Category Archives: Science

generating knowledge

[wordless] . . . . [/wordless]


Sexy Scientists


(And you wonder why you have to win an election to prove the science of alternative energy, climate change, nutrition, physical activity, health insurance for eternal life . . .)

P.S. Sexy? No contract offers from Playboy for a centerfold shot. Lollipop, you’re still All That.

Power, not Concussions, Corrupts Persuasion in the NFL

The NFL and retired players have announced an out-of-court settlement on the concussion lawsuit. The NFL will pay up to $765 million dollars to retired players for concussion-related health problems over the next 20 years. The players will no longer pursue individual or collective lawsuits on this issue. You can easily find comment on the implications of this agreement which makes it clear that no one is particularly happy with it, usually the sign of a good settlement. Past the winners-and-losers perspective, I’d like to offer a broad persuasion analysis of the case.

Loudly: No one knows all the details. It’s possible that several parties in this lawsuit are trying to avoid or hide a larger problem that is indirectly related to the issue. I’ve got to assume that what’s reported in the press is pretty much the whole truth all the while knowing it isn’t.

Almost As Loudly: I assume from my reading of the science that NFL playing is not a risk factor for future cognitive disability like Alzheimer’s. You can read prior PB posts that detail this with a particular emphasis upon the biased researchers doing Tooth Fairy Tale methods (you could start with this first post).

These are two huge assumptions, the Whole Truth and Fairy Tale Science. If you don’t buy both, the rest of this post will seem silly.

Nouveau Curves Divider

Given the assumptions, the NFL lost a billion dollars, a bad persuasion outcome no matter how you count the change. Sure, it works out to about $2 million per team amortized over 20 years, but that’s $2 million a year they could have kept. Even for a billionaire, surrendering millions when you didn’t have to, is a change you can count, very painfully.

Why did the NFL let the players and their lawyers take the unfavorable Falling Apples with the real science of NFL concussions and then turn it into the Fallen Apples of persuasion and this billion dollar win? Assuming the science says no harm, how can you possibly lose a billion dollars on the science?

When Malcolm Gladwell in 2009 wrote his New Yorker piece about NFL concussions, the Fallen Apples began. And, the NFL did nothing effective to refute or ridicule the scientific posings of a journalist FauxItAll. And, when media outlets like ESPN or the New York Times repeated that meme, again the NFL diddled and dithered. Worse still the NFL began a faux safety campaign that appeared to accept the harm the Fallen Apples from Gladwell suggested.

At minimum the NFL could have hired a bunch of paid guns like the players and their attorneys did and refuted point by point the bad science placed in the pop press. The NFL could have pursued a public campaign to highlight the inherent bias of retired players trying to get more money from a wealthy employer by hiring scientists to do science published in the New York Times. The NFL could have conducted medical investigations into the drug use and misbehavior or just plain family history of retired players with cognitive problems.

Instead the NFL stood still from 2009 until 2013, taking shots in public and doing nothing effective in reply. Why?

First, the NFL is not a monolith on this issue, but rather is composed of 32 ownership groups. That’s a lot of cooks around a persuasion pot. I can’t imagine 32 groups like this could agree on one specific, organized, and cold-hearted persuasion campaign. Each probably wanted what they thought best for themselves and as a result, nothing good happened.

This is not a trivial practical persuasion problem. Anytime you are in a large organization or coalition, planning and executing effective persuasion becomes exponentially more difficult. Assuming I’m the persuasion genius I play on this blog, then my accounts of getting zealots out of the room or out of the way explain the problem. Everyone who’s had any success in life thinks they know how to persuade Other Guys. In big organizations, the first persuasion play is on all those idiots and you’ve got to move them before you can move the real Other Guys you need to target.

Second, I think the NFL owners fell victim to the Rule: Power Corrupts Persuasion. The NFL is the living embodiment of power down to the very game itself. The NFL operates as a monopoly with a favorable nod from the US Congress. It does what it does with little regard for persuading freely choosing Other Guys.

As a result, the NFL had no idea how to react to a FauxItAll like Gladwell or the Fairy Tales from team physicians posing as scientists. More particularly, the NFL had no idea what the players’ lawyers were up to. And, that goes to the next point.

Third, a Peitho Award to the attorneys for the players. They are going to get about $200 million in fees on an issue that would be decided on science and that has no good science behind it. Can you believe that? Those attorneys completely fooled the public, the press, and the NFL owners on the science, then conducted a carefully and artfully manipulated public information persuasion campaign based on the Tooth Fairy Tales. Persuasion does not get much better than this.

Nouveau Curves Divider

Now, let’s consider the Persuasion Rules. Keep a sharp contrast between the players’ attorneys and the NFL owners.

Great Persuaders Don’t Need Rich Uncles, Kindness from Strangers, or Third Party Vote Splitters.

The attorneys invented or discovered all their persuasion plays. They didn’t need a super storm or a school shooting or an election. Notice that there were no friend of the court briefs on this suit. These attorneys went alone. They didn’t buddy up with anyone affiliated with high school or college football or any other American semi-professional leagues. They didn’t include other professions or sports that suffer concussions as part of the work.

All Persuasion Is Local.

The attorneys looked at all the elements in the Local and not just in football. They correctly saw how the current Age of Science allowed them to manipulate research to seem like science and establish a neutral ground of facts and truth. They certainly influenced either funding, participants, methods, and press releases of the research projects with NFL players. They made this case turn strongly on the court of public opinion where owner power would be powerless.

Drive with Science, Putt with Poetry.

