The Imaginary Companions Created by Children Who Have Lived in Foster Care

Abstract

This study was designed to provide some preliminary information about the imaginary companions created by children who have lived in foster care, including prevalence rates and qualitative descriptions of the imaginary companions. We were also interested in how descriptions of the imaginary companions created by children who had lived in foster care compare to those of other children. Children with a history of foster care (n =21) and children from a low socioeconomic status community sample (n =39) were interviewed about imaginary companions. Twenty-six children (43.3%) reported having imaginary companions. Although having an imaginary companion has sometimes been believed to be more common in children with a history of maltreatment, in this preliminary study, prevalence rates did not differ between the two groups of children. In addition, both groups of children described companions that were a positive source of entertainment, friendship, and social support.

In their pretend play, many young children create imaginary companions, invented characters that children interact with or talk about on a regular basis. Some imaginary companions are invisible, while others are based on special stuffed animals or dolls referred to as personified objects (Gleason, Sebanc, & Hartup, 2000; Taylor & Carlson, 1997). The prevalence of imaginary companions is typically high during the preschool and early school-age period (Davis, Meins, & Fernyhough, 2013; Pearson et al., 2001; Taylor, Carlson, Maring, Gerow, & Charley, 2004). For example, in one of the largest systematic studies of children’s imaginary companions, Pearson et al. (2001) found that approximately 39% of children aged 5 to 7 years reported having an invisible imaginary companion. When personified objects are included in the general definition, the prevalence of imaginary companions is even higher, ranging from 46% to 65% of preschool and early school-age children (Davis et al., 2013; Taylor et al., 2004).

The creation of an imaginary companion is also not limited to early childhood. Several studies have found that both older school-aged children and adolescents maintain their early childhood companions, or create new ones (Bonne, Canetti, Bachar, De-Nour, & Shalev, 1999; Fernyhough, Bland, Meins, & Coltheart, 2007; Hoff, 2005; Pearson et al., 2001; Seiffge-Krenke, 1997; Taylor et al., 2004; Taylor, Hulette, & Dishion, 2010). In the same large-scale study by Pearson et al. (2001), approximately 22% of children aged 8 to 12 years reported having an invisible imaginary companion. And in a study of 241 adolescents, Seiffge-Krenke found that 22% of adolescents kept diaries that primarily functioned as an imaginary companion. Adolescents in this study were able to describe distinctive physical, personal, and situational characteristics of the imaginary companions they addressed, either specifically in their diary or during subsequent interview questions.

In typically developing children, the creation of an imaginary companion is believed to be a healthy activity. Having an imaginary companion is often associated with positive correlates, such as advanced referential communication (Roby & Kidd, 2008) and narrative creativity (Mottweiler & Taylor, 2014; Schaefer, 1969; Trionfi & Reese, 2009). In contrast, for children who have experienced maltreatment, having an imaginary companion has been viewed with more concern and as a possible antecedent to later mental health problems (Lovinger, 1983; Lynn, Rhue, & Green, 1988; McLewin & Muller, 2006; Sanders, 1992; Trujillo, Lewis, Yeager, & Gidlow, 1996). For example, in studies of adults with dissociative identity disorder, imaginary companions are often present in early childhood (Lovinger, 1983; Sanders, 1992) and have distinctive negative characteristics. In addition, imaginary companions of maltreated children in clinical samples have sometimes been described as performing functions specific to maltreatment, such as taking pain and abuse for the child or keeping the child’s secrets (McLewin & Muller, 2006; Trujillo et al., 1996). The current study was designed to provide some preliminary information about the prevalence and characteristics of imaginary companions in a sample of children who have experienced maltreatment and subsequent disruptions in caregiving associated with being placed in foster care.

