Professional Documents
Culture Documents
Homesickness at Summer
Camp: Associations with the
Mother-Child Relationship,
Social Self-Concept, and Peer
Relati....
CITATIONS READS
14 96
3 authors, including:
All content following this page was uploaded by Kathryn Kerns on 18 May 2015.
The goal of the present study was to investigate how the quality of the mother-
child relationship, social self-concept, and the quality of peer relationships pre-
dict girls’ feelings of homesickness at a residential summer camp. We expected
that children with secure attachments to their mothers, a more positive social self-
concept, and better relationships with peers would report less homesickness at
camp. Girls were 8 to 12 years of age. Mother-child attachment was assessed
with a questionnaire and an interview completed prior to camp. Social self-
concept was assessed with a questionnaire prior to camp, and peer relationships
were assessed with questionnaires at camp. Peer relationships and social self-
concept, but not mother-child attachment, were related to homesickness at camp.
Social self-concept prior to camp and peer social support and friendship quality
at camp predicted subsequent homesickness, whereas homesickness early in
camp did not predict later assessments of peer relationships at camp. The find-
ings suggest that social self-concept and relationships with peers may be protec-
tive factors for homesickness at summer camp.
473
474 Merrill-Palmer Quarterly
interfere with children’s ability to enjoy experiences away from home and
away from family members. There is a small literature on factors related to
homesickness in children (reviewed in Thurber & Sigman, 1998). Studies
have usually included children attending boarding schools or summer resi-
dential camps. To account for individual differences in homesickness, most
prior studies have focused on child characteristics such as perceived control,
prior experience at camp, or age. For example, several studies have sug-
gested that children who have prior experiences with separation from home
may be less likely to experience homesickness (Fisher, Elder, & Peacock,
1990; Fisher, Frazer, & Murray, 1986; Thurber, 1995; Thurber & Sigman,
1998; Thurber & Weisz, 1997). The most comprehensive and well-tested
theory of homesickness (Thurber & Sigman, 1998; Thurber, Sigman, Weisz,
& Schmidt, 1999) proposes that separation experiences and child attachment
and control attitudes influence the development of a homesick disposition
(i.e., tendency to experience preseparation homesickness), which in turn is
the main factor accounting for homesickness at summer camp.
Only a few studies have considered how children’s interpersonal rela-
tionships may play a role in whether or not children experience homesick-
ness. When relationships have been examined, their presence or absence
has often been the focus. For example, Fisher et al. (1986) examined
whether the presence or absence of siblings was related to children’s adjust-
ment at boarding school. The lack of attention to the quality of children’s
relationships with others is surprising, given the extensive evidence that the
quality of children’s relationships with parents and with peers are both
strongly related to their emotional adjustment (Contreras & Kerns, 2000;
Kerns, in press; Rubin, Bukowski, & Parker, 1998; Thompson, 1998). In
addition, most studies examining the correlates of homesickness have
included adolescents or adults (see review in Thurber & Sigman, 1998), and
therefore less is known about factors related to homesickness in younger
children. The goal of the present study was to investigate how parent-child
and peer relationships are related to homesickness at summer camp. We
examined this question for children in later middle childhood, as this is typ-
ically the earliest age at which children have the experience of attending
overnight camp without parents.
One aspect of parent-child relationships that has been linked to chil-
dren’s social and emotional adjustment is the quality of children’s attach-
ments to parents (Contreras & Kerns, 2000; Kerns, in press; Thompson,
1998). Children who are securely attached are able to use their attachment
figure as a secure base from which to explore and as a haven of safety in
times of stress (Bowlby, 1982). There are several reasons why the forma-
tion of a secure attachment may in turn be associated with lower levels of
Homesickness: Associations in Middle Childhood 475
Method
Sample
The target sample was girls who were enrolled in a six-day, five-night Girl
Scout residential summer camp located in northeastern Ohio. Girls who
would be entering grades 4 to 6 in the fall following camp were invited to
participate in the study. Due to privacy concerns, the research staff did not
contact families directly. Instead, information about the study was provided
to parents with the camp materials, and interested families were asked to
contact the research staff. Families were recruited across two summers.
