Response
to
"
Comments
of
American
Chemistry
Council,
American
Forest
&
Paper
Association,
American
Iron
and
Steel
Institute,
American
Petroleum
Institute,
American
Public
Power
Association,
and
Utility
Water
Act
Group
on
EPA's
Proposed
Information
Collection
Request
on
Willingness
to
Pay
Survey
for
§
316(
b)
Phase
III
Cooling
Water
Intake
Structures"

DOCKET
ID
OW­
2005­
0006­
0007
EPA's
responses
to
the
"
Comments
of
American
Chemistry
Council,
American
Forest
&
Paper
Association,
American
Iron
and
Steel
Institute,
American
Petroleum
Institute,
American
Public
Power
Association,
and
Utility
Water
Act
Group
on
EPA's
Proposed
Information
Collection
Request
on
Willingness
to
Pay
Survey
for
§
316(
b)
Phase
III
Cooling
Water
Intake
Structures,"
the
attached
comment
"
Comments
on
EPA
316(
b)
`
Willingness
to
Pay'
Survey
Materials"
from
Lawrence
Barnthouse,
and
the
attached
comment
"
Evaluation
of
the
EPA's
316(
b)
Phase
III
Rule
Willingness
to
Pay
Survey:
Final
Report"
from
Triangle
Economic
Research
(
TER)
are
organized
to
correspond
to
the
individual
ordered
sections
of
the
comments.
These
comments
are
henceforth
referred
to
as
the
"
ACC
et
al.
Comment,"
"
Barnthouse
Comment,"
and
"
TER
Comment,"
respectively.

Responses
to
the
ACC
et
al.
Comment
Introduction
and
Overview
The
ACC
et
al.
Comment
is
summary
in
nature
and
draws
heavily
on
the
Barnthouse
Comment
and
the
TER
Comment.
In
this
spirit,
EPA
offers
the
following
general
responses
that
address
the
main
issues
presented
in
the
list
of
points
beginning
on
page
2
of
the
comment.
Detailed
responses
to
all
comments
can
be
found
in
the
Agency's
response
to
the
Barnthouse
Comment
and
TER
Comment.

The
commenter
argues
that
there
are
several
reasons
why
EPA
will
not
be
able
to
develop
reliable
estimates
of
the
non­
use
benefits
of
the
316(
b)
regulation
for
Phase
III
facilities:


"
EPA's
view
of
non­
use
values
as
presented
in
the
November
23,
2004
ICR
is
conceptually
flawed."
EPA
disagrees
with
this
statement,
which
does
not
reflect
the
wide
conceptual
acceptance
of
non­
use
values
in
the
academic
literature.
As
stated
by
Freeman,
"
Non­
use
values,
like
use
values,
have
their
basis
in
the
theory
of
individual
preferences
and
the
measurement
of
welfare
changes.
According
to
theory,
use
and
non­
use
values
are
additive"
(
Freeman,
2003).


"
The
proposed
ICR
contains
numerous
flaws
that
will
result
in
unreliable
benefit
estimates."
Again,
EPA
disagrees
with
the
commenter.
As
discussed
in
the
Agency's
detailed
responses,
below,
EPA
has
extensively
tested
the
survey
using
focus
groups
and
cognitive
interviews.
Results
indicate
that
the
survey
is
effective
at
eliciting
respondents'
opinions
about
reducing
fish
losses,
and
that
the
survey
will
produce
valid,
unbiased,
and
reliable
willingness­
to­
pay
estimates.


"
EPA
based
its
survey
on
unreliable
empirical
estimates
of
the
effect
of
cooling
water
intake
structure
regulations
on
fish
populations
and
fisheries
production."
In
response,
EPA
first
notes
that
the
survey
materials
are
based
on
the
best
biological
and
engineering
data
available.
Second,
the
stated
preference
survey
will
still
produce
results
that
are
meaningful
in
the
context
of
the
scenarios
that
are
presented
to
respondents.

The
commenter
argues
that
the
benefits
of
regulating
cooling
water
intake
structures
for
recreationally
and
commercially
significant
species
are
already
reflected
in
the
EPA's
RIA
for
the
proposed
rule
for
Phase
III
facilities.
EPA
disagrees
with
the
commenter's
statement.
As
stated
by
Freeman,
"
Non­
use
values,
like
use
values,
have
their
basis
in
the
theory
of
individual
preferences
and
the
measurement
of
welfare
changes.
According
to
theory,
use
and
non­
use
values
are
additive"
(
Freeman,
2003).
Moreover,
as
stated
by
Freeman
(
2003,
p.
137),
"[
t]
he
hypothesis
of
nonuse
values
has
gained
wide
acceptance
among
economists
working
in
the
field
of
environmental
and
resource
economics "
and
the
suppression
of
such
values
"
in
natural
resource
policymaking
could
lead
to
serious
errors
and
resource
misallocations"
(
p.
138).
The
potential
significance
of
nonuse
values
as
a
distinct
component
of
total
value 
and
the
requirement
that
stated
preference
methodologies
are
used
to
measure
these
values 
are
also
emphasized
by
the
NOAA
Blue
Ribbon
Panel
on
Contingent
Valuation
(
Arrow
et
al.
1993)
and
EPA's
guidelines
for
benefit
cost
analysis
(
US
EPA
2000).
The
Agency
notes
that
the
non­
use
benefits
of
the
proposed
regulation
are
a
completely
separate
category
of
value
from
the
commercial
and
recreational
welfare
effects
of
the
proposed
rule.
Furthermore,
results
from
the
focus
groups
and
cognitive
interviews
indicated
that
some
individuals
hold
substantial
non­
use
values
for
protecting
fish
species
with
direct
uses,
in
addition
to
any
use
values
that
they
have
for
those
species.

The
commenter
argues
that
the
presentation
materials
prepared
by
EPA
contain
inaccurate
statements
and
invalid
comparisons
that
will
lead
respondents
to
believe
that
the
benefits
of
increased
regulation
of
cooling
water
intakes
would
be
substantially
greater
than
is
actually
the
case.
EPA
agrees
that
certain
details
of
the
survey
and
presentation
materials
provided
in
the
supporting
documentation
for
the
June
9
Federal
Register
notice
required
revision.
Since
issuing
the
June
9
Federal
Register
notice,
EPA
has
conducted
five
focus
groups
and
two
cognitive
interview
sessions
to
improve
the
Agency's
understanding
of
the
public's
perceptions
and
attitudes
concerning
fishery
resources,
to
pretest
draft
survey
questions
and
presentation
materials,
to
test
for
and
reduce
potential
biases
that
may
be
associated
with
stated
preference
methodology,
and
to
ensure
that
both
researchers
and
respondents
have
similar
interpretations
of
survey
language
and
scenarios.
As
a
result
of
this
extensive
pre­
testing,
a
number
of
revisions
have
been
made
to
the
survey
that
have
significantly
improved
its
reliability
and
reduced
its
potential
for
bias.
Results
from
pretests
indicate
that
the
presentation
materials
prepared
by
EPA
do
not
lead
respondents
to
believe
that
the
benefits
of
increased
regulation
of
cooling
water
intakes
would
be
substantially
greater
than
is
actually
the
case.

The
commenter
argues
that
because
the
survey
materials
are
based
on
unreliable
population
and
fish
losses
data,
the
survey
results
will
also
be
unreliable.
In
response,
EPA
first
notes
that
the
survey
materials
are
based
on
the
best
biological
and
engineering
data
available.
Second,
the
stated
preference
survey
will
still
produce
results
that
are
meaningful
in
the
context
of
the
scenarios
that
are
presented
to
respondents.
Different
versions
of
the
survey
will
show
a
range
of
different
baseline
and
resource
improvement
levels,
where
these
levels
are
chosen
to
(
almost
certainly)
bound
actual
levels.
Different
respondents
will
be
asked
to
make
choices
over
all
possible
policy
scenarios
where
impingement
and
entrainment
reductions
range
from
0%
(
no
policy)
to
98%.
Given
that
there
will
almost
certainly
be
some
biological
uncertainty
regarding
the
specifics
of
the
actual
baselines
and
improvements,
the
resulting
valuation
estimates
will
allow
flexibility
in
estimating
WTP
for
a
wide
range
of
different
circumstances.

The
commenter
also
argues
that
the
survey
is
unnecessary
because
in
independently
conducted
verbal
protocol
interviews
with
15
individuals,
respondents
did
not
demonstrate
meaningful
values
for
marginal
changes
in
forage
fish
populations.
EPA
notes
that
this
finding
contradicts
the
results
of
the
EPA
focus
groups
and
cognitive
interviews
conducted
under
ICR
#
2155.01.
These
focus
groups
and
interviews
provided
strong
evidence
that
many
individuals
value
forage
fish
and
are
willing
to
pay
to
prevent
losses
of
all
fish
species
including
forage
fish.
Participants
in
the
EPA
focus
groups
and
interviews
cited
a
variety
of
motivations
for
preventing
fish
losses,
including
the
satisfaction
of
knowing
that
the
fish
exist,
the
desire
to
bequeath
healthy
fish
populations
to
future
generations,
and
the
desire
to
protect
the
functioning
of
aquatic
ecosystems.
Furthermore,
EPA
notes
that
the
economic
literature
advises
against
the
use
of
ad
hoc,
expert
assumptions
regarding
the
magnitude
of
non­
use
values 
particularly
assumptions
that
non­
use
values
are
trivial.
Because
values
are
inherently
subjective,
Bateman
et
al.
(
2002,
p.
75)
emphasize
that
"
it
would
be
wrong
for
experts
to
assume
that
one
resource
is
a
perfectly
good
substitute
for
another"
[
and
hence
that
non­
use
values
are
trivial
or
small].
As
further
noted
by
Bateman
et
al.
(
2002,
p.
75),
"
there
are,
therefore,
no
easy
rules
for
determining
at
the
outset"
whether
non­
use
values
are
likely
to
be
significant
or
non­
significant.
Based
on
this
clear
guidance,
EPA
believes
that
empirical
analysis
should
be
used
to
determine
the
magnitude
of
non­
use
values
in
the
316(
b)
regulation
for
Phase
III
facilities.
Finally,
the
Agency
notes
that
TER
did
not
provide
complete
transcripts
for
the
15
verbal
protocol
interviews
described
in
their
comments,
preventing
the
Agency
from
obtaining
a
complete
perspective
on
the
result
of
this
effort.

The
commenter
argues
that
the
survey
does
not
sufficiently
emphasize
the
uncertainty
in
the
estimates
of
the
fish
losses
that
would
be
prevented
by
the
proposed
policies.
EPA
points
out
that
debriefing
sessions
during
focus
groups
and
cognitive
interviews
showed
that
respondents
clearly
understood
that
the
ecological
changes
described
in
the
survey
were
uncertain.
Furthermore,
respondents
were
comfortable
making
decisions
in
the
presence
of
this
uncertainty.
Additionally,
EPA
has
modified
the
survey
and
slide
show
to
emphasize
the
uncertainty
of
the
expected
environmental
changes.
For
example,
the
following
statement
has
been
added
to
the
survey:
"
You
will
be
shown
different
policy
options 
with
different
effects
on
fish.
This
is
because
scientists
are
still
working
to
determine
what
the
exact
effect
on
fish
will
be,
so
it
is
important
to
know
how
you
would
react
to
a
wide
range
of
possible
outcomes.
Common
sense
indicates
that
preventing
the
loss
of
fish
eggs
and
young
fish
will
mean
more
adult
fish
in
future
years,
but
at
this
point
there
is
still
significant
uncertainty
regarding
the
exact
size
of
these
future
effects."
EPA
also
added
debriefing
questions
to
the
survey
instrument
that
are
designed
to
identify
respondents
whose
responses
are
based
on
incorrect
interpretation
of
the
environmental
changes
described
in
the
survey,
including
the
uncertainty
of
the
expected
changes.
1
EPA
also
notes
that
tests
of
survey
question
formats
suggested
by
TER 
incorporating
ranges
of
impacts
in
survey
questions 
performed
poorly
in
both
TER's
own
focus
group
and
in
EPA's
pretests.

1
For
detail,
see
Questions
2,
3,
8,
9,
10,
and
11
in
the
survey
instrument
provided
in
Attachment
3,
U.
S.
EPA
(
2005).
Finally,
the
commenter
expresses
the
opinion
that
the
survey
should
be
peer
reviewed
by
an
independent
panel
of
experts.
EPA
agrees
with
the
commenter's
opinion,
and
the
Agency
plans
to
convene
two
peerreview
panels.
The
first
panel
will
review
the
results
of
the
focus
groups,
the
instrument
and
the
planned
survey
sampling
design,
and
the
proposed
willingness
to
pay
estimation
methodology,
before
the
survey
is
fielded.
The
second
peer
review
panel
will
review
the
entire
survey
process,
including
EPA's
final
estimated
results
for
the
316(
b)
Phase
III
rulemaking,
after
the
survey
is
completed.

