The Effectiveness of Parent-Implemented Language Interventions: A Meta-Analysis

Abstract

Purpose
The purpose of this meta-analysis was to systematically evaluate the effects of parent-implemented language interventions on the language skills of children between 18 and 60 months of age with primary and secondary language impairments.

Method
A systematic literature search yielded 18 studies that met the predetermined inclusion and exclusion criteria. Effect sizes for each study were calculated for 7 language outcome variables and analyzed using a random effects model. Separate analyses were conducted for each language outcome and for each comparison group. Outcomes were compared for children with and without intellectual disabilities and for parent report and direct observational language measures.

Results
The results indicate that parent-implemented language interventions have a significant, positive impact on receptive and expressive language skills of children with and without intellectual disabilities. Effect sizes (g) for child measures ranged from −0.15 to 0.82 depending on the outcome measure and comparison group.

Conclusion
The results of this review indicate that parent-implemented language interventions are an effective approach to early language intervention for young children with language impairments. Critical features of parent-implemented interventions are discussed in terms of implications for practice and future research.

Full text

AJSLP
Research Article
The Effectiveness of Parent-Implemented
Language Interventions: A Meta-Analysis
Megan Y. Roberts
a
and Ann P. Kaiser
a
Purpose: The purpose of this meta-analysis was to system-
atically evaluate the effects of parent-implemented language
interventions on the language skills of children between 18
and 60 months of age with primary and secondary language
impairments.
Method: A systematic literature search yielded 18 studies
that met the predetermined inclusion and exclusion criteria.
Effect sizes for each study were calculated for 7 language
outcome variables and analyzed using a random effects
model. Separate analyses were conducted for each language
outcome and for each comparison group. Outcomes were
compared for children with and without intellectual disabilities
and for parent report and direct observational language
measures.
Results: The results indicate that parent-implemented lan-
guage interventions have a significant, positive impact on
receptive and expressive language skills of children with
and without intellectual disabilities. Effect sizes (g) for child
measures ranged from –0.15 to 0.82 depending on the
outcome measure and comparison group.
Conclusion: The results of this review indicate that parent-
implemented language interventions are an effective approach
to early language intervention for young children with language
impairments. Critical features of parent-implemented inter-
ventions are discussed in terms of implications for practice
and future research.
Key Words: language delays, parent training,
language intervention
Parents play a prominent role in their children’s lan-
guage development (Hart & Risley, 1995; Smith,
Landry, & Swank, 2000; Tamis-LeMonda, Bornstein,
& Baumwell, 2001). Parents are commonly acknowledged
as children’s first language teachers; however, their role in
the development of language is multifaceted. Four aspects
of parent–child interaction are associated with child lan-
guage development: (a) amount of parent–child interaction,
(b) responsiveness to child communication, (c) amount and
quality of linguistic input, and (d) use of language learning
support strategies. Variations in these features are associated
with differences in children’s language development both
within the population of typically developing children and
between children with and without language impairments.
Amount of parent–child interaction. Differences in the
amount of parent–child interaction are associated with dif-
ferences in language development from a very young age.
For example, Tomasello and Todd (1983) found that typi-
cally developing children whose mothers engaged in greater
amounts of joint interaction had larger productive vocabu-
laries between 12 and 18 months of age than children whose
mothers engaged in less joint interaction. In a study of
60 infants between 8 and 10 months of age, Alston and
St. James-Roberts (2005) found that mothers of infants show-
ing early signs of communication difficulties spent half as
much time interacting with their infants as compared to
mothers of normally developing infants. Hammer, Tomblin,
Zhang, and Weiss (2001) reported that parents of children
with normal language engaged in more conversational
activities with their children than parents of children with
specific language impairment (SLI).
Responsiveness. Generally, responsiveness refers to par-
ents’verbal and nonverbal responses to the child’s com-
munication attempts, play actions, and social eye contact;
however, definitions vary by study and age of child. Tamis-
LeMonda et al. (2001) examined the relationship between
maternal responsiveness and the occurrence of expressive
language milestones in typically developing children. Mater-
nal responsiveness to child vocalizations at 13 months pre-
dicted the timing of several expressive language milestones
(e.g., first 50 words, combinatorial speech, or first use of
language to talk about the past). This relationship between
parent responsiveness and child language growth also ex-
ists for children with language impairments. Girolametto,
Weitzman, Wiigs, and Pearce (1999) reported a significant
relationship between maternal contingent responsiveness
(e.g., imitation or expansion of the child’s communication)
and child language development (e.g., number of words or
a
Vanderbilt University, Nashville, TN
Correspondence to Megan Y. Roberts:
megan.y.roberts@vanderbilt.edu
Editor and Associate Editor: Laura Justice
Received June 10, 2010
Revision received January 6, 2011
Accepted March 4, 2011
DOI: 10.1044/1058-0360(2011/10-0055)
American Journal of Speech-Language Pathology •Vol. 20 •180–199 •August 2011 •AAmerican Speech-Language-Hearing Association180

word combinations) in 12 children with expressive lan-
guage delays observed before and after intervention. Yoder,
McCathren, Warren, and Watson (2001), in a descriptive
study of 58 children between 17 and 33 months of age with
developmental disabilities, found that maternal responsive-
ness to child intentional communication acts at the start of
the study was positively related to expressive and receptive
language 12 months later. Wulbert, Inglis, Kriegsmann, and
Mills (1975) compared maternal responsiveness between
mothers of children with and without language impairments
and found that mothers of children with language impair-
ments were less responsive than mothers of children with
typical language.
Language input. The relationship between amount of lan-
guage exposure during parent–child interactions and vocab-
ulary growth has been studied extensively. In a study of
22 children between 14 and 26 months of age, Huttenlocher,
Haight, Bryk, Seltzer, and Lyons (1991) found a significant
positive relationship between the amount of maternal lin-
guistic input and children’s vocabulary growth. Hart and
Risley (1995) also observed a positive relationship between
the amount of parent talk and the children’s vocabulary size
across families from different demographic backgrounds.
Rowe (2008) found a similar relation between the amount
of child-directed speech and typical children’s receptive
vocabulary at 30 and 42 months.
In addition to the quantity of language input, the content
and quality of language input also affect child language de-
velopment. The diversity of words that parents use is asso-
ciated with the size of children’s expressive vocabulary. Hoff
and Naigles (2002), in a study of 63 typically developing
2-year-olds, found that lexical richness (i.e., number of dif-
ferent words and mean length of utterance) and syntactic
complexity of maternal speech predicted children’s produc-
tive vocabulary. Weizman and Snow (2001), in a study of
316 children, found that both the number of sophisticated
words and the density of sophisticated words in parents’talk
to children predicted the children’s receptive vocabulary in
kindergarten and second grade. Furthermore, parental lin-
guistic input may also vary depending on the child’s language
ability. Harris, Jones, Brookes, and Grant (1986) found that
mothers of typically developing children labeled objects
in their child’s focus of attention more frequently than did
mothers of children with slower language development.
Language support strategies. Parents naturally use a num-
ber of language support strategies (e.g., specifying relations
between objects and actions, and expanding and recasting
child utterances to add words and syntactic structures) that
play an important role in child language development. Smith
et al. (2000) reported that support strategies occurred in 18%
of parent–child interactions, and frequency of parental use
of strategies predicted verbal skills of children in a sample of
312 children 3 years of age.
Parents of children with language impairments may not
use language support strategies in the same manner or at the
same rate as parents of children with typical language. In a
study of 18 children between 2 and 6 years of age, Conti-
Ramsden, Hutcheson, and Grove (1995) found that parents
of children with SLI used fewer simple recasts (i.e., the
parent repeated the child’s previous sentence and added
words) than parents of children with typically developing
language. Vigil, Hodges, and Klee (2005) also found that
parents of toddlers with typical language used more expan-
sions than parents of toddlers with language impairments.
While these studies support the premise that differences
in parent interaction, amount of talk, linguistic input, and use
of language support strategies are associated with children’s
language development, it is important to consider the tran-
sactional nature of interactions and the bidirectional influ-
ences of children on parents and parents on children. Children
with language impairments may differ from typical children
in rate of communication, development of joint attention
(Mundy, Kasari, Sigman, & Ruskin, 1995), clarity of com-
municative intention, intelligibility (Rice, Sell, & Hadley,
1991), and responsiveness to language (Wetherby, Prizant,
& Hutchinson, 1998). These child differences may contribute
to differences in parents. Teaching parents to modify their
interaction styles and linguistic input may be an important
element of early language intervention for young children
with language impairments.
The Need for Parent-Implemented Interventions
Given the critical role that parents play in their children’s
language development and the observed differences in parent
interaction strategies with children who have language im-
pairments, teaching parents to support language develop-
ment is an important component of effective early language
intervention. The need for interventions that include typi-
cal communication partners and are delivered in children’s
natural environments is widely recognized. The Individuals
with Disabilities Education Improvement Act of 2004 high-
lighted the importance of parent–child interactions in the
home environment by mandating that interventions for young
children with disabilities be implemented within typical and
authentic learning experiences.
Triadic intervention. Teaching parents to implement spe-
cific language intervention strategies to improve language
development in their children is a triadic intervention model;
that is, a skilled interventionist teaches parents to use specific
language intervention strategies with their children. The
success of this approach depends on parents learning and
using the strategies with sufficient frequency and accuracy to
influence their child’s development. The content of these lan-
guage support strategies varies. Generally, the strategies have
been derived from (a) descriptions of the normative parent–
child interactions discussed previously (e.g., Girolametto
et al., 1999; Tamis-LeMonda et al., 2001, Tomasello & Todd,
1983; Vigil et al., 2005; Weizman & Snow, 2001), (b) behav-
ioral learning principles (Schreibman & Koegel, 2005), or
(c) a hybrid of these (Dawson et al., 2010; Kaiser, 1993).
Studying triadic interventions requires a multimethod
approach for monitoring and measuring (a) parent training,
(b) parent implementation of intervention strategies, and
(c) child language outcomes. Methodologically strong stud-
ies measure the procedures for teaching parents specific
strategies, parents’implementation of these strategies, and
the effects of the intervention on child language development.
Parent generalization of language teaching strategies to home
interactions and maintenance of these strategies for periods
Roberts & Kaiser: Parent-Implemented Language Interventions 181