I call them Tooth Fairy Tales, but properly presented and unrefuted, those Tales turn into science. That bad science Boxed the Local, permitting certain elements while killing others. From that science, the attorneys could tell beautiful and heart breaking stories about NFL giants laid low in age by the violence of the sport. The science created a narrative structure that made idiosyncratic and unique problems seem like an inevitable outcome.

[Meme Side Bar: Never heard the smoking meme from the attorneys did you? Why not? Those violent collisions sure look awful. They’ve got to be more dangerous than Sitting or Cell Phone Driving or Sugary Soda Pop, all those things that zealots claims are Just Like Smoking. Think about it. Back to the Discipline of the Rules!]

Persuaders Can Either Be Famous or Effective, But Not Both.

Please name an NFL owner. You may only know one name, probably Jerry Jones, the ubiquitous and loquacious owner of the Dallas Cowboys. Or Robert Kraft of the Patriots for his great and public charitable work. Or maybe just the name of the owner for your nearest NFL franchise.

Name one attorney for the players. You never saw an attorney and no attorney ever became for a moment the face of the lawsuit for the players.

Nouveau Curves Divider

Let’s get out of here.

Don’t forget my assumptions. I assume we know most of the truth and that there’s no hidden crime behind the settlement. I assume the science says concussions do not cause Alzheimer’s. If these assumptions are wrong, then this analysis is largely wrong.

In shortest summary this case illustrates one Rule: Power Corrupts Persuasion. The NFL brought a gun to a word fight and lost. The owners realized too late that this was indeed a word fight, a persuasion fight, and by the time that light lit, the fight was not winnable on persuasion terms. The owners had allowed the players’ attorneys to define the science which then framed the pathos stories from each suffering player and his family in a very different narrative.

But, see the limits to persuasion.

I think the attorneys very wisely settled out of court and that this was the best outcome they could have gotten. A trial could have focused on the science which could have been fatal for the players’ attorneys. If the science is as bad as I think it is, a careful deliberation over all the evidence could have killed the players’ case. The attorneys knew the difference between Smoke and Mirrors, and Persuasion and settled before everyone smelled something funny or saw a glinting reflection.

I also think that this case will make no serious difference to America’s reigning passion for football. Little kids have been hitting each other all this summer and are doing it for scores that count as another football season begins. And, they’ve been doing it with the same passion and intensity they had in 2009 when this case began. It will probably continue for the foreseeable future. Thus, no one really believes that football is a serious risk for cognitive morbidities like Alzheimer’s . . . except when they read it in the New York Times.

This case is all and only about persuasion. Some guys are better at that than others and you can spot the difference.

Which way is the money going?

Building a Better Brain Trap

Brain Trap

Getting smart today is easy, fun, and popular. Everyone’s doing it! Through the combination of networked iGizmos, colorful apps, a few minutes of daily WATtapping, and a zeitgeist of the Age of Information, anyone and everyone can get smarter, a lot smarter. The Apple iStore proves it.

Lumosity Sells

No wonder.

“We attribute Lumosity’s growth to its broad appeal and the universality of the program’s benefits,” said Krishna Kakarala, CFO at Lumosity in an email. “Cognitive training offers something for everyone. Similar to a gym, people of all ages use Lumosity to keep their brain in shape, whether it’s to focus at work, do better at school and standardized tests, or just to stay sharp.”

Read that quote and you realize why I am not a Chief Financial Officer and only a blogger in BVDs. I had no idea that tapping an iGizmo made you smarter. This insight apparently eluded philosophers and scientists over the long sweep of education since Plato’s Academe and is only recently discovered with cognitive neuroscience, fMRI machines, and a steadfast application of the Rules of Persuasion. How did Aristotle miss the universal benefit of exercising the brain? Or Dewey? Or Skinner? Or anyone with a computer since Eniac? Why didn’t my education professors tell me about this in the 1970s for my bachelor’s or Master’s? Or the 1980s for my doc?

So, of course, WATtapping For Smarts is impossible which means the only thing you need to know about it is contained in the phrase:

. . . a steadfast application of the Rules of Persuasion.


Lumosity only teaches into the test in the very worst scientific sense, but in the best persuasion sense. With Lumosity, there’s no difference between the ExerBrain©™® activities you WATtap and the test you take to prove the effect of the ExerBrain©™® activities. The exercise is the test and the test is the exercise. Yes. It’s just another General Semantics Persuasion Play. None of the words mean the things you think they do, but they sure sound good. How good?

In 2012, the company reported $24 million in revenue, double from the previous year. “Lumosity has been growing at a quarterly rate of 20 to 25 percent and doubling every 6 to 9 months since launch,” Kakarala said.

Count the Change, baby. Count the Change.

Lumosity is a Stanford based operation that combines sheer greed, the Persuasion Rules, and various Stanford academic and medical units (and Harvard, too!). You can read honest to goodness peer review papers that prove the effect of Lumosity training on something called Executive Function. Except when you read the Methods and Results you know things are crazy. Take this one.

Women with breast cancer were randomly assigned to a wait-list control or Lumosity training for 12 weeks. All women were pre and post tested on Executive Function. After only 12 weeks the Lumosity women score higher on Executive Function.

Now, you’re probably expecting a Tooth Fairy Tale with huge sample sizes and trivial Windowpanes, but you’re wrong. This was a true experiment with random assignment of the 40 women to just one or the other group. And the Windowpane changes on Executive Function were Medium to Large! Here’s the Table with helpful yellow highlighting.

Lumosity Cures Cancer Table

If you’ve done any training in education, you know that such effect sizes are impossible. Nothing makes people “smarter” in such a short period of time doing nothing more than WATtapping colorful apps.

