Group items tagged

What, then, is the role of public (government) funding of research? Primarily, Wade argues (and I agree), to provide infrastructure for expensive research programs, such as building large colliders.

the more the government invests in basic science and infrastructure, the more winners will emerge that private industry can then capitalize on. This is a good way to build a competitive dynamic economy.

But there is a pitfall – prematurely picking winners and losers. Wade give the example of California investing specifically into developing stem cell treatments. He argues that stem cells, while promising, do not hold a guarantee of eventual success, and perhaps there are other technologies that will work and are being neglected. The history of science and technology has clearly demonstrated that it is wickedly difficult to predict the future (and all those who try are destined to be mocked by future generations with the benefit of perfect hindsight). Prematurely committing to one technology therefore contains a high risk of wasting a great deal of limited resources, and missing other perhaps more fruitful opportunities.

The underlying concept is that science research is a long-term game. Many avenues of research will not pan out, and those that do will take time to inspire specific applications. The media, however, likes catchy headlines. That means when they are reporting on basic science research journalists ask themselves – why should people care? What is the application of this that the average person can relate to? This seems reasonable from a journalistic point of view, but with basic science reporting it leads to wild speculation about a distant possible future application. The public is then left with the impression that we are on the verge of curing the common cold or cancer, or developing invisibility cloaks or flying cars, or replacing organs and having household robot servants. When a few years go by and we don’t have our personal android butlers, the public then thinks that the basic science was a bust, when in fact there was never a reasonable expectation that it would lead to a specific application anytime soon. But it still may be on track for interesting applications in a decade or two.

this also means that the government, generally, should not be in the game of picking winners an losers – putting their thumb on the scale, as it were. Rather, they will get the most bang for the research buck if they simply invest in science infrastructure, and also fund scientists in broad areas.

The same is true of technology – don’t pick winners and losers. The much-hyped “hydrogen economy” comes to mind. Let industry and the free market sort out what will work. If you have to invest in infrastructure before a technology is mature, then at least hedge your bets and keep funding flexible. Fund “alternative fuel” as a general category, and reassess on a regular basis how funds should be allocated. But don’t get too specific.

Funding research but leaving the details to scientists may be optimal

The scientific community can do their part by getting better at communicating with the media and the public. Try to avoid the temptation to overhype your own research, just because it is the most interesting thing in the world to you personally and you feel hype will help your funding. Don’t make it easy for the media to sensationalize your research – you should be the ones trying to hold back the reigns. Perhaps this is too much to hope for – market forces conspire too much to promote sensationalism.

In their appeal for more funding for scientific research, Leshner and Cooper argued that: "Across society, we don't have to look far for examples of basic research that paid off." They cite the creation of Google as a prime example of such payoffs: "Larry Page and Sergey Brin, then a National Science Foundation [NSF] fellow, did not intend to invent the Google search engine. Originally, they were intrigued by a mathematical challenge ..." The appealing imagery of a scientist who simply follows his curiosity and then makes a discovery with a large societal payoff is part of the core mythology of post-World War II science policies. The mythology shapes how governments around the world organize, account for, and fund research. A large body of scholarship has critiqued postwar science policies and found that, despite many notable successes, the science policies that may have made sense in the middle of the last century may need updating in the 21st century. In short, investments in "basic research" are not enough. Benoit Godin has asserted (PDF) that: "The problem is that the academic lobby has successfully claimed a monopoly on the creation of new knowledge, and that policy makers have been persuaded to confuse the necessary with the sufficient condition that investment in basic research would by itself necessarily lead to successful applications." Or as Leshner and Cooper declare in The Washington Post: "Federal investments in R&D have fueled half of the nation's economic growth since World War II."

Merton published his theory, called the “Matthew Effect,” in 1968. At that time, the average age of a biomedical researcher in the US receiving his or her first significant funding was 35 or younger. That meant that researchers who had little in terms of fame (at 35, they would have completed a PhD and a post-doc and would be just starting out on their own) could still get funded if they wrote interesting proposals. So Merton’s observation about getting credit for one’s work, however true in terms of prestige, wasn’t adversely affecting the funding of new ideas.

Over the last 40 years, the importance of fame in science has increased. The effect has compounded because famous researchers have gathered the smartest and most ambitious graduate students and post-docs around them, so that each notable paper from a high-wattage group bootstraps their collective power. The famous grow more famous, and the younger researchers in their coterie are able to use that fame to their benefit. The effect of this concentration of power has finally trickled down to the level of funding: The average age on first receipt of the most common “starter” grants at the NIH is now almost 42. This means younger researchers without the strength of a fame-based community are cut out of the funding process, and their ideas, separate from an older researcher’s sphere of influence, don’t get pursued. This causes a founder effect in modern science, where the prestigious few dictate the direction of research. It’s not only unfair—it’s also actively dangerous to science’s progress.

How can we fund science in a way that is fair? By judging researchers independently of their fame—in other words, not by how many times their papers have been cited. By judging them instead via new measures, measures that until recently have been too ephemeral to use.

Right now, the gold standard worldwide for measuring a scientist’s worth is the number of times his or her papers are cited, along with the importance of the journal where the papers were published. Decisions of funding, faculty positions, and eminence in the field all derive from a scientist’s citation history. But relying on these measures entrenches the Matthew Effect: Even when the lead author is a graduate student, the majority of the credit accrues to the much older principal investigator. And an influential lab can inflate its citations by referring to its own work in papers that themselves go on to be heavy-hitters.

what is most profoundly unbalanced about relying on citations is that the paper-based metric distorts the reality of the scientific enterprise. Scientists make data points, narratives, research tools, inventions, pictures, sounds, videos, and more. Journal articles are a compressed and heavily edited version of what happens in the lab.

We have the capacity to measure the quality of a scientist across multiple dimensions, not just in terms of papers and citations. Was the scientist’s data online? Was it comprehensible? Can I replicate the results? Run the code? Access the research tools? Use them to write a new paper? What ideas were examined and discarded along the way, so that I might know the reality of the research? What is the impact of the scientist as an individual, rather than the impact of the paper he or she wrote? When we can see the scientist as a whole, we’re less prone to relying on reputation alone to assess merit.

Multidimensionality is one of the only counters to the Matthew Effect we have available. In forums where this kind of meritocracy prevails over seniority, like Linux or Wikipedia, the Matthew Effect is much less pronounced. And we have the capacity to measure each of these individual factors of a scientist’s work, using the basic discourse of the Web: the blog, the wiki, the comment, the trackback. We can find out who is talented in a lab, not just who was smart enough to hire that talent. As we develop the ability to measure multiple dimensions of scientific knowledge creation, dissemination, and re-use, we open up a new way to recognize excellence. What we can measure, we can value.

