TECSE22:1 39-54 (2002) 39

The Effects of Trainer-Implemented

Enhanced Milieu Teaching on
the Social Communication of
Children with Autism

TerryB. Hancock
T his study examined the effects of Enhanced Milieu Teaching on the social com-
munication skills of preschool children wit h autism when delivered by trained
interventionists. A modified single-subject design across four children was used co
assess the children's acq uisition, maintenance, and generalized use of language targets
and social communication skills as a result of the intervention. Observational data indi-
and Ann P. Kaiser cated that all children showed positive increases for specific target language use at the end
Vanderbilt University of 24 intervention sessions, and these results were maintained thro ugh the 6 month
follow-up observations . There was also evidence of positive changes in the complexity
and diversity of language for children on observat ional measures. Three of the four
children also generalized these positive language effects to interactions with their
mothers at home, with the greatest changes seen immediately after the intervention .
Parent satisfaction with the intervention procedures and child outcomes were high.

Autism is a pervasive developmental disorder char is de- lap , J988; Lovaas, 1987; McEachin, Smith, & Lovaas,
fined by marked difficulties in social interaction and acqui - L993) . Na t uralistic teach ing methods have also been
sitio n of adequate socia l co mmu nication ski lls. Children shown to be successful i.n teaching new ski lls to ch ildren
diagnosed with autism span the range of cognitive abili- with a uti sm (Koege l, Koegel, & Surratt, 1992; Koege l,
ties; about 70% will test in the mentall y retard ed range O'Dell, & Koegel, 1987; Laski, Charlop, & Schr eibman,
(Edwa rd s & Bristol, 1991) . Alth ough individual s with 1988; McGee, Krantz, & McClannahan, 1985). Nat-
autism vary widely in th e extent to which they manif est uralistic strateg ies begin with the learner's intention to
the sympto m s associated with autism, nearly a ll childr en communicate and the trainer 's abi lity to systematically
with auti sm will requ ire intervention to support the de - provide mode ls of approp riate communication forms
velopm ent of communication sk iJls. Researc h has demon - and meanin gfu l soc ial conseq uences for communication
stra ted that systematic int ervent ion ca nnot only improve attempt s.
the language a nd socia l ski lls of child ren with autism, but It is clear that for languag e intervention ro be effec-
it can a lso positively influence the behavi ora l, social, and tive with children who have autism, it is important that
academic outcomes for these children (And erso n, Avery, (a) the inter ven tion begins early; (b) the intervention tar-
DiPietro, Edwards, & Chr istian, J 987) . Co nversely, with- gets the soc ial use of language, not just the content of
our effective ear ly soc ial communication intervention, the language; and (c) the intensity and duration of th e inter-
behavioral symptoms associated with autism are likely to vent ion are sufficient to support the acquisition and gen-
be more severe and disruptive to the ind ivid ual's deve l- era lized use of new sk ills (Wether by & Prizant, 1992).
opment and daily living (Dura nd & Carr, 1992). Alth ough there are a growing number of intervention
Studies have demonstrated that children with autism st udi es examin ing the effects on acquisitio n of la nguage
can benefit from th e application of behavioral procedures for chi ldren with au tism, no known studies to date spe-
to reach specif ic lang uage sk ills (Koe gel, O 'Dell, & Dun- cifically exp lore whether those changes in ch ildren 's Ian-

Address: Terry8. Hancock. Deportment of Special Education , Box 328 Peabody College of Vanderbilt University,
Nashville. TN37203; e-mail: [email protected]

. p,o-ed
40 Topicsin Early Childhood Special Education 22:1

guage generalize to inreractions with their parenrs. Parents were taught (e.g., labels, actions, requests); only one study
of child ren with autism frequently reporr their desire to (Scherer & Olswang, 1989) targeted early semantic rela-
communicate with their children in ways that are func- tions (i.e., two-word commenting utterances). Increases
tional for both parent and ch ild (Norton & Drew, 1994). in spontaneous language were observed as a generalized
In reality, disrupted parent-child communication inrer- effect of the intervention in some studies (McGee et al.,
acr ions arc frequent in families of children with autism, 1985; Scherer & Olswang, 1989).
and difficulty in communication can exacerbate challeng- The definition of naturalisticlanguageteaching varies
ing behaviors exhibited by some children with autism widely in reported resea rch studies from direct and inci-
(Landa, Folstein, & Isaacs, 1991; Lemanek, Stone, & dental reaching procedures in everyday settings (Anderson
Fischel, 1993). Generalization of childr en's language ski lls et al., 1987; McGee et al., 1985) to re~ponsive interac-
ro significant people in their lives cannot be assumed to tion strategies (expans ions, building turn taking, following
happen as a result of a language intervention because the child's lead; Garrner & Schultz, 1990; Hunt et al.,
ch ildr en with autism have particular difficulty in general- 1988; Schre ibman, Kaneko, & Koegel, 1991 ). Two studies
izing social rnmmunication skil ls across settings and peo- have blended prompting language responses with aspects
ple. Given the signif icance of communicat ion within the of responsive interaction (Koegel ct al., 1987; Schreib-
context of the family, it is extremely important that any man et al., 1991). None of rhe studies emp loyed the range
language intervention with children with autism also ex- of responsive interaction techniques that are foundational
plore rhe effects of that intervention with their families . to EMT. Overall, the descriptions of the intervention pro-
cedures in past studies suggests that these procedures
were Less naturalistic, more structured, and more ori-
ENHANCED MILIEU TEACHING ented toward instruction than the application of EMT in
the present study . Most important, the majority of re-
Enhanced Milieu Teaching (EMT) is a hybrid approach to ported studies focused on ly on teaching new bnguage
naturalistic, early language intervention. EMT incorpo- sk ills rather than building conversa tional interaction or
rates aspects of both behavioral and social interactionist support ing child participation in social communicative
approaches to language intervention. There are three com- interactions. None of the studies explored the effects of
ponents of EMT: changes in children's language with their parents.
In a previous study (Kaiser, Hancock, & Nietfeld,
J. environmental . arrangement to promote
2000), we examined the effects of EMT when imple-
chi ld engagement w ith activities and com-
mented by six parents of young chi ldren with autism.
munication partners (Ostrosky & Kaiser,
The parents learned to use these naturalistic language in-
1991 );
tervention st rategics during 24 individual training ses-
2. responsive interaction techniques to build
sions and genera lized cheir use of the srraregies co home
socia l, conve rsationa l interaction and to
interactions with their chi ld ren. Positive effects were ob-
model new language forms (Weiss, 198 J);
served on the use of communication targets for a ll chil-
and dren and on the comp lexity and diversity of productive
3. milieu teaching procedures to prompt,
language for most chi ldr en . Child effects generalized and
model, a nd consequate che use of new lan- maintained for four of the six ch ildren, and there was ev-
guage forms in their functional contexts.
idence of change on developmental assessments of lan-
Milieu teaching procedures include elicitive model- guage for five of six chi ldr en.
ing, mands followed by models (Warren, McQuarter, & The Kaiser et al. (2000) study provided an impor-
Rogers-Warren, L984), time delay (J-Ialle, Baer, & Sprad- tant context for the current examination of EMT imple-
lin, 1981), and incidental teaching (Hart & Risley, 1975). mented by skilled therapists, particularly in terms of
EMT has been shown to be effective in teaching new exam ining the effect of th e intervention on interactions
generalized language ski lls co childr en wit h developmen - within fami lies. The findings of Kaiser er al. indicated
ta l disabilities (Hemmeter & Kaiser, 1994; Kaiser & H es- that EMT is an effective early communication interven-
ter, 1994). tion for chi ldren with autism and that parent-chi ld in-
Studies of milieu reaching wirh child ren with autis m teractions are posicively affected by teaching parents co
have shown that naturalistic language teaching can beef- imp lement the EMT procedures. In the current study, we
fective in teaching the m new socia l communication ski lls focus on the app lication of EMT by skilled therapists
(Gartner & Schu ltz, 1990; Hunt, Alwcll, & Goetz, 1988). and examine changes in pa rent-child interactions at home
Results repo rt ed in language intervention studies with resulting from children's exposure ro the EMT proce-
chi ldren with autism have been generally positive: Chi l- dures in the clinic. Thus, we investigated rhc indirect ef-
dren learned, and in some cases, generalized the targeted fects of naturalistic communication training on children's
language responses. Typically, small set!>of language ski lls interactions in the family.
TrainerEMT 41