sufficient to affect children’s development require measure-
ment across settings and time. Thus, quantifying the dosage
of intervention received by the children is a considerable
challenge. Few studies measure all aspects of the triadic inter-
vention and both immediate and long-term child outcomes.
However, there has been a sufficient body of single-subject
and group design research to build a case for the efficacy
of parent-implemented interventions over the last 3 decades.
Effects of Parent-Implemented Interventions
Beginning in the 1970s, studies demonstrated that parents
could be taught specific strategies to support their children’s
language learning (Fey, Cleave, Long, & Hughes, 1993;
Girolametto, 1988; Tannock, Girolametto, & Siegel, 1992;
Weistuch & Lewis, 1985). Although there was early evi-
dence to suggest that parent behavior could be altered to
create a more supportive interactional context for children
(Cheseldine & McConkey, 1979; Mahoney & Powell, 1988;
Price, 1984), only more recent studies have included evi-
dence that changes in children’s communication are asso-
ciated with specific changes in parent behavior (Alpert &
Kaiser, 1992; Delaney & Kaiser, 2001; Fey et al., 1993;
Hancock, Kaiser, & Delaney, 2002; Kaiser, Hancock, &
Hester, 1998).
Findings from single-subject research. Numerous single-
subject studies have investigated the effects of teaching par-
ents to use specific language intervention strategies with
their young children with language impairments. However,
there are no comprehensive reviews of these single-subject
studies across populations and interventions. Meadan,
Ostrosky, Zaghlawan, and Yu (2009) systematically reviewed
single-subject studies of parent-implemented language inter-
ventions for children with autism. They found that all studies
reported positive outcomes for children but that interventions
varied among studies. Interventions included modified in-
cidental teaching, reciprocal imitation training, routines-based
instruction, enhanced milieu teaching, functional communi-
cation training, and the Denver Model.
Single-subject design studies typically pose two related
research questions: (a) Does parent training increase parents’
use of specific language intervention strategies, and ( b) does
parent use of specific language support strategies increase
child language skills? When parents are successfully taught
to use systematic language intervention strategies, children
show increases in their immediate use of new target forms
(Kashinath, Woods, & Goldstein, 2006). Common skills
taught to parents include (a) responding to communication,
(b) adjusting the balance of adult to child communication
turns, (c) using language models, and (d) using incidental
teaching strategies. More than 20 single-subject design stud-
ies have investigated parent-implemented language inter-
vention across populations of children with intellectual
disabilities (ID; Hemmeter & Kaiser, 1994), children with
autism (Gillet & LeBlanc, 2007), and children at risk for
language impairments (Hancock et al., 2002). Examples of
methodologically strong studies described below highlight
the contributions of these studies.
Delaney and Kaiser (2001) used a multiple-probe design
across parents and replicated across behaviors to study the
effects of a parent-implemented language intervention with
four children between 41 and 47 months of age with language
impairments and problem behaviors. In this study, parents
were trained to use three language intervention strategies:
(a) reducing amount of adult verbal turns, ( b) responding to
child communication, and (c) expanding child communica-
tion. Training included reviewing the target strategy, coach-
ing by a therapist, and feedback at the end of the session.
All four parents successfully learned each of these strategies.
Although positive changes in child language and behavior
were observed, the design of the study limited interpretation
of child effects associated with parent use of the interven-
tion components. Specifically, changes in study conditions
were based on parent behavior only, not on changes in child
language.
Hemmeter and Kaiser (1994) taught four parents to use
four language intervention strategies—(a) environmental
arrangement, (b) responsive feedback, (c) language targets,
and (d) incidental teaching—with their children with de-
velopmental delays between 25 and 49 months of age. In
this study, parent training on a new strategy was introduced
when the parent showed mastery of the previous strategy.
Parent training resulted in positive changes in parental use
of all language intervention strategies within and across
parents.
Kashinath et al. (2006) used a similar design and taught
language intervention strategies (i.e., arranging the environ-
ment, using natural reinforcement, using time delay, imitat-
ing contingently, modeling, and gestural/visual cuing) to
four parents of children with autism. All parents increased
their use of language intervention strategies and generalized
some skills to other routines. Like the previous studies, con-
clusions about the effects of parent training on child out-
comes are difficult to interpret because the independent
variable (i.e., teaching parents to use intervention strategies)
was manipulated based on changes in parent behavior rather
than on changes in child language. However in the Kashinath
et al. study, the frequency of single words was graphed for
each child in relation to the different skills parents were taught.
Across all four children, children increased their use of single
words following intervention.
Gillet and LeBlanc (2007) examined changes in child
communication following parent training. In a study of three
children with autism, they taught parents to implement the
natural language paradigm (NLP). Only after the previous
child showed changes in vocalizations was parent training
introduced to the next parent. Vocalizations per minute in-
creased for each child following parents’implementation
of these NLP procedures.
The results from the single-subject studies discussed here
indicate that parents can learn several language intervention
strategies. These results also suggest that when parents use
these strategies, their children’s language skills appear to
increase. An important contribution of single-subject studies
is the precise descriptions of parent behaviors and strate-
gies for parent training. Measurement of parents’use of lan-
guage strategies is detailed in single-subject studies because
these strategies are the dependent variable that is measured
continuously over time. It is also important to note that not
all parent-implemented language interventions have been
182 American Journal of Speech-Language Pathology •Vol. 20 •180–199 •August 2011

