January 16, 2017

Why do millennials earn some 20% less than boomers did at the same stage of life?

Filed under: Economics — Tags: , , , , — Nicholas @ 03:00

Tim Worstall explains why we shouldn’t be up in arms about the reported shortfall in millennial earnings compared to their parents’ generation at the same stage:

Part of the explanation here is that the millennials are better educated. We could take that to be some dig at what the snowflakes are learning in college these days but that’s not quite what I mean. Rather, they’re measuring the incomes of millennials in their late 20s. The four year college completion or graduation rate has gone up by some 50% since the boomers were similarly measured. Thus, among the boomers at that age there would be more people with a decade of work experience under their belt and fewer people in just the first few years of a professional career.

And here’s one of the things we know about blue collar and professional wages. Yes, the lifetime income as a professional is likely higher (that college wage premium and all that) but blue collar wages actually start out better and then don’t rise so much. Thus if we measure a society at the late 20s age and a society which has moved to a more professional wage structure we might well find just this result. The professionals making less at that age, but not over lifetimes, than the blue collar ones.


We’ve also got a wealth effect being demonstrated here. The millennials have lower net wealth than the boomers. Part of that is just happenstance of course. We’ve just had the mother of all recessions and housing wealth was the hardest hit part of it. And thus, given that housing equity is the major component of household wealth until the pension is fully topped up late in life, that wealth is obviously going to take a hit in the aftermath. There is another effect too, student debt. This is net wealth we’re talking about so if more of the generation is going to college more of the generation will have that negative wealth in the form of student debt. And don’t forget, it’s entirely possible to have negative net wealth here. For we don’t count the degree as having a wealth value but we do count the loans to pay for it as negative wealth.

January 1, 2017

Blog traffic in 2016

Filed under: Administrivia, Media — Tags: , , — Nicholas @ 03:00

The annual statistics update on traffic to Quotulatiousness from January 1st through December 31st, 2016. Overall, the traffic dropped slightly from 2015, which in turn was down a bit from the peak traffic year of 2014:

Over eight and a half million hits. That’s a pretty good number for an obscure Canadian blog.

The final count of visitors to the blog will be about 2,500-3,500 higher, as I did the screen captures at around 10:30 in the morning.

October 31, 2016

Is the “Gold Standard” of peer review actually just Fool’s Gold?

Filed under: Environment, Government, Health, Science — Tags: , , — Nicholas @ 01:00

Donna Laframboise points out that it’s difficult to govern based on scientific evidence if that evidence isn’t true:

We’re continually assured that government policies are grounded in evidence, whether it’s an anti-bullying programme in Finland, an alcohol awareness initiative in Texas or climate change responses around the globe. Science itself, we’re told, is guiding our footsteps.

There’s just one problem: science is in deep trouble. Last year, Richard Horton, editor of the Lancet, referred to fears that ‘much of the scientific literature, perhaps half, may simply be untrue’ and that ‘science has taken a turn toward darkness.’

It’s a worrying thought. Government policies can’t be considered evidence-based if the evidence on which they depend hasn’t been independently verified, yet the vast majority of academic research is never put to this test. Instead, something called peer review takes place. When a research paper is submitted, journals invite a couple of people to evaluate it. Known as referees, these individuals recommend that the paper be published, modified, or rejected.

If it’s true that one gets what one pays for, let me point out that referees typically work for no payment. They lack both the time and the resources to perform anything other than a cursory overview. Nothing like an audit occurs. No one examines the raw data for accuracy or the computer code for errors. Peer review doesn’t guarantee that proper statistical analyses were employed, or that lab equipment was used properly. The peer review process itself is full of serious flaws, yet is treated as if it’s the handmaiden of objective truth.

And it shows. Referees at the most prestigious of journals have given the green light to research that was later found to be wholly fraudulent. Conversely, they’ve scoffed at work that went on to win Nobel prizes. Richard Smith, a former editor of the British Medical Journal, describes peer review as a roulette wheel, a lottery and a black box. He points out that an extensive body of research finds scant evidence that this vetting process accomplishes much at all. On the other hand, a mountain of scholarship has identified profound deficiencies.