So. Why did this work? Well, just to be persuasion cold hearted about it, we’ve seen that people with cancer will try about anything and make it work for awhile. Remember this one with physical activity?

And crueler than that observation is the fact that the “tests” the researchers use to assess the effect of the Lumosity training are essentially the same things as the Lumosity training. Gee. People practice taking a test and they get better taking a test.


Notice that result for Digit Span, perhaps the purest measure of short term memory we’ve got.

Lumosity Cures Cancer Table Digit Span

Nothing called Executive Function exists without short term memory. Digit Span requires you to read or hear an 5 or 6 or 7 digit number, wait, then repeat it. Note that the Windowpane here is less than Small and not significant. Huh. Lumosity makes you smarter and vastly improves everything about your Executive Functioning except short term memory!

I’ll wait for the pill that delivers that result. It’s faster and easier than all that 12 weeks of WATtaping . . . except for the pill’s side effect of an erection lasting longer than 4 hours and frankly that bothers Melanie more than it does me.

P.S. Look out Lumosity. You’re gonna lose the Alzheimer’s market if reason prevails over the Rules of Persuasion. Good grief, even the NYTimes is reporting on this.

Shelli Kesler, S.M. Hadi Hosseini, Charles Heckler, Michelle Janelsins, Oxana Palesh, Karen Mustian, Gary Morrow, Cognitive Training for Improving Executive Function in Chemotherapy-Treated Breast Cancer Survivors, Clinical Breast Cancer, Volume 13, Issue 4, August 2013, Pages 299-306.

DOI: 10.1016/j.clbc.2013.02.004.

I am Not Biased, I am Right . . . Left . . . Correct . . . uhh

I have frequently noted the persuasive science from researchers who think their work is not biased when it proves that conservatives are Neanderthals. Sure, define conservative as Neanderthal then measure outcomes only Neanderthals can produce and, yep, sure enough, conservatives are Neanderthals. That’s science, baby. No persuasion here among us political scientists. And then you read a report like this.

Participants answered four questions assessing their stated willingness to discriminate against conservatives. The first two questions asked whether, when a participant was reviewing a grant application or paper, a feeling that it took a “politically conservative perspective” would negatively influence the decision to award the grant or accept the paper for publication. The third asked whether the participant would be reluctant to invite “a colleague who is generally known to be politically quite conservative” to participate in a symposium (on an unspecified topic). The fourth question asked whether, in choosing between two equally qualified job candidates for one job opening, the participant would be inclined to vote for the more liberal candidate (over the conservative).

The participants in this convenience sample are professors in social and personality psychology, the home of some of the best persuasion research you can find. And, the questions are that bald. Would you hire a conservative? Here’s the table of results to the four questions.

Not Biased Right

If you are playing along at home, you can test the prejudice here by changing the labels of conservative and liberal to something like male and female or white and minority as a control. If anyone, much less a professor, admitted a hiring preference for men over women or whites over minorities there would be obvious legal consequences. Yet, almost 4 in 10 professors of psychology will admit on an anonymous self report that they would not hire someone because of that applicant’s political beliefs. Hey, and nearly 1 in 4 would kill a grant application for the same reason; man, that’s gotta make you think since these grant applications are with the Federal government which allegedly does not discriminate!

These results are funny for at least two reasons. First, you cannot miss the obvious joke of people who are supposed to understand the science of prejudice, yet cannot see it in themselves. Second, you must see the joke that people who are prejudiced like this think they are still doing science based on such beliefs. It is overkill, unfair, and tasteless to make the comparison, but this is pretty much how dictators, tyrants, and oppressors of all stripes have operated.

We’re not biased; we’re right!

Yoel Inbar and Joris Lammers. (2012). Political Diversity in Social and Personality Psychology Perspectives on Psychological Science, 7: 496-503.


P.S. And, you’ve got to read Skitka’s commentary on the methodological weaknesses of the Inbar and Lammers survey. Somehow the narrowest definition of research only applies to results you find disagreeable. Her approach would lead one to criticize lynch mobs for the poor quality of rope they use.

Linda J. Skitka. (2012). Multifaceted Problems: Liberal Bias and the Need for Scientific Rigor in Self-Critical Research Perspectives on Psychological Science, 7: 508-511.


Chocolate WATTage or Torturing Science over Your Cocoa

NPR chocolate wattage image

The horror of science continues with the terrifying news that chocolate is once again good for you and this time it helps the brain and may – gasp – stave off Alzheimers. We’ll detail this new research in a minute, but first consider the news reports.

In a small study out this week, people with high blood pressure, diabetes and some memory challenges performed better on cognitive tests after drinking hot cocoa for a month. Drinking the cocoa also appeared to increase blood flow to their brains, according to ultrasounds. Vascular dementia, which can include confusion, trouble speaking and vision loss in addition to memory loss, is caused by reduced blood flow to the brain that deprives brain cells of oxygen and nutrients.

Small? It was an experimental design where participants got randomized to different cocoa treatments for 30 days. Brain and cognitive functions were tested at baseline, Day 1 and Day 30. There were 60 people. Small’s got nothing to do with much here. You need to replicate this research and kick all the cans you can infer from it. But small is no problem. This research is in the stream of an active river of research on brain function. But worse than small, this study is – gasp – preliminary.

The study was small and preliminary, so it should not encourage people to drink cocoa thinking it will save them from dementia, warned Heather Snyder, director of medical and scientific operations at the Alzheimer’s Association. “I don’t think we can draw any conclusions from this study about whether drinking cocoa is a potential therapy,” she said.

