Beruflich Dokumente
Kultur Dokumente
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
Wiley and Society for Research in Child Development are collaborating with JSTOR to digitize, preserve and
extend access to Child Development.
http://www.jstor.org
Claudia Lampman-Petraitis
Loyola University of Chicago
LARSON, REED, and LAMPMAN-PETRAITIS,CLAUDIA.Daily Emotional States as Reported by Children and Adolescents. CHILDDEVELOPMENT,
1989, 60, 1250-1260. Hour-to-houremotional states
reportedby children, ages 9-15, were examined in orderto evaluatethe hypothesis thatthe onset of
adolescence is associated with increased emotional variability.These youths carried electronic
pagers for 1 week and filled out reportson their emotional states in response to signals received at
randomtimes. To evaluate possible age-relatedresponse sets, a subset of children was asked to use
the same scales to rate the emotions shown in drawingsof 6 faces. The expected relationbetween
daily emotional variabilityand age was not found among the boys and was small among the girls.
There was, however, a linear relationbetween age and averagemood states,with the older participants reportingmore dysphoricaverage states, especially more mildly negative states. An absence
of age difference in the ratingsof the faces indicatedthat this relationcould not be attributedto age
differences in response set. Thus, these findings provide little supportfor the hypothesis that the
onset of adolescence is associatedwith increased emotionalitybut indicate significantalterationsin
everyday experience associated with this age period.
To evaluate the hypothesis that adolescents are more emotionally variable than
preadolescents, we examined time-sampling
reports obtained from 9-15-year-olds concerning the emotional states they experienced
during their daily lives. The vantage point for
this investigation was phenomenological; we
were concerned with the hour-to-hour range
of emotional states children and adolescents
attribute to themselves-with
their "conscious states" (Harris & Olthof, 1982)acknowledging that a different pattern of
findings might emerge were it possible to
measure components of emotion that a person
does not consciously detect, such as physiological arousal or unconscious affect. Emotional variability was defined for this study as
the standard deviation of a person's daily
mood states as assessed on a continuum from
positive to negative states. The research question is, then, do adolescents experience a
wider range of emotional states than preadolescents? In their daily lives do they attribute
to themselves the greater variation in states
that have long been attributed to them by observers? The study also addresses the related
This research was supported by NIMH grant 1 R01 MH38324, "Stressin Daily Life during
Early Adolescence," awardedto Reed Larsonand carriedout throughMichael Reese Hospital and
Medical Center. Send requests for reprintsto the first authorat the Division of Human Development and Family Ecology, University of Illinois at Urbana-Champaign,1105 West Nevada Street,
Urbana, IL 61801.
[Child Development, 1989, 60, 1250-1260. ? 1989 by the Society for Researchin Child Development, Inc.
All rightsreserved.0009-3920/89/6005-0008$01.00]
more
BackgroundResearch
One must separate the hypothesis of adolescent emotionality from the broader hypothesis that adolescence is a time of turmoil or
"storm and stress," a view that no longer enjoys general acceptance (Adelsen, 1979; Petersen, 1988; Strober, 1986). The turmoil hypothesis asserts that emotionality is part of
a normal and even desirable adolescent condition of personality disruption and change
(Blos, 1961; Freud, 1937). Research findings,
however, make it clear that such disruption is
neither normative nor desirable. The great
majority of adolescents report feeling globally
happy with their lives (Offer, Ostrov, & Howard, 1981), rates of psychiatric disorder during
adolescence are not greater than at later periods of the life span (Rutter, Graham, Chadwick, & Yule, 1976), and the typical teenager
does not experience a breach in relations with
parents (Hill & Holmbeck, 1986; Montemayor, 1986). Furthermore, when turmoil
does occur, it is associated with worse, not
better long-term adjustment (Dusek & Flaherty, 1981; Kohlberg, LaCrosse, & Ricks,
1974; Offer & Offer, 1975).
The absence of a general condition of turmoil among a majority of teenagers, however,
does not disprove the more specific hypothesis that adolescents are emotional creatures.
