PERSONALITY PROCESSES AND INDIVIDUAL DIFFERENCES
Positive Emotions in Early Life and Longevity:
Findings from the Nun Study
Deborah D. Danner, David A. Snowdon, and Wallace V. Friesen
University of Kentucky
Handwritten autobiographies from 180 Catholic nuns, composed when participants were a mean age
of 22 years, were scored for emotional content and related to survival during ages 75 to 95. A strong
inverse association was found between positive emotional content in these writings and risk of mortality
in late life (p < .001). As the quartile ranking of positive emotion in early life increased, there was a
stepwise decrease in risk of mortality resulting in a 2.5-fold difference between the lowest and highest
quartiles. Positive emotional content in early-life autobiographies was strongly associated with longev-
ity 6 decades later. Underlying mechanisms of balanced emotional states are discussed.
Longevity may be related to a variety of factors including
heredity, gender, socioeconomic status, nutrition, social support,
medical care, and personality and behavioral characteristics (Rob-
ine,
Vaupel, Jeune, & Allard, 1997). These factors might operate
throughout life or at particular life stages. Recent findings from the
Nun Study, a longitudinal study of older Catholic sisters, indicated
that linguistic ability in early life is associated with survival in late
life (Snowdon, Greiner, Kemper, Nanayakkara, & Mortimer,
1999).
In that study, the idea density (proposition, information, and
content) of autobiographies written at a mean age of 22 years was
strongly related to survival and longevity 6 decades later. Because
Deborah D. Danner and David A. Snowdon, Department of Preventive
Medicine and Sanders-Brown Center on Aging, College of Medicine,
University of Kentucky; Wallace V. Friesen, Sanders-Brown Center on
Aging, College of Medicine, University of Kentucky.
David A. Snowdon is now at the Department of Neurology and the
Sanders-Brown Center on Aging, College of Medicine, University of
Kentucky.
The study was funded by National Institute on Aging Grants
R01AG09862, K04AG00553, and 5P50AG05144, and by a grant from the
Kleberg Foundation.
This study would not have been possible without the spirited support of
the members, leaders, and health care providers of the School Sisters of
Notre Dame religious congregation. Archivists at each of the main con-
vents were instrumental in the study. We also wish to recognize the help of
Lydia Greiner in the conception of the study and Mark Desrosiers for his
valuable scientific and programming assistance. Other staff members of the
Nun Study who provided invaluable assistance on this project include
Danice Creager, Gari-Anne Patzwald, Jeanne Ray, and Mary Roycraft.
More information on the Nun Study may be obtained at http://www.
nunstudy.org.
Correspondence concerning this article should be addressed to Deborah
D.
Danner, Sanders-Brown Center on Aging, University of Kentucky, 800
South Limestone, Lexington, Kentucky 40536-0230. Electronic mail may
be sent to [email protected].
the autobiographies appeared to contain emotional content that
might be associated with idea density (Snowdon et al., 1996), we
investigated the relationship between emotional content in these
early life writings and survival in late life.
A growing body of literature has shown positive and negative
emotion-related attitudes and states to be associated with physical
health, mental health, and longevity. For example, in a longitudinal
study of Harvard graduates, Peterson (Peterson, Seligman, & Vail-
lant, 1988) found the ways in which young men explained bad
events predicted health outcome decades later. Such studies appear
to be based on assumptions that emotion-based constructs reflect
patterns of coping with negative life events and stresses that can be
harmful or beneficial to health. The assumptions of the current
longitudinal investigation of emotions and longevity are very
similar and evolved from what is known about the underlying
relationships among emotion, temperament, and physiology that
might influence longevity. This study builds on the knowledge that
there are universal, patterned emotional responses that affect phys-
iology in ways that are potentially damaging or beneficial.
Over the past 30 years, emotion researchers have identified
basic emotions such as happiness, sadness, anger, fear, and disgust
(Ekman & Friesen, 1969). More recently, these basic emotions
have been associated with differentially patterned autonomic
nervous system (ANS) responses (Ekman, Levenson, & Friesen,
1983;
Levenson, Carstensen, Friesen, & Ekman, 1991; Levenson,
Ekman, & Friesen, 1990; Levenson, Ekman, Heider, & Friesen,
1992).
The functional characteristics of the associated patterns of
emotion and ANS activation (Levenson, in press) strongly suggest
the potential for a lifelong pattern of emotional arousal affecting
health and longevity. Furthermore, numerous studies have shown
that complex emotional states, such as anxiety, produce elements
of ANS patterns associated with specific negative emotions (Laza-
rus,
1991). These same elements of elevated galvanic skin re-
sponse, heart rate, and blood pressure are found in the patterned
Journal of Personality and Social Psychology, 2001, Vol. 80, No. 5, 804-813
Copyright 2001 by the American Psychological Association, Inc. 0022-35I4/01/$5.00 DO): 10.1037//0022-3514.80.5.804
804
POSITIVE EMOTIONS
IN
EARLY LIFE
805
ANS responses
to the
arousal
of
basic emotions
and
potentially
could affect health
and
longevity.
Laboratory research also
has
found that
the
suppression
of
emotional states
can
exacerbate
ANS
responses (Gross
&
Leven-
son, 1997).
A
lifelong pattern
of
suppressing
the
expression
of
emotion
has the
potential
for
adverse effects
on
essential body
systems. Although
no
ANS pattern has been found
to be
associated
with positive emotion that differentiates
it
from baseline (Leven-
son
et al.,
1990), studies have demonstrated
the
potential muting
effects
of
positive emotion
on the
bodily responses
to
negative
emotion (Fredrickson
&
Levenson, 1998). This healing effect
of
positive emotion
may
have
the
potential
to
reduce stress
on the
cardiovascular system even
in the
face
of
inevitable negative life
events.
In
other words, constructs such
as
optimism
and
positive
attitude
may
imply
the
following sequence: Events arousing
neg-
ative affect
are
approached with confidence that
the
future holds
something positive
and
better, thus internally generating
a
positive
emotional state that mutes
the
adverse effects
of the
prolonged
arousal
of a
negative emotion.
The basic research
of
Fredrickson, Gross,
and
Levenson cited
above
has
laid
the
groundwork
for the
study
of
how sustained
and
repetitious patterns
of
emotional arousal might relate
to
physical
health
and
survival
and,
more specifically,
how the
emotion
sys-
tem
is
intimately tied
to the ANS,
which activates cardiovascular
responses that could have cumulative adverse
or
salutary effects
on
health (Krantz
&
Manuck, 1984). What
is
needed
is an
explanation
for
why a
particular pattern
of
emotional
and ANS
responses
would
be
repeated with sufficient frequency
to
produce such
cumulative effects.
As
part
of
this explanation,
it is
necessary
to
examine
the
relationship among patterns
of
emotional responsive-
ness,
temperament,
and the
development
of
personality.
Temperament,
the
biologically based propensity
for
individuals
to respond
to
events
in
particular ways,
is
considered
by
some
theorists
to
contribute
to the
development
of
personality (Izard,
Libero, Putnam,
&
Haynes,
1993;
Malatesta
&
Wilson, 1988).