October 22, 2016

Polls, voting trends, and turnouts

Filed under: Politics, USA — Tags: , , , , — Nicholas @ 02:00

Jay Currie looks at the US election polling:

Polls tend to work by adjusting their samples to reflect demographics and an estimate of a given demographic’s propensity to actually vote. On a toy model basis, you can think of it as a layer cake with each layer representing an age cohort. So, for example, if you look at younger voters 18-29 you might find that 90% of them support Hilly and 10% Trump. If there are 100 of these voters in your sample of 500 a simple projection would suggest 90 votes for Hilly, 10 for Trump. The problem is that it is difficult to know how many of those younger voters will actually go out and vote. As a rule of thumb the older you are the more likely you are to vote so now you have to estimate voting propensity.

There are two ways to get a sense of voting propensity: ask the people in your sample or look at the behaviour of people the same age but in the last couple of elections.

And now the landscape begins to shift. In 2008, nearly 50% of voters aged 18-29 voted. In 2012, 40% voted. In both elections, the youth vote was heavily pro-Obama. If you were designing a poll at this point, what sort of weighting would make sense for youth voters? Making that call will change the landscape your poll will reflect. If you want your poll to tilt Hilly you can believe that the prospect of the first woman President of the United States will be as motivating as Obama was and assign a voting propensity of 40-50%; alternatively, if you don’t see many signs of Hillary catching fire among younger voters, you can set the propensity number at 30% and create a tie or a slight Trump lead.

(The results of this are even more dramatic if you look at the black vote and turnout. In 2008 black turnout was 69.1%, 2012, 67.4% with Obama taking well over 90%. Will the nice white lady achieve anything like these numbers?)

One the other side of the ledger, the turnouts of the less educated have been low for the last two elections. 52% in 2008 and a little less than 50 in 2012. There is room for improvement. Now, as any educated person will tell you, often at length, Trump draws a lot of support in the less educated cohorts. But that support is easily discounted because these people (the deplorables and their ilk) barely show up to vote.

Build your model on the basis that lower education people’s participation in 2016 will be similar to 2008 and 20012 and you will produce a result in line with the 538.com consensus view. But if you think that the tens of thousands people who show up for Trump’s rallies might just show up to vote, you will have a model tending towards the LA Times view of things.

October 1, 2016

Here’s some fantastic news you’re not seeing in the headlines

Filed under: Economics — Tags: , — Nicholas @ 02:00

The same world poverty data, presented as absolute or relative levels of poverty:



H/T to Rob Fisher at Samizdata for the link.

June 21, 2016

World War 1 in Numbers I THE GREAT WAR Special

Filed under: Cancon, Europe, History, Military — Tags: , , , — Nicholas @ 04:00

Published on 20 Jun 2016

Special thanks to Karim Theilgaard for composing the the new theme for our brand new intro!

We are approaching the 100th regular episode and decided to surprise you with an extra special episode about the staggering numbers of World War 1.

May 26, 2016

Eighty percent of Americans surveyed favour banning things they know nothing about

Filed under: Health, Media, Science, USA — Tags: , , , — Nicholas @ 03:00

Don’t get too smug, fellow Canuckistanis, as I suspect the numbers might be just as bad if Canadians were surveyed in this way:

You might have heard that Americans overwhelmingly favor mandatory labeling for foods containing genetically modified ingredients. That’s true, according to a new study: 84 percent of respondents said they support the labels.

Survey of GMO labelling fans

But a nearly identical percentage — 80 percent—in the same survey said they’d also like to see labels on food containing DNA.

Survey of DNA labelling fans

The study, published in the Federation of American Societies for Experimental Biology Journal last week, also found that 33 percent of respondents thought that non-GM tomatoes “did not contain genes” and 32 percent thought that “vegetables did not have DNA.” So there’s that.

University of Florida food economist Brandon R. McFadden and his co-author Jayson L. Lusk surveyed 1,000 American consumers and discovered [PDF] that “consumers think they know more than they actually do about GM food.” In fact, the authors say, “the findings question the usefulness of results from opinion polls as motivation for public policy surrounding GM food.”

