The Self-Fulfilling Prophecy as an Intrafamily Dynamic

Abstract

This research examined whether parents' and children's perceptions of one another have reciprocal self-fulfilling effects on each other's behavior. Parents and their adolescent children completed self-report surveys and engaged in dyadic videotaped interaction tasks. The surveys assessed mothers', fathers', and children's perceptions of their own and the other's hostility and warmth. Observers coded the videotaped interactions to assess the actual hostility and warmth exhibited by parents and children. Data from 658 mother-child dyads were consistent with the conclusion that children had a self-fulfilling effect on their mothers' hostility but that mothers did not have a reciprocal self-fulfilling effect on their children's hostility. No other self-fulfilling prophecy effects emerged. Findings are discussed in terms of family relations and the differential power of negative versus positive self-fulfilling prophecies.

Full text

The Self-Fulfilling Prophecy as an Intrafamily Dynamic
Stephanie Madon, Max Guyll, and Richard L. Spoth
Iowa State University
This research examined whether parents’ and children’s perceptions of one another have
reciprocal self-fulfilling effects on each other’s behavior. Parents and their adolescent
children completed self-report surveys and engaged in dyadic videotaped interaction tasks.
The surveys assessed mothers’, fathers’, and children’s perceptions of their own and the
other’s hostility and warmth. Observers coded the videotaped interactions to assess the actual
hostility and warmth exhibited by parents and children. Data from 658 mother–child dyads
were consistent with the conclusion that children had a self-fulfilling effect on their mothers’
hostility but that mothers did not have a reciprocal self-fulfilling effect on their children’s
hostility. No other self-fulfilling prophecy effects emerged. Findings are discussed in terms
of family relations and the differential power of negative versus positive self-fulfilling prophecies.
keywords: self-fulfilling prophecies, hostility, family relations, interpersonal interaction, dyads
Enduring dyadic relationships emerge and develop as a
result of recurring interactions. As a result of having expe-
rienced numerous interactions in the past, individuals in an
ongoing relationship form a relatively stable set of percep-
tions of the relationship, such as whether it tends to be
characterized by hostility versus warmth, distance versus
closeness, or competition versus cooperation (Hinde &
Stevenson-Hinde, 1987). These perceptions imply expecta-
tions for future interactions, including expectations for
one’s own behavior, as well as expectations for the other
person’s behavior (Hinde, 1979; Lollis, 2003). Relation-
ships and the interpersonal perceptions and expectations
they entail are important because they provide a context
whereby one can readily interpret and assign meaning to the
events that transpire in the course of each new interaction
that occurs. As such, perceptions and expectations have the
potential to actively shape one’s behavior during an inter-
action such that it can alter the interaction itself (Lollis,
2003). If such a process were to occur reliably, then it is
conceivable that the perceptions and expectations that peo-
ple hold for one another could eventually change their
relationship. One interpersonal process through which such
change might occur is the self-fulfilling prophecy.
Self-Fulfilling Prophecies
A self-fulfilling prophecy is an erroneous belief that leads
to its own fulfillment (Merton, 1948). A self-fulfilling
prophecy consists of a sequence of three steps. First, one
person (the perceiver) must develop an inaccurate belief
about another person (the target). Second, the perceiver
must treat the target in a manner that is consistent with the
inaccurate belief. Third, the target must respond to the
perceiver’s treatment by confirming the originally inaccu-
rate belief. Numerous experimental and naturalistic studies
have demonstrated the existence of self-fulfilling prophecies
(for reviews, see Jussim, Eccles, & Madon, 1996; Snyder &
Stukas, 1999). However, this research has focused exclu-
sively on unilateral effects, investigating how the inaccurate
beliefs of one person shape the behavioral outcomes of a
second person and not considering how the second person’s
inaccurate beliefs might simultaneously shape the behav-
ioral outcomes of the first person. Although a unilateral
focus is appropriate for studies in which the outcome of
interest is only relevant to the target (e.g., children’s aca-
demic achievement), outcomes that are relevant to both
perceivers and targets are more appropriately studied with
an analytic approach that assesses bidirectional effects. For
example, both parents and children may hold inaccurate
perceptions of the others’ negative or positive affective
behaviors, and both may be affected by the others’ percep-
tion. In cases such as this, individuals are simultaneously
both perceiver and target. A unilateral approach to studying
such relations is not appropriate because it has the potential
to produce misleading results, such as overestimating the
self-fulfilling prophecy effect that one person has on anoth-
er’s behavior.
Bilateral Model of Parent–Child Relations
The likelihood of bidirectional influence has long been
recognized by both the self-fulfilling prophecy and parent–
child relations literatures (e.g., Bell, 1971; Snyder & Stukas,
1999). For example, whereas historical perspectives por-
trayed children as passive recipients of the effects of more
Stephanie Madon, Department of Psychology, Iowa State Uni-
versity; Max Guyll and Richard L. Spoth, Partnerships in Preven-
tion Science, Iowa State University.
We thank Kenneth Bollen for his thoughtful and thorough re-
sponses to our questions regarding the calculation of specific effects.
Correspondence concerning this article should be addressed to
Stephanie Madon, Department of Psychology, Iowa State Univer-
sity, Research Park Building 2, 2625 North Loop Drive, Suite 500,
Ames, IA 50010-8296. E-mail: madon@iastate.edu
Journal of Family Psychology Copyright 2004 by the American Psychological Association
2004, Vol. 18, No. 3, 459–469 0893-3200/04/$12.00 DOI: 10.1037/0893-3200.18.3.459
459

powerful parents, the bilateral model (Kuczynski, 2003)
proposes that (a) bidirectional effects exist, (b) both parents
and children are willful agents of influence, and (c) children
have resources whereby they can influence their parents,
even though the parents ultimately have more power in the
relationship. Despite this trend in theoretical perspectives,
there has been little in the way of empirical analysis that has
directly examined reciprocal effects within specific interac-
tions. Therefore, the current study investigates relations
between parents and their children during a specific inter-
action using the bilateral model as the guiding framework.
Overview of the Current Investigation and
Hypotheses
This research examines whether parents’ and children’s
perceptions of one another have reciprocal self-fulfilling
prophecy effects on the others’ behavior during mother–
child and father–child interactions. This issue is examined
within the framework of the bilateral model of parent–child
relations in that reciprocal effects between parent and child
are included in the model. To the best of our knowledge, no
previous study has examined reciprocal self-fulfilling
prophecies. The outcome variables of interest are the
observer-rated hostility and warmth that parents and chil-
dren express to one another during their interactions. The
importance of examining how self-fulfilling prophecies
might influence the affective quality of parent–child inter-
actions is demonstrated by the link between parent–child
affect and important child outcomes. For example, negative
parent–child relationships reduce parent–child closeness,
inhibit the child’s attachment to the parent, and impede the
child’s internalization of parental values—factors that in-
crease children’s risk of developing problem behaviors
(Catalano & Hawkins, 1996).
