THE I~IAGL'I;ARY
CmlPAXIOX
PH£''I;ml£.'I;OX:
A.'I; AX\LYSIS OF
PERSOXAlJ1Y
CORREUTF.S A.\1)
DEVELOP~IEXTAL
ANTECEDElI.'TS
Li~ C. Dierker. Ph.D.
Mstie F. Da\is. ~tA
BarbaJa Sanders. Ph.D.
Lisa C. Dierker, Ph.D., is a postdoctoral research associate
in me Dep..1.rt.menlofEpidcmiol<>s>'al Yale Unn'crsityin New
I-laven, Connecticut. Kristie F. Davis, j\·fA., is a graduate SllI-
dentin the Dcparlmemoflluman Development and Family
Smdies at the Universit}' ofConnecticUl, SlOrrs. Connecticut.
Barbara Sanders. Ph.D., is Professor of Psychology at the
Cni\'crsity ofConnecticul in Storrs, Connecticut.
For reprints, write Barbara Sanders, Ph.D.. Departmem of
Psychology. V-20, -l06 Babbidge Road, Uni,"crsit)' of
Connecticut, Storrs, Connecticut 06269-1020.
ABSTRACT
III the prestnt studies, both the incidence oj rualt ofall imaginm)'
cOlI/pallio" and Ihe r~membe'ml vividness oj (he experit!7IU were
assessed i" roI[,ge students. The purposeoJtI" research was 10 (lStn'-
tain lhe ex/ntt to which indittiduals in a nonclmieal population
who m:a/l having a childhood imaginary companion shnre rhar-
adms/its and negutiw life exptrimus wilh indwiduals diagnostd
(j';th Dis.sociatiw ldmtilJ Dis<mkr(DID); in (his rlinicalgroup rhild-
hood imaginary companions a~ ttportniJairl)'/fflJuently. and the
expninlceisdesaibed asextmnt·ly vivid (Sanden, 1992). Twoslud-
il'$ llln"e carried alit, In Study I, studmts of both sexes who wllem-
bered an imaginal)' comprl/lioll (IC+) wereJOl/tld (0 be more disso-
dlilille tlum those who repOl1fd /lot having a compill/ion (/~). 71/t'
let womm also scoud highn' on all Imagi'la(i~ Involvement
Invn/(01)' than IC women. This difJetence did not teach slatistiral
sig'flijicance among the male stlldmts. Stud)' 2 sCttened a 1If'W pop-
ulation offemalLstudnlls in Qrdwtocompare thrrtgroupsoJwomm:
A High 1'ividnessgroup (IN), a I.ow lividness group (Ll? and a
IllMOmjJanion group (IC). The fill group was r011lpriSLd oj IC+
UIOlIlm who said they had been able to Ii« and hear their childhood
imaginary compa'flion. mId who rrmnnbn-ed btlievillg Ihe compan-
iOI/ was real; Ott LV group cmlSMthi of le+ womell who {IlIswern/
MilO ~ to (hest' 3 T/ivicIness questions. ancI th~ IC grolll) was defined
as in Stud)' /, nle J-n' WOlllell llJi'f'tJOllnd (0 Ix sigtlijicantly high·
0' ill imaginativt: i"volvnne1/t thall the LV group. and also more
dissociative. Tht' LVgroup did not diffw signifirQlIll),Jrom the IG
group. In bothstudia. stwhnuumo repmted ~mnnberingan imag-
inary companion, even (hasi! whose experinlceu¢fp"reptuol1)' ttivid.
did lIot "port siglllficanl9' mart lonely, stressful or traumatic child-
hoods thall comparison groups.