My summary for laymen: When it comes to genetically modified food, people don’t know much, they don’t know what they don’t know, and they sure as heck aren’t letting that stop them from having strong opinions.

April 24, 2016

The “secret” of Indian food

Filed under: India, Science — Tags: , , — Nicholas @ 02:00

In an article in the Washington Post last year, Roberto Ferdman summarized the findings of a statistical study explaining why the flavours in Indian foods differ so much from other world cuisines:

Indian food, with its hodgepodge of ingredients and intoxicating aromas, is coveted around the world. The labor-intensive cuisine and its mix of spices is more often than not a revelation for those who sit down to eat it for the first time. Heavy doses of cardamom, cayenne, tamarind and other flavors can overwhelm an unfamiliar palate. Together, they help form the pillars of what tastes so good to so many people.

But behind the appeal of Indian food — what makes it so novel and so delicious — is also a stranger and subtler truth. In a large new analysis of more than 2,000 popular recipes, data scientists have discovered perhaps the key reason why Indian food tastes so unique: It does something radical with flavors, something very different from what we tend to do in the United States and the rest of Western culture. And it does it at the molecular level.


Chefs in the West like to make dishes with ingredients that have overlapping flavors. But not all cuisines adhere to the same rule. Many Asian cuisines have been shown to belie the trend by favoring dishes with ingredients that don’t overlap in flavor. And Indian food, in particular, is one of the most powerful counterexamples.

Researchers at the Indian Institute for Technology in Jodhpur crunched data on several thousand recipes from a popular online recipe site called TarlaDalal.com. They broke each dish down to its ingredients, and then compared how often and heavily ingredients share flavor compounds.

The answer? Not too often.

November 29, 2015

Does Teddy Bridgewater hold the ball too long?

Filed under: Football — Tags: , , , , — Nicholas @ 03:00

Over at Vikings Territory, Brett Anderson endangers his health, eyesight, and even his sanity by exhaustively tracking, timing, and analyzing every throw by Teddy Bridgewater in last week’s game against the Green Bay Packers. A common knock on Bridgewater is that he’s holding the ball too long and therefore missing pass opportunities and making himself more vulnerable to being sacked. It’s a long article, but you can skip right to the end to get the facts distilled:

What The Film Shows

It became clear pretty quickly that plays with larger TBH [time ball held] had a lot happening completely out of Bridgewater’s control. There were only a couple of plays where it clearly looked like Bridgewater held the ball too long while there were options downfield to target or that he hesitated to pull the trigger on guys that were open. And consistently, there were three issues I kept noticing.

  1. Receiver route depth – The Vikings receivers run a ton of late developing routes. I don’t have any numbers to back that up – we’re talking strictly film review now. But on plays ran out of the shotgun with 5-step drops or plays with even longer 7-step drops, by the time Bridgewater is being pressured (which happens about every 2 of 3 plays), his receivers have not finished their routes. And I know that just because they haven’t finished the route doesn’t mean Bridgewater can’t anticipate where they are going to be but… We’re talking not really even close to finishing their routes. It seems that a lot of the Vikings play designs consist of everybody running deep fade routes to create room underneath for someone on a short dig or to check down to a running back in the flat. So, if this player underneath is for any reason covered (or if the Vikings find themselves in long down and distance situations where an underneath route isn’t going to cut it, which… surprise, happens quite often), Bridgewater’s other receiver options are midway through their route 20 yards downfield. What’s worse? Not only are these routes taking forever to develop and typically only materializing once Bridgewater has been sacked or scampered away to save himself, but also…
  2. Receiver coverage – The Vikings receivers are typically not open. It was pretty striking how often on plays with higher TBH receivers have very little separation. (Make sure to take a look through the frame stills linked in the data table above. I tried to make sure I provided a capture for plays with higher TBH or plays that resulted in a negative outcome. Red circles obviously indicate receivers who are not open while yellow typically indicates receivers who are.) The Packers consistently had 7 defenders in coverage resulting in multiple occasions where multiple receivers are double teamed with safety help over the top. But even in plays with one on one coverage, the Vikings receivers are still having a difficult time finding space. So now, we have a situation with Bridgewater where we have these deep drops where not only are receivers not finished with their deep routes but they are also blanket covered. And why are teams able to drop so many players into coverage creating risky situations for a quarterback who is consistently risk adverse? Because…
  3. Poor offensive line play – The Vikings offensive line is not good. And it may be worse than you think. It’s no secret by this point that the Vikings offensive line had one of its worst showings of the year against the Packers. More often than not, simply by rushing four defenders, Green Bay was able to get pressure on Bridgewater within 2-3 seconds. This is a quick sack time. And more often than not, Bridgewater is having to evade this pressure by any means necessary to either give his receivers time to finish their routes or give them time to get open. (Or more frequently – both.) As a result of this, what we saw on multiple occasions against the Packers is Bridgewater being pressured quickly, him scrambling from the pocket and dancing around while stiff-arming a defender once or twice and ultimately throwing the ball out of bounds or taking a sack. Are you starting to see what the problem here?