PURPOSEOF STUDY 6-month delay in express ive languag e as mea sured by the
Sequenced inventory of Communication Development
The purpose of this study was ro examine the effects of (SICD; Hedrick , Prather, & Tobin, 1983), (c) the chi ld was
trainer-implemented EMT on the language performance verbal ly imitative, (d ) the child had an expressive vocabu-
of pre schoo l childr en with autism. Four questions we re lary of ar least 10 spontaneou s words, and (e) tbe child's
posed: hearing was within th e normal range as measured by an
audiometric assessment. In addition, the ch ildren's parents
J. What changes in child language rargecs mer the fo llowing criteria: Parent s co nsented to (a) bring
and social co mmunication ski lls will be the child to the clinic sening twi ce each week during the
observed during the clinic applications of int erve ntion phase and once each month for 6 monrhs
EMT by trained interventionists? during the follow-up phase, (b) allow project staff to
2. Will changes in child language targets and co mplete nine home visits over the course of the study,
soc ial communication skills be maintain ed and (c) be video ta ped at home with their chi ld and allow
over rime? their child to be videotaped with th e trainer at the clinic.
3. Will changes indicative of child language The children were three boys and one girl (C hi ld B)
developmenr be observed? with a mea n age of 44 months (range 35-54 months ). All
4 . Will changes in child language generalize chi ldr en had expressive and recept ive language skills in
ro interacti o ns with parents at home? the 20 - to 28-month range (Expressive Communication Age
fCAl M = 25 months, Recepti ve CAM= 2 1 months) as
tested by the SICD. The children's average Mean Length
METHOD of Utterance (MLU) at enrry was 1.29 (ra nge 1.03-2.00 ),
and their measured IQ scores ranged from< 50 to 95.
Participants Three indiv idua ls served as interventionists for the
Four preschool children with autism and their mothers four chi ldren . One interventionist who worked with two
participated in the srudy (see Table I). The child ren were of the four children (Child A and Chi ld D) had a master's
diagnosed as having aurism or pervasive developmental degree in specia l education and had co mpleted 36 hours
delay with autistic symptoms by an independenr child toward her doctorate in specia l education. The interven-
eva luation clinic befo re participating in the srudy. Each tionist for Child B had a bach elor's d egree in education
child met th e following criteria: (a) The child was be- and had comp leted 24 hours toward her master's degree
tween 2½ and 5 years of age, (b) the chil d had at least a in special education. The interventionist for Chi ld Chad

TABLE1. Participants

Language age
Child CA in mos Gender Diagnosis in months (SICD) IQ Targets
A 50 boy Autism 20-Ea < soc Agenc-acrion
20-Rb Action-object
Attribute-object
2-word request
B 38 girl POD 28-E 95d Agent-action-object
24-R Attribute-object
Preposition-object
J-wor d request
C 35 boy POD 24- E < soc Agent-action
20-R Action-object
Arrribute-object
2-word request
D 'i4 boy Autism 28- E 52d Agent-action-object
20-R Arrributc-object
3-word request

Note. CA =chronolog,cal age; PDD = pervasive developmental di~,,rclcr. '11: = Exprcs~ivc language a!(c as measured by the Sequenced ill11e11toryuf
Cum111"111a1tum Development (SICD: Hedri ck, Prarhcr, & Tobin, 19S3). ~R = Reccprive languie age Js measured by the SICD. ' Mental Developmental
Index 5core (MDI) as measured by rhe Ba)'ley Salles of /11/nnt De,,e/o/Jment-2 (flayley, 1993). Full Scale IQ sco re as measured hy the Wechsler l'reschoul
,111dl'nmary Scale of /11tel/1gence-Revised(Wechsler, 1989).
42 Topicsin EarlyChildhood Special Education 22:1

a ma ster 's degree in hum a n development and had com- th e chi ldren's lang ua ge were determined from transc ri pts
pleted 30 hours toward her docto rate in child psychol- of ch ildre n's utterances during each obse rva tion. M LU
ogy. She had 7 yea rs' ex peri ence working directly wirh and diversity w ere calcula ted using the Systemat ic Anal-
childr en with autism and l year of ex perience implem ent - ysis of Language Transcript p rogram (SALT; Miller and
ing naturali stic la nguage teac hin g pr oce dur es with p re- Chapman, 1984).
school children who had disa biliti es.
Child Language Development. Chi ld ren's language
development was assessed before baseline bega n, at the
Settings end of training, and at the end of the follow-up period.
All ch ild training sessions (base line, intervention , follow Expressive an d recep tive co mmuni cation skills were as-
up ) rook pl ace in a small p layroom at a uni versity- base d sessed using the SICD , the Peabo dy Picture Vocabulary
clini c setti ng . Th e roo m wa s eq uipped with two adult Test-Revised (PPVT-R; Dunn & Dunn, 198 1) and Ex-
chair s, a small tab le with child -size cha irs, tw o ca bin ets pressive O ne-Word Picture Vocabulary Test-Revised
for storage of toys and play mat er ials, and a video ca m- (EOWPVT-R; Gardn er, 1990) . M easures of linguistic com-
era. A variety of age-a ppropriat e toys of inte rest to th e plexity (e.g., MLU) and diversity o f vocab ulary were ob-
childr en were pro vided (e.g., blocks, ma rkers and pa per, tained from verbatim transcripts of the language sample
car s and truck s, balls a nd chut es, bubbl es). sessions. Th e SALT program was used for the ana lysis of
Prob es to assess th e genera liza tion of childr en's lan- trans cript s a nd com put ation of MLU an d diversity .
guage skills with their parent s were conducted in th e
famil ies' homes . Th e sess io ns were vide otape d in a roo m Parent Satisfaction. Pa rents were aske d ro co mplete
of the parents ' choi ce, typically in eith er the family living sa tisfa ctio n questionnaires at the end of the int erve nt ion
roo m or the chi ld's bed ro om , using the child 's coys. and again at the end of the follow- up period. The questio n-
Parent s were ask ed to sray in th e sa me room w ith th eir naire consisted of five questions answe red using a 5- po int
child durin g th e entire sess ion, with th e television turned scale an d six open -end ed questions. The questionnaire
off. Th e video ta ped sa mple of par ent-child interacti on meas ur ed parents' sat isfactio n wit h th eir child's partici -
was 15 minute s in length. pa tion in the inte rventi o n pr ogram and with the cha nges
tbey observed in th eir child 's commu nicat ion ski lls as a
result of th e intervention.
Measurement
Fo ur classes of variables wer e assessed in thi s stu dy:
lnterobserverAgreement
(a) trainer use o f the interventi on str ategies, (b) child social
co mmuni ca tion durin g o bservati ons, (c) d evelop mental lnt ero bse rver ag reement o n the Combined Milieu/
meas ures of child lan guage, a nd (d ) par ent sat isfaction Responsive Interaction Code was calc ulated for each ad ult
w ith the inter venti on program. and child behav ior. Researc h ass istants were trai ned to a
criterio n of at leas t 80% interobserver ag reemen t on the
Tra iner Implementation Measur es. Measures of im- d ata co llection syste m prior co th e co din g of data in the
plement ation by the int ervent io nists reflect th e int egrit y stud y. To deter mine int erobse rver agreemen t, a primar y
of the int ervention deli vered to th e childr en in this st ud y. observer and secon dar y o bserver independenrl y coded th e
The se measur es included the key co mp o nents o f the in- videotapes. A co mpari son of the co ded data sheets was
terv enti o n, including co rr ect use of a ll milieu pro ced ur es, made on a point -by-point bas is. An agree ment wa s score d
talk at th e child 's tar get level, expansions, balan ce of for a behavior that both o bservers reco rded, and a dis-
rums, respon sive feedback, pause error s, and not follow- ag reement was scored w hen onl y one observer scored a
ing the child 's lea d. Th e tr ainer meas ur es were coded beha vior o r when the two obse rvers differed on th ei r cat-
fro m video tapes durin g clinic sessio ns using the Com - ego riza tion of a behavio r. When th e ra tings of the two
bined Milieu/R esponsive interaction Code (Alpert, Keefer, o bservers did not m atc h, a d isag reeme nt was record ed.
& Fischer, 1992). Th e Co mbin ed Code is a co ntinu ous lnt ero bserver agreement was calc ulated using an exact
observ ation system in whi ch each adult an d child beh av- agreement pr ocedur e in which th e rota I number of ag ree-
ior is co ded while watching th e videotape of th e sessio n. ment s w as di.vided by the to ta l numb er of agreements
plus disag reement s an d mul tipl ied by 100. lnt erobse rver
Child Social Communication Skills. Child soc ia l agreement was assessed sepa ra tely for eac h ca tegor y of
co mmunic ati on skills co nsisted of th e freq uency of total par ent and chi ld behaviors. Th e ve rb at im transcriptions
child utteran ces, spo ntan eo us child utt era nces, total use of child utt eran ces were verified by a seco nd co der using
of targe ts (p rompt ed plus spo ntane ous), a nd frequen cy o f the videotap es of th e sa me language sam ple or observa -
tar gets used spo ntaneou sly. ln additi on, the diversity (nwn - tional session prio r to ana lysis u sing SALT. A license d
ber of different word roo ts) and the co mplexity (ML U) of psyc ho logist verified all test ing data and scoring .
Trainer EMT 43

