2008 Guestbook
Comments and Responses

name:	Thavaraj Subramaniam
field:	Risk Communication Technical Group for Ministry of Health and ASEAN
date:	September 29, 2008
email:	s_thavaraj@yahoo.com
location:	Malaysia

what i would add to this site:

Do you have any risk communication plan for Dengue control that I can refer to?

comment:

I have been reading your articles since 1999. I have learned a lot from you.

There is currently a debate in Malaysia re: What’s are differences and similarities between Risk Communication (RC) and Health Education (HE)? I have been saying that RC is a tool used during crisis and HE is to warn people of risks in non-crisis situations.

I may be wrong? I stand to be corrected.

peter responds:

I haven’t done any work specifically on Dengue, I’m sorry to say. But I do have a lot to say about the relationships among risk communication, crisis communication, and health education, and how these terms are defined.

As I’m sure you realize, all three terms (but especially “risk communication” and “crisis communication”) have a lot of quite different definitions. The way they are typically used has varied over time. More importantly, their use varies across disciplines. I will try to tease apart the main similarities, differences, overlaps, and confusions.

Risk communication

“Risk communication” is the newest of the three terms. It arose in the 1980s with regard to U.S. environmental controversies. Regulators and companies were concerned that members of the public were often excessively alarmed about small environmental hazards. Risk communication developed as a set of strategies to convince people that something (a nuclear power plant, a factory’s air or water emissions, a landfill) wasn’t nearly as dangerous as people thought it was.

It was obvious from the outset that risk communication was a daunting challenge. And given this definition, it was also obvious that it was a value-centric activity subject to abuse. That is, there were grounds for worry that it wouldn’t work, and grounds for worry that if it did work it would be used not just to reassure people about small risks but also to over-reassure them about serious risks.

In the three decades since the term first came into use, its definition has expanded. It is now widely used outside the environmental arena. And it is used outside the realm of reassurance. Today a doctor trying to reassure patients about the side effects of some medication is seen as doing risk communication – but so is a doctor trying to warn patients about the dangers of some lifestyle choice (eating too much, exercising too little, smoking) or a doctor trying to explain the pros and cons of two courses of action (surgery versus medication) so patients can make up their own minds. Some people use it as an extremely general term, meaning little more than “communicating about risk.”

Others – myself among them – have tried to limit the term “risk communication” to communicating about risk in a way that is grounded in research and theory, and that respectfully takes into account:

• social, cognitive, and emotional aspects of risk perception;
• the need for public involvement and candor; and
• a diagnosis of how technically serious the risk is and how upset the audience is, and a choice of communication approaches based on that diagnosis.

We welcome the broadening of the term beyond reassurance about small risks, but we resist its broadening to include communications about risk that are not purposeful and strategic; not informed by research and theory; or not respectful, honest, and two-way.

But risk communication still carries its historical association with reassurance, especially when applied to controversies. Activists, for example, very rarely use the term to describe their own behavior; to most activists “risk communication” (both the term and the activity) reeks of corporate efforts to deflect the public’s legitimate concerns.

The tendency of government and corporate sources to be more preoccupied with avoiding “undue concern” than with raising “due concern,” even in situations where the danger is serious, helps to sustain this association of “risk communication” with reassurance and even over-reassurance. Thus there are two reasons why critics of risk communication see it as a tool of powerful interests trying to quiet distress and quell dissent: because the term got its start as a label for efforts to reassure the public when a risk is small, and because powerful interests do tend to want to reassure the public even when a risk is serious.

Crisis communication

“Crisis communication” has its origins in public relations, where it meant and still means talking to unhappy stakeholders in a situation that is bad for the client’s reputation. To a PR person, a “crisis” is a reputational crisis for the client (and if the client is a profit-making organization, a profitability crisis as well). The term is used regardless of whether the unhappy stakeholders have actually been endangered or damaged or only think they have. If your milk company has allowed melamine to get into your product, giving kidney stones to thousands of babies, that’s a crisis. If a false rumor that there’s melamine in your product is circulating on the Internet, from a public relations perspective (though not mine) that’s also a crisis.

Most of public relations involves selling the client’s virtues to an uninterested public. When the public is interested, and the interest is negative, the everyday, routine PR people give way to the crisis communication specialists. Decent PR professionals certainly understand that companies should respond differently when they’re guilty than when they’re innocent – but in both cases the PR crisis is the simple fact that they have been accused.

But outside of public relations, the term “crisis communication” is widely used to mean communicating in a situation where it is urgently important for people to hear what you have to tell them so they can take appropriate precautions – that is, in a real crisis. In this non-PR use of the term, it’s a crisis for your audience, not just for you.

The term “emergency communication” is also used to apply to efforts to tell people what’s going on and what they should do in crisis situations. Insofar as the two terms are different, “emergency communication” is reserved for acute and often widespread dangers to life and limb – fires, terrorist attacks, imminent hurricanes, etc. “Crisis communication” certainly applies to these sudden emergencies, but it is used also for slower not-quite-emergencies, and for urgent threats to wellbeing even when they are not health or safety threats. What political and business leaders are saying right now about the danger of economic meltdown is clearly crisis communication; it doesn’t feel quite right to call it emergency communication.

Health education

“Health education” is the oldest and clearest of the three terms. Most people both in and out of the field use this term to mean either or both of two things:

• information about health that people in the target audience don’t know and will find useful to improve their health or protect themselves from unhealthful conditions or activities; and
• advocacy of health-promoting behaviors that people in the target audience already know to be wise but are insufficiently motivated to do.

Some educators would reserve the term “health education” for the first of these two tasks, the simple provision of health information. But in practice, health educators have long since discovered that apathy, fatalism, and other motivational deficiencies are as damaging to health as ignorance is. And most have long since realized that an audience that can pass a test on smoking or sexually transmitted diseases but continues to smoke or have unprotected sex is not an educational success story. Effective educators know how to move their audience, not just how to inform it, and self-aware educators know that behavior change is the goal and information is only the means.

And so the term “health education” has come to include advocacy, coexisting with other terms (“health promotion,” “social mobilization,” etc.) that are more candid about the advocacy goals of health education.

How I use the terms

For me, the most important fact about risk communication is the incredibly low correlation between a risk’s “hazard” (how much harm it’s likely to do) and its “outrage” (how upset it’s likely to make people). If you know a risk is dangerous, that tells you almost nothing about whether it’s upsetting. If you know it’s upsetting, that tells you almost nothing about whether it’s dangerous.

Based on this distinction, I categorize risk communication into four tasks:

• When hazard is high and outrage is low, the task is “precaution advocacy” – alerting insufficiently upset people to serious risks. “Watch out!”
• When hazard is low and outrage is high, the task is “outrage management” – reassuring excessively upset people about small risks. “Calm down.” (Outrage management may also be called for when high-hazard events are over and have entered the “blame phase.”)
• When hazard is high and outrage is also high, the task is “crisis communication” – helping appropriately upset people cope with serious risks. “We’ll get through this together.”
• When hazard and outrage are both intermediate, you’re in the “sweet spot” (hence the happy face) – dialoguing with interested people about a significant but not urgent risk. “And what do you think?”

How do my definitions map onto others’ uses of the terms “risk communication,” “crisis communication,” and “health education”?

Obviously, I have gone along with the broadened use of the term “risk communication” to apply to all kinds of communications about risk, not just reassuring ones. I use the term “outrage management” to mean what “risk communication” used to mean – reassuring people about small hazards. I have tried to imbue my approach to outrage management with respect for the audience and its reasons for overestimating the hazard. Good outrage management doesn’t mean telling people they’re foolish to be upset; it means listening to why they are upset and trying to change the things you’re doing that are making them upset. Still, the goal is to get them less upset – a goal that is honorable only if the hazard is actually small.

My use of “crisis communication” pretty much follows the non-PR convention: talking to people in situations that are crises for them. For PR people, any high-outrage situation is a crisis, because it threatens the client’s reputation. For me, a high-outrage situation is a crisis only if it is also high-hazard. It’s not enough for people to feel endangered; for the situation to be a crisis, they actually have to be endangered as well. But I use the term a little more broadly than most crisis communicators – to apply to individual as well as group risks, and to apply to chronic as well as urgent risks … any risk that is both upsetting and dangerous. In a recent consultation on autism, for example, I used the principles of crisis communication to explain how best to talk to the parents of a child who has been recently diagnosed as autistic.

I don’t use the term “health education” much. It’s all over my risk communication map:

• When the audience’s problem is mostly motivational, health education fits into what I call precaution advocacy – trying to get apathetic people more concerned about a serious risk. Example: talking to teenagers about the long-term health effects of obesity and why they should eat less and exercise more.
• When the audience is already motivated but lacks information, then health education is in my sweet spot – dialoguing with interested people about a significant risk. Example: explaining mosquito control techniques to a community that is trying to cope with its endemic malaria.
• When the risk isn’t just significant but very serious (and often urgent), and the audience knows it and is upset about it, health education turns into crisis communication – guiding appropriately upset people through a serious risk. Example: helping newly diagnosed AIDS patients master the complexities of their medication regime.
• And when the problem is that people are excessively worried about the recommended health precaution, health education requires outrage management – reassuring over-anxious people about a small risk. Example: persuading parents to vaccinate their children despite the possible side-effects of the vaccine.

name:	Julie Andrye This guestbook entry is categorized as:
field:	Volunteer activist
date:	September 28, 2008
location:	France

comment:

How essential is external communication for companies facing a crisis?

peter responds:

The answer to this question is so obvious I keep thinking it must be a trick question, or you must have some other meaning in mind. Of course external communication is essential for companies facing a crisis!

Two kinds of situations are often called a crisis.

The first is an actual crisis (high-hazard, high-outrage), when people are appropriately upset about a situation that is genuinely dangerous. In that situation, communicating with stakeholders who are upset and endangered is obviously essential. How else can you guide them through the crisis, validating their distress (outrage), helping them bear the situation and the way it makes them feel, and helping them choose wise rather than unwise precautions? See my Crisis Communication Index for more information on crisis communication.

