Defensiveness in Response to Aid From a Sibling

advertisement
Journal of Personality and Social Psychology © 1992 by the American Psychological Association, Inc.
March 1992 Vol. 62, No. 3, 422-433
For personal use only--not for distribution.
Defensiveness in Response to Aid From a Sibling
Eileen Searcy
Arizona State University
Nancy Eisenberg
Arizona State University
ABSTRACT
Defensive reactions to aid in the context of a close relationship were studied. Young
adult siblings reported on their self-esteem, the quality of the sibling relationship,
relevant demographic variables, and a variety of components of defensiveness to aid.
Consistent with theory, the components of defensiveness were interrelated. A large
percentage (52%) of the variance in defensiveness was accounted for by conflict
between siblings, dominance relative to one's sibling, low global self-esteem, high
self-esteem when comparing oneself with one's sibling, and low levels of feelings of
entitlement. Siblings were least defensive in reaction to aid from older and female
siblings. Corroboration from a subset of benefactor siblings was obtained. The results
are discussed in relation to recent thinking about the nature of receiving support in
close relationships and recipients' reactions to aid.
In times of need, individuals often do not seek help from those around them, and if
help is offered, it may be rejected or devalued. To explain this behavior, Fisher,
Nadler, and Whitcher-Alagna (1982) proposed the threat to self-esteem model.
Specifically, they have argued that most aid situations contain a mixture of selfsupportive and self-threatening elements and that certain conditions surrounding the
helping interaction determine whether a recipient views the help as primarily selfthreatening or self-supportive. These conditions include the self-relevant messages
contained in the aid itself, amount and efficacy of the aid, high or low need by the
recipient, and donor and recipient characteristics such as the quality of the preexisting
relationship and self-esteem.
According to Fisher et al. (1982) , when help is perceived as more threatening than
supportive, individuals respond with a constellation of defensive behaviors: negative
affect, negative evaluations of the donor and aid, negative reciprocity (reciprocity
motivated by unfavorable feelings toward the donor), high subsequent self-help, low
help seeking, low acceptance of aid ( Fisher et al., 1982 ), and attributions of selfserving motives for the helper ( Nadler, Fisher, & Itzhak, 1983 ; Nadler, Fisher, &
Streufert, 1974 ). These reactions are viewed as effective ways to protect the
recipient's self-esteem because they essentially devalue the source of the self-relevant
information and decrease the likelihood that the individual will be placed in the
uncomfortable recipient role. Although the results of empirical research frequently
have supported aspects of the model, much of the existing data about responses to aid
have been obtained from studies involving experimental manipulations of helping
between strangers in a laboratory. Indeed, the field has been criticized for the dearth
of studies on helping in close relationships ( Clark, 1983 ; Morse, 1972 ) in natural
settings.
Although the construct of self-esteem is central to Fisher et al.'s (1982) model,
research findings on its role in defensiveness have been equivocal. There has been
ongoing controversy concerning whether individuals with low self-esteem or
individuals with high self-esteem are more likely to respond defensively to help.
Some support has been found for the assertion that individuals with high self-esteem
are more defensive because being the recipient of help is inconsistent with their
existing self-cognitions of being competent and independent (the "consistency
prediction"; Fisher et al., 1982 ; Nadler, Fisher, & Streufert, 1976 ; Nadler &
Mayseless, 1983 ; see Nadler, 1986 ). However, some investigators have argued that
people with high self-esteem should be less sensitive than individuals who are low in
self-esteem to any negative information embedded in the receipt of aid (see Tessler &
Schwartz, 1972 , for discussion of the vulnerability hypothesis).
A basic assumption of the threat to self-esteem model is that aid that engenders
feelings of support is not threatening to one's self-esteem and, consequently, is
unlikely to elicit defensiveness. Consistent with this notion, Nadler (1986) noted that
individuals with high and low self-esteem do not differ in reactions to aid if the aid is
not perceived to be potentially threatening. In experimental studies, supportiveness
versus threat in regard to aid usually has been manipulated by varying factors such as
importance of the aid to perceived self-competence (i.e., ego relevancy), opportunity
for the recipient to reciprocate, and similarity of the benefactor to the recipient (see
Nadler & Fisher, 1986 ; Shell & Eisenberg, in press ). However, perceived
supportiveness of the receipt of aid no doubt also is influenced by the quality of the
relationship between benefactor and recipient. Thus, quality of the relationship would
be expected to be a crucial factor in determining whether recipients react defensively
to aid.
Consistent with the aforementioned argument, Wills (1991) has noted that self-esteem
may be less relevant to the receipt of support in close relationships because of a
history of mutual self-disclosure in such relationships and the assumption of a
nonjudgmental response in intimate relationships. Similarly, Clark (1983 ; Clark,
Mills, & Corcoran, 1989) has suggested that in close relationships people may be less
concerned about reciprocity and equitable exchanges than they are in interactions with
strangers; thus, people may be less defensive about many types of help from people
close to them. Indeed, in close relationships, people frequently may feel entitled to the
receipt of help ( Bar-Tall, Bar-Zohar, Greenberg, & Hermon, 1977 ; Morse, 1972 ).
High expectations of help and feelings of entitlement may decrease the likelihood of a
defensive reaction to aid because the inherent threat in being a recipient is minimized
by the attitude that one is entitled to the assistance ( Greenberg & Westcott, 1983 ).
Of course, ongoing relationships involving frequent contact or physical proximity are
not always intimate and supportive. To the degree to which there is conflict,
interpersonal comparisons, or rivalry between people, defensive reactions are likely to
dominate ( Fisher et al., 1982 ). In contrast, to the degree that the relationship is
supportive and nonconflicted, aid is likely to be viewed as social support and as
nonthreatening. However, given that relationships between adult (or nearly adult)
siblings are likely to be at least partially supportive, global self-esteem may be
somewhat less relevant to adult siblings' reactions to aid than to strangers' reactions.
Alternatively, consistent with Tessler and Schwartz's (1972) arguments, self-esteem
might be negatively related to defensive reactions to aid in close relationships because
people with high self-esteem may be less likely than people with low self-esteem to
perceive an implicit comparison (and, hence, threat) as a consequence of receiving aid
from a sibling (particularly if aid involves discussing problems and providing
assistance on everyday problems rather than on specific tasks likely to be viewed as
directly reflecting important competencies). In support of these arguments, Hobfoll
and Lerman (1989) found that parents of chronically ill children with low self-esteem
reported receiving less social support, a finding that suggests that persons with low
self-esteem may respond negatively to the receipt of aid (also see Dunkel-Schetter,
Folkman, & Lazarus, 1987) . Similarly, Cohen, Sherrod, and Clark (1986) found that
social competence and self-disclosure were related to increased perceived social
support over time. Indeed, people with high self-esteem seem to be better able than
people with low self-esteem to secure aid if the aid is not self-threatening ( Nadler,
1991 ).
In close relationships, there are at least two types of self-esteem that may play a role
in moderating recipients' reactions to aid. First, as suggested by Nadler (1986) , global
level of self-esteem may affect whether individuals are defensive about aid, but only
if the aid is potentially threatening to their self-esteem. As discussed above, given the
intrinsic association between the quality of relationships and perceptions of support,
we expected global self-esteem to be either negatively related or unrelated to
defensiveness to siblings' aid (rather than positively related as in many of the
laboratory studies), particularly if the relationship is predominantly supportive.
Because of the supportive elements in many sibling relationships, it is likely that
many types of aid between siblings are viewed as social support rather than as having
direct implications for the recipient's competence.
In addition, recipients' evaluations of themselves on the basis of direct comparison
with a given helper (in contrast to their more enduring sense of self-esteem) may
influence defensiveness. If recipients view themselves as more competent than their
siblings, help from siblings may be threatening because of the inferiority implied by
the receipt of assistance from a less competent person (regardless of the individual's
overall level of self-esteem). However, such an association would be expected
primarily if the relationship is not supportive. In other words, there may be a
difference between individuals' general feelings of self-esteem and self-evaluations
that are derived from comparisons in specific relationships.