220
J1'rotlRODUCfION
CharaClcristics associated with Dissociati\'C Identity
Disordcr (DID) havc received a gl'cal deal of auemion in
recent )'cars. Case studies suggest that individuals in this diag-
nostic calCgoryare exU'COlel)' h}pnotizable and also that the)'
exhibit man}" characteristics incorporated under the gener-
al rubricofimaginauve invoh'ement (Bliss. 1984; L)ll1l. RJme
& Green. 1988). One indication ofhigh imaginati\'e involve-
ment is the seemingly high incidencc of childhood imagi-
nary'companions. In the onlrquantitau\'e study of lhis phe-
nomenon in DID subjeclS. Sanders (1992) found lhat 14 of
the 22 adults surveycd (6-1-%) recallcd one or more imagi-
nary companions from their childhood. For nearly all of
these, thc cxpclicncc was described as exrremcl)' real and
vivid: 13 oCthe 14 (93%) said that tllC)' had been able to see
their companion. 12 (86)% could hear their companion, and
II (78%) bclie\-ed Lheir companion was real.
Although the incidence of childhood imaginat}' com-
panions in the DID group seems high (a commonl), cited esti-
mate for !.he generAl populations is aOOm 1/3 in children
between the ages oflhree and len). it is diffirull lO e-.'aluate
such data witllOul a comparison group. The prescnlsllldies
prmidc comparison dam rorone normal population. In ulese
studies, the incidence ofretrospccli\'cl)' recalled imaginary
companiol1sand the reported \i\idnessof!.he imaginalJ'com-
panion experience was assessed in college students, using
the same imaginary' companion questionnaire used with the
clinical group. In addition. we also obtained data on the per-
sonalit)' correlates of !.he experience and on childhood
antecedents. We h)'J)()(hesi1.ed thal the imaginary compan-
ion experience, panicularl)'when it isexu'emel)'\ti\'id or real.
requires capabili6cs similar to, or e\'en identical \\;th, lhose
associaled with DID. Accordingly, we assessed the degree to
which lhis experience is associated with the capacit)' to
becomc imcnsel)' absorbed in an imagined reality.J. J-lilgard's
'"imaginative inmh'emem M (Hilgard. 1974): \\itll dissociation;
and \\;tll negath'e or stressful experiences in childhood,
including severe punishment, sexual abuse. and neglecL
In Study I, college srudents who remembered ha\ing at
least one childhood imaginary companion (le+) were com-
pared \\;th students who reported not ha\ing an imaginary
companion (Ie). In Srud)' 2. a new sample ofsludents \95
D1S-'\OCl-\TIO\. \ III \1n. \0. t Drn.!ln' 199;
m • iot r-
al , or DES,IJif orin/ion. he Oi so it" i l' •
all '011
1011.
2 1
D1SS0Cl\TIO\. \'01. \'III. 'n. I. Ol'ITmbrr 19%
e<"l <C C\I to
'? ~ ... II C?
'1
:=
~
- '1 ~ C?
-a 'ti ;:s 2 c;-E
"-
ll. ~
.....
~ c X ati ~ Ho~ Envinmmmt, The hild...
-
~ or: ::'1 ~ .\bu nd Trauma cale ( T)
--
"0
-
~.
.:::: ~ nd ,,' B 'er-La: n, 1995) \\
"t' Vi d me frequency and... to m asure
:-:
'"J: :.> \ rity f .tTe and n leel in the
---
c. ,. III .-..
::: C\I ll. ,.; h m m;ronmenL The ,\T e
.f "0 r-- -::r:
'-
::
vi lh bj perception of the de:;) ct e
::: ., of \ form of maltreannent in
c C\I 0 ..., hildhood, Foreach of3 items. bjec
:': + c: C? 1.: ~ pondona~poin{~eurngmfromc- ~
-
r..
...
~n r~ to -alwa) . - to indicate the:=
c
r' :" n : of particular fi elin andIII
~ ~
occurring during childhood and
- ~:::; ::: :':l "0 ~ i:O C!c: e 1;1) 1 nce.T t-retest reliabilitY ofthe
'to
l'C :"
'""
aJ in the oll e population 9 andE v...
m im mal con istency reflect dinc:;
"iii l':I 0 C'f cD nba h' alpha tati tic is .90. Factor~ ~ ~ rt >. - .....D ...... nal} i h tablished three ub calc':r >
"9 >.:ual abu ,punishm nt andn tin:
-5 ~ h m anno phere/ne leet. Each ub-~
'- 0- j l' T core is th mean of Lh 3
"0 0;) <0 co
vi t<"l l!') e<"l
-
ir 111 ; and ub cale core in ea h ca ec: 'Jl ~ +I +1 .:!:!.0.. th means of the items in th par-
'::; ::J
l'l 0 ~ Ii 'ular sale .'::;: ...
