The Blinne Blog

Deep Impact: Gone in a Flash

This image shows the initial ejecta that resulted when NASA's Deep Impact probe collided with comet Tempel 1 at 10:52 p.m. Pacific time, July 3 (1:52 a.m. Eastern time, July 4). It was taken by the spacecraft's high-resolution camera 13 seconds after impact. The image has been digitally processed to better show the comet's nucleus.

Image Credit:

NASA/JPL-Caltech/UMD

Here's the before picture of Comet Tempel 1 six minutes before it ran over NASA's Deep Impact probe at 10:52 a.m. Pacific time, July 3 (1:52 a.m. Eastern time, July 4). The picture was taken by the probe's impactor targeting sensor.

Image Credit: NASA/JPL-Caltech/UMD

Here's the before picture of the impactor spacecraft while it was being built at Ball Aerospace & Technologies Corporation, Boulder, Colo:

Image Credit: NASA/JPL/Ball Aerospace & Technologies Corporation

Images from other observatories follow beneath the fold.

What Are the Limits of Conventional Computing?

For their 125th anniversary Science magazine asked the Top 25 science questions, one of which is in my area of expertise. That question is:  What Are the Limits of Conventional Computing?.

At first glance, the ultimate limit of computation seems to be an engineering issue. How much energy can you put in a chip without melting it? How fast can you flip a bit in your silicon memory? How big can you make your computer and still fit it in a room? These questions don't seem terribly profound.

Unless of course, you've actually been involved like I have in semiconductor design. It always amazes us that any of this stuff works. With time, the physics gets nastier and nastier. When I started my career we were designing 4 micron gates (the distance measured here are the length of the gate of the transistor). Now we are working on 90 nm gates and next year it will be 65 nm. We now have to worry about capacitance, resistance, inductance, setup times, hold times, on chip variation, power consumption, crosstalk avoidance, electromigration, latch up, charge antennas, planarization, leakage, and electrostatic discharge (and these are just the ones I can think of off the top of my head).

In fact, computation is more abstract and fundamental than figuring out the best way to build a computer. This realization came in the mid-1930s, when Princeton mathematicians Alonzo Church and Alan Turing showed--roughly speaking--that any calculation involving bits and bytes can be done on an idealized computer known as a Turing machine. By showing that all classical computers are essentially alike, this discovery enabled scientists and mathematicians to ask fundamental questions about computation without getting bogged down in the minutiae of computer architecture.

This is true. In fact, the basic architecture of a computer has not fundamentally changed since Turing described binary digits, von Neumann developed automata theory, and Shannon developed information theory before and during World War II.

For example, theorists can now classify computational problems into broad categories. P problems are those, broadly speaking, that can be solved quickly, such as alphabetizing a list of names. NP problems are much tougher to solve but relatively easy to check once you've reached an answer. An example is the traveling salesman problem, finding the shortest possible route through a series of locations. All known algorithms for getting an answer take lots of computing power, and even relatively small versions might be out of reach of any classical computer.

Mathematicians have shown that if you could come up with a quick and easy shortcut to solving any one of the hardest type of NP problems, you'd be able to crack them all. In effect, the NP problems would turn into P problems. But it's uncertain whether such a shortcut exists--whether P = NP. Scientists think not, but proving this is one of the great unanswered questions in mathematics.

First some definitions. P stands for polynomial time and NP stands for non-deterministic polynomial time. It is really, really important to find an order P solution. Given a fast enough traditional computer, this can be solved in reasonable time. An NP complete problem is in essence unsolvable for a computer. To see how many problems are probably NP complete check out this non-exhaustive list. If a mathematician can show P equals NP or P does not equal NP, it will be an instant Fields Medal for such an individual.

If it is the latter we may need to abandon the von Neumann architecture that has served us so well. But, how would we do that?