Moreover, temperament
is
proposed
to
reflect
the
degree
to
which
emotions
are
generally expressed,
as
well
as the
differing frequen-
cies with which specific emotions
or
patterns
of
emotions
are
displayed
or
suppressed (Izard
et al.,
1993). Early
and
continuing
styles
of
emotional expression
are
proposed
to
constitute some
characteristics
of
personality (Izard
et al., 1993;
Malatesta
&
Wilson, 1988). Supporting this line
of
reasoning, work
by
Headey
(Headey
&
Wearing,
1992)
suggests that individuals maintain
levels
of
positive
or
negative affect that
are
determined
by
their
personalities
and
that after emotional arousal
or
stress these levels
return
to
individual baselines (Diener, 2000). When
an
individual's
response pattern
is
frequent
or
sustained negative emotional
arousal with slow return
to a
tranquil baseline,
the
autonomic
response could prompt cardiovascular activity that accelerates
disease mechanisms such
as
atherosclerosis.
In
contrast,
a
pattern
of relatively infrequent negative emotional arousal
or one
that
rapidly returns
to a
calm baseline following negative arousal could
have beneficial effects
on
health.
Such
a
balance
of
emotional states, either
by
avoiding suppres-
sion
of the
expression
of
aroused emotion
or by
readily resolving
negative arousal,
is
compatible with Vaillant's proposal that
ma-
ture defenses work
to
promote
a
positive psychology that enhances
the ability
to
work, love,
and
play (Vaillant, 2000). Vaillant
provided evidence that earlier life manifestations
of
mature
ego
defenses that balance
and
attenuate multiple sources
of
conflict
predict enhanced physical
and
mental health
20
years later
and
suggests that mature
ego
defenses
may
reflect inborn traits.
If so,
Vaillant's proposition
may
offer
yet
another pathway
for how
potentially beneficial
or
harmful patterns
of
emotional responses
may
be
expressed
and
balanced
and may be
mediated through
patterns
of
problem solving throughout
a
lifetime, thereby influ-
encing longevity.
A pattern
of
emotional arousal
and
temperament
may be dis-
closed,
in
part,
by the
written expression
of
language. Research
by
Pennebaker
and his
colleagues
has
used written language
as a
means
of
understanding
how
emotion influences both physical
and
psychological health (Hughes, Uhlmann,
&
Pennebaker,
1994;
Pennebaker, 1993; Pennebaker
&
King, 1999).
The
early-life
au-
tobiographies
in our
study afford another opportunity
to
examine
emotional content
in
written language
and its
relationship
to
health.
If the use of
emotional content
in
these writings reflects
reactions
to
inevitable stressful life events, then these writings
may
reveal characteristic responses
to
intense
or
sustained arousal that
produces allostatic load—indicators
of
physiological response
to
stress (McEwen,
1998;
Singer
& Ryff, 1999;
Sterling
&
Eyer,
1988).
Furthermore,
if the use of
positive
and
negative emotional
content
in
writing reflects
a
general readiness
to
express emotion,
then these writings
may
indicate
a
pattern that avoids
the
adverse
effects
of
suppressing
the
expression
of
emotions.
On the
other
hand,
if the use of
positive emotional content
in
writing reflects
a
readiness
to
resolve negative arousal, then writings
may be
reveal-
ing
a
pattern of balance
in
emotional response indicating allostasis,
adaption
to
change while maintaining physiological systems
within
a
normal range (Singer
& Ryff, 1999;
Sterling
&
Eyer,
1988;
McEwen, 1998). Both
the
avoidance
of
suppression
and the
positive resolution
of
life's stresses could have beneficial effects
on health
and
longevity.
Seligman emphasizes that
an
insightful, positive attitude
in
dealing with life events,
an
optimistic explanatory style
in
contrast
to
a
pessimistic
one, can
lead
to
greater feelings
of
well-being
and
perhaps even
to
longer life (Seligman, 2000).
In
support,
a
recent
study found optimism,
as
measured
by a new
optimism-pessimism
scale
of the
Minnesota Multiphasic Personality Inventory (Swen-
son, Pearson,
&
Osborne, 1973),
was
associated with
a
lower risk
of death
in 839
Mayo Clinic patients observed over
a
30-year
period (Maruta, Colligan, Malinchoc,
&
Offord, 2000). However,
in another long-term study
of
more than
a
thousand bright Cali-
fornia school children, cheerfulness (i.e., parental judgments
of
optimism
and a
sense
of
humor)
had an
inverse relationship with
longevity during middle
and old age
(Friedman, 1999).
In the
latter
study,
the
cheerful participants also were found
to be
more likely
to engage
in
activities known
to be
risk factors
for
mortality.
On
the other hand,
in
another analysis
of
the California data, Peterson
and colleagues used
the
Content Analysis
of
Verbatim Explana-
tions technique (Peterson, Seligman, Yurko, Martin,
&
Friedman,
1998)
to
code questionnaires completed
by the
participants
in
early
adulthood
and
found evidence
of a
negative relationship between
pessimism
and
longevity.
The early-life autobiographies
and
mortality data available
for
participants
in the Nun
Study offer
a
unique opportunity
to
inves-
tigate
the
possible association
of
written emotional expression
to
longevity. Participants
in our
study
had the
same reproductive
and
marital histories,
had
similar social activities
and
support,
did not
806
DANNER, SNOWDON,
AND
FRIESEN
smoke
or
drink excessive amounts
of
alcohol,
had
similar occu-
pations
and
socioeconomic status,
and had
comparable access
to
medical care. Therefore, even though
it
may
be
difficult
to
gener-
alize from this unique population
of
Catholic sisters, many factors
that confound most studies
of
longevity have been minimized
or
eliminated.
Method
Study Population
The
Nun
Study
is a
longitudinal study
of
aging
and
Alzheimer's disease
(Snowdon, 1997; Snowdon
et
al.,
1996,
1999). Participants were members
of
the
School Sisters
of
Notre Dame religious congregation
who,
before
their retirement, lived
and
taught
in the
schools
of
cities
and
towns
in the
midwestern, eastern,
and
southern United States.
In
1991 through 1993,
all
American School Sisters
of
Notre Dame born before
1917
were asked
to
join
the Nun
Study.
Six
hundred seventy-eight women agreed
to
participate
in
all
phases
of
the study
and
gave informed written consent
to
allow access
to their archived
and
active records, participate
in
annual assessments
of
cognitive
and
physical function,
and
donate their brains
at
death.
At the
first annual exam,
the
678
participants were
75
to
102
years
old
(M
= 83).
A search
of
the
convents' archives revealed that
the
Mother Superior
of
the North American sisters,
who
resided
in
Milwaukee, Wisconsin,
had
sent
a
letter
on
September
22,
1930,
requesting that each sister write
an
autobiography.
A
mix
of
handwritten
and
typed autobiographies
for
many
of the 678 sisters
in
the
study
was
found
and
these autobiographies became
an invaluable research source. Criteria used
to
select autobiographies
for
intensive study were that
the
writers were born
and
raised
in the
United
States
and
thus
had the
opportunity
to
master
the
English language
and
that
the autobiographies were handwritten
and
therefore could
be
authenticated
as unaltered
by
clerical
staff.
We found that
the
number
of
available handwritten autobiographies
was
related
to the
convent
in
which
the
sister lived
and the
year
she
wrote
her
life story.
