Science and Informal Communication

In recent days the Nutty Ones have been in a frenzy about something they call "Climategate," a supposed controversy that validates all their delusions about global warming. Somebody hacked into a server and stole a bunch of emails among climate scientists and published them online, and now some people are asserting that the emails prove that global-warming science was faked.

It's a nice Sunday, sunny and breezy, I should be outside but instead I'll sit here at the computer and ruminate for a while with the windows open. I am thinking about this Climategate business. Specifically I am thinking about the problem of unqualified people with political motives evaluating scientific research. This piece is too long, I am aware of that, I don't expect anybody to bother reading it. I just needed to spill my thoughts about this -- maybe it will make you thankful you aren't doomed like I am to think about things like this. I'll try to find something sexier tomorrow.

Science is just a word that means knowledge.

As ordinary people, most of what we know is stuff we have been told and not seen first-hand. You have not seen the curvature of the earth's surface with your own eyes, for instance; from an individual's perspective the earth is flat, and the sun and stars move over us while the reliable surface we stand on remains firm -- but you know otherwise. Your knowledge is inconsistent with your perceptions. You have probably never seen the President of the United States but you sure know a lot about him and have beliefs about him. With our own senses we see a little bit of the world, and -- this is the amazing thing about human beings -- we put together reports from each individual point of view to compose something we call knowledge. Knowledge is different from information, it is different from accumulated data; computers can't know, only living human beings can have knowledge.

Sometimes our knowledge is wrong. In fact, if you look back at human history and prehistory, most of the time, most people have been wrong about most things they have believed about the world. The cool thing is that over time human knowledge about the world has evolved in a positive direction, facts are separated from unfounded rumors through some process of testing, validation, and discussion, and after several tens of thousands of years of homo sapiens sapiens we are relatively sapient.

It seems funny to say that knowledge can be wrong, but you have to realize that knowledge is not an entity separate from human minds, knowledge is in fact the state of a mind. Knowledge is a little bit about the truth of a belief and a whole lot about the certainty of the believer. It is entirely possible to hold a belief about which you are certain, and be wrong. Everybody understands this, really, and over time what we try to do, if we are honest, is to improve our understanding, we want to replace beliefs that correspond poorly with the state of the world with better ones. One reason, maybe the biggest reason, we want to do this is so that other people can trust us when we communicate information to them. The dependability of our statements lays the foundation for our social bonds.

And here's where science comes in. We like to say that science is a process of hypothesis formation and testing, experimentation and analysis, but sometimes it's better to think of science as a social process. Science works like this: a person has an idea, does some work, and writes a paper describing the work. The paper is submitted to a journal where an editor decides whether to send it out for review, and if it looks good he or she may send it to some individuals who they consider qualified to review the paper. The reviewers read it, comment on it, and rate it. The usual rating system is: 1. Accept as is; 2. Minor revisions and accept; 3. Rewrite and resubmit; 4. Reject. Usually three or four reviewers look at a paper, the editor takes their opinions and decides whether to publish it. Editors may totally ignore the reviewers' comments and ratings, though they usually don't. Nobody knows, usually, what the other reviewers have said, if the editor gets three rave reviews and one it-makes-me-want-to-puke he may decide to publish, and if you wrote the puke review you might feel you have been ignored, but it's all up to the editor. He may look at the negative review and decide not to publish, or he may read the article himself and decide regardless of what the reviewers have said.

Once a peer-reviewed paper has been published in a reputable journal it becomes part of the scientific literature and can be cited by future scientists. Newton called it "standing on the shoulders of giants," and that's how science works.

Listen, this is a fallible process. Science can make mistakes. There have been important moments where the scientific community chose to go one way or the other, and maybe they didn't choose correctly. Was Newton's theory of color better than Goethe's? We'll never know, because nobody paid any attention to Goethe's theory. It might as well be VHS versus Beta, eight-track versus cassettes, all that happens is that some people decide what they prefer. Except in science the people have been selected through a process of education, which is as much a matter of selecting qualified individuals as it is filling their heads with accepted knowledge. You go to graduate school to get your PhD, you take classes, seminars, conduct research, write a dissertation, take comps, but in the end it comes down to one thing. A professor shakes your hand and says, "Congratulations, doctor." It doesn't matter if you have failed to pay your fees, it doesn't matter if you flunked comps and nobody liked your dissertation, it doesn't matter if your paperwork got lost in the administration office, once they call you "doctor" you are a doctor, a peer, you are one of them, and they can't take it back. A PhD means nothing more than that you have been accepted by an elite group of scholars. You could have a perfect dissertion and get everything right in your exams, but if they don't accept you into their group you do not get to use the letters after your name.

Science is no different from other forms of knowledge. You can ask about the fundamental nature of matter, you can ask if John Gosselin cheated on his wife, it's all the same process. It all works the same way, people have a thought and share it with others, it gets discussed and passed around and refined, and in the end everybody knows that the atom is the smallest unit of matter, or that John Gosselin did or didn't cheat on his wife. And for the record, I have never seen John Gosselin on TV and have no idea who he is, I just know his name and face are in the news a lot. He has a bunch of kids and either he or his wife is having an affair, right? Scientific thoughts are scrutinized much more carefully than thoughts about television stars, scholars risk their reputations when they write or review something, their skepticism and criticism is much more intense than you see in an ordinary conversation, the scientific process is more formal than ordinary conversation but is not fundamentally different from it.