Bridgewater is not holding the ball for a length of time that should reflect poorly on his play. The data shows that Bridgewater is about average when looking just strictly at the numbers. The tape shows a quarterback who really doesn’t have a lot of options other than holding on to the ball. When Bridgewater is presented with a quick 1- or 3-step drop and his receivers run routes with lengths complementary to the length of his drop, it typically results in Bridgewater finding a relatively open receiver, making a quick decision and getting the ball there accurately. When Bridgewater is faced with longer developing plays behind an offensive line that’s a sieve and receivers who are running lengthy routes while closely covered, he tries to make a play himself. Sure, there were a couple of plays during the Packers game where it may have been a better decision for Bridgewater to take a sack when initially pressured and saving the yards he lost by scrambling backwards. However, it’s difficult to chastise him for trying to create plays when they aren’t there when it doesn’t work and applauding him when his evasiveness, deadly stiff arm and surprisingly effective spin move result in a big play.

Bridgewater has been far from perfect this season. But after this extensive exercise, I can comfortably say that the amount of time Bridgewater is holding on to the ball should not negatively reflect on his performance considering the above mentioned external factors.

November 27, 2015

Wealth, inequality, and billionaires

Filed under: Economics, Government, Politics — Tags: , , , — Nicholas @ 04:00

Several months ago, the Washington Post reported on a new study of wealth and inequality that tracked how many billionaires got rich through competition in the market and how many got rich through political “connections”:

The researchers found that wealth inequality was growing over time: Wealth inequality increased in 17 of the 23 countries they measured between 1987 and 2002, and fell in only six, Bagchi says. They also found that their measure of wealth inequality corresponded with a negative effect on economic growth. In other words, the higher the proportion of billionaire wealth in a country, the slower that country’s growth. In contrast, they found that income inequality and poverty had little effect on growth.

The most fascinating finding came from the next step in their research, when they looked at the connection between wealth, growth and political connections.

The researchers argue that past studies have looked at the level of inequality in a country, but not why inequality occurs — whether it’s a product of structural inequality, like political power or racism, or simply a product of some people or companies faring better than others in the market. For example, Indonesia and the United Kingdom actually score similarly on a common measure of inequality called the Gini coefficient, say the authors. Yet clearly the political and business environments in those countries are very different.

So Bagchi and Svejnar carefully went through the lists of all the Forbes billionaires, and divided them into those who had acquired their wealth due to political connections, and those who had not. This is kind of a slippery slope — almost all billionaires have probably benefited from government connections at one time or another. But the researchers used a very conservative standard for classifying people as politically connected, only assigning billionaires to this group when it was clear that their wealth was a product of government connections. Just benefiting from a government that was pro-business, like those in Singapore and Hong Kong, wasn’t enough. Rather, the researchers were looking for a situation like Indonesia under Suharto, where political connections were usually needed to secure import licenses, or Russia in the mid-1990s, when some state employees made fortunes overnight as the state privatized assets.

