1 Presented at the 52 nd Annual Fall Training Conference “Changing Landscapes in Emergency Management” Association of Minnesota Emergency Managers Breezy

1

Presented at the 52nd Annual Fall Training Conference“Changing Landscapes in Emergency Management”

Association of Minnesota Emergency ManagersBreezy Point, Minnesota

byDennis S. Mileti, Ph.D., Professor Emeritus

University of Colorado, BoulderSeptember 18, 2012

How can we:

Get warnings to everyone

Reduce public action delay time

Motivate appropriate public actions for…..

Dennis S. Mileti September 2012 2

3Dennis S. Mileti September 2012













Time to impact: None, minutes, hours, days

Audience to address: Single (one audience) Multiple (different actions/populations at same

time)

Protective action recommendations: Single (one action) Overlapping (multiple actions at same time) Sequenced (different actions in a row)


Reasonable conclusions about: 1. Public warning response behavior 2. Warning system reliability &

preparedness

Based on: Findings from 50 years of research


ABOUT THE RESEARCH


Half-century social science research: Hazards & disasters research literature U.S. emphasis--but not exclusively Protective actions studied:

Some a lot, others a little, some not at all Example events studied:

Natural: Hurricane Camille, Mt. St. Helens Terrorism: World Trade Center 1993 & 9/11 Hazardous Materials: Mississauga, Nanticoke Technology: Three Mile Island Building Fire: MGM Grand, Cook County

Hospital 19Dennis S. Mileti September 2012

REFERENCES: 350 page annotated bibliography available at:

http://www.colorado.edu/hazards/publications/informer/infrmr2/pubhazbibann.pdf


REFERENCES: 150 entry bibliography available at: http://www.colorado.edu/hazards/library/

BuildingsEvacBib2007.doc21Dennis S. Mileti September 2012

Studies on “hypothetical” events: Can yield wrong response conclusions:

Situational determinants of behavior NOT operating Preferences & intentions = little predictive weight

Useful for some specialized topics, e.g., Which words are/aren’t understandable

Studies of “actual” events: Yield more realistic response conclusions:

Situational determinants of behavior ARE operating Real people & events = real warnings & response


DEFINITIONS (for presentation purposes)


Definition: Get people’s attention

Old fashioned approach: Air raid sirens

Modern approach: CMAS (linked to IPAWS & EAS) Use cell phones & other mobile devices

to get people’s attention & provide messages


Definition: Emergency information that motivates

timely & appropriate public behavior

Alerting & warning are different: Alerting = get people’s attention Warning = motivate behavior Distinction being blurred in today’s

worldDennis S. Mileti September 2012 25

TWO SIDES TO THE BEHAVIOR COIN


Public warning response is predictable: About 40% explained variance (as good as it gets)

Key factors that predict it are known: Apply across hazards & events Mathematically modeled (tested & retested)

Public warning response behavior: Varies across events because of variation in the

factors that influence it Is malleable & somewhat manageable:

By managing the factors that influence it Some people will always do the wrong thing


Research also includes: Predicting the behavior of warning providers

The “sender” part of warnings Based on historical event investigations

Influences on warning provider behavior: Relatively well understood Variation across events Is malleable and manageable:

Also by managing factors that influence itDennis S. Mileti September 2012 28

PUBLIC RESPONSE BEHAVIOR


“Objective” reality for people = what they think is real What people think comes from interacting with others Most people go through life thinking they’re safe Warnings tell them they’re not & consequently Compel most people to mill around:

Interact with others & get more information & search for confirming information to form new ideas about safety & risk

“Milling” (some call it “sense-making”) intervenes between warning receipt & protective action-taking

Results in public protective action-taking delay Part of being human & will never change


Human beings are….. “the hardest animal of all on the planet to

warn”

An “exaggerated” example: While all the forest animals are running away from the

flames…..most people are talking about it with neighbors, looking at TV coverage, texting, & rubber necking trying to find out what it means & deciding what to do

Creates a public response gap: Few are skilled at shortening the time people spend

delaying protective action resulting in some unknowingly doing things that increase it




.

2 4 01 8 01 2 06 0

0.0

0.1

0.2

0.3

0.4Siren (%)Route (%)Informal (%)Media (%)

Diffussion of Warning at Nanticoke

Time

P o r t i o n W a r n e d

1 am 2 am 3 am 4 am

Audience factors impact what people hear, how they interpret it & what they do: Statuses (gender, sex, age, ethnicity, SES) Roles (children, family united, pets, kinship) Not just demographics:

Experience, knowledge, perceptions & beliefs Environmental and social cues

Effects of audience factors vary: Significant but not large with poor warning messages Many weaken in presence of strong warning messages

Some constrain communication & response: Special needs sub-populations (unique effects) Special communication channels (for sub-populations)


What to say (5 topics matter most): WHAT (Guidance): What to do & how to do it WHEN (Time): When to begin & complete the

protective action WHERE (Location): Who should & shouldn’t do

it using clear geographical boundaries WHY (Hazard): The event, the consequences &

how the protective action reduces them WHO (source): Who’s giving the message:

No single credible source, name a panel of sources


36Dennis S. Mileti February 2011

http://www.msplinks.com/MDFodHRwOi8vdGlueXBpYy5jb20=

How to say it (5 attributes matter): CLEAR: Simply worded SPECIFIC: Precise & non-ambiguous ACCURATE: Timely, accurate & complete CERTAIN: Authoritative & confident even in

face of uncertainty CONSISTENT:

Externally: Explain changes from past messages & differences from what

others are saying Internally: Never say “attack will occur soon, don’t

worry” Dennis S. Mileti September 2012 37


This is a MANDATORY EVACUATION ORDER from the Yellow County Sheriff’s Department AND Fire Authority. There’s a high risk of CATASTROPHIC MUDSLIDES & DEBRIS FLOWS due to rain on BURNT SLOPES. After consulting with the U.S. Geological Survey and the National Weather Service, WE issue a mandatory evacuation order for the following people in Yellow County: If you LIVE IN or ARE IN an area BELOW or NEAR burnt

slopes, evacuate now. Do not delay. This is a MANDATORY EVACUATION ORDER. Evacuate. Evacuate NOW.

What we mean when we say evacuate is: GET OUT OF ALL CANYONS, and get out of them NOW.

If you don’t live in or aren’t in an area below or near burnt slopes, you don’t need to do anything.


If you have CHILDREN IN A SCHOOL located in a canyon: DO NOT GO THERE TO GET THEM. They won’t

be there when you arrive. All school children in all canyon schools are

currently being evacuated to (insert the name, address, and telephone number).

You can pick them up there once you have completed your own evacuation.


There’s HIGH RISK of catastrophic mudslides and debris flows due to rain on BURNT SLOPES: Mudslides and debris flows could occur NOW, and

they could be large enough to COMPLETELY BURY homes, roads, and lives.

They can occur WITHOUT NOTICE. The amount of rain needed to start a catastrophic

mudslide or debris flow is SMALL. Don’t think you’re safe because the rainfall you see is slight.

The risk of catastrophic mudslides and debris flows below all burnt slopes in all Yellow Country is REAL.


If you LIVE IN or ARE IN an area BELOW or NEAR burnt slopes EVACUATE NOW. Evacuate WITHOUT DELAY. Evacuate NOW. This is a MANDATORY EVACUATION ORDER. There’s a Red Cross shelter at Monroe High

School in the town of Yellow. Bring your pets and keep them with you.

Larger animals can be brought to the Yellow County Fairgrounds.

If you have questions or require assistance, please call 123-456-7890.

New information will be made available as it becomes available on this same station/channel.



Headline (160 character limit) Nuclear explosion in downtown Denver. High

radiation levels in air and blowing in the direction of the wind. Take shelter indoors immediately!

Description (description + instruction = 160 word limit) A nuclear explosion occurred at 1:00 PM in

downtown Denver. Radiation is in the air, blowing in the wind, and falling to the ground. Exposure can be deadly and cause illness.