A
large number
of
autobiographies meeting criteria were found
for participants from
the
Milwaukee, Wisconsin,
and the
Baltimore, Mary-
land, convents
who
took their religious vows
and
formally joined
the
religious congregation during 1931
to
1943
(Snowdon
et
al.,
1999).
Of
the
678 sisters
in the Nun
Study,
218
took their vows
in
these
two
convents
during that time period
and
handwritten autobiographies were found
for
180 (83%)
of
these participants; that is,
101
participants from
the
Milwau-
kee,
Wisconsin, convent
and 79
from
the
Baltimore, Maryland, convent.
These
180
autobiographies were written some time between
the
ages
of
18
and 32 (M = 22)
depending
on the
age
at
which
the
sister joined
the
congregation.
At
the
time
of
writing
the
autobiographies, 82%
of
the sisters
had earned
a
high school diploma.
By the
beginning
of the
mortality
follow-up period
in
1991, 91%
had
earned
at
least
a
bachelors degree.
During
the
mortality follow-up period
of
November
13,
1991,
to
Septem-
ber
1,
2000,
the
participants ranged
in age
from
75 to 95
years
and 76
I
had
died (Milwaukee sample
=
43%, Baltimore sample
=
42%).
Autobiographies
Beginning
in
1930,
each sister
who
took
her
final vows
was
asked
to
write
a
short sketch
of
[her] life. This account should
not
contain more
than
two to
three hundred words
and
should
be
written
on a
single
sheet
of
paper
. . .
include place
of
birth, parentage, interesting
and
edifying events
of
childhood, schools attended, influences that
led to
the convent, religious life,
and
outstanding events.
Clearly,
the
instructions were
not
intended
to
influence
the
manner
in
which these life events were described
nor
were they intended
for
the
study
of emotional content, coping styles,
or
patterns
of
reasoning. Rather,
we
suspect that the autobiographies
may
have been used
in
part
by the
convent
leaders
to
gather information that might help
to
determine future educa-
tional
and
occupational paths,
as
well
as to
provide information useful
for
creating obituaries.
Despite uniformity
in the
events that were described,
the
manner
in
which
the
life facts were told
in the
autobiographies reflected individual
style
and
ranged from simply stating that these life events happened
and
when they occurred
to
elaborations
of the
simple facts that included
the
emotions experienced
by the
writer
or
others involved
in
the
life event.
The
following sentences, from
the
beginning
and
ending
of two
autobiogra-
phies,
demonstrate differences
in
emotional content:
Sister
1
(low positive emotion):
I
was born
on
September
26,
1909,
the
eldest
of
seven children, five girls
and two
boys
.... My
candidate
year
was
spent
in the
Motherhouse, teaching Chemistry
and
Second
Year Latin
at
Notre Dame Institute. With God's grace,
I
intend
to do
my best
for our
Order,
for
the
spread
of
religion
and for
my
personal
sanctification.
Sister
2
(high positive emotion):
God
started
my
life
off
well
by
bestowing upon
me a
grace
of
inestimable value...
. The
past year
which
I
have spent
as
a
candidate studying
at
Notre Dame College
has
been
a
very happy
one. Now I
look forward with eager
joy to
receiving
the
Holy Habit
of
Our Lady
and
to a
life
of
union with Love
Divine.
Coding the Autobiographies and Generating Scores
The coding system used
in
classifying
the
written autobiographies
was
designed specifically
for
this study (Danner, Friesen,
&
Snowdon, 2000).
All coding
and
review
of the
autobiographies were done without knowl-
edge
of
the health
or
functional status
of
the
study participants. Two coders
identified
all
words
in
thel80 autobiographies that reflected
an
emotional
experience
and
classified them
as
positive, negative,
or
neutral. Later,
a
third coder verified each coded word
for
accuracy
and
determined
the
specific type
of
emotional experience
or
state referenced
by
each word.
Coders were instructed
on the
distinctions between descriptions
of
possible elicitors
of
emotion (e.g., death
of a
family member),
the
emotion
that
was
experienced (e.g., sadness), subsequent behaviors (e.g., crying),
and attempts
to
control
the
overt expression
of the
emotion. They were
instructed
not to
code descriptions
of
possible elicitors,
but to
code only
words that
in
context described
the
emotion that
was
experienced
and
behaviors subsequent
to
emotional arousal. Further, they were instructed
not
to
code words such
as
good
and bad
that have positive
or
negative
connotations
or
might imply
an
emotional reaction
but do not
directly
describe
an
emotional experience.
The coders were provided with examples
of
words related
to the
expe-
rience
of the
positive emotions
of
accomplishment, amusement, content-
ment, gratitude, happiness, hope, interest, love,
and
relief;
the
negative
emotions
of
anger, contempt, disgust, disinterest, fear, sadness,
and
shame;
and
the
neutral emotion
of
surprise. The two coders,
one
with a background
in psychology
and the
other with training
in
education, then independently
read
the
autobiographies. They marked words that conveyed emotion
as
experienced
by the
writer
or
others
and
classified
the
valence
of the
emotional content
as
positive, negative,
or
neutral. When necessary
for
comprehension,
the
coders were instructed
to
identify
and
code phrases
rather than single words.
Two procedures were used
to
generate scores
for
the primary analysis
on
the basis
of
the
positive, negative,
and
neutral scoring.
The
first procedure
simply used
the raw
count
of
positive,
negative,
and
neutral emotion words
for each autobiography.
The
second procedure used these coded emotion
words
to
classify each sentence
as
containing
one or
more positive,
negative,
or
neutral words
or
as
containing
no
emotion words. The first
two
columns
of
Table
1
show
the
number
of
positive, negative,
and
neutral
emotional words
and
sentences
as
determined
by
each individual coder.
The table shows that
the two
coders identified very similar numbers
of
positive, negative,
and
neutral emotional words.
POSITIVE EMOTIONS
IN
EARLY LIFE
807
Table
1
Reliability
of
the Emotion Coding
as
Indicated
by the
Number
of
Emotion Words
and
Sentences
Scored
by
Two Coders
for
Autobiographies Written
in
Early Life
by 180 Participants
in the Nun
Study
Unit
of
analysis
and emotion
Words
Positive
Negative
Neutral
Sentences
Positive
Negative
Neutral
Coder
A
1,243
206
16
1,006
196
16
Count
Coder
B
1,242
192
17
1,017
179
17
Coders
A
and
B
a
.96
(.95,
.97)
.89
(.85,
.94)
.78 (.64, .93)
.97 (.96, .98)
.90
(.85,
.95)
.78 (.64, .93)
Correlation
of
counts
Final coding
and
Coder
A
.99 (.98, .99)
.97 (.94, .99)
.97 (.90,
1.00)
.99 (.98, .99)
.97 (.96, .99)
.97 (.90,
1.00)
each coder
a
Coder
B
.98 (.98, .99)
.94
(.91,
.97)
.82 (.69, .95)
.99 (.98, .99)
.94
(.91,
.97)
.82 (.69, .95)
Note.
For all
correlations,
p <
.0001.
a
95% confidence intervals appear
in
parentheses.
In
the
verification phase
of the
coding,
the
words scored
by the two
coders were extracted from
the
autobiographies
and a
nonredundant list
of
words
was
reviewed
by a
third person (Wallace
V.
Friesen)
for
accuracy.