Method
Participants
Participants were 658 mothers, 576 fathers, and 680 children
who were participating in a larger study.
1
At the time of the study,
children were in early adolescence (M ⫽ 12.3 years, SD ⫽ 0.5
years), all in the 7th grade; the sample included 313 (46%) girls
and 367 (54%) boys. Only one child in each family provided data.
Mothers averaged 39 years of age (SD ⫽ 5.5 years). Fathers
averaged 41 years of age (SD ⫽ 6.1 years). Mothers and fathers
averaged between 1 and 2 years of post-secondary education. The
median annual income of participating families was $40,000 (M ⫽
$43,134, SD ⫽ $25,136). Characteristic of the geographic area,
more than 98% of parents and children identified their ethnicity as
White.
Procedure
Baseline assessments. Baseline assessments were conducted
in the family’s home. Parents and the target child independently
completed questionnaires in separate parts of the residence. Family
members were reminded that their responses would be kept con-
fidential and would not be communicated to other family members.
Questionnaire completion required approximately 70 min.
Videotaped interactions. During the in-home visit, parents
and the target child were also videotaped during structured dis-
cussion tasks. The discussion tasks began as soon as project staff
had prepared for the videotaping and occurred both before and
after questionnaire completion. For dual-parent families, the vid-
eotaping included two dyadic interactions (i.e., mother–child and
father–child) and one triadic interaction (i.e., mother–father–
child). Either the mother or the father was randomly selected for
the first dyadic interaction, during which time they discussed a
series of questions concerning typical aspects of family life (e.g.,
household chores) for 15 min. Families then engaged in the triadic
interaction for 12 min, during which time both parents and their
child discussed issues of disagreement (e.g., curfews). The third
discussion task followed the same format as the first, with the
exception that the child interacted with the parent who had not
participated in the first dyadic discussion task.
2
Measures
This investigation focuses on measures that pertain to the qual-
ity of the parent–child relationship (assessed by self-report) and to
parent–child dyadic interaction behaviors (assessed by third-party
observation). These measures are described below.
Perceptions of hostility and warmth. Mothers, fathers, and
children individually responded to questions that assessed their
perceptions of their own and the others’ typical hostility and
warmth. To assess parents’ perceptions of the hostility that they
typically show toward their children, and children’s perceptions of
the hostility that their parents typically show toward them, each
parent and child reported: how often the mother (father) loses her
(his) temper and yells at the child, and how often in the past month
the mother (father) got angry at the child, shouted or yelled at the
child, and insulted or swore at the child. To assess children’s
perceptions of the hostility that they typically show toward their
parents, and parents’ perceptions of the hostility that their children
typically show toward them, each parent and child reported how
often during the past month the child: got angry at the mother
(father), shouted or yelled at the mother (father), and insulted or
swore at the mother (father).
To assess parents’ perceptions of the warmth that they typically
show toward their children, and children’s perceptions of the
warmth that their parents typically show toward them, each parent
and child reported how often during the past month the mother
(father): let the child know that she (he) cares about her or him,
acted loving and affectionate toward the child, and let the child
know that she (he) appreciates her or him. To assess children’s
perceptions of the warmth they typically show toward their par-
ents, and parents’ perceptions of the warmth their children typi-
cally show toward them, each parent and child reported how often
during the past month the child: let the mother (father) know that
1
The larger study included a control condition plus two inter-
ventions designed to prevent adolescent problem behaviors: the
Life Skills Training program (LST; Spoth, Redmond, Trudeau, &
Shin, 2002), and the Strengthening Families Program: For Parents
and Youth 10–14 (Molgaard, Kumpfer, & Fleming, 1997) in
combination with the LST. Because the current study includes only
data collected prior to implementation of the interventions, the
interventions could not have affected the results of this investiga-
tion and, therefore, are not further discussed.
2
Single-parent families engaged in two parent–child interac-
tions. The first focused on aspects of family life. The second
focused on issues of disagreements.
460 MADON, GUYLL, AND SPOTH

she or he cares about her (him), acted loving and affectionate
toward the mother (father), and let the mother (father) know that
she or he appreciates her (him).
All items were assessed on a 7-point response scale ranging
from 1 (always)to7(never) except for the items assessing per-
ceptions of parents’ tendency to lose their temper and yell at the
child, which was assessed on a 5-point response scale ranging from
1(almost always)to5(almost never). Responses to these items
were rescaled into a 7-point response scale in order to combine
them with the other hostility items. All items were reverse scored
so that higher values reflected greater perceptions of hostility and
warmth. Tables 1 and 2 present means, variances, and internal
consistencies for each scale.
Interaction behaviors. Interaction behaviors exhibited during
the mother–child and father–child dyadic discussion tasks were
coded for hostility and warmth. Trained coders used the Iowa
Family Interaction Rating Scales (IFIRS; Melby & Conger, 2001),
which has been shown to be both reliable and valid (e.g., Melby,
Conger, & Puspitawati, 1999). The trained coders provided global
assessments of the hostility and warmth displayed by mother and
child and father and child on a 9-point rating scale ranging from 1
(not at all characteristic)to9(mainly characteristic). Each vid-
eotape required approximately 3 hr to code. To assess interrater
reliability, approximately 25% of the videotapes were randomly
selected to be coded by a second coder. As detailed in Tables 1 and
2, intraclass correlation coefficients (ICCs) indicated good inter-
rater agreement for the observer ratings of the parent–child dyads
(.61 ⱕ all ICCs ⱕ .80).
Results
Descriptive Statistics
Table 1 presents the correlations and covariances among
the hostility-related variables included in the analysis, as
well as the mean and variance for each variable. Table 2
presents parallel information for the warmth-related data.
Overview of Analyses
An observed variable path-analytic approach was taken
utilizing LISREL 8.30 (Jo¨reskog & So¨rbom, 1999). Figure
1 presents the path model that was used to perform four
analyses of data acquired from the self-report questionnaires
and the observer ratings of parent–child dyadic interactions.
The four analyses correspond to the mother–child hostility
and warmth data and to the father–child hostility and
warmth data. Because the analysis estimates each path co-
efficient’s asymptotic standard error, the path coefficient is
divided by its standard error to yield a z-test for statistical
significance (Bollen, 1989). The following sections explain
how the model’s effects pertain to the issues addressed by
this investigation.
3
Predictive accuracy versus self-fulfilling influence.