The Relation Between Imaginary Companions and Coping

Qualitative and case studies of children in the general population with imaginary companions suggest that this type of pretending can help children cope with adverse situations and events such as loneliness, a family move, the birth of a sibling, and illness (Hoff, 2005; Nagera, 1969; see Taylor, 1999 for a review). For example, Murphy (1962) describes a 3-year-old boy who coped with the pain of surgery and medical care associated with reattaching and treating his amputated finger by creating an invisible elf named Woody who sat with him during his treatments.

Empirical research also associates imaginary companions with effective coping (Brinthauot & Dove, 2012; Singer, 1993; Singer & Streiner, 1966; Taylor et al., 2010). One of the most compelling studies was conducted by Sadeh, Hen-Gal, and Tikotzky (2008; also see Kushnir & Sadeh, 2012), who examined the extent to which having a personified object might reduce the stress-related symptoms of children who had lived in a war zone during the 2006 Israeli-Lebanon conflict (e.g., difficulty falling asleep, nightmares, excessive crying, anxiety and fears, clinging, and strong reactions to noise). Three- to 6-year-old children were randomly assigned to a condition in which they were asked to take care of a stuffed animal who was described as sad, lonely, and in need of a friend. At a 2-month follow-up, the 191 children who were given the stuffed animals to care for showed significant improvement in stress-related symptoms compared with a control group of 101 children. Although there was no information about the extent to which the children treated the toy as a personified object, the experimental design allows for the conclusion that the stuffed animal was causally related to children’s improvement.

While imaginary companions are often associated with coping and resilience, the imaginary companions created by children who have experienced abuse or maltreatment are sometimes viewed as red flags (McLewin & Muller, 2006). Clinical research suggests that the incidence of imaginary companions might be higher among maltreated children, and that these companions might be qualitatively different from the imaginary companions in normative samples (McLewin & Muller, 2006; Sanders, 1992; Silberg, 1998; Trujillo et al., 1996). However, the distinction between normative imaginary companions and imaginary companions that might be indicative of underlying pathology is complicated by several considerations. First, although maltreated children in clinical samples have high rates of imaginary companions (e.g., 52%, Trujillo et al., 1996) and the majority of adult patients with dissociative identity disorder report having imaginary companions as children (64%, Sanders, 1992), the creation of imaginary companions is a common childhood occurrence (Pearson et al., 2001; Taylor et al., 2004; Taylor, 1999). Having an imaginary companion—in and of itself—is unlikely to be evidence of a budding mental health problem or a response to maltreatment. Also, there have been methodological problems in research that associates imaginary companions with pathology. For example, Carlson, Tahiroglu, and Taylor (2008) found that children with imaginary companions scored higher on the child dissociative checklist but this was due to the items on this checklist that specifically asked about pretend play. In this study, having an imaginary companion was not related to measures of fear or problem behaviors. Another complication in identifying imaginary companions that might be associated with problems is that many typically developing children create imaginary companions that share some of the characteristics that have been attributed to the companions of maltreated children, such as being difficult to control, argumentative, bothersome, or unruly (Bender & Vogel, 1941; Taylor, Carlson, & Shawber, 2007). Finally, in studies with maltreated children, participants are frequently sampled from children with clinical diagnoses, making it difficult to determine more generally to what extent asymptomatic children with maltreatment histories might create imaginary companions that functionally differ from those created by typically developing children.

The Present Study

Children who have lived in foster care are a vulnerable population in the United States (Fisher, Burraston, & Pears, 2005). Many of these children have experienced a host of traumas, including severe neglect, and physical and sexual abuse (Fisher, Gunnar, Dozier, Bruce, & Pears, 2006). In addition, the stress associated with living in foster care (e.g., disruption of family relationships, changes in schools, ineffective foster parenting, and frequent transitions in and out of multiple foster care homes) can have deleterious effects on friendships (Shook, Vaughn, Litschge, Kolivoski, & Schelbe, 2009).