Exper imental Procedures ist but generally waited in a comfortab le sirring area un-
til the sessio n was comp lete.
Dara for this study we re selected from a longitudinal
study compa ring th e effects of three different models of
Interve ntion. Following the comple tio n of base line,
natural istic co mm un icatio n trainin g wirh ch ildr en who
intervention sessions were conducted twi ce eac h week
have significant language delays (Kaiser, Lambert , H an-
with each child in the clin ic playroom. Duri ng the sessions,
cock, & H ester, 1998) . In the la rger study, 76 childr en
th e interventionist and child played with age-a pp ropriate
were assigned randomly to one of three treatment gro ups
toys at a sma ll table o r on the ca rp eted floor . The proce-
(parent- impl emented EMT, n = 20; trainer-implemented
dures used in intervention are summa rized in Table 2.
EMT, 11 = 20; parent-implemented Respons ive Inter -
Based on previous implementations of these procedures,
ac tion, n = 20; o r a co ntr ol grou p, n = J 6). The data pre-
criterion levels for each measure were estab lished (see
sented here are for al l of the childr en with a utism (n = 4)
Table 3). Data were monitored week ly ro ensure thera-
who were randomly assigned to the trainer-implemented
pists reached cri terion levels and co assess chi ld progress .
EMT condi tion.
Tra ining sessions were videotaped and lasted appro xi-
In eac h of the tr eatme nt cond itions, baseline, int er-
mately 15 minut es. Ten minutes of data on trainer im-
vent ion, and follow up, phases were implemented in the
plementation of the intervent ion procedures and child
context of a sing le-subject, multiple baseline design ac-
commun ication measures were code d from the video-
ross ch ildr en (Tawney & Gast, 1984). Particip ant s were
tapes of the sessions a lways startin g wit h the first full
assigned randomly and equ ally ro a specific num ber of
minute. Parents did not observe their children during the
baselines (five, six, or seven baseline observations). Ad-
traini ng sess ion, but parents were given a gene ra l report
ditional baseline points were co llected if any measu re of
of their ch ild's participation in rhe intervention sess ion s.
adult or child behavior was not stab le at the end of th e
assigned number of baseline sessions. The multipl e base-
Fo llow up. Par en ts and childr en retu rn ed to th e
line design was replicated six times across th e 20 fam ilies
clinic once each month for 6 mont hs after the end of th e
ass igned to the traine r-imp lement ed EMT con dit ion. Th e
inte rventi on ro assess maintenance of chi ld use of socia l
four child ren with autis m ass igne d to thi s con djtion had
communi catio n skills. These follow-up sessions were iden-
baselines consist ing of five (Child A), six (Child B), and
tical to the intervention sessions with interventionjsts us-
seven (Chi ld C and C hild D ) sess ions. The grap hic darn
ing a ll o f th e intervention pro cedures w ith th e cruld in the
in chis study were analyzed using simpl e visua l inspect ion
session . Intervention ists and children p layed together in a
(Tawney & Gast, 1984).
15- minut e sess ion cha r was videotape d with 10 minut es
of these sessions coded. Parents did not observe their
Prebaseline. All childr en we re preresred on a battery ch ildr en during the sessions, bur a general report of the
of standard ized language rests at th e beginning of th e child' s part icipation in the session was given to the par-
study: the SlCD, rhe PPVT-R, and the EOWPVT-R. Two ent after every fo llow-up session.
30-rninute language sa mpl es were co llected during play
interactions between the childr en and a train ed research Generalizat io n. Ni ne obse rvatio ns we re co ndu cte d
ass istant in a clinic playroom . Based on standard ized as- in the home ro assess gen eral iza tjon of the childr en 's use
sessments and rhe language sa mpl es indj vidu al lang uage of socia l com munic at ion with their parents. Three ses-
targets for eac h child were selected (see Tab le 1). Targets sions were con du cted at the end of base line, three at the
were broad classes of early sema ntic relationship s repr e- en d of the intervention, a nd thr ee at the end of the 6-
sented in two-, three-, an d four-word utter ances (e.g., month fo llow-up period. Parenrs we re instructed to play
agent -action, agent -ac tio n-o bject, agent-act ion- loca tion). an d int eract with t beir chi ld as they usually did, using
Eac h chi ld haJ rhree or fo ur target classes, and in this toys and materials from their home. Parents and children
condition the parents were no r inform ed of their chil- were videotaped for 15 minutes eac h time with 10 min-
dren's language targets until rhe end of the study. utes of the sess ion tr anscribe d a nd coded.

