Last year Prof. James Lloyd (Cornell) published a paper that cast doubt on the key results of my Ph.D. thesis. And while it was a perfect example of science in action, let me tell you, the process didn't exactly feel super-great.
My thesis focused on studying planets around stars more massive than the Sun. By the time I started my project in 2004 there were about 200 known planets. So finding planets wasn't all that novel. However, finding planets around massive stars was brand new territory. This is because when massive stars are on the so-called main sequence, working 9-to-5 jobs fusing hydrogen into helium in their cores, they are horrible targets for planet searches. Planet hunters actively avoided them and focused instead on Sun-like stars.
At the time, most planet hunting was done with the Doppler wobble technique. Rather than detecting planets by seeing them, we can use this technique search for their gravitational tugs on their stars. For every action there's an equal and opposite reaction, and for the star to tug a planet into orbit, the planet must tug back causing the star to accelerate.
The problem is that massive stars are rapid rotators, and their rapid rotation masks their planet-induced wobbles (their absorption lines are extremely broad). For my thesis I took advantage of the effects of stellar evolution to sidestep this problem. When A-type stars like Vega or Sirius run out of hydrogen fuel in their cores, they move into retirement and become "subgiants." Subgiants are much slower rotators than their main-sequence counterparts, which makes them much better targets for planet-hunting. So a-hunting I went as a grad student, and in the 8 years since we've discovered 37 planets orbiting these "retired A stars." I also found that Jupiter-mass planets are twice as common around A stars as they are around Sun-like stars.
This tantalizing correlation between the commonality of Jupiter-mass planets and the mass of the central star has important implications for planet formation modeling, and it can be used to select targets for future surveys. For example, direct imaging surveys have started to target massive stars in the quest to take pictures of planets.
This brings us to 2011, when Prof. Lloyd noticed a feature of stellar evolution that might result in massive subgiants being exceedingly rare---so rare that my planet search program should contain only a few massive stars at most, rather than the dozens I claimed to find planets orbiting.
As stars evolve, they pass through the subgiant branch of the H-R diagram (see figure below). What Prof. Lloyd noticed is that massive stars move along the subgiant branch much faster than less massive stars. This means that at any given time, such as right now, there will be many more low-mass subgiants than high-mass subgiants throughout the Galaxy. Based on this, he argued that my target stars were not retired A stars, but rather retired Solar-mass stars. Needless to say, this would nullify my big discovery. I wasn't exactly thrilled.
I embarked on a project, along with Caltech grad student Tim Morton and Penn State's Professor Jason Wright, to see if my subgiants could possibly be as massive as I thought. It turns out that there should be a sizable number of massive stars in my survey, despite their rarity throughout the Galaxy. The problem with Prof. Lloyd's analysis is that he ignored a bias known to astronomers since 1922 known as the Malmquist bias.
The Malmquist bias is kinda like the infield fly rule in baseball: all fans know about it, but only a few understand it at a gut level. Dr. Malmquist came across this effect when studying galaxies. At the time, it looked like the further away one looked in the Universe, the more frequently they came across extremely massive, very bright galaxies. Did these more massive, brighter Galaxies dominate the universe long ago, only to be broken into smaller galaxies at the present time?
The answer turns out to be no. The reason there appears to be so many massive galaxies long ago (far away) is that all you can see are the bright ones when they're far away! Imagine standing in a pitch dark, expansive warehouse (don't ask why). Now imagine that people are milling about with three types of flashlights: faint, medium and bright. The further you look across the warehouse, the fewer faint and medium flashlights you'll see, because they're both intrinsically faint and they're far, far away. So at the other end of the warehouse all you see are the brightest flashlights, even if the people and their flashlights are evenly spread throughout the warehouse.
What does this have to do with my retired A stars? Well, more massive subgiants are way more luminous than less massive subgiants. So even though they are rare, we can see massive stars over a much larger volume! So by having a brightness-limited planet search, I have a relatively large number of massive stars on my target list.
The great thing about science is that the truth of this matter was out there available for us to figure it out. That fundamental truth didn't care about my career or my pride or my dreams. As a scientist I had to step outside of myself, set my pride aside, and seek out the truth. If the answer came back that my masses were wrong, then it would have been incumbent on me to correct my previous claims. That would suck for me personally, but science would march onward with new results and new clues about how the Universe works. But now that we've figured out that my results turned out to be correct, the ball is back in Prof. Lloyd's court to either defend or abandon his hypothesis.
Comments
For the record, there is one important element of science in action that John neglected to mention here. The next step is not necessarily for me to respond, it's for the paper to be reviewed by (usually multiple) independent and anonymous referees. The rules of the game say refereed publication trumps unrefereed publication. So if you're keeping score at home, I'm ahead on a technicality, at least for now.
I'm eagerly awaiting the refereed publication, to which I will surely craft a detailed and carefully thought out response.
In the meantime, the Johnson, Morton and Wright paper clearly points out the fundamental issues, and shows, much like my paper, the stars of interest are indeed improbably massive: "Against All Odds". Nor is there any disagreement on why the masses come out to be improbably high, it is because "the high precision of the spectroscopic parameters results in a likelihood term that dominates over the prior and favors higher masses"
So the interpretation ultimately boils down to whether or not the precision of the spectroscopic parameters is truly justified, and if there are any unaccounted for systematic errors between the spectroscopic parameters and the stellar evolution calculations.
I suspect there are, and therefore interpret the paper differently, but what I say at this point matters much less than what the referees say and what everybody else participating in "science in action" figures out as a team effort.
As for your theory that the models are so badly in error as to result in 50-100% systematic errors in our mass estimates, the burden really is on you to present empirical evidence to back that extraordinary claim.