Suppose you want to say, "Bush read Chomsky's latest book." Let's focus on just the verb, "read." To say this sentence in English, we have to mark the verb for tense; in this case, we have to pronounce it like "red" and not like "reed." In Indonesian you need not (in fact, you can't) alter the verb to mark tense. In Russian you would have to alter the verb to indicate tense and gender. So if it was Laura Bush who did the reading, you'd use a different form of the verb than if it was George. In Russian you'd also have to include in the verb information about completion. If George read only part of the book, you'd use a different form of the verb than if he'd diligently plowed through the whole thing. In Turkish you'd have to include in the verb how you acquired this information: if you had witnessed this unlikely event with your own two eyes, you'd use one verb form, but if you had simply read or heard about it, or inferred it from something Bush said, you'd use a different verb form.

Clearly, languages require different things of their speakers. Does this mean that the speakers think differently about the world? Do English, Indonesian, Russian, and Turkish speakers end up attending to, partitioning, and remembering their experiences differently just because they speak different languages?

For some scholars, the answer to these questions has been an obvious yes. Just look at the way people talk, they might say. Certainly, speakers of different languages must attend to and encode strikingly different aspects of the world just so they can use their language properly. Scholars on the other side of the debate don't find the differences in how people talk convincing. All our linguistic utterances are sparse, encoding only a small part of the information we have available. Just because English speakers don't include the same information in their verbs that Russian and Turkish speakers do doesn't mean that English speakers aren't paying attention to the same things; all it means is that they're not talking about them. It's possible that everyone thinks the same way, notices the same things, but just talks differently.

Believers in cross-linguistic differences counter that everyone does not pay attention to the same things: if everyone did, one might think it would be easy to learn to speak other languages. Unfortunately, learning a new language (especially one not closely related to those you know) is never easy; it seems to require paying attention to a new set of distinctions. Whether it's distinguishing modes of being in Spanish, evidentiality in Turkish, or aspect in Russian, learning to speak these languages requires something more than just learning vocabulary: it requires paying attention to the right things in the world so that you have the correct information to include in what you say.

Follow me to Pormpuraaw, a small Aboriginal community on the western edge of Cape York, in northern Australia. I came here because of the way the locals, the Kuuk Thaayorre, talk about space. Instead of words like "right," "left," "forward," and "back," which, as commonly used in English, define space relative to an observer, the Kuuk Thaayorre, like many other Aboriginal groups, use cardinal-direction terms — north, south, east, and west — to define space.1 This is done at all scales, which means you have to say things like "There's an ant on your southeast leg" or "Move the cup to the north northwest a little bit." One obvious consequence of speaking such a language is that you have to stay oriented at all times, or else you cannot speak properly. The normal greeting in Kuuk Thaayorre is "Where are you going?" and the answer should be something like " Southsoutheast, in the middle distance." If you don't know which way you're facing, you can't even get past "Hello."

The result is a profound difference in navigational ability and spatial knowledge between speakers of languages that rely primarily on absolute reference frames (like Kuuk Thaayorre) and languages that rely on relative reference frames (like English).2 Simply put, speakers of languages like Kuuk Thaayorre are much better than English speakers at staying oriented and keeping track of where they are, even in unfamiliar landscapes or inside unfamiliar buildings. What enables them — in fact, forces them — to do this is their language. Having their attention trained in this way equips them to perform navigational feats once thought beyond human capabilities. Because space is such a fundamental domain of thought, differences in how people think about space don't end there. People rely on their spatial knowledge to build other, more complex, more abstract representations. Representations of such things as time, number, musical pitch, kinship relations, morality, and emotions have been shown to depend on how we think about space. So if the Kuuk Thaayorre think differently about space, do they also think differently about other things, like time? This is what my collaborator Alice Gaby and I came to Pormpuraaw to find out.

To test this idea, we gave people sets of pictures that showed some kind of temporal progression (e.g., pictures of a man aging, or a crocodile growing, or a banana being eaten). Their job was to arrange the shuffled photos on the ground to show the correct temporal order. We tested each person in two separate sittings, each time facing in a different cardinal direction. If you ask English speakers to do this, they'll arrange the cards so that time proceeds from left to right. Hebrew speakers will tend to lay out the cards from right to left, showing that writing direction in a language plays a role.3 So what about folks like the Kuuk Thaayorre, who don't use words like "left" and "right"? What will they do? The Kuuk Thaayorre did not arrange the cards more often from left to right than from right to left, nor more toward or away from the body. But their arrangements were not random: there was a pattern, just a different one from that of English speakers. Instead of arranging time from left to right, they arranged it from east to west. That is, when they were seated facing south, the cards went left to right. When they faced north, the cards went from right to left. When they faced east, the cards came toward the body and so on. This was true even though we never told any of our subjects which direction they faced. The Kuuk Thaayorre not only knew that already (usually much better than I did), but they also spontaneously used this spatial orientation to construct their representations of time.

I have described how languages shape the way we think about space, time, colors, and objects. Other studies have found effects of language on how people construe events, reason about causality, keep track of number, understand material substance, perceive and experience emotion, reason about other people's minds, choose to take risks, and even in the way they choose professions and spouses.8 Taken together, these results show that linguistic processes are pervasive in most fundamental domains of thought, unconsciously shaping us from the nuts and bolts of cognition and perception to our loftiest abstract notions and major life decisions. Language is central to our experience of being human, and the languages we speak profoundly shape the way we think, the way we see the world, the way we live our lives.

The fact that even quirks of grammar, such as grammatical gender, can affect our thinking is profound. Such quirks are pervasive in language; gender, for example, applies to all nouns, which means that it is affecting how people think about anything that can be designated by a noun.

How does an artist decide whether death, say, or time should be painted as a man or a woman? It turns out that in 85 percent of such personifications, whether a male or female figure is chosen is predicted by the grammatical gender of the word in the artist's native language. So, for example, German painters are more likely to paint death as a man, whereas Russian painters are more likely to paint death as a woman.