studied using single-subject designs. For example, the Hanen
Parent Program (Manolson, 1992) is delivered in groups of
six to eight parents, making single-subject methodology more
difficult to implement.
Previous reviews of group design studies of parent-
implemented interventions. Few reviews of parent-implemented
language interventions using group designs have been pub-
lished. McConachie and Diggle (2007) conducted a system-
atic review of parent-implemented early interventions for
young children with autism spectrum disorder, but the focus
of the review included a range of developmental outcomes
(e.g., behavioral, cognitive, and linguistic). Only three stud-
ies in the review included language outcomes. In these three
studies, the effects of parent-implemented intervention were
nonsignificant for parent report of child vocabulary and
sentence length. The authors noted that the limited number
of studies with varying degrees of methodological rigor and
small sample sizes made drawing definitive conclusions
difficult.
Law, Garrett, and Nye (2004) completed a meta-analysis
on the efficacy of all types of speech and language interven-
tion (e.g., parent-implemented and therapist-implemented)
for children with language impairments and no other con-
comitant disorders. They found no statistically significant
differences between the outcomes for parent-implemented
and therapist-implemented interventions. When comparing
parent-implemented language intervention to nontreatment
control groups on standardized measures of language de-
velopment (e.g., expressive syntax, receptive syntax, and
expressive vocabulary), they found overall nonsignificant
effect sizes for all outcome measures. Effect sizes (d) for in-
dividual language constructs ranged from –0.53 for receptive
syntax to 0.83 for expressive syntax and 1.06 for expressive
vocabulary. These results suggest that parent-implemented
language interventions may have a larger effect on vocab-
ulary than receptive syntax. It may be easier for parents to
facilitate expressive vocabulary growth than receptive syn-
tactic growth and easier to facilitate expressive language
than receptive language. It is important to note that the Law
et al. (2004) review only included three parent-implemented
intervention studies. In addition, the focus of this review
was children with only language impairments and excluded
children with ID.
Both the McConachie and Diggle (2007) and the Law
et al. (2004) reviews focused on populations of children who
are a subset of young children with language impairments:
children with autism and children with language impair-
ments and no ID. As a result of the population-specific focus,
both reviews included small numbers of studies of parent-
implemented language interventions. Because the population
of children with language impairments is broader than the
specific populations examined in these previous reviews,
there is a need for a meta-analysis that includes all children
with language impairments, including ID, autism spectrum
disorders, and primary language impairments. Such a review
would allow for comparative analyses of the effects of
parent-implemented language intervention on subgroups
as well as identification of common outcomes of parent-
implemented language interventions across children with
different types of language impairments.
Rationale for Meta-Analysis
Meta-analysis methodology was chosen for this review
for several reasons. First, meta-analysis is a structured re-
search technique with documentation at each step (e.g., search
strategy, coding, and analysis) which insures that replication
of the search and analysis process is possible (Cooper &
Hedges, 2009). Second, in meta-analysis, the magnitude
and direction of results are coded for each study, rather than
only counting significant and nonsignificant results. Because
statistical significance reflects a combination of strength of
effects and sampling error, which is highly dependent on
sample size, simply examining statistical significance across
a group of studies can be misleading (Hedges, 2009). Third,
through detailed, systematic coding of study characteris-
tics, meta-analysis allows quantitative examination of the
relationship between study characteristics and study results
(e.g., cognitive ability of participants and language out-
comes; Wilson, 2009). Fourth, because all studies of parent-
implemented interventions have relatively small samples,
meta-analysis is an effective synthesis technique. In meta-
analysis, the results of the small studies are pooled, thereby
increasing the statistical power to detect differences between
experimental conditions (Lipsey & Wilson, 2001). A com-
plete description of the meta-analytic procedures used can
be found in Cooper, Hedges, and Valentine (2009).
Objectives
Although there are theoretical and policy-driven bases for
parent-implemented language interventions, the empirical
evidence across populations of children with language im-
pairments has not been systematically reviewed. Systematic
evaluation of parent-implemented language interventions is
complicated by the use of various terminologies. Many early
interventions are labeled as “family-centered,”which is de-
fined as (a) treating families with respect; (b) providing flex-
ible and individualized services; (c) sharing information with
parents; and (d) including parents in the assessment, plan-
ning, and intervention process (Dunst, 2002). Family-centered
interventions do not necessarily include the parent as the agent
delivering the intervention. Parents may play a number of
roles in the intervention process that do not include acting as
the primary interventionist. In this review, only interventions
in which the parent was the primary interventionist who im-
plemented the language intervention strategies were included
(i.e., parent-implemented language interventions). Because
previous systematic reviews of parent-implemented language
interventions have focused on specific target populations
(e.g., autism or SLI), this review focused on determining the
effects of parent-implemented interventions on language
development for young children with all types of language
impairments. The purpose of this meta-analysis was to eval-
uate the effects of parent-implemented language interventions
on the language skills of children between 18 and 60 months
of age with primary and secondary language impairments.
The following questions guided this meta-analysis:
1. Do parent-implemented interventions positively affect
language outcomes of young children with language
impairments?
Roberts & Kaiser: Parent-Implemented Language Interventions 183

2. Is early language intervention more effective when
delivered by a parent than when delivered by a therapist?
3. On which aspects of child language outcomes (e.g.,
vocabulary or syntax) do parent-implemented language
interventions have the largest effects?
4. Do the effects of parent-implemented interventions differ
for children with and without ID?
5. Does type of language measure (i.e., parent report or
direct observation) affect the magnitude of the effects?
6. Do parent-implemented interventions positively affect
parent use of language intervention strategies?
Method
Identification of Studies
Prior to the identification of studies, inclusion criteria
were selected based on study design, type of intervention,
participant characteristics, and outcome measures (see Table 1).
Using these criteria, several methods were used to identify
studies of parent-implemented language interventions. First,
a database search of CSA Illumina was conducted using
the following databases: ERIC, PsycArticles, PsycInfo, and
CSA Linguistics and Language Behavior Abstracts. The
key words used in these searches are listed in Table 2. Second,
the reference lists of previous meta-analyses and literature
reviews (Law et al., 2004; McConachie & Diggle, 2007) were
reviewed for studies that met the inclusion criteria. The refer-
ence lists of all retrieved studies were also examined. Third,
follow-up searches using the names of the first and second
authors of all included studies were conducted. Fourth, ab-
stracts of conference presentations were reviewed for papers
reporting effects of parent-implemented language interven-
tions, and authors were contacted to retrieve additional papers
or unpublished data.
As shown in Table 2, the search yielded 951 unique ab-
stracts for review. During the first selection phase, abstracts
were screened for inclusion criteria by the first author. For
a study to be definitively excluded, the abstract had to clearly
indicate that the study failed to meet one of the specified
inclusion criteria. When inclusion could not be determined
from the abstract, full text documents were retrieved for the
second selection phase. In the first phase, the majority of
studies were excluded for the following reasons: The studies
did not employ a group design, participants did not have
language impairments, and the focus of the intervention was
not language skills. During this screening phase, the number
of studies was reduced from 951 to 59.
Next, the first author retrieved full text study reports for
all 59 studies that passed the initial screening. These studies
were retrieved from the Vanderbilt Library System, inter-
library loan, University Microfilms, and the databases listed
above. The method and results sections of the full study
reports were examined in detail to determine whether each
study met the inclusion criteria. The purpose of this screen-
ing step was to obtain a fuller understanding of each article
by reviewing information that might not have been provided in
the abstract. During this second screening phase, the number
of studies was reduced from 59 to 26. Last, during the final
full coding of the 26 retained articles, an additional eight
studies were excluded. These studies and the reason for their
exclusionarepresentedinTable3.Asummaryofthe18in-
cluded studies is presented in Table 4. One study report
(Gibbard, 1994) included data from two separate studies.
A second reviewer independently determined eligibility
for the 59 studies for which the full text report was obtained.
Agreement about eligibility for inclusion between the two
reviewers was calculated using an intraclass correlation
(ICC). The results indicate high agreement between the two
reviewers (ICC = .95). Disagreements were resolved through
discussion and review of the inclusion criteria.
Data Management and Extraction
All study titles and inclusion decisions were documented
and managed using SPSS Version 16 software. When pos-
sible, PDF files were saved and organized by author and
publication year. Hard copies of print-only articles were kept
on file. After collecting the full study report, information was
extracted from each article using a detailed coding proto-
col (this protocol may be obtained by contacting the first
author). Two reviewers independently coded each full article
for the variables listed in Table 4. Agreement for variables
between the two reviewers was calculated using ICCs for
continuous variables and Cohen’s kappa for dichotomous
variables. The results indicate high agreement between the
two reviewers on all key variables presented in Table 4.
Mean kappa was .90 with a range of .25–1.0. Mean ICC was
.94 with a range of .89–1.0. Coding disagreements were
discussed and resolved such that only data with perfect
agreement were included in the analysis. All study coding,
data management, and data analysis were done using SPSS
Version 16 software.
Description of Included Studies
Characteristics of the 18 included studies are summarized
in Tables 5 and 6. All studies except for three (Baxendale
TABLE 1. Inclusion criteria.
Aspect Criteria
Design Nontreatment, “business as usual,”or
therapist-implemented comparison group
Random or nonrandom assignment
Pre/post comparison
Published, unpublished, or dissertations
Intervention Implemented only by parents
Included a component that directly affected
child communication
Excluded dialogic reading
Participants Between 18 and 60 months
Any type of language impairment (i.e.,
1SD below the mean on standardized
assessments or less than 50 expressive
words at age 2)
Outcomes At least one language outcome measure
(e.g., receptive, expressive, vocabulary,
syntax, or rate)
Any type of measure (e.g., parent report,
observational, or standardized)
184 American Journal of Speech-Language Pathology •Vol. 20 •180–199 •August 2011