Responses
to
the
Barnthouse
Comment
Introduction
The
introduction
to
the
Barnthouse
Comment
states
that
the
presentation
materials
for
the
316(
b)
survey
(
1)
overstate
both
current
impacts
of
impingement
and
entrainment
(
I&
E)
losses
and
the
biological
benefits
of
reducing
those
losses,
and
(
2)
understate
the
uncertainty
inherent
in
EPA's
estimates
of
the
magnitude
of
those
losses.
EPA
disagrees
with
these
claims.
In
response
to
the
first
point,
EPA
emphasizes
that
the
estimates
of
impingement
and
entrainment
losses
used
in
the
survey
are
scientific
estimates
specific
to
the
policy
options
valued
in
this
study.
The
policies
considered
by
EPA
in
the
context
of
the
316(
b)
regulations
would
prevent
the
loss
of
a
potentially
large
number
of
age­
one
equivalent
fish,
both
regionally
and
nationwide.
This
change
in
impingement
and
entrainment
could
lead
to
wide
range
of
biological
improvements
(
e.
g.,
effects
on
stock
size).
In
response
to
the
second
point,
EPA
notes
that
the
presentation
materials
and
the
survey
clearly
state
that
the
expected
ecological
changes
associated
with
impingement
and
entrainment
reductions
from
the
316(
b)
regulations
are
subject
to
uncertainty.
Specifically,
EPA
added
the
following
statement
to
the
slide
show
and
the
matching
script:
"
You
will
be
shown
different
policy
options 
with
different
effects
on
fish.
This
is
because
scientists
are
still
working
to
determine
what
the
exact
effect
on
fish
will
be,
so
it
is
important
to
know
how
you
would
react
to
a
wide
range
of
possible
outcomes.
Common
sense
indicates
that
preventing
the
loss
of
fish
eggs
and
young
fish
will
mean
more
adult
fish
in
future
years,
but
at
this
point
there
is
still
significant
uncertainty
regarding
the
exact
size
of
these
future
effects."
Furthermore,
given
that
there
will
be
some
biological
uncertainty
regarding
the
specifics
of
the
"
actual"
baseline
conditions
and
policy
impacts,
EPA
designed
the
survey
as
a
choice
experiment,
in
which
levels
of
attributes
will
vary
across
respondents.
The
resulting
valuation
estimates
will
allow
flexibility
in
estimating
WTP
for
a
wide
range
of
different
circumstances.

The
introduction
to
the
Barnthouse
Comment
also
notes
that
cooling
water
intake
structures
were
not
listed
as
major
concerns
in
two
recent
reports
on
threats
to
U.
S.
marine
resources
for
which
new
regulatory
policies
are
needed.
The
commenter
then
argues
that
this
implies
that
statements
in
the
survey
materials
"
that
entrainment
and
impingement
losses
are
known
to
be
causing
measurable
reductions
in
the
abundance
of
important
fish
populations,
and
that
reductions
in
those
losses
are
needed
to
prevent
continuing
long­
term
declines,"
are
false.
EPA
disagrees
with
this
statement.
The
Preamble
to
the
316(
b)
Phase
II
Rule
discusses
a
number
of
specific
examples
of
evidence
that
I&
E
at
CWIS
causes
significant
losses
of
fish
(
69
FR
41588).
For
example,
at
the
at
the
San
Onofre
Nuclear
Generating
Station,
an
estimated
57
tons
of
fish
were
killed
per
year
when
all
units
were
in
operation,
and
local
fish
densities
decreased
significantly.
Thus,
based
on
empirical
evidence
that
I&
E
at
CWIS
causes
significant
losses
of
fish,
it
is
entirely
reasonable
to
list
"
the
use
of
cooling
water
in
commercial
facilities"
as
a
factor
that
affects
fish
populations.
Moreover,
the
lack
of
emphasis
in
a
two
reports
does
not
in
any
way
imply
that
entrainment
and
impingement
losses
do
not
cause
measurable
reductions
in
the
abundance
of
important
fish
populations.

Finally,
the
introduction
points
out
that
the
survey
materials
imply
that
impingement
and
entrainment
impacts
on
freshwater
species
are
equally
severe
as
I&
E
impacts
on
marine
species.
EPA
disagrees
that
the
presentation
materials
imply
equally
severe
I&
E
impacts
on
marine
and
freshwater
species.
The
presentation
materials
clearly
emphasize
that
the
effect
of
impingement
and
entrainment
losses
on
fish
population
will
vary
depending
on
facility
location
and
that
the
estimates
presented
in
the
slide
show
discuss
average
effects
across
all
affected
water
bodies.
(
For
more
detail
see
EPA's
response
to
TER
Comments
on
Slide
13).

Specific
Comments
on
the
Slide
Show
The
commenter
provides
a
number
of
comments,
organized
by
slide.
EPA's
responses
follow
the
same
convention:


Slide
8:
The
commenter
asserts
that
commercial
and
recreational
fishing,
pollution,
and
development
of
waterfront
areas
are
"
among
the
factors
evaluated
by
the
Pew
Oceans
Commission
and
the
U.
S.
Commission
on
Ocean
Policy.
The
reports
from
these
commissions
provide
ample
documentation
of
their
harmful
effects.
No
similar
documentation
exists
for
effects
of
cooling
water
withdrawals."
EPA
disagrees
with
this
statement.
In
the
Preamble
of
the
316(
b)
Phase
II
Rule
(
69
FR
41588),
under
"
Examples
of
Environmental
Impacts
Caused
by
Cooling
Water
Intakes,"
this
passage
was
included:
"
At
the
San
Onofre
Nuclear
Generating
Station
(
SONGS),
in
a
normal
(
non­
El
Nino)
year,
an
estimated
57
tons
of
fish
were
killed
per
year
when
all
units
were
in
operation.
The
amount
lost
per
year
included
approximately
350,000
juveniles
of
white
croaker,
a
popular
sport
fish;
this
number
represents
33,000
adult
equivalents
or
3.5
tons
of
adult
fish.
In
shallow
water,
densities
of
queenfish
and
white
croaker
decreased
60
percent
within
one
kilometer
of
SONGS
and
35
percent
within
three
kilometers
from
SONGS
as
compared
to
densities
prior
to
facility
operations.
Densities
of
local
midwater
fish
decreased
50
to
70
percent
within
three
kilometers
of
the
facility."
Thus,
based
on
empirical
evidence
that
I&
E
at
CWIS
causes
significant
losses
of
fish,
it
is
entirely
reasonable
to
list
"
the
use
of
cooling
water
in
commercial
facilities"
as
a
factor
that
affects
fish
populations.


Slide
9:
The
commenter
notes
that
this
slide
does
not
point
out
that
I&
E
at
CWIS
is
not
a
primary
cause
of
the
general
decline
most
fish
stocks.
First,
EPA
notes
that
the
relative
importance
of
various
causes
of
fish
loss
varies
by
species.
Accordingly,
it
is
misleading
to
make
"
general"
statements
about
the
relative
importance
of
different
causes.
For
example,
while
overfishing
may
be
the
most
important
cause
of
decline
for
certain
highly­
prized
commercial
species,
it
is
a
negligible
cause
of
decline
for
most
forage
fish.
Second,
EPA's
focus
groups
showed
that
participants
had
a
basic
level
of
understanding
about
the
importance
of
various
factors
that
affect
fish
populations.
Most
participants
ranked
pollution
and
coastal
zone
development
as
the
main
factor
affecting
fish
population,
followed
by
commercial
and
recreational
fishing
(
for
details,
refer
to
Section
2.2.1
of
the
focus
group
report
[
Besedin
et
al.,
2005]).
EPA
also
included
a
debriefing
question
in
the
survey
that
asks
survey
respondents
to
rank
the
effect
of
various
human
and
natural
factors
in
terms
of
their
importance
to
the
decline
in
fish
populations
(
question
2,
Attachment
3,
U.
S.
EPA,
2005).
The
question
was
intended
as
a
screening
tool
that
would
identify
those
survey
respondents
who
misunderstood
the
slide
show
and
ranked
I&
E
as
a
major
factor
in
the
decline
of
fish
populations.
The
commenter
also
points
out
that
although
striped
bass
are
featured
prominently
on
a
number
of
slides,
the
population
of
this
species
has
increased
significantly
in
the
northeast
over
the
last
twenty
years.
EPA
points
out
that
pictures
of
striped
bass
are
included
in
the
slides
show
because
1.8
million
of
age­
1
equivalents
of
striped
bass
are
lost
to
impingement
and
entrainment
in
the
Northeast
region.
Whether
eliminating
impingement
and
entrainment
losses
has
any
effect
on
the
stock
size
of
this
species
will
be
determined
by
project
biologists.
The
revised
version
of
this
slide
will
include
the
estimated
change
(
if
any)
of
the
striped
bass
stock
size
from
reducing
impingement
and
entrainment
mortality
(
slide
19,
Attachment
1,
U.
S.
EPA,
2005).
Finally,
EPA
has
added
material
to
the
survey
and
slide
show
to
emphasize
that
CWIS
losses
may
not
be
the
most
significant
cause
of
fish
losses
in
many
cases.
For
example,
the
slide
show
script
now
explicitly
states,
"
It
is
important
to
recognize
that
the
effects
of
fish
losses
from
cooling
water
use
are
usually
smaller
than
the
effects
of
commercial
fishing.
Even
though
cooling
water
use
is
not
the
largest
cause
of
fish
losses
in
many
areas,
it
has
contributed
to
a
measurable
decline
in
some
fish
stocks,
in
some
locations."
This
statement
is
both
justified
based
on
scientific
evidence,
and
provides
an
appropriate
understanding
that
CWIS
losses
are
not,
in
some
cases,
the
most
significant
cause
of
fish
decline.


Slide
13:
EPA
agrees
that
the
effect
of
impingement
and
entrainment
losses
on
fish
population
will
vary
depending
on
the
baseline
condition
of
the
stock
size,
magnitude
of
impingement
and
entrainment
losses,
and
other
factors.
Therefore,
EPA
revised
slide
13
as
follows:
the
slide
shows
a
map
identifying
locations
of
major
facilities
in
the
Northeast
that
withdraw
cooling
water.
The
script
for
Slide
13
states:
"
Facilities
that
use
cooling
water
are
located
in
all
Northeastern
States.
This
map
shows
the
locations
of
major
facilities
in
the
Northeast
that
withdraw
cooling
water
in
ways
that
harm
fish.
Common
sense
indicates
that
effects
on
fish
will
be
greater
in
areas
closer
to
these
facilities,
and
smaller
in
areas
that
are
further
away.
For
the
sake
of
simplicity,
this
presentation
discusses
average
effects
across
all
affected
water
bodies"
(
Attachment
1,
U.
S.
EPA,
2005).


Slide
15:
The
commenter's
point
that
very
few
trout
are
affected
by
impingement
and
entrainment
is
well
taken.
EPA
has
removed
the
photo
of
the
trout
from
this
slide
(
slide
15,
Attachment
1,
U.
S.
EPA,
2005).


Slide
17:
EPA
disagrees
that
the
slide
show
fails
to
inform
survey
respondents
regarding
the
proportion
of
impingement
and
entrainment
losses
attributable
to
forage
fish.
Slide
16
specifically
states
that
two
thirds
of
impingement
and
entrainment
losses
are
fish
that
are
not
directly
used
by
humans
(
slide
16,
Attachment
1,
U.
S.
EPA,
2005)


Slide
18:
This
slide
has
been
removed
from
the
slide
show.


Slide
19:
EPA
has
revised
this
slide
to
compare
impingement
and
entrainment
losses
for
selected
species
used
by
humans
with
the
stock
size
and
commercial
and
recreational
harvest
for
these
species.
EPA
also
added
the
following
statement
to
the
slide
show:
"
There
are
many
billions
of
fish
that
are
not
directly
used
by
humans
in
Northeast
waters.
However,
the
exact
size
of
total
fish
stocks
for
these
species
is
not
known"
(
slides
17­
19,
Attachment
1,
U.
S.
EPA,
2005)


Slide
21:
EPA
disagrees
with
the
commenter
that
no
decline
in
fish
population
attributable
to
cooling
water
withdrawals
was
documented.
In
the
Preamble
of
the
316(
b)
Phase
II
Rule,
under
"
Examples
of
Environmental
Impacts
Caused
by
Cooling
Water
Intakes,"
this
passage
was
included:
"
At
the
San
Onofre
Nuclear
Generating
Station
(
SONGS),
in
a
normal
(
non­
El
Nino)
year,
an
estimated
57
tons
of
fish
were
killed
per
year
when
all
units
were
in
operation.
The
amount
lost
per
year
included
approximately
350,000
juveniles
of
white
croaker,
a
popular
sport
fish;
this
number
represents
33,000
adult
equivalents
or
3.5
tons
of
adult
fish.
In
shallow
water,
densities
of
queenfish
and
white
croaker
decreased
60
percent
within
one
kilometer
of
SONGS
and
35
percent
within
three
kilometers
from
SONGS
as
compared
to
densities
prior
to
facility
operations.
Densities
of
local
midwater
fish
decreased
50
to
70
percent
within
three
kilometers
of
the
facility."


Slide
24­
25:
EPA
disagrees
with
the
commenter's
interpretation
of
information
presented
in
Slide
24.
This
slide
simply
provides
an
example
of
technologies
that
may
be
implemented
to
reduce
impingement
and
entrainment
losses.
EPA
agrees
that
many
facilities
already
have
technologies
in
place
and
the
proposed
policies
will
not
result
in
further
impingement
and
entrainment
reductions
at
these
facilities.
To
avoid
any
misunderstanding
of
the
baseline
conditions,
EPA
added
the
following
statement
to
Slide
25:
"
Technologies
such
as
these
are
already
in
use
at
52%
of
facilities
that
use
cooling
water
in
the
US.
The
proposed
policies
would
require
similar
technologies
at
all
facilities
that
use
cooling
water"
(
Attachments
1
and
2,
U.
S.
EPA,
2005).
EPA
further
notes
that
the
survey
is
to
be
implemented
as
a
choice
experiment
survey.
The
goal
of
choice
experiment
methodology
is
to
obtain
choice
responses
covering
a
wide
range
of
outcomes
that
bound
(
on
the
upper
and
lower
end)
actual
possible
policy
outcomes.
Different
respondents
will
be
asked
to
make
choices
over
policy
scenarios
whose
CWIS
reductions
range
from
0%
(
no
policy)
to
98%.
This
will
provide
EPA
with
an
opportunity
for
estimating
values
for
a
wide
range
of
potential
outcomes.


Slide
31:
This
slide
has
been
removed
from
the
show.