Reactions to aid also may differ in systematic ways, depending on the type of
relationship between the helper and the recipient (e.g., parent/child, sibling, or spousal
relationship). A review of the literature on sibling relationships suggests that they are
unique in that they are potentially egalitarian and continue across the lifespan. The
sibling relationship is one in which there is a considerable amount of caretaking,
reciprocity, sharing, cooperating, and helping in both childhood ( Abramovitch,
Corter, & Lando, 1979 ; Abramovitch, Corter, Pepler, & Stanhope, 1986 ) and
adulthood ( Adams, 1968 ; O'Bryant, 1988 ; Ross & Milgram, 1982 ). Although
conflict does exist between siblings in childhood, many researchers have tried to
dispel the myth that rivalry predominates in sibling relationships ( Abramovitch et al.,
1979 ).
Although some researchers have examined sibling relationships in old age ( Gold,
1989 ; O'Bryant, 1988 ), only a few have studied the sibling relationship in early or
mid-adulthood. Adams (1968) found that adult siblings appreciated a certain amount
of tangible aid from their siblings, such as loans or gifts of money or material goods,
but resented too much help of this sort. Similarly, Ross and Milgram (1982) noted that
giving and receiving of aid between siblings resulted in both highly positive feelings
and feelings of rivalry. Moreover, although Cicirelli (1980) found that college-aged
women still preferred their mothers for active help, advice, and leadership functions,
they turned to their siblings just as much as they did to their mothers for emotional
support. The latter finding suggests that even though late adolescence through early
adulthood is a time of growing independence from the family, siblings can be
important as a source of help and emotional support. Unfortunately, however, little is
known about qualities of sibling relationships that are predictive of defensive or
positive reactions to aid. Nonetheless, it seemed reasonable to hypothesize that
siblings are less defensive in their reactions to aid if rivalry and conflict between the
siblings are low and if the relationship is nurturant (i.e., supportive).
Researchers have suggested that sibling roles become more highly differentiated in
the primary school years than in the preschool years and that interactions are
increasingly affected by the sex of the siblings and the age composition of the sibling
dyad. Brody and colleagues ( Brody, Stoneman, MacKinnon, & MacKinnon, 1985 ;
Stoneman, Brody, & MacKinnon, 1984 ) identified a coherent set of roles that
children use when interacting with their siblings: teacher, learner, manager,
managee, helper, helpee, and observer. Not surprisingly, the older sibling most often
takes the roles of teacher, manager, and helper, and older sisters take these roles more
often than older brothers. Vandell, Minnett, and Santrock (1987) identified a similar
set of roles in siblings aged 4 to 11 years and found that with age, siblings became
more equal in power and status as the level of directive guidance provided by older
siblings (e.g., giving instructions and advice) decreased. Companionship and positive
emotional tone increased with age, but so did levels of conflict. It is unclear whether
the effect of sibling constellation continues into adulthood, but it is reasonable to
propose that the "dominant" helper role remains with the older sibling until both
siblings are established in the adult world, partly because of continuity but also
because of the perceived or actual superior competence, knowledge, and experience
of the older sibling.
Consistent with this view, Rosen (1984) identified age difference as one consistent
indicator of differences in status, even in adulthood, and suggested that these status
differences have an effect on the nature of helping interactions. On the basis of
empirical findings (e.g., Druian & DePaulo, 1978 ; also see DePaulo et al., 1989 ),
Rosen argued that individuals generally will reject help from someone younger,
especially if the younger individual is lower in status, because of the implications that
age has in regard to power, experience, knowledge, and resources. Thus, in the
present study, we hypothesized that individuals would respond more defensively to
help from a younger sibling as compared with help from an older sibling.
Additionally, we hypothesized that there would be less defensiveness in response to
aid from a sibling of higher relative status (i.e. the "dominant" sibling) than from a
sibling of lower relative status.
The sex composition of the sibling dyad also was expected to affect recipients'
reactions to aid. Previous research suggests that people prefer female helpers ( Nadler,
Maler, & Friedman, 1984 ; Northman, 1978 ). Brody et al. (1985) observed older
sisters in the helper role three times as often as older brothers, and younger brothers
were never observed to be interacting with their older siblings as a helper in their
study. Thus, the previous prediction regarding the effects of age in combination with
the finding that people prefer female helpers led to the prediction that older sisters are
the most preferred helper, engendering the least defensiveness, whereas younger
brothers in the helper role produce the most defensiveness in siblings.
Method
Subjects
The subjects were 271 university undergraduates (henceforth called Sibling 1
subjects) and, for a subset of the Sibling 1 subjects, their brother or sister ( Sibling 2 ).
Sixty-one percent of the Sibling 1 subjects were women, with a mean age of 19.2
years (range = 17—25 years). Four additional subjects (3 women and 1 man) were
excluded from the analyses because of missing data in many of their questionnaires,
and 2 other subjects who showed up for the testing sessions did not meet inclusion
criteria and were excluded.
There were two subgroups of Sibling 1 subjects. The first group of 75 (24 men and 51
women) was recruited to collect test—retest data (across a 1-month period). These
subjects were not required to provide a sibling's name as part of their participation in
the study; however, 43 agreed to do so. The second group of 196 Sibling 1 subjects
was required to provide the names of full siblings whose ages were within 3 years of
their age as part of their participation (they had been informed when they signed up
for the study that the sibling would be contacted). For these subjects and those in the
test—retest group who agreed to provide a sibling's name, a full sibling whose age
was within 3 years of their age was chosen by the experimenter to act as the Sibling 2
subject. Forty-four percent of the Sibling 2 subjects were younger than Sibling 1 ( M
= 19.6 years old; range = 15—28 years) and 49% were female. Of the 239
questionnaires mailed to Sibling 2 subjects (43 from the test—retest group and 196
from the larger group), 186 (78%) were completed and returned. According to t tests,
there were no significant differences between Sibling 1 subjects whose brother or
sister responded and those whose sibling did not respond on any variables.
Measures
We systematically mixed items for all the scales in the questionnaire, except the help
frequency and evaluation scales and the demographic questions. We created two
forms of the Sibling 1 questionnaires to minimize the possible effects of order of
presentation of the questions. (There were no effects of order for any variables; thus,
order is not considered further.) Table 1 presents descriptive information regarding
each measure.
Components of the defensiveness composite score.
Four components of defensiveness were assessed. 1 First, we developed a 12-facevalid item questionnaire for this study to assess Sibling 1's tendency to refuse help
(Refusal of Help Scale). With these questions, Sibling 1 subjects were asked to rate on
a 5-point scale (1 = completely agree and 5 = completely disagree ) how likely they
would be to refuse (or accept) help from their sibling. Sibling 1 was asked how much
he or she agreed with statements such as "I am less likely to accept help from my
sibling than from other people in my life." Eight of the 12 questions asked specifically
about three types of help: directive guidance ("When my sibling offers me advice, I
don't usually follow it"), emotional support ("I generally accept it when my sibling
gives me emotional support"), and material aid ("I am happy to accept loans or gifts
from my sibling"); the remaining items did not refer to specific types of aid. A high
score on this scale was indicative of a relatively strong tendency to refuse help from
Sibling 2 ( = .86; test—retest reliability = .90).
The Help Seeking Scale assessed two elements of help seeking. First, respondents
rated six questions on a 5-point scale (1 = very likely and 5 = I would never do this )
to indicate how likely they would be to seek out help from their sibling. These six
questions referred to the same three kinds of help assessed in the Refusal of Help
Scale. The remaining five items on this scale assessed perceived discomfort in
seeking help; we adapted these items from Hobfoll and Lerman (1989) . We reworded
Hobfoll and Lerman's questions slightly to refer to the sibling relationship (e.g., "How
difficult is it for you to request emotional support from your sibling if you have
problems or are undergoing a crisis?"). A high score on the Help Seeking Scale
indicated a relatively high likelihood of help seeking and low discomfort with seeking
help from Sibling 2 ( = .90; test—retest reliability = .91).
The third component of defensiveness assessed was Sibling 1's perceptions of Sibling
2's motives for helping (the Negative Attributions of the Helper's Motives Scale). We
adapted three of the six questions in the scale from a study in which Nadler et al.