& Hesketh, 2003; Gibbard, Coglan, & McDonald, 2004;
McDade & McCartan, 1998) were randomized group experi-
ments. The effect sizes from these three studies were com-
pared to the range of effect sizes in the experimental studies.
Because the effect sizes and sample size were within the
ranges of the effect sizes in the randomized group experi-
ments, they were included in the meta-analysis.
Sample characteristics. The average sample size across
all studies was 38, with a range of 12 to 152 participants.
Seven of the identified studies included children with de-
velopmental delays (including children with autism), and
11 studies included children with language impairments
only. The majority of studies (n= 10) included chil-
dren between 24 and 36 months of age. However, age
of participants ranged from 15 to 77 months across
studies.
Intervention characteristics. The general characteristics
of the interventions and skills taught to parents are sum-
marized in Table 6. The Hanen Parent Program was the
most frequently studied parent-implemented language
intervention (n= 8). Strategies taught to parents across
studies included responding to child communication, bal-
ancing adult–child turns, and using language models. Inter-
ventions were primarily short in duration, lasted between
TABLE 3. Excluded studies and reasons for exclusion.
Study name Reasons for exclusion
Howlin, P. (1981). The results of a home-based language training
programme with autistic children. British Journal of Disorders of
Communication, 16, 73–88.
Participants’average age was greater than
60 months.
Jocelyn, L., Casiro, O., Beattie, D., Baw, J., & Kneiz, J. (1998).
Treatment of children with autism: A randomized controlled trial
to evaluate a caregiver-based intervention program in community
day-care centers. Journal of Developmental and Behavioral
Pediatrics, 19, 326–334.
Intervention was delivered simultaneously
by parent and classroom teacher.
Kot, A., & Law, J. (1995). Intervention with preschool children with
specific language impairments: A comparison of two different
approaches to treatment. Child Language Teaching and Therapy,
11, 144–162.
Intervention was delivered simultaneously
by parent and therapist.
Littleton, R. (2004). The modifiability of language input with toddlers
with expressive language delay: A study of a team approach to
parent training. Dissertation Abstracts International: Section A.
The Humanities and Social Sciences, 65, 115.
Each experimental condition had fewer
than 5 participants.
Smith, T., Groen, A., & Wynn, J. (2000). Randomized trial of intensive
early intervention for children with pervasive developmental
disorder. American Journal on Mental Retardation, 105, 269–285.
Both experimental and control conditions
included parent training.
Ward, S. (1999). An investigation into the effectiveness of an early
intervention method for delayed language development in young
children. International Journal of Language & Communication
Disorders, 34, 243–264.
Average age of participants was less than
18 months.
Wetherby, Q., & Woods, J. (2006). Early social interaction project for
children with autism spectrum disorders beginning in the second
year of life: A preliminary study. Topics in Early Childhood Special
Education, 26, 67–82.
Dependent variable was not measured
prior to intervention for the control group.
Yoder, P., & Warren, S. (2002). Effects of prelinguistic milieu teaching
and parent responsivity education on dyads involving children with
intellectual disabilities. Journal of Speech, Language, and Hearing
Research, 45, 1158–1174.
Intervention was delivered simultaneously
by parent and therapist.
TABLE 2. Search terms and respective yields.
Search source Search terms
Number
of abstracts
CSA Illumina (using ERIC, PsycInfo,
PsycArticles, and CSA Linguistics
and Language Behavior Abstracts)
(Descriptors) (“parent training”OR “parent
education”OR “parents as teachers”)
AND (“language”OR “communication”
OR “vocabulary”) AND (“intervention”
OR “effectiveness”OR “efficacy”)
936
Review of reviews References from Law et al. (2004) and
McConachie & Diggle (2007)
7
Ancestral search References from identified studies 5
Conference abstracts Parent-implemented interventions 3
Total number of abstracts reviewed 951
Roberts & Kaiser: Parent-Implemented Language Interventions 185

10 and 13 weeks, and included less than 26 total hr of parent
training. Eleven of the interventions included at least some
training at home.
Outcome measures. The most common language construct
measured as an outcome of intervention was expressive
vocabulary (n= 15). Expressive morphosyntax was measured
in 10 studies, and general receptive and expressive language
skills were each measured in nine studies. A summary of
the measures used in each language construct for individual
studies is provided in Table 7.
TABLE 4. Characteristics of parent-implemented language intervention studies.
Characteristics N% Characteristics N%
Study characteristics
Publication type Year of publication
Journal 16 89 1980–1990 1 6
Technical report 1 5 1991–2000 10 55
Unpublished data 1 5 2000–2010 7 39
Country Design
United States 2 11 Random 15 83
Canada 4 22 Nonrandom 3 17
United Kingdom 10 56 Design problems
Other European country 2 11 Yes, favors control 1 5
Type of control group No or not stated 15 83
Nontreatment control 10 46 Yes, favors treatment 2 11
Business as usual 5 23 Attrition from posttest
Therapist 7 32 0% 11 61
Treatment fidelity 1–10% 3 17
Reported fidelity 5 28 11–20% 3 17
None reported 13 72 >20% 1 5
Sample characteristics
Sample size Percentage male
10–20 2 11 50–74 8 44
21–30 9 50 75–100 8 44
31–40 4 22 Not reported 2 11
41–50 3 17 Percentage minority
Average age of child participants (months) 1–50 3 17
22 1 5 51–99 1 5
24–35 10 56 Not reported 14 78
36–47 5 28 Disability type
48–59 2 11 Autism 3 17
Socioeconomic status Language impairment 11 61
Primarily middle class 9 50 Developmental delay 4 22
Mixed lower and middle class 2 11 Intellectual disability
Not reported 7 39 No 11 61
Average parent education Yes 7 39
High school 1 5 Average parent age ( years)
More than high school 6 33 20–30 3 17
Not reported 11 61 30–40 9 50
Not reported 6 33
Intervention characteristics
Duration of intervention (weeks) Home training
10–13 11 61 No home training 7 39
20 1 5 Some home training 9 50
26 3 17 All home training 2 11
52 3 17 Described parent training strategies
Total number of training hours No 9 50
13–20 7 39 Yes 9 50
22–26 7 39 Use of specific language targets
32–36 2 11 No 9 50
Not reported 2 11 Yes 8 45
Not reported 1 5
Measure characteristics
Type of measure Construct measured
Only parent report 1 5 Overall language ability 9 50
Only observational 2 11 General receptive language 9 50
Only norm-referenced 0 0 General expressive language 8 44
Mixed measures 15 83 Expressive vocabulary 15 83
Measured parent use of strategies Receptive vocabulary 5 28
Yes 10 56 Expressive morphosyntax 10 56
No 8 44 Rate 9 50
186 American Journal of Speech-Language Pathology •Vol. 20 •180–199 •August 2011

Study quality. Study quality characteristics, such as attri-
tion, treatment fidelity, assignment of participants to conditions,
and differences between groups at pretest, are summarized in
Table 5. The majority of studies (n= 11) reported low levels
of attrition, equivalence of groups at pretest (n= 15), and
random assignment of participants (n= 15). The majority
of studies did not report a measure of treatment fidelity
(n= 13). Moderator analyses indicated that none of these study
quality variables were significantly associated with effect
sizes.
Data Synthesis
Effect size calculations and adjustments. Because sample
sizes were small in the majority of studies, and pretest and
posttest scores were available for all studies, effect sizes were
TABLE 5. Summary of included parent-implemented language intervention studies.
Study nAssignment Attrition
Mean age
in months
(range) Comparison group Diagnosis Intervention Fidelity
Parent
measures
Aldred et al.
(2004)
28 Random 0.00 45 (24–71) Community services ASD Child Talk No Responsiveness
Baxendale &
Hesketh
(2003)
37 Nonrandom 0.11 33 (29–41) Therapist, 12 sessions,
12 weeks
LI Hanen Parent
Program
No Responsiveness,
rate
Buschmann
et al.
(2008)
58 Random 0.19 25 (24–27) Nontreatment LI Heidelberg Parent-
Based Language
Intervention
No None
Drew et al.
(2002)
24 Random 0.00 23 Community services ASD Social pragmatic
joint attention
No None
Fey et al.
(1993)
30 Random 0.00 56 (44–70) Nontreatment; therapist,
22 sessions, 20 weeks,
22 total hr
LI Focused
stimulation
No None
Gibbard
(1994) 1
36 Random 0.00 34 (20–42) Nontreatment LI Parent-based
intervention
No None
Gibbard
(1994) 2
25 Random 0.00 32 (29–36) Nontreatment; therapist,
26 sessions, 26 weeks,
13 total hr
LI Parent-based
intervention
No None
Gibbard et al.
(2004)
28 Nonrandom 0.21 27 (22–32) Therapist, 26 weeks LI Parent-based
intervention
No None
Girolametto
(1988)
22 Random 0.09 37 (15–62) Nontreatment DD Hanen Parent
Program
No Responsiveness,
rate
Girolametto
et al.
(1996a)
16 Random 0.00 29 (22–38) Nontreatment LI Hanen Parent
Program and
focused
stimulation
Yes Responsiveness,
rate, language
models
Girolametto
et al.
(1996b)
25 Random 0.00 29 (23–35) Nontreatment LI Hanen Parent
Program and
focused
stimulation
Yes Rate, language
models
Girolametto
et al.
(1998)
12 Random 0.00 38 (29–46) Community services DS Hanen Parent
Program plus
focused
stimulation
Yes Rate, language
models
Green et al.
(2010)
152 Random 0.04 45 (29–60) Community services ASD Parental sensitivity
and
responsiveness
Yes Responsiveness
Kaiser &
Hancock
(1998)
73 Random 0.05 45 (30–77) Community services and
therapist, 24 sessions,
12 weeks, 24 total hr
DD Enhanced milieu
teaching
Yes Responsiveness
Law et al.
(1999)
38 Random 0.00 37 (33–39) Nontreatment; therapist,
18 sessions, 6 weeks,
27 total hr
LI Hanen Parent
Program
No Responsiveness,
language
models
McDade &
McCartan
(1998)
22 Nonrandom 0.14 24 Nontreatment LI Hanen Parent
Program
No None
Tannock
et al.
(1992)
32 Random 0.00 30 Nontreatment DD Hanen Parent
Program
No Language
models
Van Balkom
et al.
(2010)
22 Random 0.05 34 (26–37) Therapist, 12 sessions,
12 weeks, 9 total hr
LI Parent video home
training
No None
Note. ASD = autism spectrum disorders; LI = language impairment; DD = developmental delay; DS = Down syndrome.
Roberts & Kaiser: Parent-Implemented Language Interventions 187