The researchers found that some countries had a much higher proportion of billionaire wealth that was due to political connections than others did. As the graph below, which ranks only countries that appeared in all four of the Forbes billionaire lists they analyzed, shows, Colombia, India, Australia and Indonesia ranked high on the list, while the U.S. and U.K. ranked very low.

Wealth and political connections

Looking at all the data, the researchers found that Russia, Argentina, Colombia, Malaysia, India, Australia, Indonesia, Thailand, South Korea and Italy had relatively more politically connected wealth. Hong Kong, the Netherlands, Singapore, Sweden, Switzerland and the U.K. all had zero politically connected billionaires. The U.S. also had very low levels of politically connected wealth inequality, falling just outside the top 10 at number 11.

When the researchers compared these figures to economic growth, the findings were clear: These politically connected billionaires weighed on economic growth. In fact, wealth inequality that came from political connections was responsible for nearly all the negative effect on economic growth that the researchers had observed from wealth inequality overall. Wealth inequality that wasn’t due to political connections, income inequality and poverty all had little effect on growth.

November 14, 2015

The scandal of NCAA “graduation” rates

Filed under: Football, Sports — Tags: , , , — Nicholas @ 02:00

Gregg Easterbrook on the statistical sleight-of-hand that allows US universities to claim unrealistic graduation rates for their student athletes:

N.C.A.A. Graduation Rate Hocus-Pocus. [Hawaii coach Norm] Chow and [Maryland coach Randy] Edsall both made bona fide improvements to the educational quality of their college football programs, and both were fired as thanks. Edsall raised Maryland’s football graduation rate from 56 percent five years ago to 70 percent. Chow raised Hawaii’s football graduation rate from 29 percent five years ago to 50 percent.

At least that’s what the Department of Education says. According to the N.C.A.A., Hawaii graduates not 50 percent of its players but 70 percent, while Maryland graduates not 70 percent but 75 percent.

At work is the distinction between the Federal Graduation Rate, calculated by the Department of Education, and the Graduation Success Rate, calculated by the N.C.A.A. No other aspect of higher education has a graduation “success rate” — just a graduation rate. The N.C.A.A. cooks up this number to make the situation seem better than it is.

The world of the Graduation Success Rate is wine and roses: According to figures the N.C.A.A. released last week, 86 percent of N.C.A.A. athletes achieved “graduation success” in the 2014-2015 academic year. But “graduation success” is different from graduating; the Department of Education finds that 67 percent of scholarship athletes graduated in 2014-2015. (These dueling figures are for all scholarship athletes: Football and men’s basketball players generally are below the average, those in other sports generally above.)

Both the federal and N.C.A.A. calculations have defects. The federal figure scores only those who graduate from the college of their initial enrollment. The athlete who transfers and graduates elsewhere does not count in the federal metric.

The G.S.R., by contrast, scores as a “graduate” anyone who leaves a college in good standing, via transfer or simply giving up on school: There’s no attempt to follow-up to determine whether athletes who leave graduate somewhere else. Not only is the N.C.A.A.’s graduation metric anchored in the absurd assumption that leaving a college is the same as graduating, but it can also reflect a double-counting fallacy. Suppose a football player starts at College A, transfers to College B and earns his diploma there. Both schools count him as a graduate under the G.S.R.


Football players ought to graduate at a higher rate than students as a whole. Football scholarships generally pay for five years on campus plus summer school, and football scholarship holders never run out of tuition money, which is the most common reason students fail to complete college. Instead at Ohio State and other money-focused collegiate programs, players graduate at a lower rate than students as a whole. To divert attention from this, the N.C.A.A. publishes its annual hocus-pocus numbers.

October 28, 2015

The WHO’s lack of clarity leads to sensationalist newspaper headlines (again)

Filed under: Health, Media, Science — Tags: , , , , , — Nicholas @ 05:00

The World Health Organization appears to exist primarily to give newspaper editors the excuse to run senational headlines about the risk of cancer. This is not a repeat story from earlier years. Oh, wait. Yes it is. Here’s The Atlantic‘s Ed Yong to de-sensationalize the recent scary headlines:

The International Agency of Research into Cancer (IARC), an arm of the World Health Organization, is notable for two things. First, they’re meant to carefully assess whether things cause cancer, from pesticides to sunlight, and to provide the definitive word on those possible risks.