The second kind of “crisis” occurs when people are upset about a situation they believe is dangerous – but you’re pretty sure they’re wrong. If you’re responsible for the situation, it may constitute a reputational and profitability crisis for your organization, and thus an economic crisis for shareholders, employees, and others who depend on your success for their income. But it’s not a real crisis for your external stakeholders. Communicating with people who are upset but not endangered is often called crisis communication by PR practitioners. I call it outrage management (low-hazard, high-outrage), and it’s covered in my Outrage Management Index. When people mistakenly think they’re endangered, communicating with them is the most important thing you do. How could you possibly calm their outrage and change the things you’re doing that are precipitating their outrage without talking (and listening!) to them?

So if “external communication” means communicating with external stakeholders who are upset, then clearly external communication is essential, regardless of whether your external stakeholders’ outrage is technically sound (crisis communication) or technically unjustified (outrage management).

In thinking about what your underlying question might be, I thought perhaps it was this: When some people are outraged by a situation, whether rightly or mistakenly, how essential is it to tell others who are neither outraged nor endangered what’s going on? When you’re busy trying to manage a crisis or quiet a controversy, in other words, should you or shouldn’t you clue in bystanders who are “external” to the situation? That’s a much tougher question.

My advice here is to go public.

There are exceptions. Sometimes the crisis or the controversy is extremely limited. Nobody else is affected; nobody else is aware; nobody else is likely to be affected or become aware later. Okay, keep your response within the family, work team, or neighborhood – within the circle of people who are affected or aware.

But my clients routinely imagine that the crises and controversies they are managing are smaller than they actually are. When you’re under attack, it’s rare to make a mountain out of a molehill. But it’s common to fail to notice that what you devoutly hope will remain a molehill is already growing into a mountain. And over-reacting to the situation isn’t only rarer than under-reacting; it’s also safer. In other words, when you try to confine your response to a crisis or controversy to the people you think are affected or aware, typically others become aware, consider themselves affected (or at least interested), and hold you accountable for not having told them sooner.

Telling the world about your small crisis

Suppose for example that you were managing Fonterra. This huge New Zealand dairy cooperative is 43% owner of a Chinese company named Sanlu. Sanlu distributed milk contaminated with melamine, sending thousands of infants and children, mostly Chinese, to the hospital with kidney stones. Sanlu reportedly knew about the melamine contamination as early as December 2007, and certainly by spring 2008. Fonterra didn’t find out until early August 2008 – and didn’t tell anyone for more than a month, when it tipped off the New Zealand government, which started pushing Beijing for a recall.

Fonterra executives say they kept quiet so long because they believed that staying “within the system” was the best way to persuade Sanlu executives and local government officials to get the contaminated milk off the shelves. They were thinking about nothing but the children, they insist.

Like most observers, I find that hard to believe. But I find it easy to believe that Fonterra management imagined the crisis was small and would remain small. Here’s the logic I suspect (and they deny) they were following:

• We stand the best chance of forcing a recall and protecting children if we blow the whistle publicly. But that will do substantial damage to Sanlu’s reputation, and thus to our Sanlu investment. It’s good for the children but bad for our shareholders (who have children too).
• If we try to manage the crisis quietly – encouraging a recall without raising a ruckus – we may very well succeed. The crisis may stay small, get under control, and go away. It might even be going away already; maybe only one batch was contaminated, and the harm has been done. Even if there’s more bad milk in the system that will hurt more children, our best bet is still to exert quiet pressure. That way we will be able to protect the children (though perhaps not as quickly as if we went public) and also protect our investment.
• If we try to manage the crisis quietly and fail, the crisis could expand. More children could be sickened, and the chances of a major scandal could increase – damaging not just our investment in Sanlu but also our corporate reputation and even our survival. But probably the crisis will stay small. Better to keep quiet.

I have no evidence that that’s how Fonterra figured. But it is certainly how many companies figure. They roll the dice, hoping they can manage a small crisis quietly and gambling that it won’t get bigger and become public knowledge. Quite often they turn out wrong. The crisis grows, the secret emerges, and their problems are multiplied. The original crisis ends up bigger than it would have been if the company had acted more quickly; more stakeholders are both endangered and outraged. And the company ends up facing a second reputational and profitability crisis over its failure to communicate honestly about the original crisis.

Telling the world about your small crisis is thus literally conservative. You forgo the possibility of solving your problem privately; you pay the price of widespread knowledge that you have a problem; in the process you insure against the far higher price of being discovered having let your problem grow while you kept it secret. I think just about everyone agrees that it would have been more ethical for Fonterra to blow the whistle on Sanlu. My point is that it would have been better business too.

I have to acknowledge a nagging possible exception. In the Fonterra/Sanlu case, and in most cases, going public about your crisis helps you manage the crisis; it has reputational costs to you but it’s obviously good for your stakeholders who are getting hurt. (And in the long term it’s good for you too.) But sometimes going public will exacerbate your crisis. In such a case, external communication threatens not just your reputation but also your stakeholders’ wellbeing.

This is a particularly vivid dilemma right now, as we watch our financial institutions implode. Experts point out that many of the companies going down the tubes this month could have weathered the storm if the people and institutions doing business with them hadn’t got wind of their problems, precipitating a more complicated version of the old-fashioned “run on the bank.” Economies are built on confidence – which is pretty much the same thing as saying finance is a confidence game. When confidence wanes, our fear that the house of cards will tumble becomes a self-fulfilling prophesy. I don’t understand economics nearly well enough to comment further. It is often said that transparency is essential to the stability and viability of modern economies. Obviously this is true in many ways; more transparency earlier about the risks of securitized mortgages and derivative instruments might well have prevented the crisis we’re facing now. But how much transparency should we want once the crisis starts to emerge? It’s looking pretty obvious right now that there are some ways transparency about a manageable economic problem can create unmanageable economic problems, undermining not just individual financial institutions but entire economies.

Having admitted a possible exception, let me stress that exceptions are rare. Most of the time, telling the world about your small crisis doesn’t make the crisis bigger. It does make the crisis most costly to your reputation in the short run; it feels like a bigger crisis when more people know about it. But transparency helps you manage the crisis, and transparency protects you from later accusations of neglecting and hiding the crisis. If you’re really, really confident that your crisis isn’t going to grow – that only a few people are affected, you’re helping them already, and nobody else knows or cares – fine, manage it quietly. But remember that you have a long string of predecessors who made that same judgment and turned out wrong.

Telling the world about your small controversy

Now let’s turn to the second kind of “crisis,” the low-hazard, high-outrage one. Nobody’s seriously endangered. But some people are really upset. You know what you need to do to mitigate their outrage: listen to their concerns and grievances, acknowledge the validity of some of what they’re saying, apologize for your errors and misbehaviors, change some things that are exacerbating their outrage, give them the credit they deserve for the changes they urged, set up better accountability mechanisms so they can watch you more closely, etc. Isn’t this just between you and your outraged stakeholders? Is there any need to tell anyone else?

In principle, no. In outrage management, I distinguish your stakeholders and publics according to how invested they are in the controversy, dividing them into “fanatics,” “attentives,” “browsers,” and “inattentives.” (See my column on “Stakeholders.”)

• The fanatics are outraged but probably unreachable, whether because they’re too outraged or for other reasons (ideology, for example – or maybe fighting you is their job).
• The attentives are also outraged, but their outrage is more ameliorable. A lot of outrage management is interacting with fanatics while attentives watch, in hopes of showing the attentives that you genuinely have taken the fanatics’ objections onboard and made major concessions, and now the controversy is getting boring.
• The browsers aren’t really outraged, just mildly interested, at least for now. They’re not paying a lot of attention to the controversy; unlike the fanatics and attentives, they’re not very skeptical about what they learn.
• The inattentives are sitting this one out.

If you had your druthers, you would confine your outrage management efforts to the fanatics and attentives. Since the browsers aren’t interested or skeptical, there’s no reason to tell them you’re sorry you did X and you admit Y is a problem. You have to say those things to the fanatics and attentives in your effort to manage their outrage, but surely there’s no need to clue in the browsers as well!

There are two problems with this analysis.

First, most of your dealings with fanatics and attentives are going to take place with journalists in the back of the room. The journalists will inevitably report what you say to the browsers. (The inattentives will ignore that story.) So there’s simply no way to do good outrage management with the fanatics and attentives without cluing in the browsers. If you manage the attentives’ outrage well, the browsers are going to learn about X and Y too; that’s unavoidable collateral damage. Your only alternative would be to let the attentives’ outrage fester for the sake of keeping the browsers ignorant about X and Y – and that’s far too high a price to pay.

Moreover, you’re not the only one communicating with the browsers. The fanatics have their own communication effort going, trying to convince the browsers to get more outraged – that is, trying to convert them into new attentives. X and Y are some of their best ammunition. So in most cases the question isn’t whether the browsers are going to find out about X and Y or not; the question is whether you want them to find out from you or from the fanatics. That question almost answers itself. Smart outrage managers “inoculate” the browsers against the fanatics’ best (i.e., true) arguments by conceding the validity of those arguments early and often (and apologetically).

For both of these reasons, “private” outrage management works only when the controversy is extremely small. You don’t usually need to tell your neighbors about your fight with your spouse … or your apology to your spouse. There are no browsers in a tiny controversy, just participants and inattentives. But if a controversy is big enough that there are fanatics who are deeply committed to your defeat, attentives who are following closely to see if you have learned your lesson, and browsers who are tracking the situation casually, then it’s too big to keep the browsers from knowing the embarrassing things you’re admitting to the fanatics and attentives. Journalists are bound to tell the browsers what you admitted. So are the fanatics themselves. Therefore, so should you.

name:	Knut I. Tønsberg
field:	Public relations for a government agency
date:	September 25, 2008
email:	kit (at) helsedirektoratet.no
location:	Norway

comment:

I have written a book in Norwegian about risk communication, making numerous references to www.psandman.com. Thank you very much!