Does treating chairs as masculine and beds as feminine in the grammar make Russian speakers think of chairs as being more like men and beds as more like women in some way? It turns out that it does. In one study, we asked German and Spanish speakers to describe objects having opposite gender assignment in those two languages. The descriptions they gave differed in a way predicted by grammatical gender. For example, when asked to describe a "key" — a word that is masculine in German and feminine in Spanish — the German speakers were more likely to use words like "hard," "heavy," "jagged," "metal," "serrated," and "useful," whereas Spanish speakers were more likely to say "golden," "intricate," "little," "lovely," "shiny," and "tiny." To describe a "bridge," which is feminine in German and masculine in Spanish, the German speakers said "beautiful," "elegant," "fragile," "peaceful," "pretty," and "slender," and the Spanish speakers said "big," "dangerous," "long," "strong," "sturdy," and "towering." This was true even though all testing was done in English, a language without grammatical gender. The same pattern of results also emerged in entirely nonlinguistic tasks (e.g., rating similarity between pictures). And we can also show that it is aspects of language per se that shape how people think: teaching English speakers new grammatical gender systems influences mental representations of objects in the same way it does with German and Spanish speakers. Apparently even small flukes of grammar, like the seemingly arbitrary assignment of gender to a noun, can have an effect on people's ideas of concrete objects in the world.

Even basic aspects of time perception can be affected by language. For example, English speakers prefer to talk about duration in terms of length (e.g., "That was a short talk," "The meeting didn't take long"), while Spanish and Greek speakers prefer to talk about time in terms of amount, relying more on words like "much" "big", and "little" rather than "short" and "long" Our research into such basic cognitive abilities as estimating duration shows that speakers of different languages differ in ways predicted by the patterns of metaphors in their language. (For example, when asked to estimate duration, English speakers are more likely to be confused by distance information, estimating that a line of greater length remains on the test screen for a longer period of time, whereas Greek speakers are more likely to be confused by amount, estimating that a container that is fuller remains longer on the screen.)

An important question at this point is: Are these differences caused by language per se or by some other aspect of culture? Of course, the lives of English, Mandarin, Greek, Spanish, and Kuuk Thaayorre speakers differ in a myriad of ways. How do we know that it is language itself that creates these differences in thought and not some other aspect of their respective cultures? One way to answer this question is to teach people new ways of talking and see if that changes the way they think. In our lab, we've taught English speakers different ways of talking about time. In one such study, English speakers were taught to use size metaphors (as in Greek) to describe duration (e.g., a movie is larger than a sneeze), or vertical metaphors (as in Mandarin) to describe event order. Once the English speakers had learned to talk about time in these new ways, their cognitive performance began to resemble that of Greek or Mandarin speakers. This suggests that patterns in a language can indeed play a causal role in constructing how we think.6 In practical terms, it means that when you're learning a new language, you're not simply learning a new way of talking, you are also inadvertently learning a new way of thinking. Beyond abstract or complex domains of thought like space and time, languages also meddle in basic aspects of visual perception — our ability to distinguish colors, for example. Different languages divide up the color continuum differently: some make many more distinctions between colors than others, and the boundaries often don't line up across languages.

To test whether differences in color language lead to differences in color perception, we compared Russian and English speakers' ability to discriminate shades of blue. In Russian there is no single word that covers all the colors that English speakers call "blue." Russian makes an obligatory distinction between light blue (goluboy) and dark blue (siniy). Does this distinction mean that siniy blues look more different from goluboy blues to Russian speakers? Indeed, the data say yes. Russian speakers are quicker to distinguish two shades of blue that are called by the different names in Russian (i.e., one being siniy and the other being goluboy) than if the two fall into the same category. For English speakers, all these shades are still designated by the same word, "blue," and there are no comparable differences in reaction time. Further, the Russian advantage disappears when subjects are asked to perform a verbal interference task (reciting a string of digits) while making color judgments but not when they're asked to perform an equally difficult spatial interference task (keeping a novel visual pattern in memory). The disappearance of the advantage when performing a verbal task shows that language is normally involved in even surprisingly basic perceptual judgments — and that it is language per se that creates this difference in perception between Russian and English speakers.

What it means for a language to have grammatical gender is that words belonging to different genders get treated differently grammatically and words belonging to the same grammatical gender get treated the same grammatically. Languages can require speakers to change pronouns, adjective and verb endings, possessives, numerals, and so on, depending on the noun's gender. For example, to say something like "my chair was old" in Russian (moy stul bil' stariy), you'd need to make every word in the sentence agree in gender with "chair" (stul), which is masculine in Russian. So you'd use the masculine form of "my," "was," and "old." These are the same forms you'd use in speaking of a biological male, as in "my grandfather was old." If, instead of speaking of a chair, you were speaking of a bed (krovat'), which is feminine in Russian, or about your grandmother, you would use the feminine form of "my," "was," and "old."

Experts in the math of probability and statistics are well aware of these problems and have for decades expressed concern about them in major journals. Over the years, hundreds of published papers have warned that science’s love affair with statistics has spawned countless illegitimate findings. In fact, if you believe what you read in the scientific literature, you shouldn’t believe what you read in the scientific literature.

“There are more false claims made in the medical literature than anybody appreciates,” he says. “There’s no question about that.”Nobody contends that all of science is wrong, or that it hasn’t compiled an impressive array of truths about the natural world. Still, any single scientific study alone is quite likely to be incorrect, thanks largely to the fact that the standard statistical system for drawing conclusions is, in essence, illogical. “A lot of scientists don’t understand statistics,” says Goodman. “And they don’t understand statistics because the statistics don’t make sense.”

In 2007, for instance, researchers combing the medical literature found numerous studies linking a total of 85 genetic variants in 70 different genes to acute coronary syndrome, a cluster of heart problems. When the researchers compared genetic tests of 811 patients that had the syndrome with a group of 650 (matched for sex and age) that didn’t, only one of the suspect gene variants turned up substantially more often in those with the syndrome — a number to be expected by chance.“Our null results provide no support for the hypothesis that any of the 85 genetic variants tested is a susceptibility factor” for the syndrome, the researchers reported in the Journal of the American Medical Association.How could so many studies be wrong? Because their conclusions relied on “statistical significance,” a concept at the heart of the mathematical analysis of modern scientific experiments.

Statistical significance is a phrase that every science graduate student learns, but few comprehend. While its origins stretch back at least to the 19th century, the modern notion was pioneered by the mathematician Ronald A. Fisher in the 1920s. His original interest was agriculture. He sought a test of whether variation in crop yields was due to some specific intervention (say, fertilizer) or merely reflected random factors beyond experimental control.Fisher first assumed that fertilizer caused no difference — the “no effect” or “null” hypothesis. He then calculated a number called the P value, the probability that an observed yield in a fertilized field would occur if fertilizer had no real effect. If P is less than .05 — meaning the chance of a fluke is less than 5 percent — the result should be declared “statistically significant,” Fisher arbitrarily declared, and the no effect hypothesis should be rejected, supposedly confirming that fertilizer works.Fisher’s P value eventually became the ultimate arbiter of credibility for science results of all sorts

But in fact, there’s no logical basis for using a P value from a single study to draw any conclusion. If the chance of a fluke is less than 5 percent, two possible conclusions remain: There is a real effect, or the result is an improbable fluke. Fisher’s method offers no way to know which is which. On the other hand, if a study finds no statistically significant effect, that doesn’t prove anything, either. Perhaps the effect doesn’t exist, or maybe the statistical test wasn’t powerful enough to detect a small but real effect.