The purpose of this study was to provide some preliminary information about the incidence and characteristics of imaginary companions created by children who have been in the foster care system. We were also interested in how the prevalence and the characteristics of the imaginary companions created by children who lived in foster care might differ from children in a low socioeconomic status (SES) control group. Finally, we were interested in the extent to which children spontaneously describe their imaginary companions as vehicles for coping with a range of stressors, particularly for children who have experienced transition and uncertainty in the foster care system. To assess how imaginary companions helped children cope with foster care, we sampled 9- to 14-year-old children who had transitioned out of foster care and could reflect on their imaginary companions as a source for coping.

We speculated that a history of maltreatment and stress related to foster care might motivate some children to create imaginary companions who could provide stable companionship in the face of these upheavals. Thus, it is possible that the children who had experienced foster care might be more likely to have imaginary companions than children in a low SES sample. On the other hand, some researchers have found that children who have been abused are less creative in their play than nonabused children (White & Allers, 1994), which might suggest a lower prevalence of imaginary companions in the foster care group. Similarly, the imaginary companions created by children who have lived in foster care might be perceived as less creative than those of other children. However, we thought it might also be possible that the children in foster care would create more unusual imaginary companions that would therefore be rated as more creative.

In addition to prevalence, this study was designed to collect exploratory information about the extent to which a particular set of characteristics might distinguish the imaginary companions created by children who had been in foster care from those of a low SES comparison group. In intervention studies, improvement in stress-related symptoms was found both when children were given a stuffed animal who needed care (Sadeh et al., 2008) and when children were given a stuffed animal who was described as a powerful ally (Kushnir & Sadeh, 2012). How might characteristics of the imaginary companions that children spontaneously create overlap with the themes of both caretaking and alliance in these intervention studies? Because of the vulnerabilities associated with being in foster care, we speculated that maltreated children might be more motivated to create highly competent imaginary companions that could serve as powerful allies.

In addition to asking about imaginary companions, we included pilot questions about whether the child had an imaginary place or world that he or she liked to think about. Past research indicates that some children in this age range invent imaginary worlds or paracosms that can become an important imaginative activity (Cohen & MacKeith, 1991; Root-Bernstein, 2014; Root-Bernstein & Root-Bernstein, 2006; Taylor, Mottweiler, Naylor, & Levernier, 2015). Thus, we included these questions as a first step in exploring the extent to which children who have been maltreated might engage in this type of activity. We speculated that having a stable imaginary world of one’s own might appeal to children who have been in foster care, given the numerous transitions in home, school, and neighborhood that these children often experience.

Method

Participants

Participants were 60 children (mean age =11.10 years, SD = 1.31, range = 9.08–14.17 years; 31 boys and 29 girls) who had participated in a randomized efficacy trial of a preventive intervention program for preschool-aged foster care children and their caregivers (Fisher, Ellis, & Chamberlain, 1999). Twenty-one of these children (mean age =11.05 years, SD =1.36, range =9.58–13.67 years; 10 boys and 11 girls; 17 Euro-American, 2 Native American, and 2 Latino) had lived in foster care but were currently in permanent placements at the time of the interview. Of the 21 children who had lived in foster care, eight children (38.10%; 4 boys and 4 girls) had participated in the preventative intervention program; the other 13 children were in a standard intervention control group. The remaining 39 children (mean age =11.13 years, SD =1.30, range = 9.08–14.17 21 boys and 18 girls) were years; from low SES backgrounds and had never been in foster care (31 Euro-Americans, 3 African-Americans, 1 Native-American, and 4 Latinos).

For the foster care group, children who were within the selected age range and entering new foster care placements (i.e., first-time placements, reentries following failed reunifications, or transitions between placements) were referred to the original study through the local child welfare system during a 4-year period (2000–2003). The low SES comparison group was recruited by posting flyers throughout the community and advertisements in local newspapers. Families were included in the low SES comparison group if the child lived consistently with at least one biological parent, the annual household income was no more than $30,000, parental education was less than a 4-year college degree, and the family did not have any previous involvement with child welfare services as verified by child welfare services records. Children in the low SES group did not significantly differ in age from children who had lived in foster care, t(58) =0.23, p =.82. Children were given $5 for participating in this project.