Baseline. Int ervent io nists and childr en were video- Posttest ing. Fo llow ing the last inre rvenrion sessio n
taped play ing together for l5 minut es durin g each base- (average lengt h o f time between Pre and Post 1 was
Une sess ion. From these videota ped sessions, LOminut es 5.2 month s) and aga in followi ng tbe las t fo llow- up ses-
of data were coded . Parents were given genera l informa- sion (average lengt h of tim e bet ween Pre and Post 2 was
tion about their child 's participation in each sessio n 11.5 months), assessme nt s identi ca l co the prebaseline
("Yo ur ch ild rea lly enjoyed playing with th e bubbl es to- pro ced ur es were used ro determine cha nges in the child 's
day"), and the int ervent ionist would answer any ques- performance on sta ndardiz ed la nguage tests a11d in two
tio ns or concerns posed by th e parents. Parents did nor 30- minute language samp les. At both postintervention and
o bserve their child in the sess ions with the inter venti o n- post-fo llow-u p resting sessions, parents com plete d the
44 Topicsin EarlyChildhood Special Education 22:1

pa rent satisfaction sur vey to assess their satisfaction with lnterob se rver agreement for total use of targets was
the interventi on program and th eir children's progress. 87.4% (range 50%-100%) and for all child behaviors
was 94.5% (rang e 86%-100%). Low interobserver
agreement percentage s occurred only when the rate of
RESULTS behavior was very low.

lnterobseNer Agreement
Resultsin the TrainingSetting
lnt ero bserver agreement on co ding using the Combined
Code was assessed in approximately 20% of the sess ions Trainer Implementation. Seven aspects of trainer im -
for a total of LOtime s for eac h family dming the course plem entation of the int erven tion were mea sured : (a) fre-
of th e st ud y. One as sessme nt occurred for eac h fam ily quency an d perc entage of correc t use of rhe fo ur milieu
during ba seline, five during training, one assessment dur- teaching procedure s, (b) frequency of expansions and
ing the follow up, and three during the gene ralization percentage of child utterances expan ded, (c) frequency of
assessments at home. A second coder verified each tran - talk at the chi ld's target leve l, (d) balance of ad ult and
script of th e language sample and of child utteran ces dur - child turns (turn discrepancy), (e) frequency and percent-
ing the adu lt-chi ld interaction sessions and all rest data age of responsive feedb ack, (f) pause er ror s, an d (g) not
were checked for scoring accuracy by a licensed ps ychol - fo llowing th e chi ld's lead. Table 2 details the specific
ogist. co mponents of the intervention and operating definiti ons
Mean int erobserver agreement for all trainer behav - for eac h proc edur e. Criteria levels were set for all inter-
ior s was 89.5% (range 81 %- 100%) for milieu teaching vention procedures ba sed on data from past studies of
episodes, 95.1 % (range 74%- 100%) for expa nsions, intervention efficacy. Table 3 gives the means for all im-
95.2% (range 83%- 100%) for talk at the target level, plementation measur es as well as the spec ific cr iteri on
and 85.1% (range 70%-92%) for all other adult be hav- for each procedure . Although there was some va riab ility
iors. The mean interobserv er agreement for ch ild co m - among the trainer s on their use of the proce dure s, train-
munication behavior was 84.3% (range 68%-98%). ers delivered th e intervention with a high leve l of preci-

TABLE
2. EMTComponents, Measures, and Definitions

EMTcomponent Measure Definition

Milieu reaching procedures Number and percentage of correct Correct milieu reaching episodes: (1) begin with child
use of four procedures verbal or nonverbal requests; (2) follow a specific se-
quence of prompts (models, mands, or rime delay proce-
dures); (3) include corrective prompts as needed, and
(4) end with posit ive feedback, expansion of the child's
utteranc e, and access to the child's requested object
Expansions Percentage of child utteran ces Adult utterance that follows a child utterance, embeds
that were expa nded child's utterance inro a more comp lete form, and main-
tains or extends the child's meaning
Balance of adult and Turn ratio Turn ratio was ca lculated by dividing adult rum s by
ch.ild turns child turns, with a numb er larger than I indicating
more adult turns than child rums and a number less tha n
I indicating more child turn s than adult turns
Not following the Number nor following Adult verbal behavior that does not relate directly to the
child's lead child's lead child's topic or ongoing behavior (e.g., recruiting child's
attent ion t0 another coy, initiating talk about events out-
side the play comext)
Pause errors Number of pause errors Three or more consecutive adult utterances without at
least a 3-second pause for a child response
Responsive feedback Percentage of child utterances Adult verbalizations that follow a child urrerance, are
followed by responsive feed back meaningful ro child, and directly related to the child's
copic
Talk at the child's Number of adu lt utterances at Adult utterances rhar included examples from ch ildren's
target level rarget level target classes and were slightly longer (2-3 words)
rhan child's current productive language
TABLE3. Trainer Implementation of Intervention Procedures During Clinic Sessions

Child A Child B ChildC Child D All trainers

Condition Criteria BSL INT FU BSL INT FU BSL INT FU BSL INT FU BSL INT FU
# Co rrec t milieu 5- LO 0 15.4 12.3 .2 5.0 3.0 0 2.8 5.5 .I 9.6 10.3 .08 8.2 7.8
% Co rrect milieu > 80% 0 95.6 100.0 1.8 100.0 94.5 0 LOO. 98.5 J.8 98.2 100.0 1.0 98.4 98.2
% Expa nsions > 50% 2.6 91.2 90.8 4.8 87.2 72.5 0 80.2 84.3 4.4 86.4 89.8 2.9 86.2 84.4
Turn taki ng ratio J.00 2.30 .99 .88 2.30 .99 .88 4.90 .96 61.7 4.1 67.2 61.0 6.8 69.5 65.8
(ad u lt : ch ild )

% Nor following 0 14.6 0 0 40.8 0 0 16.3 0 1.05 1.60 1.03 1.09 2.16 1.01 1.00
lead

# Pa use erro rs <2 18.2 0 0 23.8 .2 0 26. 1 .4 99.3 43.1 97.6 99.3 34.9 98.1 98.5 1~
% Responsive > 80% 96.5 0 26.5 97.2 96.5 2 1.1 99 .3 15.1 0 0 20.5 .I .8
feed back

# Ad ulr rargers > 40 3.6 93.6 91.2 3.7 47.8 49.3 7.7 69.6 0 25.0 0 0 24.3 0 0

Note. Parenr rums minus child rurn<: If parenr had more rurns rha n ch ild, number is positive; if parenr had fewer rurns rhan chi ld, number is negarive. BSL = data arc the mean of three sessions at the
end of the baseline; INT= data are rhc mean of rhree sessions at the end of inrervencion; FU= dara are the mean of three sessions ar t he end of follow up.
46 Topicsin Early Childhood Special Education 22:1