But there is a realm beyond the classical computer: the quantum. The probabilistic nature of quantum theory allows atoms and other quantum objects to store information that's not restricted to only the binary 0 or 1 of information theory, but can also be 0 and 1 at the same time. Physicists around the world are building rudimentary quantum computers that exploit this and other quantum effects to do things that are provably impossible for ordinary computers, such as finding a target record in a database with too few queries. But scientists are still trying to figure out what quantum-mechanical properties make quantum computers so powerful and to engineer quantum computers big enough to do something useful.

By learning the strange logic of the quantum world and using it to do computing, scientists are delving deep into the laws of the subatomic world. Perhaps something as seemingly mundane as the quest for computing power might lead to a newfound understanding of the quantum realm.

The reason why quantum computing could be so powerful is the difference between a bit and a qubit. A qubit consists of multiple entangled quantum states. A bit can be 0 or 1. A qubit consists of superpositions of whole binary strings.  Thus we can do things in parallel which we are currently doing serially. One of the current challenges is maintaining what is know as coherence. Coherence can be lost by observing the state of a qubit as a result of the Heisenberg Uncertainty Principle.

The first half of my career has been marked by doing the same thing but making it smaller and smaller, and faster and faster. We've ridden Moore's Law far longer than I had personally anticipated. The question for the second half of my career is whether we will see Python's (Monty) Law: "And now for something completely different."

Pathological Science

Since I am in both the scientific and Evangelical communities, I've used my position to explain the scientific community to the Evangelical one, specifically the subgroup that believes in YEC and ID. Now I'm going to turn it around. When the scientific community looks at things like YEC and ID they scratch their heads. Worse, they come up with a theocratic conspiracy theory (not very parsimonious IMHO).

Irving Landmuir gave a talk in 1953 about what he termed as pathological science. This seems to explain some of the crazy scientific conclusions. Hat Tip: Dr. Randy Isaac, Executive Director of the American Scientific Affliation. His final note is well taken "No religious group has a monopoly on pathological science."

The characteristics of this Davis-Barnes experiment and the N-rays and the mitogenetic rays, they have things in common. These are cases where there is no dishonesty involved but where people are tricked into false results by a lack of understanding about what human beings can do to themselves in the way of being led astray by subjective effects, wishful thinking or threshold interactions. These are examples of pathological science. These are things that attracted a great deal of attention. Usually hundreds of papers have been published upon them. Sometimes they have lasted for fifteen or twenty years and then they gradually die away. [emphasis mine]

Now, the characteristic rules are these

Symptoms of Pathological Science:

1. The maximum effect that is observed is produced by a causative agent of barely detectable intensity, and the magnitude of the effect is substantially independent of the intensity of the cause.
2. The effect is of a magnitude that remains close to the limit of detectability; or, many measurements are necessary because of the very low statistical significance of the results.
3. Claims of great accuracy.
4. Fantastic theories contrary to experience.
5. Criticisms are met by ad hoc excuses thought up on the spur of the moment.
6. Ratio of supporters to critics rises up to somewhere near 50% and then falls gradually to oblivion.

A lot of the "science" in ID and YEC have the above characteristics. Two approaches have come out of the scientific community in how to deal with what they believe is bad science (correctly IMHO). One treats the mistake as sincere and the other seeks to do battle with conspiracies. How the scientific community dealt with the Kansas situation is an example of the former. I encourage my scientific and engineering colleagues to continue along that path. If the other path is taken and you falsely accuse people of a conspiracy or belittle their faith it will only harden the bad science.

Intelligent Design: An Evangelical Critique

Recently, I have been very critical of Intelligent Design. Why would an Evangelical Christian, even a scientific one, do that? Am I like Phillip Johnson contends, worse than an atheist? I do this because in addition to believing that Intelligent Design is scientifically unsound, I believe that it is unbiblical. For those of you who don't care about whether Intelligent Design is biblical then this is not for you. Go to Panda's Thumb. There's plenty of fodder there. I'm speaking to those who do care.