41 r- lfl <.0
a ... =s C! e<1 C!t:
-0 .2 ULT
;:a c:
0- ~ 2 .-.. ~c E C"l C'l r-- Imaginative Involvement!3 ~ c.o t- t-: ~J. ..!!. J. Th means and tandard de,iati n
-0
on the 3-1 and its thr ub-... ;~ c:; r men and women in the twO
'"
c +
... .~ co= C'f 'n 'n
,~ ... ~ t-: '-C pre emed in Table 1. A two-E -~ ...... revealed mat th I
-
oup
ificanth' muher than the I
up n me overall 3-1 (F(l.III) =
7, 1, p<. 1). a result which \\ nOl
.3 ~ alr r d when me four items comprisin
-
~ ? th im nan' companion scale werev;
~ v:. mill d from the in,' mary. The I
-::: .-::a ,..
up " also found to ore 'gnifi-...E :-: .- 11 hi her man me IC- ach
:0 ... l'C oupon
"E£ .;L c- fth three ub~ . imagination (l.
'i: r: ~ E... - .... 11 ) = 37.2-1. P < .0001). paranormal> . .- :'i:i '='1": ~
2 2
DISSOUmO\. \'01. \lIL \0. -l~r \!9j
M an. and. I nd rd I vi, lion roup n lh • 1. .', • nd ( rud 1)
Hi h Low
iviOO
I I
ariabl an .d. an .d. fean .d.
3-1 1.7 ± .6- 1
-
.33 .71 ('"
D .4.37 (_ 1 .4_ 1 •• .J .4 JI. 2 (-
).97 _0.5 O. . O.~ .71 (-'- 7
r
ur
up
P I b
n I U lh 'r. no.' cliff< r >n c r x by
intl~raclion \ pre, enl.
oup inter-
ironment
r. UI d bet\\' D th I
rOll m :an =.71) II th
llL
Di ociatiol/
( 1.113) = 7. . P<. 1) nd.· cxpc.: led. ill inan' c m-
panion ( (1,11-):=11 .2.p<. 1),confimlin III reli-
ability 0 lhe ' recuing ill 'lrUIll 'ilL cor's for me I and
J I' up n lh I n I1lr I il m.s' ere nOI i nifica.nt..l, dif-
I II.
In <tddili( II IOll1(' 'ibTllifi 'alll main effects me 0 As
I' V al i nin anl ' b' I' lip illlcract.i n on the overall
in nl I ( ],111) =..:{7.! < .0-), til· imagination· aJe
(I,ll ) = 4.1 ,p < .05). alld th imaginary companion
scale F (1,115 = .13, P < ,025). heIr po t hoc tests of
lh' in ora lion. I' 'vealed the ['0110\ ing: 1) t..Il<.~ group dif-
fer n on 111 0 n1l1 inv 'nlory was tali ticaJl reliable for
wom n (I < .00 I), b I II l r J" 111 'n; 2) t..Ile group dill renee
011 lh· imagill<l il)ll snll ' of Ill' ill 'ut( I approa hed 'ro3-
tisli ~II sign ifie nee fon III '11 SII ~j 'ts (.05 <P< .10) , where-
a lhe cliff rell f l' III ~n did n l; ) On the imaginary om-
panioll cal 'lll ' 'xI' 'l' 1 "roup liffer'n" w· . obtain cI for
hOi h - x s (p <.n I for cm<llc-, p <: .025 for male), and me
1 + \ men' l' cI 'igniii ant..l high r man Ie men (p <
. 2-), r '11 'Cling a gr 'al 'rill\' I 'Ill nL wim h 'ir imagina
olllpani II . r~~II.·· t iz '. were und to be m der-
ale.
