William M. Briggs

Statistician to the Stars!

Page 153 of 410

Scientific Truths Are Not Better Truths Than Just-Plain Truths

One of the key fallacies of scientism, in the sense of being the most destructive to common sense and personal wellbeing, is to suppose that any theory put forth in the name of science is therefore true, or certain enough to believe as true. The posited theory is, after all, scientific and, so scientism says, there is no better recommendation to truth than this.

This fallacy is field dependent, cropping up in some areas much more frequently than in others. It is rare, though still frequent enough in a mild sense, to find the speculations of chemists being refuted each new generation. But it is common as daylight to find the hypotheses put forth by sociologists, economists, and psychologists refuted not a generation after they are published, but often in the next issue of the same journal.

For example, the Weekly Standard’s Andrew Ferguson’s reviews the bad science and scientism behind the recent spate of experimentation “proving” conservatives are dumber/more inflexible/less compassionate/etc. than liberals, theories which are collected in Chris Mooney’s hagiography to scientism, The Republican Brain. (We began a collection of these studies here; please contribute. Also see Mike Flynn’s take on this.)

The studies rely on the principle that has informed the social sciences for more than a generation: If a researcher with a Ph.D. can corral enough undergraduates into a campus classroom and, by giving them a little bit of money or a class credit, get them to do something—fill out a questionnaire, let’s say, or pretend they’re in a specific real-world situation that the researcher has thought up—the young scholars will (unconsciously!) yield general truths about the human animal; scientific truths.

Although he didn’t intend it, Ferguson shows us another fallacy, which is that, in virtue of their being generated by scientists, that “scientific” truths are better than other kinds of truths, say metaphysical or logical truths. Stating it so plainly makes it obvious that if a truth is a truth, then it is a truth, and a truth is not more “truthy” because it comes from a scientist than a truth which comes from (say) a theologian.

The main problem with this summary is that often scientists use the word truth to mean “a belief which is probably but not 100% certainly, no-matter-what true.” That later creature is not a truth at all; it is a conjecture and nothing more. A conjecture which is “almost” true, or “for all practical purposes” true, is still a conjecture and not a truth. A truth is only true when it always is, when it can be deduced, when it arises as the end result of a valid argument. That is, conjectures when they are conjectures are not truths, but conjectures might become truth as new evidence arises.

Physicists make the mistake of confusing truth and conjecture just as often as sociologists and psychologists, only the physicists’ conjectures more often turn out to be truths as that new evidence arrives, so their error is of less consequence. Note that it is an error (a fallacy) to say, given evidence less than deductive, that a conjecture is a truth. The error will turn out to be more or less harmful depending on to what use the conjecture is put. If one is betting that a protein will fold a hypothesized way, one has turned a conjecture into a forecast, which is a term that acknowledges a conjecture is less than a truth.

If the conjecture turns out true, because of new evidence from an experiment, then the conjecture turns into a truth and gains are made. If the conjecture turns out false, we again know this based on new evidence, and loses are suffered. The loses are of particular interest: these are less the closer the conjecture is to the truth. Which is why the loses are greater in the soft sciences: their conjectures are much more often farther from truth.

One reason for the difference is that physicists more often than sociologists test their conjectures against reality. Another reason is that the evidence for a conjecture for the hard sciences is not just statistical, as it often is for soft-science conjectures. And any conjecture which relies primarily on statistics—given, that is, how statistics is practiced today—should not be trusted.

I’ll insert my usual plea that soft scientists act more like their hard (knock) brothers. Do not just assemble a one-time shot of data and compute some statistical model and tell us how well that model fits your data, and then assume because this fit is “good” that therefore your conjectures are true. This is formally a fallacy and is the weakest kind of evidence there is, but (almost universally) the only kind which is offered.

Instead do two things: (1) find very effort to discover evidence which refutes your conjecture (and then tell us abou it). And (2) as hard scientists (often) do, make predictions of data you have never before seen. If you do both these things, then you can ask us to believe your conjectures. Otherwise, keep quiet.

————————————————————————————

Thanks to the many readers who pointed me to Ferguson’s article, including Mike Flynn and Doug Magowan.

What Organic Boors, Swedish Pronouns, And The Exorcist Have In Common

The ExorcistNot much, except to demonstrate that the natural state of modernity is something closely resembling mass lunacy. To explain.

Swedish Pronouns

Via Sam Schulman ‏(Twitter: @Sam_Schulman) we infer that 1984 has not been translated in Swedish. This is an inference, mind, and not a direct claim. But you’ll agree it is a likely one after you learn that in Sweden, “it is now considered a distinct discrimination if one is addressed as a man or woman.” So reports Kopp Online.