Second, they are terrible at communicating their findings.


Group 1 is billed as “carcinogenic to humans,” which means that we can be fairly sure that the things here have the potential to cause cancer. But the stark language, with no mention of risks or odds or any remotely conditional, invites people to assume that if they specifically partake of, say, smoking or processed meat, they will definitely get cancer.

Similarly, when Group 2A is described as “probably carcinogenic to humans,” it roughly translates to “there’s some evidence that these things could cause cancer, but we can’t be sure.” Again, the word “probably” conjures up the specter of individual risk, but the classification isn’t about individuals at all.

Group 2B, “possibly carcinogenic to humans,” may be the most confusing one of all. What does “possibly” even mean? Proving a negative is incredibly difficult, which is why Group 4 — “probably not carcinogenic to humans” — contains just one substance of the hundreds that IARC has assessed.

So, in practice, 2B becomes a giant dumping ground for all the risk factors that IARC has considered, and could neither confirm nor fully discount as carcinogens. Which is to say: most things. It’s a bloated category, essentially one big epidemiological shruggie. But try telling someone unfamiliar with this that, say, power lines are “possibly carcinogenic” and see what they take away from that.

Worse still, the practice of lumping risk factors into categories without accompanying description — or, preferably, visualization — of their respective risks practically invites people to view them as like-for-like. And that inevitably led to misleading headlines like this one in the Guardian: “Processed meats rank alongside smoking as cancer causes – WHO.”

October 15, 2015

S.L.A. Marshall, Dave Grossman, and the “man is naturally peaceful” meme

Filed under: Books, Cancon, History, Military, USA — Tags: , , , , — Nicholas @ 03:00

The American military historian S.L.A. Marshall was perhaps best known for his book Men Against Fire: The Problem of Battle Command in Future War, where he argued that American military training was insufficient to overcome most men’s natural hesitation to take another human life, even in intense combat situations. Dave Grossman is a modern military author who draws much of his conclusions from the initial work of Marshall. Grossman’s case is presented in his book On Killing: The Psychological Cost of Learning to Kill in War and Society, which was reviewed by Robert Engen in an older issue of the Canadian Military Journal:

As a military historian, I am instinctively skeptical of any work or theory that claims to overturn all existing scholarship – indeed, overturn an entire academic discipline – in one fell swoop. In academic history, the field normally expands and evolves incrementally, based upon new research, rather than being completely overthrown periodically. While it is not impossible for such a revolution to take place and become accepted, extraordinary new research and evidence would need to be presented to back up these claims. Simply put, Grossman’s On Killing and its succeeding “killology” literature represent a potential revolution for military history, if his claims can stand up to scrutiny – especially the claim that throughout human history, most soldiers and people have been unable to kill one another.

I will be the first to acknowledge that Grossman has made positive contributions to the discipline. On Combat, in particular, contains wonderful insights on the physiology of combat that bear further study and incorporation within the discipline. However, Grossman’s current “killology” literature contains some serious problems, and there are some worrying flaws in the theories that are being preached as truth to the men and women of the Canadian Forces. Although much of Grossman’s work is credible, his proposed theories on the inability of human beings to kill one another, while optimistic, are not sufficiently reinforced to warrant uncritical acceptance. A reassessment of the value that this material holds for the Canadian military is necessary.

The evidence seems to indicate that, contrary to Grossman’s ideas, killing is a natural, if difficult, part of human behaviour, and that killology’s belief that soldiers and the population at large are only being able to kill as part of programmed behaviour (or as a symptom of mental illness) hinders our understanding of the actualities of warfare. A flawed understanding of how and why soldiers can kill is no more helpful to the study of military history than it is to practitioners of the military profession. More research in this area is required, and On Killing and On Combat should be treated as the starting points, rather than the culmination, of this process.


