Category Archives: Health

all things morbid and mortal

Can’t We All Get Along?

Paul Mueller, writing for the August 12, 2012 issue of Journal Watch, talks out of all sides of his mouth while reviewing scientific science. He reviews a research paper on sitting and notes:

Comment: In this analysis, sedentary people, on average, lived 1 to 2 years less than active people. However, given the study design, the analysis does not prove a cause-and-effect relation. Nevertheless, the results are biologically plausible, and individual patients should be advised to avoid sedentary behaviors.

So, Mueller observes there’s no science (that “cause-and-effect” thingy) behind the claim that sitting kills, but, hey, the results are plausible and physicians should tell patients to avoid sitting. And this guy is an MD treating people and claiming he’s a competent reviewer of science.

There’s no clear behavioral definition of Sedentary other than the varying proxies Tooth Fairy epidemiologist manipulate from dataset to dataset, never using the same definition of Sedentary and always finding highly adjusted and trivial Windowpanes usually about one half of a Small Effect in observational designs. Yet Mueller puts on a white lab coat and thinks that makes his analysis and recommendation scientific.

What can a physician possibly say to a patient? “Avoid sitting.” Well, no, that’s the Not TACT, so instead, “Always Stand!” “Run or Walk or Fidget.” But for how long? Even in a bath tub? On the subway or bus? What about driving? Should I stand or run when having sex?

I’m being silly for the point. Stand Don’t Sit is a General Semantics play that fools everyone except the Tooth Fairies. What thing does the word point to? Physicians can pretend that they are doing their job when they warn about the dangers of sedentary behaviors, but the word doesn’t point to a concrete, observable TACT. It’s not like blood pressure or blood sugar or any single dimension variable. Sedentary is a merely a word that describes a vague complex of interacting behaviors over a long period of time. Yet, Mueller provides support for the play and everyone eats the menu or in this case talks a good game.

He’s only and obviously persuasive. He offends no one because he acknowledges everyone’s position without critical standards of judgment. In persuasion, that’s called Strategic Self Presentation. In science, that’s called Getting Ahead.

All Bad Science Is Persuasive.

Number of Choices as WATTage Switch

Hey, everybody’s prochoice because God created humans with volition and if you ain’t got choice, what good is volition? So choice is good . . . for what?

How about messing with the Other Guys WATTage?

Merely expand or reduce the number of choices available to Other Guys and you change how They think (and act) with all those choices. It’s not the content of the choice, but the number of options in the choice. Keep that front and center.

Levav, Reinholtz, and Lin present a 5 study package of lab studies that manipulate consumer choices in one of two patterns. Increasing runs the Other Guys through a series of choices that begin with a few options and then linearly increases to many options. Decreasing runs in the opposite direction where the consumer’s first choice has the maximum number of options that decreases linearly to the fewest. Here’s an example from Study 1. Participants were given the task of creating a mix of songs on a CD for a road trip. They had to select 10 tracks to fill the CD. For each track selection they were given a list of potential songs to choose from.

Participants in the increasing condition encountered an increasing number of options from track 1 to track 10 in increments of five (five options for track 1, 10 options for track 2 . . . and 50 options for track 10). In contrast, participants in the decreasing condition encountered a decreasing number of options from track 1 to track 10 in decrements of five (50 options for track 1, 45 options for track 2 . . . and five options for track 10). Thus, we employed a 2 (sequence: increasing vs. decreasing) × 10 (choice-set size: 5, 10 . . . 50) mixed design with sequence as a between-participants factor and choice-set size as a within-participant factor.

Here’s a graph to help visualize the two choice patterns.

Across the five experiments the Other Guys always had choices that varied with this Increasing or Decreasing width of number of choices. The first study used digital music. Another used consumer products costing around $5. A third used cartoon captions. (Hey, read the paper yourself if you want all the details.)

Now. All other things being equal there should be no difference if everyone is always prochoice. Choice is a good thing whether between a few options or a lot of options. But, turns out, that this is not true. Pattern makes a difference in how Other Guys respond.

When Other Guys start with a choice that has a few options and then successive choices increase the number of options, the Other Guys tend to look at more options, take more time on options, and in general behave differently compared to Other Guys who start with lots of options that then decreases. Here’s how the researchers put it.

This suggests that an increasing sequence of choice-set sizes is effective at evoking a maximizing mind-set in participants, who otherwise would be more inclined to satisfice.

Stated in Persuasion Blog terms, with Increasing Choice, Other Guys go High WATT and engage that search and scrutiny process indicative of Central Route thinking. In contrast with Decreasing Choice, Other Guys go Low WATT and “satisfice” which means they use Cues (or Heuristics here) to short cut their way to selections along the Peripheral Route. Choice size is an elaboration moderator that affects the amount and extent of issue relevant thinking to quote somebody smart.

The Windowpanes here are not Large. Most effects range around Small to Smallish so you need a bean counter to see the effect. Stated another way, for most of the experimental outcomes, if you were the proverbial grad student chained to the oars running subjects, you’d have no idea one group of people is thinking and acting differently than another. But, there is a difference, you can Count the Change, and it flows out of a solid body of literature and that experimental method. In geekspeak, the internal validity is pretty good here, but you do make the trade on external validity: Will this work in other situations, say, like the Real World?

I suspect you could get a big Hell, Yes from all the sales guys at the Apple iTunes store or any of those other online sales sites that stone you with choice, Choice, CHOICE. If they haven’t noticed it yet, I’ll bet they could easily run their sales data through this design matrix and find, how about that!, pointy headed academic research actually works sometimes!