The possibility remains that wide emotional
extreme positive as well
swings-including
as negative emotional states (Sharp, 1980)are a normative part of the adolescent experience.
Research provides some indication that
adolescents experience stronger emotions in
their daily lives than do adults. Direct data on
daily emotional experience were obtained by
Larson, Csikszentmihalyi, and Graef (1980),
who used electronic pagers to signal adolescents and adults to report on their subjective
experiences at random times during the day.
They found that the mood states reported by a
sample of high school students had higher
standard deviations than those reported by a
sample of adults. The mean mood levels reported by the two groups were similar, indicating that their emotional states had similar
baselines; the adolescents, however, reported
more occasions both of negative and positive
extremes.
Additional evidence comes from onetime questionnaire studies showing that
youth and young adults report greater emotional extremes than older adults. Sixteen- to
1251
1252
Child Development
dren and adolescents. The onset of adolescence in the United States, where the study
was done, is typically associated with pubescence, the start of junior high school, and
other transitions that occur between the sixth
and ninth grades (Lipsitz, 1977; Steinberg,
1985). In order to obtain an adequate representation of preadolescents, we included fifth
graders; hence our sample includes students
from the fifth to the ninth grades (ages 9 to
15). The emotionality hypothesis predicts an
increase in the statistical variance of hour-tohour mood states across this age period. We
also examined age differences in average
states and the frequency of extreme states.
Following the procedure used by Larson
et al. (1980) these students rated their moods
at random times during the day in response to
signals from a pager. In order to discriminate
findings from trends that might be attributed
to response style, we also asked a subsample
of the students to rate the emotional states
they perceived in six simple drawings of faces
representing a range of emotions. Given the
quantity of evidence showing sex differences
during this age period, all analyses were performed separately for girls and boys.
Method
Subjects.-Participants in the study were
473 randomly selected children and young
adolescents in the fifth to ninth grades (ages 9
to 15) from four suburban neighborhoods in
the Midwestern United States. Two neighborhoods were generally lower middle class,
the third was middle class, and the fourth was
upper middle class. These communities were
composed almost exclusively of people with
European ancestry. To diversify possible
time-of-year effects and make the data collection manageable, the participation of the students was spread across eight waves of data
collection over 2 years. Sample selection was
stratified to obtain balanced representation by
sex, community, time of year, and grade, with
approximately equal numbers of students representing each quarter from the autumn of
fifth grade to the winter of ninth grade.
The final sample includes 68.8% of the
randomly selected students initially invited to
take part. Twenty-four percent of those invited declined to take part or failed to obtain
permission from their parents, 4.4% began the
study but failed to complete the requisite
number of self-report forms, and 2.8% completed the study but provided data that were
unusable. Sample loss was approximately
equal among boys and girls (32.2% vs. 30.2%,
respectively) and was more frequent in the
1253
Results
The analyses involved a series of parallel
multivariate analyses of variance with grade
as the independent variable and the self-reports and ratings of faces as the dependent
variables. All of the dependent variables
were subjected to homogeneity-of-variance
tests prior to the main analyses, which indicated that MANOVA would be an acceptable
method of handling these data. Each significant MANOVA was followed up with univariate tests for polynomial trends. Finally, all
analyses were performed separately for girls
and boys.
To test differences in emotional variability between age groups, we performed a
MANOVA with grade as the independent
variable and students' standard deviations on
each of the six self-ratings scales as the de-
""
CO)
in
0
cqcl
cqlCO)
c qm n
0
Cq
,v
vOQ
p
"00Po4
)-,C)0CO
C0
CO)
cd
00m CO)m
0C)~O CO)
>
W0
C ..,
...
C)) ? r-00
C)ItN00in?lcoCO)
0 m0o
0-0
14
z
cj
.I
-4CIO
UQ o~m~
m
ncI 0000i
..0
E-
>
>
ZZ
z~
cd
0n - 4-1~
0 li C)CO
00CO
bC
z CA
~4
a
T~l
L4
(I)
"I
"
k
It
CO)
0 nIT
C)
.. .C)
o . 0....