Sweden is angling to de-genderify their pronouns so that use of he or she is officially discouraged, to be replaced by something resembling it (hen). Not only does this move strip useful information from its language, the Swedes have made an important step in subtracting from a person’s humanity, since to be called an “it” is to be equated to a chair or a bug. Now that, dear reader, is true equality. And it is under the banner of Equality that these changes are being made.

Ho hum you say? Then consider that “Nyamko Sabuni, currently Minister for Integration and Gender Equality, is now trying politically to free children from the constraints of gender roles.” Sabuni is leading an effort to de-genderify names, so that if he succeeds Sweden will be no longer have the equivalent of Bills and Janes, but will be a nation of only Pats and Chrises.

There is in Sweden a kindergarten with the telling name of Egalia where “gender-free” children are taught the joys of homosexuality and to play house, imagining, for instance, that there are “two or three mothers.” In a separate article, the wardens of this institution justify their experimentation by claiming that “gender” is not something which you are born with, but is something which can be “changed at any time.” This being so, the little tots should learn early how to do this morphing.

Organic Boors

Today reports of new “research” which confirms what everybody who has shopped in an urban farmer’s market already knew (“research” is needed because everybody did not have a p-value to accompany their belief). The news is that organic food boors are often bullies, that they are often self-satisfied “snotty and arrogant” moralists.

“There’s a line of research showing that when people can pat themselves on the back for their moral behavior, they can become self-righteous,” says author Kendall Eskine, assistant professor of the department of psychological sciences at Loyola University in New Orleans. “I’ve noticed a lot of organic foods are marketed with moral terminology, like Honest Tea, and wondered if you exposed people to organic food, if it would make them pat themselves on the back for their moral and environmental choices. I wondered if they would be more altruistic or not.”

Eskine found that “organic people judged much harder” than ordinary humans, that when “it came to helping out a needy stranger, the organic people also proved to be more selfish.” The money quote:

“There’s something about being exposed to organic food that made them feel better about themselves,” says Eskine. “And that made them kind of jerks a little bit, I guess.”

The only surprise is that Eskine found his results surprising: “You’d think eating organic would make you feel elevated and want to pay it forward,” he said.

He (I assume Eskine is a “he” because his first name is Kendall) should have gone to any Whole Foods…

HT Hot Air.

The Exorcist

Stay with me for a long story made short: Health and Human Secretary and self-labeled Catholic Kathleen Sebelius was invited to speak at one of Georgetown University’s commencement ceremonies, which was, or is, kind of, affiliated with the Catholic Church. This was controversial because Sebelius instituted a “mandate” which said that Catholic employers must provide (via “health” “insurance”) their employees abortion-inducing drugs and contraception. Recall poor Sandra Fluke and her plea for somebody—anybody but herself—to fund her birth control.

Abortion and birth control, whether you are for them or not, are against Catholic doctrine, meaning that any Catholic institution had no business honoring a woman like Sebelius who knowingly “mandated” a removal of religious freedom. Sebelius, in a weak attempt to justify her curious behavior and stick in the eye of her critics, in her speech said,

[President John] Kennedy talked about his vision of religion and the public square, and said he believed in an America, and I quote, “where no religious body seeks to impose its will directly or indirectly upon the general populace or the public acts of its officials — and where religious liberty is so indivisible that an act against one church is treated as an act against us all.”

Like many other things in the Obama government, Sebellius got it exactly backwards. There was no fear the Catholic church would foist its views on an innocent and unwilling public, but instead the reality of an omnipotent government forcing the Church to abandon its core beliefs. Shame on her for misusing Kennedy’s words.

Enter William Peter Blatty, author of The Exorcist and alumnus of Georgetown, who is organizing an official protest against Georgetown, which Blatty believes (if I may) is possessed by a spirit of idiocy.

Even if you are not Catholic and earnestly desire abortion for free, Blatty’s fight is yours. Kennedy was right: “an act against one church” is “as an act against us all.” If you let the government encroach upon our freedom because you lust for free contraception, you must know that this will not be the end and the government will soon come for what you treasure.

How To Present Anything As Significant. Update: Nature Weighs In

Update The paper on which this post was based has hit the presses which was used as an occasion for Nature to opine that, just maybe, that some of the softer sciences have a problem with replication. I thought it important enough to repost (the original was on 3 March 2012).

Nature’s article is Replication studies: Bad copy: In the wake of high-profile controversies, psychologists are facing up to problems with replication. The meat is found in these two paragraphs.