[Professor Poopypants Sidebar: And how could the Big Data guys not know about this from all that fabulous Big Data New New Thing - tell me the Big Data guys know all about this . . . yeah, right. You could only figure this out if you had a good theory to make the predictions. If Big Data guys have any inkling of this, I'll bet it is tied up in knots of other variables made famous only through those little asterisks showing p < .05. But, I digress to point digital fingers. Back to the opera!]

So. Start Other Guys with a small set of choices and you can increase that set and get more work out of Them. Conversely, start Other Guys with wide choice sets, stone Their WATTage, and put Them on the Peripheral Route skipping Cues over the surface.

Let's get practical.

1. You've got to understand here that the Other Guys could not possibly see this variable. Most people probably only had the vaguest notion that the number of options was changing through the course of their work. Choice Set hides in plain sight, always a desirable feature for persuasion.

2. I'd call this a Box rather than a Play. You use this in some persuasion Local to predispose the Other Guys, to set Them up, to get Them looking here and not there. After you get Them in the Box, then you run either an Argument or Cue based play depending how you’ve thrown the WATTage switch.

3. If you have a Main Point that is complicated, use Increasing Choice sets to control your Other Guys until you get there. Start your pitch with a couple of narrow choice sets that are not the Main Point, but the set up for the Main Point. In other words, do not begin with a complicated Main Point that contains a lot of choices. Sure, that’s the McGuffin, the play for the TACT, but if you start with a wide choice set, you are hitting the Low WATT switch. That pushes your Other Guys down the Peripheral Route when your Main Point requires Central Route effort.

4. If you have to defend Weak Arguments (e.g. you’re selling a product or service that you know is not as strong as a competitor’s), bury Them in a wide choice set, then just assert you’ve got strong arguments. Throw in additional Cues (anything that shimmies, sparkles, or glows!) and move to the close.

Along with the Internet, lots of free pornography, and peace and prosperity, our world reeks with wide choice sets. See that now as a persuasion variable you can manipulate to your desired ends. Choice, like all persuasion variables, always has different effects under different Locals, Boxes, and Plays.

It’s the Combination, Stupid!

Jonathan Levav, Nicholas Reinholtz, and Claire Lin. (2012). The Effect of Ordering Decisions by Choice-Set Size on Consumer Search. Journal of Consumer Research , Vol. 39, No. 3 (October 2012), pp. 585-599.

DOI: 10.1086/664498

Picasso as WATTage Switch

The WSJ begins . . .

It’s been a moronic spring. But did it have to be? That J.P. Morgan Chase trader known as the “London Whale”—could he have been caught by a keener-eyed boss? How about the Secret Service scandal—would more astute overseers have spotted the signs? And the Pope’s butler—nobody saw that coming? Across the board, our perceptiveness has plummeted.

The writer then details more indicators of our growing inattention and then points to a potential cure. Strolling through art museums!

A “museum intervention” is now mandatory at Yale’s School of Medicine for all first-year medical students. Called Enhancing Observational Skills, the program asks students to look at and then describe paintings—not Pollocks and Picassos but Victorian pieces, with whole people in them. The aim? To improve diagnostic knack.

The article claims that this program improves diagnosis accuracy by 10% as described in an evaluation study published in the Journal of the American Medical Association. (Here’s a PowerPoint presentation of the intervention.)

We could quibble about a small and highly biased study that produces a Small Windowpane and ask whether Art Appreciation changes anything. But, since it was published in JAMA and JAMA only publishes the Real Truth, let’s take it as it is claimed. Where’s the beef with Art Appreciation?

The key skill in this effort must involve WATTage, that Willingness and Ability to Think about issue-relevant information, the Central Route as we engage the Long Conversation in the Head. Typically, we look at WATTage in the Other Guys as They deal with persuasive information, but see the general principle. When we engage higher WATTage, how we perceive and evaluate all kinds of information changes, whether for persuasion or diagnosis.

This is certainly a noncontroversial observation. Of course, High WATT produces different outcomes than Low WATT! And, usually the differences favor High WATT processing. You tend to make fewer errors in judgment, choice, and action going High WATT. Yet – and this is the interesting observation – people hate going High WATT, even physicians. It’s easier to Cue along the Peripheral Route or to go High WATT, but with Biased Processing, where you think only to confirm what you already know.

And, never forget Whitehead’s admonition about WATTage.

A society advances by the number of operations it can perform without thinking. Operations of thought are like cavalry charges in a battle — they are strictly limited in number, they require fresh horses, and must only be made at decisive moments.
Alfred North Whitehead, Introduction to Mathematics, 1911.

Thus, Art Appreciation appears to enhance not only observational skill, but more importantly, it enhances that WATTage switch. The museum stroll strengthens your WATTage. Let’s continue our stroll with that thought.

The practical problem with WATTage is not that people cannot engage High WATT, Objective Processing, or even that they don’t particularly like going High WATT. The problem is that they aren’t good at determining when to do it. It takes WATTage to know that you need to use more WATTage. You see the conundrum. To control WATTage, you need WATTage. Thus, you are asking something to not only accomplish a function or a goal, but to also assess itself while it is working. It’s like an athlete who is monitoring and evaluating every physical action during a skilled task like hitting a baseball. The more he thinks during the task about how well or poorly he’s doing the task, the more likely he will fail at the task.

Everyone, even smart people, thinks that they are thinking properly for the moment. It requires enormous discipline to control your WATTage on the fly and make adjustments. And, many situations will not permit that needed excess of WATTage to realize that you need more WATTage. Sure, if you are extremely smart and experienced, you will have a lot of WATTage, but when the situation creates more demand like a busy cockpit with tired pilots in a storm when there is an odd instrument reading, even with smart and experienced people, WATTage capacity will break.