During the first summer 90 families volunteered for the study, and 73 fami-
lies participated in the precamp data collection. The most common reasons
why families who initially volunteered did not participate in the study were
that the home visit could not be scheduled prior to camp or the families
lived too far away from the university (more than 50 miles). There were 224
spots at camp across the entire summer, but some girls were enrolled for
more than one session (the exact number was not available), and some girls
dropped out prior to or during camp. We estimated that approximately one-
third of the eligible families participated in the study. In the second year of
data collection the same recruitment method was used, with the added crite-
rion that families who had already participated in the study could not partic-
ipate again. An additional 41 families responded, and 33 of these families
participated in the precamp data collection. Of the 106 families for whom
we had precamp data, 97 of the girls attended camp. The latter subsample of
girls and their mothers constituted the sample for the present study.
The girls were between 8 and 12 years old, with a mean age of 10 years
and 3 months (SD = 10.97 months). Ninety-seven percent of the sample
was Caucasian (similar to the demographics for the camp), and most girls
were from intact families (84%). Some girls attended on scholarships, and
thus there was some variability in family income. The family’s socioeco-
nomic status (SES) score was computed using Hollingshead’s Four Factor
Index of Social Status (Gottfried, 1985), which is based on information
about the education level, current occupation, and gender of each parent as
well as marital status. The potential range of score is 8 to 66, with unskilled
laborers ranging from 8 to 19 and business professionals ranging from 55 to
Homesickness: Associations in Middle Childhood 479
66. The mean SES level of our sample was 50 (business managers, lesser
professionals), with a SD = 11 and a range of 17–66.
Procedure
Prior to attending camp, girls and their mothers completed one session with
trained interviewers. The sessions were conducted either at the participant’s
home or at our lab, depending on the family’s preference. At the beginning
of the session mothers signed consent forms indicating their willingness to
participate and to have their daughters participate, and daughters signed
written assent forms. The consent and assent forms stated that the purpose
of the study was to understand how family factors and child characteristics
are related to girls’ adjustment at summer camp. During this visit girls also
completed an attachment security scale, a homesickness scale, and a social
self-concept scale. Furthermore, girls were administered a doll story inter-
view task to assess attachment patterns. Reporting about their daughters,
mothers completed a social self-concept scale and a temperament scale. In
addition, mothers provided information about girls’ previous experience at
camp and demographic data.
Girls attended the camp without parents. While the girls were at camp,
interviewers made three short visits on the first, third, and fifth days of their
six-day stay. The girls completed a homesickness scale on the first, third,
and fifth days of camp. In addition, girls were administered a social support
scale on the third day and a friendship quality scale on the last day of camp.
Number of friendships was assessed on the first and fifth days of camp.
Means, standard deviations, and minimum and maximum values for the
main study measures are included in Table 1.
Precamp Measures
Mother-child attachment. This study included two attachment-based
measures of the mother-child relationship: a questionnaire completed by the
girls and a semiprojective interview administered individually to the girls.
Children completed the Security Scale, a 15-item questionnaire
(Kerns, Aspelmeier, Gentzler, & Grabill, 2001) designed to assess chil-
dren’s perceptions of a particular attachment relationship during middle
childhood. The items were administered using Harter’s “Some kids . . .
other kids . . .” format in which children are presented information about
two types of kids, asked to decide which one is most like them, and then
asked to indicate whether their answers are really true or sort of true. A
sample item from the Security Scale presents is “some kids go to their mom
480 Merrill-Palmer Quarterly
Table 1. Means and Standard Deviations
when they are upset, but other kids do not go to their mom when they are
upset.” Each item was scored from 1 to 4, with a score of 1 indicating lower
security and a score of 4 indicating higher security. The mean of the
answers to all 15 items was computed. Previous research has demonstrated
very good psychometric properties of this measure (Kerns, Klepac, & Cole,
1996; Kerns, Schlegelmilch, Morgan, & Abraham, 2005; Kerns, Tomich,
Aspelmeier, & Contreras, 2000). Alpha for the present study was .81.
Homesickness: Associations in Middle Childhood 481
In addition, all girls completed a doll story interview task (Granot &
Mayseless, 2001). The Doll Play Interview is a story stem task in which an
experimenter begins a story with an attachment theme, using dolls to act out
events (e.g., reunion after a separation from parent), and the child com-
pletes the story. Using the dolls, the child tells what she thinks happens in
the story. The target child, the mother, and siblings are represented in the
stories. The interview includes one warm-up story (birthday party) and five
stories with a parent-child theme: child spills juice at dinner, child gets hurt
after falling off a rock, child sees something in her room after going to bed,
child is left with the sitter after mother departs for three days, and reunion
of mother and child after a three-day separation.
