“If you go through a lot of hammers each month, I don’t think it necessarily means you’re a hard worker. It may just mean that you have a lot to learn about proper hammer maintenance.” –Jack Handey
The most common type of question I get asked by people genuinely wanting to know more about the Universe goes something like, “Hey, I saw such-and-such-a-story about some fanciful-sounding-theory, and that could be the explanation for this-weird-thing-that-we-see. What do you think about that?”
Well, here’s the thing.
We’ve got a set of laws of nature, rules, observed and ordered phenomena, etc., that’s the sum total of how we presently make sense of the world. It’s pretty damn good, and it explains the vast majority of phenomena that we know. If we want to learn something new about the Universe, this is the standard we have to check it against!
In other words, that’s what the null hypothesis is: a statement that all the observed phenomena we see can be explained by the laws we already know are true.
When the null hypothesis does the job, there’s no reason to postulate any sort of alternative, or novel, hypothesis; there’s simply no merit to it. Hyping an idea that makes predictions no different from the null hypothesis is not only silly, it’s pointless. If you want to learn something new about the Universe, you need a hypothesis that leads to different quantifiable predictions — even if you can’t yet measure those differences — than the things you already know.
And when the null hypothesis doesn’t do the job, that’s when it’s clear that there’s a gap in our understanding, and that’s where science really gets interesting. Examples of this abound in both our past and present understanding of things.
Back in the 1800s, the “null hypothesis” about the motion of the heavens was that all the planets in our Solar System orbited the Sun according to Newton’s Law of Universal Gravitation. And — to the best of our measurements — this was true for all of the planets (once Neptune was discovered, that is), except for Mercury! That observation was key, as the null hypothesis — that our Solar System consisted of eight planets governed by Newtonian Gravity — no longer held up.
Why was Mercury’s orbit precessing at the rate observed? Three alternate hypotheses came up:
- there was an inner planet to Mercury, which was causing the perihelion advance,
- Newton’s law of gravity needed to be slightly modified; perhaps instead of a 1/r2 law, it was actually 1/r(2 + ϵ), or
- Newtonian gravity needed to be replaced with a more complete theory of gravitation.
As you all know, it was this final option that turned out to be correct, as Einstein’s General Theory of Relativity explained not only Mercury’s advance, but also a host of other phenomena. But it was the observation of Mercury — an observed violation of the null hypothesis — that made the question worth asking in the first place.
And it’s the question you always need to ask yourself. If you want to know whether these nebulae are within your own galaxy, like everything else you know and see, or beyond it, you need something that contradicts the null hypothesis. In the case of this particular spiral nebula, it was the observation of Cepheid variable stars that established this object’s distance to be millions of light years distant, placing it far outside our own galaxy.
And this is true across the board, in all fields of science.
The null hypothesis would be that special relativity holds, and applies to all particles in the Universe. So when we get an observation that contends neutrinos travel faster than light, we are justifiably skeptical. So many observations have supported this null hypothesis over such a long period of time that, to overturn a null hypothesis this good, we’ll require overwhelming evidence that these new experiments do, verifiably, contradict it! (And if they do, it’s terribly interesting, but that wouldn’t be my first guess!)
When someone claims that the new evidence for the Higgs supports Supersymmetry, we need to ask what the laws of nature without Supersymmetry predict: that’s the null hypothesis. So far, the standard model with no Supersymmetry is just as good as it with Supersymmetry, so we go with the null — i.e., non-SUSY — hypothesis.
And when we look at the Universe on the largest scales, from galaxy clusters and supercluster to the cosmic microwave background, we find that a Universe full of protons, neutrons, electrons, neutrinos, photons, and all the other known particles can’t explain what we see. In other words, the null hypothesis is invalid, and so we need something else (i.e., dark matter).
The burden of proof is always on the new hypothesis, the one that postulates a new effect or new phenomenon. This is true in all types of science, from astro- and particle physics to chemistry, biology, psychology and the social sciences.
So when there are claims like this,
Sensationalized claims like this one almost always compare poorly to the null hypothesis, so please pardon me when I don’t comment on every one of them. Most of them don’t get very much attention, and I feel no need to draw extra attention to such claims by publicly spotlighting them just to compare them with the null hypothesis, and show how unfavorable it is.
Comparison with the null hypothesis is a great starting point for anyone who knows what the current state of affairs — and hence what the null hypothesis — actually is. It’s also why I don’t believe that — as you can surely tell — women are underrepresented because they’re somehow inherently inferior to men as scientists and science professors. (The null hypothesis being, of course, that they’re inherently and intellectually equal.)
The contention normally goes as follows:
- More men want to be scientists and science professors than women,
- Men have an inherently better aptitude at math than women at the high end,
- This aptitude is what determines whether they make good scientists/science professors, and
- Sexism and gender discrimination is a negligible factor in explaining the existing gender gap in the profession.
If you want to make this contention — that women are inherently inferior to men at this — you need to establish all four of these over the null hypothesis. (It surprises me every time I see a study done on the second item on the list, as though that somehow settles the issue.)
As to my own personal experience, I’ve seen #1 appear to be true, but only after women encounter some form of sexism and/or gender discrimination. I’ve seen a large number of women just as competent (at the high end) in math and science skills as any of the men, yet who were treated differently by many (not all, but many) of the professors. And I’ve seen what can only be described as a total lack of correlation between item #2 and item #3 on that list. Being “good enough” at math to get your foot in the door is important, but that is hardly the determining factor as to whether someone makes a good scientist or science professor. And the existence of sexism and gender discrimination — particularly in physics, my field — is incredibly well-documented.
I’m not saying that the alternative hypothesis, that there might be some inherent differences between men and women, is necessarily wrong, but I am saying that until you can demonstrate that the null hypothesis is invalid, you’ve got nothing.
Once you get there, we’ve got to figure out what the new, correct explanation is, of course, and that’s not only terribly interesting, it’s how science advances. But if you can’t prove the null hypothesis invalid, you’re not even science yet; you’re just hype.