-4
.iQn
. ...0.It. oc
.cq
.C)
. N.. C.1Z
P..
.
"0"~0
1
O S.1
bo
C) )
ItfCO)f0C)
1
..
...c. .
"
..
>
IIC-1
-
0(
m"in
'I0-.?coC?-I
)co
t-0rco
?-I
?I'm
Cco
C)tP-4
u a D0 Ecd?I -I~0~I?00
in00inIt0) C C ;2
d>
aa o
"c
*..
P.
..
4.
CO)
cd
. . . . .D
.
U
1$
P.
cd
r
rA
A) f*
bCbO
t. PO
m
~?m
a)
1255
TABLE 2
AGE TRENDS IN THE STANDARD DEVIATION OF REPORTED DAILY EMOTIONALSTATES
GRADE
Girls:
Happy ..............
Cheerful ............
Friendly ............
Alert ................
Strong ..............
Excited .............
N ...................
Boys:
Happy ..............
Cheerful ............
Friendly ............
Alert ................
Strong ..............
Excited .............
N ...................
NOTE.-The
t FOR
LINEAR
TREND
CORRELATION
WITH
GRADE
1.21
1.08
1.12
1.44
1.05
1.46
49
1.18
1.20
1.15
1.44
1.12
1.49
54
1.27
1.27
1.22
1.48
.98
1.58
52
1.25
1.22
1.27
1.52
1.19
1.46
48
1.36
1.21
1.23
1.50
1.12
1.40
37
1.70
2.27
1.68
.85
.81
- .65
.09
.03
.01
.40
.42
.52
.11
.15
.17
- .06
.05
- .03
1.24
1.27
1.24
1.37
1.12
1.44
49
1.18
1.17
1.15
1.43
1.11
1.40
54
1.28
1.25
1.24
1.56
1.17
1.49
52
1.11
1.21
1.16
1.50
1.06
1.42
48
1.12
1.12
.99
1.42
1.00
1.27
37
-1.48
- 1.06
-1.94
.62
-1.09
- 1.23
.14
.29
.05
.53
.28
.22
-.10
-.06
-.06
.05
-.07
-.07
body of the table shows mean standard deviations for each grade level.
1256
Child Development
TABLE 3
AGE TRENDS IN THE MEANS OF REPORTED DAILY EMOTIONALSTATES
Girls:
Happy ...........
Cheerful ..........
Friendly .........
Alert .............
Strong ...........
Excited ..........
Boys:
Happy ...........
Cheerful .........
Friendly .........
Alert .............
Strong ...........
Excited ..........
35
CORRELATION
WITH
GRADE
t FOR
LINEAR
TREND
5.61
5.34
5.60
4.45
4.46
4.77
5.43
5.25
5.41
4.58
4.68
4.64
5.10
4.92
4.98
4.24
4.28
4.25
5.14
5.11
5.12
4.31
4.34
4.44
4.98
4.83
5.11
4.48
4.42
4.42
-4.22
- 2.64
-3.19
- .42
- 2.07
- 2.07
.001
.009
.002
.68
.04
.04
- .28
-.17
- .21
- .04
-.14
-.14
5.27
4.94
5.26
4.67
4.92
4.53
5.18
4.90
4.93
4.77
4.99
4.40
4.96
5.01
4.88
4.47
4.89
4.38
4.86
4.60
4.78
4.26
4.52
4.26
4.90
4.46
4.75
4.12
4.58'
4.44
-2.81
- 3.03
- 3.04
- 3.30
- 2.62
- 1.08
.005
.003
.003
.001
.009
.48
-.19
-.19
- .20
- .21
-.17
- .06
GRADE
Percentage of Self-Reports
35
30
30
25
25
20
20
15
16
10
10
6
0
-3
-2
Negative
Neutral
+3
+2
+1
-1
Positive
Grade
5th
6th
7th
8th
9th
of self-reported states by grade: girls. (Note: Table shows the mean frequency
FIG. 1.-Frequency
with which students used each of the gradations on the scales to identify their daily emotional states.)