(1983) examined whether attributions of self-serving motives for a helper would be
related to a more defensive reaction to aid. Nadler et al.'s original questions were
rephrased slightly to refer to the sibling relationship in particular (e.g., "My sister
helps me because she wants me to help her in the future"). Three questions were
added to improve the psychometric properties of the scale, two of which broadened
the scope of the scale by asking the subject about helping motivated by indebtedness
or a commitment to reciprocity (e.g., "My sister helps me because she is paying me
back for helping her"). Sibling I subjects were asked to rate how much they agreed
with each statement on a 5-point scale (1 = completely agree ; 5 = completely
disagree). A high score indicated Sibling 1 scored relatively high on negative
attributions about Sibling 2's motives for helping him of her ( = .79; test—retest
reliability = .84).
We assessed the final component of defensiveness, evaluation of Sibling 2's help,
with a helping inventory we developed for this study. The inventory was based on
Barrera and Ainlay's (1983) model of socially supportive interpersonal interactions.
For this study, we included the highest loading items from Barrera and Ainlay's three
factors: five Tangible Assistance items (physical assistance and providing objects of
value), eight Directive Guidance items (providing information and advice), and eight
Emotional Support items (intimacy, esteem, and trust). Subjects were asked to
indicate how frequently each helping behavior had been provided by their sibling.
(The ratings of frequency of help provided were used as the Frequency of Help Scale,
which was not included in the defensiveness composite score, but was examined
separately;  = .93, test—retest reliability = .91.) If it was a type of aid that their
sibling had provided, then they rated how helpful it was on a 5-point scale from very
helpful to not at all helpful. A 2-year time frame was specified in an attempt to
measure a variety of sibling helping interactions and to minimize the effects of
geographic proximity. The ratings of helpfulness of the aid were used as the
Evaluation of Help Scale. A high score on this scale was indicative of a positive
evaluation of Sibling 2's help ( = .92; test—retest reliability = .88).
Sibling Relationship Questionnaire.
We used three of the factor scales from Furman and Buhrmester's (1985) Sibling
Relationship Questionnaire (SRQ) as measures of the quality of the sibling
relationship. The 12-item Relative Status/Power scale assessed nurturance of Sibling
1 by Sibling 2 (e.g., "How much does your sister show you how to do things you don't
know how to do?"), dominance of Sibling 1 by Sibling 2 (e.g., "How much does your
sister make you do things?"), nurturance of Sibling 2 by Sibling 1, and dominance of
Sibling 2 by Sibling 1. In accordance with Furman's procedures, we computed the
Relative Power/Status scale score by subtracting Sibling 1's perception of nurturance
and dominance received from Sibling 2 from Sibling 1's perception of his or her own
nurturance and dominance of Sibling 2; hence a high Relative Power/Status scale
score reflected the degree of status or power of Sibling 1 over Sibling 2 ( = .83;
test—retest reliability = .90; Furman, Jones, Buhrmester, & Adler, 1989 ). The 9-item
Conflict scale included questions about quarrelling, antagonism, and competition
between the siblings ( = .89; test—retest reliability = .88). The 6-item Rivalry scale
tapped Sibling 1's perceptions of maternal partiality and paternal partiality. These 27
items from the SRQ were unchanged from Furman and Buhrmester's original items (
= .62; test—retest reliability = .87).
Self-concept measures.
We used Rosenberg's (1965) Self-Esteem Scale as a measure of global self-esteem (
= .85; test—retest reliability = .87). A second measure, the Comparative Self-Worth
Scale, was developed to assess Sibling 1's comparison of themselves with Sibling 2
on potentially ego-relevant dimensions: physical attractiveness and athletic ability
(physical domain), intelligence and academic achievement (cognitive domain),
likability in social situations (social domain), competence, and independence. For
each of the 14 items on the Comparative Self-Worth Scale, we asked subjects how
they would rate themselves and their sibling on the dimension on a 1—10 scale, with
a high score indicating that the rating was very positive on that dimension. For
example, Sibling 1 was first asked, "How likeable are you in social situations?" then,
"How likeable is your brother in social situations?" By deliberately placing each of
the sibling-rating questions immediately after each of the self-rating questions, we
structured the measure in such a way as to promote a comparison between the self and
the sibling. The Comparative Self-Worth Scale score was obtained by summing
Sibling 1's ratings of himself or herself and subtracting the sum of the ratings of
Sibling 2 ( = .78; test—retest reliability = .93). 2
Expectations of help.
The four-item Expectations of Help Scale was designed to assess the possibility that
defensiveness is attenuated when siblings feel they are entitled to help from their
brothers and sisters ( Greenberg & Westcott, 1983 ). Items assessed Sibling 1s'
attitudes regarding whether siblings should be expected to help one another and
whether they expected Sibling 2 to help them (e.g., subjects rated on a 5-point scale
how much they agreed with statements such as "My sister helps me more than I
expect her to" and "I don't really expect my sibling to help me" [reversed items];  =
.57; test—retest reliability = .86).
Demographic questions and social desirability.
A series of demographic questions was used to assess the age and sex of Sibling 1 and
Sibling 2, whether they lived with each other or not, and how much contact they had
with each other if they did not live together. Sibling 1s were asked how far they lived
from Sibling 2 geographically (1 = same city and 5 = different state or country ) and
how often they had face-to-face contact with Sibling 2 in the past 2 years (1 = once
per week or more and 5 = never). In addition, we included a 22-item version of the
Marlowe—Crowne Social Desirability Questionnaire ( Crowne & Marlowe, 1964 ) to
examine whether our various measures might be influenced by a subject's tendency to
design their responses with the gaining of social approval in mind ( = .72; test—
retest reliability = .75).
Sibling 2 measures.
To increase the probability of Sibling 2 responding to our questionnaires, only a small
subset of scales were sent to them. Thus, we tried to choose a representative sampling
of measures. To assess defensiveness, we selected the two measures that to some
degree assessed the behavioral components of defensiveness (the Refusal of Help and
the Help Seeking Scales) rather than requiring Sibling 2 to infer their sibling's internal
cognitions about the receipt of aid. The number and content of the items on these
Sibling 2 scales were the same as on the Sibling 1 scales except that Sibling 2 was
asked to report on his or her perceptions of Sibling 1's attitudes and behaviors; s =
.88 for both). In addition, a three-item scale measuring the frequency of help provided
by Sibling 2 to Sibling 1 was also included; a shorter version was used for Sibling 2 to
increase the probability that Sibling 2s would fill out the questionnaires. Sibling 2 was
asked to indicate how frequently (on a 5-point scale from never to very often ) he or
she actually provided emotional support, material aid, and advice or information to
Sibling 1 ( = .70). The final Sibling 2 measure included was the Relative
Status/Power scale from the SRQ ( = .82). 3
Summary of measures.
Thus, the major measures for Sibling 1 included (a) the four components of
defensiveness (Refusal of Help, Help Seeking, Negative Attributions of Helper's
Motives, and Evaluation of Help Scales), (b) the three measures of quality of sibling
relationships (Status/Power, Conflict, and Rivalry Scales), (c) the two measures of
self-esteem (Global Self-Esteem and Comparative Self-Worth Scales), (d) the
Expectations of Help Scale, and (e) the Frequency of Help Scale. In addition, social
desirability and several demographic and sibling constellation variables were
assessed. For Sibling 2, two components of defensiveness (Help Seeking and Refusal
of Help), as well as the Frequency of Help and the Power/Status Scales were
administered.
Formation of the Defensiveness Composite Score
To determine whether we could compute an aggregate Defensiveness Composite
Score, we intercorrelated the four measures of defensiveness (help seeking, negative
attributions of the helper's motives, refusal of help, and a negative evaluation of the
help; see Table 2 ). Note that high Negative Attributions of the Helper's Motives Scale
scores and high Refusal of Help Scale scores were viewed as reflecting defensiveness
whereas low Help Seeking Scale scores and low (or negative) Evaluation of Help
Scale scores were seen as reflecting defensiveness. The intercorrelations were
relatively high, a finding that justified computing a composite measure of
defensiveness. Moreover, according to a principle-components factor analysis, a
single factor best described the data (eigenvalue = 2.88; percentage of factor variance
accounted for = 71.9%).