We’ve seen people jump off the cliff for Fairy Tales and Windowpanes that are barely cracked open and here we’ve got a randomized controlled experiment that shows both physiological and cognitive improvements from cocoa consumption. We can draw conclusions from Fairy Tales, but not experiments? Is chocolate that big a threat?

Remember the hilarious research which showed that consuming sugary drinks increased WATTage thus enabling Other Guys to hit the Central Route? Jeepers, you gotta do something bad like drink sugary soda pop to understand that drinking sugary soda pop is bad for you! Now, it appears if you consume evil cocoa for 30 days your brain works better and your mind solves puzzles faster which means consuming chocolate helps you understand why consuming chocolate is dangerous for you!

What are people supposed to do?

Let’s start with the actual research. The results are much more complicated than any of the news reports understand. Researchers recruited 60 older Other Guys for the study which required them to show up for testing and to drink cocoa twice a day for 30 days. The researchers provided all the cocoa in easy to use packets and therein lies the experiment with randomization and control.

See, while everyone got normal cocoa, half the Other Guys got a flavanol-enhanced dose. A considerable amount of prior science demonstrates that some of the chocolate effect comes from the flavanol, so the researchers increased that amount to see what kind of boost you’d get.

(Experimental Side Bar: Why not cocoa without flavanol instead? That could be reasonably interpreted as harming people. Flavanol helps. You could have done this study with either reducing or enhancing the flavanol amount and you’d get a reasonable control. However, cutting the flavanol would be some kind of unethical. Does that make sense?)

So. You’ve got an experiment with older Other Guys all drinking cocoa twice a day with half randomly getting that flavanol boost. And, we’ve tested everyone before, Day 1 and Day 30 on a variety of physiological and cognitive measures. Now it gets complicated.

The experimental variable with flavanol did nothing. The experiment is blown and did not work. The extra flavanol had no effect on brain function or cognitive testing. Now, it gets simple.

Just run the analysis on all Other Guys making this a longitudinal observational Fairy Tale with 60 participants. Now concerns about sample size and preliminary results come flying in. This ain’t the plan and we’re testing post hoc hypotheses. Now, it gets complicated again.

The researchers “adjust” the analysis for the participants neurovascular coupling status. This is a standard Tooth Fairy play where you start diddling your analysis with different combination of variables. This was designed as a simple experimental t-test, but that blew up. Now, let’s add other variables. What is neurovascular coupling?

Neuronal activity is linked to oxygen and glucose delivery. An increase in metabolic demand leads to an increase in blood flow. The close functional and spatial relationship between neuronal activity and cerebral blood flow has been termed neurovascular coupling (NVC) or functional hyperemia. Impaired NVC has been associated with significant pathology.

So, let’s now divide the Other Guys – not on the experimental variable of flavanol amount – but on their NVC score and break everyone into two groups, Intact and Impaired NVC. Now, look at the effect of 30 days of drinking cocoa (with or without extra flavanol). Here’s a pretty picture.

Cocoa NVC Cognitive Results

Cocoa had no impact on cognitive outcomes for people with Intact NVC function (the left side). Cocoa did have a beneficial effect on cognitive function for people with Impaired NVC status (the bars on the right side). In this case, the lower scores are better because it means you are getting faster on this measure of cognition called Trails B which requires you to connect a bunch of dots scattered on a page. The faster you go the better the test.

So, the pretty picture looks good, but there are a big buncha buts with it. See, only 17 people had Impaired NVC function. And, yeah, whoopee, the effect is significant at p < .007 which is a kinda cool number, but come on. We start with 60 people, randomize into groups of 30, but that blows up. We then adjust the data and find 17 people who are 007 at connecting dots.

And, the Other Guys took 3 measures of cognitive function: Mini-Mental State Examination and Trail Making Test A and B. The pretty picture shows the results from Trail B, but the report is silent on the other 2 measures. Think about that. Start with 60 people, randomize and control, but, oops. Now find 17 with a good story on NVC, run 3 tests of cognitive function on them and report the one that delivers and omit the others.

Yeah. Cocoa drinking reduces Alzheimers. Probably. Maybe if you have impaired NVS function and are connecting a lot of dots on a daily basis in the Local called Trails B. Then. Yeah. There’s a significant relationship between cocoa drinking and cognitive function.

The most interesting element of this story is not the science, but the persuasion. I have no idea why this thing hit journalists as something to tout as the New New Thing, but they sure did. And, if you do a search on Goggle News you can find all those cheerfully stupid press reports about chocolate and Alzheimers.

And even more persuasion interesting was the response of folks who represent science. Our more careful look at Methods and Results shows that this is an interesting little study with some tantalizing outcomes that are obviously weak and shaky, but consistent with the literature and suggest more focus on NVS status and cocoa consumption. The Need For Future Research here shouts for experiments with impaired NVS participants randomly assigned to cocoa conditions with measures of physiological and cognitive function. If I was on that study section or running a grant funding operation, I’d be favorably disposed toward that given this study.

But, the media statements of these science guys didn’t take that tack. Instead they all worried about people eating chocolate for diet and weight gain reasons. Here’s the lead author.

“I do not recommend that people add chocolate or cocoa to their diet at this point,” Sorond told The Salt via email. “Our results are preliminary and adding the extra calories, sugar and fat that comes with chocolate and cocoa carries additional health hazards which may offset any possible brain benefits.”

You could put it like that. Or even like this from the actual Discussion of the paper.

Moreover, we show that NVC may be modifiable. Four weeks of cocoa consumption resulted in improved NVC and Trails B scores in those with impaired function at baseline . . . Our study shows that cocoa consumption resulted in higher NVC and that individuals with higher NVC had better cognitive function and greater cerebral white matter structural integrity.