There are two non–mutually exclusive reasons why a per-
ceiver’s perception might predict a target’s behavior, one
relating to predictive accuracy and one relating to self-
fulfilling influence. Because a perception can be both par-
3
As shown on the left-hand side of Figure 1, the model allowed
for intercorrelations among parents’ and children’s self-reports of
both their own and the other’s typical behavior. However, these
relationships are of only secondary interest vis a´ vis the issues
examined in the current investigation. Therefore, we devote text
neither to their meaning (which in any case is straightforward) nor
to their corresponding results (which are provided in Figure 1).
Table 1
Descriptive Data for Self- and Observer Reports Relating to Interpersonal Hostility: Covariances, Correlations, Means,
Variances, and Reliability
Variable 123456
Mother–child dyads (N ⫽ 658)
1. C typical hostility to M (C) (0.86) 0.52 0.66 0.39 0.31 0.31
2. C typical hostility to M (M) 0.45 (0.88) 0.45 0.70 0.26 0.32
3. M typical hostility to C (C) 0.62 0.43 (1.04) 0.46 0.28 0.29
4. M typical hostility to C (M) 0.32 0.59 0.42 (0.80) 0.16 0.31
5. C actual hostility to M (O) 0.52 0.45 0.53 0.26 (3.30) 0.46
6. M actual hostility to C (O) 0.53 0.55 0.55 0.52 1.56 (3.44)
M 2.11 2.30 2.42 2.80 2.22 2.68
␣
a
.69 .76 .80 .83 .73 .61
Father–child dyads (N ⫽ 576)
1. C typical hostility to F (C) (0.63) 0.24 0.54 0.18 0.32 0.32
2. C typical hostility to F (F) 0.39 (0.62) 0.25 0.37 0.23 0.37
3. F typical hostility to C (C) 0.67 0.31 (1.04) 0.36 0.26 0.44
4. F typical hostility to C (F) 0.28 0.58 0.43 (0.67) 0.08 0.39
5. C actual hostility to F (O) 0.25 0.18 0.16 0.06 (2.66) 1.29
6. F actual hostility to C (O) 0.23 0.27 0.25 0.27 0.45 (3.10)
M 1.81 2.03 2.22 2.58 1.91 2.27
␣
a
.61 .76 .80 .83 .80 .63
Note. C ⫽ child; M ⫽ mother; O ⫽ observer; F ⫽ father. Variances appear in parentheses on the diagonal, and correlations and
covariances appear above and below the diagonal, respectively. (C), (M), and (F) denote variables based on child, mother, and father
self-report, respectively. O denotes variables based on observer ratings of actual behaviors exhibited during parent–child dyadic
interactions.
a
Values represent internal consistency scores as assessed by Cronbach’s
␣
for questionnaire scales. Values for observer ratings (O)
represent interobserver reliabilities as assessed by intraclass correlation coefficients.
461SELF-FULFILLING PROPHECIES

tially accurate and partially inaccurate, the magnitude of the
total bivariate relationship between the perceiver’s percep-
tion and a target’s behavior can include a combination of
both predictive accuracy and self-fulfilling influence
(Jussim, 1991). Therefore, when testing for self-fulfilling
prophecy effects, one must attempt to control for predictive
accuracy.
Accuracy is the extent to which a perceiver’s perception
accurately and directly predicts a target’s behavior without
causing the behavior. The model shown in Figure 1 controls
for accuracy by including the direct effects of child percep-
tions of child behavior on child behavior (Path a), parent
perceptions of child behavior on child behavior (Path b),
child perceptions of parent behavior on parent behavior
(Path e), and parent perceptions of parent behavior on parent
behavior (Path f). To the degree that either parent or child
perceptions can directly predict either parent or child be-
havior, that portion of the variance in actual behavior will be
attributed to accuracy and will, therefore, be unavailable for
providing support for the existence of a self-fulfilling proph-
ecy (see Jussim et al., 1996).
Self-fulfilling influence, by contrast, is the extent to
which a perceiver’s perception indirectly predicts a target’s
behavior because the perception first influences perceiver
behavior which, in turn, influences target behavior. That is,
the effect of the perceiver’s perception on the target’s be-
havior cannot be direct, but must be indirect and mediated
solely by the perceiver’s behavior. For example, in the
current model, a child-based self-fulfilling prophecy effect
on a parent is associated with the indirect effect of com-
pound Path cg, and a parent-based self-fulfilling prophecy
effect is associated with the indirect effect of compound
Path dh. As noted above, only the portion of a target’s
behavior that cannot be accounted for by accuracy (i.e., the
direct effects of both interactants’ perceptions of the target’s
behavior) is available for providing support for a self-
fulfilling prophecy. In addition, only the portion of the
perceiver’s behavior that cannot be attributed to accuracy
(i.e., the direct effects of both interactants’ perceptions of
the perceiver’s behavior) is available for mediating the
indirect effect of the perceiver’s perception on the target’s
behavior and thereby providing support for a self-fulfilling
prophecy.
Reciprocal effects. Paths g and h reflect the reciprocal
effects anticipated in the interaction, in which one person’s
behavior in the interaction affects the other person’s behav-
ior. The self-fulfilling prophecy effects are transmitted by
compound paths that include these reciprocal effects.
4
Hostile Behavior
Model fit. The model presented in Figure 1 provided a
good fit to the hostility-relevant data for both mother–child,
4
Calculating reciprocal effects between parent and child behav-
iors requires summing across an infinite series of terms to calculate
each self-fulfilling prophecy effect. For example, the child-based
effect is [(c ⫻ g)] ⫹ [(c ⫻ g) ⫻ (h ⫻ g)] ⫹ [(c ⫻ g) ⫻ (h ⫻ g)
2
]
⫹ [(c ⫻ g) ⫻ (h ⫻ g)
3
] ⫹ (etc.).