sion and were well above the established criteria level for four childr en during the inte rvention. Children B and C
all procedures. Figures I and 2 show session-by-sess ion showed modest increases in MLU from intervention to
data for the baseline, inte r vention, and follow-up co ndi - follow up. Three of the four children (A, B, C) increased
tions for use of milieu teaching and expansions . The in- slightly in their diversity ar the follow up.
terventionists depressed their use of the procedures during
the basel ine phase to resemble levels observed in baseline Generalization Settings
sessions with par ents in the othe r two parent-implemented
intervention groups in the larger study (Kaiser et al., Data showing the mean use of targets, spontaneous ut-
1998) . Parents may demonstrate some of rhe behaviors terances, MLU, and diversity for the four children during
being measuretl tluring baseline but not at the optimum baseline, treatment, and the follow -up phase in genera l-
level needed to affec t chi ld language . ization sessio ns at borne with their mothers are shown in
fn an average intervention session, interventionists Table 5 . Generalization to home setting was observed for
engaged the children in eight mil ieu episodes to prompt three of the four children, with the greater changes occur-
target lan gua ge, responded to 98% of rhe children 's ut- ring immediate ly afte r the intervention than at the 6-monrh
terances with responsive-type feedback, and expanded fo llow up. Two of t he four child ren (B and CJshowed
86% of the children's utterances. T bey a lso used th e chil- large increases on their rota ! use of ta rgets and sponta-
dren's targeted lan guage forms an average of 69 times in neous use of targets after intervention. The other rwo
a session . Turn taking between the trainer and child was chi ldr en (A and D) showed modest increases in use of
exp ressed as a ratio of adult to chi ld turns with a num- targets ana lyzed at the 6-monrh follow up. Three of the
ber larger than one reflecting more adu lt than chi ld turns fou r childr en (8, C, D) increased the frequency of their
and a number less than one reflecting more child th an total utt erances and their spontaneous utterances after
adult turns. Int ervention sessions revealed approximate ly th e intervention. Children B and D had additiona l in-
equa l turn taking between the trai ners and chi ldr en (M = creases in total and spo ntan eous utterances at the follow-
1.01 ). During mo nthly follow- up sessions in the clinic, up assessment.
trainers close ly approximated the rares of intensity and Childr en also showed generalized changes in two
precision of implementation delivered dur ing the inter- genera l measures of language complexity: MLU and num-
vent io n sessions . ber of different word roots (diversity). MLU increased
for rwo chi ldr en (B and D) during the postintervention
Ch ild Socia l Communica tion. After the intervention home gene ral ization assessme nts, and modest increases
was introduced, a ll four ch ildr en showed increases in in th is measure were seen at the fo llow-up home gener-
their roral use of targets (prompted plus unprompt ed), as alization assessments for the other rwo ch ild ren (A and
shown in Figure 3. Th e strongest effects were seen for C) . A similar pattern of effects was observe d for diversity
Childr en B and C. Childr en A and D had low er levels of of word roots . Increases in diversity of word roots were
target use throughout the int ervention. Child A increased observed during home general ization observations at the
his target use in the last half of the intervention cond i- end of t he intervention phase for all children, with three
tion. All four children continued to use their targets dur- of rhe four childr en (B, C, D) approximate ly doubling
ing th e clinic follow-up assessments. Three of the fo ur their diversity from pre- ro posrintervenrion assessments .
children (A, B, C) showed clear changes in spontaneous Chi ldr en maintained t heir higher levels of diversity in the
use of targets (see Figure 4). Child D showed small changes fo llow-up observa tions with their mothers.
in use of spontaneo us targets during the intervention. All
four childr en maintained simil ar rares of sponta neous Language Development Measures
target use durin g the follow-up period.
Cha nges in ch ildr en's comm uni cation are summa- Six measures of language development were taken at the
rized in Tab le 4 . All child ren increased in t heir frequency prebaseline, postintervention, and post-follow-up periods.
of total utt erances from base line to intervention, with The resu lts of these measurements are shown in Table 6.
Child C showi ng the lar gest gains. Two childr en (A and Alth ough th ere is evidence indicating developmenta l
B) showed further increases in total utterances at fo llow cha nge concurrent with the intervention period, this evi-
up. Two children (B and C) increased in th eir sponta- dence is variab le across rhe four chi ldren . Child B showed
neous utterances from baseline to intervention, with changes on all measures ar the postintervention assess-
Chi ld C making the lar gest gains. Only Chi ld B showed ment , whereas Child A cha nged positively on five of the six
an increase in frequency of spontaneous utterances from measures. Chi ld D changed on two measures. The post-
intervention to fo llow up; the ot her three childr en main - follow-up assessmem revealed a sim ilar positive develop-
tained inrervention levels at follow up. Changes in MLU mental change for the chi ldren, altho ugh not quire as
occur red for three cbild ren (A, B, C), and cha nges in di- strong as was meas ur ed pre- to postintervention. Ch ild B
versity (numb er of different words) were observed for all showed changes on all measures. Two children (C and D)
Trainer EMT 47

Baseline Intervention Follow-up

40 r-- i-- - - 100
A - 80
30

40 100

30
B 80
,__
60

40

20

40 - 100

80
~
.._30
G) C 60 Cl)
C
.020 (J
E 40 ..
Cl)
~10 20 Q.

t- ,. 100
,__

so D -ao
80
2SI

1 2 3 4 I I 7 1 2 3 4 5 I 7 8 9 10 11 12 13 14 15 11 17 18 111 20 21 22 23 24 1 2 3 4 5 8
Sessions
Total Number Percent
Number Correct Correct
-+- -e- D
FIGURE 1. Frequency and percen t age of correct use of the four milieu teaching procedures by interventionists in
the clinic setting .

changed on four of the six measures a nd mainta ined the ot her measures . Although rhe overa U pattern is one of pos-
change made fro m Pre co Pos t 1 on the ot her two mea- itive changes in deve lopmen tal measures associated wi th
sures. Child A made changes on rwo measures and main - the interventio n, cha nge on developme ntal measures is
tained the change from pre- to post inte rvention on two difficult co inte rpret in the abse nce of a con trol group.
 48

Baseline Intervention Follow-up

• 1•

71

•
-
•
.. ..
-
11
..

• 1111

71 ..

• 1111

,.
-
- -..
.. ..
11

1 2 3 4 5 6 7 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 1 2 3 4 5 6
Sessions
Number Number % Opportunities
Opportunities Expansions Expanded
-+- :: CJ

FIGURE2. Frequency of expansions and percentage of chi ld utterances that were expanded by interventionists in
the clinic setting.
Trainer EMT 49

Parent Satisfaction que st ions. The intervention ha s co nverted 'babbling' into

real speec h. " An o ther parent report ed that the interven-
Parent raring s of th eir satisfaction w ith rhe training were
tion had helped her child "go from one- and nvo-wo rd ut-
very po sit ive. Using a 5-po int scale, the mean rating for
terances to ha ving meaningful co nversat ion includin g up
five scaled item s for the four parents wa s 4.88. All par - to nine -word sentences," w ith an added benefit that her
ents rated their satis faction with improvement in their
ch ild now "ge ts frustrat ed less frequ ently " a nd is " more
child's co mmunication sk ills an d their overal l satisfac -
enga ged in th e environment and the peo ple in it." The
tio n with the interv enti on as 5 .0. Th e on ly qu estion that
ma in suggestio n offe red fo r improvin g the training was
par en ts rat ed with less than a 5-poinr response conce rn ed
havi ng th e training last for more sess ions and fo r longe r
the rim e requir ement involved in th e traini ng wi th par- than 15 to 20 minute s a sess ion. Beca use of the design of
ent s wanting mor e sessio ns for longer t ime periods.
chis study, we could not train the parents in rhe crain er-
Par ent s also gave very po sitive response s ro th e six ope n-
implemented co nditi on until the lase fo llow-up sess ion
ended questions. All four parent s d escribed specific, pos- w as co mpleted. One parent sa id, ''The wa it to be trai ned
itive changes in their childr en's language . One par ent as a parent " wa s w hat she liked the least abo ut her
sa id, "Now my chi ld speak s in sentences and can a nswe r
child 's tr a ining, "alt hough I rea lize it had ro be that way
due to resea rch. J can 't wait to learn ' the sec ret. " '
TABLE4 . Child Language in Clinic Sessions