Additionally, we used confirmatory factor analysis with maximum likelihood
estimation procedures. A chi-square of 10.86 ( df = 2, p < .01) was obtained. A
significant chi-square indicates that the pattern of observed data differs significantly
from the pattern predicted by the hypothesized model. However, chi-square is
extremely sensitive to sample size. As sample size increases, so does the sensitivity of
the test to even small differences between the pattern of observed and predicted data (
Bentler & Bonett, 1980 ; Jöreskog, 1973 ). Thus, an alternative index was used, the
Normed-Fit Index (NFI; Bentler & Bonett, 1980 ), which is not biased when the
sample size is large (i.e., greater than 200). Bentler and Bonett have suggested that
NFIs > .90 indicate adequate fit. An NFI of .98 was obtained, indicating good fit.
To form the composite score, we converted the four scale scores to z scores and
summed them, with help seeking and evaluation of help being negatively coded
(alpha for the four z scores was .87; mean inter-item correlation = .62). 4
Procedure
The questionnaires described above were administered in written form to the Sibling 1
subjects in groups of 10—15 individuals. All Sibling 1 subjects except those in the
test—retest groups were informed before signing up for the study that they would be
asked to provide the name and address of a full sibling whose age was within 3 years
of their age.
At the beginning of the testing session, subjects were required to fill in a form with
the names and ages of siblings whose ages were within 3 years of their age. On the
basis of these data, the experimenter selected which sibling would serve as Sibling 2
(and thereby tried to balance the number of older and younger, and male and female,
siblings), and Sibling 1 was asked to answer the questions in regard to this particular
sibling. In addition, information regarding the age of subjects was obtained so
subjects who did not meet the criterion could be dropped from the study. Subjects in
the test—retest groups returned for a second testing session approximately 1 month
later.
Information was collected from the Sibling 2 subjects using Dillman's (1978)
procedures. Within 1 week after the collection of the Sibling 1 data, a cover letter,
questionnaire, and postage-paid return envelope were mailed to Sibling 2. For those
Sibling 2 subjects under the age of 18 (approximately 40% of the total sample), the
packet was addressed to their parents and the parents were asked to pass the
questionnaire along to Sibling 2. One week after the questionnaire packet was mailed,
a reminder postcard was sent. In their instructions, both Sibling 1 and Sibling 2 were
assured that their answers were totally confidential and would not be shared with their
sibling.
Results
Our analytic strategy consisted of several steps. First, we examined the reliability of
our indexes and the relations between measures completed by both Sibling 1 and
Sibling 2. Next we examined the relations of social desirability and our demographic
variables to defensiveness; these analyses were used to test hypotheses regarding the
relations of sibling constellation and sex to defensiveness and to decide which
variables should be covariates in subsequent analyses. Third, to examine hypotheses
regarding the relations of self-esteem, qualities of the sibling relationship, and
expectations of help to defensiveness in reaction to aid, we performed correlational
and regression analyses. Fourth, in supplemental analyses, we examined the relation
of frequency of help to defensiveness as well as the potential interaction between
supportiveness of the sibling relationship (i.e., conflict) and self-esteem in predicting
defensiveness. Finally, as partial replications, we examined the relations among
Sibling 2 variables and the degree to which Sibling 1 variables predicted Sibling 1's
defensiveness to aid as reported by Sibling 2.
Preliminary Analyses
Means, standard deviations, internal consistency scores, test—retest correlations, and
correlations between Sibling 1 and Sibling 2 reports are presented in Table 1 . The
test—retest correlations were all .75 or higher. Two scales failed to reach an alpha
value of .70, the commonly used criterion of acceptable internal consistency (
Nunnally, 1978 ). However, one of the scales, Expectations of Help, had only four
items, so it was deemed to have acceptable internal consistency ( = .57) in light of
the small number of items ( Nunnally, 1978 ) and its high test—retest reliability
(which is probably a better index of reliability for short scales). The relatively low
reliability for the Rivalry scale (.62) appeared to be due to low variability in the
scores. The Rivalry scale was structured so that a neutral response indicative of no
feelings of rivalry was possible on each question and nearly half (46%) of the subjects
responded with a neutral response set on all items. This severely restricted the
distribution of scores on the Rivalry scale, so conclusions drawn from results for this
scale must be interpreted with caution. 5
As a check of the validity of several of the scales, Sibling 1 measures were correlated
with Sibling 2 variables when possible. The correlations between Sibling 1 and
Sibling 2 reports were moderate to high (.32 to .60; see Table 1 ). (Here, as
everywhere, two-tailed tests were used for the correlational analyses.) However, in
evaluating the correlation for the Frequency of Help Scale, it is important to note that
this scale differed for the 2 siblings. The Sibling 1 Frequency of Help Scale contained
21 items, whereas the Sibling 2 version contained only 3 items.
Demographic Variables
Because of the necessity of some form of contact in order for help to be provided, we
examined the relations of the measures of frequency of contact between Sibling 1 and
Sibling 2 to the major variables in the study. Specifically, subjects who lived with
their siblings and those who did not were compared in regard to the other variables by
means of t tests. There were significant effects for sex of Sibling 2 and constellation, t
s(269) = 2.10 and 2.14, p s < .05, respectively, with female Sibling 2s and younger
Sibling 2s more likely to be living with Sibling 1. Living together was associated with
a higher Sibling 1 report of defensiveness, t (269) = 3.01, p < .01, and higher
comparative self-worth, t (269) = 2.53, p < .05, and with both Sibling 1 and Sibling 2
reports of higher relative status/power for Sibling 1, t (269) = 2.48, and t (184) = 2.11,
p s < .05, respectively. Living together was not significantly related to the other
variables in the study. For siblings who did not live together, closer geographic
distance between the siblings and higher frequency of face-to-face contact likely
served to facilitate the provision of help. These two variables were highly correlated
with one another, r (269) = .61, p < .001, and there was a low, positive correlation
between defensiveness and close geographic distance between the siblings, r (269) =
.15, p < .05. We summed geographic distance and face-to-face contact to form a twoitem Contact Scale ( = .61); this scale was positively correlated with rivalry and
defensiveness, r s(269) = .16 and .13, p s < .01 and .05. As might be expected, more
contact between siblings was correlated with higher rivalry and defensiveness.
Finally, social desirability was significantly correlated with four of the predictor
variables: global self-esteem, conflict, relative power/status, and comparative selfworth, r s(269) = .31, .22. .12, and .13, p s < .001, .001, .05, and .05, respectively,
but not defensiveness r (269) = .09. Because of their relations to the variables of
interest, we used social desirability, whether Sibling 1 and Sibling 2 lived together or
not, and contact (as well as constellation and sex of the siblings) as covariates in
subsequent correlational and regression analyses.
Effects of Sex and Constellation
We performed a 2 (sex of Sibling 1) × 2 (sex of Sibling 2) × 2 (constellation) analysis
to examine the role of sex of the siblings and constellation in relation to
defensiveness. There was a large main effect for sex of Sibling 1; women reported
much less defensiveness ( M = .56) than did men ( M = .91) in response to help from
their siblings, F (1, 263) = 12.78, p < .001. The effect for sex of Sibling 2 was also
significant and in the hypothesized direction; male Sibling 2s (i.e., male donors)
engendered higher defensiveness from Sibling 1 ( M = .58) than did female Sibling 2s
( M = .23), F (1, 263) = 3.89, p < .05. In addition, there was a marginally significant
main effect for constellation in the hypothesized direction; younger Sibling 2s
engendered more defensiveness ( M = .57) than did older Sibling 2s ( M = .22), F (1,
263) = 3.64, p < .06. Finally, there also was a marginally significant interaction
between constellation and sex of Sibling 2; younger male siblings engendered the
most defensiveness ( M = .58), and older female siblings engendered the least ( M =
1.01), F (1, 263) = 3.56, p < .07. As hypothesized, the mean defensiveness scores for
older brothers and younger sisters fell between those of older sisters and younger
brothers ( M s = .27 and .55, respectively). Moreover, according to a planned
comparison of the four cells (older sister, older brother, younger sister, and younger
brother), the mean defensiveness score for the older sister cell was significantly lower
than the mean in the other three cells, F (1, 263) = 3.88, p < .05. The mean
defensiveness scores when Sibling 2 was an older brother, younger brother, or
younger sister were not significantly different. Because of their relation to
defensiveness, sex of Sibling 1, sex of Sibling 2, and constellation were used as
covariates (along with the demographic variables and social desirability) in all
subsequent correlational and regression analyses.