Soon after Fisher established his system of statistical significance, it was attacked by other mathematicians, notably Egon Pearson and Jerzy Neyman. Rather than testing a null hypothesis, they argued, it made more sense to test competing hypotheses against one another. That approach also produces a P value, which is used to gauge the likelihood of a “false positive” — concluding an effect is real when it actually isn’t. What  eventually emerged was a hybrid mix of the mutually inconsistent Fisher and Neyman-Pearson approaches, which has rendered interpretations of standard statistics muddled at best and simply erroneous at worst. As a result, most scientists are confused about the meaning of a P value or how to interpret it. “It’s almost never, ever, ever stated correctly, what it means,” says Goodman.

experimental data yielding a P value of .05 means that there is only a 5 percent chance of obtaining the observed (or more extreme) result if no real effect exists (that is, if the no-difference hypothesis is correct). But many explanations mangle the subtleties in that definition. A recent popular book on issues involving science, for example, states a commonly held misperception about the meaning of statistical significance at the .05 level: “This means that it is 95 percent certain that the observed difference between groups, or sets of samples, is real and could not have arisen by chance.”

That interpretation commits an egregious logical error (technical term: “transposed conditional”): confusing the odds of getting a result (if a hypothesis is true) with the odds favoring the hypothesis if you observe that result. A well-fed dog may seldom bark, but observing the rare bark does not imply that the dog is hungry. A dog may bark 5 percent of the time even if it is well-fed all of the time. (See Box 2)

Another common error equates statistical significance to “significance” in the ordinary use of the word. Because of the way statistical formulas work, a study with a very large sample can detect “statistical significance” for a small effect that is meaningless in practical terms. A new drug may be statistically better than an old drug, but for every thousand people you treat you might get just one or two additional cures — not clinically significant. Similarly, when studies claim that a chemical causes a “significantly increased risk of cancer,” they often mean that it is just statistically significant, possibly posing only a tiny absolute increase in risk.

Statisticians perpetually caution against mistaking statistical significance for practical importance, but scientific papers commit that error often. Ziliak studied journals from various fields — psychology, medicine and economics among others — and reported frequent disregard for the distinction.

“I found that eight or nine of every 10 articles published in the leading journals make the fatal substitution” of equating statistical significance to importance, he said in an interview. Ziliak’s data are documented in the 2008 book The Cult of Statistical Significance, coauthored with Deirdre McCloskey of the University of Illinois at Chicago.

Multiplicity of mistakesEven when “significance” is properly defined and P values are carefully calculated, statistical inference is plagued by many other problems. Chief among them is the “multiplicity” issue — the testing of many hypotheses simultaneously. When several drugs are tested at once, or a single drug is tested on several groups, chances of getting a statistically significant but false result rise rapidly.

Recognizing these problems, some researchers now calculate a “false discovery rate” to warn of flukes disguised as real effects. And genetics researchers have begun using “genome-wide association studies” that attempt to ameliorate the multiplicity issue (SN: 6/21/08, p. 20).

Many researchers now also commonly report results with confidence intervals, similar to the margins of error reported in opinion polls. Such intervals, usually given as a range that should include the actual value with 95 percent confidence, do convey a better sense of how precise a finding is. But the 95 percent confidence calculation is based on the same math as the .05 P value and so still shares some of its problems.

Statistical problems also afflict the “gold standard” for medical research, the randomized, controlled clinical trials that test drugs for their ability to cure or their power to harm. Such trials assign patients at random to receive either the substance being tested or a placebo, typically a sugar pill; random selection supposedly guarantees that patients’ personal characteristics won’t bias the choice of who gets the actual treatment. But in practice, selection biases may still occur, Vance Berger and Sherri Weinstein noted in 2004 in ControlledClinical Trials. “Some of the benefits ascribed to randomization, for example that it eliminates all selection bias, can better be described as fantasy than reality,” they wrote.

Randomization also should ensure that unknown differences among individuals are mixed in roughly the same proportions in the groups being tested. But statistics do not guarantee an equal distribution any more than they prohibit 10 heads in a row when flipping a penny. With thousands of clinical trials in progress, some will not be well randomized. And DNA differs at more than a million spots in the human genetic catalog, so even in a single trial differences may not be evenly mixed. In a sufficiently large trial, unrandomized factors may balance out, if some have positive effects and some are negative. (See Box 3) Still, trial results are reported as averages that may obscure individual differences, masking beneficial or harm­ful effects and possibly leading to approval of drugs that are deadly for some and denial of effective treatment to others.

nother concern is the common strategy of combining results from many trials into a single “meta-analysis,” a study of studies. In a single trial with relatively few participants, statistical tests may not detect small but real and possibly important effects. In principle, combining smaller studies to create a larger sample would allow the tests to detect such small effects. But statistical techniques for doing so are valid only if certain criteria are met. For one thing, all the studies conducted on the drug must be included — published and unpublished. And all the studies should have been performed in a similar way, using the same protocols, definitions, types of patients and doses. When combining studies with differences, it is necessary first to show that those differences would not affect the analysis, Goodman notes, but that seldom happens. “That’s not a formal part of most meta-analyses,” he says.

Meta-analyses have produced many controversial conclusions. Common claims that antidepressants work no better than placebos, for example, are based on meta-analyses that do not conform to the criteria that would confer validity. Similar problems afflicted a 2007 meta-analysis, published in the New England Journal of Medicine, that attributed increased heart attack risk to the diabetes drug Avandia. Raw data from the combined trials showed that only 55 people in 10,000 had heart attacks when using Avandia, compared with 59 people per 10,000 in comparison groups. But after a series of statistical manipulations, Avandia appeared to confer an increased risk.

combining small studies in a meta-analysis is not a good substitute for a single trial sufficiently large to test a given question. “Meta-analyses can reduce the role of chance in the interpretation but may introduce bias and confounding,” Hennekens and DeMets write in the Dec. 2 Journal of the American Medical Association. “Such results should be considered more as hypothesis formulating than as hypothesis testing.”