Table 2
Descriptive Data for Self- and Observer Reports Relating to Interpersonal Warmth: Covariances, Correlations, Means,
Variances, and Reliability
Variable 123456
Mother–child dyads (N ⫽ 658)
1. C typical warmth to M (C) (1.79) 0.47 0.70 0.39 0.24 0.17
2. C typical warmth to M (M) 0.80 (1.62) 0.39 0.72 0.27 0.20
3. M typical warmth to C (C) 1.21 0.65 (1.66) 0.40 0.17 0.19
4. M typical warmth to C (M) 0.51 0.90 0.50 (0.96) 0.20 0.25
5. C actual warmth to M (O) 0.46 0.49 0.32 0.28 (2.07) 0.18
6. M actual warmth to C (O) 0.42 0.45 0.45 0.43 0.47 (3.28)
M 5.31 4.91 5.51 5.60 2.15 6.10
␣
a
.87 .89 .87 .89 .67 .78
Father–child dyads (N ⫽ 576)
1. C typical warmth to F (C) (2.26) 0.81 1.78 0.52 0.37 0.26
2. C typical warmth to F (F) 0.42 (1.67) 0.72 1.04 0.31 0.49
3. F typical warmth to C (C) 0.86 0.40 (1.91) 0.50 0.36 0.28
4. F typical warmth to C (F) 0.32 0.74 0.33 (1.20) 0.22 0.54
5. C actual warmth to F (O) 0.17 0.17 0.18 0.14 (2.07) 0.58
6. F actual warmth to C (O) 0.09 0.20 0.11 0.25 0.21 (3.75)
M 5.20 4.48 5.35 4.92 2.06 5.65
␣
a
.91 .88 .90 .90 .61 .78
Note. C ⫽ child; M ⫽ mother; O ⫽ observer; F ⫽ father. Variances appear in parentheses on the diagonal, and correlations and
covariances appear above and below the diagonal, respectively. (C), (M), and (F) denote variables based on child, mother, and father
self-report, respectively. O denotes variables based on observer ratings of actual behaviors exhibited during parent–child dyadic
interactions.
a
Values represent internal consistency scores as assessed by Cronbach’s
␣
for questionnaire scales. Values for observer ratings (O)
represent interobserver reliabilities as assessed by intraclass correlation coefficients.
462 MADON, GUYLL, AND SPOTH

Figure 1. Path model and results relating parent and child self- and other-reports to observer ratings of interpersonal hostility and warmth
actually exhibited during an interaction. Results pertaining to hostility for mother–child dyads (N ⫽ 658) and father–child dyads (N ⫽
576) are identified by MH and FH, respectively. Likewise, results pertaining to warmth for mother–child and father–child dyads are
identified by MW and FW, respectively. * p ⱕ .05. ** p ⱕ .01.
463SELF-FULFILLING PROPHECIES

␹
2
(1, 658) ⫽ .04, p ⫽ .84, and father–child,
␹
2
(1, 576) ⫽
2.02, p ⫽ .16, dyads.
Accuracy-based prediction of hostile behaviors. Figure
1 presents standardized effects and p levels for Paths a, b, e,
and f, which account for the effects of accuracy in predict-
ing hostile behavior for both mother–child (MH) and
father–child (FH) dyads. These results reveal that the child
and both parents were accurate in predicting child hostility
(Paths a and b). Results further indicated that whereas both
parents were accurate in predicting their own hostility (Path
f), child perceptions were not accurate predictors of parent
hostility (Path e).
Expectancy-based effects on hostile behaviors. Within
the context of the analytic model, the potential for one
person’s behavior to be affected by her or his inaccurate
perception about the other person is represented by Paths c
and d in Figure 1. As reported in the figure, the only
expectancy-based effect was that children’s perception of
their mothers’ typical hostility was positively associated
with children’s hostility during the interaction (Path c). No
such effect was observed for mothers, nor for either children
or fathers in the father–child dyads.
Reciprocal effects of parent and child hostile behaviors.
The possibility that parent and child could influence each
other’s behavior during the interaction is represented by
Paths g and h. As reported in Figure 1, findings reveal that
greater child hostility in the interaction predicted both
greater mother hostility and greater father hostility (Path g).
Hostility exhibited by the parent did not predict child hos-
tility (Path h).
Self-fulfilling effects of parent and child perceptions of
hostility. The findings reported above indicate that chil-
dren’s perception of their mothers’ typical hostility was
positively associated with children’s hostility during the
interaction, which, in turn, was positively associated with
mothers’ hostility during the interaction. These findings
appear to suggest that children’s perceptions influenced
mothers’ hostile behaviors via a self-fulfilling prophecy.
However, the fact that both of the constituent direct effects
(i.e. Paths c and g) are significant does not necessarily mean
that the corresponding indirect effect is also significant
(Bollen, 1987). Moreover, in order to test the self-fulfilling
prophecy hypothesis, it is necessary to test the specific
indirect effect of children’s perception on their mothers’
hostile behavior that is specifically conveyed through the
children’s behavior (i.e., solely via Path c, and not via Path
e).
5
To calculate the magnitude and significance of this
specific indirect effect, we used Bollen’s (1987) procedure,
which makes use of the delta method (Sobel, 1986) to
generate the necessary standard errors. The interested reader
may obtain excellent treatments of these topics from the
citations provided.
Results from these analyses indicated that children’s per-
ceptions of their mothers’ typical hostility affected how
hostile they themselves behaved during the interaction,
which, in turn, affected how hostile their mothers actually
behaved during the interaction,
␤
⫽ .11, z ⫽ 2.58, p ⫽ .01.
This pattern is consistent with a self-fulfilling prophecy and
occurred in the context of a model that attempted to control
for accuracy. That is, the model controlled for that portion
of mothers’ hostility that could be predicted by the ability of
both children’s and mothers’ perceptions to directly predict
mothers’ hostility (Paths e and f).
We also tested for a self-fulfilling prophecy originating in
mothers’ perceptions of children’s hostility, but analyses
yielded no support for such an effect (z ⬍ 1). Results for
father–child dyads did not support the idea that fathers or
children conveyed self-fulfilling prophecy effects on the
others’ behavior, (zs ⬍ 1).
Warmth Behaviors
Model fit. The model presented in Figure 1 provided a
good fit to the warmth-relevant data for both mother–child
dyads,
␹
2
(1, N ⫽ 658) ⫽ .08, p ⫽ .77, and father–child
dyads,
␹
2
(1, N ⫽ 576) ⫽ 0.25, p ⫽ .61.
Accuracy-based prediction of warmth behaviors. For
the mother–child dyads, findings indicated that children’s
and mothers’ perceptions of children’s warmth exhibited a
measure of accuracy (Paths a and b), as did both children’s
and mothers’ perceptions of mothers’ warmth (Paths e and
f). For the father–child dyads, only fathers’ perception of
fathers’ warmth evidenced any accuracy, insomuch as it
predicted fathers’ actual warmth during the interaction
(Path f).
Expectancy-based effects on warmth behaviors. Results
provided no support for the idea that either mothers’, fa-
thers’, or children’s perceptions of the other person’s
warmth affected their own warmth during the interaction
(Paths c and d).
Reciprocal effects of parent and child warmth interaction
behaviors. Analyses yielded no evidence for bidirectional
effects relating to warmth exhibited during the interaction
for either mother–child or father–child dyads (Paths g
and h).
Self-fulfilling effects of mothers’ and children’s percep-
tions of warmth. As detailed above, testing the self-
fulfilling prophecy hypothesis requires calculating and test-
ing a specific indirect effect. However, because none of the
constituent direct effects (i.e., Paths c, d, g, or h) associated
with either child- or parent-based self-fulfilling prophecies
for warmth approached significance, the delta method is not
appropriate for calculating standard errors for the corre-
sponding specific indirect effects (Bollen, 1987). Hence, the
statistic for determining the significance of the specific
indirect effects cannot be computed in rigorous fashion.