Identification of Predictors of Defensiveness
Six variables (besides the demographic variables) were hypothesized to be predictors
of defensiveness in response to sibling aid: high conflict, high rivalry, disparity in
relative status/power, low expectations of help, disparity in comparative self-worth,
and high global self-esteem. The intercorrelations of the six predictors and their
correlations with defensiveness (partialing constellation, social desirability, and the
various demographic variables) are presented in Table 3 . High conflict and high
relative status/power were both positively correlated with defensiveness, partial r
s(263) = .48 and .20, p s < .001, respectively. Consistent with the prediction that high
expectations of help are related to low defensiveness, the Expectations of Help Scale
was negatively correlated with defensiveness, partial r (263) = .41, p < .001. Rivalry
was not significantly correlated with defensiveness.
Global Self-Esteem was negatively correlated with defensiveness, partial r (263) =
.16, p < .01, whereas comparative selfworth was positively correlated with
defensiveness, partial r (263) = .34, p < .001. 6 On first inspection, the zero-order
correlation of comparative self-worth with defensiveness appeared to be much larger
than the analogous correlation for global self-esteem, r (269) = .36, p < .001, and r
(269) = .15, p < .05, respectively. However, with regression analyses, we discovered
that the two self-concept variables had mutual suppression effects ( Cohen & Cohen,
1983 ) on one another in relation to defensiveness. This limits the conclusions that can
be drawn about the two variables in isolation and precludes a so-called comparison of
their abilities to predict defensiveness, as interpretation of correlations when
suppression effects exist can be misleading.
Cohen and Cohen (1983) cautioned that zero-order effects may be deceptively small
when a suppressor relation exists among three variables, and they suggest that the beta
coefficients are a more accurate reflection of the size of the effect. With the set of
control variables entered on the first step and both global self-esteem and comparative
self-worth entered on the second step in a regression equation (with defensiveness as
the dependent variable), for global self-esteem,  = .34, p < .001, and for
comparative self-worth,  = .46, p < .001. Thus, their relationships with defensiveness
may be closer in magnitude than was indicated by the correlational analyses. The two
variables appear to be measuring different, but related, constructs.
We also computed a regression equation to determine what percentage of the variance
in defensiveness could be accounted for when all the predictor and demographic
variables and covariates were entered in two sets. Results of this analysis are
presented in Table 4 . The set entered on Step 1 was the set of control variables; it
included sex of Sibling 1, sex of Sibling 2, sibling constellation, siblings living
together or not, social desirability, and contact. Four of these six variables had
statistically significant beta weights when entered on Step 1 as a set: sex of Sibling 1
( = .20, p < .001), siblings living together or not ( = .18, p < .05), sex of Sibling
2 ( = .13, p < .05), and Sibling 2 older or younger ( = .12, p < .05). Defensiveness
was higher for male helpers, male recipients, older recipients, and siblings who lived
together. This set of variables accounted for 11% of the variance in defensiveness.
We entered the six predictor variables as a set in the second step of the regression
analysis. Four of the predictors had statistically significant beta weights: conflict ( =
.36, p < .001), expectations of help ( = .31, p < .001), comparative selfworth ( =
.31, p < .001), and global self-esteem ( = .21, p < .001). The level of defensiveness
reported by Sibling 1 was higher when Sibling 1 reported higher conflict with Sibling
2, lower expectations of help from Sibling 2, higher comparative self-worth, and
lower global self-esteem. The six predictor variables accounted for 41% of the
variance in defensiveness over and above the control variables. The control variables
and the predictor variables together accounted for 52% of the variance in
defensiveness. 7
Quality of the Sibling Relationship as a Moderator of the Relation Between SelfEsteem and Defensiveness
Because the relation of self-esteem to defensiveness might be expected to be higher in
conflicted (i.e., nonsupportive) sibling relationships than in harmonious ones, we
examined whether quality of the relationship moderated the relation between our
measures of self-esteem and defensiveness. In two regression equations, the six
control variables were entered in the first step; conflict and one index of self-esteem
(global or comparative) was entered in the second step, and the interaction term
between the given measure of self-esteem and conflict was entered on the third step.
The interaction term did not approach significance for global self-esteem, p < .87;
however, the interaction term was marginally significant for comparative selfworth, F
(1, 261) for change in r 2 = 2.98, p < .085. Because of the conceptual significance of
this interaction for understanding the relation between self-esteem and defensiveness,
we examined the nature of this marginally significant interaction. In the highly
conflicted relationships (i.e., for subjects 1 SD above the mean on conflict), there was
a strong positive correlation between comparative self-esteem and defensiveness,
partial r (35) = .58, p < .001 (partialing the six control variables). For subjects in less
conflicted relationships, this correlation was more modest, albeit also positive, partial
r (220) = .28, p < .001. These findings are consistent with the argument that the
quality of the relationship partially moderates the association between self-esteem and
defensiveness.
Frequency of Help and Defensiveness
Frequency of help was found to be highly positively correlated with evaluation of
help, partial r (263) = .89, p < .001. Frequency of help also was moderately or highly
correlated with the other components of defensiveness: help seeking, partial r (263) =
.70, p < .001 (partialing social desirability, constellation, and the other demographic
variables), negative attributions of the helper's motives, partial r (263) = .41, p <
.001, and refusal of help, partial r (263) = .61, p < .001. In fact, frequency of help
and evaluation of help had virtually identical patterns of relations with the other
variables in the study (see Table 2 ). As might be expected, higher frequency of help
was correlated with lower defensiveness, partial r (263) = .77, p < .001; however,
this strong relation between frequency of help and defensiveness may have been due
to evaluation of help being one component of defensiveness. Frequency of help was
given only minor attention in the analyses because of its lack of independence from
evaluation of help.
Sibling 2 Data
As described previously, all four of the Sibling 2 scales were moderately to highly
correlated with their Sibling 1 counterparts and provided corroboration for Sibling 1's
report. 8 Intercorrelations of the Sibling 2 scales are presented in Table 5 . The pattern
of relations among the four Sibling 2 scales was similar to the pattern of
intercorrelations of their Sibling 1 counterparts. Help seeking and help refusal were
highly negatively correlated, and frequency of help was highly positively correlated
with help seeking and highly negatively correlated with help refusal. The primary
difference between the relations among the Sibling 2 variables and the relations
among the Sibling 1 variables was the lack of significant correlations between Sibling
2s' report of relative power/status and the indicators of defensiveness (help seeking
and help refusal).
In addition, we examined the relations of Sibling 2's assessment of Sibling 1's
defensiveness with the other Sibling 1 predictors. With such an analysis, information
regarding the predictors and the dependent variable were from different informants,
although it should be kept in mind that the Sibling 2 defensiveness composite
contained only two of the four components of defensiveness and the sample size was
smaller in this analysis than in the analogous analysis with all Sibling 1 data. In this
equation, the same six control variables were entered on the first step, F (1, 173) for
change in r 2 = 4.63, p < .001, r 2 change = .13. As in the analysis using all Sibling 1
data, sex of Sibling 2 (being male,  = .22, p < .002) was significantly related to
defensiveness; marginally significant relations were found for sex of Sibling 1 (being
male,  = .13, p < .07) and constellation ( = .14, p < .06). Expectations of help,
rivalry, conflict, power/status, global self-esteem, and comparative self-worth were
entered on the second step, F (1, 173) for change in r 2 = 4.63, p < .001, r 2 change =
.21. Expectations of help was negatively related to defensiveness ( = .22, p < .002),
whereas conflict was positively related ( = .25, p < .002). The measures of selfesteem, rivalry, and power/status did not contribute significantly to prediction. The
total multiple correlation coefficient for the equation was .46 ( p < .0002).
Discussion
The results of this study confirmed the hypothesis that features of the sibling
relationship, as well as an individual's self-esteem, are associated with young adults'
reactions to aid from their siblings. Specifically, 52% of the variance in defensive
reactions could be accounted for by a combination of demographic, self-concept, and
sibling relationship variables. On the basis of these data, one can conclude that adult
sibling relationships are a meaningful context in which to study reactions to aid and
that characteristics of close relationships relate to reactions to aid. These findings are
particularly impressive given the corroboration from a second reporter for some of the
indexes and the fact that relevant demographic variables as well as social desirability
were controlled in the analyses.