Hmmm. So the results are preliminary and hazardous while showing that NVC can be modified with cocoa which then improves NVC function and cognitive function.

No persuasion here. Just science.

Farzaneh A. Sorond, Shelley Hurwitz, David H. Salat, et al. Neurovascular coupling, cerebral white matter integrity, and response to cocoa in older people. Neurology published online August 7, 2013.

DOI 10.1212/WNL.0b013e3182a351aa

P.S. When you do science with a PR kit you will always find persuasion.

P.P.S. And just for fun read this.

The healing power of chocolate. The researchers weren’t quite ready to call chocolate a health food — they cruelly reminded the audience of its fatty content — but they did have good news about the flavanols found in cocoa (particularly some dark chocolates).

Norman Hollenberg of Harvard Medical School has documented that central American Indians who consume large quantities of cocoa have low rates of hypertension and of vascular dementia (caused by restriction of blood flow in the brain). At the AAAS meeting, he reported on a experiment showing people given flavanol-rich cocoa enjoyed a “a significant increase” in cerebral blood flow. “We hope,” he noted, “to explore the potential of flavanol-rich cocoa in preventing or ameliorating the vascular dementias.”

Another researcher, Ian Macdonald of the University of Nottingham, scanned the brains of women who’d been given flavanol-rich cocoa. He found it increased “cerebral blood flow to gray matter.” He and Dr. Hollenberg didn’t urge listeners to go out and gorge on chocolate, but they did raise the possibility of flavanols being used to help aging brains, perhaps being administered in the form of vitamins. Let’s hope these vitamins are the chewable variety.

So says a New York Times report from . . . 2007, six years prior to this paper.

Climate Changes the Windowpane and Science!

You might have caught the unusually long news cycle of a research report in Science about the causal relationship between climate change and human violence both small (interpersonal) and large (intergroup). That attention was due in no small part to the finding: Climate change causes more violence. Of course, the researchers had to invent a new kind of meta-analysis, a new kind of effect size, and a new standard of scientific evaluation to prove this, but hey, that’s the TACT with persuasive science!

Really. Climate change science requires a totally new science to prove that weather makes people kill each other. Start with the new kind of meta-analysis.

We then collect all available candidate studies and – guided by previous criticisms that not all correlations imply causation (9–11) – focus only on those quantitative studies that can reliably infer causal associations (9, 12) between climate variables and conflict outcomes. The studies we examine exploit either experimental or natural-experimental variation in climate, where the latter term refers to variation in climate over time that is plausibly independent of other variables that also affect conflict.

Do a quantitative review of the literature, but only pick the “good” studies, not all the studies. And in particular pick studies that experimentally manipulate climate or show a natural experiment. Now, you might ask, as I did, how can anyone experimentally manipulate climate or weather in the sense we mean here (the great outdoors)? You cannot, of course. You can only run the General Semantics Persuasion Play and eat the menu. And see the General Semantics Persuasion Play with the words, Natural Experiment. Sure, it appears to be a natural assignment to treatment or control, rainy days versus dry days for example, but it’s still not under experimenter control and thus not an experiment. So call these carefully selected “good” studies experiments and, voila, they are!

Now, the new effect size.

In some prior studies, authors have argued that a particular estimated effect is “unimportant” based on whether a climatic variable substantially changes goodness-of-fit measures (e.g., R2) for a particular statistical model, sometimes in comparison to other predictor variables (14, 22–24). We do not use this criteria here for two reasons. First, goodness-of-fit measures are sensitive to the quantity of noise in a conflict variable: more noise reduces goodness-of-fit – thus, under this metric, irrelevant measurement errors that introduce noise into conflict data will reduce the apparent “importance” of climate as a cause of conflict, even if the effect of climate on conflict is quantitatively large. Second, comparing the goodness-of-fit across multiple predictor variables often makes little sense in many contexts since (i) longitudinal models typically compare variables that predict both where a conflict will occur and when a conflict occur and (ii) these models typically compare the causal effect of climatic variables with the non-causal effects of confounding variables, such as endogenous covariates.

So, explained variance is not scientific or at least not scientific enough for climate change science because it includes “noise” (i.e. the error term)! And because the climate data is too noisy (i.e. too much error because you don’t really understand what’s going on) we’ll simply ignore noise or random variation or unexplained variance or the error term or whatever you want to call that. And, since different researchers use different variables with different measurement theories which makes meta-analysis more difficult, rather than work harder, we’ll work easier. And this.

However, in principle the “signal” is a relationship that exists in the real world and cannot be affected by the researcher, whereas the level of “noise” in a given study’s finding (i.e., its uncertainty) is a feature specific to that study – a feature that can be affected by a researcher’s decisions, such as the size of the sample they choose to analyze.

You cannot control the “signal” yet they analyze nothing but experiments or natural experiments which, by definition, control the “signal?” When you are building the dataset for analysis you restrict it to only those articles that, as experiments, must control the “signal,” but when you analyze the data from the experiments, you won’t count that change because you cannot control it? And, double-wow, you think you control the “noise” or the error term with a larger sample size?

Quick summary of the new climate science. To execute the New New Meta, first pick only studies that do your new science – especially ones that eat the menu, substituting words or numbers for the actual phenomena. And, since the Old Old Meta uses outdated quantitative methods that produce disconfirming statistics, invent New New Effect Sizes that avoid things like signal and noise ratios or variance explained.

And, then the New New Evaluation.

Evaluating whether an observed causal relationship is “important” is a subjective judgement that is not essential to our scientific understanding of whether there is a causal relationship . . . Our preferred measure of importance is to ask a straightforward question: Do changes in climate cause changes in conflict risk that an expert, policy-maker or citizen would consider large?