There are numerous passages of Scripture that deals with God as designer. One of them is Psalm 139.

139:14 I will give you thanks because your deeds are awesome and amazing.²²You knew me thoroughly;²³

139:15 my bones were not hidden from you,when²⁴ I was made in secret and sewed together in the depths of the earth.²⁵

Before I get to my main point, I would like to address an issue with interpretation of Genesis in light of YEC. Note that I am using the Net Bible translation. I'm using this one because it has a good technical commentary and also has good conservative credentials (being endorsed by Chuck Swindoll and Wayne Grudem). I would like to point out note 25.

²⁵ sn The phrase depths of the earth may be metaphorical (euphemistic) or it may reflect a prescientific belief about the origins of the embryo deep beneath the earth's surface (see H. W. Wolff, Anthropology of the Old Testament, 96-97). Job 1:21 also closely associates the mother's womb with the earth.

When the science is truly uncontroverted (embryos do not grow in the ground) the possibility of metaphor and pre-scientific accommodation is allowed, even in Evangelical exegesis.  Note that kind of phraseology is used in both Genesis 1:24 where animals were produced by the earth and Genesis 2:19 where Adam is formed out of the ground. Thus, the YEC interpretation of Genesis 1 and 2 is not a necessary one. In order to show a non-literal interpretation to be the correct one (or more to the point which non-literal interpretation is correct) requires more support. My limited point here is that such an interpretation is possible.

Back to my main point. Regardless of the use of metaphor or pre-scientific accommodation, Psalm 139 is clear that the development of the embryo is designed by God. Neo-Darwinism makes the argument that embryonic development and evolution are driven by the same underlying genetic and epigenetic mechanisms. Jonathan Wells got his PhD in developmental biology. He told his colleagues at Berkeley:

Introduction to Jonathan Wells
To: Participants in the Evolutionary Biology Virtual Discussion Group
From: Jonathan Wells

Thank you for inviting me to participate in your discussions. Let me begin by introducing myself.

I first became interested in evolution as an undergraduate at Princeton, where I studied geology. Years later, I also became interested in religion and studied theology at Yale, where I wrote a Ph.D. dissertation on the 19th-century Darwinian controversies. [1] At that time, I learned that the historical conflict between Christian theology and Darwinian evolution had almost nothing to do with biblical chronology, but turned primarily on the issue of design. Pursuing my interest in evolution, I came to Berkeley in 1989 to study embryology, since understanding how organisms develop is obviously relevant to understanding how they evolve. I am now finishing a second Ph.D. dissertation on mechanisms of early development in frog embryos. [emphasis mine]

Back then Jonathan Wells saw a connection. Now he participates in seminars that says there is a discontinuity of development and evolution. Not only does he seek to go against evolution but also against development. Note this quote:

To be sure, genes (DNA sequences) affect development, but many lines of research suggest that body plans and other morphological features are laid down prior to and largely independently of gene expression.

Excuse me? How come identical twins are identical? Identical genes by and large produce identical morphology. In his nonsensical mess, he does give a decent definition (if you take out the allegedly and note that natural selection is not the only mechanism for genetic change):

In neo-Darwinism, evolution is explained by the natural selection of small mutational changes in the "genetic programs" that allegedly control embryonic development.

Why am I making such a big deal about the connection between development and evolution?  Biblically speaking, development is designed. If something else uses the same or similar mechanism as something already shown to be designed then you have an effective argument that the evolutionary process is designed also.

There is one very big practical problem with my argument, though. It is not a scientific argument, but it's a Biblical one. The apologetic value is slight because it argues for evolution to Evangelicals but doesn't argue the Gospel to evolutionists. This should not be sneezed at however because there are many young Evangelical scientists who might not now lose their faith because they are persuaded that the evolutionary mechanism is true.