Positive results in psychology can behave like rumours: easy to release but hard to dispel. They dominate most journals, which strive to present new, exciting research. Meanwhile, attempts to replicate those studies, especially when the findings are negative, go unpublished, languishing in personal file drawers or circulating in conversations around the water cooler. “There are some experiments that everyone knows don’t replicate, but this knowledge doesn’t get into the literature,” says Wagenmakers. The publication barrier can be chilling, he adds. “I’ve seen students spending their entire PhD period trying to replicate a phenomenon, failing, and quitting academia because they had nothing to show for their time.”

These problems occur throughout the sciences, but psychology has a number of deeply entrenched cultural norms that exacerbate them. It has become common practice, for example, to tweak experimental designs in ways that practically guarantee positive results. And once positive results are published, few researchers replicate the experiment exactly, instead carrying out ‘conceptual replications’ that test similar hypotheses using different methods. This practice, say critics, builds a house of cards on potentially shaky foundations.

Ed Yong, who wrote this piece, also opines on fraud, especially the suspicion that this activity has been increasing. Yong’s piece, the Simmons et al. paper, and the post below are all well worth reading.

Thanks to James Glendinning for the head’s up.

———————————————————————————————————

false positives My heart soared like a hawk when I read Joseph P. Simmons, Leif D. Nelson, and Uri Simonsohn’s “False-Positive Psychology : Undisclosed Flexibility in Data Collection and Analysis Allows Presenting Anything as Significant“, published in Psychological Science1.

From their abstract:

[W]e show that despite empirical psychologists’ nominal endorsement of a low rate of false-positive findings (≤ .05), flexibility in data collection, analysis, and reporting dramatically increases actual false-positive rates. In many cases, a researcher is more likely to falsely find evidence that an effect exists than to correctly find evidence that it does not. We…demonstrate how unacceptably easy it is to accumulate (and report) statistically significant evidence for a false hypothesis.

Preach it, brothers! Sing it loud and sing it proud. Oh, how I wish that your colleagues will take your admonitions to heart and abandon the Cult of the P-value!

Rarely have I read such a quotable paper. False positives—that is, false “confirmations” of hypotheses—are “particularly persistent”; “because it is uncommon for prestigious journals to publish null findings or exact replications, researchers have little incentive to even attempt them.” False positives “can lead to ineffective policy changes.”

Many false positives are found because of the “researcher’s desire to find a statistically significant result.” That researchers are “are self-serving in their interpretation of ambiguous information and remarkably adept at reaching
justifiable conclusions that mesh with their desires” “Ambiguity is rampant in empirical research.”

Our goal as scientists is not to publish as many articles as we can, but to discover and disseminate truth.

I am man enough to admit that I wept when I read those words.

Chronological Rejuvenation

The authors include a hilarious—actual—study where they demonstrate that listening to a children’s song makes people younger. Not just feel younger, but younger chronologically. Is there nothing statistics cannot do?

They first had two groups listen to an adult song or a children’s song and then asked participants how old they felt afterwards. They also asked participants for their ages and their fathers’ ages “allowing us to control for variation in baseline age across participants.” They got a p-value of 0.033 “proving” that listening to the children’s song made people feel younger.

They then forced the groups to listen to a Beatles song or the same children’s song (they assumed there was a difference), and again asked the ages. “We used father’s age to control for variation in baseline age across participants.”

According to their birth dates, people were nearly a year-and-a-half younger after listening to “When I’m Sixty-Four” (adjusted M = 20.1 years) rather than to [the children's song] (adjusted M = 21.5 years), F(1, 17) = 4.92, p = .040.

Ha ha! How did they achieve such a deliciously publishable p-value for a necessarily false result? Because of the broad flexibility in classical statistics which allows users to “data mine” for the small p-values.

Ways to Cheat

The authors list six major mistakes that users of statistics make. They themselves used many of these mistakes in “proving” the results in the experiment above. (We have covered all of these before: click Start Here and look under Statistics.)

“1.Authors must decide the rule for terminating data collection before data collection begins and report this rule in the article.” If not, it is possible to use a stopping rule which guarantees a publishable p-value: just stop when the p-value is small!

“2. Authors must collect at least 20 observations per cell or else provide a compelling cost-of-data-collection justification.” Small samples are always suspicious. Were they the result of just one experiment? Or the fifth, discarding the first four as merely warm ups?

“3. Authors must list all variables collected in a study.” A lovely way to cheat is to cycle through dozens and dozens of variables, only reporting the one(s) that are “significant.” If you don’t report all the variables you tried, you make it appear that you were looking for the significant effect all along.