I have also tried to find studies and empirical research supporting the seesaw principal. I found one study that presumably contradicts it – or does it? People were asked to imagine that they had lung cancer, and asked to choose between surgery and radiation. Some were presented with cumulative probabilities of “dying” rather than of “surviving.” When “dying” was used, the number choosing surgery dropped from 44% to 18% (McNeil, Pauker, Sox and Tversky, “On the Elicitation of Preferences for Alternative Therapies,” New England Journal of Medicine 1982, 306:1259–62).

Trying to find examples that contradict the seesaw principle or situations where talking about the worst case should be avoided could perhaps help us with more guidelines for the balancing act of “riding the seesaw.” Have you elaborated on situations when talking about the worst case should be avoided?

peter responds:

The study you cite is one of the classic studies of framing effects, a literature pioneered by Daniel Kahneman and Amos Tversky. In this particular study, patients, grad students, and physicians were all asked to think through how they might decide whether to have surgery for a hypothetical case of lung cancer. For all three groups, surgery was significantly more attractive when it was framed as offering a 68% chance of living for more than one year than when it was described as posing a 32% chance of dying within a year.

Framing effects are very important. The many framing studies that have been done teach us that equivalent ways of expressing a problem can have very different impacts on our choice of solution – one of many ways human information processing falls short of economic models of rationality. Thus a 68% chance of living feels more optimistic than a 32% chance of dying; an “escalation” feels more dangerous than a “surge”; etc. The huge impact of framing leaves communicators (and particularly risk communicators) only three choices:

• Ignore the data on which option the particular frame you pick is likely to encourage.
• Pick the frame that is likely to encourage the option you want to encourage.
• Use several counterbalancing frames in order to minimize any framing effects.

The third option is most objective and most respectful of the audience; the second option is most effective if persuasion is the goal; the first option is most common among doctors and other de facto risk communication practitioners, who have seldom mastered the principles of framing.

But I don’t see how framing is a counterexample to the risk communication seesaw.

The seesaw principle says that when people are ambivalent, they tend to emphasize the side of their ambivalence that is inadequately represented in the communication environment. Suppose we continued the framing study you cited by identifying respondents who were torn between surgery and radiation, and asking this subgroup to imagine a conversation with an opinionated friend. The seesaw principle predicts that if the friend urges surgery, all the reasons for sticking with radiation will come to mind, and respondents will become more inclined to pick radiation – whereas if the friend pushes hard for radiation, ambivalent respondents will rebound toward surgery.

Obviously, sometimes ambivalent people abandon their autonomy and just do whatever somebody else recommends; this is quite likely if a trusted doctor strongly urges surgery or radiation. But when we are trying to make up our own minds and have powerful inclinations in both directions, a one-sided argument often boomerangs.

The seesaw and framing are both powerful communication phenomena. The two are neither identical nor antithetical.

Like you, I have searched for empirical research – especially risk communication research – bearing on the seesaw concept. Like you, I haven’t found any. I referenced some of the non-empirical literature in a 2007 Guestbook entry on “Origins of the risk communication seesaw principle.” (That was in response to a comment from you! Your interest in the seesaw and your frustration at the absence of empirical data on it hasn’t diminished.)

Your comment also mentions worst case scenarios, and asks whether there are situations when it’s better not to tell people how awful things might get. One obvious example comes to mind: when your audience is fragile and very frightened already, and you’re worried that talking about the worst case might propel them into denial. I’m sure there are other exceptions as well. But they’re exceptions. The “rule” in pre-crisis communication is to split your attention about equally between the likeliest outcomes and the most alarming ones that aren’t vanishingly unlikely. As my column on worst case scenarios argues, that’s what you should do when you’re trying to alarm people, and it’s also what you should do when you’re trying to calm people. The last paragraph of the column reads:

And so we have come full circle. My advice to those who wish to warn us is to acknowledge how unlikely the worst case scenario is, even as they insist that it is too awful to bear. My advice to those who wish to reassure us is to acknowledge how awful the worst case scenario is, even as they insist that it is too unlikely to justify precautions. If both sides do good risk communication, they’re going to come out sounding a great deal more alike than they usually do today.

The seesaw goes a long way toward explaining why that advice is sound.

name:	Jonathan Waldron
field:	Dentist/former naval aviator
date:	September 22, 2008
email:	jcwdentist@hotmail.com
location:	Georgia, U.S.

comment:

I thought I would give you some feedback as your site wisely requests same.

While I was on active duty and in the Navy Reserve, one of our tenets was to avoid discussions of any sort about the “Big Three” – politics, religion and sex – in the course of business or even in the wardroom. The obvious reason: People have emotional feelings about all three that are rarely influenced by rational discussion.

A number of people I talked to about your work objected to your endorsement for Obama on your website. [See “Risk Communication Talking Points for Hillary Clinton: Some Primary Principles for This Post-Primary Moment.”] The objection was not so much about who your endorsement was for. People felt that it was not a good decision to make any endorsement in your business capacity (especially as most small business people are Republicans – primarily over the tax and regulation issues); also, it was felt that some readers might object for emotional reasons.

I have my own choice for President, but I am not telling any of my patients.

peter responds:

You’re right, of course, that my political preference is nobody’s business. Nor is my opinion on anything else, except on risk communication.

But occasionally when writing about the risk communication implications of something, I decide that I have to acknowledge some substantive opinion so people can assess my risk communication opinion in that context. So when Jody Lanard and I wrote a column on the risk communication challenge facing Hillary Clinton as she sought to endorse Obama (assuming she meant it), we felt we had to tell readers which candidate we preferred personally.

It wasn’t meant as an endorsement of Obama, but as an acknowledgment of possible bias as we analyzed what Clinton ought to do.

The problem comes up fairly often. When I write about the risk communication implications of the vaccination/autism controversy, for example, I feel obliged to say that I think getting vaccinated is safer than not getting vaccinated.

Of course politics is arguably different. Or maybe not; many anti-vaccination activists feel far more strongly about that than about any political candidate ever.

I do appreciate your feedback. I will think further about the right balance between acknowledging private opinions that might influence my risk communication judgments and letting private opinions intrude on my risk communication articles.

name:	Knut I. Tønsberg
field:	Public relations for a government agency
date:	August 4, 2008
email:	kit(at)shdir.no
location:	Norway

comment:

In the 6th point of the “Responding to Rumors” section of your column on “Rumors – Information Is the Antidote,” you write: “Good science is always tentative, and so is good risk communication.”

Risk communication is tentative – that’s understandable and one of your basic thoughts as I have understood. Could you please elaborate a little on why “good science” also is tentative? Most people would perhaps reply that good science gives us facts, principles to hold onto, that science is not tentative but universal.

peter responds:

Science is indeed the pursuit of “universal facts and principles.” But scientific findings are always a rough draft of those universal facts and principles, subject to amendment as new data identifies errors or exceptions, different facts or better principles.

Non-scientists tend to miss this important truth, seeing science as definitive rather than tentative. That’s partly our hunger for certainty at work. But it’s mostly the fault of scientists, who make two fundamental risk communication mistakes again and again:

• Scientists try to augment the authority and prestige of the scientific enterprise by implying that science builds on a firm foundation of prior science. They know better. In fact, science builds on parts of the foundation; it keeps tearing down other parts to rebuild them better.
• Scientists try to augment the authority and prestige of particular scientific claims by implying that those claims are firm. Once again, they know better. When they write for their peers, they are at pains to address the competing claims of other scientists. But when talking to the public, they all too often ignore those claims or disparage them as not scientific at all. The more controversial the claim, unfortunately, the likelier a scientist is to commit this sin on behalf of his or her side of the controversy.

Falsifiability is the fundamental premise of all science. If there is no possibility that you could be proved wrong by new data, you’re in the realm of faith, not science. If there is such a possibility, your findings are tentative by definition. (This is an oversimplification of Karl Popper’s notion of falsifiability, but I think it captures the essence of his point.)

Of course some scientific claims have stood firm for centuries, and are likely to stand firm forevermore. Likely – but not guaranteed. Scientific claims that stood firm for centuries occasionally crumble; consider what Einstein did to Newtonian physics.

Many scientific claims have never stood firm. Rather, they are endlessly under attack. Even claims that a majority of scientists consider well-established frequently coexist with a competing viewpoint advanced by a scientific minority. If 90% of the evidence supports one position and only 10% supports the other, most scientists are likely to go with the 90% – and a smart layperson should too. But in science as in horse-racing, sometimes the long shot comes from behind and wins. And depending on the cost of erring in each direction, sometimes it makes sense to hedge your bets – for example, to take precautions against a horrific risk even though most (but not all) experts believe it can’t happen.

As a practical matter, moreover, new science is extremely tentative. Different scientists often advance competing and mutually incompatible claims simultaneously, all of them grounded in data. Further science must determine which of these claims are valid and which are false, or must find a new interpretation that integrates them and shows that they’re not mutually incompatible after all. Sometimes that further science turns up in a matter of weeks or months. Other times that further science relies on methodological or conceptual breakthroughs that haven’t yet occurred; we can (and often do) wait for decades without a good basis for choosing among competing scientific claims.

Even if there are no competing claims, a new scientific claim has “stood firm” for minutes, not for centuries. At the very moment it is published a half-dozen other scientists may be hard at work on relevant studies that can confirm or disconfirm the new study – or, most likely, complicate its interpretation without quite confirming or disconfirming it.

Wise scientists therefore take a new finding with more than a grain of salt, waiting to see whether additional studies will emerge that support or rebut it. Wise laypeople should do the same.

And good risk communicators should help them do so – which is why communicating about uncertainty is so important. (See my 2004 column on “Acknowledging Uncertainty.”)

The uncertainty of scientific findings and the inevitability of scientific controversies don’t have to lead to paralysis. In fact, once you understand that all science is uncertain, it becomes clear that uncertainty isn’t an acceptable reason (or excuse) for inaction.