September 5, 2015

The subtle lure of “research” that confirms our biases

Filed under: Health, Science — Tags: , , — Nicholas @ 02:00

Megan McArdle on why we fall for bogus research:

Almost three years ago, Nobel Prize-winning psychologist Daniel Kahneman penned an open letter to researchers working on “social priming,” the study of how thoughts and environmental cues can change later, mostly unrelated behaviors. After highlighting a series of embarrassing revelations, ranging from outright fraud to unreproducible results, he warned:

    For all these reasons, right or wrong, your field is now the poster child for doubts about the integrity of psychological research. Your problem is not with the few people who have actively challenged the validity of some priming results. It is with the much larger population of colleagues who in the past accepted your surprising results as facts when they were published. These people have now attached a question mark to the field, and it is your responsibility to remove it.

At the time it was a bombshell. Now it seems almost delicate. Replication of psychology studies has become a hot topic, and on Thursday, Science published the results of a project that aimed to replicate 100 famous studies — and found that only about one-third of them held up. The others showed weaker effects, or failed to find the effect at all.

This is, to put it mildly, a problem. But it is not necessarily the problem that many people seem to assume, which is that psychology research standards are terrible, or that the teams that put out the papers are stupid. Sure, some researchers doubtless are stupid, and some psychological research standards could be tighter, because we live in a wide and varied universe where almost anything you can say is certain to be true about some part of it. But for me, the problem is not individual research papers, or even the field of psychology. It’s the way that academic culture filters papers, and the way that the larger society gets their results.

August 29, 2015

We need a new publication called The Journal of Successfully Reproduced Results

Filed under: Media, Science — Tags: , , , , — Nicholas @ 04:00

We depend on scientific studies to provide us with valid information on so many different aspects of life … it’d be nice to know that the results of those studies actually hold up to scrutiny:

One of the bedrock assumptions of science is that for a study’s results to be valid, other researchers should be able to reproduce the study and reach the same conclusions. The ability to successfully reproduce a study and find the same results is, as much as anything, how we know that its findings are true, rather than a one-off result.

This seems obvious, but in practice, a lot more work goes into original studies designed to create interesting conclusions than into the rather less interesting work of reproducing studies that have already been done to see whether their results hold up.

Everyone wants to be part of the effort to identify new and interesting results, not the more mundane (and yet potentially career-endangering) work of reproducing the results of older studies:

Why is psychology research (and, it seems likely, social science research generally) so stuffed with dubious results? Let me suggest three likely reasons:

A bias towards research that is not only new but interesting: An interesting, counterintuitive finding that appears to come from good, solid scientific investigation gets a researcher more media coverage, more attention, more fame both inside and outside of the field. A boring and obvious result, or no result, on the other hand, even if investigated honestly and rigorously, usually does little for a researcher’s reputation. The career path for academic researchers, especially in social science, is paved with interesting but hard to replicate findings. (In a clever way, the Reproducibility Project gets around this issue by coming up with the really interesting result that lots of psychology studies have problems.)

An institutional bias against checking the work of others: This is the flipside of the first factor: Senior social science researchers often actively warn their younger colleagues — who are in many cases the best positioned to check older work—against investigating the work of established members of the field. As one psychology professor from the University of Southern California grouses to the Times, “There’s no doubt replication is important, but it’s often just an attack, a vigilante exercise.”


Small, unrepresentative sample sizes: In general, social science experiments tend to work with fairly small sample sizes — often just a few dozen people who are meant to stand in for everyone else. Researchers often have a hard time putting together truly representative samples, so they work with subjects they can access, which in a lot of cases means college students.

A couple of years ago, I linked to a story about the problem of using western university students as the default source of your statistical sample for psychological and sociological studies:

A notion that’s popped up several times in the last couple of months is that the easy access to willing test subjects (university students) introduces a strong bias to a lot of the tests, yet until recently the majority of studies disregarded the possibility that their test results were unrepresentative of the general population.

Older Posts »

Powered by WordPress

%d bloggers like this: