WHEN SHOULD I LEAVE FOR THE AIRPORT?
Applying EPA's Risk Assessment
Methodology to the Risk of Missing
an Airline Flight
t is instructive to compare EPA's Superfund risk
assessment process to the approach that people
commonly use in making everyday decisions
such as: When should I leave for the airport? Assume
for example that Beth Moore, a newcomer to Wash-
ington, D.C., is planning to catch the 9:30 a.m. Ozark
Airlines flight from Washington's Dulles Airport to
Little Rock, Arkansas. Beth has never used Dulles
Airport before and she must decide when to leave
downtown D.C. to catch the flight. The variables that
might affect her decision include:
(1) distance to the airport,
(2) various routes to the airport,
(3) speed limits on such routes,
(4) existence of construction or delays on the high-
(5) general availability of taxis willing to drive to
(6) familiarity of taxi drivers with the appropriate
(7) the mechanical condition of the taxi,
(8) timing and relative congestion caused by the D.C.
(9) the length of the lines at the Ozark Airlines ticket
(10)delays associated with the airport security check-
points and the Dulles shuttle bus system.
Each of these factors is relevant to Beth's "travel
Beth consults two friends to get information for
her travel assessment. Jim Jones, another newcomer
to Washington, has never made the trip to Dulles, but
offers Beth his best assumptions about D.C. traffic.
Jim says that it is his understanding that Dulles is 45
miles away, the speed limits are low and strictly
enforced, the D.C. roads are full of potholes and
always under construction, the D.C. taxi drivers are
notoriously unfamiliar with the best routes, the taxis
are in poor mechanical condition, the rush hour lasts
until 10:00 a.m., the Dulles airline counters are always
congested, and the airport security and shuttle bus
system are subject to interminable delays. Based on
Jim's independent estimates of each potential delay,
the following times and delays could affect Beth's
Normal driving time 1 hour
Construction delays 2 hours
Taxi driver loses way 1 hour
Taxi flat tire 3 hours
Rush hour delay 45 minutes
Ticket line 45 minutes
Security & shuttle 45 minutes
TOTAL TIME: 9 hours 15 minutes
Thus, if all ofJim's worst case estimates are added
together, the total travel time to catch the 9:30 a.m.
flight would be 9 hours and 15 minutes. Beth would
therefore need to leave for the airport at 12:15 a.m.
Next, Beth consults another friend, Monica Men-
tor. Monica regularly takes the 9:30 a.m. Ozark flight
from Dulles to Little Rock. Monica says that in order
to catch the flight, she normally leaves downtown
D.C. at 8:15 a.m. Monica made this trip just last week,
and found the highway excellent and the drive unaf-
fected by the rush hour because it is against traffic.
Hazardous Waste Cleanup Project
Monica always uses Dupont Cab Company, which has
efficient drivers and reliable taxis.
SArmed with this information, Beth's decision as
to when to leave for the airport is based on a two-step
process. First, Beth will make a "travel assessment,"
estimating the amount of time that it will probably
take her to get to the airport. Next, Beth will make a
"travel management" decision, involving Beth's per-
sonal preferences (policy choices) as to how much of
a "safety margin" she likes to add to her estimated
travel time. These two steps in Beth's decision are
parallel to the Superfund steps of risk assessment and
If Beth were to use Jim's conservative estimates
as the basis for her travel plans, she would need to
leave for the airport at an extremely early hour.
Although she would have an enormous "safety mar-
gin" against missing the 9:30 flight, she would waste
many hours at the airport. The flaws in Jim's ap-
proach include the following:
Use of Assumptions Instead of Actual Data
Beth can either rely on Jim's second-hand infor-
mation about the bad reputation of D.C. traffic, roads,
and cabdrivers, or she can use the actual data pro-
vided by Monica. Actual data are always preferable
to generic assumptions.
Use of Excessively Conservative
Many of Jim's estimates, such as three hours to
replace a flat tire, are based on assumptions (e.g., no
spare and no alternative cab) that, although theoreti-
cally possible, are excessively conservative and quite
unrealistic. In addition, Jim's overconservative ap-
proach stems from his unstated personal aversion to
missing any flight and his assumption that catching
the 9:30 flight is probably crucial to Beth's career.
Jim's inflated numbers thus incorporate unstated
value judgments that are totally inappropriate at the
initial travel assessment stage.
An additional problem with Jim's approach is that
he simply took each worst-case assumption and
added them together to produce a predicted travel
time of 9 hours and 15 minutes. This is absurd-it is
extraordinarily unlikely that every element in the
travel time to Dulles would be subject to its worst-case
delay. Instead, any reasonable travel assessment
would factor in the probabilities associated with each
type of delay. For example, for the ticket counter
"factor," the traveler might know that (1) 10% of the
time there is no line at the counter so that purchasing
a ticket only takes three minutes, (2) 80% of the time
there is a line and it takes ten minutes; and (3) 10%
of the time there is a terrible line and a 45-minute
delay. Likewise, the flat tire "factor" might occur in
only 1% of the taxi rides. A person would more
typically use these probabilities in her decisionmak-
ing and recognize that the probability of both a
terrible line at the ticket counter and the cab getting
a flat tire is remote. No reasonable person would
assume the worst for every element and leave for the
airport at 12:15 a.m.
Unfortunately, EPA's Superfund risk assessment
process commonly makes the same mistakes as Jim,
leading to similar waste and misallocation of effort.
The Superfund approach, however, is even more
conservative than the airport example because most
of the factors in the Superfund risk assessment proc-
ess are multiplied together rather than added.*
In the airport example, the most conservative departure time is calculated by adding together each possible delay. In a risk
assessment, the most conservative expression of a factor, such as exposure, is calculated by multiplying each variable together.
For example, to calculate lifetime intake of water from a waste site, one would multiply the following: (daily intake of water) x
(percentage of daily intake of water derived from the waste site) x (number of years spent living near the site). Adding the numbers,
such as 5 + 5 + 5 15, does not result in as high, or conservative, a number as multiplying the numbers, such as 5 x 5 x 5 125.
EXAGGERATING RISK 7