Measures

To enable participation for children who had moved or who were living in other counties, participants were interviewed on the phone following the imaginary companions interview. The imaginary companions interview is a semistructured interview adapted from other interviews conducted by Taylor et al. (2004) with school-age children about their imaginary companions. This interview was designed to obtain thorough descriptions of the imaginary companions to assess their nature and function, including vividness, competency, and behaviors.

Procedure

Families were contacted and given a brief explanation of the project. If the parent or guardian indicated that his or her child was interested in participating, consent and assent forms were either mailed to families or signed during a visit to the center. Once signed consent forms were received, families were contacted by phone to schedule a date and time for a telephone interview with the child.

On the scheduled date and time, an interviewer (author N. A. or C. M.) called families, and verbal assent was obtained prior to the start of the interview. To establish rapport with participants, the interviewers first began with causal, warm-up questions, such as inquiries about the child’s day or weekend plans. The interviews lasted approximately 5 to 25 minutes, with no single interview exceeding 25 minutes. All interviews were recorded with a digital telephone recorder and were later transcribed by research assistants.

Coding and Data Analysis

Imaginary companions

Two independent coders who had not been involved in data collection and were blind to the child’s foster care status categorized the child as having an imaginary companion if the child reported having an imaginary companion currently or in the past and provided a description of either an invisible friend or a personified object. Children were categorized as not having an imaginary companion if (a) they reported not having an imaginary companion, (b) they reported having a current imaginary companion but did not provide any information about it, or (c) they described an imaginary companion based on a stuffed animal or doll (i.e., a personified object), but the description did not go beyond physical characteristics (i.e., the toy appeared to function more as a comfort object than as a friend). The agreement for the two coders was 97% for invisible friends and 100% for personified objects; disagreements were resolved by discussion. Note that although it is often desirable to collect information from parents to corroborate the presence and status of children’s imaginary companions, this step would have been challenging as many of the participants in the foster care group had have lived in multiple foster care homes with different guardians. However, parent corroboration is not as important for children this age as it is for younger children because parents are often unaware of the imaginary companions of children older than 5 or 6 years of age (Taylor et al., 2004).

Valence, creativity, competency, and relationship types

For children who described imaginary companions, the coders categorized the valence of the description (mostly positive, a mix of positive and negative characteristics, neutral, or mostly negative). There was 84% overlap in these categorizations, with disagreements resolved by discussion. The coders also used a 5-point scale to rate the descriptions of the imaginary companions for creativity (from not creative to highly creative), competency (from incompetent to highly competent), and type of relationship from the child caring for the imaginary companion (IC) (a score of 1) to the IC caring for the child (a score of 5). The correlations for the coders’ ratings were .81 (creativity), .54 (competence), and .78 (relationship), ps < .05.

Paracosms

In addition to the questions about imaginary companions, we asked children if they “ever thought about a pretend place.” If they said yes, we asked for a description. Two independent coders who had not been involved in data collection and were blind to the child’s foster care status categorized children as having a paracosms if they reported that they had made up a “special pretend place,” and gave a description that was not based solely on a book (e.g., “Hogwarts” from the Harry Potter series). Agreement was 88%; disagreements were resolved by discussion.

