Virtual Research on Imaginary Companions:
Using the Internet to Gather Adult Retrospective Accounts
of Imaginary Companions in Childhood.

Lorraine Ball & Kimberly Wright Cassidy
Bryn Mawr College

Christopher Lalonde
The University of Victoria

Address for correspondence:

Lorraine Ball
Psychology Department / West House
Bryn Mawr College
101 N. Merion
Bryn Mawr, PA 19010


Presented at the 28th Annual Symposium of the Jean Piaget Society, Chicago IL, June 11-13, 1998

Draft Copy: Please do not quote without permission


Up to one-third of young children indicate that they have an imaginary playmate (Manosevitz, Prentice & Wilson, 1973; Taylor & Carlson, 1996). Nevertheless, little is currently known about factors associated with the creation and abandonment of ICs or about the functional roles ICs play in children’s lives. The present study used Internet-based questionnaires as a means of gathering retrospective self-reports and current or retrospective parent-reports of the nature and functional role of ICs in either their own or their children’s lives. Questionnaires asked individuals to articulate, with the help of target questions, their own thoughts about their (or their child’s) particular IC and the circumstances surrounding its existence. Questions of interest included 1) whether such questionnaire data can be collected over the Internet, 2) if that the Internet data on function can be reliably categorized into 7 distinct but non- exclusive functional roles, 3) whether the functional roles attributed to ICs by self-reporters over the Internet were similarly spread over functional categories as the functional roles attributed to ICs by a traditional non-Internet tested sample of undergraduates, and 4) whether parent-reports and self-reports of functional roles emphasized different features.



The questionnaire consisted of a set of hypertext documents made available on the Internet ( Though the documents were stored on a computer in Vancouver, Canada, participants located anywhere could access the questionnaire from any computer connected to the World Wide Web. Recruitment messages were sent to various Internet news groups thought to attract persons with relevant interests (e.g. groups devoted to children, parenting, etc.), inviting them to participate in the study by completing the on-line questionnaire (see hard copy). At the WWW location, participants read a short introduction to the study and could indicate whether they, or their child (or grandchild), had an IC or not. Based on their response, 1 of 3 questionnaire versions appeared (IC-Parent; IC-Self; IC-No).

Each of the questionnaires was tailored to suit the respondent’s choice: The IC-Self and IC-Parent questionnaires were identical except that the questions were phrased either with reference to a child or oneself. IC-No contained only the demographic data sections of the IC-Parent and IC-Self questionnaires.


A total of 81 self-selected participants have completed the questionnaires to date. Of these, 22 indicated they did not have an IC (IC-No group), 18 reported that their child or grandchild had an IC (IC-Parent group) and 41 reported that they had an IC (IC-Self group). Mean age of the self-report group was 26.5 years (range 13-61 years). Parents were not asked their age. Interestingly, although an estimated 60% of Internet users are male, 78% of respondents to the questionnaire were female.

Reported education levels were high with 72.7% having at least some college education, and 42% having at least a BA. Almost half of the sample were married or living with a partner (48.9%) and the majority were working full-time (68.7%), with the percentage of part-timers and unemployed at 17.5% and 13.7% respectively. No differences were observed between the groups in level of education or occupation type (F-T, P-T, or unemployed). Respondents were from the U.S. (64.6%), Canada (11%), the U.K.(8.5%), Europe (6.1%), Australia (4.9%), and Latin America (4.9%). Information on family constellation indicated that 17% were "only" children. Of those with siblings, 33% were the oldest, 20.4% were "middle" children, and 29.5% were youngest children.


Responses were directed by e-mail to the authors. Respondents who gave their e-mail address were thanked for their participation. Responses to the more quantitative categories were directly imported to a database file for later analysis. Responses to qualitative questions were first coded and then entered into the database.

Functional categories

Responses across several separate questions about the ICs were coded for functional role. The categories of function were originally derived through a review of relevant research and an analysis of the responses from previously collected pilot data from a college sample using a pen and paper version of the same questionnaire. Responses from the present Internet sample were able to be coded into the same 6 original categories. A seventh category was also added for the present study but college data were not re-scored. The final 7 categories are described in brief below:

CONTROL: IC serves as something to control, take charge of, parent, boss around, etc.

COMPANION: IC provides companionship, alleviates loneliness

NURTURANCE: IC provides comfort, understanding, support, unconditional love, dependability

EMOTIONAL OUTLET: IC expresses emotions child can’t, or facilitates expression of emotion

ALTER EGO: IC is or does things child can’t; IC takes blame for or from child

FACILITATES FANTASY: IC provides vehicle for fantasy play; fuels child’s imagination

GUIDANCE: IC gives advice, guides; helps child work through problems or emotions



© 1998-1999 Lorraine Ball, Kimberly Wright Cassidy & Christopher Lalonde
Address comments about this page to: Chris Lalonde <>