TABLE 6. Summary of parent-implemented intervention characteristics.
Study Intervention
Amount of
parent training
Home
sessions
Quantity of
linguistic input
Quality of
linguistic input Responsiveness
Support
strategies Other strategies
Aldred et al. (2004) Child talk 9 sessions and
maintenance,
52 weeks
0 General
language
models
Joint attention, routines, expanding
play, communication teasers
Baxendale &
Hesketh (2003)
Hanen Parent Program 11 sessions,
11 weeks
3 Adult–child
turn ratio
Imitations Expansions,
recasts
Buschmann et al.
(2008)
Heidelberg Parent-
Based Language
Intervention
8 sessions,
13 weeks,
17 total hr
0 General
language
models
Child-oriented interactions
Drew et al. (2002) Social pragmatic joint
attention
9 sessions,
52 weeks,
26 total hr
9 Behavioral management, joint
attention, turn-taking games,
gestures, routines
Fey et al. (1993) Focused stimulation 19 sessions,
20 weeks,
36 total hr
3 Increased
opportunities
to hear
targets
Specific
language
models
Recasting Contingent queries
Gibbard (1994) 1 Parent-based
intervention
11 sessions,
26 weeks,
14 total hr
0 Specific
language
models
Gibbard (1994) 2 Parent-based
intervention
11 sessions,
26 weeks,
14 total hr
0 Specific
language
models
Gibbard et al. (2004) Parent-based
intervention
11 sessions,
26 weeks,
16.5 total hr
0 Specific
language
models
Girolametto (1988) Hanen Parent Program 12 sessions,
11 weeks,
32 total hr
3 Turn taking General
language
models
Responding to child
communication
Expansions Following the child’s lead, play
and music, sharing books
Girolametto et al.
(1996a)
Hanen Parent Program
and focused
stimulation
13 sessions,
13 weeks,
20 total hr
3 Turn taking Specific
language
models
Responding to child
communication
Expansions
Girolametto et al.
(1996b)
Hanen Parent Program
and focused
stimulation
11 sessions,
11 weeks,
22 total hr
3 Turn taking Specific
language
models
Responding to child
communication
Expansions
Girolametto et al.
(1998)
Hanen Parent Program
and focused
stimulation
13 sessions,
13 weeks,
26 total hr
4 Turn taking Specific
language
models
Responding to child
communication
Expansions
(table continues)
188 American Journal of Speech-Language Pathology •Vol. 20 •180–199 •August 2011

TABLE 6 (continued ).
Study Intervention
Amount of
parent training
Home
sessions
Quantity of
linguistic input
Quality of
linguistic input Responsiveness
Support
strategies Other strategies
Green et al. (2010) Parental sensitivity and
responsiveness
18 sessions,
52 weeks,
36 total hr
0 General
language
models
Responding to child
communication
Expansions Action routines, pauses
Kaiser & Hancock
(1998)
Enhanced milieu
teaching
24 sessions,
12 weeks,
24 total hr
0 Turn taking Specific
language
models
Responding to child
communication
Expansions,
prompting
Law et al. (1999) Hanen Parent Program 10 sessions,
10 weeks,
25 total hr
3 Turn taking General
language
models
Respond to child
communication
Expansions
McDade &
McCartan (1998)
Hanen Parent Program 12 sessions,
12 weeks
3 Turn taking General
language
models
Respond to child
communication
Expansions
Tannock et al.
(1992)
Hanen Parent Program 12 sessions,
12 weeks,
25 total hr
3 Turn taking General
language
models
Respond to child
communication
Expansions
Van Balkom et al.
(2010)
Parent video home
training
3 sessions,
12 weeks,
9 total hr
6Involvement in child activities,
conversational coherence,
reciprocity of communication,
conversational interactions,
attachment, referencing,
relevance, connectivity
Roberts & Kaiser: Parent-Implemented Language Interventions 189

TABLE 7. Included measures by construct for each study.
Study Overall language Receptive language Expressive language
Expressive
vocabulary
Receptive
vocabulary
Expressive
morphosyntax Rate
Aldred et al. (2004) Vineland Communication
Subscale
a
CDI words said CDI words
understood
Communication
acts
Baxendale &
Hesketh (2003)
PLS total score MLU
Buschmann et al.
(2008)
SETK
b
word production,
SETK sentence
production
CDI words said CDI syntax, CDI
morphology
Drew et al. (2002) CDI words said CDI words
understood
Fey et al. (1993) Developmental
sentence score
Gibbard (1994) 1 Reynell Receptive and
Expressive Subscales
c
Reynell Receptive
Subscale
Reynell Expressive
Subscale
Mother reported
total words
Mother reported
phrase length, MLU
Total number of
utterances
Gibbard (1994) 2 Reynell Receptive and
Expressive Subscales
Reynell Receptive
Subscale
Reynell Expressive
Subscale
Mother reported
total words
Mother reported
phrase length, MLU
Total number of
utterances
Gibbard et al.
(2004)
Reynell total score PLS Receptive
Subscale
PLS Expressive
Subscale
Mother reported
total words
Mother reported
phrase length, MLU
Girolametto (1988) SICD Expressive and
Receptive Subscales
SICD Expressive
Subscale
SICD Receptive
Subscale
NDW Percentage of
verbal turns
Girolametto et al.
(1996a)
CDI words said
Girolametto et al.
(1996b)
CDI words said,
NDW
CDI sentence complexity,
multiword combinations
WPM, total number
of utterances
Girolametto et al.
(1998)
CDI words said
Green et al. (2010) PLS Expressive and
Auditory Subscales
PLS Receptive
Subscale
PLS Expressive
Subscale
CDI words said Child initiations
Kaiser & Hancock
(1998)
SICD Expressive and
Receptive Subscales
SICD Receptive
Subscale
SICD Expressive
Subscale
CDI words said,
NDW
PPVT MLU Total number of
utterances
Law et al. (1999) PLS Expressive and
Receptive Subscales
Reynell Receptive
Subscale, PLS
Receptive
Subscale
PLS Expressive
Subscale
CDI words said,
use of nouns
PPVT MLU, CDI sentence
complexity
Communication
acts
McDade &
McCartan (1998)
Information
carrying words
CDI words
understood
Tannock et al.
(1992)
SICD Receptive
Subscale
CDI words said,
NDW
Verbal turns
Van Balkom et al.
(2010)
Reynell Receptive
Subscale
MLU, grammaticality
Note. CDI = MacArthur–Bates Communicative Development Inventories (Fenson et al., 1993); PLS = Preschool Language Scale (Zimmerman et al., 1992); MLU = mean length of utterance;
SICD = Sequenced Inventory of Communication Development (Hedrick et al., 1984); NDW = number of different words obtained through a language sample; WPM = words per minute obtained
through a language sample; PPVT = Peabody Picture Vocabulary Test (Dunn & Dunn, 1997).
a
From Vineland Adaptive Behavior Scales (Sparrow et al., 1984).
b
German norm-referenced test (Grimm, 2000).
c
From Reynell Developmental Language Scales (Reynell & Gruber, 1990).
190 American Journal of Speech-Language Pathology •Vol. 20 •180–199 •August 2011