Some studies show dramatic effects that don’t require sophisticated statistics to interpret. If the P value is 0.0001 — a hundredth of a percent chance of a fluke — that is strong evidence, Goodman points out. Besides, most well-accepted science is based not on any single study, but on studies that have been confirmed by repetition. Any one result may be likely to be wrong, but confidence rises quickly if that result is independently replicated.“Replication is vital,” says statistician Juliet Shaffer, a lecturer emeritus at the University of California, Berkeley. And in medicine, she says, the need for replication is widely recognized. “But in the social sciences and behavioral sciences, replication is not common,” she noted in San Diego in February at the annual meeting of the American Association for the Advancement of Science. “This is a sad situation.”

Most critics of standard statistics advocate the Bayesian approach to statistical reasoning, a methodology that derives from a theorem credited to Bayes, an 18th century English clergyman. His approach uses similar math, but requires the added twist of a “prior probability” — in essence, an informed guess about the expected probability of something in advance of the study. Often this prior probability is more than a mere guess — it could be based, for instance, on previous studies.

it basically just reflects the need to include previous knowledge when drawing conclusions from new observations. To infer the odds that a barking dog is hungry, for instance, it is not enough to know how often the dog barks when well-fed. You also need to know how often it eats — in order to calculate the prior probability of being hungry. Bayesian math combines a prior probability with observed data to produce an estimate of the likelihood of the hunger hypothesis. “A scientific hypothesis cannot be properly assessed solely by reference to the observational data,” but only by viewing the data in light of prior belief in the hypothesis, wrote George Diamond and Sanjay Kaul of UCLA’s School of Medicine in 2004 in the Journal of the American College of Cardiology. “Bayes’ theorem is ... a logically consistent, mathematically valid, and intuitive way to draw inferences about the hypothesis.” (See Box 4)

In many real-life contexts, Bayesian methods do produce the best answers to important questions. In medical diagnoses, for instance, the likelihood that a test for a disease is correct depends on the prevalence of the disease in the population, a factor that Bayesian math would take into account.

But Bayesian methods introduce a confusion into the actual meaning of the mathematical concept of “probability” in the real world. Standard or “frequentist” statistics treat probabilities as objective realities; Bayesians treat probabilities as “degrees of belief” based in part on a personal assessment or subjective decision about what to include in the calculation. That’s a tough placebo to swallow for scientists wedded to the “objective” ideal of standard statistics. “Subjective prior beliefs are anathema to the frequentist, who relies instead on a series of ad hoc algorithms that maintain the facade of scientific objectivity,” Diamond and Kaul wrote.Conflict between frequentists and Bayesians has been ongoing for two centuries. So science’s marriage to mathematics seems to entail some irreconcilable differences. Whether the future holds a fruitful reconciliation or an ugly separation may depend on forging a shared understanding of probability.“What does probability mean in real life?” the statistician David Salsburg asked in his 2001 book The Lady Tasting Tea. “This problem is still unsolved, and ... if it remains un­solved, the whole of the statistical approach to science may come crashing down from the weight of its own inconsistencies.”

government financing agencies as the National Institutes of Health and the National Science Foundation are like the managers of a stock index fund: they buy everything in the market, and the few spectacular winners make up for all the disasters. But just as index fund managers often go astray when they try to improve on the index’s performance by overweighting the stocks they favor, the government can go wrong when it tries to pick winners.

when researchers took up these questions experimentally, they did not end up confirming the traditional view. They did not find that people overwhelmingly favoured objectivism. Instead, the results consistently point to a more complex picture. There seems to be a striking degree of conflict even in the intuitions of ordinary folks, with some people under some circumstances offering objectivist answers, while other people under other circumstances offer more relativist views. And that is not all. The experimental results seem to be giving us an ever deeper understanding of why it is that people are drawn in these different directions, what it is that makes some people move toward objectivism and others toward more relativist views.

consider a study by Adam Feltz and Edward Cokely. They were interested in the relationship between belief in moral relativism and the personality trait openness to experience. Accordingly, they conducted a study in which they measured both openness to experience and belief in moral relativism. To get at people’s degree of openness to experience, they used a standard measure designed by researchers in personality psychology. To get at people’s agreement with moral relativism, they told participants about two characters – John and Fred – who held opposite opinions about whether some given act was morally bad. Participants were then asked whether one of these two characters had to be wrong (the objectivist answer) or whether it could be that neither of them was wrong (the relativist answer). What they found was a quite surprising result. It just wasn’t the case that participants overwhelmingly favoured the objectivist answer. Instead, people’s answers were correlated with their personality traits. The higher a participant was in openness to experience, the more likely that participant was to give a relativist answer.

Geoffrey Goodwin and John Darley pursued a similar approach, this time looking at the relationship between people’s belief in moral relativism and their tendency to approach questions by considering a whole variety of possibilities. They proceeded by giving participants mathematical puzzles that could only be solved by looking at multiple different possibilities. Thus, participants who considered all these possibilities would tend to get these problems right, whereas those who failed to consider all the possibilities would tend to get the problems wrong. Now comes the surprising result: those participants who got these problems right were significantly more inclined to offer relativist answers than were those participants who got the problems wrong.

Shaun Nichols and Tricia Folds-Bennett looked at how people’s moral conceptions develop as they grow older. Research in developmental psychology has shown that as children grow up, they develop different understandings of the physical world, of numbers, of other people’s minds. So what about morality? Do people have a different understanding of morality when they are twenty years old than they do when they are only four years old? What the results revealed was a systematic developmental difference. Young children show a strong preference for objectivism, but as they grow older, they become more inclined to adopt relativist views. In other words, there appears to be a developmental shift toward increasing relativism as children mature. (In an exciting new twist on this approach, James Beebe and David Sackris have shown that this pattern eventually reverses, with middle-aged people showing less inclination toward relativism than college students do.)

People are more inclined to be relativists when they score highly in openness to experience, when they have an especially good ability to consider multiple possibilities, when they have matured past childhood (but not when they get to be middle-aged). Looking at these various effects, my collaborators and I thought that it might be possible to offer a single unifying account that explained them all. Specifically, our thought was that people might be drawn to relativism to the extent that they open their minds to alternative perspectives. There could be all sorts of different factors that lead people to open their minds in this way (personality traits, cognitive dispositions, age), but regardless of the instigating factor, researchers seemed always to be finding the same basic effect. The more people have a capacity to truly engage with other perspectives, the more they seem to turn toward moral relativism.

To really put this hypothesis to the test, Hagop Sarkissian, Jennifer Wright, John Park, David Tien and I teamed up to run a series of new studies. Our aim was to actually manipulate the degree to which people considered alternative perspectives. That is, we wanted to randomly assign people to different conditions in which they would end up thinking in different ways, so that we could then examine the impact of these different conditions on their intuitions about moral relativism.