points. For both girls and boys, the most positive scale point (+3) was used to describe
their states half as often in ninth grade as in
fifth grade. For girls, the linear correlation between grade and percentage in this category
was r = -.21 (p < .001) and for boys was r =
-.22 (p < .001). In place of these extreme
positive states, the older girls reported more
frequent experience of both mildly negative
and mildly positive states, as indicated by
significant linear correlations between frequency and grade level for the scale points
marked +1 (r = .20, p < .01), -1 (r = .21,
p < .001) and -2 (r = .14, p < .05). Older
Percentage of Self-Reports
35
30
30
25
25
20
20
15
15
10
10
5
5
0
1257
0
-3
-2
-1
Negative
+1
Neutral
+2
+3
Positive
1258
Child Development
among older than younger subjects. The absence of comparable age differences in ratings
of emotions in drawings of faces indicates that
these findings cannot be attributed to age differences in response bias.
Two explanations can be considered for
these differences between children's and adolescents' representations of their daily affective states. One explanation is that there are
differences in how younger and older youth
interpret what may be very similar internal
and external daily affective experiences. A
second explanation is that there are real
changes in these internal and external experiences.
The interpretive explanation posits an
adolescent who has become more critical and
discerning in reading internal and external
emotional cues, who may be less willing to
label his or her experience with naively positive superlatives, and who is more able to
detect cues indicative of mildly negative
states. Laboratory research indeed demonstrates that, with increasing age, children's
emotional understanding deepens (Nannis,
1988a, 1988b): their lexicon of emotional
terms expands, they make more complex differentiations of the emotions appropriate to
different situations (Harter, 1980; Schwartz &
Trabasso, 1984; Weiner & Graham, 1984), and
they more frequently consider internal processes in their attributions of emotion (Harris,
Olthof, & Terwogt, 1981; Wolman, Lewis, &
King, 1971). It is conceivable, then, that the
less frequent extreme positive ratings among
the adolescents and the slightly increased
variability with age among the girls reflects an
ability to make finer discriminations of emotional experience. This laboratory research,
however, does not easily account for the overall downward shift in average emotional
states.
The second explanation attributes the
age differences in reported states to real
changes in the affective composition of daily
life across the transition to adolescence. That
is, the downward shift in average states may
reflect an alteration in the balance sheet of
positive and negative affective cues that make
up the individual's daily experience. Such an
alteration might be related to a host of normative changes associated with early adolescence: increasing stressful events (BrooksGunn & Warren, 1989; Coddington, 1972),
the hormonal changes of pubescence (Nottelmann et al., 1987), complex interactions occurring with pubertal development (BrooksGunn & Warren, 1989; Petersen & Taylor,
1980; Simmons, Blyth, Van Cleave, & Bush,
References
Adelson, J. (1979, February).Adolescence and the
generalizationgap. Psychology Today, 12, 3337.
Becker, H. S. (1964). Personal change in adult life.
Sociometry,27, 40-53.
Benedict, R. (1938). Continuities and discontinuities in cultural conditioning. Psychiatry, 1,
161-167.
Blos, P. (1961). On adolescence. New York:Free
Press.
Bradburn,N. (1969). The structureof psychological
well-being. Chicago: Aldine.
Brooks-Gunn,J., & Warren,M. P. (1989).Biological
and social contributions to negative affect in
young adolescent girls. Child Development,
60, 40-55.
Campbell, A. (1981). The sense of well-being in
America: Recent patterns and trends. New
York:McGraw-Hill.
1259
213.
Csikszentmihalyi, M., & Larson, R. (1987). Validity
and reliability of the experience-sampling
(Eds.), Social cognition: Studies in the development of understanding (pp. 188-209). Chicago: University of Chicago Press.
Harris, P. L., Olthof, T., & Terwogt, M. M. (1981).
Children's knowledge of emotion. Journal of
1260
Child Development