Since I like strolling through art museums and value WATTage, I’ll gladly jump on board the Med School training that includes a trip to the Metropolitan. I suspect that Melanie would suggest a stroll through a garden, hot house, or arboretum to build observational skills. And you doubtless have your own ideas. But we must all realize that merely having WATTage does not solve the problem of using it effectively. And I suspect we never will because of all those persuasion operators out there daily doing their best to divert, distort, or otherwise mangle all the Other Guys WATTage.

That and human nature.

Illuminating Marriage with Persuasion

It’s about the Other Guy, Stupid, and most particularly Changing the Other Guy. You can Count the Change on four dimensions: Magnitude, Persistence, Resistance, and Prediction. You can aim at a little bit of Change that lasts long enough to get a glance, one night, or the Long Run. You get all my point and see the variety of Change you seek in Other Guys. And, of course, the variety you seek also determines your persuasion Box and Play. Lingering glances versus one night stands versus the Long Run each require different persuasion.

So, consider that for marriage.

I asked whether society should consider something like a 20-year marriage contract, my own modest proposal that, as in the one from Mexico, acknowledges the harsh truth that nearly half of marriages in the United States end in divorce and many others are miserable. The rough idea: two people, two decades, enough time to have and raise children if that’s your thing; a new status quo, a ceremony with a shelf life, till awhile do us part. But despite having proposed it, whimsically, as a journalistic expedition, I found myself surprised and even unnerved by the extent to which some experts I spoke with say there is a need to rethink an institution that so often fails.

Marriage is designed for the Long Run, you know, that until death us do part thing. And some people wish to take the Long Run and make it sprints of predetermined lengths and see how it goes after each finish line. Marriage is no longer a marathon, but a staccato of dashes that require a new registration after each burst.

As a persuasion guy, if I wanted to design persuasion that ensured no one ever made the Long Run, I would change the Box from a Marathon to a Sprint. Don’t permit the Other Guy to see the transaction as a Long Run, but always get Her thinking about intervals. Did she like the Honeymoon Race? How about the Newlywed Race? Now, the House ‘N Kids Race? How about the MidLifeCrisis Race? Empty Nest?

Get more specific. The He’s Drinking Again Sprints. Or, She Likes Them Younger Now. How about, But California Would Be Better Than Here. The ambition race: I Got A Better Job Offer There.

Make the Long Run nothing but a series of episodes, challenges, DIY projects with this important proviso: You can Change the Marriage after each Race.

You don’t need much persuasion savvy to see how the Short Runs Box and Play produces different Change compared to the Long Run Box and Play. If you want to diminish loyalty, commitment, persistence, perseverance, trust, tolerance, toughness, creativity, innovation, and, oddly enough, self-reliance, change the Box from Long Run Change to Short Run Change.

Now, take the Negativity Effect or the Bad Is Stronger Than Good Effect and run that through all those short run trials, tribulations, and finish lines. Simply give all the Married Other Guys regular, predictable, and unavoidable opportunities to engage Agonizing Reappraisal after each end point. Our human nature predisposes us to over-seek and over-weigh negative information. Think that might play any role in Short Runs Marriage? See all the naturally and required opportunities to finish an interval with bad news. And, you think most Other Guys will want to re-up for another race?

Marriage in the Short Runs is a persuasion dream for stimulating selfishness, counterfactual thinking, short sidedness, self destruction through justification, and on and on with the failures that arise from always having to evaluate. When people decide whether they like or dislike something, the very nature of the evaluation process plays against the Long Run. Merely provide recurring attitude evaluations of Marriage at various intervals and you will guarantee a Change.

What looks like freedom or rationality or even PostModern2.0 is nothing more than persuasion in action. Hey, if you are a divorce lawyer, you need to get behind this idea in a big way!

Interference Penalty – a Persuasion Dysquotation

The Feds are changing regulations regarding school lunch programs in part because the regulations are causing a lot of trouble and are producing no Change worth Counting. This is good according to the Food Police who argued for these regulations and their obvious beneficial effects, however unCountable.

Margo Wootan, a nutrition lobbyist for the Center for Science in the Public Interest, says the change is minor and the new guidance shows that USDA will work with school nutrition officials and others who have concerns. “It takes time to work out the kinks,” Wootan said. “This should show Congress that they don’t need to interfere legislatively.”

Congress doesn’t need to interfere legislatively?

How else can They interfere? The only thing Congress can do is legislate. That’s it. That’s the power the Constitution gives the Congress, and any other kind of interference would be illegal.

And, how can any action Congress takes regarding the laws it passed be considered “interference?” Congress enacted the laws the Food Police wanted, the laws, as turned into Executive action through regulation, did not work as predicted and now it would be “interference” if Congress passed new laws or amended old ones?

Look. The Other Guys here are not journalists and citizens. The Other Guys here are those interfering fools in Congress who passed the laws Wootan wanted in the first place. To tell your Other Guy that She’s interfering when She’s doing Her job is not the sharpest persuasion play to run here.

P.S. Gee whiz. When the newspapers are calling the Director of Nutrition Policy at the Center for Science in the Public Interest, a Nutrition Lobbyist, you’ve got trouble in River City (YouTube). You got trouble, I say. Trouble. Trouble. Trouble.

Mavens! Run, Don’t Walk, for This One

Of course, there’s no science behind this and that’s why it’s your persuasion opportunity!

Yeah. Some people benefit from exercise, others don’t, and some even die from it! How to tell the difference? Genetic Testing!

