09 June 2015

A question for particle physicists

What does a person have to do to get a doctorate in particle physics these days?

I ask because I am thinking about authorship. I still intend to write a follow-up post to an earlier one about a biology paper with over a thousand authors. Such large numbers of authors are increasingly becoming normal for papers of experimental particle physics.

My understanding of a doctorate is that it is supposed to represent the new contribution of a single individual to the scientific enterprise. I have never heard of any doctoral dissertation that was authored by multiple people.

If the papers in the field of particle physics are typically authored by huge collaborative teams, how can a grad student show that he or she has done independent research? Is that just not considered necessary for certain experimental fields in physics?

Related posts

When does authorship stop meaning anything useful?

1 comment:

practiCal fMRI said...

First up, we have a difference of opinion on what constitutes a good PhD. I have always thought it meant learning the skills to do a literature review, read & interpret the pertinent literature, design an experiment, conduct it, interpret the results, design a follow-up study, and put the whole thing in context. If a novel finding happens along the way, bonus! But placing the emphasis on the destination rather than the journey risks an overall paucity of training that I expect in anyone who has a PhD in a scientific discipline.

So now I can speak to what I have observed in the hundreds of grad students over dozens of fields and departments I've had the pleasure to interact with over the past two decades. Some of my collaborators hail from physics (condensed matter, not quite particle but close*), others from physiology, neurology, economics, even anthropology. It is always risky to generalize, but there are a few trends that are pertinent to your question. In physics the grad students are trained to be totally and utterly self-reliant. They still have to do shop training so they can build widgets, for example. So when the massive teams of physicists are brought together on a project like a particle collider, what you have is a vast number of (generally) competent individual scientists aiming for gestalt. When it comes down to specific tasks there may well be an individual physicist charged with task X. Thus, the opportunities for independent research - how to make task X or device Y optimal, etc. - are generally guaranteed. And the training generally guarantees that collaboration is useful but not essential. Your average physicist doesn't flounder if she doesn't know something or gets stuck. In my experience she buggers off to the literature until she puts me in the shade in my own field.

Contrast this experience with what can happen in some endeavors closer to my day job: fMRI. In some grad programs there can be a tendency to build houses of cards out of black boxes (to mix metaphors). Again, generalizations are risky, but in psychology and neuroscience there is less of an insistence that all black boxes be opened and probed for their contents before use than if a physics grad student was using them. I know personally of many two- and three-author fMRI papers for which the authors couldn't describe the acquisition protocol or many of the processing steps except in the vaguest, hand-wavy manner. There are people doing fMRI as their primary way to study the brain yet they don't actually know how fMRI works. Really. The grad programs and labs who don't allow black boxes unless there is no alternative, stand up and take a bow.

To return to your original point, then, if you cut into a multi-author physics paper more than skin deep you will find many silos filled with many layers filled with small teams, all of which were vital to the successful operation of the experiment. There will be many individually novel contributions. (It may be recognition of these multi-faceted, broad-based fundamental contributions that renders it necessary to have hundreds of authors.) What won't be in doubt, in my experience, is the caliber of grad training during the process. There won't be many passengers. The counter example I already went into, for fMRI. So, I don't see the number of authors on a paper as having but perhaps a fleeting correlation with the input or the training. I'll finish by noting that physics PhD programs don't worry me at all for their level of sophistication in training. But I do have more than a passing concern for many neuroimaging/neuroscience/psych programs, especially those using fMRI.

*I was contacted by a particle physics group from LBNL about measuring paint thickness on the inner surface of a scale model detector to be used for calibration on LHC. I did my best to assist them but I didn't actually collaborate with them past email exchanges. Still, my impression of them was not at all unlike my daily interactions with the condensed matter physics crowd.