calculated by subtracting pretest group differences from
the posttest. This method was used to control for differences
in the outcome variables between groups at the pretest
(Morris, 2008):
ESsm ¼ðXG1post XG1preÞðXG2post XG2post Þ
Sp
:
In this formula, XG1is the mean of the outcome measure
for the intervention group, XG2is the mean of the outcome
measure for the comparison group, and S
p
is the pooled
standard deviation of both the intervention and comparison
groups at posttest. Applying this method yielded a more con-
servative estimate of the effects of the parent-implemented
intervention. For example, the pretest effect size for the over-
all language construct was g= 0.31, p= .06, 95% CI [–0.01,
0.63] and approached significance, indicating that children
in the parent-implemented experimental condition had sub-
stantially higher overall language skills than children in the
therapist-implemented condition prior to intervention. Failing
to account for these differences would have inflated the effect
size in favor of the parent-implemented group. Using only
the posttest to compute the effect size yielded an overall
effect size of g=0.53, p=.08, 95% CI [–0.07, 1.12], which
approached significance. However, adjusting for pretest
differences yielded an effect size of g=0.24,p=.35,
95% CI [–0.25, 0.73], which was not significant.
In addition to the main analyses, raw mean difference ef-
fect sizes were calculated for the most commonly used mea-
sure across studies (i.e., parent report of child number of
expressive vocabulary words) in order to quantify interven-
tion effects in a clinically meaningful metric. The raw mean
difference was calculated by subtracting the means of the
raw score of the comparison group from the mean of the
parent-implemented intervention group. The raw mean differ-
ence was only calculated for those studies that used this
common measure.
All effect size computations were adjusted using Hedges’s
(1981) small sample correction (g), such that the unbiased
effect size estimate was:
ES0
sm ¼13
4N9

ESsm:
Effect sizes were not weighted by an index of methodological
quality, as current meta-analytic standards recommend using
study quality characteristics as control or moderator vari-
ables, rather than weighting each effect size, which hides the
relationship between study quality and observed effect sizes
(Lipsey & Wilson, 2001).
Effect sizes and sample sizes that were greater than
1.5 times the interquartile range were Winsorized. To assess
the possibility of publication bias, the “trim and fill”proce-
dure was conducted using Comprehensive Meta-Analysis
software (Version 2). This procedure trims excessively large
studies by reducing the influence of such studies and imputes
small studies that may be missing. Visual inspection of the
funnel plot for each outcome variable and examination of
Egger’s regression intercept pvalues indicated that the ob-
served effect sizes for all outcome variables were likely
based on an unbiased set of studies.
Many studies included multiple comparison groups,
reported results of multiple outcome measures (e.g., vocab-
ulary or syntax), and measured outcomes for the same par-
ticipants over time. To limit the impact of a single study on the
aggregated outcome analysis, several steps were taken to
ensure that each study only contributed one effect size in each
analysis. First, two separate analyses were conducted based
on the comparison condition: (a) “business as usual”or
nontreatment control (parent-implemented vs. control) and
(b) therapist-implemented intervention (parent-implemented
intervention vs. therapist-implemented intervention). Four
studies included three comparison conditions (i.e., parent-
implemented, therapist-implemented, and control). Only
one comparison was used in each analysis. If two parent-
implemented interventions were compared to a control group,
the effect sizes for the two parent-implemented interventions
were averaged. If a parent-implemented intervention was
compared to two control groups, the group that most closely
resembled business as usual was included. Second, when
participants were measured at multiple time points, only the
first time point after the intervention was included in the
analysis, as this was the most common point of measurement
across studies. Third, effect sizes for different language con-
structs (e.g., expressive vocabulary and receptive language)
were analyzed separately. When different measures for the
same language construct were reported within the same study,
the most commonly used measure across studies for a given
construct was chosen. If multiple common measures of a
language construct were used in a given study, these effect sizes
were averaged together for that study. Table 7 lists the measures
included for each language construct for individual studies.
Statistical model for analysis. A random effects model
was used to analyze effect sizes. The standard error for each
effect size was calculated as follows:
SE ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffi
vqþvi
p:
In this equation, v
i
is the estimate of the variance associated
with subject level sampling error:
vi¼nG1þnG2
nG1nG2þðES0
smÞ2
2ðnG1þnG2Þ;
and v
q
is the estimate of the random or between-studies vari-
ance component; v
q
was estimated through SPSS meta-analysis
macros (Lipsey & Wilson, 2001). The mean effect size for the
random effects model was calculated by weighting each
adjusted effect size (ES
i
) by the inverse of its variance (w
i
):
w
i¼1
vqþvi
;
where v
i
is the estimate of the variance associated with subject-
level sampling error, and v
q
is the estimate of the random or
Roberts & Kaiser: Parent-Implemented Language Interventions 191

between-studies variance component. Analysis of the effect
sizes was conducted separately for each outcome construct,
and effects are summarized in Table 7. Moderator analyses
were also conducted for disability type (i.e., ID vs. no ID) and
measure type (i.e., direct observation vs. parent report) using
a random effects model. A moderator analysis for disability
type was only conducted for those constructs for which there
were at least two studies enrolling children with ID and two
studies enrolling children without ID (i.e., overall language,
expressive language, receptive language, expressive vocabu-
lary, and rate). A moderator analysis for disability type was not
conducted for studies that included a therapist-implemented
intervention group because only one study included children
with ID. A moderator analysis for measure type (i.e., direct
observation vs. parent reported) was conducted for only
expressive vocabulary and expressive syntax, as these were
the only two constructs for which there was a sufficient num-
ber of studies in each group.
Results
Do Parent-Implemented Interventions Positively
Affect Child Language Outcomes?
This question was answered by examining the effect sizes
for each of the seven language constructs when parent-
implemented language intervention was compared with
a control group. The effect sizes ranged from g=0.35,
95% CI [0.05, 0.65] for receptive language to g=0.82,
95% CI [0.37, 1.38] for expressive syntax. Effect sizes for six
of the seven language constructs were positive and statisti-
cally significant. Overall effect sizes for each outcome are
presented in Table 8. These results indicate that parent-
implemented intervention had positive, significant effects
on receptive and expressive language skills, receptive and
expressive vocabulary, expressive morphosyntax, and rate
of communication when compared to a control group.
In addition to these standardized mean difference results,
a raw mean difference analysis was also calculated for parent
report of expressive vocabulary. A forest plot summarizing
weighted effect sizes for the raw mean difference for each
study is provided in Figure 1. Forest plots illustrate the effect
size, significance level, and confidence interval for each
study as well as showing the overall aggregated effect size.
The size of the markers representing individual effect sizes
displays the individual weight of each study in the analysis,
with larger studies receiving a greater weight and a larger
marker. Parents in the parent-implemented intervention group
reported that their children said 52 more words, p=.01,
95% CI [12.03, 93.12], than parents in the control group.
Is Intervention More Effective When Delivered
by a Parent or by a Therapist?
When parent-implemented intervention was compared
to a therapist-implemented intervention, effect sizes were
smaller and mostly nonsignificant. Effect sizes ranged from
g=–0.15, p= .48, 95% CI [–0.56, 0.27] for rate of com-
munication to g= 0.42, p= .02, 95% CI [0.08, 0.76] for
expressive morphosyntax. There was no difference between
parent- and therapist-implemented intervention for five of
the seven language constructs. Significant positive effects
were only present for receptive language, g= 0.41, p= .02,
95% CI [0.08, 0.76], and expressive syntax, g= 0.42, p= .02,
95% CI [0.08, 0.76]. These results indicate that differences
between parent- and therapist-implemented interventions are
variable depending on the language construct.
In addition to these standardized mean difference results,
a raw mean difference analysis was calculated for parent
report of child vocabulary. A forest plot summarizing weighted
effect sizes for the raw mean difference for each study is
provided in Figure 2. Parents in the parent-implemented
intervention group reported that their children said 22 more
words, p= .42, 95% CI [–3.87, 74.25], than parents in the
therapist-implemented intervention group.
Which Child Language Outcomes Have
the Largest Effects?
Examination of the effect sizes and confidence intervals
in Table 8 reveals that expressive morphosyntax had the
largest effects for both the control comparison, g= 0.82,
p= .00, 95% CI [0.37, 1.38], and the therapist-implemented
comparison, g=0.42,p= .92, 95% CI [0.06, 0.79]. However,
TABLE 8. Mean effect sizes, confidence intervals, significance levels, and sample sizes for all language outcome
constructs.
Parent versus control Parent versus therapist
g95% CI pn g 95% CI pn
Child measures
Overall language 0.45 [–0.02, 0.92] .06 7 0.24 [–0.26, 0.73] .35 5
Expressive language 0.61 [0.00, 1.21] .05 7 0.25 [–0.43, 0.93] .47 4
Receptive language 0.35 [0.05, 0.65] .02 7 0.41 [0.08, 0.76] .02 5
Expressive vocabulary 0.48 [0.24, 0.73] .00 14 0.14 [–0.25, 0.54] .69 4
Receptive vocabulary 0.38 [0.10, 0.66] .01 5 0.19 [–0.26, 0.64] .41 2
Expressive morphosyntax 0.82 [0.37, 1.38] .00 7 0.42 [0.06, 0.79] .02 7
Rate 0.51 [0.18, 0.84] .00 9 –0.15 [–0.56, 0.27] .48 3
Adult measures
Parent responsiveness 0.73 [0.26, 1.20] .00 7
Rate 0.26 [–0.13, 0.64] .19 5
Use of language models 0.38 [–0.03, 0.80] .07 5
192 American Journal of Speech-Language Pathology •Vol. 20 •180–199 •August 2011