However, given that the constituent direct effects are non-
significant and that the magnitudes of the specific indirect
effects pertaining to the child- and parent-based self-
fulfilling prophecy were negligible in all cases (all 兩
␤
兩s ⬍
.02), it is reasonable to conclude that if it had been possible
5
A specific effect is “that part of the indirect or total effects
transmitted by a path or combination of individual paths” (Bollen,
1987, p. 55). Thus, a specific indirect effect is that part of an
indirect effect that is transmitted specifically by a particular path or
combination of paths.
464 MADON, GUYLL, AND SPOTH

to perform the calculation, the self-fulfilling prophecy ef-
fects would not have been statistically significant.
Modeling Mother–Child Effects as Unidirectional:
The Potential for Invalid Conclusions
The results reported above were based on analyses that
modeled the effects between parents and children as bidi-
rectional (i.e., simultaneous inclusion of both Path g and
Path h in Figure 1). Simultaneous inclusion of these paths
was essential to capture the bidirectional nature of parent–
child effects, as advised by the bilateral model of parent–
child relations (Kuczynski, 2003). To evaluate the impor-
tance of conforming to the bilateral framework in the
context of the current investigation, we next report findings
from an analysis that only accounts for unidirectional ef-
fects between parents and children and, thereby, ignores the
bidirectional nature of interpersonal effects within the dy-
adic interactions.
Because the bidirectional model yielded the most inter-
esting findings with respect to interpersonal hostility for the
mother–child dyads, and because we consider the unidirec-
tional model to be inferior, we limit our reporting of results
from the unidirectional model to those generated by a re-
analysis of the hostility-relevant data from mother–child
dyads. Specifically, we repeated the analysis depicted in
Figure 1 two times. The first re-analysis modeled interaction
effects as being unidirectional from children to mothers and
provided a significantly worse fit to the data compared with
the bidirectional model, ⌬
␹
2
(1, N ⫽ 658) ⫽ 6.04, p ⫽ .01.
Otherwise, this unidirectional model tended to yield a pat-
tern of findings consistent with the results yielded by the
bidirectional model. The one exception was that the unidi-
rectional model indicated that children’s perceptions of their
mothers’ typical hostility directly predicted their mothers’
actual hostility during the interaction (Path e), whereas the
bidirectional model indicated that they did not.
The second re-analysis modeled the interaction effects as
being unidirectional from mothers to children and also
provided a comparatively poor fit to the data, ⌬
␹
2
(1, N ⫽
658) ⫽ 17.49, p ⬍ .01. It also yielded a number of results
that diverged in critical respects from those produced by the
bidirectional model. First, the unidirectional model indi-
cated that children’s perceptions of their mothers’ typical
hostility accurately predicted mothers’ actual hostility (Path
e), whereas the bidirectional model indicated that they did
not. Second, the unidirectional model indicated that moth-
ers’ perceptions of their children’s typical hostility did not
accurately predict children’s actual hostility (Path b),
whereas the bidirectional model indicated that they did.
Third, the unidirectional model indicated that children’s
perceptions of their mothers’ typical hostility did not predict
children’s actual hostility (Path c), whereas the bidirectional
model indicated that they did. Fourth, the unidirectional
model indicated that mothers’ perceptions of their chil-
dren’s typical hostility predicted mothers’ actual hostility
(Path d), whereas the bidirectional model indicated that they
did not. Fifth, the unidirectional model indicated that moth-
ers’ hostility during the interaction predicted children’s
hostility during the interaction (Path h), whereas the bidi-
rectional model indicated that they did not. And, most
interestingly, the unidirectional model supported the exis-
tence of a self-fulfilling prophecy originating in mothers—
that is, mothers’ perceptions of their children’s typical hos-
tility was mediated by its proximal effect on mothers’ actual
hostility and ultimately predicted their children’s actual
hostility during the interaction (via compound Path dh), a
pattern not supported by the bidirectional model. The dif-
ferent results that emerged when the reciprocal nature of the
interaction effects was eliminated from the model empha-
size the importance of accounting for the bidirectional na-
ture of interpersonal effects in the context of parent–child
interactions.
Reversal of Direction of Effects
As noted in the Method section, videotaping of the dyadic
interactions commenced as soon as the project staff had
completed the necessary preparations. As a result, some
family members participated in the interaction before com-
pleting the perception items. This raises the concern that
rather than perceptions influencing subsequent interaction
behaviors, the interaction behaviors might have influenced
subsequent questionnaire responses that were used to assess
family members’ perceptions. Because data regarding the
order of these tasks were not recorded, we addressed this
concern by re-analyzing the data using a model in which the
direction of the effects represented by Paths a through f was
reversed. Thus, this re-analysis assessed how well the data
are fit by a model in which the interaction behaviors are
modeled as causes of the questionnaire responses. In com-
parison with the original model (i.e., with the direction of
effects as shown in Figure 1), which provided a good fit for
all four analyses, all
␹
2
s ⱖ 0.04, all ps ⬎ .16, the new model
provided a poor fit to the data,
␹
2
s ranged from 6.96 to 32.7,
ps ⬍ .01 (for both models, df ⫽ 1, Ns ⫽ 658 mother–child
dyads, 576 father–child dyads), suggesting that children’s
and parents’ perceptions predicted and influenced interac-
tion behaviors rather than the reverse.
Additional Analyses
We also evaluated whether the results might have been
affected by child gender and whether the dyadic interaction
was the first or third videotaped task in which the parent and
child participated. We conducted a series of two-group
stacked LISREL analyses to test for differences in path
coefficients related to child gender. Across the four analy-
ses, the only gender difference to emerge was for the
father–child hostility data, in which the association between
the children’s perception of their own typical hostility (Path
a) was more predictive of children’s actual hostility for girls
than for boys. A second series of four stacked two-group
LISREL analyses tested whether the order of the interaction
affected the results. Analyses revealed only a single differ-
ence. Mothers’ perceptions of their own typical hostility
were more strongly linked to their actual hostility (Path f)
when the interaction was the third task in which they en-
465SELF-FULFILLING PROPHECIES

gaged. However, the child-based self-fulfilling prophecy
effects were not affected by task order: self-fulfilling proph-
ecy effects were nearly identical in magnitude when the
mother–child hostility data were analyzed separately by
task order.