The new test, which is being sold by a British company called XRGenomics, is available to anyone through the company’s Web site and involves rubbing inside your cheek with a supplied swab and returning the tissue sample to the company. Results are then available within a few weeks. It is based on a body of research led by James Timmons, a professor of systems biology at Loughborough University in England, and colleagues at the Pennington Biomedical Research Center in Louisiana and other institutions.

Any company called XRGenomics is already leaning forward towards persuasion. And, it’s led by a real live professor, so you’ve got the expert in the lab coat. Plus, he’s got the Gene Screen©™® to tell you whether you will run and: Live or Die or just Huff n’ Puff.

But, like I said, there’s no science behind this Gene Screen©™®. An expert I respect who doesn’t make many public appearances in a lab coat notes:

Their actual predictive value, based on the best currently available genetic science, “is approximately zero,” says Claude Bouchard, a professor of genetics at Pennington, who was the senior author of the 2010 study with Dr. Timmons, but has no involvement with XRGenomics.

Gee, a coauthor calls you out in public. That’s gotta hurt. Except we need to observe this about Bouchard.

He is a paid consultant for another company, Pathway Genomics, that offers gene tests only through physicians.

So while Bouchard is smart enough to avoid lab coat public appearances, maybe he understands the inside deal with consultancy?!?

You can make that inference if you wish, but Bouchard’s work doesn’t require his public pronouncements. His data speak for themselves. We simply don’t know enough about the genetics of exercise to offer a useful predictive test for most people. Yeah, maybe there are the OGODs, the One Gene One Disease tests that bear on exercise, but the kind of test implied by the NYTimes story? No way.

Don’t take my word for it. Take the numbers from Pennington at XRGenomics.

In the original 2010 gene study, the authors concluded that the gene profile they’d uncovered accounted for at least 23 percent of the variation in how people responded to endurance training, which, in genetic terms, is a hefty contribution.

Stated another way, that 23% explained variation is almost a Large Windowpane, a 25/75 effect. Yet, in this context even that Large Windowpane is almost meaningless in practical circumstances.

1. Hey, 77% of the variation in exercise effectiveness is nonGenetic, meaning lifestyle factors like specific exercise routine and diet and substance use (steroids, anyone?).

2. That 23% is not your individual mileage, but rather the group average for the sample under study. Among all people, the Gene Screen©™® accounts for about one quarter of the population effect. You cannot generalize that group effect to a specific individual.

3. And it’s all observational. You cannot randomly assign people to different genes (and lifestyles) then Count the Change. These are correlational designs with all the trials and tribulations of bias and threats to internal and external validity.

When I arrived at NIOSH in the late 1990s, molecular genetics was in full swing and I worked with some world class researchers dying on the bleeding edge of genetics work. What impressed me most from talking with them was their bemusement at the New New Thing hype of Big Genetics and how so many people were seeing the work as a Revolution that would Revolutionize everything. Even back then the guys in the caves were already just doing the grind that is basic research and none of them expressed anything remotely like a zealot’s belief.

Genetics has not yet produced the Revolution that will Revolutionize everything and it may yet. The OGODs clearly demonstrate that, but we knew about OGODs without the Human Genome Project.

But, get yourself an expert with a lab coat, a groovy sounding company name, and best of all a Gene Screen©™®, and, baby, you can do some persuasion!

Getting TACTful about the Weather

Sure, If You Can’t Count It, You Can’t Change It, but!

If you can’t even define It, specifically as a TACT, you’re not even doing persuasion. As with the latest effort from the UN conference on Climate Change.

Delegates from more than 190 nations agreed to extend the increasingly ineffective Kyoto Protocol a few years and to commit to more ambitious — but unspecified — actions to reduce emissions of climate-altering gases.

The NYTimes article details the continuing inability of UN climate changers to agree on anything remotely approaching a TACT, the Who does What When and Where or the Target Action Context and Time. Consider this another way.

I have science. Cold hearted, irrefutable science that conclusively says X causes Y, but I can’t get a group of like minded scientists to agree on a TACTful statement of that science.

See the failure of both the Falling Apples and the Fallen Apples. There is no science and no persuasion here or else you would be stepping on TACTs like a barefoot visitor in a nail factory. Sincerity, however, abounds!

And lest anyone think the persuasion problem resides only within the professional meeting attenders at the UN, consider the mavens in this area.

I’ll take just one example of a persuasion campaign for Climate Change suggested by the good folks at Breakthrough Strategies. I don’t know these guys from Adam, Eve, or the Serpent and have no axe to grind, stone to throw, or eye to gouge. It sounds pretty much like other free campaigns I’ve seen from similar organizations whether from government, NGO, or even the charitable arm of forprofit units. They do persuasion for Good Hearted People with Money, including Climate Change Affirmers. They have published a free Guide (pdf) based on their persuasion analysis of the problem.

Here’s their point of entry.

While many have approached climate change cautiously, Americans have grown impatient with the lack of action and the excessive influence of fossil fuel interests. They respond positively to honest, confident problem-solvers who will lead our communities, nation and world to a stronger and safer future. This guide is for those leaders.

And how can you move the People? A persuasion campaign based on three elements.

A large body of recent research shows a solid majority of voters respond favorably to confident, pro-clean energy, climate leadership messages grounded in three core American values:

Responsibility. Stepping up to the climate challenge is the right and responsible thing to do.

Patriotic pride. America can rise to the challenge and succeed.

Accountability. We can’t allow the billionaire Koch Brothers and Big Oil to continue to rig the system and block clean energy solutions.

These mavens suggest a straight Central Route Arguments play for High WATT processors. Argue with Responsibility, Patriotism, and Accountability. People will engage the Long Conversation in the Head as they elaborate on Responsibility, Patriotism, and Accountability which in turn will produce favorable attitude Change which will then lead to favorable behavior Change.