In coping with scientific uncertainty, the layperson has a fourfold job:

• Figure out what the weight of the evidence says. Or if that’s too difficult, figure out what the majority of the experts think – bearing in mind that their judgments may be biased or simply mistaken.
• Figure out how much residual uncertainty there is. Science is always uncertain – but there’s a real difference between the theoretical uncertainty of “all claims are falsifiable” and the practical uncertainty of “this claim is brand new and hotly debated.”
• Figure out what the cost of being wrong in each direction is likely to be. Often (but not always) over-caution about a risk does less damage than recklessness. When the disparities in outcome are big and the choice is uncertain, it’s not stupid to choose the option that won’t kill you if you’re wrong.
• Figure out whether you can afford to wait for new evidence that may reduce (or increase) the uncertainty. If new evidence is expected soon and the situation isn’t urgent, fine, wait. But when the time comes to act, take your best shot despite the uncertainty. And then keep looking for new evidence; you may want to change your mind or reverse your course.

The risk communicator’s job is to help with these four tasks – or at least not to make them more difficult by implying that science is certain or that action should wait for certainty.

As if all this weren’t difficult enough, I need to add one final factor. People’s decisions are rightly and inevitably grounded in more than just scientific evidence. Values play a role. So does outrage. So do a variety of other aspects of the situation. (See my January 2008 column on “Who’s Irrational: When People ‘Ignore’ Risk Data.”) Risk communicators not only have to help people cope with scientific uncertainty. We have to help them integrate what they know about the science with the rest of what they know, think, and feel.

name:	Robbie Lamons
field:	Volunteer activist, geologist
date:	July 14, 2008
email:	robbielamons@yahoo.com
location:	California, U.S.

comment:

My neighborhood is an extremely charming old mercury mining town, New Almaden. The EPA, the California Regional Water Quality Board, and several other organizations are naturally worried about mercury pollution coming from our town, and the park behind it where most of the mines were located.

Several years ago the County Parks Department dug up, dumped and buried tons of dirt that had more mercury (in several different chemical forms) than was judged to be safe. The main fear seemed to be that park visitors would sue the county if they went hiking, claiming their medical problems were caused by their visit to the park. People in this town haven’t shown any clusters of diseases, so they weren’t too worried. The “cleanup” was pretty messy. It left scars in the park and distributed mercury in the town. Also, there was no subsequent reduction in methylmercury in fish.

Methylmercury is the really toxic form of mercury that resulted in deaths around Minimata Bay in Japan. Methylmercury is produced in nature by anoxic bacteria in stagnant bottom waters and soils of reservoirs, lakes and bays.

The three reservoirs around Almaden were found to have fish with high levels of methylmercury. A diverse group of people – water district employees, fish and game people, a geologist, and others – was convened to decide on an action plan called a Total Maximum Daily Load plan, or TMDL.

No one wanted to take responsibility for writing up the TMDL, so the state water board staff volunteered to do it. When the TMDL came out, it required that the suspended sediments have less than half as much mercury (of all chemical species) as the background level in non-mining areas, and far less than the EPA standard for methylmercury in fish. It also required that property owners along the creek test their property for mercury, and for erodability. This requirement caused rather a stir among the few people who were aware of it, especially the two who had property along the creek.

Tonight, we have a meeting with the principal author of the TMDL. We want to convince her to table the plan in the light of new evidence that the water circulators installed almost two years ago have reduced the amount of methylmercury in one reservoir by 95%. The fish have not yet been tested to see if they also show a reduction in methylmercury. If they do, the TMDL plan should be rewritten to require dissolved oxygen circulators in all of the stagnant areas of our watershed, and hold up on the far more expensive work of moving and burying dirt.

Tonight I will try to make sure everyone’s opinions are heard, and will make my own as clear as possible – or as my brother says, “Don’t sound like a wacko.”

peter responds:

I just picked up your comment – and it sounds like your TMDL meeting is probably over by now. Are you still interested in a response from me?

I found the mercury debate in New Almaden (as you described it) fascinating. If you decide to post an updated riskcomm-related comment/question on this situation – one that isn’t so time-sensitive – I would be delighted to respond.

And even if you decide not to, I’d love to know how the meeting went!

robbie lamons responds:

Our meeting did not go well, but we learned some things about the other side, and they learned some things about us.

They had not come to the meeting to receive comments, or to discuss changes to the document, because (they told us) as bureaucrats, they cannot accept comments or suggestions after the comment period has ended. So no negotiations could take place. They came only to try to get the community to accept their action plan. The only hope that property owners have is that many of their deeds say that they don’t own the mineral rights (and we hope they don’t own the responsibilities either).

The engineer who wrote the document had a pie-in-the-sky idea of how the community could get the U.S. government to pay for the cleanup. She did not understand that much of the huge expense of cleaning up the sediments may not be necessary because circulating dissolved oxygen using solar-powered circulators or river power has taken 95% of the methylmercury out of the water. (Methylmercury in fish is the entire reason for the cleanup.) Her own graph shows that methylmercury levels in fish are high in reservoirs and low in the streams above and below them. The total mercury bedload is highest in the stream below the reservoir.

She said to me, “You and I are going to disagree forever.” Well, that is true. She was there with her new boss, a geologist. Part way into the meeting she said to him, “I tried running this meeting.” Then she said to one of the community members, “And now you’re raising your hand. I give up.”

Her boss apologized to her! He said they would not delay the approval of the plan so that fish could be tested in the reservoirs that have had lower methylmercury levels for one and a half years. It almost seems as if the fish haven’t been tested on purpose.

We all left the meeting frustrated and nervous. The geologist thanked me for shaking his hand! Why was that? Did he feel guilty?

Our next negotiating strategy choices seem to be limited. We could notify the media, get the fish tested, talk to our legislators, talk to the board members who have to approve the plan, go over their head to the California and national EPAs, or wait for it to pass and then sue them.

Have you got any ideas for changing the path of a bureaucracy?

peter responds:

Your situation in a nutshell: The bureaucrats are convinced that your community’s mercury problem is a serious hazard. You’re pretty sure the mercury hazard is low and controllable with much less costly measures than the ones they’re proposing. You believe their remedy poses a higher hazard (especially a financial hazard) to the community than the mercury problem itself.

I’m not qualified to judge which of you is right about the mercury hazard. It’s not rare for communities to be insufficiently concerned about a genuinely serious risk, especially if the remedy poses a financial threat. Nor is it rare for government agencies to overreact to a small risk, especially if they’re under pressure or excited about the chance to write new regs. So there’s a long history of controversies where outside bureaucrats want tougher environmental protection than the locals.

Of course there is also a long history of controversies where communities are convinced some contaminant endangers everybody’s health, while regulators keep trying to tell them the risk is minimal. That pattern is a lot more frequent than the one you’re experiencing. But neither is rare.

What’s most interesting from a risk communication perspective is the way outrage is playing out in this controversy. Whether or not your mercury problem is technically serious, it’s easy to see some of the reasons why the problem is low-outrage for most community people. Most importantly, it’s familiar – part of the comfortable and perhaps even revered history of your “extremely charming old mercury mining town” (to use the words of your first comment). And the last time bureaucrats tried to do something about it, they scarred your park and spread mercury through your town. So you’re disposed to experience the remedy as a bigger outrage than the problem … and thus to suppose that the remedy is also a bigger hazard than the problem.

Judging from your description of your recent meeting, the plan’s sponsors are acting in ways that are pretty much guaranteed to exacerbate your outrage at them, and your conviction that they’re determined to disrupt your community with arbitrary overregulation. Their disinclination to listen to your concerns, their lack of interest in your data on the effectiveness of dissolved oxygen, and their unwillingness to assess methylmercury levels in the fish were all examples of very poor outrage management.

I also get the impression from your second comment that they were experiencing their own high outrage – at you – which obviously made it much harder for them to do a decent job of addressing your outrage at them.

Despite your description of your strategic choices as “limited,” your list of possible next steps looks pretty impressive to me. It might help to organize your options into three categories:

• Strategies aimed at arousing more local support for your campaign against the plan – greater use of the media, for example. Your goal here would be to increase your neighbors’ outrage (anger) at bureaucratic arbitrariness and intransigence, as well as your neighbors’ outrage (fear/concern) about the damage an unreasonable and unnecessary mercury cleanup could do to your community.
• Strategies aimed at overturning the new plan by going over the heads of its sponsors – to the state water board honchos, to the legislature, to the state and federal EPAs, to the courts. Most of the strategies on your list so far are in this category. This approach may or may not be compatible with the first approach, arousing more local outrage – it depends on whether you have a better shot at getting the plan reversed through public pressure or through quiet diplomacy.
• Strategies aimed at changing the minds of the plan’s sponsors. It may be too late for this; the die may be cast as far as the plan’s sponsors are concerned, either because they have committed themselves legally or because they have backed themselves into an emotional corner. But you ought to consider whether there might still be time to try to reduce their outrage and seek reconsideration of the plan at their level. Note that this approach is incompatible with the other two; arousing stronger community opposition to the plan and going over the heads of its sponsors will almost certainly solidify their unwillingness to reconsider.

One possible step toward reducing the outrage both sides are feeling would be to offer to conduct a collaborative study to measure methylmercury in the fish. If you do that, I would advise negotiating the implications of the study results before you launch the study. Agree in advance what findings would justify changing the cleanup plan, and what findings would justify reaffirming it.

name:	Marcus
field:	Environmentalist, Division of Solid and Hazardous Materials, New York State Department of Environmental Conservation
date:	July 10, 2008
location:	New York, U.S.

comment:

I read your paper on “Risk Communication: Evolution and Revolution.”

In the Appendix, the EPA’s “Seven Cardinal Rules of Risk Communication,” a portion of #7 reads: “Risk communication will be successful only if carefully planned and evaluated.”

My question to you is this: In evaluating risk communication, does one give a test to the audience? If yes, does one give a pretest, and another test at the end of the presentation/project?

peter responds:

Two distinctions are crucial in thinking about evaluating any sort of communication, including risk communication.

• The distinction between methodologically rigorous evaluation research and seat-of-the-pants, cut-the-corners-you-have-to-cut pragmatic evaluation.
• The distinction between evaluating pieces of a communication campaign you’re still drafting (“formative evaluation”) or still conducting (“process evaluation”) and evaluating the whole campaign after it’s over (“outcome evaluation”).