Slide
42:
EPA
emphasizes
that
the
proposed
survey
versions
will
allow
respondents
to
choose
among
a
wide
variety
of
hypothetical
policy
options,
some
with
larger
and
other
with
very
small
changes
in
fish
populations
and
harvest.
More
generally,
the
survey
is
to
be
implemented
as
a
choice
experiment
survey,
in
which
levels
of
attributes
will
vary
across
respondents.
The
experimental
design
will
explicitly
allow
for
variation
in
baselines
and
improvements.
As
long
as
the
range
of
baseline
and
improvement
levels
present
in
the
experimental
design
bounds
the
actual
levels,
the
choice
experiment
survey
will
be
able
to
estimate
appropriate
values 
even
though
some
respondents
will
see
numbers
that
are
"
not
right."
This
is
a
fundamental
property
of
the
choice
experiment
method.
Our
goal
is
to
construct
survey
versions
showing
a
range
of
different
baseline
and
resource
improvement
levels,
where
these
levels
are
chosen
to
(
almost
certainly)
bound
the
"
actual"
levels.
Given
that
there
will
almost
certainly
be
some
biological
uncertainty
regarding
the
specifics
of
the
"
actual"
baselines
and
improvements,
the
resulting
valuation
estimates
will
allow
flexibility
in
estimating
WTP
for
a
wide
range
of
different
circumstances.
EPA
refers
the
commenter
to
the
numerous
sources
in
the
literature
that
address
the
uses
and
methods
of
the
choice
experiment
method,
including
Louviere
et
al.
(
2000),
Bateman
et
al.
(
2002),
and
Bennett
and
Blamey
(
2001).


Slide
44:
See
EPA's
response
to
TER
Comments
on
Slide
42.

Comments
on
the
Fact
Sheet
The
commenter
first
states
that
the
fact
sheet
provided
with
the
survey
questionnaire
overstates
the
potential
benefits
of
reducing
entrainment
and
impingement
losses.
The
commenter
then
lists
specific
issues
associated
with
the
information
presented
in
the
fact
sheet.
EPA
notes
that
the
fact
sheet
was
dropped
from
the
final
version
of
the
survey.
Findings
from
the
focus
groups
conducted
in
August­
September
2005
showed
that
the
slide
show
and
the
introductory
materials
in
the
survey
provided
sufficient
information
to
answer
survey
questions
and
that
the
fact
sheet
was
unnecessary.
For
example,
one
participant
stated:
"
I
found
it
unnecessary.
[ ]
I
thought
I
picked
out
enough
from
the
slide
show
in
order
to
answer
the
questions
that
were
being
asked"
[#
9/
3­
2].
Another
stated:

M:
You
were
also
handed
a
fact
sheet.
Did
you
use
the
fact
sheet?

R:
No,
I
didn't.
No.

M:
So
basically,
what
you
are
telling
me,
that
the
only
information
you
used
was
the
slide
show.

R:
Right,
exactly.
I
was
listening
to
him
giving
us,
you
know,
what
he
was
reading
from,
and
then
looking
at
the
show.
It
was
okay.

M:
And
you
felt
that
you
had
enough
information
to
respond
to
the
survey.

R:
Right,
exactly.
[#
9/
2­
11]

Therefore,
EPA
decided
not
to
include
the
fact
sheet
as
additional
reference
material.

Conclusion
The
commenter
states
that
the
informational
slide
show
and
fact
sheet
prepared
by
EPA
contain
numerous
inaccurate
statements
and
invalid
comparisons
and
that
the
survey
respondents
are
likely
to
be
misled
to
believe
that
the
benefits
of
requiring
closed­
cycle
cooling
or
other
advanced
intake
technologies
are
substantially
greater
than
is
actually
the
case.

EPA
disagrees
with
this
conclusion.
First,
EPA
emphasizes
that
the
estimates
of
impingement
and
entrainment
losses
used
in
the
survey
are
scientific
estimates
specific
to
the
policy
options
valued
in
this
study.
The
policies
considered
by
EPA
in
the
context
of
the
316(
b)
regulations
would
prevent
the
loss
of
a
potentially
large
number
of
age­
one
equivalent
fish,
both
regionally
and
nationwide.
EPA,
however,
notes
that
the
estimated
reductions
in
impingement
and
entrainment
could
lead
to
wide
range
of
biological
improvements
(
e.
g.,
effects
on
stock
size).

In
addition,
the
presentation
materials
and
the
survey
clearly
state
that
the
expected
ecological
changes
associated
with
impingement
and
entrainment
reductions
from
the
316(
b)
regulations
are
subject
to
uncertainty.
Specifically,
EPA
added
the
following
statement
to
the
slide
show
and
the
matching
script:
"
You
will
be
shown
different
policy
options 
with
different
effects
on
fish.
This
is
because
scientists
are
still
working
to
determine
what
the
exact
effect
on
fish
will
be,
so
it
is
important
to
know
how
you
would
react
to
a
wide
range
of
possible
outcomes.
Common
sense
indicates
that
preventing
the
loss
of
fish
eggs
and
young
fish
will
mean
more
adult
fish
in
future
years,
but
at
this
point
there
is
still
significant
uncertainty
regarding
the
exact
size
of
these
future
effects."

Given
that
there
will
be
some
biological
uncertainty
regarding
the
specifics
of
the
"
actual"
baselines
conditions
and
policy
impacts,
EPA
designed
the
survey
as
a
choice
experiment,
in
which
levels
of
attributes
will
vary
across
respondents.
The
proposed
survey
versions
will
allow
respondents
to
choose
among
a
wide
variety
of
hypothetical
policy
options,
some
with
larger
and
other
with
very
small
changes
in
fish
populations
and
harvest.
As
long
as
the
range
of
baseline
and
improvement
levels
present
in
the
experimental
design
bounds
the
actual
levels,
the
choice
experiment
survey
will
be
able
to
estimate
appropriate
values
corresponding
to
the
policy
outcomes.
The
resulting
valuation
estimates
will
allow
flexibility
in
estimating
WTP
for
a
wide
range
of
different
circumstances.
Focus
groups
conducted
in
August­
September
2005
showed
that
focus
group
participants
clearly
understood
that
the
316(
b)
regulations
will
have
only
a
small
impact
on
long­
term
fish
populations,
and
that
the
estimates
provided
in
the
survey
were
uncertain
(
Besedin
et
al.,
2005).
For
example,
participants
made
statements
such
as:
"
I
don't
think
a
three
percent
population
thing
with
the
fish
is
a
big
deal"
[#
9/
8­
4],
and
"
It
didn't
show
much
either
way,
to
me.
From
65
percent
to
68
percent,
maybe
in
the
big
scheme
of
things
three
percent
is
a
lot,
that
I
don't
know"
[#
9/
10­
2].

Responses
to
the
TER
Comment
1
Introduction
and
Overview
The
introduction
to
the
TER
Comment
describes
EPA's
proposed
survey,
and
then
provides
a
summary
of
later
comments.
In
the
spirit
of
these
summary
comments,
EPA
offers
the
following
summarized
responses.
Details
and
rationale
underlying
these
responses
are
provided
in
later
sections
of
this
commentary,
corresponding
to
the
appropriate
numbered
sections
of
the
TER
Comment.
The
primary
elements
highlighted
in
this
section
of
the
TER
Comment,
and
EPA's
summary
comments,
are
found
below.


Comment:
"
Respondents
are
not
valuing
marginal
changes
in
forage
fish
populations.
[ ]
The
pattern
of
such
responses
is
consistent
with
other
well­
documented
biases
(
such
as
embedding
or
payment
vehicle).
[ ]
The
fact
that
EPA's
survey
elicits,
in
many
instances,
values
other
than
the
nonuse
value
of
forage
fish
is
a
fatal
flaw."
EPA
Response:
This
statement
reveals
a
clear
misunderstanding
of
the
purpose
of
the
survey,
which
is
to
measure
total
willingness
to
pay
to
prevent
losses
of
both
forage
and
commercial
and
recreational
species.
Furthermore,
despite
the
commenter's
claims,
EPA's
pretesting
suggests
that
symbolic
biases
such
as
the
warm
glow
effect
will
not
be
a
major
concern.
Additionally,
the
survey
instrument
provides
a
means
to
test
for
and
eliminate
responses
based
on
such
motives.
More
detailed
responses
to
this
comment
are
provided
in
Section
2.1,
below.


Comment:
"
The
survey
responses
reveal
a
consistent
pattern
of
hypothetical
bias."
EPA
Response:
EPA
agrees
with
the
commenter
that
addressing
hypothetical
bias
is
an
important
survey
design
issue,
and
the
Agency
has
taken
several
steps
to
minimize
hypothetical
bias.
The
survey
has
been
explicitly
designed
to
maximize
the
consequentiality
of
choice
experiment
questions,
thereby
maximizing
incentive
compatibility,
and
it
includes
numerous
debriefing
questions
to
assess
the
rationale
underlying
the
choices
made
by
respondents,
and
to
assess
whether
hypothetical
bias
is
an
issue.
EPA
also
emphasizes
that
the
proposed
survey
follows
the
guidelines
of
the
NOAA
Blue
Ribbon
Panel
on
Contingent
Valuation
(
Arrow
et
al.
1993)
with
regard
to
mechanisms
for
reduction
of
hypothetical
bias.
Based
on
the
results
of
EPA's
focus
groups,
EPA
believes
that
these
measures
have
been
effective
at
limiting
hypothetical
bias
in
the
survey.
More
detailed
responses
to
this
comment
are
provided
in
Section
2.2,
below.


Comment:
"
The
study
results
offer
little
or
no
evidence
that
the
stated
preference
format
used
in
this
survey 
is
any
less
subject
to
the
pervasive
biases."
EPA
Response:
EPA's
focus
groups
have
shown
that
the
survey
respondents
are
aware
of
their
budget
constraints,
evaluate
the
policy
choices
realistically,
and
understand
the
environmental
changes
they
are
being
asked
to
value.
Furthermore,
the
referendum
survey
format
allows
EPA
to
evaluate
willingness
to
pay
for
a
range
of
policy
scenarios.
More
detailed
responses
to
this
comment
are
provided
in
Section
2.3,
below.

Comment:
"
The
survey
responses
deviate
from
standard
economic
principles
for
measuring
benefits
and
interpreting
survey
data.
[ ]
Respondents
are
answering
the
questions
conditional
on
the
responses
given
to
previous
questions,
which
violates
the
independence
assumption
used
in
the
standard
statistical
analysis."
EPA
Response:
EPA
believes
that
the
survey
will
generate
economically
consistent
responses.
Although
the
resource
changes
associated
with
the
policies
presented
in
the
survey
will
affect
respondents
in
different
ways,
the
three­
way
factorial
survey
format 
in
which
each
policy
is
represented
as
having
separate
effects
on
fish
losses,
long
term
populations,
and
commercial
and
recreational
catch
rates 
is
explicitly
designed
to
allow
separate
calculation
of
the
values
that
respondents
assign
to
each
aspect
of
resource
change,
both
independently,
and
though
interaction
terms.
Furthermore,
the
results
of
EPA's
focus
groups
revealed
that
most
respondents
answered
the
choice
questions
separately.
More
detailed
responses
to
this
comment
are
provided
in
Section
2.3,
below.


Comment:
"
Respondents'
answers
are
entirely
dependent
upon
the
information
provided.
[ ]
Barnthouse
(
2005)
shows
that
the
information
contained
in
the
EPA
survey
materials
is
inaccurate.
[ ]
Consequently,
any
responses
based
on
such
data
are
biased
and
not
suitable
for
inclusion
in
a
RIA."
EPA
Response:
The
data
presented
in
the
survey
represents
the
best
scientific
estimates
of
fish
losses
and
population
changes
available.
The
fact
that
respondents'
answers
are
influenced
by
this
information
provided
is
expected,
and
is
consistent
with
the
academic
literature.
As
the
commenter
points
out,
the
"
NOAA
panel
identified
the
inclusion
of
accurate
and
complete
information
in
the
CV
questionnaire
as
a
critical
characteristic
of
a
CV
survey."
More
detailed
responses
to
this
comment
are
provided
in
Section
2.4,
below.


Comment:
"
EPA
also
has
failed
to
include
information
concerning
the
inherent
uncertainty
of
the
effects
of
CWIS
on
the
environment,
which
further
limits
the
usefulness
of
the
survey
responses.
The
respondents
in
the
verbal
protocol
survey
found
it
very
difficult
to
answer
questions
that
include
uncertain
outcomes.
Such
a
finding
is
supported
by
the
recent
Resources
for
the
Future
(
RFF)
study
(
Banzhaf
et
al.
2004),
which
found
that
respondents
could
not
provide
answers
to
questions
when
uncertainty
was
present.
The
RFF
solution
to
this
problem
was
to
misrepresent
the
factual
information,
which
is
akin
to
no
solution
at
all."
EPA
Response:
EPA
disagrees
with
this
comment.
The
survey
materials
clearly
indicate
the
uncertainty
involved
in
forecasting
the
impacts
of
CWIS.
EPA's
focus
groups
demonstrated
that
survey
respondents
clearly
understood
that
the
ecological
changes
described
in
the
survey
were
uncertain.
Furthermore,
most
respondents
were
comfortable
making
decisions
in
the
presence
of
this
uncertainty.
More
detailed
responses
to
this
comment
are
provided
in
Section
2.4,
below.


Comment:
"
Many
respondents
indicated
that
they
found
the
survey
process
to
be
long,
difficult
and
confusing.
Such
a
finding
increases
the
chances
of
significant
nonresponse
bias
in
the
survey.
The
evident
confusion
in
respondents'
answers
is
yet
another
source
of
statistical
noise
that
further
lowers
the
likelihood
that
this
survey
would
yield
useful
information
for
a
RIA."
EPA
Response:
EPA
has
made
a
number
of
efforts
to
simplify
the
presentation
of
information
and
the
format
of
the
survey.
Results
from
EPA's
focus
groups
confirm
that
respondents
understood
the
questions
and
were
able
to
complete
the
survey
without
difficulty.
More
detailed
responses
to
this
comment
are
provided
in
Section
2.5,
below.