This review
was
done without knowledge
of
whether
one or
both coders
had scored the word or how frequently the word was scored. Words that did
not meet
the
original criteria
for an
emotional experience were removed
from the list. The 1,598 words retained
in
the final scoring constituted
1.8%
of the total words
in the
autobiographies
and
95%
of
the words scored
by
one
or
both coders.
Of
these emotional words,
84%
were classified
as
positive, 14%
as
negative,
and 1% as
neutral.
As
described above
for the
single coders,
the
verified coding
of
emotion words was used
to
determine
the number
of
sentences with
one or
more positive, negative,
and
neutral
emotion words.
As
a
part
of the
verification process, each unique emotion word
was
classified
as
referring
to a
specific type
of
positive
or
negative emotions
(only one emotion, surprise,
was
scored
in the
neutral category). Initially,
the purpose
of the
categorization
was to aid in the
verification
of the
positive, negative, and neutral scoring
of
Coders
A
and B.
If a
word could
not
be
categorized,
its
validity
as an
emotion word
was
questionable.
We
carefully reviewed this categorization
of
the emotion words
and
disagree-
ments were discussed and arbitrated. The final list
of
subcategories and
the
number and percentage
of
sentences containing one
or
more words
in
each
category
is
presented
in
Table
2.
Intercoder Reliability
Two types
of
intercoder reliability were assessed:
the
overall agreement
in selecting
and
classifying the valence
of
emotional words
and the
degree
to which the coders' scoring and the verified scoring
of
the
autobiographies
were correlated. Kappa coefficients were used
to
assess overall agreement
between
the two
coders
on the
selection
and
classification
of
emotion
words. The coefficient values were .83 and .84, .85, and .79
for
all emotion,
positive, negative,
and
neutral words respectively, indicating
a
satisfactory
level
of
intercoder reliability both overall
and for the
individual types
of
emotion words. Additional analyses indicated that most differences
be-
tween
the two
coders were
due to one
coder identifying
a
word that
the
other failed to detect and that this occurred with similar frequencies
for
the
two coders. Examination
of
these disagreements indicated that
the
coder
who failed
to
code apparently simply
did not see the
word because
the
same word
was
identified
and
classified identically
by the
errant coder
in
different places
in
the autobiographies.
In
other words, had the errant coder
noticed
the
word when reading
the
autobiography,
it
almost certainly
would have been scored
in
agreement with
the
accurate coder.
In addition
to the
kappa coefficients
of
agreement, each coder's scoring
and
the
verified scoring were used
to
generate positive, negative,
and
neutral counts
for
both words
and
sentences
for
each autobiography.
Correlations were used
to
test the comparability
of
the three sets
of
coding.
The resulting correlations
are
shown
in the
three columns
on the
right
of
Table
I. It
can
be
seen here that the correlations between Coders
A and B
and verified counts
of
the numbers
of
emotional words and sentences were
very high, indicating that virtually identical results would have been
obtained
in
subsequent survival analysis had either Coder A's
or
Coder
B's
scoring been used
in
place
of the
final verified scores.
Linguistic Measures
Recent findings from studies
of
the same 180 autobiographies indicated
that linguistic ability
in
early life
was
associated with survival
in
late life
(Snowdon
et
al., 1999).
In
that study,
the
idea density (proposition, infor-
mation, and content)
of
these autobiographies was associated with survival
and longevity
6
decades later. Idea density
of
the early-life autobiographies
also
had a
strong inverse association with Alzheimer's disease (Snowdon
et al., 1996; Snowdon, Greiner,
&
Markesbery, 2000). Because idea-dense
sentences
of
the autobiographies were observed to contain emotional words
(Snowdon
et al.,
1996), idea density
and
grammatical complexity were
used
as
control variables
in one of the
analyses
in the
current study.
The
following
is a
brief description
of how
idea density
and
grammatical
complexity were measured.
Without
the
linguistic coders' knowledge
of
the
age or
cognitive func-
tion
of
each sister during late life, each autobiography
was
scored
for two
indicators
of
linguistic ability: idea density (Kintsch
&
Keenan,
1973;
Turner
&
Greene, 1977)
and
grammatical complexity (Cheung
&
Kemper,
1992).
Mean idea-density
and
grammatical-complexity scores were
com-
puted from
the
last
ten
sentences
of
each autobiography. Idea density
was
defined
as the
average number
of
ideas expressed
per ten
words. Ideas
corresponded
to
elementary propositions, typically
a
verb, adjective,
ad-
verb,
or
prepositional phrase. Complex propositions that stated
or
inferred
causal, temporal,
or
other relationships between ideas also were counted.
Grammatical complexity
was
computed using
the
Developmental Level
metric originally developed
by
Rosenberg
and
Abbeduto (Rosenberg
&
Abbeduto,
1987) and
modified
by
Cheung
and
Kemper (1992).
The De-
velopmental Level metric classifies sentences according
to
eight levels
of
808
DANNER, SNOWDON, AND FRIESEN
Table 2
Distribution of the Different Types of Emotion Sentences in the
Autobiographies Written in Early Life by 180 Participants
in the Nun Study
Type of emotion
Positive
Happiness
Interest
Love
Hope
Gratefulness
Contentment
Unspecified
Accomplishment
Relief
Amusement
Negative
Unspecified
Sadness
Afraid
Disinterest
Confused
Anxiety
Suffering
Shame
Hopelessness
Frustration
Disgust
Anger
Contempt
Neutral
Surprise
No.
Milwaukee
convent
109(6.10)
160(8.95)
36(2.01)
21 (1.17)
6 (0.34)
19(1.06)
11(0.62)
7 (0.39)
2(0.11)
0 (0.00)
19(1.06)
8 (0.45)
4 (0.22)
7 (0.39)
5 (0.28)
1 (0.06)
8 (0.45)
4 (0.22)
2(0.11)
1 (0.06)
1 (0.06)
0 (0.00)
0 (0.00)
3(0.17)
(and %) of sentences
Baltimore
convent
341 (12.51)
281 (10.31)
131 (4.81)
30(1.10)
41 (1.50)
21 (0.77)
14(0.51)
15(0.55)
4(0.15)
1 (0.04)
36(1.32)
46(1.69)
18(0.66)
13 (0.48)
13 (0.48)
16(0.59)
9 (0.33)
6 (0.22)
2 (0.07)
1 (0.04)
1 (0.04)
2 (0.07)
1 (0.04)
14(0.51)
Both
convents
450 (9.97)
441 (9.77)
167(3.70)
51(1.13)
47(1.04)
40 (0.89)
25 (0.55)
22 (0.49)
6(0.13)
1 (0.02)
55 (1.22)
54(1.20)
22 (0.49)
20 (0.44)
18(0.40)
17 (0.38)
17 (0.38)
10 (0.22)
4 (0.09)
2 (0.04)
2 (0.04)
2 (0.04)
1 (0.02)
17 (0.38)
Note. A small percentage of sentences contained more than one type of
emotion word. Nonspecific positive and negative emotion words were
classified as unspecified (e.g., words such as liked and filled with emotion).
These are words that definitely refer to an emotional experience but might
refer to several different basic or complex emotional states.
grammatical complexity, ranging from 0 (simple one-clause sentences) to 7
(complex sentences with multiple forms of embedding and subordination).
