquark – in which he claimed there could also be hints of the Higgs.
Notwithstanding their excitement, however, Dorigo and Conway tried to emphasize the uncertainty inherent in their results. Particle physicists usually say that they have evidence for a new particle when the bump in question lies at least three standard deviations or “3- sigma” away from a statistical fluctuation (equivalent to a 1 in 800 chance that what they are seeing is merely a statistical fluke), while 5- sigma constitutes a cast-iron discovery (with the probability of a fluke being merely 1 in 3.5 million).
Achieving such high levels of certainty generally requires enormous amounts of data, but is particularly difficult when searching for the Higgs boson because the mass of this particle (or particles) is unknown. Gauging the significance of the results from such a “bump hunt” must therefore allow for the fact that a fluctuation could occur at any mass. Taking all this into account, Conway explained in a detailed follow-up post that the significance of the bump in the tau-lepton search was 2.1- sigma, below the threshold for evidence but corresponding to a probability of only about 1 in 50 that the bump was caused by a fluctuation.
While this at first sounds convincing, both bloggers pointed out that such “2-sigma effects” come and go in particle physics all the time. There are so many different analyses of the data being carried out that 1 in 50 chances will crop up quite frequently (i.e. a 2- sigma effect would be expected to occur due to random effects once in every 50 searches). Furthermore, Dorigo estimated that his bottom-quark data only had a significance of less than 1- sigma, far below what particle physicists would normally consider to be interesting.
Mixed messages
Such statistical nuances are not the easiest thing to convey to the wider world, however, as Dorigo admits. “Particle physicists know that a 2- sigma effect is bound to wash away,” he says. “But that was not represented in the New Scientist article.” However, he does accept some of the blame himself for not adequately portraying all of the uncertainty.
Some of the other members of the collaboration agree that the press articles overstated the case for a Higgs sighting and fear that the affair could damage the collaboration’s image. Conway also points to the danger of a “boy who cried Higgs” effect, with the media having lost interest by the time the Higgs really is discovered.
Another bone of contention was the identification of individuals’ names with the Higgs searches, in a field of physics that emphasises the collective. Conway calls the media “somewhat myopic in this regard, when they exaggerate the contributions of just a few, however compellingly they have recorded their work in blogs”. Meanwhile, Robert Roser, one of the two CDF spokesmen, believes that Dorigo and Conway “did not cross the line” in their blogs. “Perhaps they didn’t make it clear how much of the exuberance was personal as opposed to that of the collaboration as a whole, but that is a small point,” he adds.
Blogs also threaten the traditional ways of releasing scientific results at seminars and conferences. Before CDF results can be made public they go through a lengthy period of internal scrutiny known as “blessing” – something that Dorigo has criticized on his blog before as “baroque, bordering on grotesque”. But even after that stage, the collaboration prefers to control the way in which results are released – for example, earlier this year Dorigo was asked to remove CDF’s new measurement of the mass of the W boson from his blog until after a seminar at Fermilab.
Conway says he is “baffled” as to why the media found the story interesting, but says he is concerned that blogs might lead to distortions. “Any science reporter whose primary source of scientific information is blogs is about as responsible as the student who, writing a paper, draws primarily from [the online encyclopedia] Wikipedia,” he says.
Dorigo, however, is more optimistic about the role blogs have to play. He decided to apologise to the CDF collaboration, in an open letter that he also posted on his blog, but he maintains that he did nothing wrong. He has always hoped that journalists would pick up on blogs as he thinks that “there is a major lack of communication from the experiments to the media”. He even thinks CDF and other large collaborations could set up their own official blogs to supplement their existing “arid, static and unfriendly” webpages. “Why not have a blog where collaborators can write what they think about the results?” he adds.
As the researchers at Fermilab make their last attempts to pluck the Higgs from their data, and results from the LHC begin to emerge over the next couple of years, there are sure to be plenty more bumps and blog rumours, whatever the leaders of the experiments think about it. “The spreading of scientific information through blogs cannot be ignored by those involved in large experiments,” says Dorigo. Indeed, the CDF collaboration is currently discussing whether to set up guidelines for blogging researchers. “The very fact that I am talking to Physics World is a sign that scientific magazines and newspapers have understood that information is easily accessible through well-maintained blogs,” says Dorigo.