Do you really believe this will work?

1. Nothing in the Guide says anything about creating High WATT processing. They just advise Arguments without offering any tactics for getting the kind of willingness and ability to think. And, I’d like to see a lot more about their data on Argument quality for Responsibility, Patriotism, and Accountability. Good grief, no one is against concepts like these, but the devil is in how you state them. Breakthrough Strategies provides these Mom, Apple Pie, and Flag vapories without any thought about how to express them in a way that generates a favorable Long Conversation in the Head. Worse still, this has been a primary persuasion play from Climate Changers for the past 18 years since the first UN meeting on this topic. In other words, there’s no new persuasion here.

2. Color me stupid, but I fail to see any TACT, Who does What When and Where in the Guide. Sure, the People want Change, but that’s not a TACT, a specific statement of Target Action Context and Time. That’s a vague, feel-good statement that has no concrete, observable behavior. So, what behavior will the three Arguments produce? It might attract those Good Hearted People with Money which is good for a consultant, but does it change the Other Guys who will then do something about Climate Change? I cannot see a simple straightline from Breakthrough campaigning to TACTful Climate Change.

The lesson here is not about specific people, organizations, and issues but about persuasion skill. All persuasion must begin with a good TACT. Without that, you have no direction, no control, no count. You’re in a 1930 MGM movie musical with Judy Garland and Mickey Rooney that says, hey, I’ve got ten dollars and a pair of tap shoes, let’s put on a show that raises money to save the orphanage! If you are staging musical productions, then study all the Climate Change Affirmers! Of course, while they are putting on a show, they ain’t saving the planet or even an orphanage.

Even before It’s about the Other Guy, Stupid, it’s about the TACT.

Nudging in Living Color

Read the good news.

Now there may be another approach, based on the same logic that maintains order at traffic intersections around the world: people respond to color. A new study published Aug. 6 in the American Journal of Preventative Medicine suggests that when it comes to nutrition, people may pay more attention to simple color-coded warnings: Green for healthy, yellow for less so, and red for calorie bombs.

Yeah. Simply add color to your warning label and it works. Why?

In part, Levy said, it’s that many people just aren’t very good at math. “One of the reasons why we’re interested in doing it this way is that the data out there on numeracy and literacy shows that people, even with high literacy and numeracy, don’t necessarily understand nutritional labeling particularly well,” he said. “And certainly folks with lower numeracy and literacy would have more difficulty putting calorie numbers into context.”

Gee. Who knew you had to take off your shoes when dining? Turns out you gotta know math to eat correctly. And color coded math warning labels allow people who aren’t good with numbers to find the right food.

Math. That sounds like Counting the Change, doesn’t it? Let’s count the Change here.

Start with the intervention.

The labeling intervention began in March 2010. During 1 weekend, all food and beverages were labeled red, yellow, or green on the menu board located either directly over the individual food station, directly over the shelf where the food was located, or directly on the packaging. The labeling intervention was advertised as the Massachusetts General Hospital Choose Well, Eat Well program, and the message to cafeteria patrons focused on making a better choice. We posted new signage to describe the labeling on a wall in the cafeteria as well as on 2 large columns in the middle of the cafeteria. This signage highlighted that green meant “consume often,” yellow meant “consume less often,” and red meant “there is a better choice in green or yellow.” Rather than tell patrons to stop for red items, we used a positive yet clear message to redirect patrons toward a better choice. During the first 2 weeks, a dietician was available in the cafeteria to answer questions about the labels. Throughout both phases 1 and 2, we supplied the cafeteria with pocket-sized pamphlets containing information about the labeling as well as the specific amount of calories and fat in all items.

This is a flat out information persuasion campaign in a large health center cafeteria. You are providing Arguments, information that bears on the central merits of the attitude object, in this case, food, and its nutritional content. The color coding simplifies all the information in a Cue-based fashion and if the Other Guys trust the source of information, They don’t have to check each food item to make sure the color code is properly applied or even get into a food fight over whether that Yellow should be a Red because it is genetically modified.

So, this is a nice setup by my lights. You’re running on the Central Route with lots of information about nutrition plus you layer a simple, familiar kind of Cue with that Red Yellow Green color coding that accurately reflects all the math in your Arguments. Kind of a Ding-Donged Central Route play, right? Ding-dong the Color with the Math so that people can easily and quickly associate complex nutritional information with reliable and honest Cues that simplify choice and action right at the point of TACT.

But, wait. There’s more.

Phase 2: Choice architecture intervention. In June 2010, we began the choice architecture intervention. We made the changes for this phase over a weekend and did not advertise them. The main target items for phase 2 were cold beverages, premade sandwiches, and chips. We chose cold beverages because they represented a large portion of cafeteria sales (20% of overall sales), and we hypothesized that location and convenience would influence beverage purchases. We also hypothesized that location and convenience would influence the sales of chips and premade sandwiches because cafeteria patrons who do not have a lot of time to spend in the cafeteria are likely to purchase these items.

The Nudge hits the cafeteria! Through the strategic arrangement of choice, you Change the Other Guys. Here’s the Nudge detailed.

We rearranged all 5 beverage refrigerators so that the green beverages (including water, diet beverages, and low-fat dairy products) were located at eye level and yellow and red beverages were located below eye level. We defined eye level as a height between 5 and 6 feet. During baseline and phase 1, bottled water was available in 2 refrigerators that were not centrally located in the cafeteria, similar to the cafeteria layout before the study started (Figure 1). During phase 2, we added bottled water to the other 3 beverage refrigerators and added 5 baskets of bottled water throughout the cafeteria near the food stations (Figure 1). We rearranged the premade sandwich refrigerator so that the green sandwiches were located at eye level and the yellow and red sandwiches were below or above eye level. Chips were located on 2 adjacent racks, and we placed the yellow chips on the higher eye level racks and the red chips on the bottom (no chips were rated green).