Data Analysis
The dependent variables in the analyses were simple measures of all-
cause mortality such as the percent who died by the end of an approxi-
mately 9-year follow-up period and the mortality rate (i.e., deaths per
person-years of observation) for that same period of time. The primary
multivariate method used to investigate mortality was Cox proportional
hazards regression (Allison, 1995). This regression yielded the relative risk
of death, which refers to the ratio of mortality rates (or, more exactly, to the
ratio of hazard functions). Age was adjusted in these analyses by using age
as the time scale for the regression (Allison, 1995). Educational level at the
time the autobiographies were written in early life was adjusted by includ-
ing it as an ordinal variable in the regression. Age- and education-adjusted
survival curves (the probability of a 75-year-old surviving to different
advanced ages) were created using the baseline feature of the Cox regres-
sion procedure in the SAS statistical program (Allison, 1995).
In the regression analyses, ordinal variables were used to characterize
the percentile ranking of each type of emotional expression; that is, the
number of positive emotional words. Binary variables were used in the
regression to characterize the quartile rankings of each type of emotional
expression. These percentile and quartile rankings of emotional-word us-
age were derived using the distribution within each of the two convents.
This was done to obtain comparable scales of emotional expression across
convents because the distribution of emotion-word usage differed between
convents (see Table 2). The primary analyses used three measures of
emotion word usage: (a) the percentile or quartile rankings derived from
the number of sentences containing one or more positive or negative
emotion words or no emotion words; (b) percentile and quartile ranks
derived from the simple counts of positive emotion words; and (c) percen-
tile and quartile ranks of a diversity score generated by counting the
number of different positive emotion categories (see Table 2) scored in
each autobiography.
Results
The current study included 180 participants from the Milwau-
kee,
Wisconsin, and Baltimore, Maryland, convents of the School
Sisters of Notre Dame. Handwritten autobiographies composed
when the sisters were a mean age of 22 years were scored for
positive, negative, and neutral emotional content. When these
autobiographies were written in early life, 82% of the participants
had earned a high school diploma. At the beginning of the Nun
Study in 1991, approximately 58 years later, 91% of them had
earned at least a bachelors degree. During the 9-year mortality
surveillance period, the 180 participants ranged in age from 75
to 95 years and 76 (42%) of them had died (Milwaukee sample =
43%,
Baltimore sample = 42%).
Compared with the Baltimore participants, the Milwaukee par-
ticipants had a lower mean number of positive emotion sentences
(Milwaukee = 3.2, Baltimore = 9.7; p < .001), negative emotion
sentences (Milwaukee = 0.6, Baltimore = 2.0; p < .001), and
nonemotion sentences (Milwaukee =
14.1,
Baltimore = 23.5;/? <
.001).
(Given the very low frequency of neutral emotions, shown
in Tables 1 and 2, their possible relationship to mortality was not
examined.) Although the exact reasons for the differences between
convents in written emotional expression is not known, the differ-
ences in lengths of the autobiographies could simply reflect more
time allowed for the Baltimore sisters to complete the task. Be-
cause of differences in the distribution of these measures between
convents, all analyses were based on percentile and quartile rank-
ings within each convent.
Four basic types of analyses were conducted and all were age
and education adjusted. The first examined the relationship be-
tween risk of mortality and the percentile ranking of the number of
positive emotion sentences, negative emotion sentences, and non-
emotion sentences in the autobiographies from early life. The
second examined the relationships between the risk of mortality
and the quartile ranking of the number of positive emotion sen-
tences, positive emotion words, and different types of positive
emotion words (i.e., categories). The third analysis examined the
age-adjusted survival curves (length of life) as a function of the
quartile rankings of positive-emotion sentences, positive-emotion
words, and different categories of positive emotion words. A
fourth analysis examined the relationships between positive emo-
tion usage and survival after controlling for linguistic ability
demonstrated in the early-life autobiographies, the level of educa-
tion attained at the time the autobiographies were written, and the
lifetime occupation of the participants.
The first Cox regression model we used to investigate mortality
used the percentile ranking of positive emotion sentences, negative
POSITIVE EMOTIONS IN EARLY LIFE
809
emotion sentences, or no emotion sentences and was adjusted for
age and education. The results of these analyses are presented in
Table 3 (Model I). Statistically significant inverse associations
were found between the percentile ranking of the number of
positive sentences in the early-life autobiographies and the risk of
mortality in late-life within each of the convents and in both
convents combined. For example, for every 1.0% increase in the
number of positive-emotion sentences there was a 1.4% decrease
in the mortality rate (i.e., the hazard function from the Cox
regression model). In contrast, there were no statistically signifi-
cant associations between the risk of mortality and the percentile
rankings of the number of negative emotion sentences or the
number of nonemotion sentences.
In another regression model that included age, education, and
the percentile rankings of all three types of sentences (Model II;
see Table 3), the strength of the associations with mortality were
statistically unchanged from those above (Model I). Overall, the
findings from these regressions suggest that positive and negative
content reflected different aspects of written emotional expression.
Because of these findings, the remaining analyses focused on
positive emotions.
We further explored the association between positive emotion
content and survival using quartile rankings of positive emotion
sentences. Both the percent who had died and the mortality rate
had inverse associations with the quartile rankings of the number
of positive-emotion sentences (see Table 4). Findings from age-
and education-adjusted Cox regression analyses also indicated that
the relative risk of death increased in a stepwise fashion as the
quartile ranking of positive emotion sentences decreased, with
a 2.5-fold difference in mortality between the lowest and highest
quartiles of positive emotional expression. Two other methods of
characterizing positive-emotion content, the number of positive-
emotion words and the number of different positive emotions, also
had strong inverse associations with mortality (see Table 4).
Cox regression also was used to create age- and education-
adjusted survival curves, that is, probabilities of a 75-year-old
surviving to different advanced ages. Figure 1 shows a strong
association between the quartile rankings of the number of positive
emotion sentences and survival: The median age at death was 86.6
years for those in the lowest quartile for the number of positive
emotion sentences, 86.8 for the second quartile, 90.0 for the third
quartile, and 93.5 for those in the highest quartile, that is, a
difference of 6.9 years between the highest and lowest quartiles of
positive emotion sentences. Survival curves for the other two
measures of positive emotion content (not shown) indicated even
stronger associations with survival; in other words, the difference
in the median age at death between the highest and lowest quartiles
was 9.4 years for the number of positive emotion words and 10.7
years for the number of different positive emotions.
Other analyses indicated that there were no material changes in
the association between positive emotion content and survival after
controlling for measures of linguistic ability as demonstrated in the
autobiographies; that is, the 2.5-fold difference in risk of mortality
between the lowest and highest positive emotion sentence quartiles
in Table 4 was a 2.2-fold difference in risk when adjusted for idea
density. Furthermore, the relationship between positive-emotion
content and survival was still apparent after limiting the analyses
to 162 college-educated, lifetime teachers.
Discussion
This study found a very strong association between positive
emotional content in autobiographies written in early adulthood
and longevity 6 decades later. Such a finding is congruent with
other studies by investigators that have found relationships be-
tween longevity and emotion-related concepts. Features of the
current study differ from other studies that have investigated
relationships between emotion-relevant behaviors and longevity or
mortality and may account for the strength of the relationship
observed in the current study: the population sample and the
technique used to measure emotion.