Instruction (description + instruction = 160 word limit) You can increase your chances of surviving and reducing

injury by IMMEDIATELY going inside the nearest tall building, basement, underground garage or tunnel and staying there. Shelters of brick, concrete, and earth will give the most protection. If you are in a different kind of building, stay there unless you can reach a better shelter in a few minutes or less. DO NOT EVACUATE. Drivers should stop their cars and take shelter. You will receive less radiation inside than outside no matter how fast you drive. If you are not in the area, STAY OUT. Keep listening to the media for more instructions. We will ask you to take other actions later. This “Shelter Order” is based on advice from nuclear scientists and local, state, and federal emergency managers.


Number of communication channels: More channels work better than fewer channels Some subpopulations need unique channels

Type of communication channels: Personal delivery channels work best Channel “diversity” (multi-media) helps too

Frequency of communications: The more its repeated & heard the better:

Repetition fosters confirmation which yields taking action



48

STATUS

ROLES

EXPERI-ENCE

CUES

INFORECEIVED

MILLING

KNOW-LEDGE

PERCEIVEDRISK

ACTION

INFO BELIEF

Dennis S. Mileti February 2011

49

STATUS

ROLES

EXPERI-ENCE

CUES

INFORECEIVED

MILLING

KNOW-LEDGE

PERCEIVEDRISK

ACTION

IINFO BELIEF

Dennis S. Mileti February 2011

Models are represented by equations: Called “simultaneous multiple regression equations”

Equations enable us to determine: Effect of every factor while controlling for the effects of

everything else (good science) The result is:

Distinguish between what’s really important & what isn’t

When to get excited: When different studies reach the same conclusions That’s where we are with research on public response

to warnings for hazardous events


X4 = β41X1 + β42X2 + β43X3 + e4 X5 = β51X1 + β52X2 + β53X3 + β54X4 + e5 X6 = β61X1 + β62X2 + β63X3 + β64X4 + β65X5 + e6 X7 = β71X1 + β72X2 + β73X3 + β74X4 + β75X5 + β76X6 + e7

*Averill, J. D., D.S. Mileti, R.D. Peacock, E.D. Kuligowski, N. Groner, G. Proulx, P.A. Reneke, and H.E. Nelson. 2005. Federal Building and Fire

Safety Investigation of the World Trade Center Disaster: Occupant Behavior,

Egress, and Emergency Communications. Report NCSTAR 1-7, National Institute of

Standards and Technology, Gaithersburg, MD. Available at:

http://wtc.nist.gov/NISTNCSTAR1-7.pdf


Not just about official warning messages: Public gets information from many sources

Public in an “information soup” when warned: Many formal & informal information sources Some information is correct & some is not Inconsistencies slow protective action-taking

What works best = deliver warnings and manage the soup: Put good information in & take bad information out


Managed warning information includes: Use evidence-based messages (pre-scripted & vetted) Take audience factors into account for delivery Actions to reduce public milling & response delay

Match messages across information providers Distribute messages repetitively over diverse channels Send the messages to other providers + JIC

Inform people not at risk to reduce “response creep” Monitor public response (people at & not at risk) Listen for wrong information & then Re-warn with adjusted messages based on what

people are + aren’t doing, wrong information, & any changed protective actions recommendations plus

Q & A provide & staff a call-in number


Even great warning messages: Aren’t silver bullets that work well on their

own

Messaging impacts public response most effectively when: “Its a process of public messaging &

information management based on plans, procedures and training”

Bottom line: Emergency communication planning works,

not planning doesn’t work quite as wellDennis S. Mileti September 2012 54

WARNING PROVIDER BEHAVIOR


Warnings come from a system of people, agencies & organizations: A systems perspective helps “see” all the

parts

System-level preparedness helps to: Design, plan, train & exercise to create a more

“highly reliable integrated warning system” In place long before an actual event occurs