Condition
DISCUSSION
Child Measure BSL INT FU
A Total utt erances 48.4 68 .0 82.3 Th ese resuJts ex tend the app licat io n of hybrid natural istic
Spontaneo us utt era nces 4 l.2 46.2 47.5 lang uage int ervention proc edur es ro ch ildren wit h
MLU 1.32 1.65 1.70 a uti sm and offer fur ther supp or t for the effectiveness o f
Diversiry 24.4 35.6 46.8 na tur alistic langu age intervention s imp lemented with
B Total utterances 44.0 52.2 69.7 this populat ion of chi ldren. In thi s study , chi ldren demon-
Spo nt aneo us urrera nccs .8 3.8 10.2 str ated positive changes in socia l co mmun icat ion as-
MLU 1.94 1.94 3.62 sesse d across se tti ngs and mea sur es, and these chan ges
Diver sity 50.4 3 l.4 77.5
genera lized to int eractions with their parents. Th ese find-
C Total utterances 16.8 68.8 58 .8 ings a re especially imp ortant give n the socia l and com-
Spontaneo us utteran ces 9.2 45.6 43.5 muni carive difficulties o f mo st children w ith autis m and
MLU l.54 2.29 2.29
Diversiry 11.0 38 .2 44.2 the likelih oo d tbar the y will need cont inued support to
acq uire an d use functional communication skills.
D Tota l utt era nces 35 .2 50.6 46.2 Of parti cu lar interest in chis stud y is that when a hy-
Spo ntane o us utterances 30.4 27.6 3 1.0
MLU J. 65 1.56 2.06 brid natural istic la nguage intervention is del ivered in a
Diversity 20 .8 30 .6 29 .7 precise, high -qu a lit y ma nn er, th e actual time it rak es to
effec t chan ges in chi ldr en's language ma y nor be as long
Note. BSL = data arc the mean of rhrce session s at the end of baseline;
as tradit iona lly th o ught. In this study, children we re en-
INT= darn Me che mean of rhrce sessions ar the end of int crvenrion; FU =
data arc the mean of thrcl" sessions at the end of follow up; ML U = mean gaged in int ervent ion for 24 15 -minut e sessions for a to-
length of urrcrance; Divcrsiry = mean n umber of diffcrenr word roots in tal of 6 ho urs of dir ect intervention. In this particu lar
each session.
intervention , the emphasis on respons iven ess shifted the

TABLE5. Child Language in Home Generalization Sessions

Child A Child B ChildC ChildD

Condition BSL INT FU BSL INT FU BSL INT FU BSL INT FU
Spontaneous targets .7 .7 5.3 2.7 25.0 17.3 0 10.0 9.7 .7 1.0 2.0
Toral tar gets .7 .7 5.3 3.7 27.0 19.0 .3 19.0 14.0 .7 1.0 l. 7
Total utteranc es 34.7 32 .3 37.3 64.0 94.3 107.0 6 1.0 105.3 85.3 33.0 62 .0 75.3
Spontaneou s utterance s 26.3 26.0 29.3 44.7 81.0 98.3 30.3 58.7 45.3 22.7 32.7 58.0
MLU 1.34 J.34 I..7 1 L.94 3.54 3.69 1.33 1.56 J.82 1.49 2. 15 2.13
Diversiry 159 .0 35.0 50.0 52.0 98.3 99.7 26.J 56.3 64.7 21.3 46.0 54.3
Note. BSL = data arc the mean of rhrce sessions a 1 the end of base line; INT= dara are rhc mean of three sessions ar rhc end of imcrvcnt.ion; FU = dara are
rhe mean of rhree ~essions at the end of follow up; MLU = mean length of urrerancc; Diversiry = mean munbcr of differe nt word roors in each session .
 50

Baseline Intervention Follow-up

100

ao
A
'°
~

20

100

80

60

"°
20

l.-100
CD
.0
IO
C
E
z:, so

"°
20
9
··..ii;/
0

100

80 D

1 2 3 4 5 8 7 1 2 3 ♦ 5 8 7 8 9 10 11 12 13 14 15 16 17 18 18 20 21 22 23 24 1 2 3 4 5 6
Sessions
Adult Child
Targets Targets
--- --9-
FIGURE3. Frequency of adult and child communication total target use in the clinic setting.
 51

Baseline Intervention Follow-up

- --
A

40 B
-- 3

30

20

10
- -- 1

'-
G) 40
.0 .... - -- - - 4

E C
~ 30 3
z
20
- i 2

10 1

0 ---- 1-

40 - - -
D
30
- -- - 2
20 -

10 -
- - ~ "" - ,--- ~ - 1

.... - ...-__,-~ ..~~.....1-.

0 i__._-._.Ap-y-~-
, ..... -~•...Jt-e..~~•~•~L..•~-~~-
-..l4-~-~-L..-..i,.1 r r 1 1..,.
,._..b"'"'~-,..__w...J~~a.",..
, 0
1 2 3 4 5 6 7 1 2 . 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 1 2 3 4 5 6
Sessions

Number Number
of Targets Different Targets
--- D
FIGURE4. Frequency of child communication, spontaneous target use, and number of differen t classes of ta rgets
used In the clinic settings.
52 Topicsin Early Childhood Special Education 22:1

TABLE6. Child Performance on Developme ntal Language Measures at Preintervention, Postintervention, and
Follow Up
SICD-R SICD-E EOWPVT-R PPVT-R MLU Diversity
Child Pre Post FU Pre Post FU Pre Post FU Pre Post FU Pre Post FU Pre Post FU

A 20 28 28 20 24 24 Nh Nh 29 Nb 24 23 1.58 1.76 2.0 l 42 109 93

B 28 36 40 24 36 40 36 54 62 13 40 44 1.79 3.20 4.02 I 04 l43 173

C 24 28 28 20 36 36 Nh 31 34 23 26 30 1.33 1.79 2.08 30 64 77

D 28 28 28 20 20 32 Nh Nh 30 Nb 23 26 1.64 1.61 2.37 67 108 94

Nole. SICl>-R = Seq11c11c!'d/1wentory of Co111111w11 clltiw1 Dc1•do/1111e111

(I lcdnck, Prather. & Tobin, I 981) rcccpnvc co111111umc,1t1on
score in rnonrh,;
SICD E = cxprcs;il'e cornrnuni<aling ,.:ore in month,; f'OWl'VT -R = 1:x/,resswe 011e-Word 1'1c/1tr,
• Vocalml,1rvTe.st-Re,,,sed (Gardner , 19'10) score in
• Test-Re1•1sed(Ounn & Dunn. 19~ I ) score in month~; Ml.U = mc,rn leng1h of utrcrnncc from rwo rn-
monlhs; l'l'Vl-R = Peabody f'1.c1t1reVoc.1/111/,1,;
111inu1c
hrnguagc samples; Divcrs,rv = mean tl,versit) of vocabul:iry from two JO minurc h111gWt/!C samples; I U = follow up: Nh = no h:1salt•s1,1hli,hed.