Comment:
" 
the
EPA
survey
design
does
not
try
to
determine
whether
people
value
protecting
all
fish
from
all
forms
of
predation
and
whether
the
value
of
reducing
the
impacts
of
CWIS
on
forage
fish
is
a
subset
of
that
broader
valuation.
At
a
minimum,
this
survey
presents
a
classic
illustration
of
the
conundrum
as
to
whether
respondents
have
preferences
for
reducing
the
effects
of
CWIS
on
forage
fish
or
whether
such
preferences
are
merely
an
artifact
of
the
survey
process."
EPA
Response:
Results
from
EPA's
focus
groups
indicate
that
respondents
feel
differently
about
fish
losses
they
perceive
as
"
natural",
and
fish
losses
that
are
caused
by
humans.
Furthermore,
the
focus
groups
revealed
that
many
respondents
have
large
values
for
preventing
fish
losses;
these
values
are
not
likely
to
be
an
artifact
of
the
survey
process.

2
Results
of
Verbal
Protocol
Many
of
the
claims
presented
in
the
TER
Comment
are
based
on
the
results
of
a
series
of
verbal
protocol
interviews
conducted
by
TER.
In
these
interviews,
which
were
held
in
July
2005,
fifteen
subjects
were
asked
to
"
think
aloud"
while
they
completed
a
modified
version
of
EPA's
choice
experiment
survey.
A
moderator
then
asked
a
series
of
questions
intended
to
determine
how
subjects
processed
and
reacted
to
the
information
and
survey
questions.
The
commenter
notes
that
the
results
of
such
interviews
are
qualitative,
and
appropriately,
the
TER
Comment
"
does
not
attempt
to
provide
any
statistical
analysis
of
the
responses
because
of
the
small
sample
size."
EPA
emphasizes
that
the
TER
Comments
are
based
on
interviews
with
only
fifteen
participants.
Given
such
a
small
sample,
it
is
difficult
to
generalize
these
findings
into
the
broad
claims
made
by
TER.
Moreover,
the
demographic
characteristics
of
the
TER
sample
may
not
be
representative
of
the
general
population
(
e.
g.,
2/
3
of
focus
group
participants
are
females).
This
stands
in
contrast
to
the
extensive
testing
and
improvement
of
the
survey
instrument
conducted
by
EPA,
which
involved
twelve
focus
groups
with
120
participants.
While
EPA
nonetheless
takes
these
comments
seriously,
it
believes
that
comments
of
greater
validity
and
applicability
are
generated
by
the
12
focus
groups
conducted
by
the
agency.

At
various
points
throughout
the
TER
Comment,
the
commenter
provides
"
paraphrases
of
respondents'
answers"
that
support
specific
claims
made
by
the
commenter.
Many
of
these
excerpted
comments
from
the
TER
verbal
protocol
interviews
are
of
great
interest
to
EPA.
However,
it
is
difficult
for
EPA
to
evaluate
these
excerpted
comments
properly,
given
that
most
are
provided
without
relevant
contextual
information,
such
as
the
moderator
questions
to
which
respondents
were
replying,
or
previous
statements
made
by
the
same
respondents.
EPA
would
greatly
appreciate
the
opportunity
to
examine
the
full
transcripts
from
each
interview,
including
moderator
questions.
In
the
absence
of
this
complete
information,
it
is
difficult
to
assess
the
validity
of
the
claims
that
TER
draws
from
these
transcripts 
even
within
the
small­
sample
caveats
offered
by
TER.

At
the
end
of
the
introduction
to
Section
2,
the
TER
Comment
states
that
"
the
results
from
TER's
verbal
protocol
study
show
that
the
respondents
took
the
survey
seriously.
They
spent
considerable
effort
to
both
process
the
information
provided
in
the
slide
show
and
use
that
information
in
their
attempt
to
answer
the
paired
comparisons
as
well
as
the
other
survey
questions."
This
result
is
consistent
with
the
findings
from
EPA's
focus
groups.
However,
the
comment
continues:
" 
despite
their
efforts,
a
number
of
fundamental
concerns
arose
which
seriously
undermine
EPA's
ability
to
draw
any
meaningful
or
reliable
conclusions
regarding
the
nonuse
benefits
of
the
Phase
III
316(
b)
Rule
based
on
this
survey
questionnaire."
EPA
disputes
the
accuracy
of
this
statement.
As
discussed
in
EPA's
detailed
responses,
below,
a
number
of
issues
raised
in
the
TER
Comment
reflect
a
misunderstanding
of
the
purpose
of
the
survey,
which
is
to
evaluate
total
willingness
to
pay
(
WTP)
to
reduce
losses
of
all
species
affected
by
impingement
and
entrainment 
including
forage
species
as
well
as
commercial
and
recreational
species.
Furthermore,
the
results
from
the
focus
groups
and
cognitive
interviews
carried
out
under
EPA
ICR
#
2155.01
contradict
many
of
the
findings
from
the
TER
verbal
protocol
interviews.
Without
access
to
the
full
transcripts
of
the
TER
interviews,
EPA
is
unable
to
determine
whether
this
discrepancy
is
due
to
differences
in
the
verbal
protocol
used
by
TER,
the
specific
form
of
questioning
used
by
the
TER
moderator,
random
variation
in
attitudes
of
the
small
number
of
participants
in
the
TER
interviews,
or
selective
presentation
of
results
by
the
commenter.
Given
that
EPA
conducted
twelve
focus
groups
with
120
participants,
the
Agency
prefers
to
rely
on
the
overall
weight
of
evidence,
which
suggests
that
(
1)
respondents
understand
the
information
presented,
and
respond
to
the
survey
as
they
would
a
real
vote,
(
2)
the
survey
is
not
subject
to
serious
methodological
or
informational
biases,
and
(
3)
the
survey
effectively
measures
respondents'
total
willingness
to
pay
for
reductions
in
fish
losses.
Finally,
EPA
emphasizes
that
a
review
of
the
leading
scholarly
journals
in
the
environmental
economics
field
(
e.
g.,
Journal
of
Environmental
Economics
and
Management,
Land
Economics,
American
Journal
of
Agricultural
Economics,
etc.)
clearly
reveals
that
the
extent
of
testing,
development
and
revision
given
to
the
survey
in
question
far
exceeds
that
reported
for
the
vast
majority
of
surveys
reflected
in
the
peer
reviewed
literature.
Moreover,
the
information
provided
to
respondents
is
more
extensive
than
that
in
most
surveys
reflected
in
the
literature,
and
focus
groups
indicate
that
respondents
are
able
to
process
this
information
appropriately
when
answering
choice
experiment
questions.

Despite
these
concerns
of
EPA
with
the
verbal
protocol
analysis
results
selectively
reported
by
TER,
the
Agency
nonetheless
takes
very
seriously
the
possibility
of
the
reported
biases.
As
a
result,
the
suggestions
of
the
TER
Comment
were
explicitly
explored
and
tested
in
seven
focus
groups
conducted
after
the
TER
Comment
was
received
by
the
Agency.
As
noted
below,
these
tests
showed
that
the
TER
claims,
in
the
opinion
of
the
Agency,
are
largely
unsupported
by
the
weight
of
the
evidence.

2.1
Respondents
are
not
valuing
marginal
changes
in
forage
fish
populations.

This
section
of
the
TER
Comment
begins
with
the
statement:
"
EPA
has
developed
measures
of
recreational
and
commercial
fishing
values
in
the
regulatory
impact
analysis
for
the
proposed
Phase
III
rule,
so
any
respondent
who
is
valuing
sport
or
commercial
fish
and
not
forage
fish
would
be
valuing
the
wrong
commodity."
This
statement
shows
the
commenter's
confusion
about
the
multiple
ways
in
which
commercially
and
recreationally
valuable
fish
may
influence
the
utility
of
individuals.
Specifically,
it
fails
to
recognize
that
there
are
two
legitimate
sources
of
economic
value
related
to
these
fish:
use
benefits
and
non­
use
benefits.
As
noted
in
the
TER
Comment,
commercial
and
recreational
anglers
benefit
from
catching
these
species.
However,
only
a
small
percentage
of
the
recreationally
and
commercially
valuable
fish
whose
losses
would
be
prevented
by
the
316(
b)
regulation
will
be
caught
by
anglers.
The
remaining
fish
would
enter
individuals'
utility
functions
by
providing
non­
use
benefits 
through
their
existence 
that
have
little
or
no
relationship
to
the
role
of
these
species
in
recreational
and
commercial
fisheries.
EPA
emphasizes
the
well­
known
tenet
of
economic
theory
which
states
that
benefits
or
costs
from
multiple
direct
(
primary)
effects
on
markets
should
be
counted
within
a
benefit­
cost
analysis;
this
is
not
double
counting,
and
is
indeed
appropriate
(
Just
et
al.
1982).
Comprehensive
estimates
of
total
resource
value
should
include
both
use
and
non­
use
values,
such
that
the
resulting
total
value
estimates
may
be
compared
to
total
social
cost
.
"
Non­
use
values,
like
use
values,
have
their
basis
in
the
theory
of
individual
preferences
and
the
measurement
of
welfare
changes.
According
to
theory,
use
values
and
non­
use
values
are
additive"
(
Freeman,
2003,
p.
154).

Both
the
market
valuation
and
revealed
preference
methods
used
to
estimate
benefits
of
improved
commercial
and
recreational
harvest
allow
EPA
to
estimate
direct
use
values
only.
These
methods
are
not
suitable
for
estimating
total
values
(
use
and
nonuse)
of
harvested
species
and
therefore
would
underestimate
total
value
(
Freeman
2003).
Both
recreational
and
commercial
and
forage
fish
attributes
may
enter
individuals'
utility
functions
directly,
providing
non­
use
benefits
that
bear
little
or
no
relationship
to
(
and
are
not
subsumed
by)
use
benefits
generated
by
recreational
and
commercial
fish.
EPA
refers
to
Freeman
(
2003),
who
provides
a
clear
distinction
between
use
benefits
(
indirect
and
direct)
and
true
non­
use
benefits.
Finally,
EPA
notes
that
nearly
all
(
96
percent)
of
impingement
and
entrainment
losses
at
CWIS
consist
of
either
forage
species,
or
non­
landed
recreational
and
commercial
species
that
do
not
have
direct
uses
or,
as
a
result,
direct
use
values.
Therefore,
EPA
believes
that
(
1)
only
a
small
fraction
of
potential
benefits
associated
with
the
316(
b)
regulation
is
quantified
in
EPA's
benefit
cost
analysis
of
the
proposed
316(
b)
regulation
for
the
Phase
III
facilities
and
(
2)
developing
comprehensive
quantified
benefit
estimates
for
the
section
316(
b)
regulation
requires
consideration
of
non­
use
values.

The
TER
Comment
also
argues
that
"
in
order
to
empirically
measure
these
nonuse
values,
survey
respondents
must
be
able
to
conceive
of
a
value
of
forage
fish
for
its
own
sake
and
to
be
able
to
separate
such
values
from
recreation
or
commercial
use
values."
This
statement
reveals
a
clear
misunderstanding
of
the
purpose
of
the
survey,
which
is
to
measure
total
willingness
to
pay
to
prevent
losses
of
both
forage
and
commercial
and
recreational
species.
The
referendum
format
of
the
survey
allows
respondents
who
do
not
hold
non­
use
values
for
fish
to
choose
policy
options
that
reflect
their
preferences;
respondents
need
not
value
the
existence
of
fish,
nor
do
they
need
to
be
able
to
separate
their
use
and
non­
use
values.
In
fact,
the
peer­
reviewed
literature
is
highly
skeptical
of
stated
preference
surveys
that
attempt
to
decompose
use
and
nonuse
values,
or
to
measure
only
nonuse
values
(
for
example,
see
Cummings
and
Harrison
1995).
This
is
the
rationale
behind
the
estimation
of
nonuse
values
as
the
total
willingness
to
pay
(
WTP)
of
nonusers
(
Cummings
and
Harrison
1995;
Johnston
et
al.
2003).
A
primary
advantage
of
this
approach
is
the
general
conceptual
acceptance
of
nonuser
values,
even
among
those
highly
critical
of
surveys
that
attempt
to
isolate
nonuse
values
among
the
general
population
(
Cummings
and
Harrison
1995).
In
early
survey
versions
tested
in
focus
groups,
EPA
attempted
to
assess
values
for
only
one
subset
of
fish
(
e.
g.,
only
forage
fish).
Such
surveys
were
protested
by
respondents,
and
were
often
received
with
confusion.
For
example,
respondents
had
difficulty
understanding
how
regulations
designed
to
prevent
fish
losses
in
CWIS
could
only
"
target"
species
with
no
human
uses.
As
a
result
of
both
the
above
arguments,
EPA
feels
that
the
current
conceptual
approach 
assessing
values
for
the
full
range
of
fish
affected
by
the
316b
rule 
is
both
the
most
appropriate
and
most
practical
approach
towards
stated
preference
value
estimation
in
this
case.

Finally,
the
Agency
notes
that
these
comments
are
nearly
identical
to
those
already
addressed
by
EPA
in
prior
responses
to
TER.
In
these
prior
responses,
EPA
noted
that
the
TER
assertion,
"
EPA
should
focus
only
on
the
nonuse
value
of
fish,"
contradicts
both
the
economic
literature
and
its
own
statement
on
page
7,
"
telling
respondents
to
think
only
about
the
nonuse
part
will
not
work."
As
noted
above,
"
what
matters
for
policy
purposes
is
total
value,
regardless
of
how
it
is
divided
between
use
and
non­
use
values"
(
Freeman
2003,
p.
154).
It
is
generally
recognized
in
the
economic
literature
is
preferable
to
estimate
total
value
of
a
given
resource
rather
that
to
estimate
components
of
the
total
value
(
e.
g.,
indirect
use
vs.
non­
use
values)
and
then
aggregate
them.
EPA
further
notes
that
both
estimates
of
use
values
associated
with
improved
commercial
and
recreational
harvest
and
the
total
value
of
reduced
fish
losses
due
to
the
316(
b)
regulation
provide
valuable
information
to
decision
makers.