Well, you can call that a choice architecture, but you’re just moving the cookie jar where the kid or the dog or the spouse can’t easily find it. I have a little trouble when you call Hiding It something like Choice Architecture. You’re not Nudging here in the sense of doing something that pushes or pulls the Other Guy ever so slightly. You’re putting up barriers, obstacles, and costs to frustrate, confuse, or distract a behavior. For me, calling this Choice Architecture suggests a different Other Guy for persuasion: Me or you, but not those folks in line at the cafeteria.

But, let’s not quibble over mere words, definitions, and concepts. We know what everyone is doing. Phase 1 is a Central Route Argument plus that layered Ding-Dong color Cue. Phase 2 is . . . a Nudge.

And all this makes a difference as we know from the news reports. Do we really need to Count the Change? Mark Bittman doesn’t! Heck. Since we’re all propeller-head, bean-counters, here, how about a little data . . .

. . . which takes us to a nice feature about this intervention: How they Count. Employees of the health center have Dining Cards they present at checkout. The researchers therefore have the same kind of information Big Food and Big Marketers collect on you in the grocery store. They know to the penny what you bought and who you are in that Anonymized Data way. The researchers got all cafeteria sales data for 3 months before the intervention and collected the same data during and after. It would be nice to have another similar work site as a control and nicer still to execute this on a hundred cafeterias randomly assigned to treatment or control, but this is still a strong data design. You’ve got pre-post data that is about as good a Count as you can get with this TACT.

Click to enlarge.

The reporting is a bit confusing. If you navigate merely by statistical significance (those p values), it looks like nothing worked because you don’t see a bunch of “p < .05″ or smaller indications. If you read the fine print, you learn that all numbers in boldface are p < .05 which means something happened. Roughly speaking the Phase 1 Central Route intervention produced about a 11% decrease (Small Windowpane) in Red food items and about a 7% increase (half a Small Windowpane) in Green food items. The Phase 2 Nudge produced another 4% decrease in Red food items and, this isn’t good, a 2% decrease in Green food items. Clearly very Small Windowpanes here, but statistically significant because you have over 4,000 cases in a repeated measures design which provides power out the wazoo.

Now, the researchers also present analysis from a segment on Beverage purchases and finds better effects. Those show Small+ Windowpanes consistently across all Phases. I discount the importance of this finding because nothing in the intervention was aimed at producing more Change in beverage purchase compared to other items. It’s all about Red, Yellow, or Green regardless of food category. When you start grabbing the best findings from the total, you will always see larger effects but that’s because you are cherry picking. Believe me, if sales of a different category like sandwiches or soups had popped larger effect sizes, you’d be reading all about that instead.

And, you can partition these sales data however you wish. The authors make a big deal about ethnicity and job classification differences with concerns of education and literacy. And, you can see some variation there, but for me it’s just Gertrude Stein’s Oakland with not much there, there (except for the glory days of Da Raiders).

And, in fact, the intervention taken as a whole doesn’t make much Change. The main point for me is a Small Windowpane effect on Red items. The minor point is that the Nudge clearly had costs that hurt the minimal gains. Hiding the cookie jar clearly had some Reactance type effects as it decreased purchase of Green items. Sure, you can find something special with the Left Handed Gay Armenian segment, but the intervention is aimed at everybody seated at the Table of Brother and Sister Hood.

Now, the pop press takes this as a New New Thing in the Global War Against Food Terrorism when the data suggest a rather weak intervention that is a standard public health education in a health center cafeteria. In other words, you’re doing an Old Old Thing in the most favorable circumstance you can find, rather like preaching in the soup kitchen. You’ll get a lot of Amen, Brother, but lose the soup and you lose the congregation.

And you get a Small effect for all the nagging, oops, the education on the Central Route. Imagine doing this on a food truck that hits construction work sites. Guys would buy carrots just to throw them at the signs. And imagine Hiding The Cookie Jar under these conditions. You’d earn combat pay for playing interventionist on that battlefield. And, you’d probably lose the war and most of the battles, too.

I appreciate the effort and concern here, but I disagree with the Change and its Count. The numbers clearly show the Nudge had negative effects along with its small benefits (which is consistent with the Nudge literature if you actually read the Methods and Results section). The small overall changes flow from a standard public health education model as implemented in a congenial public health setting, a health center. Nothing in this report is exaggerated or manipulated the way you see in Tooth Fairy Tales and everything is exactly and precisely correct. It just doesn’t Count for much Change.

P.S. Choice Architecture Nudge? Do this. Stand in front of a bunch of folks with low education and literacy who have trouble thinking for themselves about food and show and tell your Choice Architecture Nudge. After a few minutes, stop, look around the room, and feel the vibe. Maybe you’ll understand why Romney lost the election.