Construct of Defensiveness
Consistent with Fisher et al.'s (1982) predictions, low help seeking, negative
attributions regarding a helper's motives, refusal of help, and a negative evaluation of
help tended to be intercorrelated, and all were correlated with a lower level of
reported receipt of help. Thus, as was not the case in previous research, we were able
to construct a multidimensional measure of defensiveness. Using this composite
index, young adults' defensiveness was positively associated with perceived conflict
in the sibling relationship, high status/power of the recipient of aid, and high reported
competence in comparison with that of the helper sibling (i.e., comparative selfesteem). Defensiveness was relatively low for individuals who expected that their
sibling should help.
It is of interest to note that the means for the components of the defensiveness
composite score generally were near or below the midpoint for the scale. In other
words, the study participants as a group did not report high levels of defensiveness in
response to aid from a sibling. This pattern of means is consistent with the conclusion
that individuals are not highly defensive in response to aid from a sibling. It is likely
that they are less defensive than are strangers or persons from dysfunctional families;
however, further research is needed to determine the veracity of this hypothesis. It
should be noted that the subjects used in this study were primarily white students, and
the results of the study may not generalize to other groups.
Self-Esteem and Defensiveness
There was a small but significant association between defensiveness and low global
self-esteem (the magnitude of the association was larger when comparative self-worth
was controlled); hence, our data are not consistent with prior laboratory research
indicating that high self-esteem subjects are more defensive in reactions to aid than
are persons with low self-esteem. Perhaps the differences between our findings and
those discussed by Nadler and Fisher (1986) are due to differences in the samples–
specifically, the use of subjects with a close, long-term relationship rather than
strangers in our study. Thus, it is possible that the types of help assessed in this study
and in the aforementioned studies (i.e., the types of help that are exchanged in close
relationships) are less threatening by nature ( Wills, 1991 ) and, consequently, are
relatively unlikely to induce feelings of threat, even in people with high self-esteem
(see Nadler, 1986 ).
Perhaps, consistent with Clark's (1983) assertions, indebtedness, or pressure to
reciprocate, often is not a relevant construct in the context of helping between siblings
(indeed, the indebtedness scale we used was not very reliable; see Footnote 1 ). Nadler,
Mayseless, Peri, and Chemerinski (1985) proposed that high self-esteem is related to
higher defensiveness because persons with high self-esteem are more committed to
the norm of reciprocity. For individuals with high self-esteem, being the recipient of
help is incongruous with their self-concept; as a result, they feel a strong pressure to
reciprocate or otherwise decrease the threat of help. If the norm of reciprocity does
not operate (or is a less potent motivator) in sibling relationships because of their
close nature, then the relation between high self-esteem and defensiveness may not
exist or may be much weaker.
Similarly, high expectations of help were correlated with low defensiveness,
supporting the notion that people are relatively nondefensive in reaction to aid they
feel entitled to receive ( Greenberg & Westcott, 1983 ). Because helping often is a
part of the socially sanctioned role in close relationships (see Bar-Tal, Bar-Zohar,
Greenberg, & Hermon, 1977 ), the receipt of help within the context of a relationship
may not influence the recipient's self-evaluation. These results may in part explain
why indebtedness did not relate to the other components of the defensiveness (see
Footnote 1
).
Of course, not all sibling relationships are intimate even if siblings are in close
physical proximity. If the relationship is not supportive, aid is likely to be viewed as a
commodity to be reciprocated and, consequently, as potentially self-threatening,
particularly if reciprocation is impossible or difficult. However, we did not find that
quality of the sibling relationship (i.e., level of conflict) moderated the relation
between global self-esteem and defensiveness, although it did so to some degree for
comparative self-worth. Comparative self-worth was more highly correlated with
defensiveness in highly conflicted than in less conflicted and more supportive
relationships.
Although our two measures of self-esteem were positively correlated, they appeared
to measure different, but related, constructs. Whereas global self-esteem was
negatively correlated with defensiveness, comparative self-worth (particularly the
subject's evaluation of his or her sibling's competence, which was only very modestly
correlated with global self-esteem) was positively correlated with defensiveness.
Perhaps people tend not to value help from others they perceive as less competent
than themselves. Alternatively, some people may maintain a positive view of
themselves by devaluing others with whom they compare themselves; such people
also may be especially prone to view helping situations as potentially threatening (i.e.,
as contexts in which there is self-relevant information). These people may be
individuals who are both defensive to aid and falsely score relatively high on selfesteem measures because of their defensiveness (see Dion & Dion, 1985 ; Nadler,
1986 ); in other words, they do not truly have high self-esteem. In addition, it is
possible that self-esteem in the context of a close relationship is somewhat
independent of individuals' global self-esteem (as is suggested by the modest
correlation between the two) and that feelings of self-esteem that are attached to or
derived from a particular dyadic relationship may be especially relevant to
individuals' defensiveness in reaction to the receipt of aid from the other person in the
dyad.
Quality of the Relationship and Defensiveness
As noted previously, Sibling 1's report of the level of conflict in his or her sibling
relationship was found to be positively correlated with defensiveness in response to
sibling aid, as was the recipients' status or dominance over Sibling 2 (in the
correlational analyses). Status implies some degree of greater competence on the part
of the dominant sibling. Thus, as predicted, help from an "inferior" or "subordinate"
sibling was associated with greater defensiveness, and older siblings were more likely
to be rated as dominant than were younger siblings.
It could be argued that the association between sibling conflict and defensiveness
existed because both reflect a negative, antagonistic sibling relationship. Certainly, a
supportive relationship would be expected to increase the likelihood that aid from one
member of a dyad to another is viewed as supportive. However, some researchers
have argued that the relation between these two variables is more complex. Tesser
(1980) has proposed that friction between siblings is due in part to attempts to reduce
the threat to self-esteem that takes place when social comparisons are made between
siblings. He found that friction was higher when one sibling outperformed the other as
compared with when their performances were equal. There also was more friction
when a younger sibling outperformed an older sibling as compared with the reverse
situation. Tesser suggested that conflict occurs to decrease the threat of sibling
comparison by decreasing "closeness" and thereby decreasing the relevance of the
social comparison. Thus, conflict may be related to defensiveness because it is one
mechanism by which the threat of being the recipient of help from a sibling can be
attenuated.
Contrary to expectations, the Rivalry scale of the SRQ was not significantly
correlated with defensiveness. It is unclear whether this pattern of findings was due to
inadequacies in the scale itself (see Footnote 5 ) or whether the construct of rivalry is not
a relevant one for young adults. The latter explanation is inconsistent with Ross and
Milgram's (1982) finding that 71% of adult respondents described current feelings of
rivalry toward siblings. In contrast, the former explanation is consistent with their
finding that rivalry in childhood was generally rooted in parental favoritism whereas
rivalry in adulthood was more likely to occur as a result of overt comparison of
individual accomplishments. Changes in the bases of sibling rivalry as siblings mature
were not reflected in our rivalry scale; thus, the scale may have been a relatively weak
measure of rivalry in adulthood.
Sex, Age, and Siblings' Defensive Reactions to Aid
Consistent with the limited literature ( Balls & Eisenberg, 1986 ; Nadler et al., 1984 ;
see McMullen & Gross, 1983 ), sex of the recipient was related to siblings' reactions
to aid. Specifically, women, in comparison with men, reported higher levels of help
seeking, a more positive evaluation of help, lower levels of help refusal, and lower
levels of defensiveness overall. Because of societal stereotypes ( Martin, 1987 ),
women may be more comfortable than men in the dependent role (indeed, they are in
the role of recipient more often than are men; Eagly & Crowley, 1986 ). Alternatively,
it is possible that women, who appear to value intimacy more than men do (
Buhrmester & Furman, 1987 ), responded more positively to aid because intimacy is
associated with greater satisfaction with social support ( Hobfoll, Nadler, &
Lieberman, 1986 ). Such an explanation is supported by the finding that women had
more contact with their siblings.
Both men and women reported less defensiveness when the helper sibling was female.