2.1.1
Respondents
are
valuing
recreationally
or
commercially­
caught
fish,
not
forage
fish.

The
commenter
argues
that
that
because
participants
in
TER's
verbal
protocol
study
frequently
mentioned
recreational
and
commercial
species,
individuals
do
not
hold
significant
nonuse
values
for
forage
fish.
EPA
disagrees
with
this
statement.
First,
EPA
points
out
that
the
commenter's
agument
is
based
on
a
logical
fallacy.
Absence
of
evidence
that
participants
value
forage
fish
is
not
equivalent
to
the
presence
of
evidence
that
participants
do
not
value
forage
fish.
This
is
particularly
relevant
here,
given
that
TER
has
not
provided
a
complete
transcript
of
their
focus
group
results 
a
full
transcript
which
might
well
provide
evidence
of
values
for
forage
fish.
Second,
many
participants
in
the
focus
groups
conducted
by
EPA
stated
clearly
that
they
held
significant
non­
use
values
for
forage
species.
For
example,
when
asked
what
types
of
fish
they
were
imagining
as
they
filled
out
the
questionnaire,
participants
gave
responses
such
as:
"
All
of
them:
little
minnows
to
big
whales,"
"
I
am
more
worried
about
the
types
of
fish
that
are
not
directly
used
by
us,"
and
"
I
think
about
them
as
being
living
creatures."
When
directly
questioned
in
both
written
debriefing
questions
and
verbal
focus
group
questions,
the
majority
of
respondents
indicated
that
their
survey
responses
reflected
values
for
all
fish,
including
forage
fish.
EPA
notes
that
one
possible
explanation
for
the
discrepancy
between
the
results
of
the
TER's
verbal
protocol
interviews
and
EPA's
focus
groups
is
that
seven
of
the
fifteen
individuals
in
the
TER
study
stated
that
they
participate
in
recreational
angling.
Inclusion
of
such
a
disproportionately
high
percentage
of
anglers 
who
presumably
hold
large
use
values
for
species
affected
by
the
proposed
policies 
would
naturally
result
in
a
higher
percentage
of
responses
mentioning
recreational
species.
However,
the
fundamental
fallacy
of
TERs
argument
is
grounded
on
the
fact
that
frequent
mention
of
recreational
and
commercial
fish
among
their
respondents
in
no
way
implies
that
values
for
forage
fish
are
negligible.

2.1.2
Respondents
are
giving
a
value
for
improving
the
environment,
not
forage
fish.

In
this
section,
the
commenter
expresses
a
concern
that
the
survey
may
measure
"
the
feeling
of
satisfaction
of
helping
to
improve
the
environment,
and
not
the
value
of
the
enhanced
fish
populations
per
se.
Based
on
the
results
of
verbal
protocols,
respondents
did
report
what
amounts
to
warm
glow
as
a
reason
why
they
chose
particular
programs
in
the
paired
comparisons."
As
evidence,
the
commenter
provides
several
statements
made
by
respondents
about
their
motivations
for
answering
the
survey.
EPA
notes
that
not
all
of
the
statements
support
the
hypothesis
of
a
warm
glow
effect.
For
example,
the
participant
who
stated
that
"
I
have
a
conscience.
Don't
want
to
see
anything
wiped
out,"
was
clearly
thinking
about
fish
mortality,
not
about
a
general
desire
to
benefit
the
environment.
EPA
also
notes
that
the
lack
of
full
transcripts
from
TER's
focus
group
is
important
here,
as
the
critical
issue
when
assessing
symbolic
effects
is
whether
these
effects
contribute
to
specific
answers
to
stated
preference
responses,
not
whether
such
issues
are
mentioned
casually
by
respondents
in
focus
group
discussions.

More
importantly,
the
evidence
from
EPA's
focus
groups
and
cognitive
interviews
shows
that
the
warm
glow
effect
is
not
a
significant
source
of
bias
in
the
survey.
Many
participants
specifically
stated
that
their
responses
to
the
choice
questions
were
based
on
the
value
of
preserving
fish,
not
on
a
general
desire
to
help
the
environment
or
to
send
an
environmental
message.
Some
participants
felt
that
fish
were
worth
saving
for
their
own
sake.
Others
felt
that
preserving
fish
was
important
because
changes
in
fish
populations
can
affect
aquatic
ecosystems,
anglers,
and
fish
markets.
For
example,
one
participant
stated,
"
You
know,
if
you
were
telling
me
you
were
saving
spiders,
who
cares?
You
know
what
I
mean?
So,
it
is
not
just
to
do
the
right
thing.
I
think
that
fish
are
important."

The
commenter
continues
by
quoting
one
participant
who
expressed
hesitation
over
paying
money
to
reduce
fish
losses
when
there
are
other
important
issues
that
she
felt
strongly
about.
The
commenter
argues
that
her
statements
demonstrate
how
respondents'
answers
"
do
not
conform
to
the
basic
economic
assumptions
that
are
required
to
measure
benefits."
EPA
disagrees
with
this
point.
As
provided
in
the
TER
Comment,
the
participant's
answers
do
not
reveal
any
inconsistencies.
Instead,
her
responses
show
a
strong
awareness
of
her
own
budget
constraint 
a
critical
point
for
any
stated
preference
survey.
The
commenter
also
cites
the
following
statement
about
the
programs
for
which
the
respondent
stated
that
she
was
willing
to
pay:
"
Air
pollution,
something
having
to
do
with
forests 
not
cutting
down
forests,
what
else?
I
would
be
for
polices
like
this
or
anything
having
to
do
with
preserving
wildlife
or
animals
or
fish
and
plant
life.
Education
would
definitely
be
a
top
issue."
This
comment
shows
that
the
respondent
is
aware
of
a
variety
of
environmental
and
social
programs
that
the
government
may
support
and
that
she
has
well
defined
preferences
for
these
programs.
EPA
notes
that
survey
respondents
are
expected
to
answer
survey
questions
based
on
the
specific
information
provided
and
their
own
preferences.
Thus,
the
provided
excerpts
shows
that
respondents
are
likely
to
vote
differently
for
different
programs
and
that
their
responses
do
conform
to
the
basic
economic
assumptions
required
to
measure
benefits
(
i.
e.,
responses
are
based
on
well
defined
preferences
and
take
the
budget
constraint
into
account).

On
page
9,
the
TER
Comment
states
that
"
respondents
are
likely
to
consider
that
their
valuation
responses
would
affect
either
other
fish
populations
besides
those
that
would
be
described
in
the
contingent
valuation
(
CV)
survey
or
other
environmental
factors
such
as
habitat
quality
that
limit
fish
populations."
The
commenter
then
provides
excerpts
from
statements
made
by
respondents.
EPA
notes
that
three
of
the
four
statements
do
not
support
the
commenter's
argument:


One
participant
in
the
TER
verbal
protocol
interviews
stated
that:
"
It's
important
to
protect
the
ecosystem.
Fish
are
a
part
of
that
[
ecosystem].
If
you
kill
all
the
fish
in
a
lake,
the
lake
is
going
to
change
a
lot.
And
other
plants
or
animals
in
the
lake
might
not
be
able
to
survive."
This
statement
clearly
demonstrates
the
participant's
understanding
that
fish
are
part
of
a
larger
aquatic
ecosystem,
and
that
fish
losses
may
harm
that
ecosystem.
While
the
phrase
" 
kill
all
the
fish
in
a
lake "
is
inconsistent
with
the
survey
scenario,
it
is
not
clear
from
the
excerpt
whether
the
respondent
actually
believed
that
the
proposed
policies
would
have
this
effect,
or
whether
the
respondent
was
simply
using
an
extreme
example
to
illustrate
a
point
to
the
moderator.


Another
participant
stated:
"
I
think
it's
to
determine
what
we
should
do
as
far
as
saving
the
fish.
We
did
lose
a
lot
with
that
oil
spill.
That
one
in
Alaska,
and
we
had
one
somewhere
else
not
long
ago."
This
statement
provides
no
evidence
that
the
respondent
was
imagining
that
the
policies
would
affect
anything
except
fish
populations.
Instead,
it
shows
that
the
respondent
is
aware
of
previous
situations
in
which
fish
have
been
killed.


The
fourth
participant
stated:
"
Coastal
development
in
my
lifetime
has
hurt
fishing
a
whole
lot.
You
can't
catch
fish
beside
a
high
rise
condominium.
Part
of
fishing
is
environment
(
the
fisherman's
surroundings).
I
have
a
better
experience
on
an
island
fishing
than
at
Virginia
Beach
or
Myrtle
Beach.
Now
it
is
row
after
row
of
box
buildings.
You
don't
want
to
fish
beside
the
storm
water
sewer
outlet.
It
degrades
the
whole
experience."
This
statement
reflects
the
participant's
opinion
about
the
negative
effects
of
coastal
development,
but
at
no
point
does
the
participant
indicate
that
cooling
water
intake
structures
are
the
cause
of
this
problem,
or
that
the
proposed
survey
policies
would
address
coastal
development.

Thus,
only
one
of
the
responses
provided
in
the
TER
Comment
provides
evidence
that
a
participant
misunderstood
the
effects
of
the
proposed
policies.
This
is
consistent
with
the
results
from
EPA's
focus
groups,
in
which
the
majority
of
respondents
indicated
that
they
completed
the
survey
questions
based
solely
on
the
impingement
and
entrainment
changes
described
in
the
survey.
Although
a
few
respondents
indicated
that
they
thought
that
the
proposed
policies
might
also
reduce
pollution
or
have
other
environmental
effects,
most
respondents
considered
only
changes
related
to
fish
losses,
fish
populations,
and
commercial/
recreational
catch
rates.
For
example,
when
asked
what
kind
of
environmental
changes
would
occur
as
the
result
of
the
policies,
respondents
answered,
"
Just
fish.
And
the
fishing
industry;
it's
not
just
the
fish.
But
that's
about
it"
(#
8­
7),
and
"
I
understood
it
as,
this
would
only,
the
only
effect
of
this
would
be
to
increase
the
catch
for
local
fisheries"
(#
9/
5­
4).
Furthermore,
EPA
has
added
the
following
statement
to
the
survey:
"
Sometimes
people
taking
this
survey
imagine
that
the
policies
being
discussed
would
affect
such
things
as
water
quality,
pollution,
or
other
environmental
issues.
They
will
not.
The
only
environmental
effect
of
these
regulations
is
on
the
quantity
of
fish
in
the
water."
Testing
of
this
statement
in
later
focus
groups
indicated
that
it
was
effective
at
eliminating
any
confusion
about
the
results
of
the
proposed
policies.

However,
while
finding
that
the
preponderance
of
evidence
does
not
point
to
the
substantial
"
warm
glow"
biases
suggested
by
TER,
EPA
does
take
the
possibility
seriously.
As
a
result,
the
revised
survey
contains
numerous
debriefing
questions
to
identify
explicitly
those
answering
based
on
symbolic
or
warm
glow
considerations.
For
example,
there
are
four
questions
in
the
revised
survey
(
questions
9,
11,
12
and
13)
that
are
designed,
among
other
things,
to
identify
respondents
answering
based
on
symbolic
considerations
rather
than
based
on
a
personal
value
for
fish.
2
In
addition,
the
survey
slide
show
materials
provide
extensive
text
to
focus
respondents
on
specific
changes
to
fish
attributes,
and
to
eliminate
considerations
of
broader
environmental
issues.
For
example,
the
slide
show
script
states:
"
Now,
before
you
answer
the
survey
questions,
there
are
a
few
very
important
things
that
need
to
be
emphasized.
First,
sometimes
people
taking
this
survey
imagine
that
the
policies
being
discussed
affect
such
things
as
water
quality,
pollution,
or
other
environmental
problems.
They
will
not.
The
only
significant
environmental
effect
of
these
regulations
is
on
the
quantity
of
fish
in
the
water.
There
are
no
other
significant
effects
on
the
environment."
Focus
groups
and
cognitive
interviews
clearly
indicate
that
this
text
was
effective
at
minimizing
symbolic
responses.
Hence,
although
EPA's
pretesting
suggests
that
symbolic
biases
will
not
be
a
major
concern,
the
survey
instrument
provides
a
means
to
test
for
and
eliminate
responses
based
on
such
motives.