focus from proactive reaching to responsive teaching and creases in acqu1s1non, maintenance, and generalizatio n
embedded prompting into conversations and engaged of language skills. These results are similar to those with
play interaction:,. When children are engaged with adults parents who we re trained to use EMT when interacting
in activities of interest to them, there are many opporru- with thei r children with autism (Kaiser et al., 2000). It is
nitie s to teach functional skil ls without disrupting rhe possible that the children who were le:,s dela)'ed in their
Flow of social interaction. It is possible that for children language abi lities relative to their age required less adult
with autism who may resist a didactic approach and even support both from the interventionist and parent to
direct instruction:,, the use of a responsive engagement make positive gains in their language abi lities. The chi ld
strategy may be especially powerful. who did nor make the same gains as the other three chi l-
Although results of this study indicate this interven - dren may have needed more than 24 sessions or sessions
tion was effective, it is important co remember that the longer than 15 minutes tO promote positive changes in
interventionists who worked with these childr en were his lan guage. The limited number of children in this
highly educated, had years of experience working with study does not allow a formal analysis of child charac-
children before this study, and received intensive training reristil:s that might predict child response to treatment
before and weekly supervision during the intervention protocol. Future stud ies cou ld explore the intensit y and
period. Also, rhe children had to meet minimum criteria dosage of language interventions needed to effect posi-
robe included in the study, specifically the ability to im- tive changes g iven the different profiles of abi lities dem-
itate and produce lan guage of at least JO words. All chi l- o nstrated by chi ldren.
dren were selected to be in rhe range of language abilities Ir is important ro note that although there was some
that previously has predicted strong positive outcomes in variability in language performance among the chi ldr en,
a milieu reaching intervention (Yoder, Kaiser, Alpert, & there was al~o variability across observations for each
Fischer, 1993), so we expected rhar all children wou ld chi ld. For examp le, in baseline sessions there was a fairly
benefit from rhe intervention. Results with childre n in large range in MLU across these observations for all four
sim ilar interventions could definitely vary depending on childr en (A, range= 1.13-1.64; B, range= 1.39-2.62; C,
the incervenrionisr's training and abilities an d the march range= l.0 0- 1.86; D, range= 1.23-2 .28). Professionals
between intervention content and chi ldr en's entry ski lls. need to be aware that inconsistency in behavior is com-
The development of intervention and measurement pro- mon in children with aut ism (Lord, Bri:.tol, & Schopler,
tocols ro fir individual chi ldren poses a considerab le 1993), so individualized programming for young chil -
challenge for both researche rs and practitioner:, bur is dren with autism should always include multiple assess-
needed to more accurately chara<.:rerize the effects and ef- ments and multip le observations to produce the clearest
fectiveness of naturalistic language intervention. picture of a child's abilities. Inconsistent responding by
As is the case in many studies of language intervention children with autism also .1ffects the results of the stan -
with children with a uti sm, the effects of the inrervenrion dardized test and dcvelopmenral language measures re-
were va riab le across individual childr en. C hildr en in the ported in this st udy. Because rhere is no way to know in
current study va ried nccordi11g to age, tested LQ, and lan- the cllrrem design if deve lopmenr may have contributed
guage ability at rhe beginning of the inrervenrion. In gen- to the changes in child MLU, diversity, and performance
era l, younger child ren whose language development was on standardized tests, rhese results must be interpreted
less delayed relative ro their age showed rhe greate!>t in- cautious ly. Replication with a randomly assigned control
 Trainer EMT 53

group of children wi rh a uti sm would be req uir ed ro con- AUTHORS

' NOTES
firm the developmenral findin gs reported here.
I. Thi~ re\earch wa~ ,upportcd in p,m hy Grams HD 15051 and
Ch ildr en in this stu d y general ized pos itive chan ges HD27563 from NICHHD.
in their lan guage s kills to int eractio ns with tbeir un- 2. We thank the par ent~ and children who parcicipateJ 111 this study.
trained mothers immediately a fter the inter ve ntion. Al- We al,o gratefully acknowledge the technical support for the con-
th oug h these changes were main tained a t the end of th e duL, of the midy :rnd ,inalysi\ of data provided h) rhe Milieu
Teaching Group at Vanderbilt Univer,ity, esptciallr Maureen
follow-up period, on ly one child showe d evidence o f Keefer.
co ntinu ed acce lerated change in co mmuni cation at the
follow up. If parents had been trained ro use simpl e re-
sponsive interaction <;trat eg ies with th eir childr en at REFERENCES
home during the fo llow-up period, they might have been
ab le to conti nu e co suppo rt their childr en's cont inu ed Alpert. C. I ., "-eefer, .\1. A.. & fohcr. R. (1992). Cr,111bi11ed wde:
Milieu, respll11siue-111teract111e and hy/1nd lc111 gut1ge teacbmg.
language growth. Nashville, TN: Peabody College nf Vanderbilt Universil}-
Th e design used in th e curr ent st ud y has limi tatio ns. Andcrson, S. R., Avery, D. I.., DiPi..rro, E. K.• Edwards, G. L., &
In order to implement th e singl e subject design wi rhin the Chrisrian, W. P. ( 1987). Intensive home-based early intervcnnon
randomly assigned treatment groups, we needed to re- wi1h ,1uti~11 c children. Ed11c<1/11111
a11dTreatment of Cluldren. 10(4),
JH-366.
strict t he numb er of base line sess ions ro cont rol for pre -
Bayle}, N. ( 1993). B,1yleyscales of i11/t111/deitelvpment {lnd ed.). San
treatment expe rience in th e clinic across con diti ons . We Antonio, TX: Psychological Corp.
selec ted five to seven baselines based o n ou r previous Dunn, L. M., & Dunn, I.. M. (198 1). Pe.ibody picture 1mc,1b11/ary
single subject implementations of EMT indicating that frst-Rc1 1ised. Circle Pines, MN: American Guidance 'ien ice.
chi ld behavior is typica lly stabl e after five sess io ns. We Durand, V. M., & Carr, E. G. ( 1992). An analysis of maintenance fol-
lowing funcrional communication training. j1111ma/of Applied
ini tiated treatment after the prescribed number of ses-
Be/1<11111JrAnalysis, 25(4), 777-794.
sions ; however, in eve ry case, a clea r shi ft in the incer- F.dwards, D. l"., & Bristol, M. M. ( 199 1). Autism: Early identification
ventionist's implementation of the intervention occ urr ed and management in family prncnce. Amer,ca11 family Physicia11.
immediately (see Figures I & 2), a nd modest change s in 4-1(5), 1755- 176.,.
child 's rotal use of targets (Figure 3) were immediate ly G.irdner, M. F. { 1990). Express111eone-word pict11reuoca/mlary test-
Rm11sed.Novaro, CA: Ai.:adcmii.:Therapy.
obse rved . The use of combi ned si ngle subj ect and group
Carmer, D •• & !:ii.: hultz, N. 1\1. ( 1990). Esrahlishing the firs1 stai;cs of
designs has many advantages bur may also co nstr a in th e early rei.:ipmcalinteracuon~ between mothers and their :iuristic chil-
inrerpretat ion of the si ngle subj ect data. Replications of Jrcn. Wlome11 ,md Therapy, f()_ 159-167.
t he effects of the EMT intervention were seen with other Halle, J. W.. Baer, D. M., & Spradlin. J. F. ( 198 1). Teachers' gcncra l-
ch ildr en assigned to this treatment co nditi on six separate 11,cclU,l' of dday :1~a ,11mulu;,control procedure to increase lan-
guage us~ in handicapped children. Jo11mal o/ Applied Bebat'l<>r
three-tie red multipl e baseline app licat io ns of the treat -
A11alys1s,1--1.]89-409.
ment (Kaiser et a l., 1998). Notab ly, the effects with chil - Harr, It ,\ I.. & R1slei. T. R. ( I 975). lncidemal teaching nf l:inguage in
dren w ith autism were very simil ar to th ose wit h othe r rhc preschool. Jo11mt1/n/ Aflplied Beh,wu,r Analysis. 8. 4 11--420.
chi ldren who had a range of different developmental dis - Hedrick, D. L., Prather, £. ,VI., & Tobin, A. R. ( 1983 ). Seq11e11,ed i11
-
abili ti es a nd nearly identical language sk ills at entry. l'entory of n,1111111micatio11 de11elopme11t.!'>carrlc:University of
Washingron Press.
In the Kaiser et al. (20 00 ) stud y, the children w ith
l lemmcwr. ~I. L.. & Kaiser, A. I'. ( 1994). F.nhanced milieu tcachuig:
autism showed simi lar changes in language use in train- An ana lysis of pMcnt-implerncnred languag<' interven1ion. j o11m<1/
ing sessio ns with their parents and slightl y stronger gen- nf F'..irly /11tcrl'e11t1011
. 18(3), 269-189.
eralization ro the home settin g at th e end of inrervencion Hum, I'., Alwcll, M., & Gocr1.,I . ( 1988). Acqu1\ition of conversarion
:rnd follow up. Child ge nera liza ti on ro th e ho me set tin g skills and the reduction of inappropriate: ~ocial interaction hehav-
10rs.fASH. 1 l( I ), 20-27.
appeared to be relat ed to parents' co ntinu ed use of t he
Kaiser, A. P.• Hancock, T. B., & Nierfeld, J. P. (2000). The effects of
EMT procedures . When parents maint ained t heir use of p~rtnt-irnplcmcnrcd enhanced milieu tca~·hing on rhe socia l com-
the proced ur es, c h_ildr en s howed furth er ga ins in la n- municJtion of children who have autism. Early F.d11cati011 and
guage use. Developmenl, 11(4). 42 3-446.
For child ren with au t ism, it is imp ortant to co nsider Kaiser, A. P., & Hester, P. P. ( 1994 ). The generalized effect;, of enhanced
milieu teaching. Jo11mal of Speec/J and Hearing Research, 17(6),
the range of possib le suppor ts for developing co mmu -
I .HO-I .HO.
nication skills . lmervenrions th at empl oy bo th skilled in - Kaiser, A. I'., Lambcrr, W., Hancock, T. 8., & Hester, P. P. ( 1998 ,
terventionist s (e.g., speech- lang uage clinicians, ea rly March/. Di(/erc11twl011tn1111es of 11'1/ltralisticla11g11agc
inter1•c11-
chi ldh ood specia l education teac hers) a nd parents may tio11.Paper presenreJ at the 3 1st annua l Caclinhurg Conference nn
help chi ldr en acce lerate their language development and Research and Theory in Mental Retardatton and Devclopmenral
DisJhi linc~, Charlc~ton, SC.
use. In vestiga ti ons of multiple partner interventions co
Koegel, R. L..,Koegel, L. K., & Surra tt, A. ( 1992) . Language interven-
support communica tion in every day settings a nd to pro- tion and disruptive behavior in preschool children with aut ism.
mote genera lization and maintenance are needed. ♦ Jo11r11al of Autism ,md Deuclopmemal D1sorders, 22(2), 141- 153.
54 Topicsin Early Childhood Special Education 22:1