Our findings are compatible with recent longitudinal studies that
suggest that optimism is associated with longer life (Maruta et al.,
2000;
Peterson et al., 1998), but incompatible with another study
indicating that cheerfulness measured in early life was not asso-
ciated with longer survival (Friedman, 1999). In the latter study,
the investigators reported that there were behaviors related to risk
and substance abuse in late-life activities of the more cheerful
participants that may account for their findings (Friedman, 1999).
These types of behaviors should be less of an issue in our study of
Catholic sisters given the relative homogeneity of their adult
lifestyles and environments.
Table 3
Percent Change in the All-Cause Mortality Rate Per Single Percentile Change
in the Ranking of the Number of Sentences
Model I
Sentence type Milwaukee convent Baltimore convent
Both convents
Model II, both
convents
Positive emotion -1.4 (-2.5,-0.2)* -1.4 (-2.7,-0.1)* -1.4 (-2.3,-0.6)*** -1.4 (-2.3,-0.5)**
Negative emotion -0.7 (-1.9, 0.6) -0.7 (-1.9, 0.6) -0.7 (-1.5, 0.2) -0.2 (-1.2, 0.8)
Noemotion 0.5 (-0.6, 1.6) -0.6 (-1.9,0.7) -0.1 (-0.9,0.7) 0.3 (-0.6, 1.2)
Note. 95% confidence intervals appear in parentheses. The mortality rate refers to the hazard function from
Cox regression. Both Models I and II were adjusted for age by using age as the time scale in the regression. The
level of education achieved at the time in early life when the autobiography was written was included as an
ordinal variable in both Model I and II regressions. Three regressions were used for Model I, that is, each
included only one sentence-type variable, as well as age and education. One regression was used in Model II,
that is, it included each of the three sentence-type variables, as well as age and education.
*p£.05.
**p<.01.
***/?<.
001.
810
DANNER, SNOWDON, AND FRIESEN
Table 4
Positive Emotion Expression in Autobiographies Written in Early Life and the Risk of All-Cause Mortality
in Late Life for 180 Participants in the Nun Study
Categories and quartiles
Positive emotion sentences
I (low)
II
III
IV (high)
Positive emotion words
I (low)
II
III
IV (high)
Different positive emotions
I (low)
II
III
IV (high)
Age at follow-up
80.1
81.1
80.1
79.4
79.9
81.1
79.7
79.9
81.3
80.4
80.2
79.4
No.
of
participants
Dead
25
23
18
10
23
30
13
10
11
26
29
10
At-risk
46
40
52
42
42
51
40
47
17
58
65
40
Survival"
253.3
221.8
333.4
296.9
233.0
279.5
255.6
337.3
91.0
331.0
388.0
295.5
% died
54
58
35
24
55
59
33
21
65
45
45
25
Mortality rate
b
10
10
5
3
10
11
5
3
12
8
7
3
Relative risk of mortality
(and 95% CI)
2.5(1.2,5.3)**
2.4(1.1,5.2)*
1.4(0.6,3.0)
1.0
3.2(1.5,6.8)**
3.1 (1.5,6.4)**
1.6(0.7,3.7)
1.0
4.3(1.7, 10.4)**
2.3(1.1,4.7)*
2.2(1.1,4.6)*
1.0
Note. CI = confidence interval. The relative risks were adjusted for age and education (i.e., age was adjusted by using it as the time scale in the Cox
regression analyses, and the level of education achieved at the time in early life when the autobiography was written was included as an ordinal variable
in regressions). The quartiles are not equal size groups because of the distribution of the variables. Positive emotion sentences refer to the number of such
sentences, and different positive emotions refer to up to 10 different types of positive emotions.
" In person-years.
b
Deaths per 100 person-years.
*/7<.05.
**/><.01.
There were, however, other important differences between the
studies of Maruta et al. (2000), Peterson et al. (1988), and Fried-
man (1999) and the current study in addition to the sample pop-
ulations. Maruta used items from the first Minnesota Multiphasic
Personality Inventory to develop a scale of optimism. Peterson
measured globality of explanatory style (a tendency to ascribe a
single cause across negative life events) in samples of writing.
Friedman, studying the same population as Peterson, used parental
reports of children's degree of cheerfulness as a measure of pos-
itive affect. Thus, there were notable differences in the sources of
information and the constructs that were measured as predictor
variables in previous research. The current study used a different
source of data and an emotion-specific measurement technique.
Given the unique lifestyle and culture of our study population of
Catholic sisters and the fact that the autobiographies were written 6
decades ago, we created a coding system appropriate for this
sample of writings. It has face validity and good reliability. Based
on knowledge of emotion research and theory, it was designed to
identify positive and negative emotional content in writings and
requires little or no inference to apply. However, such coding does
not attempt to measure more complex reactions to life events such
as long-term positive or negative attitudes, forward thinking, types
of explanatory style, mature ego defenses, pessimism, or opti-
mism. Further research is required to discover how the use of
emotion words in written text is related to other constructs that
have been found to be related to better health and longevity.
Although the scoring tool used in this study was designed to
measure both positive and negative emotion, the emotional content
of the writings describing the early lives of the participants in our
study was overwhelmingly positive. This finding does not differ
substantially from what other research studies using similar mea-
surement tools have found (Pennebaker & Francis, 1999). More-
over, an examination of cross-cultural and developmental data
for 13 cultures by Boucher and Osgood (1969) found a universal
tendency to learn positive words earlier than negative words in the
acquisition of language, to more readily retain positive words in
tests of memory, and to use more positive than negative words
when communicating.
Contextual factors also may have influenced the use of positive
emotion words in this set of writings. For example, the Catholic
sisters in our study may have been aware, or at least believed, that
the content of the autobiographies would be used by their superiors
to determine their careers in the religious congregation, and there-
fore they may have been cautious about revealing memories of
negative emotion. Even more likely to have influenced the tone of
these autobiographies was that they were written during a period of
time when the sisters would be expected to feel happy and positive
about the future; namely, when they were about to leave the
convent and begin working (mainly teaching) in the community.
Having completed years of study and preparation for entry into the
religious order, the sisters wrote these autobiographies just prior to
taking their final vows. A goal toward which they had worked was
being realized. Yet, despite the forces that may have resulted in
predominantly positive content that was relatively constant for all
of the authors of the autobiographies, there were individual dif-
ferences in the use of positive emotion words that predicted
longevity.
Because there was relatively little negative emotional content in
the autobiographies, it was not possible to address directly ques-
tions related to the underlying mechanisms responsible for the
current findings. Whether a generalized suppression of emotional
expressiveness presented a risk factor for longevity (Gross &
POSITIVE EMOTIONS IN EARLY LIFE
811
1.0
0.9
0.8
m
jrvii
0)
"5
ity
bil
n
rob
a.
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
75
Quartile 1
Quartile 2
Quartile 3
Quartile 4
80 85
Age
90
95
Figure
1.
Quartile rankings of the number of positive emotion sentences in autobiographies written in early life
and the probability of survival in late life for 180 participants in the Nun Study. (Note that the survival curves
for Quartiles 1 and 2 are virtually overlaid on each other.)