2.2
Respondents'
answers
show
substantial
hypothetical
bias.

This
section
of
the
TER
Comment
begins
with
a
description
of
the
well­
known
problem
of
hypothetical
bias.
The
comment
then
states,
"
Respondents
in
the
verbal
protocol
survey
explicitly
recognized
this
difference
between
what
they
said
they
would
do
in
the
survey
and
what
they
currently
do
or
would
actually
do
in
the
future."
The
comment
follows
with
several
statements
from
participants
whose
answers
seem
to
indicate
potential
for
hypothetical
bias.
EPA
agrees
with
the
commenter
that
addressing
hypothetical
bias
is
an
important
survey
design
issue.
Thus,
EPA
has
taken
several
steps
to
minimize
hypothetical
bias:

1.
The
survey
has
been
explicitly
designed
to
maximize
the
consequentiality
of
choice
experiment
questions,
thereby
maximizing
incentive
compatibility
(
i.
e.,
reducing
strategic
and
hypothetical
biases)
(
Carson
et
al.
1999;
Johnston
and
Joglekar
2005).
Elements
specifically
designed
to
maximize
consequentiality
include
a]
explicit
mention
of
the
agency
involved,
b]
explicit
mention
that
this
survey
is
associated
with
considerations
of
actual
policies
that
are
being
considered,
c]
numerous
details
provided
in
the
slide
show
and
survey
concerning
the
proposed
policies,
d]
emphasis
that
some
sort
of
policy
WILL
be
enacted 
and
that
the
type
of
policy
enacted
will
depend
in
part
on
survey
results.
The
consequentiality
and
policy
implications
of
the
survey
are
highlighted
in
both
the
survey
script
and
the
slide
show.
In
sum,
respondents
are
presented
with
a
situation
in
which
they
are
very
much
aware
that
their
survey
responses
will
influence
impending
policy
decisions 
which
has
been
shown
to
ameliorate
problems
with
hypothetical
bias
in
past
work.
Evidence
from
focus
groups
suggests
that
most
2
For
detail,
see
Questions
9,
10,
11,
and
12
in
the
survey
instrument
provided
in
Attachment
3,
U.
S.
EPA
(
2005).
respondents
feel
that
the
survey
is
consequential,
and
that
they
are
responding
as
if
voting
in
a
binding
referendum.
EPA
believes
that
the
emphasis
on
the
budget
constraint,
the
consequentiality
of
the
survey,
and
realism
of
the
payment
mechanism
is
as
extensive
as
virtually
any
survey
reflected
in
the
peerreviewed
literature
to
date.
Moreover,
the
survey
as
currently
specified
clearly
meets
the
guidelines
of
the
NOAA
Blue
Ribbon
Panel
on
Contingent
Valuation
(
Arrow
et
al.
1993)
with
regards
to
the
emphasis
given
to
the
budget
constraint.
2.
The
survey
includes
numerous
debriefing
questions
to
assess
whether
hypothetical
bias
is
an
issue.
For
example,
the
survey
includes
the
question:
"
If
Questions
3­
5
were
asked
in
a
real
referendum,
how
certain
would
you
feel
about
the
answers
you
provided?
(
circle
one)".
Focus
groups
indicate
that
this
question
is
quite
successful
at
identifying
respondents
for
whom
hypothetical
bias
is
an
issue.
Previous
research
has
shown
that
such
"
certainty"
questions
can
be
used
to
mitigate
hypothetical
bias
(
Champ
et
al.
2004).
Pretests
regarding
this
question
reveal
that
the
majority
of
respondents
feel
very
certain
of
their
stated
preference
responses,
and
indicate
that
they
would
answer
the
same
way
in
a
binding
referendum 
a
key
means
of
addressing
the
criterion
validity
of
stated
preference
responses
(
Vossler
and
Kerkvliet
2003).
3.
The
survey
also
includes
debriefing
questions
that
assess
the
rationale
underlying
the
choices
made
by
respondents
(
why
they
answered
the
way
that
they
did),
and
to
assess
whether
symbolic
or
hypothetical
bias
issue
are
present.
These
questions
are
detailed
in
the
draft
survey
instrument.

Because
of
these
measures,
EPA
does
not
believe
that
hypothetical
bias
will
be
a
significant
issue
for
this
survey.
This
conclusion
is
supported
by
the
results
of
EPA's
focus
groups
and
cognitive
interviews.
Participants
took
the
survey
questions
seriously,
as
illustrated
by
the
following
excerpt
from
one
of
the
focus
groups:

M:
Well,
when
you
answered
this
survey,
did
you
consider
that
these
results
might
actually
be
used
by
the
government?

[
voices
of
assent]

M:
Did
that
affect
how
you
answered
the
survey?

R:
No,
I
think
I'd
answer
it
the
same.
If
this
is
just
an
exercise
to
see
how
I
would
react,
I'd
still
answer
the
same
as
if
the
government
really
used
it
to
create
policy.
The
reason
I
used
it,
I
would
use
no
matter
what,
no
matter
who's
going
to
use
the
information,
yes.

M:
But
do
you
feel
that
the
survey
will
have
an
impact
on
actual
policies?

R:
We
hope
so.

[
others
repeat]

R:
It
would
be
a
waste
of
our
time
if
it
doesn't.
[#
8­
8]

Although
some
participants
had
mixed
opinions
about
whether
their
responses
would
actually
affect
government
policy,
most
indicated
that
they
would
have
chosen
the
same
answers
if
the
survey
had
been
part
of
a
real
vote.
For
example,
one
participant
stated,
"
I
take
it
seriously
but
I
have
my
doubts
whether
it
will
affect
government
policy"
(#
6­
16).
Even
those
participants
who
did
not
believe
that
their
answers
would
have
any
effect
said
that
they
expressed
their
opinion
honestly.
For
example:

M:
And
I
guess
the
last
thing
before
I
let
you
go
is,
did
you
think
that 
when
you
were,
you
know 
obviously
we're
testing
the
survey
today.
But
were
you
to
get
this
and
take
it,
would
you
think
that
your
answers
would
actually
affect
anything?
Or
would
this
just
be
a
survey,
okay,
it's
another
survey?
Do
you
think
that
your
answers
would
make
a
difference?
Were
you
thinking
that
someone
would
actually
make
use
of
the
answers?

R:
No.
R:
Not
mine,
anyway.
Maybe,
"
Okay,
his
is
not
what
we're
looking
for."
But
like
you
say,
there's
no
wrong
or
right
answers.
I
was
just
giving
honest
answers.

R:
I
think
that
it
gave
you
the
impression
that
they
will
take
this
seriously
the
way
they
said,
regarded
as
a
real
vote.
So
it
gives
you
that
impression
that
they
are
going
to
use
it.
[#
10­
18]

The
TER
Comment
also
argues
that
"
the
verbal
protocol
results
revealed
that
many
respondents
either
did
not
realize
that
they
could
choose
neither
option
(
opt
out)
or
felt
some
social/
moral
responsibility
not
to
do
so."
This
argument
is
not
supported
by
EPA's
observations
during
the
focus
groups
conducted
under
ICR
#
2155.01.
In
these
focus
groups,
many
respondents
indicated
that
they
would
not
vote
for
the
proposed
policies.
For
example:

R:
If
I
had
to
make
the
choice
between
the
three
options,
that's
unfortunately,
if
I
was
going
to
go
into
a
voting
booth 

M:
That's
how
you'd
vote.

R:
I
would
not
vote
for
any
of
them.
[#
9/
8­
8]

The
section
of
the
TER
Comment
concludes
with
a
summary
of
results
from
previous
studies
of
hypothetical
bias,
and
a
discussion
of
the
magnitude
of
such
bias
in
the
current
survey.
EPA
notes
that
such
discussion
is
highly
speculative.
EPA
emphasizes
that
these
arguments
in
the
TER
Comment
represent
largely
unsupported
empirical
claims
that
were
clearly
and
repeatedly
addressed
within
EPA's
focus
group
and
survey
design
process.
EPA
emphasizes
that
the
peer
reviewed
literature
provides
very
few
examples
of
surveys
provided
with
survey
design
and
testing
as
extensive
as
the
survey
proposed
by
EPA
for
the
Phase
III
316(
b)
analysis.
Given
that
the
results
of
EPA's
focus
groups
indicated
that
most
survey
respondents
answered
the
questions
as
honestly
as
they
could 
as
if
they
were
really
voting
in
a
binding
referendum 
the
Agency
does
not
believe
that
hypothetical
bias
will
significantly
affect
the
results
of
the
current
survey.
As
one
participant
stated:
"
I
think
I'd
answer
it
the
same.
If
this
is
just
an
exercise
to
see
how
I
would
react,
I'd
still
answer
the
same
as
if
the
government
really
used
it
to
create
policy.
The
reason
I
used
it,
I
would
use
no
matter
what,
no
matter
who's
going
to
use
the
information"
[#
8­
8].

2.3
Respondents'
answers
violate
key
economic
assumptions.

In
this
section
of
the
TER
Comment,
the
commenter
argues
that
"
the
manner
in
which
the
respondents
answer
the
questions
must
be
conceptually
consistent
with
the
way
EPA
intends
them
to
be
answered.
Any
substantial
divergence
between
the
respondents'
answers
and
the
key
economic
assumptions
used
in
the
statistical
analysis
of
the
data
will
either
bias
the
estimated
benefits
or
create
such
a
large
amount
of
statistical
noise
that
the
results
are
not
economically
meaningful."
While
EPA
agrees
that
there
should
be
general
correspondence
between
the
survey
scenario
and
the
actual
316(
b)
scenario,
exact
correspondence
is
not
necessary,
and
in
fact,
may
not
be
desirable.
As
is
common
in
surveys,
scenarios
must
be
presented
in
simplified
form
to
facilitate
respondent
comprehension,
and
to
encourage
respondents
to
focus
on
the
most
important
policy
characteristics,
balancing
the
desire
to
provide
detailed
policy
information
with
respondents'
cognitive
abilities
to
simultaneously
consider
a
large
number
of
attributes
(
Louviere
et
al.
2000).
Policy
attributes
that
are
suppressed
include
the
specific
technology
used
to
prevent
fish
losses
and
the
exact
location
of
cooling
water
intake
structures.
As
an
additional
note,
EPA
emphasizes
that
the
level
of
"
statistical
noise"
suggested
by
TER
will
be
reflected
in
the
variance
of
model
parameter
estimates.
This
will
influence
the
statistical
significance
of
WTP
measures,
but
will
not,
as
suggested
by
TER,
render
results
"
not
economically
meaningful."
The
confusion
between
theoretical
appropriateness
and
statistical
variance
is
an
additional
flaw
in
TER's
reasoning.

The
commenter
states
that
"
information
obtained
through
the
verbal
protocol
indicates
that
many
of
the
respondents
are
valuing
vastly
different
commodities
both
from
what
the
EPA
survey
intends
and
from
one
another."
The
commenter
then
provides
several
examples
of
different
environmental
changes
for
which
respondents
in
the
interviews
expressed
values,
including
improvements
in
commercial
and
recreational
fishing,
reductions
in
losses
of
fish,
and
increases
in
overall
fish
populations.
EPA
agrees
that
it
likely
that
the
resource
changes
associated
with
the
policies
presented
in
the
survey
will
affect
respondents
in
different
ways.
However,
the
three­
way
factorial
survey
format 
in
which
each
policy
is
represented
as
having
separate
effects
on
fish
losses,
long
term
populations,
and
commercial
and
recreational
catch
rates 
is
explicitly
designed
to
allow
separate
calculation
of
the
values
that
respondents
assign
to
each
aspect
of
resource
change,
both
independently,
and
though
interaction
terms.
Thus,
EPA
believes
that
opinions
expressed
by
participants
in
the
TER
verbal
protocol
interviews
are
consistent
with
the
purpose
and
format
of
the
survey.

TER's
comments
suggest
a
lack
of
understanding
of
the
capacity
of
choice
experiments
to
isolate
the
marginal
effects
of
different
attributes
on
stated
choices
and
hence
on
WTP 
a
methodology
clearly
described
by
numerous
sources
(
e.
g.,
Louviere
et
al.
2000).
For
example,
TER
states
that
"
some
respondents
are
valuing
improvements
in
recreational
and
commercial
fishing 
while
others
are
simply
valuing
reductions
in
the
number
of
fish
killed."
EPA
is
not
at
all
surprised
by
these
findings 
indeed
this
is
the
specific
reason
that
the
choice
experiment
survey
distinguishes
attributes
for
recreational
and
commercial
fishing
and
total
quantity
of
fish
losses.
The
survey
is
specifically
designed
to
address
the
patterns
observed
by
TER.
The
fact
that
TER
finds
such
patterns
in
conflict
with
appropriate
stated
preference
value
estimation
belies
a
lack
of
understanding
of
current
methods
for
choice
experiments,
reflected
in
Louviere
et
al.
(
2000)
and
Bennett
and
Blamey
(
2001),
among
others.

The
commenter
also
argues
that
the
results
of
the
survey
should
not
be
used
to
estimate
national
WTP
to
prevent
fish
losses
because
(
1)
respondents
are
thinking
about
local
water
resources,
and
(
2)
respondents
to
the
survey
may
not
actually
live
near
a
Phase
III
facility.
This
comment
reflects
a
misunderstanding
of
the
survey
statistical
sampling
design.
In
response
to
the
first
point,
EPA
notes
that
the
survey
will
allow
calculation
of
different
WTP
values
for
respondents
in
different
regions
of
the
country.
Such
geographic
differentiation
will
allow
EPA
to
control
for
regional
differences
in
water
resources,
fish
species,
and
CWIS
impacts.
In
response
to
the
second
point,
the
Agency
notes
that
the
survey
will
be
implemented
using
a
national
probability
sample.
Regardless
of
whether
respondents
actually
live
near
a
Phase
III
facility,
their
answers
will
reveal
information
about
their
preferences
for
saving
fish.
However,
when
EPA
estimates
non­
use
benefits
in
the
context
of
the
316(
b)
regulation
for
Phase
III
facilities,
EPA
will
estimate
non­
use
benefits
only
for
the
relevant
population
of
households
that
are
likely
to
hold
non­
use
values
for
losses
attributable
to
Phase
III
facilities.

On
page
14,
the
commenter
again
demonstrates
a
clear
misunderstanding
of
the
purpose
of
the
survey
by
stating:
"
Moreover,
to
the
extent
that
users
are
expressing
a
use
value,
the
relevance
of
such
a
value
to
the
nonuse
value
of
marginal
changes
in
forage
fish
populations
is
questionable,
at
best."
EPA
points
out
that
the
purpose
of
the
survey
is
to
measure
total
WTP
for
the
changes
discussed
in
the
survey 
including
both
use
and
non­
use
values.
When
the
results
of
the
survey
are
used
to
estimate
benefits
in
the
context
of
the
316(
b)
regulation,
EPA
will
approximate
non­
use
values
associated
with
the
policy
changes
based
on
total
WTP
of
non­
users 
a
method
characterized
as
" 
one
of
the
more
defensible
means
of
distinguishing
use
and
nonuse
values
in
a
stated
preference
context"
(
Johnston
et
al.
2005).
A
primary
advantage
of
this
approach
is
the
general
conceptual
acceptance
of
nonuser
values,
even
among
those
highly
critical
of
surveys
that
attempt
to
isolate
nonuse
values
among
the
general
population
(
Cummings
and
Harrison
1995).
In
early
survey
versions
tested
in
focus
groups,
EPA
attempted
to
assess
values
for
only
one
subset
of
fish
(
e.
g.,
only
forage
fish).
Such
surveys
were
protested
by
respondents,
and
were
often
received
with
confusion.
For
example,
respondents
had
difficulty
understanding
how
regulations
designed
to
prevent
fish
losses
in
CWIS
could
only
"
target"
species
with
no
human
uses.
As
a
result
of
both
the
above
arguments,
EPA
feels
that
the
current
conceptual
approach 
assessing
values
for
the
full
range
of
fish
affected
by
the
316b
rule 
is
both
the
most
appropriate
and
most
practical
approach
towards
stated
preference
value
estimation
in
this
case.