Formative evaluations and process evaluations almost always cut methodological corners. Outcome evaluations usually do too. There is rarely enough budget to do a rigorous evaluation.

I am a big, big believer in “sloppy” formative and process evaluation, rather than no evaluation at all. Early in the design of a campaign, it’s important to talk about your preliminary plans with people who are representative of your target audiences. A bit later, when you have some draft messages, it’s important to try them out – once again on groups typical of the people you’re trying to reach. And later still when the campaign is ongoing, it’s important to do some testing to guide your mid-course corrections.

Formative and process research typically makes use of focus groups – a methodology purists rightly insist is appropriate for generating hypotheses but not really for testing them. The samples are “samples of convenience” – that is, people it was easy to round up and talk to – rather than random samples of the populations you’re targeting, so statistical generalizations about those populations are unjustified.

Nonetheless, the difference between sloppy evaluation and no evaluation at all is bigger, in my judgment, than the difference between sloppy evaluation and rigorous evaluation. Even in a crisis, when there’s obviously no time (and probably no budget) for rigorous research, you can still take a few hours to test your draft messages on whoever’s handy and revise based on what you learn. And once you have started rolling out your campaign, you’d be crazy not to make some effort to see what’s working and what’s backfiring, and then adjust accordingly.

For more on formative and process evaluation that doesn’t necessarily dot all the methodological i’s and cross all the methodological t’s, see the “Evaluation and Coaching” summary my wife and colleague Jody Lanard prepared as part of the development of the World Health Organization’s outbreak communication guidelines.

Although formative and process evaluations almost always need to cut some methodological corners, try hard not to cut the corners that matter most. In message testing, for example, I think it’s crucial to test what you really want to know. Asking people how much they liked a message or measuring how much information they learned from it is a very poor stand-in for whether the message made them more inclined to take a specific action (quit smoking, say, or get their flu shot). Similarly, message testing often needs to be done in the context of other messages likely to be emanating from other sources.

Both of these issues come up again and again in risk communication message testing. Typically, I will urge clients to acknowledge some information they’d prefer not to mention (prior errors, current uncertainties, the seriousness of the risk, etc.). A focus group test of my recommended message against the client’s preferred message will very likely find that people would rather not hear disquieting information. For example, people may well “like” over-reassuring messages more than appropriately alarming ones. But over-reassuring messages don’t “work” if the goal of the campaign is to motivate people to take precautions. And when over-reassuring messages are tested in the context of alarming information from other sources (embedding the test messages in a mock radio newscast, for example), people’s preference for over-reassurance dissipates, and its devastating impact on trust and credibility becomes clear.

I am focusing on formative and process evaluation because I believe they are so important. Your question, however, focused on post-campaign outcome evaluation.

Here methodologically sophisticated studies may be somewhat more feasible. But only somewhat. To evaluate a campaign rigorously, you obviously need a control group. It’s usually possible to randomly assign target communities to two groups: the one that receives your campaign messages and the ones you skip. When the campaign is over, if the targeted communities differ systematically from the ones you skipped in ways relevant to your campaign, that’s pretty solid evidence that the campaign was responsible for the differences. Similarly, you can develop three or four different campaigns, randomly decide which communities get which ones, and test the differential effects.

But this means lots of communities aren’t getting your best shot. They’re getting the “alternative” campaign you’re trying to prove works less well, or they’re getting no campaign at all. And it means a big share of your budget is going into two peripheral tasks: putting out suboptimal campaigns, and testing to see if they really worked less well than your best shot.

The long-term payoff of careful evaluation research is immensely worthwhile. (See my Precaution Advocacy Index for some of the evaluation research Neil Weinstein and I have done (with colleagues) on radon communication campaigns.) If your goal is figuring out how to do more effective risk communication about smoking or HIV or radon, rigorous evaluation research is the way to go.

But if you think you already have a handle on how to do effective risk communication, and your goal is actually warning people about smoking or HIV or radon, then evaluation isn’t your core task. You’ll probably end up doing less-than-rigorous evaluation so you can focus more of your efforts on getting the word out.

Once you decide to do a rigorous study (if you do), one of many questions you will face is the one you mentioned: whether to do a pretest or not. As you probably know, there are pros and cons to a pretest-posttest study, as opposed to posttest-only. Posttest-only relies on randomization to assure that the groups started out equivalent. Pretest-posttest measures how they started out, which makes it easier to show a change as a result of your campaign. But the pretest sensitizes people to the campaign, and thus might alter its effect. The methodological gold standard is the so-called “Solomon four-group design,” which merges the two – but it’s rare to be able to afford it.

Some of the issues that arise when testing risk communications in particular are discussed in a 1993 article I wrote with Neil Weinstein, “Some Criteria for Evaluating Risk Messages,” Risk Analysis,13:1, pp. 103–114. It’s fairly heavy reading (I wouldn’t have attempted it without Neil), and it’s not on this website.

name:	Bhuvaneswari Jayaraman This guestbook entry is categorized as:
field:	Student
date:	July 4, 2008
location:	California, U.S.

comment:

I am a student of health communication and am currently working on a report on the avian influenza outbreak in India.

I was wondering if people can be prone, with frequent episodes of avian influenza, to becoming apathetic in response to warnings. The heuristic would be something like this: The authorities and the media warned about the deadly outbreak; we were advised not to eat chicken and egg; and nothing happened – there were no human casualties. [Therefore there is nothing to worry about.]

Is this scenario likely? What precautionary measures can risk communicators take to prevent such a scenario?

peter responds:

People notice when warnings don’t come true, and if warnings don’t come true repeatedly people sometimes stop taking the warnings seriously. This phenomenon is in the popular culture (in the U.S., anyway) as “the boy who cried wolf.” It is in the research literature as “warning fatigue.”

Warning Fatigue

What’s most noticeable about warning fatigue, however, is how weak it is. When weather forecasters warn that a hurricane is coming, most people in the predicted path prepare; when the hurricane changes course, most people are relieved; the next time there’s a hurricane warning, most people prepare again. Similarly, activists have long known that they’re pretty safe warning that a particular industrial facility is likely to explode or its emissions are likely to increase the cancer rate; if the explosion doesn’t happen and the cancer rate stays stable, the activist group simply moves on the next issue, undeterred and undamaged.

There’s a good reason why warning fatigue is weak. People intuitively understand that a false alarm is a lot smaller problem than a disaster they weren’t warned about. We understand that it’s a minor irritation if a smoke alarm goes off when there’s no fire, but a catastrophe and a scandal if there’s a fire and no alarm. So we calibrate smoke alarms to be oversensitive; we tolerate their going off too much in order to be fairly confident that they won’t miss a fire. We “calibrate” activists and weather forecasters to be similarly conservative in their warnings – that is, to err on the alarming side.

Still, warning fatigue does happen. And warnings have other costs too; picture a major airport shut down for hours because of a bomb threat. The authorities shouldn’t warn us any more often or more urgently than necessary in order to protect us.

And when they warn us, they should be candid that the warning may turn out unnecessary. The hurricane may change course. The factory may not explode. The bird flu outbreak may be quickly contained without any human fatalities. In the last section of my column on “Worst Case Scenarios,” I make the case that warnings that dramatize the magnitude of a risk are less vulnerable to warning fatigue than those that overstate its probability. An insurance salesperson who reminds you how awful it would be if your house burned down is likely to get more renewals than one who keeps insisting your house will probably burn down.

All Warnings Begin and End in Apathy

Warning fatigue is all about the danger that the authorities will sound the alarm, nothing will happen, and people will respond to the false alarm by becoming resistant to future warnings and apathetic about the risk.

Okay, but suppose the authorities sound the alarm and then the thing they warned against does happen. People are upset – grateful to have been warned, but unhappy to have a new risk to cope with. And then things get better. The crisis has passed. For a while, people are on guard, worried that it might happen again and interested in figuring out what precautions and preparations are appropriate. Then newer problems arise and memories start to fade. Pretty soon people are apathetic again – not as apathetic as they were in the first place, but a lot more apathetic than they were at the height of the crisis, or even at the height of the pre-crisis warnings. The “New Normal” includes the risk of another such crisis … which won’t feel quite as much like a crisis as the first one did.

Or suppose it’s a chronic risk rather than an acute crisis. It comes and stays; it becomes endemic. For a while, people are focused on trying to get rid of it. Once they start to realize they can’t, they pay some serious attention to learning to live with it. And then they get used to it. Once again, apathy reasserts itself. The “New Normal” routinizes the new risk, which doesn’t feel so risky anymore.

In short, all warnings begin and end in apathy. If the bad thing doesn’t happen, people get apathetic again. If it happens and ends, people get apathetic again. If it settles in forever, people get apathetic again. Apathy is the default position. What varies is how much people learn, and put in the back of their minds, during the warning phase and the crisis phase – before they become apathetic again.

An effective warning rouses people out of their apathy for a while. They go through an “Oh my God!” moment, an adjustment reaction. During the adjustment reaction, people rehearse for the crisis to come – cognitively, emotionally, logistically, and socially:

• They pay a lot more attention to the new risk, watching for signs that it’s approaching.
• They have an emotional reaction – mostly concern/fear about the risk itself, but sometimes mixed with irritation at the authorities for burdening them with a new problem.
• They learn about the risk and the recommended precautions, decide how best to prepare, and start preparing.
• They may communicate their own suggestions – at least to family and friends, and maybe even to the authorities.
• They may take some premature precautions, as if imagining that a future crisis were already happening.

The adjustment reaction to warnings doesn’t last long. If a little time passes and nothing bad has happened yet, most people stop worrying. They put the problem out of their minds. They relax their vigilance. They forget some (but not all) of what they learned, and abandon some (but not all) of their preparations. And yes, some people scoff at the original warnings, complaining that the authorities got people all upset for nothing.

But post-warning apathy is less apathetic than pre-warning apathy. If the predicted crisis eventually arrives, people will find it less shocking than if they had never been warned, and will gear up more quickly and more effectively to cope with it.