Results

Imaginary Companions

Prevalence and creativity

Three of the 60 children reported that they might have had an IC in the past but could not remember anything about it. Because the IC status of these children was ambiguous, we did not include them in the analyses. Of the remaining 57 children, 26 children (45.6%; 14 boys and 12 girls) met the criteria for having an imaginary companion (15 invisible friends and 11 personified objects). Having an IC was not related to group; 9 of the 20 foster care children (45%) and 17 of the 37 children in the community sample (46%) had imaginary companions, χ² (1, n =57) =.005, p > .05. Having an imaginary companion was also not related to age, t(55) =1.49, p > .05, or sex, χ² (1, n =57) =.17, p > .05. In addition, the creativity of children’s imaginary companions was not related to having been in foster care; M (foster care) = 3.33, SD =1.27, M (community) =2.74, SD =1.24, t(24) 1.16, p > .05. Girls’ descriptions (M =3.46, SD =1.32) were rated as somewhat more creative than boys’ descriptions (M = 2.5, SD, =1.06), but this difference was not statistically significant t(24) =2.06, p > .05. Creativity was not related to age, r = −.21, p > .05.

Valence

The descriptions of children’s imaginary companions were diverse with many idiosyncratic details (see Table 1 for examples). However, they were strikingly similar in valence. Twenty-three of the 26 imaginary companions (88.5%) were categorized as having mostly positive characteristics, and 3 (1.2%) had a mix of positive and negative characteristics (e.g., an invisible bear named Louie who was funny and often cheered the child up but also enjoyed sneaking up on the child and scaring her). None of the imaginary companions were described as mostly negative.

Table 1.

Examples of Imaginary Companions Created by Children Who Had Lived in Foster Care and by Children in the Community Sample.

Participant	Description of the imaginary companion
Male 9 years 7 months Foster care	An invisible Siberian tiger described as a “good best friend.” Helps the child feel better when he is angry or lonely. Described as having power swipes, but also needs the child to help him feel better when he is sad, and seeks comfort during rainy nights.
Male 13 years 5 months Foster care	A stuffed gorilla described as “funny” because of the clothes he wore. Helped the child feel better when his grandfather passed away. He was “pretty good” at things, but child did not like that they would sometimes disagree about activities (e.g., whether to go to the park).
Female 9 years 11 months Foster care	An invisible milk carton described as “very kind and kind of like a conscience.” Helped the child feel better when her sister did things she was not allowed to do. Child helped Milk learn to run, jump rope and hula-hoop. She liked that the milk carton was “not like a human being.” “I learned a lot about Milk and Milk learned a lot about me.”
Female 13 years 5 months Foster care	A stuffed pony described as really good at telling jokes. Helped the child by doing what a “mom would do … try and calm someone down.” Pony was described as a secret agent with X-ray vision who was really good at everything.
Male 9 years 5 months Community sample	An invisible boy who can “shape-shift into any kind of animal” so he “won’t get caught.” Described as “nice, generous, safe, crazy” because he “eats weird things like bugs.” Can also read “a 150 page book in a minute.” Helps child feel better by talking to him when he’s upset. Likes that he’s “super friendly,” but sometimes argues with him about activities.
Female 9 years 1 months Community sample	A tiny invisible boy who is “always there for … giving me ideas” and “very rarely needs me to cheer him up.” A mind reader who “finds out your secrets” and who is good at being “in very, very tight places.” Lives in a little house in a wall in her bedroom. Helps child feel better when her sister is being mean. Likes that “he’s kind and generous” and dislikes that he “sometimes pulls my hair.”
Female 10 years 2 months Community sample	A stuffed lion described as “nice and almost always happy” and “likes to talk.” Sometimes feels “left out” and “gets mad at me when I … talk to people about him.” Helps the child feel better by talking to him when she is lonely. Really good at “swinging on the swings and … .playing board games.” Likes that “we always play together.”
Female 13 years 5 months Community sample	An invisible girl who could fly. Described as “really nice”, “funny” and “like a superhero” that “succeeded … followed through and helped me.” Helped the child by hanging out with her when she was sad. Was “a part of my family and was there all the time.” Liked that “she was there when I needed her.”