“4. Authors must report all experimental conditions, including failed manipulations.” Self explanatory.

“5. If observations are eliminated, authors must also report what the statistical results are if those observations are included.” Or: there are no such things as outliers. Tossing data that does not fit preconceptions always skews the results toward a false positive.

“6. If an analysis includes a covariate, authors must report the statistical results of the analysis without the covariate.” This is a natural mate for rule 3.

The authors also ask that peer reviewers hold researchers’ toes to the fire: “Reviewers should require authors to demonstrate that their results do not hinge on arbitrary analytic decisions.”

Our trio is also suspicious of Bonferroni-type corrections, seeing these as yet another way to cheat after the fact. And it is true that most statistics textbooks say design your experiment and analysis before collecting data. It’s just that almost nobody ever follows this rule.

Bayesian statistics also doesn’t do it for them, because they worry that it increases the researchers’ “degrees of freedom” in picking the prior, etc. This isn’t quite right, because most common frequentist procedures have a Bayesian interpretation with “flat” prior.

Anyway, the real problem isn’t Bayes versus frequentist. It is the mania for quantification that is responsible for most mistakes. It is because researchers quest for small p-values, and after finding them confuse them for holy grails, that they wander into epistemological thickets.

Now how’s that for a metaphor!

——————————————————————————————–

Thanks to reader Richard Hollands for suggesting this topic and alerting me to this paper.

1doi:10.1177/0956797611417632

Holocaust Survivor Compares Climate Skeptics To Hitler Deniers

Micha Tomkiewicz is a physicist who tells us he was a child in the Warsaw Ghetto and Bergen-Belsen during World War II when the Nazis (and Soviets) were gleefully murdering millions. He said that the Holocaust “killed most of my family and deprived me of my childhood.”

His is one more awful story from a century filled will awful stories of what happens when people assume Utopia can be had by all-powerful central government. His story and the story of his fellow survivors becomes far worse when we consider that there are some who deny the Holocaust occurred, that there are exist people who actively impugn evidence that is plain to the simplest idiot.

We hear these denials, but all of us know that these statements aren’t denials at all. It is clear that the people who deny that millions upon millions of Jews, Gypsies, Slavs, and others who were slaughtered by state power think it a fine thing that these souls died. Holocaust deniers, as everybody knows, don’t deny and instead have a secret (and sometimes an open) desire that the killing should begin anew.

Tomkiewicz knows what we know. He understands what the term denier means; he knows it is a code-word for evil.

But even knowing all this, and living the life he as lived, he cannot stop himself from using this word to describe people who do not fret as much as he does about climate change. He is so consumed with his passion that he was able to write this:

In 1933, very few people believed that Hitler would seriously try to accomplish what he preached and almost no one could imagine the consequences of his deadly reign. Although there was evidence available — Hitler was clear about what he wanted to do in Mein Kampf — why did people not pay attention? These “deniers” might as well have been called skeptics in their day.

This is well worth spending a moment to unpack. He begins with a truth: it is true that in 1933 “very few people believed that Hitler” would become the menace he was to become. Tomkiewicz follows this true by claiming the truth was false and that there was enough evidence for all, or at least a majority, to have predicted with certainty that Hitler would eventually come in third as Leader With The Highest Body Count (Mao still holds the title, followed closely by Stalin; thank you socialism!). With loving hindsight, Tomkiewicz condemns the world for being filled with “deniers.”

Which makes it strange that he next says,

But what I am suggesting is that even though it’s hard to see a genocide — any genocide — coming. The future is hard to predict, but we can see this one coming. This genocide is of our own making, and it will effect everyone, not just one group or country.

It is hard to tell that climate change will be a genocide, but it is also easy to tell. Just as he claims it was hard to see that the Holocaust was coming but also easy to see. Just as he paradoxically claims that skeptics, whom he calls “deniers”, cannot see as sharply as he can. He says that skeptics pine for “unattainable certainty” about the coming “climate change genocide.” But he also claims to possess this certainty, or enough of it so that he can demand the government “do something.”

To call a skeptic a “denier” is rank abuse, because as we have seen the word is a stand-in for vile intent. To compare “climate genocide” “deniers” with those who—what exactly? Supported Hitler? Enabled the man? Remember Tomkiewicz implied “deniers” in 1933 were responsible for Hitler—ah, the whole thing is asinine.