no single language construct was significantly larger than
another, as indicated by nonoverlapping confidence inter-
vals. In addition, the magnitude of effects by language
construct varied substantially by comparison group. For ex-
ample, receptive language had the smallest effect, g=0.35,
p=.02, 95% CI [0.05, 0.65], in the control comparison but
the largest effect for the therapist-implemented comparison,
g= 0.42, p= .02, 95% CI [0.06, 0.79].
Do the Effects Differ for Children With
and Without ID?
A moderator analysis between children with and without
ID was only conducted for the control comparison group,
as there was an insufficient number of studies enrolling
children with ID in the therapist-implemented comparison.
Analyses were conducted for overall language, expressive
language, receptive language, expressive vocabulary, and rate
of communication. Effect sizes were only significantly differ-
ent between children with and without ID for expressive
vocabulary, Q
b
= 7.62, p= .01. The effect for children without
ID was g= 0.80, p=.00, 95% CI [0.50, 1.1], and the effect for
children with ID was g= 0.23, p= .10, 95% CI [–0.04, 0.50].
These results indicate that for the majority of language
constructs (i.e., overall language, expressive language,
receptive language, and rate of communication), there were
no significant differences between children with and with-
out ID.
Does Type of Language Measure Affect
the Magnitude of the Effects?
Moderator analyses examining the difference between
parent report and observational measures of expressive
FIGURE 1. Forest plot for raw mean difference in vocabulary for the parent-implemented intervention versus control comparison.
FIGURE 2. Forest plot for raw mean difference in vocabulary for the parent-implemented intervention versus therapist-implemented
intervention comparison.
Roberts & Kaiser: Parent-Implemented Language Interventions 193

vocabulary and morphosyntax were conducted for both
comparison groups. There were no significant differences in
effect sizes for parent-reported measures and observational
measures of expressive vocabulary for control (Q
b
= 0.08,
p= .78) or therapist-implemented comparisons (Q
b
= 0.66,
p= .42). There were also no significant differences for ex-
pressive morphosyntax for control (Q
b
= 0.25, p= .61) or
therapist-implemented comparisons (Q
b
= 1.0, p= .76).
These results indicate that measurement type of expressive
vocabulary and morphosyntax did not significantly affect
the magnitude of the effects in either comparison (i.e., con-
trol or therapist-implemented).
Do Parent-Implemented Interventions Positively
Affect Parent Outcomes?
Ten of the 18 studies reported the effects of parent train-
ing on parents’use of specific language support strate-
gies. The three most commonly measured parent strategies
were (a) parent responsiveness (n= 7), (b) use of language
models (n= 5), and (c) adult rate of communication (n= 5).
These constructs were measured only in studies that com-
pared a parent-implemented intervention group to a control
group.
In general, parent training had a positive impact on
parent–child interaction style in terms of responsiveness,
use of language models, and rate of communication. Parents
who received parent training were significantly more respon-
sive than parents who were not trained, g= 0.73, p= .00,
95% CI [0.26, 1.20]. A forest plot summarizing weighted
effect sizes for parent responsiveness for each study is
provided in Figure 3. Parents who received parent training
also used more language models than parents who were not
trained, g= 0.38, p= .07, CI [–0.03, 0.80]. In addition to
changes in responsiveness and language modeling, parent
training may have increased rate of adult communication,
but the difference between trained parents and parents in
the control condition was not significant, g= 0.26, p= .19,
95% CI [–0.13, 0.64].
Discussion
The objective of this meta-analysis was to understand the
extent to which parent-implemented language interventions
were effective in increasing the language skills of children
between 18 and 60 months of age with primary and secondary
language impairments. Because the research base was small
(i.e., 18 studies), and all studies included small samples,
meta-analysis was an effective synthesis technique. In meta-
analysis, the results of the small studies are pooled, thereby
increasing the statistical power to detect differences between
experimental conditions.
In general, parent-implemented language interventions
had positive effects on children’s language outcomes when
compared to a control group, as indicated by significant effect
sizes across six of the seven language outcome constructs.
The relative effects of parent-implemented interventions were
larger in comparison to a control group than in comparison
to a therapist-implemented intervention group. While effects
for all seven constructs were significant or near significant
for the control comparison, only receptive language and ex-
pressive syntax were significant for the therapist-implemented
comparison. Measures that were more proximal to the inter-
vention (e.g., expressive morphosyntax) had larger effect
sizes than more distal measures of developmental outcomes
(e.g., overall language). Parent-implemented language inter-
ventions were effective for both children with and without
ID. There were no differences in language outcome effects for
children with and without ID, except for expressive vocab-
ulary. There were no differences in the magnitude of effects
between parent report and direct-observational measures of
vocabulary and syntax, suggesting that intervention effects
were not simply the result of changes in parental perceptions
of their children’s language skills, which might have been
reflected in parent report measures. Parent training also had
positive effects on parent use of intervention strategies, with
the largest effect on parent responsiveness.
While the results clearly indicate that the overall effects of
parent-implemented language interventions were positive
and significant, there are several weaknesses in this body of
FIGURE 3. Forest plot of responsiveness for the parent-implemented intervention versus control comparison.
194 American Journal of Speech-Language Pathology •Vol. 20 •180–199 •August 2011