And then after it is over, or after they’re used to it, they’ll go back to the default position: apathy. That’s why it’s so important to use a crisis as a “teachable moment.” There is a narrow window of opportunity to debrief the lessons learned and lock in commitments to prepare better for next time before everyone (including the authorities) loses interest.

Over-Reassurance: Warning Fatigue’s Evil Twin

So if you warn people, you stave off apathy for a while, and then it returns (though not completely) – whether what you warned them about happens or not.

What if you don’t warn people? What if you warn them insufficiently? What are the effects of official over-reassurance, over-optimism, and failure to warn?

I have written about this pretty often before. See particularly “Tsunami Risk Communication: Warnings and the Myth of Panic,” written with my wife and colleague Jody Lanard.

In a nutshell, here’s what happens when officials over-reassure:

• Before we know whether things are going to get bad or not, people who are unaware of the risk or who are very trustful of the authorities are reassured by official over-reassurances. They stay calm (presumably the goal), but too calm: They fail to get ready. They don’t have an adjustment reaction, don’t rehearse emotionally, don’t prepare logistically.
• People who are more aware of the risk and more inclined to be mistrustful have a seesaw reaction to official over-reassurances. Even though nothing bad has happened yet, they smell a rat. So they become all the more concerned – in a particularly undesirable way. They feel abandoned by the authorities, left alone with their fears, misled and patronized. So they are likely to overreact (because the government is under-reacting), and likely to think things are worse than they are. Paradoxically, they might even panic (though panic is rare even in the face of official over-reassurances).
• If the situation turns out benign in the end, officials will probably get away with their failure to warn. Those who bought into the official over-reassurances and took no precautions feel they were well-led. Those who smelled a rat rightly feel the officials just got lucky this time – but they’re unlikely to be able to generate much outrage at official misbehavior that turned out okay. Officials, of course, feel all the more encouraged to over-reassure again the next time some risk arises.
• If the situation deteriorates and the officials’ over-optimism is proved wrong, on the other hand, mistrust runs rampant. Those who suspected as much all along are confirmed in their judgment that they’re on their own. Those who swallowed the over-reassuring official line must cope not only with their own unpreparedness for the situation at hand but also with their feelings of betrayal by the authorities. Just when people most need official guidance on how to handle the crisis, officials have forfeited their credibility and thus much of their ability to lead the public.
• Over the long haul, a government that has been routinely over-reassuring in the past faces a new risk without essential tools. If the new situation is genuinely not very serious, they find themselves unable to be convincingly reassuring. (After all, they sound the same when something is serious as when it is not.) By under-warning about serious risks in the past, they have set their people up to overreact to small risks now. And if they now try to warn people about a serious risk, they face confusion and incredulity. What does it mean when a normally over-reassuring government suddenly says something alarming? Is the situation so dire that officials finally had to admit there’s a problem? Or do they have ulterior motives for being falsely alarming now, just as they were falsely reassuring in the past?

Official over-reassurance leading to mistrust is a bigger problem than official over-alarm leading to warning fatigue. A much, much bigger problem. It’s bigger in general. It’s bigger in communications about bird flu. And it’s bigger in India, whose government (like most governments) has a long history of being publicly over-reassuring about many risks.

Bird Flu Warnings

Consider the announcement of India’s first known outbreak of the H5N1 bird flu strain, in Maharashtra in 2006. This isn’t a horrible example, just a typical one.

The first announcement was, by definition, delayed. The government announced confirmation of H5N1 in birds. That means that several days earlier, the government had to have known (and not said) that there was some deadly bird flu strain killing local poultry. A couple of days after that, they had to have known that it was H5 (because it takes less time to identify the “H” than the “N”). Only after the lab tests were complete for both the hemagglutanin and the neuraminidase did the authorities announce the presence of bird flu.

This means they missed the chance, for several days, to warn poultry farmers to ramp up their biosecurity precautions in order to reduce the risk of spread. It also means consumers were buying potentially diseased birds without having been warned.

On the other hand, once they made the announcement the Indian authorities did not understate the seriousness of the outbreak. One early Reuters story (February 18, 2006) quoted local animal husbandry minister Anees Ahmed as saying, “We are treating it as an emergency.”

This is much better than India’s risk communication during Delhi’s 2005 human meningococcal disease outbreak. As that situation got progressively worse, officials repeatedly insisted it was under control, was just “a few sporadic cases,” could be treated easily with antibiotics, didn’t necessitate a vaccination program, etc.

Avian influenza poses at least three quite separate risks:

• The risk to farmers that their poultry will be infected. A bird flu outbreak is an economic disaster not just for farmers whose flocks get sick and die, but also for nearby farmers whose flocks need to be culled in an effort to stop the spread of the virus. There are serious follow-on effects on the whole local economy, and on the diets of poor people who rely on chicken as a major source of protein.
• The risk that someone will come into contact with a sick bird and contract the disease. This is a low-probability risk even for farmers and their families. Avian influenza passes from birds to humans only with considerable difficulty; with millions of opportunities, it has successfully done the trick only a few hundred times so far. The probability of bird-to-human transmission is lower still for people who don’t spend a lot of time with birds. Still, contact with the carcass of a diseased chicken can be sufficient, so the cook faces a small but real risk. Contact with an incompletely cooked chicken dinner might be sufficient as well, so the rest of the family faces a tiny but still conceivable risk. No matter how small the probability that a particular person will catch bird flu, the magnitude of the risk is huge; roughly half the diagnosed cases have died.
• The risk that a bird flu virus that has never before circulated among humans will mutate or reassort in a way that makes human-to-human transmission efficient. Each time this happens – which is roughly three times a century – it causes a pandemic and kills millions of people. The prospect that it could happen to the H5N1 bird flu virus is the risk public health experts are focused on – a risk of huge magnitude and unknown probability. Controlling bird flu in birds is a good way to reduce the probability of this disaster. But if it happens, people won’t have to worry about contact with birds anymore; they’ll be worried about contact with their neighbors.

So every government – the Indian government included – has to make decisions about three warnings, not one:

• how urgently to warn farmers that their flocks might get the disease;
• how urgently to warn various subgroups that they might get it themselves from contact with diseased birds; and
• how urgently to warn all of us that a pandemic might devastate the entire world.

Most experts agree that the first and third risks are a lot more serious than the second one. That is, H5N1 poses a huge risk right now to poultry, a huge risk maybe someday to us all … and a small risk right now to people who come into contact with diseased birds.

In each case, excessive warnings might lead to warning fatigue and to apathy: “You told us it was coming and nothing happened!”

And in each case insufficient warnings might lead to a host of problems: apathy and lack of preparedness among those who feel more reassured than the situation justifies; mistrust among those who believe the situation is worse than the reassurances imply; outrage on everyone’s part if things get bad and we demand to be told why we weren’t properly warned.

The risk communication goal is to communicate all three risks in ways that are proportionate to the actual hazard. Tougher still, the goal is to communicate all three risks in ways that feel proportionate to the hazard – that feel neither excessively alarming nor excessively reassuring … not just now, but also in hindsight after bad things happen or after they don’t.

It’s not easy to walk this risk communication tightrope. For most officials who are trying to walk it, excessive warnings aren’t the main problem. Insufficient warnings are.

[Note: My wife and colleague Jody Lanard collaborated on this response.]

name:	Ahmad Rasoul Mofleh This guestbook entry is categorized as:
field:	Medical doctor
date:	July 3, 2008
location:	Kabul, Afghanistan

comment:

I am a medical doctor working in Afghanistan. Next week we will conduct a communication workshop for some of our emergency response teams, focusing on avian influenza, and I am expected to lecture on both risk communication and outbreak communication.

Could you please answer this question: What is the difference between risk communication and outbreak communication?

peter responds:

There are really three questions here: the origins and development of “risk communication”; its relationship to “outbreak communication”; and the role of “avian influenza communication.” Let me take them one at a time.

The Origins and Development of “Risk Communication”

The term “risk communication” was introduced in the U.S. in the 1980s. Its first use was with regard to environmental controversies where the actual technical risk was fairly small but the affected public was very upset. A typical example would be a polluting factory; even though most experts may judge that a particular factory’s emissions are unlikely to cause significant health effects, the factory’s neighbors may be understandably skeptical about this judgment and fearful that the emissions could end up giving their children cancer.

The first conference with “risk communication” in its title (as far as I know) was held in 1986, sponsored by the U.S. Environmental Protection Agency and others. It focused on trying to figure out how best to reassure people who are excessively upset about small environmental hazards. The conference organizers understood from the outset that telling people to “calm down” wasn’t going to do the trick – that effective risk communication would need to be respectful rather than patronizing, and two-way rather than one-way. But we had a lot to learn about how to foster the sort of mutually respectful dialogue that might actually be experienced as reassuring.

While we were busy creating this new field of risk communication, there was already a well-established field of “health communication,” which generally made the opposite assumption: that the audience was insufficiently upset about serious hazards. Health communicators aimed to persuade apathetic people to eat less, exercise more, quit smoking, etc. A closely allied field that also existed already was “safety communication”; like health communicators, safety communicators tried to arouse more concern and motivate more precaution-taking, focusing on driving, crime, workplace accidents, and similar threats.

And there was already another well-established field of “emergency communication” that focused on yet a third problem: how to guide appropriately upset people through serious situations – natural disasters, epidemics, wars, etc.

In my vocabulary, health communication and safety communication focused on high-hazard, low-outrage risks. Emergency communication focused on high-hazard, high-outrage risks. Risk communication initially evolved to address low-hazard, high-outrage risks.

But over the next two decades the term “risk communication” expanded to encompass all three tasks – alerting insufficiently upset people to serious risks, guiding appropriately upset people through serious risks, and calming excessively upset people about small risks. Today, even a technical expert who simply wants to educate the public about risk data is likely to be described as doing risk communication.

This expansion of the meaning of “risk communication” is regrettable, I think, because it obscures the huge differences in goals and methods among these various tasks. But it happened, and it’s probably irreversible.