The results of the study showed a systematic difference between conditions. In particular, as we moved toward more distant cultures, we found a steady shift toward more relativist answers – with people in the first condition tending to agree with the statement that at least one of them had to be wrong, people in the second being pretty evenly split between the two answers, and people in the third tending to reject the statement quite decisively.

If we learn that people’s ordinary practice is not an objectivist one – that it actually varies depending on the degree to which people take other perspectives into account – how can we then use this information to address the deeper philosophical issues about the true nature of morality? The answer here is in one way very complex and in another very simple. It is complex in that one can answer such questions only by making use of very sophisticated and subtle philosophical methods. Yet, at the same time, it is simple in that such methods have already been developed and are being continually refined and elaborated within the literature in analytic philosophy. The trick now is just to take these methods and apply them to working out the implications of an ordinary practice that actually exists.

Morris also makes heavy use of the “taking a quote out of context” strategy favored by creationists. His quotes are often from secondary sources and are incomplete.

A more scholarly (i.e. intellectually honest) approach would be to cite actual evidence to support a point. If you are going to cite an authority, then make sure the quote is relevant, in context, and complete.

And even better, cite a number of sources to show that the opinion is representative. Rather we get single, partial, and often outdated quotes without context.

(nature is not, it turns out, cleanly divided into “kinds”, which have no operational definition). He also repeats this canard: Such variation is often called microevolution, and these minor horizontal (or downward) changes occur fairly often, but such changes are not true “vertical” evolution. This is the microevolution/macroevolution false dichotomy. It is only “often called” this by creationists – not by actual evolutionary scientists. There is no theoretical or empirical division between macro and micro evolution. There is just evolution, which can result in the full spectrum of change from minor tweaks to major changes.

Morris wonders why there are no “dats” – dog-cat transitional species. He misses the hierarchical nature of evolution. As evolution proceeds, and creatures develop a greater and greater evolutionary history behind them, they increasingly are committed to their body plan. This results in a nestled hierarchy of groups – which is reflected in taxonomy (the naming scheme of living things).

once our distant ancestors developed the basic body plan of chordates, they were committed to that body plan. Subsequent evolution resulted in variations on that plan, each of which then developed further variations, etc. But evolution cannot go backward, undo evolutionary changes and then proceed down a different path. Once an evolutionary line has developed into a dog, evolution can produce variations on the dog, but it cannot go backwards and produce a cat.

Stephen J. Gould described this distinction as the difference between disparity and diversity. Disparity (the degree of morphological difference) actually decreases over evolutionary time, as lineages go extinct and the surviving lineages are committed to fewer and fewer basic body plans. Meanwhile, diversity (the number of variations on a body plan) within groups tends to increase over time.

the kind of evolutionary changes that were happening in the past, when species were relatively undifferentiated (compared to contemporary species) is indeed not happening today. Modern multi-cellular life has 600 million years of evolutionary history constraining their future evolution – which was not true of species at the base of the evolutionary tree. But modern species are indeed still evolving.

Here is a list of research documenting observed instances of speciation. The list is from 1995, and there are more recent examples to add to the list. Here are some more. And here is a good list with references of more recent cases.

Next Morris tries to convince the reader that there is no evidence for evolution in the past, focusing on the fossil record. He repeats the false claim (again, which I already dealt with) that there are no transitional fossils: Even those who believe in rapid evolution recognize that a considerable number of generations would be required for one distinct “kind” to evolve into another more complex kind. There ought, therefore, to be a considerable number of true transitional structures preserved in the fossils — after all, there are billions of non-transitional structures there! But (with the exception of a few very doubtful creatures such as the controversial feathered dinosaurs and the alleged walking whales), they are not there.

I deal with this question at length here, pointing out that there are numerous transitional fossils for the evolution of terrestrial vertebrates, mammals, whales, birds, turtles, and yes – humans from ape ancestors. There are many more examples, these are just some of my favorites.

Much of what follows (as you can see it takes far more space to correct the lies and distortions of Morris than it did to create them) is classic denialism – misinterpreting the state of the science, and confusing lack of information about the details of evolution with lack of confidence in the fact of evolution. Here are some examples – he quotes Niles Eldridge: “It is a simple ineluctable truth that virtually all members of a biota remain basically stable, with minor fluctuations, throughout their durations. . . .“ So how do evolutionists arrive at their evolutionary trees from fossils of organisms which didn’t change during their durations? Beware the “….” – that means that meaningful parts of the quote are being omitted. I happen to have the book (The Pattern of Evolution) from which Morris mined that particular quote. Here’s the rest of it: (Remember, by “biota” we mean the commonly preserved plants and animals of a particular geological interval, which occupy regions often as large as Roger Tory Peterson’s “eastern” region of North American birds.) And when these systems change – when the older species disappear, and new ones take their place – the change happens relatively abruptly and in lockstep fashion.”

Eldridge was one of the authors (with Gould) of punctuated equilibrium theory. This states that, if you look at the fossil record, what we see are species emerging, persisting with little change for a while, and then disappearing from the fossil record. They theorize that most species most of the time are at equilibrium with their environment, and so do not change much. But these periods of equilibrium are punctuated by disequilibrium – periods of change when species will have to migrate, evolve, or go extinct.

This does not mean that speciation does not take place. And if you look at the fossil record we see a pattern of descendant species emerging from ancestor species over time – in a nice evolutionary pattern. Morris gives a complete misrepresentation of Eldridge’s point – once again we see intellectual dishonesty in his methods of an astounding degree.

Regarding the atheism = religion comment, it reminds me of a great analogy that I first heard on twitter from Evil Eye. (paraphrase) “those that say atheism is a religion, is like saying ‘not collecting stamps’ is a hobby too.”

Morris next tackles the genetic evidence, writing: More often is the argument used that similar DNA structures in two different organisms proves common evolutionary ancestry. Neither argument is valid. There is no reason whatever why the Creator could not or would not use the same type of genetic code based on DNA for all His created life forms. This is evidence for intelligent design and creation, not evolution.

Here is an excellent summary of the multiple lines of molecular evidence for evolution. Basically, if we look at the sequence of DNA, the variations in trinucleotide codes for amino acids, and amino acids for proteins, and transposons within DNA we see a pattern that can only be explained by evolution (or a mischievous god who chose, for some reason, to make life look exactly as if it had evolved – a non-falsifiable notion).

The genetic code is essentially comprised of four letters (ACGT for DNA), and every triplet of three letters equates to a specific amino acid. There are 64 (4^3) possible three letter combinations, and 20 amino acids. A few combinations are used for housekeeping, like a code to indicate where a gene stops, but the rest code for amino acids. There are more combinations than amino acids, so most amino acids are coded for by multiple combinations. This means that a mutation that results in a one-letter change might alter from one code for a particular amino acid to another code for the same amino acid. This is called a silent mutation because it does not result in any change in the resulting protein.