Moreover, older siblings engendered less defensiveness in recipients than did younger
siblings, with older sisters engendering the least. Interestingly, neither men nor
women reported expecting more help from sisters than from brothers, which suggests
that the increased help that is sought or accepted from sisters as compared with
brothers may be due to a preference for help from a sister rather than a sense that
sisters are more obligated to help. Because women generally are viewed as more
nurturant than men in our culture ( Martin, 1987 ), help from a woman may seem
natural, especially if the help is in the form of emotional support ( Eagly & Crowley,
1986 ). Older siblings and women may have provoked less defensiveness because
girls and older children tend to enact the helper role in the home more often than boys
or younger children during childhood ( Brody et al., 1985 ; Stoneman et al., 1984 ;
Whiting & Whiting, 1975 ), with the consequence that siblings are used to receiving
assistance from their sisters, particularly their older sisters. Moreover, siblings may be
less defensive when their helper is an older sibling because the older sibling may be
perceived as having greater knowledge and experience ( Druian & DePaulo, 1978 ;
Rosen, 1984 ).
Summary
The results of this study provide support for aspects of Nadler and Fisher's (1986)
model, particularly for their construct of defensiveness. Nonetheless, the data
highlight the importance of considering the type of relationship involved when
examining reactions to aid. The quality of the relationship may moderate recipients'
reactions to aid; moreover, the relation of self-esteem to defensiveness to aid may
vary depending on the nature of the relationship between recipient and benefactor.
Several limitations of this study should be noted. First, some of the siblings discussed
the study with one another, which may have affected some of the data (see Footnote 8 ).
Moreover, the data were all self-report or other-report data, and all information was
obtained from relatively young college students and siblings who were fairly close in
age. It is possible that the findings may differ for a differently aged sample or if
behavioral measures were obtained (i.e., self-reported measures may not accurately
reflect reality). Indeed, it is likely that both the quality of sibling relationships and
siblings' reactions to one another's aid vary across the life span. Moreover, if
similarity increases defensiveness ( Nadler & Fisher, 1986 ), it is possible that more
widely spaced siblings are less defensive than closely spaced siblings, particularly in
reaction to aid from an older sibling. Thus, the results of the present study emphasize
the need for research on helping in close relationships and the role of quality of the
relationship in recipients' reactions to aid. Finally, it is clear from our findings that
various facets of self-esteem need to be examined in future research on reactions to
aid.
References
Abramovitch, R., Corter, C. & Lando, B. (1979). Sibling interaction in the home.
Child Development, 50, 997-1003.
Abramovitch, R., Corter, C., Pepler, D. J. & Stanhope, L. (1986). Sibling and peer
interaction: A final follow-up and comparison. Child Development, 57, 217-229.
Adams, B. N. (1968). Kinship in an urban setting. (Chicago: Markham)
Balls, P. & Eisenberg, N. (1986). Sex differences in recipient's reactions to aid. Sex
Roles, 14, 69-79.
Barrera, M. & Ainlay, S. L. (1983). The structure of social support: A conceptual and
empirical analysis. Journal of Community Psychology, 11, 133-143.
Bar-Tal, D., Bar-Zohar, Y., Greenberg, M. S. & Hermon, M. (1977). Reciprocity
behavior in the relationship between donor and recipient and between harm-doer and
victim. Sociometry, 40, 293-298.
Bentler, P. M. & Bonett, D. G. (1980). Significance tests and goodness of fit in the
analysis of covariance structures. Psychological Bulletin, 88, 588-606.
Brody, G. H., Stoneman, Z., MacKinnon, C. E. & MacKinnon, R. (1985). Role
relationships and behavior between preschool-aged and school-aged sibling pairs.
Developmental Psychology, 21, 124-129.
Buhrmester, D. & Furman, W. (1987). The development of companionship and
intimacy. Child Development, 58, 1101-1113.
Cicirelli, V. G. (1980). A comparison of college women's feelings toward their
siblings and parents. Journal of Marriage and the Family, 42, 111-118.
Clark, M. S. (1983). Some implications of close social bonds for help-seeking.(In B.
M. DePaulo, A. Nadler, & J. D. Fisher (Eds.), New directions in helping: Vol. 2.
Help-seeking (pp. 205—233). San Diego, CA: Academic Press.)
Clark, M. S., Mills, J. R. & Corcoran, D. M. (1989). Keeping track of needs and
inputs of friends and strangers. Personality and Social Psychology Bulletin, 15, 533542.
Cohen, J. & Cohen, P. (1983). Applied multiple regression/correlation analysis for
the behavioral sciences. (Hillsdale, NJ: Erlbaum)
Cohen, S., Sherrod, D. R. & Clark, M. S. (1986). Social skills and the stressprotective role of social support. Journal of Personality and Social Psychology, 50,
963-973.
Crowne, D. P. & Marlowe, D. (1964). The approval motive: Studies in evaluative
dependence. (New York: Wiley)
DePaulo, B. M., Tang, J., Webb, W., Hoover, C., Marsh, K. & Litowitz, C. (1989).
Age differences in reactions to help in a peer tutoring context. Child Development, 60,
423-439.
Dillman, D. A. (1978). Mail and telephone surveys: The total design method. (New
York: Wiley)
Dion, K. K. & Dion, K. L. (1985). Personality, gender and the phenomenology of
romantic love.(In P. Shaver (Ed.), Review of personality and social psychology (Vol.
6, pp. 209—239). Beverly Hills, CA: Sage.)
Druian, P. R. & DePaulo, B. M. (1978). Asking a child for help. Social Behavior and
Personality, 5, 33-39.
Dunkel-Schetter, C., Folkman, S. & Lazarus, R. S. (1987). Correlates of social
support receipt. Journal of Personality and Social Psychology, 53, 71-80.
Eagly, A. H. & Crowley, M. (1986). Gender and helping behavior: A meta-analytic
review of the social psychological literature. Psychological Bulletin, 100, 283-308.
Fisher, J. D., Nadler, A. & Whitcher-Alagna, S. (1982). Recipient reactions to aid.
Psychological Bulletin, 91, 27-54.
Furman, W. & Buhrmester, D. (1985). Children's perceptions of the qualities of
sibling relationships. Child Development, 56, 448-461.
Furman, W., Jones, L., Buhrmester, D. & Adler, T. (1989). Children's, parents', and
observers' perspectives on sibling relationships.(In P. G. Zukow (Ed.), Sibling
interaction across cultures (pp. 165—183). New York: Springer-Verlag.)
Gold, D. T. (1989). Sibling relationships in old age: A typology. International
Journal of Aging and Human Development, 28, 37-51.
Greenberg, M. S. & Westcott, D. R. (1983). Indebtedness as a mediator of reactions to
aid.(In J. D. Fisher, A. Nadler, & B. M. DePaulo (Eds.), New directions in helping:
Vol. 1. Recipient reactions to aid (pp. 85—112). San Diego, CA: Academic Press.)
Hobfoll, S. E. & Lerman, M. (1989). Predicting receipt of social support: A
longitudinal study of parents' reactions to their child's illness. Health Psychology, 8,
61-77.
Hobfoll, S. E., Nadler, A. & Lieberman, J. (1986). Satisfaction with social support
during crisis: Intimacy and self-esteem as critical determinants. Journal of Personality
and Social Psychology, 51, 296-304.
Jöreskog, K. (1973). A general method for estimating a linear structural equation
system.(In A. Goldberger & O. Duncan (Eds.), Structural equation models in the
social sciences (pp. 85—112). New York: Seminar Press.)
Martin, C. L. (1987). A ratio measure of sex stereotyping. Journal of Personality and
Social Psychology, 52, 489-499.
McMullen, P. A. & Gross, A. E. (1983). Sex differences, sex roles, and health-related
help-seeking.(In B. M. DePaulo, A. Nadler, & J. D. Fisher (Eds.), New directions in
helping. Vol. 2. Help-seeking (pp. 233—263). San Diego, CA: Academic Press.)
Morse, S. J. (1972). Help, likability, and social influence. Journal of Applied Social
Psychology, 2, 34-46.
Nadler, A. (1986). Self-esteem and the seeking and receiving of help: Theoretical and
empirical perspectives.(In B. Maher & W. Maher (Eds.), Progress in experimental
personality research (Vol. 14, pp. 115—163). San Diego, CA: Academic Press.)
Nadler, A. (1991). Help-seeking behavior: Psychological costs and instrumental
benefits.(In M. Clark (Ed.), Review of personality and social psychology: Vol. 12.