Trained raters later scored the videotaped interview using the manual
for coding the Attachment Doll Story Completion Task (Granot & Mayse-
less, 1999). First, each story was rated secure or insecure based on four cri-
teria. The first criterion is the expression of emotion. In a more secure
relationship, the child is able to express both positive and negative emo-
tions. The second criterion refers to the nature of the mother-child relation-
ship (e.g., is the mother available and responsive to the child’s needs). The
third criterion was whether the events depicted in the stories are resolved in
a positive way. The final criterion is the coherence of the narrative. In a
more secure relationship the narrative is logical and integrated. In insecure
relationships the stories are superficial and pedestrian, contain irrelevant
details, or lack consistency and continuity. After rating each individual
story as secure or insecure, the child was assigned a 5-point rating for four
attachment patterns (security, avoidance, ambivalence, and disorganiza-
tion). A more secure attachment was indicated by at least three stories rated
as secure, including the reunion story. A more avoidant attachment was
assigned if a minimizing strategy, in which a child distances herself, was
used in most of the stories. Children who rated higher on ambivalence dis-
played a heightening strategy involving high levels of distress by the child
and inconsistency in the response of the mother. Children who rated higher
on disorganization presented a chaotic strategy, with either no particular
strategy identifiable or story themes in which catastrophic events occurred
that threatened the availability of the mother. (See the manual for additional
coding details.)
Three raters coded the videotapes. Of the 106 interviews coded, 34
were rated by two observers to check the observer agreement with interrater
agreement (rs = .75—.87). These values are close to those reported by Gra-
not and Mayseless (2001). Previous studies have demonstrated good valid-
ity and test-retest reliability of the measure (Granot & Mayseless, 2001;
Kerns et al., 2005).
482 Merrill-Palmer Quarterly
Social self-concept (in relation to peers). Mothers and their daughters
separately completed Harter’s (1982, 1985) scales of perceived compe-
tence, which assess self-concept in several different domains. The 36-item
measure consists of six scales that measure self-concept (i.e., perceived
competence) in the areas of academics, peers, athletics, physical appear-
ance, behavior, and global self-esteem. Only the 6-item peer scale is rele-
vant for this essay. Items on this scale ask about peer skills (e.g., find it easy
to make friends), the availability of friends (e.g., have a lot of friends), and
peer acceptance (e.g., others their age like them). A high score reflects that
a child has positive perceptions of his or her interactions with peers. The
scale is based on the “Some kids . . . other kids . . .” format. The girls com-
pleted the items with the standard instructions, that is, they were asked to
report their self-perceptions. For each item, after picking which kids they
were like, girls indicated whether the statement was really true or sort of
true for them. Thus, each item was rated on a 4-point scale. The mothers
separately completed the questionnaire. Mother questionnaires were scored
using the same 4-point scale. Mothers were instructed to answer how they
thought their daughters would answer about themselves, even if that was
not how the mothers perceived their daughters. These instructions were
used because in the context of the larger study, this instrument was also
scored for other variables. The alphas for the current study were .77 for the
perceived peer acceptance scale completed by girls and .88 for the same
scale completed by mothers (with mothers reporting how the child per-
ceives herself). Child and mother reports of perceived peer acceptance were
correlated (r = .56, p < .01) and were aggregated to create a single score.
Precamp homesickness. During the precamp session, the girls com-
pleted Thurber and Sigman’s (1998) Rate Your Day measure, which includes
a 3-item homesickness scale. Following the procedures and instructions of
Thurber and Sigman (1998), to obtain a measure of precamp homesickness
we instructed girls to rate their feelings for the prior two weeks. When used
prior to a separation experience, the measure is thought to tap anticipated
homesickness as well as reactions to everyday separations (e.g., during the
school day). Each item was scored from 0 to 10, with a higher score indicat-
ing greater homesickness. Previous research has shown that the measure has
adequate internal consistency and predicts homesickness at camp (Thurber
& Sigman, 1998). The alpha for the current study was .60.