So these days I like to talk about three principal “risk communication paradigms”:

• When hazard is high and outrage is low, the task is “precaution advocacy” – alerting insufficiently upset people to serious risks. “Watch out!”
• When hazard is high and outrage is also high, the task is “crisis communication” – helping appropriately upset people cope with serious risks. “We'll get through this together.”
• When hazard is low and outrage is high, the task is “outrage management” – reassuring excessively upset people about small risks. “Calm down.”

This terminology isn’t universal (though I keep trying). What I call precaution advocacy is sometimes called social marketing or social mobilization. What I call outrage management is sometimes called crisis communication. The proliferation of labels is by no means over.

“Risk Communication” versus “Outbreak Communication”

So where does “outbreak communication” fit in? This term was coined just a few years ago by the World Health Organization, when it was developing its first set of “Outbreak Communication Guidelines.” ( 453kB, off site)   In part, the new label was simply a branding exercise. WHO packaged together those aspects of risk communication that were paramount in coping with infectious disease outbreaks – informing early, acknowledging uncertainty, warning people that the situation was likely to change and early information might turn out wrong, etc.

Most or all of the outbreak communication guidelines are identical to crisis communication best practices. Nonetheless, “outbreak communication” is turning out to be a useful term to describe the way crisis communication principles ought to be applied under outbreak conditions. Infectious disease outbreaks are arguably different from other kinds of emergency situations in some ways – they tend to last longer; the fear of contagion complicates people’s emotional responses; etc. And some of what’s in the WHO guidelines addresses the particular crisis communication problems of national governments and international health agencies.

I had a hand in helping draft the WHO guidelines. And my wife and colleague Dr. Jody Lanard was hired by WHO to draft a background document on which the guidelines were built. That background document, based on a literature review and interviews with experienced outbreak communicators, is in three parts:

• The main paper ( 154kB on this site)
• A set of nine appendices ( 183kB on this site)
• A primer on evaluation and coaching ( 126kB on this site)

Bottom line, in my view: outbreak communication is a kind of crisis communication; crisis communication is a kind of risk communication.

But this over-simplifies the situation. An infectious disease outbreak is clearly a crisis, and so it makes conceptual sense to see outbreak communication as a subset of crisis communication. But in practice, outbreak communicators need precaution advocacy skills and outrage management skills too. In the midst of an epidemic, some people may remain apathetic. Alerting them to the risk and persuading them to protect themselves is obviously both an outbreak communication task and a precaution advocacy task. Others in mid-epidemic may be angry or frightened about the wrong things – they may inappropriately blame their neighbors or outside doctors or government officials, for example; or they may fear the wrong disease vector, imagining an airborne disease to be foodborne. Calming misplaced anger and fear is obviously both an outbreak communication task and an outrage management task.

And of course outbreak communication isn’t confined to communicating in mid-outbreak. Warning people about a possible future outbreak (or a possible future second wave of the current one) is surely precaution advocacy. When the outbreak has passed, reassuring people that it’s really over, reducing the stigma problems of those who were exposed or infected, and dealing with recriminations about how the situation was handled are all surely outrage management.

So while the WHO outbreak communication guidelines are grounded in crisis communication principles, it is probably more accurate to see outbreak communication as a subset of risk communication, not of crisis communication only.

Where Does Avian Influenza Communication Fit In?

And now a question you didn’t ask: What kind of risk communication is “avian influenza communication”? All kinds, I think, depending on exactly what you’re doing:

• When you warn the general public that there may someday be an influenza pandemic and people should try to prepare themselves and their families to cope with it when it happens, you’re doing precaution advocacy.
• When you warn farmers to be on the alert for bird flu in their poultry flocks, and to take various steps to lessen the likelihood of their flocks becoming infected, you’re also doing precaution advocacy.
• If there is a local bird flu outbreak (in birds) and you have to explain to affected farmers that you must kill their birds and devastate their livelihoods in order to try to contain the spread of infection, you’ll be doing crisis communication – and probably some outrage management as well.
• During the same bird flu outbreak, when you have to explain to consumers that it is safe for them to keep eating chicken as long as they cook it thoroughly, you’ll be doing outrage management.
• If a novel influenza virus crosses the species boundary from birds to people, starts spreading human-to-human, and launches a devastating pandemic, we’ll all be doing crisis communication.
• If that happens, after the pandemic runs its course there will be endless investigations of why we were so unprepared – and the risk communication task will shift to outrage management. If it doesn’t happen for a long time and there are investigations into why we wasted precious resources preparing unnecessarily, that will call for outrage management too.

Which of these six are outbreak communication? Certainly #5. Probably #3. Arguably #1 (unless you can tolerate the term “pre-outbreak communication”). Maybe all of them.

For sure, all of them are risk communication. And all of them are avian influenza/pandemic influenza (AI/PI) communication.

But what’s most important here is that an AI/PI communicator is ultimately going to need three toolkits: a precaution advocacy toolkit, an outrage management toolkit, and a crisis communication toolkit.

name:	Ulrich Schnabel
field:	Science Editor, Die ZEIT
date:	June 12, 2008
location:	Hamburg, Germany

comment:

As a science editor of Germany’s biggest weekly paper, Die ZEIT, I read with great interest your articles on the public perception of risks and how organizations can deal with them.

As I’m currently working on an article about the ups and downs of media coverage on different risks, I will surely cite your findings.

But there is one open question left for me: What would be good advice for terrified readers of articles on new (and maybe unknown) risks? As you point out, it doesn’t help to tell them: “Stay calm. The real risk has just a probability of xx percent.…”

What would be a better strategy? Or in other words: What is YOUR OWN strategy, when you read about a risk that is unknown to you? How would you deal with it, if you didn’t have time to verify all the scientific or technical details (as it is the case for most of our readers)? Just wait and see whether it is still in the news four weeks later? Or call an authority whom you trust – and which one would that be?

If you have any personal advice, I would be happy to include it in my article.

peter responds:

Note to readers: I received and answered this inquiry back in April, but waited till June 12 hoping the Die ZEIT article would be published. It finally appeared (in German, obviously) on June 19. My response starts with a review of some risk perception basics, but then moves on to topics I haven’t written about much before: how I advise the public to cope with outrage and uncertainty, and how I cope with them myself.

Let me start with some basics you probably already understand.

Outrage trumps hazard in people’s perception of risk.

I distinguish a risk’s “hazard” (how much harm it’s likely to do) from its “outrage” (how upset it’s likely to make people). The essence of my “Risk = Hazard + Outrage” formula is the fact that both hazard and outrage are part of what people mean by risk.

In fact, outrage determines people’s hazard perception more than hazard does. That is, whether or not people take a risk seriously and choose to take precautions (or demand that the government or some company take precautions) depends mostly on how much outrage that risk arouses.

If a risk is coerced, dreaded, and immoral, for example, it will be considered more dangerous than if it’s voluntary, not dreaded, and morally irrelevant. Similarly, if the people responsible for the risk are dishonest and unresponsive, the risk will be considered more dangerous than if they’re honest and responsive. “Outrage components” like these five – voluntariness, dread, moral relevance, trustworthiness, and responsiveness – pretty much determine how seriously we take a risk.

So high-hazard, low-outrage risks, such as driving and smoking, leave people insufficiently concerned, while low-hazard, high-outrage risks, such as mobile telephone towers and pesticide residues on foods, tend to arouse excessive concern.

People are right to focus on outrage.

This emphasis on outrage rather than just hazard isn’t foolish. It reflects our shared understanding that outrage matters. We want to live in a world where we decide for ourselves which risks to accept; where we don’t have to endure the anxiety of facing highly dreaded risks; where institutions behave morally, tell the truth, and are responsive to our concerns. These outrage components – voluntariness, dread, moral relevance, trustworthiness, and responsiveness – are not alien, Martian values. They are our values. By making these values part of what we mean by risk, we are insisting that society take these values seriously in its decisions about risk policy and risk management.

Here’s a hypothetical example I like to give.

Imagine that every ten years or so a sniper climbs up onto an overpass with a high-powered rifle and shoots and kills a passing motorist. Then he’s good for another decade. Finally, after 30 years and three deaths, he is caught and brought to trial. Here’s his defense: “During the 30 years during which I shot and killed three passing motorists, thousands of people died on our nation’s highways as a result of drunk driving, not wearing their seatbelts, poor highway design, and poor automotive design. Sniping is an infinitesimal part of the highway death toll. In picking on me, a mere sniper, the government is distorting the public’s understanding of the real priorities of highway safety. The money the government is spending catching me, trying me, and imprisoning me could save far more lives if the government made the rational risk management decision to let me continue killing a mere one person per decade, and reallocated the money to repainting the lane markers on highways.”

My hypothetical sniper is right on the data. Nonetheless, no jury would vote to acquit. Even after they study the data, normal people support spending more money per life saved catching and punishing snipers than repainting lane markers.

It is true, and important, that outrage leads people to misperceive hazard. If a risk is upsetting, I will imagine it’s dangerous even if it really isn’t. But it is also true, and even more important, that outrage matters apart from its effect on hazard perception. Suppose you succeed in convincing me that some very upsetting risk isn’t really as dangerous as I thought. I will still care that the risk is coerced, dreaded, and immoral; and that the responsible institutions are dishonest and unresponsive. I will still want to avoid such a high-outrage risk, and I will still want the government to regulate it strictly. In other words, people don’t just want to live in a world that minimizes hazard. We want to live in a world that minimizes outrage too.

Many of my clients are companies whose activities impose a low-hazard, high-outrage risk on their neighbors or customers. They keep asking me how to convince people that the hazard is low. I keep answering that they should focus on reducing the outrage instead.

It helps to keep outrage separate from hazard in your mind.