We assess science with the opinion of experts, policy-makers, or citizens? So, create an artificially complicated research project that I doubt not one person in the world not on this research team could independently replicate, then let the People decide on its scientific importance?

All Bad Science Is Persuasive.

P.S. Here’s the effect size.

The precision-weighted average effect on interpersonal conflict is a 2.3% increase for each 1σ change in climatic variables (s.e.= 0.12%, p < 0.001, Fig. 4 and table S1) and the analogous estimate for intergroup conflict is 11.1% (s.e.= 1.3%, p < 0.001, Fig. 5 and table S1).

I’m not close to confident about exactly how these guys did the numbers, so, I implore you, take off your shoes and use your own fingers and toes to do the math. That noted, I find expressing percent change as a function of standard deviation change to be more than a bit weird. Assuming this is all linear, then you always get a 2.3% or 11.1% increase in violence with each standard deviation increase in climate change.

I think (maybe, could be, sounds reasonable until an Interested Reader corrects me) I think this can be Windowpaned traditionally as around Smallish effects. A 10% change would be Small, so that 2.3 thing for interpersonal violence is one quarter of a Small effect. And, then the intergroup estimate of 11% is clearly Small.

So, we’ve got purely observational data (i.e. Tooth Fairy Tale) despite the General Semantics Persuasion Play from the authors about experimental studies. No one controlled or randomized anything anytime anywhere. And, these are all convenience samples; the data comes to you as it pleases. Yeah, sure, it is tree rings and ice cores and geological strata, but you are not in control of gathering that data. You get what you can get. And, you’ve got some ginned up math that boldly denies past method as Your Father’s Oldsmobile.

Then you get effects that are barely detectable above random variation. Man, that is some persuasive science.

P.P.S. Pretty pictures help!

Pretty Picture of Climate Violence

Jeepers, those models go to hell at the extremes, don’t they? All that fuzzy error on the hard left or right with tighter bands through the middle. Maybe there’s some nonlinearity in Mother Nature, but that requires more math effort. Besides, dogma says climate change is all and only linear, unless it’s a hockey stick. Or not. Whatever. Just spell everyone’s name correctly!

Solomon M. Hsiang, Marshall Burke, and Edward Miguel. Quantifying the Influence of Climate on Human Conflict. Science, Published Online August 1 2013.

DOI: 10.1126/science.1235367

Alzheimer’s as Bolivian Bank

No. I don’t wish Alzheimer’s on anyone and the world will be a better place when we find The Cure and can find new uses for ribbons, but until then you need to distinguish between the science and the persuasion of this phenomenon. Here’s today’s lesson.

. . . in a significant shift highlighted at the conference, leading Alzheimer’s researchers are identifying a new category: “subjective cognitive decline,” people’s perception of fledgling changes still invisible to others.

This from a NYTimes report on a panel at the Alzheimer’s Association International Conference 2013. Five teams on one panel reported results on the relationship between Subjective Cognitive Decline and later Alzheimer’s . . . or Mild Cognitive Impairment. The headline from the panel is that people can detect their own cognitive changes that will eventually lead to Alzheimer’s . . . or MCI . . . while these cognitive changes remain invisible to others.

So. I know something about myself that you cannot detect and that knowledge will turn into disease that both of us will then see for ourselves. Subjective Cognitive Decline. And, you can read the science of SCD here in the abstracts from the panel presentations at the conference.

At the risk of appearing a cold hearted persuasion analyst, I’ll note the very bad math in these abstracts with only one providing anything remotely approaching a Windowpane which is a Small+ effect size in a necessarily observational design (any volunteers to get assigned to the Alzheimer’s group?). The rest of the abstracts seem to do their science with persuasion and a relentless focus on words over numbers. Like this.

We, as well as another research group, have reported a significant relationship between SCC and cortical PiB binding in CN individuals.

In both ApoE4 carriers and non-carriers, increasing number of subjective memory symptoms at baseline was related to worse baseline verbal memory (P for trend<0.001 in both groups) and to higher rates of verbal memory decline over time (P for trend = 0.006 and <0.001 in ApoE4 carriers and non-carriers, respectively), after adjustment for age and depression.

One team does talk tough . . .

The group concludes that the currently available data is too limited and too heterogeneous to define SCD in preclinical AD as a clear-cut entity

. . . but then . . .

. . . highlights the need for intensified research on this topic.

So. There’s Nothing Going On, but We Need To Keep Funding Grants.

Stated in persuasion terms, when you Count the Change, there is not much scientific evidence of Change unless you have an incentive like grant funding. Then this Subjective Cognitive Decline appears real, important, and predictive. And, remember, all of this science is only in conference presentations with all that preference for words over numbers and Small effects in observational designs and none of it has even survived peer review.

Now consider this only from a persuasion angle. Look at the Local. The Other Guys are getting old, over 50 or 60. Just as everything in the body begins to break down with age, so does the brain and the mind and memory. Alzheimer’s is a frightening disease where the brain ages wildly more rapidly than the rest of the body, leaving a relatively healthy shell with nobody in charge. Thus, older Other Guys are experiencing normal aging which is a cognitive decline that everyone experiences except . . .

Experts step in and say that some aging Other Guys are more sensitive to their cognitive decline and can accurately detect abnormal cognitive decline thus providing an early warning about impending Alzheimer’s . . . or Mild Cognitive Impairment. Some people can tell you they are getting Alzheimer’s before you can detect it. Of course, we can’t really Count these guys, but if we get more grants, we’ll come up with . . . something.