223
D1ssoamo\'. \'oJ. \1JI. :\0. t lHmnbrr I~j
~lea and randard \i.ation r up on we 3-1 and i al (rudy 2
Mean and tandard
An
'arion for we \ l\idn, r up on D and i ub cal (rudy 2)
High ow
Vividne Vivicln s
IC+ I + I
ariabl Mean .d. Man .d. an .d.
2 .1 (_ 14. 4) 21. (_ r.1 ) 17.3] ("" 10.53)
1maginauye
hl\'ol\"cmen t
cale 39. (... .24 (= 21.4-5 -.15 ± 16.09)
224
I ~
.07
<=14..7
(± . 6)
7. (... 9.10)
(_ 9.75)
-. -1
7. (_ 10.12)
Dlssoa·mo~.rul. \lll. \0. t Dl'rt'mhrr 1995
r up
. ,\11
all rcliabk Sc diflt'n'n > that \ re found for this fa t r
rru IlIre, group eli 'll'IICl' \\cre c. amincd in thi " lu 'i" '1)'
r, male ampl
n and
tao i-
'\ 'rc: idcl1lificd a' b illg IIV1(: 'lIb-
thod
ni
natk
amu
Dl'iSOCLmO\. \01. \111. \ll. t~I~,;
IMAGINARY COMPANION
) is beu..'een the imaginau,<e inyolvemCIll factors from bOlh
me DES and 3-1 gavc this finding further support (r =.63).
The ~imaginati\"c in\'ol\'emenl~questions contained in the
DES seem 10 address a kind ofdissociative absorption r<llhcr
than the more general itwol\'cmenl in imaginath"c actidtics
mcasun.-d b>' the 3-1 factor.
Negative Home E"v;rommmi
An .''-".....OVA revealed no significant. differences among the
HVIC. group (mean::: I.02), me LVIC+ group (mean; .91)
and the Ie group (mean = .91) on the CAT scale or all its
subscales.
DISCUSSION
ll1inccn percent of male college srudents and 23% of
females reponed ha\'ingan imaginar)'companioll asachild.
and 10% ofthe men and 19% orthe \,'omen reponed a firSl-
person memory of their a companion. As expected. these
percentages aTe much lo\,'er than that reportcd by mc clin-
ical group survcyed b)' Sanders (1992) in \\'hich 64% per-
cellt of the DID respondents reported having an imaginal)'
companion whom the)' could remember firsthand.
The Iwpothesis that college studcnts who remember a
childhood imaginal')' companion would be more dissocia-
tive as well as higher in imaginati\'e invoh'cmcntthan those
who do not was dead)' confirmed for womcn. and was p.'l.r-
tiall)' SUpportL-d for men, Though the IC+ men were signifl-
C'dntl)' more dissociati\'e than IC- men. the difference in imag-
inative ill\'Olvement associated \,ith rCC'.dll of an imaginary
companion did not reach statistical significance. Howc\'er.
the sample of 10. men \,<15 extremel)' small. Additional
research \\ith largcr samples and equal numbers of 10. sru-
dents of both sexes is needed to pursue the suggestion that
the correlatesoft.he imaginal}'companion phcnomcnonmay
be different in men and women.
The degree of vividness I... ith which imaginary compan-
ions are experienced has been given little systcmatic research
attention, yet the LOpic has oftcn been discussed as an inter-
esting facet afthe phenomenon. lmaginaf'\'companjons rna}
be vCl")' "hidl)' imagined. and possess a sense of realit), for
their creators. Through retrospective report. Hudock and
Burstein (1932) found that81 % offemalesubjens\\ith imag-
inal") companions and 60% of male subjects testified that
their companions were reallo lhem. "'hile 79% of females
and 43% ofmales imagined com'ersing \\;th their imaginal'}'
companions. Svcndson (1934) reported that 330f the 40chil-
dren IC+ he studied. (Le., m'er 80% of lhem) played with
their companions in a WlI)' lhal suggested tbatthe compan-
ion occupied space. -The)' were spoken to direCl.l)', were
chased in games. and \\'crc brought to lhe llIble where a place
was set for them - (p. 995). While some children describe lheir
companions as endowed I\ith apparent reality and marked
\;vidness. others describe characters Ihat arc much less per-
226
manem and \i\id Oersild. ~IarkC)" &JersOld. 1933).
