Crackpots, geniuses, and how to tell the difference

Over at Download the Universe, Ars Technica science editor John Timmer reviews a science ebook whose science leaves something to be desired. Written by J. Marvin Herndon, a physicist, Indivisible Earth presents an alternate theory that ostensibly competes with plate tectonics. Instead of Earth having a molten core and a moveable crust, Herndon proposes that this planet began its existence as the core of a gas giant, like Jupiter or Saturn. Somehow, Earth lost its thick layer of gas and the small, dense core expanded, cracking as it grew into the continents we know today. What most people think are continental plate boundaries are, to Herndon, simply seams where bits of planet ripped apart from one another.
The problem is that Herndon doesn't offer a lot of evidence to support this idea.

Once the Earth was at the center of a gas giant, Herndon thinks the intense pressure of the massive atmosphere compressed the gas giant's rocky core so that it shrunk to the point where its surface was completely covered by what we now call continental plates. In other words, the entire surface of our present planet was once much smaller, and all land mass.
I did a back-of-the-envelope calculation of this, figuring out the radius of a sphere that would have the same surface area as our current land mass. It was only half the planet's present size. Using that radius to calculate the sphere's volume, it's possible to figure out the density (assuming a roughly current mass). That produced a figure six times higher than the Earth's current density — and about three times that of pure lead. I realize that a lot of the material in the Earth can be compressed under pressure, but I'm pretty skeptical that it can compress that much. And, more importantly, if Herndon wants to convince anyone that it did, this density difference is probably the sort of thing he should be addressing. He's not bothered; the idea that the continents once covered the surface of the Earth was put forward in 1933, and that's good enough for him.

Herndon's book came out with the help of a vanity publishing house and Timmer uses it as an example of why peer review is important—it forces scientists with interesting ideas to actually present evidence and go through a process of answering questions about and explaining holes in that evidence. Even though peer review can be flawed, it's a much better system than not having any kind of vetting process available.
I noticed something else here, as well: The similarities in the way different kinds of badly done science often work. Even though Herndon can't present evidence supporting his theory, he can tell a good story about it. If I'm honest, the idea that, once upon a time, Earth was a gas giant is pretty appealing. As a story. It makes our planet seem more impressive. It gives a sense of a secret history known only to a few. It connects to familiar sounding things: Gas giants and Earth. And, if you don't know all the astronomical background that Timmer does, it sounds plausible.
That reminds me of something Pesco posted about recently: Creationist textbooks that teach kids that the Loch Ness Monster might be a surviving dinosaur and therefore evolution must be wrong. I learned high school biology from one of these textbooks. (In fact, that such arguments exist is one of those facts I have forgotten is not widely known information. My reaction when Pesco posted that story was to think, "Oh, right. I guess most of our readers don't know that already, do they?")
In a lot of ways, the Loch Ness Monster hypothesis is a lot like Herndon's Gas Giant Earth hypothesis. They both have storytelling appeal, especially a great sci-fi hook. They both offer access to secret knowledge. They both propose a connection between familiar ideas—a tactic that makes these hypotheses seem more accessible to lay people than the ideas they propose to replace. They both do a lot of hand-waving and mumbling when you start asking questions about the details.
I think that it can be legitimately really hard to tell the difference between science and pseudoscience. We want to know about the world around us. We often need scientific data to make useful decisions in our lives. But we can't just go out and do all the research ourselves because we have other stuff to do. We're each busy with our own area of expertise and don't have time to become experts in every question we're ever going to need an answer for. Specialization of labor is a bitch like that. At a certain point, we have to trust people who are experts in a given field to tell us what they've learned.
So how do we know who to trust?
I don't think I have a perfect answer for that, but looking at books like Herndon's and those Creationist biology texts, I have a couple suggestions:
1) If it makes a really nice story, ask for the details. (Good science usually makes a bigger deal out of the evidence than it makes out of the story. In fact, that's actually a problem many legit scientists have—they're better at talking about the details and data then they are at telling stories. But most of us respond to stories better than we respond to details and data.)
2) If the proof seems self-evident (i.e., it's just good common sense), ask more questions.
3) If believing the idea will make you smarter than the official experts, be suspicious. Experts aren't always right. But they do know their fields and experience does matter. Chances are, you're an expert in something. Say you knew how to bake pies really well. You'd be pretty suspicious if somebody who didn't bake (or didn't even really cook much) told you that you were making pies all wrong—and that they had a secret pie recipe that was better than yours. They might be right. It's worth taking a look at their evidence. But it also worth being skeptical.
4) If the studies used to prove it are really old, or if there's only a few of them, dig deeper. What looks like truth when you look at five research papers can very quickly become completely untrue when you look at 500. What sounds like a good idea when presented by it's originator can turn out to be terrible when you talk to a few other people. Try to get a sense of what the bulk of evidence is saying.
5) If you're told you can't trust any other sources of information (especially because of Big Conspiracy, or because so-and-so expert is a bad person in other areas of his or her life), be cautious. Replication is a powerful tool. It helps us get past accidental and intentional biases to see something closer to the truth. Suppressing replication is also powerful, because it leaves you with no way to check against bias.
Obviously, all these rules come with caveats. But I think they're a good place to start.

Read John Timmer's full review of Indivisible Earth at Download the Universe