Competence and care-giving

The mean competence rating for of the imaginary companions was 4.02 on a 5-point scale (SD = .87; range = 1.5–5). Although there were a few imaginary companions who were described as needing help (e.g., an invisible panda who needed the child to play music to cheer him up when he was sad), most were highly competent and often had super powers (e.g., the ability to fly, shape shift, etc.). Competency was not related to having been in foster care; M (foster care) =4.00, SD =.66, M (community) =4.03, SD =.98, t(24) =.08. In addition, competency was not related to sex: M (girls) = 3.83, SD =1.11; M (boys) =4.18, SD =.58; t(24) =1.01, p > .05, or age (r = .04).

Children were consistent in describing their imaginary companions as allies who provided the child with care and support. In Table 2, scores of 1 to 2 were categorized as indicating that the child looked after or cared for the IC, scores from 2.5 to 3.5 were categorized as indicating that the child and IC had an egalitarian relationship in which each helped the other, and scores of 4 to 5 were categorized as indicating that the IC was a supportive ally who helped the child. None of the children described a relationship in which the child looked after a vulnerable imaginary companion. All of the relationships were categorized as either reciprocal or as hierarchical, in which the imaginary companion provided care for the child. The mean ratings did not vary as a function of group: M (foster care) =3.5, SD =.71; M (community) = 3.88, SD = .92; and were not related to age, r (25) =.32, p =.12.

Table 2.

Types of Relationships With Imaginary Companion as a Function of Foster Care Status.

	Type of relationship
	Child cares for IC	Reciprocal relationship	IC cares for child
Foster care children (n =9)	0	6 (66.7%)	3 (33.3%)
Community children (n =17)	0	8 (47.1%)	9 (52.9%)
Total (N =26)	0	14 (53.8%)	12 (46.1%)

IC = imaginary companion.

Descriptions

During the interviews, children who had been in foster care and children from the community described their imaginary companions as helping them with a host of experiences, ranging from fights with family members to a death in the family. For example, a 13-year-old boy who had lived in foster care described his stuffed gorilla as helping him when his grandfather passed away, “… when I would get upset about something that happened that I didn’t want to happen – like … when our grandpa died - he helped me with that.” Most children described their imaginary companions as helping them cope with or regulate negative emotions. For example, one 13-year-old girl who had lived in foster care described how her stuffed pony had functioned the way a mother would when she was sad or angry. She said, “Whenever I got sad or mad or … I wanted to throw something and break it … it was … like what a mom would do, try to calm someone down.” Another 9-year-old boy from the community sample described how an invisible boy would talk with him about what was bothering him when he felt upset, “… when I’m upset he makes me … talk sometimes. He makes me feel a little bit better.” A number of children also described their imaginary companions as helping them in other ways, such as working with them on homework. For example, one 9-year-old girl from the community sample explained that her tiny invisible friend “was there for giving me ideas” and a 9-year-old boy from the community mentioned his invisible friend liked to read books to him.

Paracosms

Eleven of the 60 children (18.3%) described paracosms: 4 of the 21 children (19.04%) who had been in foster care (e.g., a place where there was lots of candy, no rules, and the child could get anything he wanted) and 7 of the 39 children (17.9%) in the community sample (e.g., a giant magical rain forest where fairies live in little castles and the child lives in a pink castle with pink furniture). Having a paracosm was not related to group, sex, age, or the report of having an imaginary companion. These preliminary data suggest that a substantial minority of children living in low SES homes and children who have been in foster care respond positively when asked about imaginary worlds. The brief descriptions children provided did not allow for content analyses but there were hints that children might design imaginary worlds that address issues in their own lives (e.g., access to material possessions). Taylor et al. (2015) have recently developed an interview strategy for collecting comprehensive information about paracosms that might be useful in future research.