There is a reason why science cannot answer this question. ID needs to understand a theological concept known as concursus. Concursus is the doctrine that God in his Providence acts through and with second causes.  (Second causes are those things in which God acts indirectly.)  God's design while not hidden from His children is thus hidden from science. Trying to prove design from science is a fool's errand. Or put differently, if you want to know if something is designed, talk to the designer.

Not So Privileged Planet

Just as Privileged Planet  premieres June 23 at the Smithsonian comes this discovery of a rocky exoplanet as reported by news @ nature.com:

The hunt for worlds outside our Solar System has found its smallest planet yet: it weighs in at just seven-and-a-half times the size of the Earth.

Astronomers have already found more than 150 extrasolar planets, also known as exoplanets. But all of them are larger than Uranus, which has 15 times Earth's mass. The recent find is so small that it is likely to be rocky, its discoverers say, rather than a gas giant.

"This is the smallest extrasolar planet yet detected," says team member Paul Butler of the Carnegie Institution in Washington DC. "It's like Earth's bigger cousin."

Is there more to come? Probably.

"The fact that you have a rocky planet inside two gas giants makes it look a lot like our own Solar System," agrees Alan Boss, an astronomer at the Carnegie Institution who was not involved in the research.

Boss adds that the prospects of finding similar planetary systems are good, because stars like Gliese 876 are extremely common. "They're all over the place," he says, adding that of the 400 or so stars within 33 light years of Earth, about 300 of them are in the same class as Gliese 876

The team now hopes to find rocky planets around other red dwarfs in our Galaxy.

Artwork Credit: Trent Schindler/NSF

Stone Cold Mars

Last year, I reported about a possibility of microbes on Mars because of a large amount of methane. Another possibility has come up -- dare I say more parsimonious?

In today's Nature:

When methane was found in Mars's atmosphere last year, the media (and many scientists) seized on the idea that it was a whiff of life from martian bacteria. As estimates of the gas grew, so did column inches on the hopes of finding microbes.

But the real source may be far more mundane: chemistry. Geologists have calculated that given a supply of water and carbonates, just 80,000 tonnes of the mineral olivine would be enough to generate a year's worth of methane. Another paper reports a Cuba-sized olivine field on Mars's surface, and suggests there may be more beneath. Could the prospects for life be stone dead?

A Sample of One

AIDS researchers in New York apparently cried wolf. The Los Angeles Times reported:

The announcement from New York health officials in early February was chilling: A single patient had progressed from HIV infection to AIDS in months rather than years, and his strain of the HIV virus seemed impervious to normally effective medicines.

The patient, a gay man in his 40s, had unprotected anal intercourse with scores of partners. Headlines of a potential new killer spread around the world.

"This case is a wake-up call," Dr. Thomas R. Frieden, New York City's health commissioner, said at a news conference where he issued a warning for physicians to prepare for a possible new phase in the epidemic.

Yet several AIDS experts immediately questioned the importance of the case and the strategy of publicizing it so widely.

Months later, those doubts seem to have been confirmed.

No super-strain has emerged. The patient, whose name has been withheld, has responded to drug therapy. No one — not even the man's known sexual partners — was found to be infected with the same HIV strain.

The problem was there was only one patient that sparked the panic. The LA Times continues:

Some AIDS specialists now say the New York announcement was scientifically naive and needlessly alarmist — risking the effectiveness of future prevention efforts.

"Does it do good to [mislead] people and exaggerate?" asked Dr. Robert Gallo, co-discoverer of the virus that causes AIDS. He condemned Frieden's far-reaching conclusions as "scientifically, completely invalid, without a shred of evidence."

Dr. Gallo has an excellent point. Using a sample of one is not scientifically valid. Given the danger of a super-strain of AIDS I understand why the researchers dropped the ball. Nevertheless, this story is a cautionary tale about having too small a sample set.