Levenson, 1997), or, conversely, whether those persons using
more positive emotion words were in fact more expressive of all
emotions and thereby reduced allostatic load by avoiding the
detrimental effects of suppression could not be tested. Although
negative life events were sometimes mentioned in the autobiogra-
phies,
the participants had not been instructed to include such
events or to elaborate on their resolution. The absence of negative
emotion words in relating negative incidents did not allow a direct
test of whether positive emotion might have been a factor in
muting the adverse effects of negative emotional arousal
(Fredrickson & Levenson, 1998). Finally, the relative absence of
negative emotional content limited the statistical power to detect
associations with mortality. However, the analysis that we could
perform indicated that in this context written negative emotional
content is not the opposite of positive emotional content but,
rather, is a reflection of something different. This finding that
positive emotion may be a different phenomenon from negative
emotion (depression) also was reported by Ostir and colleagues
(Ostir, Markides, Black, & Goodwin, 2000).
Our investigation raises questions about why the positive emo-
tional content in early-life writings might have such a powerful
relationship to longevity. Unfortunately we had no independent
measures of temperament, personality, or emotional tendencies for
participants, and we can only speculate that individual differences
in emotional content in the autobiographies reflect life-long pat-
terns of emotional response to life events.
A pattern of emotional expression that accentuates positive
affect undoubtedly has behavioral correlates that could enhance or
disrupt the positive effects on physiology and health. One behav-
ioral pathway is suggested by the study by Friedman (1999) in
which cheerful participants were more likely to engage in behav-
iors that are health risks such as excessive drinking and smoking.
Such a pathway would be expected to disrupt the potential phys-
iological benefits of a pervading pattern of positive emotional
responsiveness. In contrast, all participants in the current study had
lived a lifestyle in which such health-risk behaviors were improb-
able and therefore the physiological impact of a positive emotional
style was almost certainly enhanced. Because many alternative
paths that might be the consequence of a positive style were not a
part of this study, generalization of the current findings is limited.
Many of the limitations of our study also could be considered
strengths. As mentioned earlier, participants in our study were all
female, had the same reproductive and marital histories, had sim-
ilar social activities and support, did not smoke or drink excessive
amounts of alcohol, had virtually the same occupation and socio-
economic status, and had comparable access to medical care.
Furthermore, the 180 participants had successfully completed a
lifetime within their careers and living situations and many had
lived beyond average life expectancy for their generation by the
time they were enrolled in this study. Although it may be difficult
to generalize from this unique population of Catholic sisters, the
findings of the study should not be minimized. Despite factors in
these sisters' lives that are known to extend life and that might
have overwhelmed any contribution of the mechanisms underlying
our findings, the phenomenon represented by the use of positive
emotion words in early-life writings effectively added to longevity.
It could be argued that the results of this study may not apply to
a sample of participants less than 75 years of age. We are in the
process of searching the convent archives for the autobiographies
of sisters who died prior to the beginning of the study and in
812
DANNER, SNOWDON, AND FRIESEN
particular those who died before age 75. This information will
allow us to examine the possibility that what was found in this
study was the late stages of a relationship between the use of
positive emotion words and longevity that was evident years
earlier. Furthermore, increasing the sample size will increase the
statistical power of future analyses and allow the investigation of
relationships between survival and different types of positive emo-
tional words, such as interest, love, and hope. Finally, our contin-
ued follow-up of the population will allow us to determine whether
this association continues beyond age 95.
Finding such a strong association of written positive emotional
expression to longevity indicates a need for research that sheds
light on the underlying mechanisms and mediators responsible for
and associated with this relationship. Within the context of the Nun
Study, evidence that the expressive patterns observed in the early-
life autobiographies were stable over time would help substantiate
a relationship between emotional expression and temperament and
personality. In future research, we will study late-life writings and
spoken speech samples from the sisters for consistency of the
expressive patterns found in early life.
Archived records of medical history and career path will be
examined for evidence of social and health-related patterns asso-
ciated with what has been observed in the autobiographical writ-
ings and that might suggest pathways taken by participants differ-
ing in their use of positive emotional words that might have
contributed to their longevity or mortality. Considering the poten-
tial impact of positive expressiveness on relationships, we feel that
research is needed to examine possible differences in social and
professional behaviors that may have amplified the effects of a
positive style on longevity.
Given that there have been annual examinations of cognitive
and physical functioning, it will be possible to study relationships
between the emotional content of the early-life writings and late-
life capacities. Also, the results of neurological examinations will
allow study of relationships with neurological functioning and
related disease and disability. Finally, because there will be brain
autopsies on all participants, it will be possible to study relation-
ships between written emotional expressions and neuropathology
and brain structure. These future studies hold promise for identi-
fying underlying mechanisms and mediators that may account for
the findings of the current study.
References
Allison,
P. D.
(1995). Survival analysis using the SAS system:
A
practical
guide. Cary, NC: SAS Institute.
Boucher,
J., &
Osgood,
C. E.
(1969).
The
Pollyanna Hypothesis. Journal
of Verbal Learning
and
Verbal Behavior,
8, 1-8.
Cheung,
H., &
Kemper,
S.
(1992). Competing complexity metrics
and
adults' production
of
complex sentences. Applied Psycholinguistics,
13,
53-76.
Danner, D. D., Friesen, W.
V.,
& Snowdon, D.
A.
(2000). Written emotion
expression code. Unpublished manuscript, University
of
Kentucky, Lex-
ington.
Diener,
E.
(2000). Subjective well-being:
The
science
of
happiness
and a
proposal
for a
national index. American Psychologist,
55,
34-43.
Ekman, P., & Friesen, W. V. (1969). The repertoire
of
nonverbal behavior:
Cite%OTO4,
CK\%\ns,,
\\4Age,,
ra\&
eoAmg. Semiotica,
],
49-9%.
Ekman, P., Levenson, R. W., & Friesen, W. V. (1983). Autonomic nervous
system activity distinguishes among emotions. Science,
221, 1208-
1210.
Fredrickson,
B. L., &
Levenson,
R. W.
(1998). Positive emotions speed
recovery from
the
cardiovascular sequelae
of
negative emotions.
Cog-
nition
and
Emotion,
12,
191-220.
Friedman,
H. S.
(1999). Personality
and
longevity: Paradoxes.
In J. -M.
Robine, B. Forette, C. Franceschi, & M. Allard (Eds.), The paradoxes of
longevity (pp. 115-122). Berlin, Germany: Springer-Verlag.
Gross,
J. J., &
Levenson,
R. W.
(1997). Hiding feelings: The acute effects
of inhibiting negative
and
positive emotion. Journal of Abnormal
Psy-
chology, 106, 95-103.
Headey, B.,
&
Wearing,
A. J.
(1992). Understanding happiness:
A
theory
of subjective well-being. Melbourne, Australia: Longman Cheshire.
Hughes,
C. F.,
Uhlmann,
C, &
Pennebaker,
J. W.
(1994).
The
body's
response
to
processing emotional trauma: Linking verbal text with
autonomic activity. Journal
of
Personality,
62,
565-585.
Izard, C.
E.,
Libero,
D. Z.,
Putnam,
P., &
Haynes,
O. M.
(1993). Stability
of emotion experiences
and
their relations
to
traits
of
personality. Jour-
nal
of
Personality
and
Social Psychology,
64,
847-
860.