Discussion

The main goals of this study were to provide some preliminary information about the incidence of imaginary companions created by children who have lived in foster care, and to compare the characteristics of these imaginary companions to those created by children in a low SES community control group. The prevalence rates of imaginary companions did not differ by group, and the overall rate (43%) was comparable to the general rates reported for 9 to 12 year olds for both past and current imaginary companions (22%–52%; Pearson et al., 2001). This preliminary finding suggests that having an imaginary companion is not more common in maltreated children, which is consistent with retrospective data showing that adults who described having an imaginary companion as a child were no more likely to have experienced abuse or trauma than other adults (Dierker, Davis, & Sanders, 1995). In addition, the descriptions of imaginary companions reported by children who had lived in foster care were not remarkably different from the descriptions reported by low SES community children. Both groups of children created companions that were rated as equally creative and described as good friends who were fun to be around and helpful allies, providing them with steadfast companionship and support.

The similarity in both the incidence and descriptions of imaginary companions for the two groups of children contrasts with assumptions that imaginary companions are more common in maltreated children and that there are characteristics that differentiate the imaginary companions of these children. Perhaps the clearest results from this preliminary study concern the valence of the descriptions and the types of relationships children reported having with their imaginary companions. Although it is sometimes assumed that the imaginary companions of maltreated children are likely to have negative characteristics, children in both groups described their imaginary companions in positive ways. Not a single child described an imaginary companion that was primarily negative. In addition, not a single child (from the foster care or community group) described a relationship in which the child was the caregiver of the imaginary companion. When imaginary companions were described as needing care, children also mentioned ways in which these imaginary companions helped them in return. And in many cases, the imaginary companion was described as a highly competent individual who helped the child, rather than the reverse.

However, there are limitations in what can be concluded from this exploratory study. First, our sample sizes were small; without replication, these preliminary findings cannot be generalized to broader populations of children living in low SES homes and children who have lived in foster care. Second, it is possible that our control group of low SES children included at least some individuals who had been maltreated, making the two groups of children more similar with respect to abuse than expected. Third, it is also possible that imaginary companions might be particularly likely for a certain type of abuse. If so, the variation in the nature of the maltreatment (e.g., neglect, sexual abuse, or physical abuse) that was represented in our small sample of foster care children is an important limitation of the study. Finally, up to 25% of children who live in the foster care system are not successfully transitioned into permanent homes (Akin, 2011). Because children in our foster care group had participated in an intervention and were all living in permanent homes, they are not representative of the full range of experiences and transition outcomes associated with living in foster care.

Concluding Comments

Engaging with imaginary companions has sometimes been viewed as a maladaptive coping strategy or as indicative of emotional, social, or psychological problems (Benson & Pryor, 1973). However, Singer (1993) has argued that the presence of an imaginary companion for a child in therapy should be taken as a positive sign that the child is using his or her own resources to cope with distress. The children in this preliminary study provide some examples of how their companions helped them cope with negative affect or difficult experiences (e.g., loss of a family member). In future research, it will be important to identify how an imaginary companion fits into a child’s larger set of strategies for coping with the stress of maltreatment and how the comfort and support gained from an imaginary relationship might be related to resilience.

Biographies

Naomi Ruth Aguiar, PhD, is a postdoctoral fellow in the Children’s Digital Media Center at Georgetown University. Her research examines how children conceptualize opportunities for relationships with real and imaginary others, including real peers, imaginary companions, and artificially intelligent agents (virtual characters and social robots).

Candice M. Mottweilier, MS, is a clinical intern at Kansas University School of Medicine–Wichita. She studies how imagination and creativity can help children cope with stressful life events.

Marjorie Taylor, PhD, is a professor Emerita of Psychology at the University of Oregon. She is author of “Imaginary Companions and the Children Who Create Them” and editor of the “Oxford Handbook of the Development of Imagination.”

Philip A. Fisher, PhD, is Philip H. Knight chair and professor of Psychology. His research focuses on developing and evaluating early childhood interventions in socially and economically marginalized communities, and on translating scientific knowledge regarding healthy development under conditions of adversity for use in social policy and programs.