Additional
comments
by
TER
in
this
area
suggest
additional
confusion
among
the
commenter
with
regard
to
the
appropriate
measurement
of
values
in
the
316(
b)
policy
case.
For
example,
TER
states
"
even
respondents
who
are
not
users
of
the
resource
may
still
be
valuing
recreational
and
commercial
species,"
as
an
argument
that
the
valuation
proposed
by
EPA
is
conceptually
flawed.
As
noted
above,
however,
it
is
perfectly
valid
for
nonusers
to
maintain
nonuse
values
for
recreational
and
commercial
fish
that
are
in
addition
to
the
use
values
that
might
be
realized
by
users.
That
is
fact
is
not
noted
by
the
commenter
suggests,
again,
a
fundamental
flaw
in
reasoning
and
understanding
of
accepted
conceptual
models
for
benefit
cost
analysis
(
Boardman
et
al.
2001).

Finally,
the
commenter
stated
that
the
TER
verbal
protocol
interviews
showed
that
respondents
were
not
answering
the
questions
independently.
EPA
takes
this
concern
seriously,
and
has
hence
explicitly
assessed
the
potential
for
non­
independence
in
focus
group.
Based
on
the
results
of
these
tests,
EPA
finds
little
evidence
that
this
is
a
serious
problem
in
the
current
survey.
As
shown
in
the
following
excerpt
from
the
Agency's
focus
groups
and
cognitive
interviews,
most
respondents
considered
the
choice
questions
independently.

M:
When
you
answered
these
questions
three,
four,
and
five,
were
you
answering
them
separately
or
did
you
compare
them
across
 

R:
No,
separately.
[#
11/
2­
9]

Moreover,
EPA
notes
that
there
are
numerous
well­
known
statistical
methods
to
test
for 
and
if
present,
account
for 
the
potential
for
non­
independence
of
responses
among
individual
respondents
(
e.
g.,
McFadden
and
Train
2000).
Hence,
while
focus
groups
provide
little
evidence
of
concern
in
this
area,
statistical
options
exist
even
if
unanticipated
independence
is
discovered
ex­
post.

2.4
Respondents'
answers
are
conditioned
by
inaccurate
and
misleading
information.

The
commenter
states
that
the
TER
verbal
protocol
interviews
show
that
respondents'
answers
were
informed
by,
and
conditional
on,
the
information
provided
in
the
slide
show.
EPA
believes
this
result
is
entirely
expected,
and
is
consistent
with
the
academic
literature.
It
is
well
known
that
information
can
influence
economic
values
(
Bergstrom
and
Stoll
1989;
Bergstrom
et
al.
1989;
Hoehn
and
Randall
2002).
As
the
commenter
points
out,
the
"
NOAA
panel
identified
the
inclusion
of
accurate
and
complete
information
in
the
CV
questionnaire
as
a
critical
characteristic
of
a
CV
survey."

The
commenter
also
states
that
some
verbal
protocol
respondents
felt
that
the
survey
was
a
"
marketing
effort
and
sales
pitch"
designed
to
convince
them
to
vote
for
the
survey.
This
finding
contradicts
the
results
from
EPA's
focus
groups
and
cognitive
interviews,
which
demonstrated
that
most
participants
felt
that
the
slide
show
presented
objective,
unbiased
information.
Participants
felt
that
the
slide
"
didn't
push
you
either
way"
(#
10­
4)
and
that
the
information
presented
"
seemed
objective"
(#
11/
7­
1).
Furthermore,
the
participants
who
felt
that
the
slide
show
did
not
provide
objective
information
had
mixed
opinions
about
which
way
it
was
biased.
For
example,
one
individual
explained
that
"
I
think
by
saying
that
the
cost
will
be
passed
on
to
the
consumer,
without
discussing
the
benefit
other
than
just
more
fish,
[
it]
leaves
out
the
portion
that
crab
prices
could
go
down,
cod
prices
could
go
down,
there
would
be
more
abundance
of
fish"
(#
12­
16).
Another
individual
gave
an
opposite
opinion:
"
I
can't
remember
what
it
said
at
the
end
there,
that
seemed
a
little
bit
like,
you
know,
`
We're
trying
to
protect
everything'"
(#
6­
5).
Additionally,
many
of
the
individuals
who
felt
that
the
factual
information
in
the
slide
show
was
biased
revealed
through
their
responses
that
the
factual
information
presented
was
startling
(
i.
e.,
the
number
of
fish
killed
was
very
high),
but
that
the
presentation
of
the
materials
was
not
itself
slanted.
For
example,
one
participant
stated,
"
They
push
you
out
there,
1.8
billion
or
12.8 
they
make
it
sound
like
it's
so
much.
Like
you're
killing
this
amount
of
fish,
you
know.
And
almost
anything
that
you're
killing,
you
know,
you
don't
want
to
do
that"
(#
10­
4).
Finally,
some
individuals
who
indicated
that
the
slide
show
was
biased
also
indicated
that
this
perceived
bias
did
not
influence
their
decisions.
For
example:

R:
I
thought
they
were
trying
to
tell
me
to
vote
so
they
don't
use
their
cooling
systems.

M:
Okay.
Did
it
actually
get
you
to
vote
a
different
way
than
what
you
would
have?

R:
No.
(#
11/
8­
1)

On
page
16,
the
TER
Comment
claims
that
presenting
costs
on
a
monthly
basis
is
deceptive.
EPA
does
not
believe
that
this
is
true.
First,
the
Agency
notes
that
most
consumers
have
ample
experience
with
monthly
payments
for
items
such
as
rent,
mortgages,
loans,
utility
bills,
big­
ticket
items,
and
insurance.
Furthermore,
to
ensure
that
respondents
carefully
evaluate
the
cost
of
the
policies
relative
to
their
budget
constraints,
the
survey
also
shows
costs
on
a
yearly
basis.

On
page
17,
the
commenter
questions
EPA's
rationale
for
identifying
EPA
as
the
sponsor
of
the
survey,
and
argues
that
doing
so
may
affect
participants'
responses.
EPA
notes
that
this
choice
improves
the
consequentiality
of
the
survey,
as
clearly
recommended
by
Carson
et
al.
(
1999).
Respondents
are
presented
with
a
scenario
in
which
they
are
told
that
the
results
of
the
survey
"
will
influence
real
policy
changes,"
and
as
evidence,
EPA
is
cited
as
the
sponsor
of
the
survey.
Results
of
EPA's
focus
groups
show
that
this
treatment
is
effective,
and
that
participants
took
the
survey
questions
seriously.
For
example,
one
respondent
stated,
"
If
this
is
just
an
exercise
to
see
how
I
would
react,
I'd
still
answer
the
same
as
if
the
government
really
used
it
to
create
policy.
The
reason
I
used
it,
I
would
use
no
matter
what,
no
matter
who's
going
to
use
the
information,
yes."
Although
the
commenter
cites
a
few
respondents
who
felt
that
identifying
the
Agency
as
the
survey
sponsor
biased
their
responses,
the
majority
of
participants
in
EPA's
much
more
extensive
focus
groups
did
not
think
that
the
survey
was
biased.

In
the
following
paragraph,
the
commenter
refers
to
arguments
made
in
the
Barnthouse
Comment
about
the
validity
of
EPA's
estimates
of
fish
losses
and
long­
term
fish
population
impacts.
EPA's
responses
to
these
comments
are
discussed
in
the
Agency's
response
to
the
Barnthouse
Comment,
above.

Finally,
the
commenter
provides
a
number
of
excerpts
from
the
TER
verbal
protocol
interviews
that
demonstrate,
in
the
commenter's
opinion,
that
the
survey
provides
in
inadequate
treatment
of
uncertainty.
EPA
disagrees
with
this
comment.
EPA's
focus
groups
demonstrated
that
survey
respondents
clearly
understood
that
the
ecological
changes
described
in
the
survey
were
uncertain.
Most
respondents
were
comfortable
making
decisions
in
the
presence
of
this
uncertainty
(
although
they
were
less
comfortable
about
political
uncertainties,
such
as
whether
their
money
would
be
used
effectively).
As
one
respondent
explained,
"
It
sounds
like
they're
talking
from
a
certain
degree
of
certainty,
and
they're
not
putting
any
figures
out
there
that
they
can
actually
prove,
but
based
on
theory
we
know
that
fish
populations
are
declining
and
so
these
are
all
estimates.
And
I
accept
them
as
estimates"
(#
7­
7).
Another
stated,
"
If
they
had
stated
figures
as
though
these
figures
are
it 
that
I
would
question"
(#
7­
7).
However,
EPA
acknowledged
that
a
few
participants
did
express
protest
when
told
that
biologists
were
uncertain
as
to
the
long­
term
effects
of
programs
on
fish
populations.
For
example,
one
participant
stated,
"
I
instantly
thought,
oh,
here
we
go
again!
We
are
going
to
spend
the
money
and
they
are
going
to
tell
us
later
on
that
they
were
wrong,
but
we
are
not
going
to
say
that
we
were
wrong.
We
are
just
going
to
do
another
study."
(#
3­
17).
EPA
emphasizes
that
the
slide
show
clearly
states
that
results
of
the
proposed
policies
are
uncertain:
"
You
will
be
shown
different
policy
options 
with
different
effects
on
fish.
This
is
because
scientists
are
still
working
to
determine
what
the
exact
effect
on
fish
will
be,
so
it
is
important
to
know
how
you
would
react
to
a
wide
range
of
possible
outcomes.
Common
sense
indicates
that
preventing
the
loss
of
fish
eggs
and
young
fish
will
mean
more
adult
fish
in
future
years,
but
at
this
point
there
is
still
significant
uncertainty
regarding
the
exact
size
of
these
future
effects."

Beyond
these
comments,
EPA
also
highlights
that
many
of
TER's
comments
regarding
respondents'
reactions
to
uncertainty
are
based
on
a
version
of
the
survey
invented
by
TER
itself 
and
not
the
version
of
the
survey
developed
by
EPA.
EPA
emphasizes
the
clear
and
significant
differences
between
the
two
survey
versions,
as
highlighted
in
the
TER
Comment
and
the
Appendices.
EPA
as
a
result,
finds
these
results
questionable,
as
they
are
not
based
on
the
survey
instrument
proposed
by
EPA,
but
rather
a
"
straw
man"
survey
designed
explicitly
by
TER.
EPA
believes
that
these
comments
also
show
the
relative
success
of
the
EPA
survey
at
addressing
issues
of
uncertainty.
Unlike
the
version
of
the
survey
developed
by
TER,
which
was
met
with
skepticism
and
protests
by
their
respondents,
the
survey
versions
developed
by
EPA
were
found
to
appropriately
capture
uncertainty
by
the
vast
majority
of
focus
group
and
cognitive
interview
respondents.
For
example,
participants
made
statements
such
as:
"
This
is
what
the
scientists
project.
It's
not
for
certain.
I
knew
that
anyway
from
the
presentation"
[#
11/
1­
4],
"
It
was
said
in
the
slide
show,
that
they
really
don't
know.
[ ]
Scientific
speculation,
see,
so
there's
a
scientific
aspect
of
it"
[#
11/
2­
7],
and
"
Even
in
the
slide
show,
when
we
heard
it
and
watched
it,
a
lot
of
these
things
are
estimates.
They're
not
exact
sciences.
So,
basically
taking
these
questions
with
a
grain
of
salt
in
that
respect
as
far
as
the
averages
and
the
figures
and
all
that"
[#
11/
7­
4].

To
address
the
problem
of
uncertainty,
the
comment
suggests
modifying
the
survey
to
present
a
range
of
possible
effects.
EPA
agrees
this
is
a
valid
suggestion,
and
in
fact,
in
one
of
the
groups
of
cognitive
interviews,
EPA
tested
a
version
of
the
survey
that
did
show
a
range
of
effects.
The
respondents
were
explicitly
asked
whether
the
ranges
were
helpful
in
understanding
the
uncertainty
of
estimates
presented
in
the
choice
question
or
whether
they
were
a
source
of
confusion.
Seven
out
of
the
eight
respondents
interviewed
indicated
that
that
ranges
were
more
confusing,
that
the
original
presentation
of
resource
changes
was
more
clear,
and
that
they
clearly
understood
that
the
ecological
changes
described
in
the
survey
were
uncertain.
Furthermore,
respondents
were
comfortable
making
decisions
in
the
presence
of
this
uncertainty.
Moreover,
and
more
importantly,
the
respondents
in
cognitive
interviews
showed
that
different
people
anchor
their
responses
to
different
values
within
the
provided
range.
This
indicates
that
the
model
suggested
by
the
commenter
does
not
match
respondents'
behavior,
as
few
of
the
cognitive
interviews
showed
that
respondents
were
using
expected
values
to
determine
their
responses.
For
example,
some
respondents
anchored
their
responses
to
either
upper
or
lower
bound
of
the
presented
ranges,
while
others
used
a
mid­
point
estimate.
Still
others
had
no
clear
understanding
of
the
meaning
of
the
presented
ranges.
Therefore,
the
use
of
range
midpoints
for
estimation
(
or
any
other
point
for
that
matter)
will
result
in
errors­
in­
variables
bias
(
Greene
2003).