Now, shift to another set of Other Guys, the kids and younger friends and colleagues of these aging Other Guys. These outside Other Guys can see the cognitive decline, but They have no internal perspective and don’t know what that decline feels like. When these internal Other Guys make complaints about cognitive decline to the external Other Guys who consider all the persuasive expert opinion, what will They think?

You see the Bolivian Bank as both the aging Other Guys and the external Other Guys become easy, ripe, and luscious for persuasion. Some experts are claiming that self reports of Subject Cognitive Decline herald not normal aging, but a disease like Alzheimer’s . . . or Mild Cognitive Impairment. Recall that phrase from the NYTimes report:

while these cognitive changes remain invisible to others.

So, if you are a kid with an aging parent and Mom starts complaining about her cognitive decline then Mom is going to develop Alzheimer’s. Hey, if you are Mom and you start noticing cognitive decline then you are going to develop Alzheimer’s. That’s what this very weak collection of unpublished abstracts claims.

How can such subjectivity not create a Bolivian Bank?

Now, some more science about Alzheimer’s. Recent evidence suggests that rates of Alzheimer’s are now decreasing. While this sounds strange, if you think about it, the decline is reasonable.

We’ve just started a new era in human aging where a lot more people are living longer compared to historical expectations. That first wave of older survivors is not as healthy compared to later waves. Think only about smoking rates in the Greatest Generation who won World War II compared to their Boomer children now hitting 60. People are living longer and healthier, too, which means that some disease rates may decline as populations get smarter about health as they age.

So while fewer people are getting the disease and this decline is likely to continue for at least a few more years, Other Guys may think They are more susceptible to Alzheimer’s because of this Subjective Cognitive Decline persuasion. And, believe me, if you live long enough you will notice an objective cognitive decline that will shock the hell out of you.

Regardless of the bad science, see the great persuasion possibilities here. Remember the Mozart Effect which proved that listening to K.448 made Other Guys smarter? The science on that one was no better than the Subjective Cognitive Decline effort and it sold a lot of good music to no practical effect. You can do the same thing with Alzheimer’s. Shootfire, Lumosity lights the way!

If you are unfortunately Sincere about helping Other Guys, you probably want to forget this post because it strongly suggests that most Alzheimer interventions are pure persuasion even among some folks wearing white lab coats. Such feelings might interfere with your ability or motivation to pitch various in-home tests of Subjective Cognitive Decline or various inventions that slow Subjective Cognitive Decline.

Jeepers, science is telling everyone It’s All In Your Mind which is the whole point of persuasion, isn’t it?

The Optogenetics Persuasion Play

Stop the WordPresses!

Optogenetics BBC Story

And, if you take the time to read the Science report (and especially the Supplemental Materials), you know that the Headline ain’t no Fallen Apple. Researchers experimentally manipulated neurons in mice that were the actually memories of prior learning trials. The method impresses as much as the result. How do you physically managed to do something to a few neurons in a mouse brain that affects the memory of past events?

That is a marvel of engineered science with a physical probe that is placed into a mouse brain (as depicted in the news story picture). That probe then produces a controlled burst of light onto a specific neuron which then changes the memory contents of that neuron. This process is called Optogenetics and you can get a good lay level intro here at the public wiki and a deeper one here at a science wiki.

Turns out that light affects neuron activity and status and if you hit the right neurons you can change memory. The specific illustration in this research involves memory for fear-conditioned locations. Transgenetic mice were run through standard fear-conditioning trials in four different mazes with painful, but not injurious, foot shocks. Normally after such learning, mice will avoid the locations where they received the shocks and display obvious behavioral signs of agitation and fear when placed in those locations. However, the researchers used those light bursts from the brain probes to change the neurons in the brains of the mice so conditioned. When they returned the optogenetically modified mice back in the mazes, the mice showed no fear of the shock locations.

Memory stores both information about the location and the shock, but, apparently, in different neurons. Using optogenetics, the researchers were able to alter the neurons for the location memory or the shock memory and manipulate them independently. Thus, a mouse could be made to forget a true shock location or misremember a safe location as shocking all through that light probe aimed at specific neurons.

The researchers make a fuss about the False Memory implications of this. Elizabeth Loftus demonstrated a long time ago that people could be psychologically manipulated to believe they remembered things that were not true and never happened. This research employs a physical rather than psychological manipulation to accomplish the same outcome. However, I don’t think that False Memory is the biggest deal here.

Good grief, we can target specific neurons in the living mammalian brain that affect learning and memory. That is astounding for me. Sure, we can manipulate brain function with drugs, for example, but to find, hit, and change neurons that are encoding learning and memory is a giant step. That we can do this is not surprising. Everyone working with learning and memory knows that the brain must be the source, but to get to this level of detail to affect such a large behavior (i.e. fear conditioning versus chemical changes in neuron activity) is a big whoop. Stated another way the theory and research here is fabulous.

Practical persuasion, you ask? Well, you can imagine your own Optogenetic Skull Cap that somehow penetrates the human brain with a light probe to manipulate attitude and belief neurons just as your Other Guy walks by your product in the store aisle. Turn on a beam of light and she hits the TACT. Yet, while we can imagine this, making the reality is probably like breaking the speed of light barrier: Unlikely.

Sure, you can imagine a cap ‘n probe and one that might even work, but it would never be legal even if it was safe and effective. Such a cap would alter human behavior beyond all recognition. Assume you could wrap that cap around someone’s head and that the cap could not be removed. Whomever controlled the light probe would become Geppetto and the Other Guy in the cap would be Pinocchio. Geppetto could install whatever memories Geppetto deemed useful and Pinocchio would dance. And, now, we are skating on the ice of Hollywood pitch meetings.