Prosocial behavior (pp. 290—311). Newbury Park, CA: Sage.)
Nadler, A. & Fisher, J. D. (1986). The role of threat to self-esteem and perceived
control in recipient reaction to help: Theory development and empirical validation.
Advances in experimental social psychology (Vol. 19, pp. 81—123).(San Diego, CA:
Academic Press)
Nadler, A., Fisher, J. D. & Itzhak, S. B. (1983). With a little help from my friend:
Effect of single or multiple act aid as a function of donor and task characteristics.
Journal of Personality and Social Psychology, 44, 310-321.
Nadler, A., Fisher, J. D. & Streufert, S. (1974). The donor's dilemma: Recipients'
reactions to aid from friend or foe. Journal of Applied Social Psychology, 4, 275-285.
Nadler, A., Fisher, J. D. & Streufert, S. (1976). When helping hurts: The effects of
donor-recipient similarity and recipient self-esteem on recipient reactions to aid.
Journal of Personality, 44, 392-409.
Nadler, A., Maler, S. & Friedman, A. (1984). Effects of helper's sex, subject's
androgyny, and self-evaluation on males' and females' willingness to seek and receive
help. Sex Roles, 10, 327-339.
Nadler, A. & Mayseless, O. (1983). Recipient self-esteem and reactions to help.(In J.
D. Fisher, A. Nadler, & B. M. DePaulo (Eds.), New directions in helping: Vol. 1.
Recipient reactions to aid (pp. 167—189). San Diego, CA: Academic Press.)
Nadler, A., Mayseless, O., Peri, N. & Chemerinski, A. (1985). Effects of opportunity
to reciprocate and self-esteem on help-seeking behavior. Journal of Personality, 53,
23-35.
Northman, J. E. (1978). Developmental changes in preference for help. Journal of
Clinical Child Psychology, 7, 129-132.
Nunnally, J. C. (1978). Psychometric testing. (New York: McGraw-Hill)
O'Bryant, S. L. (1988). Sibling support and older widows' well-being. Journal of
Marriage and Family, 50, 173-183.
Rosen, S. (1984). Some paradoxical status implications of helping and being
helped.(In E. Staub, D. Bar-Tal, J. Karylowski, & J. Reykowski (Eds.), Development
and maintenance of prosocial behavior (pp. 359—378). New York: Plenum Press.)
Rosenberg, M. (1965). Society and the adolescent self-image. (Princeton, NJ:
Princeton University Press)
Ross, H. G. & Milgram, J. I. (1982). Important variables in adult sibling relationships:
A qualitative study.(In M. E. Lamb & B. Sutton-Smith (Eds.), Sibling relationships:
Their nature and significance across the life span (pp. 225—249). Hillsdale, NJ:
Erlbaum.)
Shell, R. & Eisenberg, N. (in press). A developmental model of recipients reactions to
aid. Psychological Bulletin., ,
Stoneman, Z., Brody, G. H. & MacKinnon, C. (1984). Naturalistic observations of
children's activities and roles while playing with their siblings and friends. Child
Development, 55, 617-627.
Tesser, A. (1980). Self-esteem maintenance in family dynamics. Journal of
Personality and Social Psychology, 39, 77-91.
Tessler, R. C. & Schwartz, S. H. (1972). Help seeking, self-esteem, and achievement
motivation: An attributional analysis. Journal of Personality and Social Psychology,
21, 318-326.
Vandell, D. L., Minnett, A. M. & Santrock, J. W. (1987). Age differences in sibling
relationships during middle childhood. Journal of Applied Developmental
Psychology, 8, 247-257.
Whiting, B. B. & Whiting, J. W. M. (1975). Children of six cultures A Psychocultural
analysis. (Cambridge, MA: Harvard University Press)
Wills, T. A. (1991). Social support and interpersonal relationships.(In P. S. Clark
(Ed.), Review of personality and social psychology: Vol. 12. Prosocial behavior (pp.
265—289). Newbury Park, CA:Sage.)
1
A measure of perceived indebtedness adapted from Nadler, Mayseless, Peri, &
Chemerinski's (1985) "commitment to the norm of reciprocity" scale was originally
tested as part of the Defensiveness score, but was excluded because of low internal
consistency and a lack of corroboration from Sibling 2.
2
Sibling 1 subjects were also asked to rank each of the dimensions in order of
importance to them as a measure of "ego relevance" of the dimensions. A comparison
was made between the overall score and scores obtained from only the four highest
ranked dimensions. There were no significant differences between the two scores in
their correlations with any of the other variables (the difference in the values of the r s
ranged from .00 to .08); hence, the overall comparative self-esteem score was used in
subsequent analyses.
3
We pilot-tested the scales described here on 53 subjects before conducting this study.
As a result of the pilot study, two items were added to the Negative Attributions of the
Helper's Motives Scale, increasing the number of items from four to six and
increasing the value of alpha from .64 to .79. Two items with low inter item
correlations were deleted from the Expectations of Help Scale, increasing the value of
alpha from .42 to .57.
4
Additionally, we factor analyzed all the major Sibling 1 variables. For factor loadings
of .50 or higher, the first factor contained Refusal of Help (.88), Help Seeking (.87),
Evaluation of Help (.77), Negative Attributions of Help (.75), Expectations of Help
(.58), and Conflict (.56). The second factor included Global Self-Esteem (.78) and
Comparative Self-Worth (.79). The third factor contained Rivalry (.64), Conflict (.52),
and Power/Status (.52).
5
Furman and Buhrmester (1985) and Furman et al. (1989) have acknowledged the
weakness of the Rivalry scale, as the eigenvalue of the Rivalry factor in the original
factor analysis of the SRQ was actually below the "generally acceptable" cutoff level
for inclusion as a separate factor. In addition, the issue of parental partiality was
prominent for only 20% of the school-aged siblings they interviewed in developing
their questionnaire. They have since undertaken to develop a revised Rivalry measure
( Furman et al., 1989 ).
6
Because of the importance of level of self-esteem in theory on recipients' reactions to
aid and the possibility that persons with both relative high and low self-esteem might
react defensively to aid, we also used regression analyses to examine whether the two
self-concept variables had curvilinear relationships with Defensiveness. The quadratic
effects were not significant.
7
We also examined the relations of the two components of the Comparative SelfWorth Scale (self- and other-ratings) to Defensiveness. Only ratings of the sibling
were significantly correlated with Defensiveness, partial r (263) = .42 (for the selfratings, partial r [263] = .04). Moreover, the ratings of the sibling functioned
similarly to the total Comparative Self-Worth Scale in the regression analyses, albeit
their prediction was slightly lower. Global self-esteem was highly related to selfratings, r (269) = .61, p < .001, but not ratings of one's sibling, r (269) = .16, p < .01.
8
Forty-one percent of the Sibling 2 subjects reported that Sibling 1 discussed the study
with them before they completed the questionnaire (possibly to inform them that the
questionnaires would be sent). Results of t tests comparing those Sibling 2 subjects
who discussed the study with Sibling 1 and those who did not showed a significant
difference, p < .05, on only one scale, Refusal of Help. However, when alpha was
adjusted for the number of comparisons made to avoid a Type I error (.05/12 =
.0004), the means for the two groups were not significantly different on any of the
scales.
Eileen Searcy is now at the Medical Center Rehabilitation Hospital, Child Evaluation
and Treatment Program, University of North Dakota.
This study was completed by Eileen Searcy in partial fulfillment of the requirements
for a doctoral degree.
This research was supported by National Science Foundation Grant BNS8807784,
National Institute of Child Health and Development Career Development Award K04
HD00717, and National Institute of Mental Health Research Scientist Development
Award K02 MN00903-01 to Nancy Eisenberg. We would like to express our
appreciation to Linda Ruehlman for her statistical assistance and to Paul Karoly,
Carol Nemeroff, and Linda Ruehlman for their input.
Correspondence may be addressed to Eileen Searcy, Child Evaluation and Treatment
Program, University of North Dakota, University Station, Grand Forks, North Dakota,
58202.
Received: February 19, 1991
Revised: October 24, 1991
Accepted: November 5, 1991
Table 1.
Table 2.
Table 3.
Table 4.
Table 5.
Download