Douglas E. Levy, PhD, Jason Riis, PhD, Lillian M. Sonnenberg, DSc, RD, Susan J. Barraclough, MS, RD, Anne N. Thorndike, MD, MPH. (2012). Food Choices of Minority and Low-Income Employees: A Cafeteria Intervention. American Journal of Preventive Medicine, Volume 43, Issue 3, September 2012, Pages 240–248.

doi: 10.1016/j.amepre.2012.05.004

Thorndike, A. N., Sonnenberg, L., Riis, J., Barraclough, S., & Levy, D. E. (2012). A 2-Phase Labeling and Choice Architecture Intervention to Improve Healthy Food and Beverage Choices. American Journal Of Public Health, 102(3), 527-533.

doi: 10.2105/AJPH.2011.300391

A Null Fairy Tale or Take This Pill, Act IV

As I often demonstrate on the Persuasion Blog, many empirical health and safety studies are exercises in applied CSI:Seinfeld as researchers make their data a Show About Nothing. Through the magic of story telling with sophistical statistics, they entertain us with strong narrative in a detective story with many false leads, but always ending with a compelling CSI case against sitting or sugary soda pop or cellphone driving or vending machines and on and on.

But sometimes their data truly is a Show About Nothing and the Tooth Fairies admit it. Consider this recent episode in the continuing series called Take This Pill with men randomly assigned to multivitamins or placebo and that impact on various cardiovascular outcomes. Of course, you don’t get a simple cross tab table of pill versus placebo by dead versus alive even though you’ve got an experimental design that tests exactly that proposition. No. To reach a null conclusion you need to do this first: Adjust.

We initially compared baseline characteristics by multivitamin treatment assignment to ensure that randomization equally distributed baseline characteristics by active vs placebo groups. As done in previous PHS II trial analyses,22,33 Cox proportional hazards models estimated hazard ratios (HRs) and 95% CIs, comparing event rates in the multivitamin and placebo groups. For each prespecified end point, we stratified on the presence of CVD at randomization and adjusted for PHS II study design variables, including age (in years), PHS cohort (original PHS I participant, new PHS II participant), and randomized vitamin E, vitamin C, and beta carotene assignments. For analyses of total major cardiovascular events, all new events were included, regardless of whether the participant had a baseline history of CVD. Analyses of individual cardiovascular end points did not censor men on occurrence of another cardiovascular end point. For analyses of total and cardiovascular mortality, we included all 14 641 PHS II participants; for total mortality, we additionally stratified by history of cancer at randomization.

Whew. But, wait. There’s more!

We tested the proportional hazards assumptions by including an interaction term for multivitamin treatment with the logarithm of time; this assumption was not violated for major cardiovascular events, total MI, and total stroke (P > .05 for each). Cumulative incidence curves compared the overall effect of the multivitamin component on major cardiovascular events, total MI, and total stroke using a crude log-rank test. We investigated the effect of adherence to the multivitamin intervention on our primary results using sensitivity analyses with censoring and stratification.

Man. They’re still going and going like the statistical energizer bunny! Wasn’t this an experimental design with random assignment to pill or placebo? Doesn’t an experimental design kill the need for adjusting? Apparently not.

We then conducted additional exploratory analyses on the effect of the multivitamin intervention on major cardiovascular events, total MI, and total stroke after excluding the first 2 or 5 years of follow-up to explore a possible early or late benefit associated with long-term multivitamin use. We also conducted subgroup analyses stratified by major risk factors, parental history of MI at ages younger than 60 years, and selected coronary biomarkers and dietary factors available in a subgroup of PHS II participants. We evaluated the effect of a daily multivitamin within the prespecified subgroups of 754 men with and 13 887 men without a baseline history of CVD. Effect modification was assessed using interaction terms between subgroup indicators and randomized multivitamin treatment assignment.

I’m exhausted just from the cut and paste. I can’t imagine how tired you are if you read everything like a good grad student. Okay. We’re fully adjusted. What about the results of the experiment?

Among this population of US male physicians, taking a daily multivitamin did not reduce major cardiovascular events, MI, stroke, and CVD mortality after more than a decade of treatment and follow-up.

And, they’re not kidding. Here’s a summary of the main statistical findings.

During a median follow-up of 11.2 (interquartile range, 10.7-13.3) years, there were 1732 confirmed major cardiovascular events. Compared with placebo, there was no significant effect of a daily multivitamin on major cardiovascular events (11.0 and 10.8 events per 1000 person-years for multivitamin vs placebo, respectively; hazard ratio [HR], 1.01; 95% CI, 0.91-1.10; P = .91). Further, a daily multivitamin had no effect on total MI (3.9 and 4.2 events per 1000 person-years; HR, 0.93; 95% CI, 0.80-1.09; P = .39), total stroke (4.1 and 3.9 events per 1000 person-years; HR, 1.06; 95% CI, 0.91-1.23; P = .48), or CVD mortality (5.0 and 5.1 events per 1000 person-years; HR, 0.95; 95% CI, 0.83-1.09; P = .47). A daily multivitamin was also not significantly associated with total mortality (HR, 0.94; 95% CI, 0.88-1.02; P = .13). The effect of a daily multivitamin on major cardiovascular events did not differ between men with or without a baseline history of CVD (P = .62 for interaction).

Nothing but non significant P.

What’s funny about this is that if you closely read many other Fairy Tales, you’ll find ratios of the same magnitude (that HR of .93 for myocardial infarction, for instance) that can be adjusted into statistical significance in other data sets. It’s the same effect size, but since they couldn’t make sophistical statistics achieve significance, they now interpret it as a Null Finding of no effect.

This may be the deadest dataset to ever achieve publication status in any journal in the history of peer review science. Imagine this comparative thought experiment. You create a dataset of 100 variables and for each variable you create a score from a random number generator. Let’s do this for 15,000 cases. Then for each case, randomly assign them to 1 for Pill and 0 for Placebo. Now. Run as many regression models as you’d like on what is entirely a dataset of random variables. You know what? You’ll find more statistically significant effects for the randomly assigned treatment than these researchers did with their multivitamin study.