Earlier on the Blinne Blog I commented on the politics of the movie, Privileged Planet, but I refrained from commenting on the content of the speculation. (I call it a speculation rather than a hypothesis because I don't know how this is falsifiable.) Like the AIDS researchers, much of the idea of a rare earth depends on a sample size of one. How do we know that life needs Carbon or water? Given extremophiles (life that lives in extreme environments, e.g. ingests arsenic, lives at extreme temperatures or without oxygen) how do we know that Earth is in a Goldilocks-style sweet spot? The answer is we don't because we have a sample of one. We can give all the probabilities in the world but they mean nothing if they are plus or minus 100%.

Is life unique on Earth? We don't know and the only way we will know if we find life somewhere else and thus prove life is not unique. SETI has the same problem as the Privileged Planet. The fact that E.T. hasn't phoned home means absolutely nothing. One of the motivations for Privileged Planet was to refute Carl Sagan's notion of our planet's mediocrity. Carl Sagan was in love with SETI. It's somewhat ironic that both fell into the same pit.

Christian Apologist Lead Astray By Intelligent Design Movement

Update: A reader notified me that the way I quoted was unclear. I've reformatted my entry to make it clear who is talking.

As you will see from the following interview Lee Strobel has a subtle, penetrating, mind and this has served him well in his apologetics works. But, he's not a scientist and he is dependent on scientists who are Christians to give him good information. Unfortunately, he has not been served well by them. Hat Tip: Short Attention Span.

Strobel is interviewed by  Christianity Today Magazine.

Christianity Today: You start your book with a scene with you as a young reporter. You're sent to West Virginia, where a bunch of religious townspeople are protesting the teaching of evolution in their textbooks. I was wondering if you thought that some of the things going on in public schools today would be similar to that.

Lee Strobel: If you look at public opinion polls, the public at large is generally skeptical about Darwinism.  It just doesn't ring true to a lot of people. There's an underlying widespread skepticism that neo-Darwinism could explain the diversity of life.

I take a different approach to that than some people do. I want more evolution to be taught, not less. What I mean by that is, right now, students are only getting one side of the coin. They're only getting a cursory overview of what neo-Darwinism is and being told some facts that some people believe support it. I want them to hear more about it. I want them to hear the evidence that challenges neo-Darwinism. I want students to be able to critically think about whether or not this makes sense. I want them to be free to follow the evidence wherever it points. That, to me, is academic freedom, that they should be able to pursue the evidence.

I'm not saying that Intelligent Design ought to be taught in public schools. I am saying that kids ought to be open to possibilities and pursue the evidence wherever it points, including in that direction.

The Blinne Blog: Give Strobel some points for not just cheer leading for Intelligent Design. You get the sense (and this is near-universal in the Evangelical community) that if you laid all the cards on the table Intelligent Design has the winning hand. Other more subtle proponents might say Intelligent Design doesn't have a strong hand, but neo-Darwinism has a weaker one. This is the root of the average Christian's opposition to how Darwinism is taught in the Public Schools.

And it is a pack of goods.

First Young Earth Creation and now Intelligent Design have been telling Christians that Darwinism is on the run. This is simply not true. There are thousands of papers describing in detail a host of evolutionary pathways. These pathways include many of Behe's supposedly irreducibly complex systems. Many of these papers are very recent. Not only do scientists know that the mechanisms of evolution work, but there is greater understanding how it works. The latter part could be used in a design argument because the genetic mechanism is quite beautiful. (Listen to the Francis Collins keynote on the ASA web site.) Anthony Flew is now being tauted by the Intelligent Design community. He stopped being an atheist because he believed neo-Darwinism, not because he doubted it. Design and Darwinism are not by necessity mutually-exclusive categories.  Even if the Darwinists are terribly self-deceived it is an utterly wrong to characterize them with the  "on-the-run" self-perception. Unfortunately, it is the Christian lay people who have been deceived. This produces the slander of the scientists that they are just a bunch of atheistic materialists. Some are, but many are not.