In the present studies the imaginal} comp..'lnion expe-
rience "'"'<is designated as \;\id if the smdent reported being
able to see as well as to hear the imaginary companion. and
also believed (in childhood) thaI tlle companion was rcal.
This was tlle sort ofexpeli.ence described b}' nearl}' all oftlle
DID respondents who reported imaginary companions. This
\'ivid experience was reported far less frequentl)' in lhe nor-
mal college population. ani}' about 5% of the women Stu-
dellts sUl"Ye)'cd - roughly one quaner of the IC+ women -
endorsed all three vividness ilems. Those who did were sig-
nificanuy morc dissocialive than IC- women. xotsurprisingl),
those whose experience was \i\;d and real were also lllorc
i1wolved "iill their companions: talking to them, taking them
along \,ith lhem. and being comforted by them. Our results
suggCSt lhal individuals from the nonnal population \\"110
have a \i,;d imaginaJ)' companion experience arc similar to
om groups in having e1evaled dissociativc and imaginative
capacities.
Ja}lles (1977) reponed Ulal about halfof the <ldlllts that
be slln'cyed said lhe)' had ocen able 10 hear the miceoflheir
imaginary companion. Our findings indicate that hili esti-
mates may ha\'e been high, but confirm his obsen'<ltion that
for maIl}', the experience is indeed \'ery real. Marjorie
Taylor's recent \','ork (Taylor. Cartwright. & Carl.son. 1992),
,md obscn.uions of others (Svend.son. 1934. in particular)
auestto the \ivid imaginarycompanioll experience in man)'
children. II. should be noted, hOI\'e~;er, that. the 'i\idness of
the imaginaf'\' companion experience ma) be defined in SC\'-
era! ways. Earlie.' studies have included di.-.cussion ofthe indi-
\'idual'sabilil) 10 tOuch the companion. the indi\idual'sabil-
it)' to describe Ihe companion in detail or the companion's
ability to occup)' \,<lriOlls lypes of space. FUl.llre research on
tlle vividness of lhe imaginar)'companion experience should
not be Iimiled to a restricted set ofvi\'idncss variables, but
should encompass various dimensions of the experience.
All oful(~ results suggesting a relationship between a vivid
imaginar)' companion experience and DID, including our
0\\'1, are based on retrospective reports. and the possibilit)'
remains ulat. it is thc memoryof the experience which iscrit-
ical rather than the experience itself. Particularl), sincc one
characteristic of imaginati\'c im'olvemem is \i\id autobio-
graphical memory, this possibilit\'should nO[ be minimil.ed.
despite the difficulty of resoh'ing the issue empiricall)'.
Procedures for invcstigating subjective aspects of the imag-
inar)' companion experience in roung children (Ta)'lor et
aI., 1992) offer a means ofaddressing the question b)'dctcr-
mining correlatcs of the vivid imagina'1' companion expe-
rience in childhood.
Despite conflicting views regarding lhe degree ofadtlp·
lation of children with imaginal}' companions. the creation
of imaginary companions, C\'en in normal children, is offen
SCen as a defensive phenomenon (Nagera, 1969: Merer &
Tuber, 1989), Nagera (1969) notes that the most apparent~
D1i~mO\.\01. \11L \"'~.~ I"j
and bcha\;or.l1 Of·
1111·
nali(\' and E n-
lao n? Bnlt hjoflmal
fLo
). D 'I'do m~nt. reliabilitv and
TIll' journal oj Xl'tl/O! and ,(Imlal
Bent L ill, ;., '. PUIII,IIII,
I. lidil, 01 ~l di~ 0 i<tLioll .
/), fI\t', I i-l( (2). 727-i:n.