Discussion
This research examined whether parents’ and children’s
perceptions of the other’s hostility and warmth led to self-
fulfilling prophecy effects during dyadic interactions. Re-
sults supported the existence of self-fulfilling prophecies
among the mother–child dyads with respect to the hostility-
relevant data. Children’s perceptions of their mothers’ typ-
ical hostility predicted children’s hostility during the inter-
action, which, in turn, predicted mothers’ hostility during
the interaction. Results did not, however, support reciprocal
self-fulfilling prophecy effects originating in mothers.
Mothers’ perceptions of their children’s hostility did not
predict children’s hostility during the interaction by virtue
of their relation to mothers’ interaction hostility. Results
also did not support the existence of self-fulfilling prophe-
cies among the mother–child dyads with respect to warmth-
relevant data, nor among the father–child dyads for either
the hostility- or warmth-relevant data. These findings sug-
gest that the self-fulfilling prophecy effects observed in this
study were child-based rather than parent-based, were re-
stricted to mother-child interactions, and occurred for per-
ceptions of negative affective behaviors but not positive
affective behaviors.
Mother–Child Differences in Self-Fulfilling Prophecy
Effects
Analyses of the hostility-relevant data pertaining to the
mother–child dyads indicated that self-fulfilling prophecy
effects operated through children but not through mothers.
There are at least two explanations for these differences, one
relating to predictive accuracy and a second relating to the
potential for different effects of agency in parents and
children. Focusing first on predictive accuracy, the more
accurate one’s belief for another is, the less opportunity
there is for the belief to create a self-fulfilling prophecy
(Jussim, 1991). In this study, the relation between mothers’
perceptions of their children’s typical hostility and chil-
dren’s actual hostility (Path b) was larger than the corre-
sponding relation between children’s perceptions of their
mothers’ typical hostility and mothers’ actual hostility (Path
e). Consistent with this observation, a supplemental analysis
in which the effects pertaining to mothers’ and children’s
accuracy for the others’ hostility were constrained to be
equal (i.e., Path b ⫽ Path e) produced a significant reduction
in model fit, ⌬
␹
2
(1, N ⫽ 658) ⫽ 6.73, p ⫽ .01, thereby
indicating significantly greater accuracy on the part of
mothers. Accordingly, one reason that children’s percep-
tions might have produced stronger self-fulfilling prophecy
effects than mothers’ perceptions is because children’s per-
ceptions were less accurate.
Self-fulfilling prophecy effects may also have been me-
diated through children’s behavior but not through mothers’
behavior because of differences between mothers’ and chil-
dren’s agency in pursuing their own goals. Children who
perceived their mothers to be hostile may have preemptively
expressed hostile behaviors within the interaction as a
means of demonstrating autonomy. By contrast, mothers
may have intentionally behaved in ways that they believed
would be most effective for attaining more long-term goals,
such as maintaining positive relationships with their chil-
dren (Kuczynski & Lollis, 2001).
It is also important to note that these explanations are not
mutually exclusive. Although mothers’ perceptions of chil-
dren’s typical hostility were more accurate than were chil-
dren’s perceptions of mothers’ typical hostility, mothers’
perceptions did not perfectly predict their children’s actual
hostility, indicating that they were partially inaccurate. Yet,
the inaccurate portion of their perceptions did not produce a
self-fulfilling prophecy effect on their children’s behavior.
Thus, the differing patterns of mothers’ and children’s self-
fulfilling prophecy effects most likely reflect a combination
of predictive accuracy operating in conjunction with differ-
ences in mother and child agency.
Negative Versus Positive Self-Fulfilling Prophecy
Effects
The self-fulfilling prophecy effects found in this research
were limited to hostile behaviors. This suggests that when it
comes to expressions of positive and negative affect, neg-
ative self-fulfilling prophecies may be more powerful than
positive ones. The tendency for perceptions of negative
affective behaviors to create more powerful self-fulfilling
prophecies than perceptions of positive affective behaviors
has important implications for family interventions seeking
to improve parent–child relationships. In particular, it sug-
gests that interventions may benefit from emphasizing the
value of developing positive perceptions of family members
and providing skills training to help family members better
communicate their positive perceptions to each other. Inter-
ventions may also benefit by raising children’s awareness of
how preexisting beliefs might affect their behavior and
encourage them to respond to parents’ actual interaction
behaviors. Such efforts should reduce negative interactions,
increase positive interactions, and promote parent–child
bonding, which facilitates positive adolescent outcomes
(Catalano & Hawkins, 1996).
Mother–Father Differences in Self-Fulfilling
Prophecy Effects
A child-based self-fulfilling prophecy emerged between
children and mothers but not between children and fathers.
Because the self-fulfilling prophecy could only occur if an
emotional expression was elicited from the target, the dif-
ference between the parent–child dyads might have oc-
curred if the mother–child dyads were characterized by
greater emotional expressiveness. Consistent with this idea,
paired-samples t tests did show that mothers exhibited more
hostility and warmth than fathers, ts(553) ⱖ 3.89, ps ⱕ .01,
466 MADON, GUYLL, AND SPOTH

and that children exhibited more hostility with their moth-
ers, t(553) ⫽ 3.23, p ⱕ .01. Previous research has docu-
mented similar differences in dyadic emotional expressive-
ness and suggests that they arise because mothers and
children spend more time with each other and share a
broader range of experiences (e.g., Buhrmester, Camparo,
Chistensen, & Gonzalez, 1992).
Limitations
Interpretation of naturalistic data. When investigating
self-fulfilling prophecies with naturalistic data, analyses
must control for that portion of perceivers’ perceptions that
accurately predict, without influencing, targets’ behavioral
outcomes. When a valid predictor is omitted from an ana-
lytic model, accuracy and self-fulfilling prophecy effects are
under- and overestimated, respectively. Although all natu-
ralistic studies are susceptible to the omitted variable prob-
lem, the current data were only interpreted as reflecting a
self-fulfilling prophecy to the extent that perceivers’ per-
ceptions had an indirect effect on targets’ behavioral out-
comes that was specifically mediated by the perceivers’ own
behaviors during the interaction. Thus, in the current anal-
ysis, if valid predictors of targets’ behavioral outcomes had
been omitted from the model, to the degree that they cor-
related with perceiver perceptions, that accuracy-related
portion of the relation between the perceivers’ perceptions
and target outcomes would have been captured entirely by
the direct effects between these variables (i.e., Paths b and
e; Jussim et al., 1996).
However, it is possible that the analytic model omitted a
variable that both correlated with perceivers’ perceptions
and had a causal effect on perceivers’ behavior during the
interaction. In this case, the effect that perceivers’ percep-
tions of the target had on their own behavior would have
been overestimated, raising the possibility that their self-
fulfilling influences could have been smaller than suggested.