Difficult temperament. We measured difficult temperament prior to
camp so that we could control for general tendencies to experience distress-
ing emotions in our analyses predicting homesickness at camp. Difficult
temperament was measured using McClowry et al.’s (1993) 11-item scale.
The items were based on the Chess and Thomas conceptualization of diffi-
Homesickness: Associations in Middle Childhood 483
Camp Measures
Peer relationships. During the camp visits, four variables were meas-
ured: attending camp with friends, peer social support, quality of best
friendship at camp, and number of friends gained at camp. On the first day
of camp children were given a list of all campers in their group (approxi-
mately 20 girls) and were asked to indicate which other campers they con-
sidered to be a friend. Data were highly skewed in that 52% of girls did not
attend camp with a friend, 28% attended with 1 friend, and 20% attended
with 2 or more friends. This variable was coded as 0 or 1 (1 = attended camp
with at least one friend).
On the third day of camp girls rated social support from their peers
using Malecki, Demaray, and Elliot’s (2002) social support scales. The
scale consists of 12 items that assess four types of social support (emo-
tional, informational, appraisal, and instrumental). Each item is scored
from 1 to 3, with a score of 3 indicating greater perceived social support
from peers. Sample items are “treat me nicely,” “give good advice,” “nicely
tell me when I have made mistakes,” and “ask me to join activities.” The
mean score of all the items was included in the analysis (α = .92). There
was more missing data from Day 3 than Day 1 or Day 5 due to some unan-
ticipated disruptions on that day of camp (during one summer session there
was a severe storm and the camp lost electricity, and during another session
the researcher did not visit due to an illness outbreak at camp that day).
On the fifth day of camp, girls rated the quality of their best friendship at
camp (Bukowski, Hoza, & Boivin, 1994). Two items were removed from the
scale, one from the companionship subscale and one from the helpfulness
subscale, as these items related to events that would not occur while at camp
(i.e., spending time together after school and offering to lend lunch money).
There were 5 subscales: companionship, α = .59, 3 items; helpfulness, α =
.79, 4 items; security, α = .64, 5 items; closeness, α = .62, 5 items; conflict
484 Merrill-Palmer Quarterly
subscale, α = .71, 4 items. The girls rated on a scale from 1 (not true) to 5
(really true) statements such as “My friend would help me if I need it” and
“My friend and I can argue a lot.” Because companionship, helpfulness, secu-
rity, and closeness subscales were highly correlated (rs = .49 – .70), they were
aggregated to obtain a positive friendship qualities score (α = .85).
Finally, on the fifth day of camp children again nominated their friends.
Children were given a list of all campers in their group and marked which
other campers they considered to be a friend. The number of friends gained
at camp was calculated by subtracting the Day 1 total from the Day 5 total.
Homesickness at camp. Girls again completed the Rate Your Day scale,
which included a 3-item homesickness scale (Thurber & Sigman, 1998),
reporting their feelings on Days 1, 3, and 5 of camp. Because of missing
data for homesickness on Day 3 and the high correlation with homesick-
ness measured at Day 3 and Day 5 at camp, we used only Day 1 and Day 5
measures of homesickness in our analyses (αs = .82 and .93, respectively).
T-test analyses indicated that there were differences in the levels of home-
sickness between Day 1 and Day 5. Specifically, homesickness increased
from the first day of camp to the fifth day of camp (t[82] = –3.42, p = .001;
see Table 1 for means).
Results
Preliminary Analyses
The first analysis examined associations among the different attachment-
based measures of the mother-child relationship. Security scores from the
questionnaire measure of attachment were not correlated with the ratings of
attachment patterns from the interview. Ratings from the interview were all
correlated significantly, with the absolute values of correlations ranging
from .21 to .36 except for a higher correlation between security and disor-
ganization (r = –.59). In other research with children, adolescents, or adults,
questionnaire and interview measures have not correlated consistently
(Crowell, Fraley, & Shaver, 1999; Kerns et al., 2007), suggesting that the
two approaches tap different conceptualizations of attachment (i.e., con-
sciously available perceptions vs. unconscious or less readily accessible
beliefs). The security score from the questionnaire and the interview pattern
ratings were analyzed separately in subsequent analyses.