So some scientists sent emails to each other about global warming, big deal. They're people. I remember in graduate school sitting in seminars with a bunch of research assistants where everybody would cheer if the results came out "our way" and boo if they didn't. We were all happy if an experiment went well and we had a good chance of getting the results published. This was a little bit before email was common so you can't steal those hopeful comments and put them on the Internet, but we talked to each other about the studies, we hoped it would come out, we shared ideas about the hypothesis, the methods, the statistical analysis, we might talk about how to assign the degrees of freedom in the ANOVA to get a significant result, what to do about outliers, we might have noticed confounding factors or conflicting results in another paper, or anything, we might hope the results come out badly because the professor gave us a bad grade on something or because he was on the outs with the other professors. We talked about the scientists with rival theories, dwelling on their unclean personal habits. Scientists are just people, motivated in the same ways everyone is motivated. Nowadays they use email, and somebody broke into the server and stole it.

I can talk about this because I've been part of it. I am an originator of a scientific topic that has been the subject of thousands of scientific papers. I have personally seen the birth of an idea, shared the results with other scientists, presented at conferences, published in journals, edited journals and reviewed hundreds of papers, there are now entire journals committed to the subject I originated, I have seen the paradigm veer off track and come back, I have seen popular novels based on the subject, management consultants citing the research ... I am in a privileged position to comment on the social process of science.

Scientific knowledge is essentially the same as all other knowledge, it is the state of a thinking mind, what cognitive psychologists call the "feeling of knowing" is a matter of perceptual fluency, speed of processing, salience and familiarity, we say we know something when we feel like we know something. Scientists are no different from other people, they do everything the rest of us do. But amazingly, science as a formal system has been astoundingly successful at evolving the quality of knowledge, it is the best process the human species has found. Non-scientists might think there is something pristine and perfect going on in those ivy-covered buildings but it's just people figuring out stuff and talking about it, it's trial-and-error group learning. They specialize in tiny niches -- I met a guy once who studied hypothermia in ruby-throated hummingbirds, that's it, that was his life. Trust me, he knew more about that topic than you or I do, he knew it in a way we can't imagine. You couldn't read his email to his friend and conclude his research was a hoax, even if he was discussing his research design or statistical analysis.

It's not quite right that scientists are exactly like the rest of us. A scientist is a person who has first of all been accepted to a graduate school, second of all gotten through the program, and finally has conducted research that met the approval of his or her peers. Getting into graduate school is an exercise in ass-kissing, you need to have good grades as an undergraduate, good scores on the GRE, and you need impressive references and a nice letter that shows you to be sincere and intelligent and ambitious in ways that please the faculty. Anybody who has been to grad school can tell you about people who got in for the wrong reasons, and getting in is the most important thing, the hardest thing. Getting through it is a lot of work but basically once you're there they'll help you unless you're lazy, and then they can't be bothered. You learn the rules of the scientific method in fine-grained detail, you submit some papers as a co-author with your professor, whose reputation increases the chances of papers being accepted, and when you get out you will hopefully have had your name on enough of them that an editor will recognize your bloodline when you submit one as first author. And so it goes, you continue the tradition and make a name for yourself.

Scientists are different from other people because they have been selected by the academic process to become doctors of philosophy and have stepped over the thresholds of that process. If you look at scientists individually you will see that they stand out, there are not going to be a lot of two-digit IQ's in that crowd, for instance. These are smart, hard-working people. If they are successful it means they have learned how the game of science is played and have used their intelligence and motivation to place themselves prominently in their field. They have figured out how to get positive reviews from journal referees, how to get funding for their research, how to get tenure.

None of this means that scientific results are "factual." Let's make a distinction here, you might not be used to thinking this way but let me point out the distinction between deductive and inductive logic. Mathematics is deductive. Ignoring Gödel, a mathematical fact is true no matter where it is applied or who thinks of it. The method used for identifying mathematical truths is the proof, which is a strictly defined series of logical steps, each building on the previous one, where the previous one has its own proof, and the final result is deductively valid and true. Mathematical facts are true even if there is nothing in the real world that corresponds to them, even if no one has ever thought of them before.

On the other hand, inductive reasoning builds up facts from observations. The method for determining causality inductively is the experimental method, where factors are carefully controlled, an independent variable is presented in two or more levels, the value of a dependent variable is measured, and values of the dependent variable are observed to be affected by values of the independent variable. There is almost always some error in measurement of the dependent variables and other factors affecting the outcome, and so some statistical analysis is used to determine whether the outcome is due to chance or whether it is due to the experimental manipulation. Based on all this fudginess, a conclusion can be drawn about the condition of the world with some degree of certainty. Through some clever statistical manipulations, the certainty of a conclusion can be pretty well estimated.