Christianity Today: How can Intelligent Design get past the creationist label?

Lee Strobel: It's always the Darwinists who bring that up. I've done this on my TV show, Faith Under Fire, where we'll have a debate between someone who is convinced of Intelligent Design versus a Darwinist. The Intelligent Design person brings up scientific data and arguments based on scientific evidence to support his or her beliefs. And then it goes to the other side, and that person is immediately accused of injecting faith and injecting religion and trying to be a subterfuge to teach the Bible in schools.

Well, time out here, who's bringing up religion? I didn't hear the Intelligent Design advocate bring up religion. It's being brought up by the other side. It's an ad hominem argument that Darwinists use to throw sand in people's eyes to suggest that this is just biblical creationism in another disguise. What I'd like to see is the debate centered on the evidence and the data. Why are people so afraid of evidence that happens to point toward an affirmation of what the vast majority of people on the Earth believe in the first place?

The Blinne Blog: As a Christian who believes in intelligent design (the concept, not the movement) I say start by  stopping the slander of scientists.  The mis-appraisal of the state of science has caused a huge mutual misconception between scientists and Evangelicals. Evangelicals say they only want the science taught. If that was so, then it would be neo-Darwinism, because the science taught by Intelligent Design is utter, complete garbage. The scientists try to figure us out and only can figure that this is just re-packaged creationism, not realizing that we really think a sow's ear is a silk purse. We really believe that in a fair fight we would win. If we don't do a brutally honest self-appraisal and fast, good men like Lee Strobel will lose their hard-fought reputation and the good they have done will be tainted.

Teaching the Controversy: A Proposal

A constant refrain from intelligent design proponents is that we need to teach the controversy. On the other side we have science educators who say, no we don't. I say use the controversy to teach good science. My proposal is a result of a challenge by Henry Gee in Nature.

What we really need is a cadre of religious scientists who can both stand up for their beliefs, and realize that they don't need intelligent design to stick their faith and work together.

These brave scientists should be prepared to expose intelligent design for the nonsense it is.

Fair enough. This Evangelical scientist will attempt to fill the gap. Here is what is needed in science education:

1. Have a clear definition of both science and methodological naturalism. Science is a natural tool seeking natural answers. There may exist supernatural answers but this is beyond the ken of science. Methodological naturalism assumes there exists a natural answer to a natural problem. If such an answer does not exist you are attempting to prove a negative and thus it cannot be done in science.
   
2. Distinguish metaphysical inferences about nature from science. I believe both in evolution and a designer. I look at the order of the Universe and the beauty of how evolution elegantly searches a solution space and I see a Designer behind it. Non-religious scientists see a self-existent Universe run by Natural Law. Science cannot tell us which metaphysical inference is true. Good grief, there's a reason why Aristotle called called this discipline "above physics". Let science be science and answer the "what" question and leave the "why" question to other disciplines. If the reason why certain topics are not brought up is because of the intentional self-limitation of the scientific method, then the conflict can be resolved without  crushing  students' faiths. The way many religious people view science is a bunch of elitists looking down and judging their faith. By stressing the deliberate limitations of science here, the misconception can be diffused.
3. Acknowledge the controversy but teach the science. This is the core of my proposal. Some of Intelligent Design theory is scientifically testable. Use these to show how science advances and what characterizes good and bad science.
   
   Behe had a number of concerns about Evolution in that Irreducible Complexity was something that Evolution couldn't overcome. Since Behe originally came up with his thesis the genomic revolution in Biology occurred. Biology students need to know how much progress has been made in Evolutionary Biology in the last decade. If you listen to the ID proponents you would think it was going the other way. By not addressing the concerns it gives science students the wrong impression that ID is simply being silenced.
   
   Here's how it would work in practice. Take some of his specific concerns about the blood clotting cascade, whale transitional forms, the vertebrate immune system etc. and show how they were addressed. These are actually good case studies of how Evolutionary Biology is done in the 21st Century and how it is different from the 19th. When you are done, stress that proving IC does not prove ID and disproving IC does not disprove ID. Science cannot answer that question.
   