2.5
Respondents
have
difficulty
completing
the
survey,
which
could
lead
to
nonresponse
bias.

In
this
section,
the
commenter
makes
that
argument
that
the
information
presented
in
the
survey
is
complex,
which
may
lead
to
respondent
fatigue
and
non­
response
bias.
This
argument
is
contradicted
by
observations
from
EPA's
focus
groups.
Participants
made
statements
about
the
slide
show
such
as:
"
I
thought
it
was
good.
I
liked
it,"
"
Just
about
right,"
and
"
It
seems
interesting.
Informative."
No
participants
in
EPA's
focus
groups
indicated
that
they
would
not
have
completed
the
survey
because
it
was
long
or
difficult.
Furthermore,
EPA
has
made
a
number
of
efforts
to
simplify
the
presentation
of
information
and
the
format
of
the
survey.
Participants
responded
positively
to
these
changes.
For
example,
one
participant
stated:
"
On
option
B,
looking
at
it,
I
saw
that
the
blue
was
where
we
were
presently,
and
the
purple
would
be
how
much
more
we
would
increase
the
population
of
fish.
So
that
was
pretty
straightforward.
It
wasn't
very
difficult
to
understand."
[#
11/
7­
7].
Another
participant
stated,
"
The
pie
graphs,
I
could
understand
them."
[#
11/
4­
1].

The
commenter
also
argues
that
participants
failed
to
understand
some
of
the
information
presented
in
the
slide
show
and
survey.
EPA
notes
that
in
any
survey,
some
participants
will
fail
to
understand
or
remember
certain
pieces
of
information.
However,
in
EPA's
own
focus
groups,
most
participants
seemed
to
comprehend
and
remember
information
presented
in
the
slide
show
and
survey.
Participants
displayed
a
high
degree
of
understanding
with
regard
to
fish,
fish
ecosystem
services,
and
aquatic
habitat.
Furthermore,
no
participants
indicated
that
they
had
trouble
completing
the
survey
because
they
didn't
remember
or
understand
information
from
the
slide
show.
Finally,
EPA
has
included
specific
questions
in
the
survey
to
test
respondents'
recollection
and
understanding
of
specific
elements
of
information
from
the
slide
show.
3
These
questions
will
assist
in
the
identification
of
respondents
for
whom
recall
of
critical
information
is
a
concern.

On
pages
21
and
22,
the
commenter
discusses
reasons
why
participants
in
the
TER
protocol
interviews
were
uncertain
about
their
answers,
such
as
insufficient
information
or
misunderstanding
the
choice
sets.
In
EPA's
focus
groups,
participants
generally
felt
comfortable
with
their
answers,
although
some
expressed
uncertainty.
For
example,
participants
made
statements
such
as:
"
I
was
pretty
certain"
(#
7­
17),
and
"
I
would
be
very
certain
that
I
would
answer
in
a
real
referendum
the
way
I
did
it
here,
if
this
was
the
same
exact
survey
that
went
out
to
the
people,
and
it
was
presented
this
way
with
the
presentation"
(#
9/
7­
6).
Some
participants
expressed
hesitation:
"
I
wasn't
real
sure.
I
don't
know
that
much
about
it.
If
I
was
going
to
be
voting
on
it
I'd
do
some
research
on
it
before
I
would
vote
on
it"
[#
8­
6]).
Most
participants
who
were
unsure
of
their
answers
cited
a
lack
of
information
as
the
primary
reason.
These
participants
often
stated
that
they
wanted
more
details
about
how
the
policies
would
be
implemented,
particularly
3
For
detail,
see
Questions
2
and
3
of
the
survey
instrument
provided
in
Attachment
3,
U.
S.
EPA
(
2005).
what
technology
would
be
used
("
You
would
need
to
go
into
a
whole
lot
more
specifics
with
the
technology
you're
going
to
use"
[#
10­
17]).
As
discussed
previously,
EPA
has
included
several
debriefing
questions
in
the
survey
to
identify
respondents
who
are
very
uncertain
about
their
answers.

2.6
Summary
The
commenter
summarizes
the
major
points
of
the
comment
as
follows:

"
The
verbal
protocol
study
demonstrates
that
the
stated
preference
survey
proposed
by
EPA
to
measure
the
nonuse
value
of
certain
marginal
changes
in
forage
fish
populations
will
not
produce
reliable
results."
As
discussed
in
the
previous
sections
of
this
response,
most
issues
identified
by
the
commenter
reflect
a
misunderstanding
about
the
purpose
of
the
survey,
or
have
been
addressed
in
more
recent
versions
of
the
survey.

"
The
verbal
protocol
analysis
demonstrates
that
the
study
is
unnecessary.
Respondents
do
not
have
a
separate
meaningful
value
for
marginal
changes
in
forage
fish
populations.
To
the
extent
that
respondents
care
about
fish,
and
perhaps
would
actually
pay
to
protect
fish,
they
are
thinking
of
recreationally
and
commercially
significant
species.
The
benefits
of
CWIS
on
those
species
are
already
reflected
in
EPA's
RIA
for
the
Phase
III
proposed
rule."
As
discussed
in
EPA's
responses
above,
this
conclusion
reflects
a
misunderstanding
of
the
purpose
of
the
survey,
which
is
to
measure
total
use
values
for
reduction
in
impingement
and
entrainment
of
all
species 
including
both
forage
species
and
commercial
and
recreational
species.

Appendix
A:
Description
of
Verbal
Protocols
and
Respondents
First,
EPA
notes
that
the
TER
sample
is
very
small
(
15
participants)
to
use
to
draw
general
conclusions
based
on
the
focus
group
findings.
EPA
notes
that
the
demographic
information
provided
in
Appendix
A
shows
that
the
TER
sample
may
not
be
representative
of
general
population.
Some
demographic
groups
are
overrepresented
in
the
TER
sample.
For
example,
2/
3
of
the
participants
are
females,
and
half
of
the
participants
are
recreational
anglers.

EPA
also
emphasizes
that
the
pretest
conducted
by
TER
is
not
necessarily
a
"
clean"
pretest
of
the
survey
instrument,
in
that
additional
materials
and
questions
are
presented
to
respondents
prior
to
viewing
the
slide
show
and
taking
the
stated
preference
survey.
This
is
distinct
from
the
approach
used
in
actual
survey
implementation,
in
which
respondents
are
only
shown
the
slide
show
and
then
given
the
survey.
For
example,
as
noted
by
TER,
"
Respondents
completed
a
set
of
questions
at
the
start
of
the
interview
inquiring
about
their
knowledge
of
the
effects
of
cooling
water
intake
structures
on
fish
populations.
They
were
also
asked
to
indicate
whether
or
not
enough
money
is
spent
on
other
problems
in
this
country
and
then
rated
the
importance
of
specific
issues,
ranging
from
preserving
aquatic
life
to
reducing
taxes.
A
slide
show
was
presented
to
the
respondents
that
gave
background
information
on
the
effects
of
cooling
water
intakes
on
fish
populations.
The
slide
show
also
explained
the
decision
that
the
EPA
is
trying
to
make
regarding
regulations
for
cooling
water
intake
structures
and
how
the
technological
advancements
might
affect
the
respondent
financially.
The
respondents
were
then
given
the
SP
survey."
This
divergence
between
TER's
pretest
methods
and
actual
survey
implementation
methods
invalidates
much
of
the
pretest
results
with
regard
to
the
specific
survey
developed
by
EPA.
EPA
also
emphasizes
that
TER
apparently
did
not
test
the
specific
version
of
the
survey
proposed
by
EPA,
but
rather
a
"
similar"
survey.
Again,
any
differences
between
the
survey
proposed
by
EPA
and
the
survey
tested
by
TER
will
reduce
the
validity
of
focus
group
results
with
regard
to
the
specific
survey
instrument
proposed
by
EPA.

Appendix
B:
Verbal
Protocol
Material
No
response
is
required.

References
Arrow,
K.
,
R.
Solow,
E.
Leamer,
P.
Portney,
R.
Rander,
and
H.
Schuman.
1993.
"
Report
of
the
NOAA
Panel
on
Contingent
Valuation."
Federal
Register
58(
10):
4602­
4614.

Bateman,
I.
J.,
R.
T.
Carson,
B.
Day,
M.
Hanemann,
N.
Hanley,
T.
Hett,
M.
Jones­
Leee,
G.
Loomes,
S.
Mourato,
E.
Ozdemiroglu,
D.
W.
Pierce,
R.
Sugden,
and
J.
Swanson.
2002.
Economic
Valuation
with
Stated
Preference
Surveys:
A
Manual.
Northampton,
MA:
Edward
Elgar.

Bennett,
J.
and
R.
Blamey,
eds.
2001.
The
Choice
Modeling
Approach
to
Environmental
Valuation.
Northampton,
MA:
Edward
Elgar.

Bergstrom
J.
C.,
and
J.
R.
Stoll.
1989.
"
Application
of
Experimental
Economics
Concepts
and
Precepts
to
CVM
Field
Survey
Procedures."
Western
Journal
of
Agricultural
Economics
14(
1):
98­
109.

Bergstrom,
J.
D.,
J.
R.
Stoll,
and
A.
Randall.
1989.
"
Information
Effects
in
Contingent
Markets."
American
Journal
of
Agricultural
Economics
71(
3):
321­
323.

Besedin,
Elena,
Robert
Johnston,
Matthew
Ranson,
and
Jenny
Ahlen.
2005.
"
Findings
from
2005
Focus
Groups
Conducted
Under
EPA
ICR
#
2155.02."
Memo
to
Erik
Helm,
U.
S.
EPA/
OW,
October
18,
2005.

Boardman,
A.
E.,
D.
H.
Greenberg,
A.
R.
Vining,
and
D.
L.
Weimer.
2001.
Cost­
Benefit
Analysis:
Concepts
and
Practice,
2nd
edition.
Upper
Saddle
River,
NJ:
Prentice
Hall.

Carson
R.
T.,
T.
Groves,
and
M.
J.
Machina.
1999.
"
Incentive
and
Informational
Properties
of
Preference
Questions."
Plenary
Address
at
European
Association
of
Resource
and
Environmental
Economists,
Oslo,
Norway,
1999.

Champ,
P.
A.,
R.
Moore,
R.
C.
Bishop.
2004.
"
Hypothetical
Bias:
The
Mitigating
Effects
of
Certainty
Questions
and
Cheap
Talk."
Selected
Paper
prepared
for
presentation
at
the
American
Agricultural
Economics
Association
Annual
Meeting,
Denver,
Colorado,
2004.

Cummings,
R.
G.
and
G.
W.
Harrison.
1995.
"
The
Measurement
and
Decomposition
of
Non­
use
Values:
A
Critical
Review."
Environmental
and
Resource
Economics
1995(
5):
225­
247.

Federal
Register.
2004.
"
National
Pollutant
Discharge
Elimination
System 
Final
Regulations
To
Establish
Requirements
for
Cooling
Water
Intake
Structures
at
Phase
II
Existing
Facilities;
Final
Rule."
69
FR
41588.
Freeman
III,
A.
M.
2003.
The
Measurement
of
Environmental
and
Resource
Values:
Theory
and
Methods.
Washington,
DC:
Resources
for
the
Future.

Greene,
W.
H.
2003.
Econometric
Analysis.
5th
ed.,
Prentice
Hall,
Upper
Saddle
River,
NJ.

Hoehn,
J.
P.
and
A.
Randall.
2002.
"
The
Effect
of
Resource
Quality
Information
on
Resource
Injury
Perceptions
and
Contingent
Values."
Resource
and
Energy
Economics
24(
ER1­
2):
13­
31.

Johnston,
R.
J.,
and
D.
P.
Joglekar.
2005.
"
Validating
Hypothetical
Surveys
Using
Binding
Public
Referenda:
Implications
for
Stated
Preference
Valuation."
American
Agricultural
Economics
Association
(
AAEA)
Annual
Meeting,
Providence,
July
24­
27.

Johnston,
R.
J.,
J.
J.
Opaluch,
M.
J.
Mazzotta,
and
G.
Magnusson.
2005.
"
Who
Are
Resource
Nonusers
and
What
Can
They
Tell
Us
About
Nonuse
Values?
Decomposing
User
and
Nonuser
Willingness
to
Pay
for
Coastal
Wetland
Restoration."
Water
Resources
Research
41(
7):
W07017.

Johnston,
R.
J.,
E.
Y.
Besedin,
and
R.
F.
Wardwell.
2003.
"
Modeling
Relationships
Between
Use
and
Nonuse
Values
for
Surface
Water
Quality:
A
Meta­
Analysis."
Water
Resources
Research
39(
12):
1363.

Just,
R.
E.,
D.
L.
Hueth,
and
A.
Schmitz.
1982.
Applied
Welfare
Economics
and
Public
Policy.
Englewood
Cliffs,
NJ:
Prentice
Hall.

Louviere,
J.
J.,
D.
A.
Hensher,
and
J.
D.
Swait.
2000.
Stated
Preference
Methods:
Analysis
and
Application.
Cambridge,
UK:
Cambridge
University
Press.

McFadden,
D.
and
K.
Train.
2000.
"
Mixed
Multinomial
Logit
Models
for
Discrete
Responses."
Journal
of
Applied
Econometrics
15(
5):
447­
470.

U.
S.
EPA.
2000.
Guidelines
for
Preparing
Economic
Analyses.
USEPA
240­
R­
00­
003.
U.
S.
EPA,
Office
of
the
Administrator,
Washington,
DC,
September
2000.

U.
S.
EPA.
2005.
"
Supporting
Statement
for
Information
Collection
Request
for
Willingness
to
Pay
Survey
for
§
316(
b)
Phase
III
Cooling
Water
Intake
Structures:
Instrument,
Pre­
test,
and
Implementation."

Vossler,
C.
A.,
J.
Kerkvliet.
2003.
"
A
Criterion
Validity
Test
of
the
Contingent
Valuation
Method:
Comparing
Hypothetical
and
Actual
Voting
Behavior
for
a
Public
Referendum."
Journal
of
Environmental
Economics
and
Management
45(
3):
631­
649.