Science, no matter what anybody says about null hypothesis testing, uses inductive logic. Scientific results are never completely certain, if they were certain they would be common knowledge already, the results of an experiment almost always have some room for doubt or something that can be challenged. That's why science doesn't just wrap it up and say, Okay, we're done. The cool thing is that knowledge is always increasing, beliefs are improving, because people selected by the academic process use methods that their peer group approves of to write papers that will be published and get them funding and tenure.

So now we have the shocking revelation that some climate scientists have sent emails to each other, hoping their research will look good to journal editors and department chairs. Woo, scientists are people, who would have guessed? It's a fake controversy, it's like finding out that priests have sexual fantasies, that bankers are greedy, that teenagers are self-conscious. Sure, climate scientists get excited when they look at a graph of data that no one has ever seen before, and notice a trend that has never been documented, their first thought is that maybe Science or Nature will accept this, and if there is something imperfect in the data they will see if they can improve it. They email their friends and colleagues and discuss it, not in terms of how it will read in the inevitable Wikipedia page that will be quoted by the inevitable anonymous troll, but how it will affect their career. What can they do to make this new-found discovery palatable to journal editors and reviewers, how should they describe it so they sound like they are supporting and extending the prevailing paradigm rather than looking like some cranky eccentric? Of course they talk about that, just like you and I send email to our friends about things that matter to us. Except nobody steals our emails and puts them on the Internet where we will look stupid and human.

The current case, climate change, matters because there is a political angle to it, because there is money involved. There is a bizarre alliance between the wealthiest capitalists in our society and the hardest-working, lowest paid, least educated citizens. The capitalists understand that climate science could cost them money. If their factories and products are destroying the environment then somebody is going to demand that they clean up, and that means spending a lot of money. They don't want to have to do that, and so it is a foregone, self-serving conclusion for them that climate science must find that humans are not making the planet warmer. And because they contribute to political campaigns and own media outlets they are able to get their wishful thinking disseminated as fact to a poorly-educated population that is not too critical of its information sources. The television audience figures that if the announcer is good-looking and well dressed, what they're saying is probably right, or if there is a colorful banner moving at the bottom of the screen, a dynamic background, and some music sometimes then the news is probably accurate. The people with the money can get the masses to believe anything, that's just the way it goes, that's what we're up against these days.

Somebody captured a lot of emails among some scientists and put them on the Internet, and some of it sounds bad. But you and I are not climate scientists, we don't know how they conduct their research, we don't know how they talk about it, we are unqualified to judge. Let me give you an example and a counterexample. One scientist emailed another guy saying, "I've just completed Mike's Nature trick of adding in the real temps to each series for the last 20 years (ie, from 1981 onwards) and from 1961 for Keith's to hide the decline." That sounds bad, but I admit I don't know the field well enough to know what he's really saying here, and neither do you. Now let me tell you something you might not realize: every survey you see, if it's any good, has fake data in it. Google for the words "survey" and "imputation" and you will find nearly two million hits. Imputation in a survey means that you make up data. It has to do with variance and statistical certainty, and it happens every day. Sometimes the data are generated from a statistical distribution, sometimes they "hot deck" using data from a case that is similar in some way to a missing one. Nobody criticizes a survey for using fake data, but if you saw some captured emails between statisticians you might conclude shockingly that survey results are faked. Government surveys, political surveys, marketing surveys, it actually makes a survey more valid to use made-up data, but you could paint that in a way to make it sound bad -- that's because you don't know anything about survey methods, you don't understand how margins of error are calculated, you don't understand the function of imputation. In the same way, you can take statements out of these climate emails and make something out of them. You don't know any more about climate science than you know about statistical surveys, you certainly do not know enough to read some emails -- or more likely read some out-of-context statements extracted by some teabagger and circulated around the Internet -- and conclude that the whole field is a hoax.

We teach kids that the "scientific method" is something about hypothesizing, experimenting, analyzing, but in reality science is like other socially constructed knowledge. That is not to say that scientific knowledge is of the same low quality as ordinary knowledge. There is no question that scientists do discover facts and evolve theories that are more refined and more comprehensive than the theories held by ordinary people. That is because scientists are specially selected by the academic process, specially trained in the theories and methods of their field, and their research is scrutinized and filtered through a rigorous peer-review publication process. Technical scientific work can not be debated in the popular press, because you and I don't have the special training required to evaluate it. You can assume that when scientific research appears in the newspapers and television shows there is money behind it, and you can assume that the news story simplifies and slants the research in a direction that is desirable to their own funding sources.

Thousands of articles have been published on a subject like global warming, and though it is not our field we can expect that there is a scientific consensus and we can expect that there are challenges to the consensus. That's how science works, you can call it paradigms and revolutions, whatever, it is a fascinating process of human beings who are trained to be skeptical and objective in their thinking, criticizing one another's work, looking for the errors, and at the same time building on one another's work. Each one is struggling for funding and academic standing, which are granted on the basis of the scientist's reputation, which is in turn based on his or her record of successful participation in the community of researchers.

In sum, some scientists talked freely among themselves, somebody stole the email and posted millions of lines of it, someone with an agenda tried to pick sentences out of it and make a case that the scientists were trying to pull off a hoax. I'm sorry, there's no story here.