Although we cannot rule out this possibility, our self-
fulfilling prophecy interpretation is consistent with a long
history of experimental findings demonstrating that perceiv-
ers’ perceptions influence targets’ behaviors by means of
self-fulfilling prophecies (see Snyder & Stukas, 1999, for a
review). The convergence of our findings with those of
previous investigations increases confidence that our results
reflect self-fulfilling prophecies between mothers and
children.
Timing of videotaping. Parent–child interactions were
videotaped before and after family members completed the
questionnaire items. Consequently, some participants re-
ported their perceptions after they had already engaged in
the dyadic interaction tasks. This procedural aspect of the
study raises the concern that the interaction behaviors might
have influenced family members’ perceptions rather than
the reverse. Unfortunately, information regarding the pre-
cise sequence of the completion of the perception items
relative to the interaction task was not recorded. Although
this fact precluded us from specifically testing whether the
timing of the videotaping influenced participants’ responses
to the perception items, there are several reasons why we
believe that the direction of effects was from perceptions to
interaction behaviors. First, nearly all of the items assessing
family members’ perceptions of their own and the others’
typical hostility and warmth inquired about behaviors that
had occurred during the past month. Thus, the wording of
these items instructed parents and children to report their
perceptions on the basis of behaviors that had preceded the
interaction. Second, had the direction of effect been re-
versed (with interaction behaviors influencing perceptions),
one would have expected to see very large direct effects
from perceptions to behaviors. That is, the self-reported
perceptions would seem to have been extremely accurate
because they would have been based on the actual interac-
tions that had just occurred. In the current data, a self-
fulfilling prophecy was indicated for hostility in mother–
child interactions and originated in the child’s perception of
the mother. However, the direct effect of children’s percep-
tions on mother’s behavior was nearly zero for this analysis,
thereby providing no support for the idea that mother inter-
action hostility influenced child perceptions. In addition,
even if interaction behaviors had influenced self-reported
perceptions, these effects would have been attributed to
accuracy (i.e., direct effects) and not to a self-fulfilling
prophecy (i.e., specific indirect effects). Third, as reported
in the Results section, we re-analyzed the data modeling the
direction of causal influence as being from interaction be-
haviors to perceptions. In comparison with the original
model, the new model was found to provide a poor fitto
both the hostility- and warmth-relevant data for both
mother–child and father–child dyads. Thus, the wording of
the perception items, the lack of accuracy in child reports of
mother’s hostility, and the comparison of empirical results
generated by the two alternative models all argue against the
idea that interaction behaviors influenced family members’
perceptions. On the basis of this evidence, it seems most
plausible that the direction of effect was from self-reported
perceptions to behaviors.
Perception items. The perception items used in this
research reflected family members’ perceptions of their own
and the others’ typical affective behaviors. Family mem-
bers’ actual affective behaviors, in contrast, were assessed
during a specific and circumscribed situation. Thus, the
wording of the perception items was not perfectly matched
to the affective behaviors that were assessed. It is possible,
therefore, that family members’ perceptions may have more
accurately predicted parent–child affective behaviors in the
context of mundane parent–child interactions than they
predicted parent–child affective behaviors in the contrived
context of the dyadic interaction tasks. Although this pos-
sibility does not undermine our conclusion that children’s
perceptions of their mothers’ typical hostility predicted their
mothers’ actual hostility during the interaction task via the
process of a self-fulfilling prophecy, it does highlight the
importance of exercising caution when generalizing our
findings to other less contrived situations in which family
members’ perceptions of one another’s behaviors may be
more accurate than they were in this research.
A second issue related to the perception items and the
generalizability of our findings is whether similar patterns
467SELF-FULFILLING PROPHECIES

would have emerged had the perception items directly as-
sessed family members’ perceptions about affective behav-
iors pertaining specifically to the context of the interaction
task. Although people’s perceptions tend to be more accu-
rate for circumscribed behaviors than for global behaviors
(Swann, 1984), the dyadic interaction tasks in which parents
and children engaged were unfamiliar to most family mem-
bers. Therefore, any perceptions that parents and children
did have regarding the affective behaviors each would ex-
hibit in this context would probably have been less accurate
than are most perceptions regarding circumscribed behav-
iors. Thus, had we assessed perceptions that were more
closely tied to the behaviors occurring during the interaction
task, we may have found stronger self-fulfilling prophecy
effects than reported. Of course, our discussion of this issue
is highly speculative and requires empirical validation be-
fore firm conclusions can be drawn.
Conclusion
This research tested for reciprocal self-fulfilling prophecy
effects among parents and their adolescent children. Al-
though such effects were not found among father–child
dyads, data were consistent with a self-fulfilling prophecy
effect among the mother–child dyads with respect to
hostility-relevant data. Children’s perceptions of their moth-
ers’ typical hostility predicted how hostile they behaved
toward their mothers during an interaction, which, in turn,
predicted how hostile their mothers behaved toward them.
Although mothers’ perceptions did predict children’s hos-
tility, this relationship was due to mothers’ accuracy and not
to a self-fulfilling prophecy. Thus, the self-fulfilling proph-
ecy effects observed in this study were unidirectional from
child to mother and not vice versa. However, additional
analyses based on an inferior model that considered only
unidirectional effects suggested that mothers had a self-
fulfilling prophecy effect on their children’s behavior.
These differing conclusions highlight the importance of
using a bilateral framework to investigate interpersonal
relations.
The tendency for children to exert stronger self-fulfilling
prophecy effects than mothers contrasts with past work
showing that the person with greater power in a relationship
tends to have greater self-fulfilling influences (Snyder &
Stukas, 1999). The current finding likely reflects the unique
character of the parent–child relationship and suggests that
the present pattern of results may not be typical of relation-
ships in general. Indeed, patterns of self-fulfilling influence
between parents and their children have previously been
shown to differ from patterns typically observed in the
context of nonfamilial relationships (Madon, Guyll, Spoth,
Cross, & Hilbert, 2003). Further investigation of how these
patterns of interpersonal influence vary across time, in re-
sponse to interventions designed to improve parent–child
relationships, and how they vary with the characteristics of
the individual family members are necessary to enhance
understanding of relationship dynamics within the family.
References
Bell, R. Q. (1971). Stimulus control of parent or caretaker behavior
by offspring. Developmental Psychology, 4, 63–72.
Bollen, K. A. (1987). Total, direct, and indirect effects in structural
equation models. In Clifford C. Clogg (Ed.), Sociological meth-
odology (Vol. 17, pp. 37–69). Washington, DC: American So-
ciological Association.
Bollen, K. A. (1989). Statistical equations with latent variables.
New York: Wiley.
Buhrmester, D., Camparo, L., Christensen, A., & Gonzalez, L. S.
(1992). Mothers and fathers interacting in dyads and triads with
normal and hyperactive sons. Developmental Psychology, 28,
500–509.