A far less serious crime to logic is his begging of the question. Skeptics claim, via arguments and evidence, to be less certain about climate change than Tomkiewicz. Tomkiewicz claims to be more than sure; he says he is certain. But he also implies that because he, Tomkiewicz, is sure then everybody should be, when the point at issue is how certain anybody should be. To attempt to bypass this debate by casting foolish aspersions and distasteful comparisons is a sign of weakness.

Update See this and this.

Dutch Doctors’ Strange New House Calls

Dr Overlijden of Levenseindekliniek
Dutch doctors

A new contest is in order, if it doesn’t already exist. Each year we should recognize and award the person or group who invents the best new Orwellian Phrase. This is a word or words which appear to say one thing but which are in reality their own antonym.

I bruited this idea with a colleague, who was enthusiastic and suggested for a prize the same which Winston received at the closing of 1984. I was sympathetic but thought this lacked charity. But he pointed out that this was the point: the prize is not one which one would wish to win.

Except perhaps by the folks who are responsible for my entry for 2012. These are the Dutch doctors behind Nederlandse Vereniging Voor Een VrijWillig Leveseinde, an organization which will dispatch a van to your very own house filled with doctors itching to kill you. Oh, yes. Leveseinde, you see, means “life’s end” and the docs of NVVE want to be the ones who take you on that journey.

The phrase I enter is NVVE’s “Life’s End Clinic” or Levenseindekliniek. It would not be much of an entry, I admit, except for the addition of “clinic,” an act which abuses that word in a savage manner. Incidentally, this “clinic” (which is not yet fully operational) is a place for those who would rather not have the van park in front of their home and thus frighten the neighbors.

I envision it as the kind of building into which Edward G. Robinson strolled near the end of Soylent Green. Which means we now we have to wonder about Holland’s food supply chain.

NNVE’s front page proudly announces that “there were over 200 applications from people with a euthanasia request at this clinic. There are reportedly twice as many women as men” (“waren er meer dan 200 aanmeldingen van mensen met een euthanasiewens bij deze kliniek. Er meldden zich twee maal zo veel vrouwen aan als mannen.”). For some reason this brings to mind the phrase “Minorities and women strongly encouraged to apply” which appears at the bottom of every academic job announcement.

The leftward news site MSNBC called the Vans of Death “mobile euthanasia units”, which isn’t Orwellian but is at least sufficiently bureaucratic. “Old Jones was heard groaning last night. Dispatch the M.E.U.!”

The reporter claims all sorts of safeguards are in place. The death supplicant must cry not once, not twice, but thrice or more, “Kill me!” Various folks are interviewed. Documents are signed. And then a doctor—a person once thought to be have sworn a duty to preserve life—slips in a needle which kills you. Well, to be fair, it only puts you to sleep. Once your eyes are closed, that’s when the knife comes out and the Dutch “doctor” slits your throat.

Kidding! Actually, knives are far too messy so “a second injection follows, which will stop [your] breathing and heart beat.” A very long-winded way to say that the doctor actively kills you, that he commits homicide, that he purposely deprives you of your life, that the doctor has stepped just this side of being a murderer. That he gives a whole new meaning to “house call.”

MSNBC reports “The teams”—squads of white-coated killers—”would be allowed one procedure a week because of the emotional toll that each visit takes.”

I’m not sure what to think of that. Does it imply that these killers—for this is what they are—have retained some small scrap of humanity? It must be small because it only takes seven short days for their consciences to quiet enough so that they can kill again. Or is the one-scalp-a-week bag limit imposed so that the “teams” don’t begin to love their jobs too well? And don’t you claim that this isn’t possible, for all history tells us that it is not difficult to find individuals who enjoy killing.

Something has to qualify these people. Perhaps a desire to kill those who claim that they want to be killed will be a prerequisite for aspirants who wish to enter the growing specialty of euthanasia. As this field develops, we can imagine classes in Emergency Euthanasia covering topics like what to do when the needle is not at hand: see pillows, suffocation; tall buildings and the lower lumbar shove; the Ty Cobb upper lumber cranial contact.

Update I originally ended this piece with the unfortunate “Remind me not to visit the Netherlands anytime soon.” By this, believe it or not, I meant something along the Soylent Green line. I stupidly thought it implicit and did not add this. I hope it is obvious that I do not disparage all peoples of the Netherlands. I beg all your pardons. I blew it. I am sorry.

I do, however, disparage the doctors who ride around in shiny white (I’m assuming the color) vans willfully killing people. The Soylent Green business comes in from asking: what happens to the bodies?

« Older posts Newer posts »

© 2014 William M. Briggs

Theme by Anders NorenUp ↑