Kintsch, W.,
&
Keenan,
J.
(1973). Reading rate and retention
as a
function
of the number
of
propositions
in the
base structure
of
sentences.
Cog-
nitive Psychology,
5,
257-274.
Krantz, D.
S.,
& Manuck, S. B. (1984). Acute psychophysiologic reactivity
and risk
of
cardiovascular disease:
A
review
and
methodologic critique.
Psychological Bulletin,
96,
435-464.
Lazarus,
R. S.
(1991). Emotion
and
adaptation.
New
York: Oxford
Uni-
versity Press.
Levenson,
R. W. (in
press). Autonomic specificity
and
emotion.
In R. J.
Davidson,
K.
Scherer, & H.
H.
Goldsmith (Eds.), Handbook of affective
sciences.
New
York: Oxford University Press.
Levenson,
R. W.,
Carstensen,
L. L.,
Friesen,
W. V., &
Ekman,
P.
(1991).
Emotion, physiology, and expression
in old
age. Psychology and Aging,
6,
28-35.
Levenson,
R. W.,
Ekman,
P., &
Friesen,
W. V.
(1990). Voluntary facial
action generates emotion-specific autonomic nervous system activity.
Psychophysiology,
27,
363-384.
Levenson, R. W., Ekman, P., Heider, K., & Friesen, W. V. (1992). Emotion
and autonomic nervous system activity
in the
Minangkabau
of
West
Sumatra. Journal
of
Personality
and
Social Psychology,
62,
972-988.
Malatesta,
C. Z., &
Wilson,
A.
(1988). Emotion cognition interaction
in
personality development:
A
discrete emotions, functionalist analysis.
British Journal
of
Social Psychology,
27,
91-112.
Maruta,
T.,
Colligan,
R. C,
Malinchoc,
M., &
Offord,
K. P.
(2000).
Optimists
vs
pessimists: Survival rate among medical patients over
a
30-year period. Mayo Clinic Proceedings,
75,
140-143.
McEwen,
B. S.
(1998). Stress, adaptation,
and
disease: Allostasis
and
allostatic load. Annals
of the New
York Academy
of
Sciences,
840,
33-44.
Ostir,
G. V.,
Markides,
K. S.,
Black,
S. A., &
Goodwin,
J. S.
(2000).
Emotional well-being predicts subsequent functional independence
and
survival. Journal
of
the American Geriatrics Society,
48,
473-478.
Pennebaker,
J. W.
(1993). Putting stress into words: Health, linguistic
and
therapeutic implications. Behaviour Research
and
Therapy,
31, 539-
548.
Pennebaker,
J. W., &
Francis,
M. E.
(1999). Linguistic Inquiry
and
Word
Count (LIWC). Mahwah,
NJ: LEA
Software
and
Alternative Media/
Erlbaum.
Pennebaker,
J. W., &
King,
L. A.
(1999). Linguistic styles: Language
use
as
an
individual difference. Journal
of
Personality
and
Social Psychol-
ogy,
77,
1296-1312.
Peterson,
C,
Seligman,
M. E. P., &
Vaillant,
G. E.
(1988). Pessimistic
e*p\atiatorj sty\e
« a
risk factor
fox
pYi^sVcaX
fflness-.
A
ftnrty-fwe year
longitudinal study. Journal
of
Personality
and
Social Psychology,
55,
23-27.
Peterson,
C,
Seligman, M.
E.
P., Yurko, K. H., Martin, L.
R.,
& Friedman,
POSITIVE EMOTIONS IN EARLY LIFE
813
H. S. (1998). Catastrophizing and untimely death. Psychological Sci-
ence, 9, 127-130.
Robine, J. -M, Vaupel, J. W., Jeune, B., & Allard, M. (1997). Longevity:
To the limits and
beyond.
Berlin, Germany: Springer-Verlag.
Rosenberg, S., & Abbeduto, L. (1987). Indicators of linguistic competence
in the peer group conversational behavior of mildly retarded adults.
Applied Psycholinguistics, 8, 19-32.
Seligman, M. E. P. (2000). Optimism, pessimism, and mortality. Mayo
Clinic Proceedings, 75, 133-134.
Singer, B., &
Ryff,
C. D. (1999). Hierarchies of life histories and associ-
ated health risks. Annals of the New York Academy of Sciences, 896,
96-115.
Snowdon, D. A. (1997). Aging and Alzheimer's disease: Lessons from the
Nun Study. Gerontologist, 37, 150-156.
Snowdon, D. A., Greiner, L. H., Kemper, S. J., Nanayakkara, N., &
Mortimer, J. A. (1999). Linguistic ability in early life and longevity:
Findings from the Nun Study. In J. -M. Robine, B. Forette, C. Franches-
chi,
& M. Allard (Eds.), The paradoxes of longevity (pp. 103-113).
Berlin, Germany: Springer-Verlag.
Snowdon, D. A., Greiner, L. H., & Markesbery, W. R. (2000). Linguistic
ability in early life and the neuropathology of Alzheimer's disease and
cerebrovascular disease: Findings from the Nun Study. Annals of the
New York Academy of
Sciences,
903, 34-38.
Snowdon, D. A., Kemper, S. J., Mortimer, J. A., Greiner, L. H., Wekstein,
D.
R., & Markesbery, W. R. (1996). Linguistic ability in early life and
cognitive function and Alzheimer's disease in late life: Findings from
the Nun Study. Journal of the American Medical Association, 275,
528-532.
Sterling, P., & Eyer, J. (1988). Allostasis: A new paradigm to explain
arousal pathology. In S. Fisher & J. Reason (Eds.), Handbook of life
stress, cognition, and health (pp. 629-649). New York: Wiley.
Swenson, W. M., Pearson, J. S., & Osborne, D. (1973). An MMPI source
book:
Basic item, scale, and pattern data on 50,000 medical patients.
Minneapolis: University of Minnesota Press.
Turner, A., & Greene, E. (1977). The construction and use of a proposi-
tional text
base.
Boulder: Institute for the Study of Intellectual Behavior,
University of Colorado.
Vaillant, G. E. (2000). Adaptive mental mechanisms: Their role in a
positive psychology. American Psychologist, 55, 89-98.
Received October 6, 2000
Accepted October 30, 2000 •
Low Publication Prices for APA Members and Affiliates
Keeping you up-to-date. All APA
Fellows,
Members, Associates, and Student Affiliates
receive—as part of their annual dues—subscriptions to the American Psychologist and
APA Monitor. High School Teacher and International Affiliates receive subscriptions to
the APA Monitor, and they may subscribe to the American Psychologist at a significantly
reduced rate. In addition, all Members and Student Affiliates are eligible for savings of up
to 60% (plus a journal credit) on all other APA
journals,
as well as significant discounts on
subscriptions from cooperating societies and publishers (e.g., the American Association for
Counseling and Development, Academic Press, and Human Sciences Press).
Essential resources. APA members and affiliates receive special rates for purchases of
APA books, including the Publication Manual of
the
American Psychological Association,
and on dozens of new topical books each year.
Other benefits Of membership. Membership in APA also provides eligibility for
competitive insurance
plans,
continuing education programs, reduced APA convention fees,
and specialty divisions.
More information. Write to American Psychological Association, Membership Services,
750 First Street, NE, Washington, DC 20002-4242.