The social self-concept measure and the different measures of peer
relationships also were not highly related to one another (6 of 15 correla-
tions were significant; significant rs ranged from .25 to .35). Girls who
reported a more positive social self-concept in regard to peers prior to camp
Homesickness: Associations in Middle Childhood 485
reported receiving more support from peers at camp (r = .27, p < .05),
reported more positive qualities in their best friendship at camp (r = .25, p <
.05), and made more new friends at camp (r = .35, p < .01). Girls who
reported more positive friendship quality at camp also perceived more sup-
port from peers (r = .35, p < .01), reported less conflict in the friendship (r =
–.27, p < .05), and gained more friends at camp (r = .29, p < .01). Going to
camp with a friend was not related to any of the other measures of peer rela-
tionships. The modest overlap for the peer relationship measures suggests
that they capture somewhat different facets of peer relationships, and we
did not aggregate the measures so that we could examine relations between
homesickness and specific aspects of peer relationships.
Child age, prior experience at camp, difficult temperament, and pre-
camp homesickness were assessed at the precamp home visit. These vari-
ables were included in the study because they might plausibly predict
distress at camp. These child characteristics were not related to one another
except that children with a greater homesick disposition were less likely to
have previously attended a one-week residential camp (r = –.24, p < .05),
and older girls had more prior experience at camp and more often had been
to a one-week residential camp (r = .22, p < .05) or had attended the target
camp (r = .24, p < .05). Prior camp experience and precamp homesickness
were both related to homesickness at camp. Specifically, children who had
attended a one-week residential summer camp before reported less home-
sickness on Day 5 of camp (r = –.21, p < .05). Finally, precamp homesick-
ness was significantly correlated with homesickness on Day 1 and Day 5 of
camp (rs = .40 and .34, respectively, ps < .01).
We also examined associations between the control child characteristics
and our measures of mother-child attachment and peer relationships. There
were four significant correlations for attachment. Older girls were rated as
more secure (r = .37, p < .001) and less disorganized (r = –.38, p < .001) on
the attachment interview. Children rated as more secure or less ambivalent on
the interview were also more likely to have attended the target camp before
(rs = .25 and –.20, respectively, ps < .05). The child characteristics were not
related to peer relationships except that children with more difficult tempera-
ments were lower on perceived peer competence (r = –.24, p < .05).
In the following section we present associations between the mother-
child and peer measures and homesickness at camp. In addition to present-
ing correlations for individual measures, we used regression analyses so
that we could control for some child characteristics when examining these
associations. The regressions also allowed us to examine the total variance
accounted for by the mother-child relationship variables and the peer self-
esteem and peer relationship variables.
486 Merrill-Palmer Quarterly
Mother-Child Relationship, Social Self-Concept, and Peer
Relationships: Correlations with Homesickness at Camp
As a first step to testing our hypotheses, we calculated correlations of the
mother-child and peer relationship measures and social self-concept with
homesickness. Table 2 displays the zero-order correlations between the
main variables and children’s homesickness ratings on the first and fifth
days of camp. Homesickness on Day 1 and Day 5 were not significantly
related to either questionnaire or interview assessments of mother-child
attachment. Because some of the attachment pattern ratings were correlated
with age we also calculated partial correlations between attachment and
homesickness, controlling for age, but the findings did not change.
By contrast, social self-concept and measures of peer relationships
were related to homesickness. Social self-concept in relation to peers (rated
by girls and their mothers), assessed prior to the start of camp, was related
to homesickness on Day 1 and Day 5. Homesickness on Day 1 did not cor-
relate with assessments of peer relationships taken on Day 3 (social sup-
port) or Day 5 (friendship) of camp. However, some of the camp peer
relationship measures were correlated with homesickness on Day 5. More
specifically, children were less likely to experience homesickness on Day 5
if they reported receiving more social support from peers at camp on Day 3
and if they reported a higher quality best friendship at camp on Day 5. Thus,
measures of social self-concept and peer relationships predicted subsequent
feelings of homesickness, but homesickness did not predict subsequent
peer relationships at camp. Although homesickness was related to the qual-
ity of girls’ peer relationships, it was not related to the presence of a home
friend at camp or to the number of friendships gained at camp.1
1 Due to missing data the Ns for the correlations in Table 2 were not the same for Day 1 and
Day 5, nor were they the same for the different peer variables (see Table 2 for pairwise Ns). We
checked to see whether this affected the results by recalculating the peer relationship correlations in
Table 2 including only those girls for whom we had complete peer data as well as homesickness
data for both Day 1 and Day 5 (N = 65). The results did not change.