But you’re not asking what my clients should do about the low-hazard, high-outrage risks they are imposing on people. You’re asking what my clients’ neighbors and customers should do about those low-hazard, high-outrage risks? I have three answers:

• Give yourself permission to be upset about high-outrage risks even though their hazard may be low. We human beings are hard-wired to respond that way – and it’s good for society that we are, because it incentivizes government and industry to work to make those low-hazard risks lower in outrage. (Of course it also incentivizes them to reduce the hazard of risks that are already low-hazard, a much less desirable effect.) Feel free to get upset when some risk situation is managed in a way that seems coercive, dreaded, immoral, untrustworthy, unresponsive, etc. And feel free to demand action – preferably action to reduce the risk’s outrage, rather than its hazard.
• Try to remember that a high-outrage risk isn’t necessarily high-hazard. Even as you are giving yourself permission to get upset anyway, bear in mind that you may not actually be endangered – that the anger portion of your outrage may be right on target but the fear portion may be misplaced. Aim for double-entry bookkeeping in your mind, taking outrage seriously without letting it control your judgment about hazard.
• Pay attention to high-hazard, low-outrage risks too. These are the risks that aren’t likely to upset you – but can kill you (or hurt you, or damage the environment). They’re the risks your loved ones keep harassing you about: smoking, driving, eating too much, exercising too little, not going to the doctor or not taking the doctor’s advice, etc. Since these risks lack the crutch of outrage to capture your emotions, they need some concentrated attention from your mind. (And policymakers should do what they can to increase the outrage of such high-hazard risks.)

Coping with uncertainty is hard – and expert overconfidence makes it harder.

Everything I’ve written so far assumes you can tell whether a risk’s hazard is high or low. But often you can’t tell. Some experts (or presumed experts) are saying the risk is dangerous, others are saying it’s safe, and others are saying they don’t really know. Now what can you do?

Uncertainty is part of an outrage factor called knowability. Especially when outrage is high already for other reasons (such as dread and coercion), uncertainty makes it higher: “How dare you make me the unwilling subject of your risk experiment! If you’re not sure it’s safe, how dare you expose me to it!”

Because the experts know that uncertainty makes people anxious, they often pretend more certainty than their science justifies. That almost always backfires on them. For one thing, it leads to a worse knowability problem: expert disagreement. When different experts are making conflicting claims, all of them sounding certain, the public’s outrage goes through the roof. Expert disagreement is such a powerful source of outrage that people are often more upset when half the experts say something is dangerous and half say it’s safe than when they all say it’s dangerous.

Aside from triggering disagreement, expert over-confidence also arouses the public’s mistrust. Here’s a typical pattern in risk controversies. Most of the experts think X is safe, and most of the evidence shows they’re right. But a few experts dissent, and they have a few studies on their side too. Instead of acknowledging the dissenters, the anomalous studies, and the possibility that X might be dangerous after all, the mainstream experts pretend they’re certain and ignore everything that suggests otherwise. The dissenters rightly point out that they and their data are getting ignored. The public rightly senses that the mainstream experts are being overconfident and even dishonest. Fueled by mistrust and expert disagreement, public outrage increases – and so does the public’s perception that X is a serious hazard … which it probably isn’t.

This pattern is characteristic of many recent and current risk controversies, such as bovine growth hormone in milk and the possible link between vaccination and autism.

Sometimes the mainstream experts are on the alarming side of a risk controversy (global warming, for example). Other times they are on the reassuring side (mobile telephones, say). It can be hard to tell which side is the mainstream side. Dissenters tend to communicate more aggressively than the mainstream – so journalists may pay them as much attention, and their websites may be easier to find. Distinguishing dissenting experts from dissenting non-experts can also be a challenge.

Still, with a little research it is usually possible to figure out where the majority of the experts stand. That’s certainly a lot easier than reaching an independent judgment about which side is right. So figure out which side the mainstream experts are on, and then follow these three decision rules:

• The mainstream experts usually turn out mostly right in the end (though less completely right than they were claiming). Of course the times when the mainstream was dead wrong and the dissenters were right are memorable and very important. Paradigm shifts do happen – but they’re relatively infrequent. Bet on the mainstream.
• It’s usually a lot more damaging to think something is safe when it turns out dangerous than to think something is dangerous when it turns out safe. So the alarming side in a risk controversy deserves the benefit of the doubt, even when the mainstream experts are on the reassuring side. If most experts say X is safe but a few say it’s deadly, and if you can live nicely without X, it’s not crazy to go with the dissenters – even though you know you’re probably taking an unnecessary precaution. When the cost of being mistakenly cautious is low and the cost of being mistakenly unconcerned is high, in other words, it may be sensible to be cautious even though the mainstream experts are telling you caution is unnecessary. This strategy does have a social cost in the aggregate. If a lot of people are unnecessarily cautious about X, the X industry will pay the price. If a lot of people are unnecessarily cautious about a lot of things, we will all pay the price: excessive spending on trivial hazards and reduced attention to more serious ones. Nonetheless, erring on the alarming side makes sense.
• The second rule has exceptions. Sometimes it can be very damaging to think something is dangerous when it turns out safe. Wrongly thinking a vaccine is dangerous, for example, can leave you unvaccinated and at risk of catching a disease everyone knows is dangerous. The vaccine risk is a bigger outrage than the disease risk – but the disease risk is a bigger hazard. When the cost of being mistakenly cautious is higher than the cost of being mistakenly unconcerned (or when it’s about as high), and the mainstream experts are unconcerned, go with the mainstream – even if it means setting aside your own outrage.

How do I respond to risk?

You asked how I personally respond to risk.

For health and environmental risks, I usually follow my first decision rule. That is, rather than try to figure out the science myself or call a scientist I trust, I usually decide where most of the experts stand – and that’s where I stand too. Once in a while I rely on the second rule, deciding to go with the cautious minority on the grounds that I’d rather be wrongly cautious (and safe) than wrongly unconcerned (and endangered). And once in a while I smell a rat and decide that the mainstream experts simply can’t be trusted. (An example of smelling a rat: I got myself a prescription for Tamiflu to have in case of a pandemic even though most experts recommended against it.) But the vast majority of the time I take the precautions the mainstream experts recommend, shrugging off the dissenters’ warnings.

I am unusual – statistically weird – in that I find this fairly easy to do. Most people are more responsive to a risk’s outrage than I am. And most people are less comfortable than I am playing the odds – that is, relying on the mainstream experts while knowing that once in a while they’ll be mistaken or dishonest. My atypical risk response is probably why I became interested in risk perception and risk communication issues in the first place.

That’s for health and environmental risks. In other risk venues I am oversensitive to outrage and totally impervious to expert reassurance. For example, I avoid situations that pose a high risk (in my mind) of social embarrassment – sports and dancing come immediately to mind – even though I know the real hazard to my reputation or self-esteem is very low.

name:	Goatchowder
field:	Activist
date:	June 11, 2008
location:	California, U.S.

comment:

Your column on concession speech recommendations for Hillary Clinton provided a great analysis, but you’re not a very compelling speechwriter. Your article would have been stronger without “examples” of what you thought Clinton should have said.

As it happens, it seems that she did take your (or someone else’s) advice and did pretty much what you said, to stunning results. I think she did a great job (and I supported Kucinich this year). Her speechwriters did a better job of implementing your suggestions.

Perhaps you could replace your “example” passages with actual well-written quotes from her speech? That might be helpful to other politicians, and might make your article a stronger, more useful, and more inspiring reference in the future.

peter responds:

I accept the point that Sen. Clinton has better speechwriters than Peter Sandman and Jody Lanard. But I don’t agree that her June 7 speech “suspending” her campaign did what we were recommending.

I think she did a good job of endorsing Sen. Obama – which was, of course, the sine qua non for the speech. I give her full credit for that. Well, almost full credit. The endorsement language, while strong, was surrounded by an awful lot of language celebrating not just her supporters (which I think was appropriate) but also her causes and in some ways herself. (I didn’t count the number of times she mentioned her tally of 18 million – but I thought it was too many.) And, perhaps inevitably, her face and voice were a lot more enthusiastic during those other passages than during the Obama endorsement passages. I think the speech reads like a stronger endorsement than it sounded like, and it sounded like a stronger endorsement than it looked like.

Still, I give her credit for a strong endorsement. That can’t have been easy. And it wasn’t long-delayed, coming just three days after his delegate count went over the top. That’s pretty short, especially given the contortions surrounding the assignment of Michigan and Florida delegates. Usually when a candidate gets enough delegates, that is truly the end. But it must feel to Sen. Clinton, and to her supporters, like Sen. Obama kind-of sort-of got enough delegates – so conceding and endorsing were harder. (It would have felt the same to Sen. Obama and his supporters if Sen. Clinton had been put over the top by superdelegate endorsements.)

What Sen. Clinton didn’t do on June 7 – and what the column urged her to do – was to address explicitly her own anger, anguish, and ambivalence, and the anger, anguish, and ambivalence of her followers. The closest she came to that – and it was a terrific line – was the first thing she said: “Well, this isn’t exactly the party I’d planned….”

She said she was supporting Obama, but she didn’t say it was painful for her. She said she wanted her followers to support Obama, but she didn’t say she knew it would be painful for them. These acknowledgments would have helped her followers make the difficult transition she was making. They would have addressed her followers’ outrage.

Of course these acknowledgments can still do so in the weeks to come. One of my favorite bloggers said it best: Reconciliation is a process, not an event.

There is still plenty of time for both Sen. Clinton and Sen. Obama to consider the outrage management task they face – time for them to ask themselves how best to help Clinton Democrats reconcile themselves to becoming Obama Democrats (or at least Democratic Party Democrats), and time for them to realize that this task calls for a different sort of communication than traditional political cheerleading. There’s no big hurry.

While I wanted Sen. Clinton to make a start in her June 7 speech, arguably that was too early. We should all recall the large cohort of people on the Republican right whose early reaction to Sen. McCain’s success in wrapping up the Republican nomination was to vow never to support him. Most of them, today, are supporting him. It didn’t take a psychologically sensitive endorsement from, say, Mike Huckabee to win them over. It took time.

Nonetheless, the responses to Clinton’s June 7 speech were bimodal in a way I found illuminating.

If you read the reactions of Obama supporters on Daily Kos, for example, it’s clear that most thought it was a wonderful speech. Their enthusiasm about the speech she gave makes me wonder how Obama supporters would have responded to the speech Jody and I had urged her to give. It’s possible they would have li