57

DETECTIONMonitoring

Risk DetectionData Assessment &

AnalysisPredictionInforming

MANAGEMENTInterpretation

Decision to WarnWarning Content &Protective Action

SelectionWarning Method &

ChannelResponse Monitoring

Warning Feedback

PUBLIC RESPONSEInterpretation

Confirmation & MillingResponse

Warn Others

RISKNatural Environment

TechnologicalCivil

Dennis S. Mileti September 2012

58

RISKNature

TechnologyTerrorists & more

DETECTIONScientific AgenciesLaw Enforcement

(Police, DHS, CIA, FBI)Public

MANAGEMENTGovernment

(Local, State, Tribal)Building Operators

RESPONSEGeneral PublicRacial & Ethnic

MinoritiesVisitors & TransientsSpecial Needs Groups

Organizations & Facilities

Dennis S. Mileti September 2012

Warning system failures can occur anywhere in the system: Many links across functions & actors Historical examples of non-failures & failures Reasons for historical failures documented

Warning preparedness: Integrates all parts of the system resulting in

a “more reliable” system with lower odds of failing


SYSTEM DESIGN FLAWS: Warning system design, preparedness, training lacking Un-reliable system linkages, e.g., detectors to

managers Actor’s personality not removed with procedures Fail safe solutions for technological problems missing Problems of non-communication not addressed

MESSAGING FLAWS: Evidence-based messages not used Everyone at risk not reached People not at risk not communicated to Repetitive message dissemination absent Message management missing


The link between: Risk detectors & local warning providers

Ready local warning providers to receive information from risk detectors with: “Planned triggers & procedures” about when

to warn and what public protective actions to recommend to whom

Ad hoc approaches have historically been a root cause of warning system failures Dennis S. Mileti September 2012 61

Warning messages should be short People may panic One-way delivery is communication People will understand the message Messages can’t be changed There’s one public A credible message source exists People blindly follow instructions One channel delivery works Great messages guarantee great response


Don’t confuse with preparedness education

Pre-event public “warning” education: Doesn’t much influence response in an actual event Why: warning response is largely determined “in situ”

Use to teach people: Hazard exists, warning system & source, etc.

And to acquaint people with: Protective actions, e.g., don’t pick kids up at school

In other words: It can prime the public by removing surprises and

reducing confusion in future warning eventsDennis S. Mileti September 2012 63

Community warning metric: Assess research knowledge implementation

Measured in several UASI areas: Washington, D.C., New York, & Los Angeles

Key findings: Application lags behind knowledge What is applied is done so unevenly

Narrowing the gap: Plan development & training for practitioners Modernized guidance Pre-scripted (& pre-vetted) warning messages


GAME CHANGERS


New approaches & new technology: CAP (Common Altering Protocol) IPAWS (Integrate Public Alert & Warning

System) CMAS (Commercial Mobile Alert System) All hold great promise

Message length limited: By carrying capacity of local distribution

systems, e.g., Emergency Alert System (EAS) character limits

Research is neededDennis S. Mileti September 2012 66

“Sirens in our pockets” Combines alerting & warning:

Blurs distinction (calls them both alerting) Message length limits:

90 characters (not words) long (not long) Holds promise & raises hypotheses:

Decrease diffusion time? Increase milling & response delay time? Enhance risk personalization? Research is needed


Won’t change some things: How people are “hard wired” Strong impact of message factors on public response behavior

Will change other things (hypotheses): Accelerate milling, confirmation, informal

notification How public response can be monitored Evidence so far = is mixed (about actual use) Role & use likely to change over time Holds promise Research is needed


END NOTES


We only “hit the highlights”: More could be said about everything:

This was a speech (not a course)

Social science knowledge can’t: Provide guarantees about public response or Solve all public warning & response problems

But it can: Help solve some problems and Point to planning & training needs


“The key determinant of public

warning response has more to do with what public information providers say to the public than anything to do with the public itself”


QUESTIONS? [email protected]

303-520-3400


Documents

1 Presented at the 52 nd Annual Fall Training Conference “Changing Landscapes in Emergency Management” Association of Minnesota Emergency Managers Breezy