Rl1l1l·,JW.. <' L 1111, S,l. (l9R7 . F;lIlIasy prollelle '~: DcI' -I pm 'l1lal
alll('n'dl'III~. jfJllmnl of 1'1'/ 1////fllil)l. --( I ), 12]·137.
alld 'I s. Ij., . ;iol(ls, M. ( 199 I). 1)isso ial.ion and hildhood U'all-
lila ill ps c1llllogicall disltlr!wd adolf'S Ill,. AIIlI'r;aw.lourl/{l! of
P,\'Vr!lia II)', J./8( 1). ;,o·!),'I.
I3Ii ,1-•• (I. 4). nl, Il(-'()LI! <;t'1-h\'pno~i ill ~1PD. P"dltalnCC!IIIIr.
ofSortII Immra. 7. I:{).. II .
I'urlc ~. ',B... Bur Lcill. ~l. II :l2). Thl' intagUl~"playmaL : -\
qll ti lllMin' III h./mmllli of llrllC P')dlOll . -fl, •• 0-. _2.
1'1I11I.lm, F.. (;11101 ,J., ill '1 II IAll , E., Barball. L.•. Po<t R. (19' 1 .
'1 h c1ini ,II ph 'JH 1IIt'lIolo!y or IlIl1hipl per onality di rder:
R('\1 'wof 100 I' 'nl (' . ../otllllrll o! 1ill/cal Pf.l'rluallJ' -17(6).2 5-
~l:t
'a lid 'rs. B. (II C 2), II· illlagiltar}' cOlllpanion cxp ri nee ill 11I111-
liple per IInalitv eli VI d '.. 1JI..\()(.f,I'I'IO,\'. -(3 , 16()-16~.
,and('r, B.. :, 'II,J.. I i "I.. 'I. L., Da\'i.. K., . Giola.~. M. (IYY~ .
Imflf!l'lflilfl/ll/l IIlIclli/JIt, /)('1 IIII/flli/\" diwrdn: DI'vl'lo/nnm/ ofan lInagtnflhv
IIIT/OIr'1I1I1'II1 III IlIlIfII). l'lIpllbli l1('d III.ulli cript. Ililer iL\' 01
,olllll'ninll at Storr,
Sant'l .B.. ·B'I..t:I-Lau't('Il,f.{I1.5),rhcm urement rp~­
",1 i al 1lI<llu ';lUn 'Ill: F lrh clam Oil the child abu e and trolll-
al . (111M \bll "mi. f'Klr11. I' :\)' 315-: 2..
aJICI" ,B., . ;n'('II,j.\.(l.. ~).The aClOr Lruetllr o[thcdi
dathl'l"pt:llcn t' aklll It t'~llIdenu.Df Da.lJJo.\'.7(I) •
.) -
- _I.
uard tb
an inl" !!lnary om-to
Baird.tin, Illa/l/lld-
1".11/1'
')-
__ I
DISSOClATIO\, \ol \llt \0. t Ikttmbrr 199;
I~IAGINARY COMPA.\,ION
Tallof-. M.. Um.Tl~hL B.. & Carbon. S. (1993). A ck\-dopmenQ!.
IIIn:sog:oiItion of c.hiktren's imaginU'\ companions. lJft!ftoprttntlaJ
P1)dwiog]. 29. 27&285.
Wilson, S.. & Barber. T. (1981). Vivid famasyand ImllucinalO!1'abil-
ities in the life hi~torie~ of excdlent hypnotic subjects (~somnam.
bulcsT Prcliminaf) report \\ith female 5ubjec15. 111 E. Klinger (Ed.),
I~ l'o",~_2:Colltt'jJU.muflsandapplimtjORJ Cpp. 133-149). Ne-..'
York; Plenum PreY.
228
Dl\SOQH10\ \ __ \lIL' ... -l.~I~;