It also means that there are very many possible codes for any individual protein. The question is – which codes out of the gazillions of possible codes do we find for each type of protein in different species. If each “kind” were created separately there would not need to be any relationship. Each kind could have it’s own variation, or they could all be identical if they were essentially copied (plus any mutations accruing since creation, which would be minimal). But if life evolved then we would expect that the exact sequence of DNA code would be similar in related species, but progressively different (through silent mutations) over evolutionary time.

This is precisely what we find – in every protein we have examined. This pattern is necessary if evolution were true. It cannot be explained by random chance (the probability is absurdly tiny – essentially zero). And it makes no sense from a creationist perspective. This same pattern (a branching hierarchy) emerges when we look at amino acid substitutions in proteins and other aspects of the genetic code.

Morris goes for the second law of thermodynamics again – in the exact way that I already addressed. He responds to scientists correctly pointing out that the Earth is an open system, by writing: This naive response to the entropy law is typical of evolutionary dissimulation. While it is true that local order can increase in an open system if certain conditions are met, the fact is that evolution does not meet those conditions. Simply saying that the earth is open to the energy from the sun says nothing about how that raw solar heat is converted into increased complexity in any system, open or closed. The fact is that the best known and most fundamental equation of thermodynamics says that the influx of heat into an open system will increase the entropy of that system, not decrease it. All known cases of decreased entropy (or increased organization) in open systems involve a guiding program of some sort and one or more energy conversion mechanisms.

Energy has to be transformed into a usable form in order to do the work necessary to decrease entropy. That’s right. That work is done by life. Plants take solar energy (again – I’m not sure what “raw solar heat” means) and convert it into food. That food fuels the processes of life, which include development and reproduction. Evolution emerges from those processes- therefore the conditions that Morris speaks of are met.

But Morris next makes a very confused argument: Evolution has neither of these. Mutations are not “organizing” mechanisms, but disorganizing (in accord with the second law). They are commonly harmful, sometimes neutral, but never beneficial (at least as far as observed mutations are concerned). Natural selection cannot generate order, but can only “sieve out” the disorganizing mutations presented to it, thereby conserving the existing order, but never generating new order.

The notion that evolution (as if it’s a thing) needs to use energy is hopelessly confused. Evolution is a process that emerges from the system of life – and life certainly can use solar energy to decrease its entropy, and by extension the entropy of the biosphere. Morris slips into what is often presented as an information argument.  (Yet again – already dealt with. The pattern here is that we are seeing a shuffling around of the same tired creationists arguments.) It is first not true that most mutations are harmful. Many are silent, and many of those that are not silent are not harmful. They may be neutral, they may be a mixed blessing, and their relative benefit vs harm is likely to be situational. They may be fatal. And they also may be simply beneficial.

Morris finishes with a long rambling argument that evolution is religion. Evolution is promoted by its practitioners as more than mere science. Evolution is promulgated as an ideology, a secular religion — a full-fledged alternative to Christianity, with meaning and morality . . . . Evolution is a religion. This was true of evolution in the beginning, and it is true of evolution still today. Morris ties evolution to atheism, which, he argues, makes it a religion. This assumes, of course, that atheism is a religion. That depends on how you define atheism and how you define religion – but it is mostly wrong. Atheism is a lack of belief in one particular supernatural claim – that does not qualify it as a religion.

But mutations are not “disorganizing” – that does not even make sense. It seems to be based on a purely creationist notion that species are in some privileged perfect state, and any mutation can only take them farther from that perfection. For those who actually understand biology, life is a kluge of compromises and variation. Mutations are mostly lateral moves from one chaotic state to another. They are not directional. But they do provide raw material, variation, for natural selection. Natural selection cannot generate variation, but it can select among that variation to provide differential survival. This is an old game played by creationists – mutations are not selective, and natural selection is not creative (does not increase variation). These are true but irrelevant, because mutations increase variation and information, and selection is a creative force that results in the differential survival of better adapted variation.

What's the answer for environmentalists? Change the message and frame the issue in a way that doesn't trigger unconscious opposition among so many Americans. That can be a simple as using the right labels: a recent study by researchers at the University of Michigan found that Republicans are less skeptical of "climate change" than "global warming," possibly because climate change sounds less specific. Possibly too because so broad a term includes the severe snowfalls of the past winter that can be a paradoxical result of a generally warmer world. Greens should also pin their message on subjects that are less controversial, like public health or national security. Instead of issuing dire warnings about an apocalyptic future — which seems to make many Americans stop listening — better to talk about the present generation's responsibility to the future, to bequeath their children and grandchildren a safer and healthy planet.

Group identification also plays a major role in how we make decisions — and that's another way facts can get filtered. Declining belief in climate science has been, for the most part in America, a conservative phenomenon. On the surface, that's curious: you could expect Republicans to be skeptical of economic solutions to climate change like a carbon tax, since higher taxes tend to be a Democratic policy, but scientific information ought to be non-partisan. Politicians never debate the physics of space travel after all, even if they argue fiercely over the costs and priorities associated with it. That, however, is the power of group thinking; for most conservative Americans, the very idea of climate science has been poisoned by ideologues who seek to advance their economic arguments by denying scientific fact. No additional data — new findings about CO2 feedback loops or better modeling of ice sheet loss — is likely to change their mind.

The bright side of all this irrationality is that it means human beings can act in ways that sometimes go against their immediate utility, sacrificing their own interests for the benefit of the group.

Our brains develop socially, not just selfishly, which means sustainable behavior — and salvation for the planet — may not be as difficult as it sometimes seem. We can motivate people to help stop climate change — it may just not be climate science that convinces them to act.

Anyone who really values the progressive civilising potential of science should argue (in a qualified way as here) against Beddington's intemperate call for "complete intolerance" of scepticism. It is the social and human realities shared by politicians, non-government organisations, journalists and scientists themselves, that make tolerance of scepticism so important. The priorities pursued in scientific research and the directions taken by technology are all as fundamentally political as other areas of policy. No matter how uncomfortable and messy the resulting debates may sometimes become, we should never be cowed by any special interest—including that of scientific institutions—away from debating these issues in open, rational, democratic ways. To allow this to happen would be to undermine science itself in the most profound sense. It is the upholding of an often imperfect pursuit of scepticism and tolerance that offer the best way to respect and promote science. Such a position is, indeed, much more in keeping with the otherwise-exemplary work of John Beddington himself.Stirling's eloquent response provides a nice tonic to Beddington's unsettling remarks. Nonetheless, Beddington's perspective should be taken as a clear warning as to the pathological state of highly politicized science these days.