Catalano, R. F., & Hawkins, J. D. (1996). The social development
model: A theory of antisocial behavior. In J. D. Hawkins (Ed.),
Delinquency and crime: Current theories (pp. 149–197). New
York: Cambridge University Press.
Hinde, R. A. (1979). Toward understanding relationships. Lon-
don: Academic Press.
Hinde, R. A., & Stevenson-Hinde, J. (1987). Interpersonal rela-
tionships and child development. Developmental Review, 7,
1–21.
Jo¨reskog, K. G., & So¨rbom, D. (1999). LISREL 8.30: User’s
reference guide. Chicago: Scientific Software.
Jussim, L. (1991). Social perception and social reality: A
reflection–construction model. Psychological Review, 98, 54–
73.
Jussim, L., Eccles, J., & Madon, S. (1996). Social perception,
social stereotypes, and teacher expectations: Accuracy and the
quest for the powerful self-fulfilling prophecy. In M. P. Zanna
(Ed.), Advances in experimental social psychology (Vol. 28, pp.
281–388). San Diego, CA: Academic Press.
Kuczynski, L. (2003). Beyond bidirectionality: Bilateral concep-
tual frameworks for understanding dynamics in parent–child
relations. In L. Kuczynski (Ed.), Handbook of dynamics in
parent–child relations (pp. 3–24). Thousand Oaks, CA: Sage.
Kuczynski, L., & Lollis, S. (2001). Four foundations for a dynamic
model of parenting. In J. R. M. Gerris (Ed.), Dynamics of
parenting (pp. 445–462). Appledoorn, the Netherlands: Garant.
Lollis, S. (2003). Conceptualizing the influence of the past and the
future in present parent–child relationships. In L. Kuczynski
(Ed.), Handbook of dynamics in parent–child relations, (pp.
67–87). Thousand Oaks, CA: Sage.
Madon, S., Guyll, M., Spoth, R. L., Cross, S. E., & Hilbert, S.
(2003). The self-fulfilling influence of mother expectations on
children’s underage drinking. Journal of Personality and Social
Psychology, 84, 1188–1205.
Melby, J. N., & Conger, R. D. (2001). The Iowa Family Interaction
Rating Scales: Instrument summary. In P. Kerig & K. Lindahl
(Eds.), Family observational coding systems: Resources for sys-
tematic research (pp. 33–58). Mahwah, NJ: Erlbaum.
Melby, J. N., Conger, K. J., & Puspitawati, H. (1999). Insider,
participant observer, and outsider perspectives on adolescent
sibling relationships. In F. M. Berardo (Series Ed.) & C. L.
Shehan (Vol. Ed.), Contemporary perspectives on family re-
search: Vol. 1. Through the eyes of the child: Re-visioning
children as active agents in family life (pp. 329–351). Stamford,
CT: JAI Press.
Merton, R. K. (1948). The self-fulfilling prophecy. Antioch Re-
view, 8, 193–210.
Molgaard, V., Kumpfer, K., & Fleming, B. (1997). The Strength-
ening Families Program: For parents and youth 10–14. Ames,
IA: Iowa State University Extension.
468 MADON, GUYLL, AND SPOTH

Snyder, M., & Stukas, A. A., Jr. (1999). Interpersonal processes:
The interplay of cognitive, motivational, and behavioral activi-
ties in social interaction. Annual Review of Psychology, 50,
273–303.
Sobel, M. (1986). Some new results on indirect effects and their
standard errors in covariance structure models. In N. B. Tuma
(Ed.), Sociological methodology (pp. 159–186). Washington,
DC: American Sociological Association.
Spoth, R. L., Redmond, C., Trudeau, L., & Shin, C. (2002).
Longitudinal substance initiation outcomes for a universal pre-
ventive intervention combining family and school programs.
Psychology of Addictive Behaviors, 16, 129–134.
Swann, W. B., Jr. (1984). Quest for accuracy in person perception:
A matter of pragmatics. Psychological Review, 91, 457–477.
Received June 6, 2003
Revision received November 17, 2003
Accepted January 6, 2004
䡲
New Editors Appointed, 2006–2011
The Publications and Communications Board of the American Psychological Association an-
nounces the appointment of seven new editors for 6-year terms beginning in 2006. As of January
1, 2005, manuscripts should be directed as follows:
• Experimental and Clinical Psychopharmacology (www.apa.org/journals/pha.html), Nancy K.
Mello, PhD, McLean Hospital, Massachusetts General Hospital, Harvard Medical School, 115
Mill Street, Belmont, MA 02478-9106.
• Journal of Abnormal Psychology (www.apa.org/journals/abn.html), David Watson, PhD, De-
partment of Psychology, University of Iowa, Iowa City, IA 52242-1407.
• Journal of Comparative Psychology (www.apa.org/journals/com.html), Gordon M. Burghardt,
PhD, Department of Psychology or Department of Ecology & Evolutionary Biology, University
of Tennessee, Knoxville, TN 37996.
• Journal of Counseling Psychology (www.apa.org/journals/cou.html), Brent S. Mallinckrodt,
PhD, Department of Educational, School, and Counseling Psychology, 16 Hill Hall, University
of Missouri, Columbia, MO 65211.
• Journal of Experimental Psychology: Human Perception and Performance (www.apa.org/
journals/xhp.html), Glyn W. Humphreys, PhD, Behavioural Brain Sciences Centre, School of
Psychology, University of Birmingham, Edgbaston, Birmingham B15 2TT, United Kingdom.
• Joural of Personality and Social Psychology: Attitudes and Social Cognition section
(www.apa.org/journals/psp.html), Charles M. Judd, PhD, Department of Psychology, Univer-
sity of Colorado, Boulder, CO 80309-0345.
• Rehabilitation Psychology (www.apa.org/journals/rep.html), Timothy R. Elliott, PhD, Depart-
ment of Psychology, 415 Campbell Hall, 1300 University Boulevard, University of Alabama,
Birmingham, AL 35294-1170.
Electronic submission: As of January 1, 2005, authors are expected to submit manuscripts
electronically through the journal’s Manuscript Submission Portal (see the Web site listed above
with each journal title).
Manuscript submission patterns make the precise date of completion of the 2005 volumes uncertain.
Current editors, Warren K. Bickel, PhD, Timothy B. Baker, PhD, Meredith J. West, PhD, Jo-Ida C.
Hansen, PhD, David A. Rosenbaum, PhD, Patricia G. Devine, PhD, and Bruce Caplan, PhD,
respectively, will receive and consider manuscripts through December 31, 2004. Should 2005
volumes be completed before that date, manuscripts will be redirected to the new editors for
consideration in 2006 volumes.
469SELF-FULFILLING PROPHECIES