Homesickness: Associations in Middle Childhood 487
Table 2. Correlations of Mother-Child Attachment, Social Self-Concept, and Peer
Relationships with Homesickness on Day 1 and 5 of Camp
obtained prior to camp. When the peer variables were entered as a block at
Step 2, the test for R2 change was not significant (F[6, 54] = 1.20, p = .32),
although as shown in Table 3 social self-concept did uniquely predict Day 1
homesickness after accounting for the control variables. Social self-concept
and peer relationship measures were more strongly related to homesickness
at Day 5 (Table 4), with the test for R2 change on Step 2 significant (F[6, 66] =
4.93, p < .001). The social self-concept and peer relationship measures
accounted for an additional 26% of the variance in homesickness ratings after
controlling for age, temperament, prior experience at camp, and dispositional
ratings of homesickness (adjusted R2 = .22). As shown in Table 4, the peer
measures collectively accounted for more of the variance in Day 5 homesick-
ness than did the control child characteristics.2
2 Some children were missing homesickness ratings from Day 1 because they arrived at camp
late the first day. As suggested by a reviewer, we checked to see if the different results for the Day 1
and Day 5 regressions were due to different cases in the two analyses. The Day 5 results did not
change when we eliminated the girls who had not contributed data on Day 1, thus eliminating this
explanation for the different results.
Homesickness: Associations in Middle Childhood 489
Discussion
Contrary to our hypotheses, we did not find an association between mother-
child attachment and homesickness at summer camp. Our findings are not
consistent with Thurber and Sigman (1998) and Thurber et al. (1999), who
reported a small but significant association between self-reports of attach-
ment and homesickness. Negative results are difficult to interpret, and it is
important to consider how our study design and methods may have affected
the results. The measures of attachment used in this study, chosen specifi-
cally because they were developed for children in middle childhood, were
different from those used in earlier mixed-age samples. The questionnaire
measure of attachment we used, the Security Scale, has been extensively
validated. For example, in prior research security scores have been found to
be related to other measures of attachment, parenting, and children’s social
and emotional adjustment (for reviews, see Dwyer, 2005; Kerns et al.,
2005). As reported elsewhere, in the present study attachment was related
to other measures of children’s social and emotional adjustment at camp
490 Merrill-Palmer Quarterly
(Abraham & Kerns, 2005; Kerns & Abraham, 2005), further suggesting
that the findings for homesickness are not due solely to measurement prob-
lems. The Doll Play Interview has only been used in two published studies
(Granot & Maysless, 2001; Kerns et al., 2007). Those studies provide some
evidence for the measure’s validity, although findings were much stronger
in an Israeli sample (Granot & Mayseless, 2001) than in a U.S. sample
(Kerns et al., 2007). One important difference in design between the present
study and earlier studies that found an association between self-reports of
attachment and homesickness is that attachment was assessed prior to camp
in our study. It is possible that in studies in which attachment was assessed
at camp (Thurber & Sigman, 1998; Thurber et al., 1999), children’s initial
affective reactions to camp may have influenced their ratings of attachment.
In addition to these methodological considerations, it may be important
as well to consider whether other variables (e.g., peer relationships) have a
more salient influence than does attachment on children’s homesickness at
camp. That is, just as attachment may not predict all aspects of peer rela-
tionships (Ainsworth, 1990), there may also be limits on the contexts in
which attachment predicts emotional adjustment (i.e., the theory does not
exclude the possibility that other factors may be more influential on chil-
dren’s sense of well-being in some contexts). It is also possible that there is
a small true association between attachment and homesickness, but a mod-
est sample size in the present study limited the power to detect this effect.
Our findings do suggest that the importance of attachment for homesick-
ness (e.g., Brewin et al., 1989; Thurber et al., 1999) may have been overem-
phasized in the literature.
The findings were more consistent with the peer social support hypoth-
esis. Homesickness at camp was related to perceptions of social self-
concept obtained prior to camp as well as to the supportiveness of peer
relationships at camp. By contrast, the presence of a friend at the start of
camp and the number of friendships formed at camp were not related to
homesickness. In regression analyses, the social self-concept and peer
relationship measures predicted a substantial amount of the variance in
homesickness at the end of camp, even after controlling for several child
characteristics including children’s tendencies to experience homesickness.