We could get close to this in the name of prosocial persuasion. Consider one of the terrible problems with Post Traumatic Stress Disorder: Intense, uncontrollable memories of the trauma. Gee, maybe we could identify the specific neurons and illuminate them with an optogenetic treatment. How about the awful hallucinations of various mental disorders or from certain drug treatments? Find those neurons and light them up. Now, after that success move on to other applications. How about training and learning? Maybe we can create learning cartridges in the form of neuron stacks that be inserted with optogenetics.

Even if we do move the science in that direction, I see no chance of the Optogenetic Skull Cap getting released into the natural Local of most Other Guys. Hey, Skinner demonstrated what you can do with a Box and the When-Do-Get Play and we don’t allow scientists to do this to us or our children or even our pets. We’ve got pretty good science on how to change the Other Guys and we’re not allowed to use it to best effect precisely because it takes volition away.

The practical persuasion lesson here is the science. What you do with persuasion does operate on the brain at the cellular level. While most of us had figured that out a long time ago, this kind of detailed support is useful. And, optogenetics may prove a therapeutic modality of the first rank, but always with a very wary look over the shoulder at the most pressing concern: Volition.

P.S. Can’t wait for those Hollywood movies that employ optogenetics . . . wait. They’ve already done it. Twice! Total Recall with Arnold (boom) and Total Recall with Colin (bust).

Total Recall 1 and 2

A persuasion parable! Take the same persuasion Play, but put it in a different Box. and It Depends! If only the remakers had had the Google Bottom Line Regresser Regression!

Genetics as Bolivian Banks: Exercise Motivation

Talk about easy, ripe, and luscious.

At some point, Dr. Booth says, scientists conceivably could develop a test that would reliably inform someone whether he or she is genetically predisposed to being physically lazy, or the reverse.

As if you don’t know already that you don’t like to exercise even without the ExerGene Test©™®.

As if the results from the ExerGene Test©™® would make any difference in your workout.

You see the persuasion implications. Even if the Other Guys are genetically lazy or just plain old lazy, we are still talking about volition. The Other Guys always have choice even knowing Their motivation. For some it is easier and for others it is harder, but it is always volition. Hey, that’s the nature of sin, right?

But see the persuasion opportunity here. You can hide that volition from the Other Guys with ExerGene Test©™®, then sell Them your ExerGeneStimulator©™® energy drink or DVD or app or the any-day-now Google Glasses Vizulator©™®.

When the world makes demands on slothful Other Guys, you’ve got a Bolivian Bank!

Buzz Phrenology in Living Color

Do you make your living as a persuasion consultant running on the bleeding edge of innovation? Then you’re using fMRI pretty pictures. Got one for you to measure Buzz.

fMRI Buzz

Cool, huh?

Now, the details.

Buzz or Word Of Mouth is one desideratum of advertising. Buzz refers to the direction and intensity of communication your advertising drives. You want Other Guys to say good things about your product or service to still Other Guys in a two step flow, thus multiplying the reach of your message. Get some Guys to Buzz and hopefully a lot of Other Guys will also start Buzzing and you’ll be rich and famous.

A team of persuasion scientists tested Buzz and obtained these pretty pictures of your brain on buzz. It’s real. It’s statistically significant. It’s in living color. So, use it to sell your Buzz services. Because we can’t use it for Buzz science. Here’s their description of the study.

In this study, message communicators (who pretended to be interns at a television studio) viewed ideas for television pilots during an fMRI scanning session and considered whether they would pass the ideas on to message recipients (producers) for further consideration. After scanning, interns gave video interviews about each pilot-show idea. These interviews, but not the original pilot descriptions, were then shown to producers in a separate behavioral testing session. On the basis of only the videotaped interviews, producers indicated whether they would pass the idea on to other individuals.

There is no Buzz in that description. None. Some people think about Buzzable ideas while lying in the fMRI tomb. The researchers take pretty pictures, then correlate those pictures with other ratings. But, no Buzz. No Buzzing. Maybe Buzzability.

A pretty picture of brains makes no Buzz sense. Looking at the color of the bumps in brains certainly measures something, but it can’t measure Buzz because Buzz is an aggregated behavior like number of positive tweets about a product or service. Regardless of what the brain looks like – with or without color – we know absolutely nothing about the communication of Buzz, just the color of the bumps in the average head stuffed in a claustrophobic tunnel thinking thoughts. Yet, the researchers claim:

More specifically, in examining neural activity associated with the ideas that the interns most successfully spread to producers (such that the producers also had the intention to spread the ideas further; i.e., facilitating the buzz effect), we observed activation of the regions most commonly associated with mentalizing (DMPFC, TPJ), as well as VS in the reward system.

So, we take pretty pictures of people who are not actually Buzzing, but thinking about ideas and if other people accept those ideas then those were pictures of Buzz. I’m waiting for an offer about a bridge.

But you’ve got the brand of Psychological Science saying that picture is a picture of Buzz which you can then cite in your glossy brochure or groovy website for persuasion consulting! You can actually make money with these pretty pictures and some pretty persuasive science, too!

Emily B. Falk, Sylvia A. Morelli, B. Locke Welborn, Karl Dambacher, and Matthew D. Lieberman. Creating Buzz: The Neural Correlates of Effective Message Propagation. Psychological Science, first published on May 30, 2013.

DOI: 10.1177/0956797612474670

P.S. Can’t resist this. I’m guessing that this paper got reviewed and accepted after Psych Science recently published a special issue on the perils and promise of fMRI theory and research. This paper exactly demonstrates all the peril and none of the promise as described in that review issue.