   Address why ID does not show up in peer-reviewed journals. Namely, it is because of bad methodology. Americans think science is "democratic" and simply don't understand how real science is done. Here the controversy can be used to teach them. Pick up an issue of Science or Nature and go to the section on methods. Explain how a peer reviewer examines and critiques the methods to make sure that they are valid. Show how a peer reviewer looks for transparency and lack of conflicts of interest in order for the experiment to be replicated by others. A good case study can be found here. In this case study Randy Isaac plays a peer reviewer for an  allegedly censored paper by John Baumgartner. Randy Isaac is Executive Director of the ASA, a group of working scientists who are also Christians.

In conclusion, I believe that reasonable science educators and reasonable Christians can get together to improve the quality of the science education in our country. The level of ignorance of science in this country is quite high and it is not limited to just Conservative Christians. The current controversy can be taken from a huge distraction to an occasion to really teach Americans how science really works and how it is not a threat to people of faith.

Clinical Trials in Need of Reform

More on the ongoing saga of suppressing negative results of drug trials. From the front page of today's  New York Times.

When the drug industry came under fire last summer for failing to disclose poor results from studies of antidepressants, major drug makers promised to provide more information about their research on new medicines. But nearly a year later, crucial facts about many clinical trials remain hidden, scientists independent of the companies say.

Within the drug industry, companies are sharply divided about how much information to reveal, both about new studies and completed studies for drugs already being sold. The split is unusual in the industry, where companies generally take similar stands on regulatory issues.

Eli Lilly and some other companies have posted hundreds of trial results on the Web and pledged to disclose all results for all drugs they sell. But other drug makers, including Merck and Pfizer, release less information and are reluctant to add more, citing competitive pressures.

As a result, doctors and patients lack critical information about important drugs, academic researchers say, and the companies can hide negative trial results by refusing to publish studies, or by cherry-picking and highlighting the most favorable data from studies they do publish.

But that's not the half of it. In PlosMedicine Richard Smith, former editor of BMJ, cites other way drug companies corrupt the peer-review system for evaluating clinical trials:

Examples of Methods for Pharmaceutical Companies to Get the Results They Want from Clinical Trials

• Conduct a trial of your drug against a treatment known to be inferior.

• Trial your drugs against too low a dose of a competitor drug.

• Conduct a trial of your drug against too high a dose of a competitor drug (making your drug seem less toxic).

• Conduct trials that are too small to show differences from competitor drugs.

• Use multiple endpoints in the trial and select for publication those that give favourable results.

• Do multicentre trials and select for publication results from centres that are favourable.

• Conduct subgroup analyses and select for publication those that are favourable.

• Present results that are most likely to impress—for example, reduction in relative rather than absolute risk.

Given his twenty five years of experience what's Smith's cure?

How might we prevent journals from being an extension of the marketing arm of pharmaceutical companies in publishing trials that favour their products? Editors can review protocols, insist on trials being registered, demand that the role of sponsors be made transparent, and decline to publish trials unless researchers control the decision to publish. I doubt, however, that these steps will make much difference. Something more fundamental is needed.

Firstly, we need more public funding of trials, particularly of large head-to-head trials of all the treatments available for treating a condition. Secondly, journals should perhaps stop publishing trials. Instead, the protocols and results should be made available on regulated Web sites. Only such a radical step, I think, will stop journals from being beholden to companies. Instead of publishing trials, journals could concentrate on critically describing them.

I am no fan of bigger government but for information as critical as this we need a truly independent evaluation. If that cannot be done then at least the post-marketing safety analysis should be done and paid for by the government. Letting the pharmaceutical companies do these seems to be cheaper but in the long run the savings are not worth it. The temptation to rig the system is just too great.