Finally, to assist interpreting the direction of effects, we examined associa-
tion between peer relationships measures and subsequent homesickness as
well as whether homesickness predicted subsequent peer relationships. We
found that peer relationships and social self-concept predicted subsequent
homesickness, whereas homesickness did not predict later assessments of
peer relationships. Although the study design does not allow for strong con-
clusions regarding causal direction of influence, the pattern of results is
Homesickness: Associations in Middle Childhood 491
References
Abraham, M. M., & Kerns, K. A. (2005). Attachment: Links to the expression and
regulation of emotion in preadolescence. Paper presented at the Society for
Research in Child Development, Atlanta, GA.
Homesickness: Associations in Middle Childhood 495
Kerns, K. A., Klepac, L., & Cole, A. (1996). Peer relationships and preadolescents’s
perceptions of security in the child-mother relationship. Developmental Psy-
chology, 32, 457–466.
Kerns, K. A., Schlegelmilch, A., Morgan, T. A., & Abraham, M. M. (2005). Assess-
ing attachment in middle childhood. In K. A. Kerns & R. A. Richardson (Eds.),
Attachment in Middle Childhood (pp. 46–70). New York: Guilford.
Kerns, K. A., Tomich, P. L., Aspelmeier, J. E., & Contreras, J. M.. (2000). Attach-
ment based assessments of parent-child relationships in middle childhood.
Developmental Psychology, 36, 614–626.
Kerns, K. A., Tomich, P. L., & Kim, P. (2006). Normative trends in children’s per-
ceptions of availability and utilization of attachment figures in middle child-
hood. Social Development, 15, 1–22.
Malecki, C. K., Demaray, M. K., & Elliot, S. N. (2002). A working manual of the
development of the Child and Adolescent Social Support Scale. Dekalb: North-
ern Illinois University.
McClowry, S. G., Hegvik, R. L., & Teglasi, H. (1993). An examination of the con-
struct validity of the Middle Childhood Temperament Questionnaire. Merrill-
Palmer Quarterly, 39, 279–293.
Rubin, K. H., Bukowski, W., & Parker, J. G. (1998). Peer interactions, relation-
ships, and groups. In W. Damon (Ed.), Handbook of Child Psychology (5th
ed.), pp. 619–700. New York: Wiley.
Schneider, B. H., Atkinson, L., & Tardif, C. (2001). Child-parent attachment and
children’s peer relations: A quantitative review. Developmental Psychology,
37, 86–100.
Sheldon, A. (1992). Conflict talk: Sociolinguistic challenges to self-assertion and
how young girls meet them. Merrill-Palmer Quarterly, 38, 95–117.
Thompson, R. A. (1998). Early sociopersonality development. In W. Damon (Ed.),
Handbook of child psychology (pp. 25–104). New York: Wiley.
Thurber, C. A. (1995). The experience and expression of homesickness in preado-
lescent and adolescent boys. Child Development, 66, 1162–1178.
Thurber, C. A. (2005). Multimodal homesickness prevention in boys spending 2
weeks at a residential summer camp. Journal of Consulting and Clinical Psy-
chology, 73, 555–560.
Thurber, C. A., & Sigman, M. D. (1998). Preliminary models of risk and protective
factors for childhood homesickness: Review and empirical synthesis. Child
Development, 69, 903–934.
Thurber, C. A., Sigman, M. D., Weisz, J. R., & Schmidt, C. K. (1999). Homesick-
ness in preadolescent and adolescent girls: Risk factors, behavioral correlates,
and sequelae. Journal of Clinical Child Psychology, 28, 185–196.
498 Merrill-Palmer Quarterly
Thurber, C. A., & Weisz, J. R. (1997). “You can try or you can just give up”: The
impact of perceived control and coping style on childhood homesickness.
Developmental Psychology, 33, 508–517.
Waters, H. S., Rodgrigues, L. M., & Ridgeway, D. (1998). Cognitive underpinnings
of narrative attachment assessment. Journal of Experimental Child Psychol-
ogy, 71, 211–234.
Zimmerman, D. R., & Bijur, P. E. (1995). Homesickness and the use of a camp
infirmary: A preliminary report. Developmental and Behavioral Pediatrics,
16, 187–191.