Koegel, R. L., O'Del.l, M. C., & Dunlap, G. ( 1988). Producing speech Nonon, P., & Drew, C. { 1994). Autism and potential family stw,sors.
use in nonverbal aurisric children by reinforcing arrempts. journal The American journal of Fa111rlyTbera/J)', 22( I), 67-76.
of Autism and Dei•e/opme111alDisorders. 18, 525-538. Ostrosky, M. M ., & Kaiser, A. P. (199 l ). Preschoo l classroom envi-
Koegel, R. L., O'Dell, ~ I. C., & Koegel, L. K. ( 1987). A natura l lan- ro nments tha t promote commun ication. Teaching Exceptional
guage paradigm for reaching non-verbal au tistic child ren. Joumal Children, 23(4), 6- 10. (Reprinted in Ammal editions: Ed11cating
of A uti sm and Develof1111e 11tal Disorders, 17. 187- 199. excep tional children, 6th ed., pp. 158- 161, hy K. L. 1-reiberg. Ed.,
Land,1, R., Folstein, S. E., & Isaacs, C. ( 199 J ). Spontaneous narrative- 1995, Guilford, CT: Dushkin.)
discourse performance o f parents of autistic ch ildren. Journal nf Scherer, N. J., & Olswang , L. B. (1989). Using structured discou rse as
Speech and /-tearing Research. H, 1339-1345. a language intervention technique with autistic children. Jo umal of
Laski, K. E., Charlop, M. H., & Schreibman, L. ( J 988). Training par- Speech and Hearing Disorders, 'i./, 383-394.
ents to use the natural paradigm ro increase their autistic children's Schreibman, L., Kaneko, W. M .. & Koegel, R. L. (199 J ). Positive af-
speech. ]01mu1I of App lied Behav ior Ana lysis, 21 (4), 39 1-400. fect of parent~ of autistic children: A compa rison across rwo reach-
Lemanek, K. L.. Srone, W. L., & Fischel, P. T. ( 1993). Parent-child in- ing techniques. Behavior Therapy, 21, 479-490.
teractions in handicapped preschoolers: The relatio n herween patenr Tawney, J. W., & Gast, D. L. ( 1984 ). Single subiect research i,1 special
behavio rs and comp liance. journal of Clinical Child Psychology, ed1tcc1tion. Columbus, OH: Merrill.
22( I ), 68- 77. Warren, S. F., McQuarter, R. J., & Rogers-Warren, A. K. (1984). The
Lord, C., Bristol, M. M., & Schopler, F.. (l 993). Early intervention for effecrs of reacher mands and model on the speech of unresponsive
children with autism an d related developmenta l disorders. In E. language-de layed ch ildren. ]011rnal of Speech and Hearing Re-
Scbopler, M. E. VanBourgondien, & M . M. Bristol (Eds.). search, 49, 43-52.
Preschool issues i11autism (pp. 199-221 ). New York: Plenum Press. Wechsler,D. ( 1989). Wechsler preschool and primary scale of intelligc11ce-
Lovaas, O. R. ( I 987). Behavioral treatment and normal cducmio nal Reuised. San Antonio, TX: Psychological Corp .
and intellectual functioning in young autistic children. Jo urnal of Weiss, R. S. {198 1). INR EAL intervention for language handicapped
Co 11s11lti11gand Clinical Psycholol,')\ SS, 3-9. and bilingual children. journal of tbe Drvision for Early Childhood,
McEachin, J.J., Smith, ·r., & Lovaas, 0. L (J 993). Long-term outcomes 4. 40-52.
for children wirh aunsm who received ear ly inrensive behav ioral Wetherby, r\. M., & Prizanr, B. M. ( 1992). Facilitating language and
uearmcnt. Am erican journal on Mental Retardation, 97(4). 359-372. communicat iun development in autism: Assessment and interven-
McGee, G. C., Krantz, P. J., & McC lannahan, L. E. S. {l 985). The fa- tion guidelines. In D. E. Bcrkell (Ed.), Autis m: Tdentification, ed u-
cilirarivc effects of incidental reaching on preposition use by autis- cation and treatment. (pp. 107-134) Hillsdale, NJ: Erlbaum.
tic children. journal of App lied Behavior Analysis, 18, 17-31. Yoder, P. J., Kaiser, A. P., Alpert, C.. & Fischer, R. ( l 993). Following
Miller, J., & Chapman, R. ( 1984). SALT: Systematic analysis of lan- the child's lead when reaching nouns to preschoolers with mental
guage transcri/1ts. Baltimore: University Park Press. retardation . .Journal of Speech and Hearing Research. 36, [58- 167.