Medical science, when it works well, tends to progress from basic science, to small pilot studies, to larger randomized studies, and then–only then–to those big, rigorous, insanely expensive randomized, double-blind, placebo-controlled trials. Dr. Spector mentions hierarchies of evidence, but he seems to fall into a false dichotomy, namely that if it’s not Level I evidence, it’s crap. The problem is, as Mark pointed out, in medicine we often don’t have Level I evidence for many questions. Indeed, for some questions, we will never have Level I evidence. Clinical medicine involves making decisions in the midst of uncertainty, sometimes extreme uncertainty.

Dr. Spector then proceeds to paint a picture of reckless physicians proceeding on crappy studies to pump women full of hormones. Actually, it was more than a bit more complicated on than that. That was the time when I was in my medical training, and I remember the discussions we had regarding the strength (or lack thereof) of the epidemiological data and the lack of good RCTs looking at HRT. I also remember that nothing works as well to relieve menopausal symptoms as HRT, an observation we have been reminded of again since 2003, which is the year when the first big study came out implicating HRT in increasing the risk of breast cancer (more later).

I found a rather fascinating editorial in the New England Journal of Medicine from more than 20 years ago that discussed the state of the evidence back then with regard to estrogen and breast cancer: Evidence that estrogen increases the risk of breast cancer has been surprisingly difficult to obtain. Clinical and epidemiologic studies and studies in animals strongly suggest that endogenous estrogen plays a part in causing breast cancer. If so, exogenous estrogen should be a potent promoter of breast cancer. Although more than 20 case–control and prospective studies of the relation of breast cancer and noncontraceptive estrogen use have failed to demonstrate the expected association, relatively few women in these studies used estrogen for extended periods. Studies of the use of diethylstilbestrol and oral contraceptives suggest that a long exposure or latency may be necessary to show any association between hormone use and breast cancer. In the Swedish study, only six years of follow-up was needed to demonstrate an increased risk of breast cancer with the postmenopausal use of estradiol. It should be noted, however, that half the women in the subgroup that provided detailed data on the duration of hormone use had taken estrogen for many years before their base-line prescription status was defined. The duration of estrogen exposure in these women before the diagnosis of breast cancer was probably seriously underestimated; a short latency cannot be attributed to estradiol on the basis of these data. Other recent studies of the use of noncontraceptive estrogen suggest a slightly increased risk of breast cancer after 15 to 20 years’ use.

even now, the evidence is conflicting regarding HRT and breast cancer, with the preponderance of evidence suggesting that mixed HRT (estrogen and progestin) significantly increases the risk of breast cancer, while estrogen-alone HRT very well might not increase the risk of breast cancer at all or (more likely) only very little. Indeed, I was just at a conference all day Saturday where data demonstrating this very point were discussed by one of the speakers. None of this stops Dr. Spector from categorically labeling estrogen as a “carcinogen that causes breast cancers that kill women.” Maybe. Maybe not. It’s actually not that clear. The problem, of course, is that, consistent with the first primary reports of WHI results, the preponderance of evidence finding health risks due to HRT have indicted the combined progestin/estrogen combinations as unsafe.

technology in the classroom fails with respect to each of the three criteria: (a) technology is not a causal factor in learning in the sense that more technology means more learning, (b) assessment of educational outcome sis itself difficult and contested, much less disentangling various causal factors, and (c) the lack of evidence that technology leads to improved educational outcomes means that there is no such standardized technological core.

This conundrum calls into question one of the most significant contemporary educational movements. Advocates for giving schools a major technological upgrade — which include powerful educators, Silicon Valley titans and White House appointees — say digital devices let students learn at their own pace, teach skills needed in a modern economy and hold the attention of a generation weaned on gadgets. Some backers of this idea say standardized tests, the most widely used measure of student performance, don’t capture the breadth of skills that computers can help develop. But they also concede that for now there is no better way to gauge the educational value of expensive technology investments.

absent clear proof, schools are being motivated by a blind faith in technology and an overemphasis on digital skills — like using PowerPoint and multimedia tools — at the expense of math, reading and writing fundamentals. They say the technology advocates have it backward when they press to upgrade first and ask questions later.

[D]emand for educated labour is being reconfigured by technology, in much the same way that the demand for agricultural labour was reconfigured in the 19th century and that for factory labour in the 20th. Computers can not only perform repetitive mental tasks much faster than human beings. They can also empower amateurs to do what professionals once did: why hire a flesh-and-blood accountant to complete your tax return when Turbotax (a software package) will do the job at a fraction of the cost? And the variety of jobs that computers can do is multiplying as programmers teach them to deal with tone and linguistic ambiguity. Several economists, including Paul Krugman, have begun to argue that post-industrial societies will be characterised not by a relentless rise in demand for the educated but by a great “hollowing out”, as mid-level jobs are destroyed by smart machines and high-level job growth slows. David Autor, of the Massachusetts Institute of Technology (MIT), points out that the main effect of automation in the computer era is not that it destroys blue-collar jobs but that it destroys any job that can be reduced to a routine. Alan Blinder, of Princeton University, argues that the jobs graduates have traditionally performed are if anything more “offshorable” than low-wage ones. A plumber or lorry-driver’s job cannot be outsourced to India.

patents are designed to make that sort of thing more difficult, because it assumes that the initial act of invention is the key point, rather than all the incremental innovations built on top of it that all parties can benefit from.

the report points to numerous studies that show, when given the chance, many companies freely share their ideas with others, recognizing the direct benefit they get.

Even more importantly, the paper finds that due to technological advances and the ability to more rapidly and easily communicate and collaborate widely, these forms of innovation (innovation for direct use as well as collaborative innovation) are becoming more and more viable across a variety of industries, which in the past may have relied more on the old way of innovating (single company innovative for the profit of selling that product).

because of the ease of communication and collaboration these days, there's tremendous incentive for those companies that innovate for their own use to collaborate with others, since the benefit from others improving as well help improve their own uses. Thus, the overall incentives are to move much more to a collaborative form of innovation in the market. That has huge implications for a patent system designed to help the "old model" of innovation (producer inventing for the market) and not the increasingly regular one (collaborative innovation for usage).

no one is saying that producer-based innovation (company inventing to sell on the market) doesn't occur or won't continue to occur. But it is an open policy question as to whether or not our innovation policies should favor that model over other models -- when evidence suggests that a significant amount of innovation occurs in these other ways -- and that amount is growing rapidly.