Make sure you get that. A constructed dataset of random variables will find 5 statistically significant effects if you run 100 tests of variables. And, of course, you can run considerably more than 100 tests of statistical significance with 100 variables and all their combinations, doing what these researchers did with their dead multivitamins.

Stated in plain terms, they turned the diamond of their data, looked at every facet, and had to conclude it was only a lump of coal after all when even a random dataset with nothing more than dumb luck and randomization would provide a glint of light.

P.S. They did find one statistically significant effect. And they noted:

The reduction observed in fatal MI (P = .048) may have been attributable to chance.

A p value less than .05 and it may be attributable to chance!?! The mind boggles at this observation. Do the Tooth Fairies truly understand the meaning of statistical significance after all? Doesn’t this admission nullify all their other Fairy Tales with datasets producing just a couple of “p < .05″ notations that are just as likely a chance finding from running a 100 tests?

Sesso HD, Christen WG, Bubes V, et al. Multivitamins in the Prevention of Cardiovascular Disease in Men: The Physicians’ Health Study II Randomized Controlled Trial. JAMA. 2012;308(17):1751-1760.

doi:10.1001/jama.2012.14805.

MOOC as Aspiration without Perspiration Persuasion Play

The NYTimes provides another nice look at MOOCs, those massively open online courses from elite universities taught by elite professors.

MOOCs have been around for a few years as collaborative techie learning events, but this is the year everyone wants in. Elite universities are partnering with Coursera at a furious pace. It now offers courses from 33 of the biggest names in postsecondary education, including Princeton, Brown, Columbia and Duke.

Elite! And, MOOCs are growing faster than Facebook!

The paint is barely dry, yet edX, the nonprofit start-up from Harvard and the Massachusetts Institute of Technology, has 370,000 students this fall in its first official courses. That’s nothing. Coursera, founded just last January, has reached more than 1.7 million — growing “faster than Facebook,” boasts Andrew Ng, on leave from Stanford to run his for-profit MOOC provider.

So, MOOCs are elite, big, growing, and new. Exactly the criteria that produce high quality learning . . .

And, that’s the Main Point with MOOCs. They are a General Semantics persuasion play when it comes to learning. MOOCs are just the Word and not the Thing, the Map and not the Territory, the Menu and not the Dish, the Appearance and not the Reality of learning. Please just think with only common sense. Why didn’t instructional media with TV or radio produce MOOCs that Educated The World? Sure, you can argue about technical differences between these different media, but why didn’t Instructional TV survive past a few boutique programs in the 1950s and 60s?

Because it doesn’t work.

You cannot produce high quality learning this way. People do not acquire competence in knowledge, application, or analysis with MOOCs. Anyone with training in some pedestrian field like educational or learning psychology can point to a voluminous literature littered with both Rules and Laws of Learning that explain this.

Now, here’s the cool part of that Truth. Most people won’t believe it, even elite professors at elite universities. Many of them will sincerely believe that the MOOC Changes Everything because of that networked computer fantasy. And that sincerity is useful for pushing MOOCs out into the public where foolish people can waste their time and end up with little more than a t-shirt proclaiming, I Did Harvard.

Read this.

Nick McKeown is teaching one of them, on computer networking, with Philip Levis (the one with a shock of magenta hair in the introductory video). Dr. McKeown sums up the energy of this grand experiment when he gushes, “We’re both very excited.” Casually draped over auditorium seats, the professors also acknowledge that they are not exactly sure how this MOOC stuff works. “We are just going to see how this goes over the next few weeks,” says Dr. McKeown.

Hi. I’m an elite expert teaching at an elite university doing a MOOC on computer networking and I have no idea how MOOCs work, but I am an elite educational expert at an elite university.

There it is, hiding in plain sight, the persuasion play behind MOOCs. Elite professors at elite universities publicly admit that they are idiots and don’t know either the Laws or Rules of Learning, but they still attract millions of online participants. Clearly elite MOOCs are all Appearance and No Reality when the teachers themselves don’t know what or how or why about their educational efforts.

MOOCs succeed precisely because they come from elite universities with a great brand of teaching and learning. This is pure Cue for Low WATT NYTimes readers. Grab the elite brand through a MOOC and viola, you’ve got an elite education. Free no less. And completed while wearing your underwear and favorite slob suit!

MOOCs even steal from of all those Do It Yourself videos on the Internet. And that’s a good thing.

Thanks to Khan Academy’s free archive of snappy instructional videos, MOOC makers have gotten the memo on the benefit of brevity: 8 to 12 minutes is typical.

The instructional benefit of brevity. The 8-12 minute barrier. Brevity is a Law of Learning? Even a Rule of Learning? Anyone who knows anything about Learning knows that the single best predictor of learning is Time On Task. The longer learners spend on the thing-to-be-learned, the better they learn it.

Yet the MOOC-ers seem to operate under the impression that brevity is not only the soul of wit, but also learning. The stupidity and ignorance of that belief is exceeded only by the arrogance of those who express it.

For me, MOOCs are borderline fraud and the only reason elites won’t go to jail for it is because they are free. Which means that caveat emptor applies: Buyer beware! MOOCs are delivered as is, with no representations about learning, quality, or value.

Count the Change. Show me concrete, observable evidence that people who take MOOCs have the same competence as people who took a standard course with a teacher. Then replicate it. Then show it over many different kinds of students and many different kinds of learning.

You cannot. The evidence does not exist. Yet elites present MOOCs with an implicit seal of approval, a transparent Authority Cue, delivered to Low WATT Other Guys who’ll gladly eat the screen in lieu of actually grinding out the Time On Task the Falling Apples of Learning require.

Hey, when scientists won’t show you their Windowpanes of effectiveness, you have to keep your head on a swivel and your hand on your wallet.