research studies. First, the majority of studies (72%) did
not measure treatment fidelity, and half of the studies (50%)
did not adequately describe the parent training procedures.
Little information was provided about the personnel who
delivered the parent training, strategies used to teach parents,
or how much parent training occurred within the parent and
child sessions. Parents’implementation of the language in-
tervention techniques during and after training was measured
in only 10 of the 18 studies. Without specific descriptions
of the parent training intervention and how it was actually
implemented, it is difficult to determine what specific charac-
teristics of the independent variable (i.e., parent training)
resulted in changes in the dependent variable (i.e., child lan-
guage outcomes). Thus, while the effects of parent-implemented
language interventions are positive, it is unclear (a) how
parent training is best accomplished and (b) which specific
language strategies or key intervention components are effec-
tive at improving language skills for children with language
impairments. Future research should include detailed descrip-
tions and direct measures of the parent training procedures. In
addition, systematic measurement of parents’use of inter-
vention strategies as an outcome of parent training procedures
is also necessary. Only when measures of parent training and
parent implementation are included in studies will it become
possible to examine particular parental characteristics (e.g.,
education or buy-in to the intervention) that may influence
parental use of language intervention strategies and child
language outcomes.
Second, while all of the studies reviewed included chil-
dren with language impairments, not all profiles of young
children with language impairments were represented in this
data set. The majority of studies included children with lan-
guage impairments and intellectual or developmental delays
or children who had expressive language delays and typical
intellectual abilities. Only one study (Law, Kot, & Barnett,
1999) included participants who had both receptive and ex-
pressive language impairments with no intellectual impair-
ment. In this study, there were negative effects for receptive
(g=–0.17) and expressive (g=–0.62) language. It is impor-
tant to note that this was the only study in which effect sizes
were calculated based on standard scores rather than raw
scores. Because standardized scores are adjusted for age and
designed to be stable estimates, differences in standard scores
over a period of 10 weeks would not be expected.
Third, it is not yet clear whether the effects of parent-
implemented intervention differ for receptive and expressive
child language outcomes. Across all studies in which parent-
implemented interventions were compared to a control group,
the average effect sizes for general receptive language skills
and general expressive language skills were statistically
different than zero (g=0.35 and 0.61, respectively). In gen-
eral, the effect sizes were larger for general expressive lan-
guage and smaller for receptive language in the majority of
studies. Because parent-implemented interventions focused
primarily on improving children’s expressive language skills,
these results may indicate that parent-implemented language
interventions might be strengthened by including a compo-
nent that specifically targets receptive language abilities. In
particular, the addition of strategies to improve receptive
skills might be needed for children who exhibit both receptive
and expressive language impairments.
Fourth, only four studies followed children over time
(Baxendale & Hesketh, 2003; Buschmann et al., 2008; Law
et al., 1999; van Balkom, Verhoeven, Van Weerdenburg, &
Stoep, 2010). In a 6-month follow-up, Buschmann et al.
(2008) found that effect sizes remained positive and signif-
icant over time, with an effect size of g=0.68 for expressive
vocabulary and g=0.63 for expressive morphosyntax at
the 6-month follow-up when compared to a control group.
Van Balkom et al. (2010) also found that language skills
maintained over time. The effect for mean length of utterance
was g= 0.12 at posttest and g= 0.21 at follow-up; the effect
size for receptive language was g=0.37 at posttest and
g=0.71 at follow-up when comparing a parent-implemented
intervention to a therapist-implemented intervention. The
Law et al. (1999) study that had negative effects for expressive
language at the posttest showed positive effects for expres-
sive language at the 6-month follow-up. General expressive
language ability, which had a negative effect size ( g=–0.46)
at posttest, had a positive effect size at the 6-month follow-up
assessment (g=0.37) when compared to a therapist-
implemented intervention group. However, receptive lan-
guage outcomes that had a negative effect size (g=–0.56)
at posttest remained negative at the 6-month follow-up as-
sessment (g=–0.25). Thus, it appears that while long-term
expressive language gains were made for children in the
parent-implemented intervention group, receptive language
changes over time were less than those made by the therapist-
implemented group.
Studying the longitudinal effects of parent-implemented
interventions is important for several reasons. One rationale
for parent-implemented interventions is that intervention
effects are expected to maintain over time. Findings from
Buschmann et al. (2008) and Law et al. (1999) support this
premise. As noted previously, Law et al. included children
with expressive and receptive language impairments, while
Buschmann et al. included younger children with only ex-
pressive language delays. The difference in population may
have contributed to differences in the long-term outcomes
in these studies. Additional research examining growth of
child language skills as well as parents’use of intervention
strategies over time is necessary to determine the long-term
effects of parent-implemented language interventions for
specific populations of children with language impairments.
Limitations of the Review
There are several limitations to this review. First, as men-
tioned previously, minimal information was provided about the
two critical components of parent-implemented intervention:
(a) parent training (e.g., how parents were trained and how
much training occurred) and (b) subsequent parental use of
language support strategies. This lack of information limits
the conclusions that may be drawn from this review. The
results of this review indicate that parent-implemented lan-
guage interventions resulted in positive language outcomes
for children with language impairments. The results are some-
what ambiguous as to which intervention components led
Roberts & Kaiser: Parent-Implemented Language Interventions 195

to increases in language skills. Thus, no recommendations
can be made regarding specific effective features of parent-
implemented language interventions.
Second, the results of a meta-analysis are only as strong as
the studies that are included in the analysis. The methodo-
logical rigor of the studies varied substantially, as reflected in
the study characteristics presented in Table 5. Seven studies
had sample sizes of less than 25 participants. Two studies
exhibited nonequivalence between groups on variables that
were not accounted for statistically in the outcome analysis.
This nonequivalence is not surprising given the small sample
sizes, but failure to statistically correct for nonequivalence
is problematic. These methodological limitations should be
considered when interpreting the results.
Third, in addition to the small sample sizes in individual
studies, this review included a total of only 18 studies. The
small number of studies restricts both the generalizability of
the findings and the number of methodological variables that
can be analyzed statistically to explore moderators of inter-
vention effects. Secondary analyses are further limited by the
range of outcome measures that were used across studies. The
number of studies contributing to any one analysis ranged
from two to 14. Analyses that included fewer than three stud-
ies should be interpreted with caution.
Fourth, the external validity of these results is limited be-
cause the socioeconomic status for the majority of the study
participants was middle class, and all parents volunteered
to participate in parent training and to learn the intervention
strategies. It is unclear whether the positive results of parent-
implemented language interventions would generalize to
parents and children from a lower socioeconomic status or to
parents who were provided parent training as part of routine
speech-language therapy. Additional research that includes
a more diverse sample of participants is needed to under-
stand for which populations of parents and children parent-
implemented language interventions is effective.
Implications for Practice
In addition to providing empirical support for the practice
of parent-implemented language interventions, several spe-
cific implications for practices related to parent-implemented
language interventions may be drawn from this review:
(a) interventions should focus on socially communicative
interactions between parents and children, ( b) parents should
be taught to increase their use of specific linguistic forms
through models and expansions, (c) parents should be trained
at home and across everyday routines, (d) parent-implemented
interventions may be effective for children with a range of
intellectual and language skills, and (e) training parents about
once per week may be sufficient to improve child language
outcomes. While specific parent training techniques were
not clearly described in this set of studies, general interven-
tion principles were common across the majority of parent-
implemented language interventions. The consistent use of
common principles across parent-implemented interventions
and the quantitative outcomes of this review together sup-
port these recommendations for practice.
Most of these experimental interventions included a social-
communication focus. The context for parent-implemented
language teaching was the interaction between the child and
the parent occurring during play, routines, and everyday
activities. The focus of many parent-implemented interven-
tions was enhancing turn-taking interactions to increase child
communicative initiations and amount of parent–child inter-
action. Increasing parent responsiveness to child communi-
cation was targeted in 10 of the 18 studies, and in the seven
studies that measured parent strategy use, there was clear
evidence of changes in parent responsiveness resulting from
training. In general, increasing parent–child turn taking in
interactions and improving parent responsiveness to commu-
nication appear to be associated with positive child language
outcomes.
Parent-implemented interventions focused on increasing
two types of linguistic input to the child: what the parent says
(i.e., language models) and how much the parent says. For
example, in focused stimulation, the child is exposed to mul-
tiple examples of a linguistic target form during a parent–
child interaction. In the Hanen Parent Program, parental
linguistic input is increased by expanding what the child says
to include more complex models of language contingent upon
child verbalizations. For example, if the child says, “Dog
eats,”the parent is instructed to say, “The dog eats the bone.”
Enhancing parental input by increasing exposures to lan-
guage targets and increasing complexity of the models pro-
vided to the child by expanding the child’s utterances appears
to have positive effects on children’s language development.
In some studies, parents were taught to use specific lan-
guage targets. From the results of this review, it is unclear if
teaching parents to use child-specific language targets (e.g.,
modeling specific words or word combinations) is more
effective than simply increasing parent responsiveness or
changing their rate of talking. The degree of individualization
of child language targets needed to affect child outcomes is
likely influenced by child characteristics (e.g., cognitive level
and language skills).
Several parent-implemented language interventions used
everyday routines that occurred in the child’s natural environ-
ment (usually the home) as a context for intervention. Over
half of the included studies (n= 11) included at least some
training in the home. By training parents in the home, it is
more likely that intervention will affect both the parent and
the child in their everyday interactions and that these effects
will generalize to other settings and maintain over time. Fu-
ture research is needed to test this hypothesis empirically.
Thus, these studies suggest that training occur at home when-
ever possible. The duration of parent training in the included
studies was 36 hr or less, with a mean of 23 hr and a range
of 9–36 hr. The majority of interventions were fewer than
26 hr, which is equivalent to 1 hr of parent training per week
for 6 months. This is a relatively small amount of direct
intervention with the parent and child given the magnitude
and consistency of the effects on child language outcomes.
These results indicate that parent-implemented language
intervention is an effective approach and may be an efficient
intervention for children with language impairments.
Parent-implemented interventions were used with parents
of children with varying degrees of cognitive and language
impairments. In general, parent-implemented interventions
were effective for children with and without ID. These results
196 American Journal of Speech-Language Pathology •Vol. 20 •180–199 •August 2011

indicate that parent-implemented interventions may be ef-
fective for several populations of children with language
impairments, including children with autism, developmental
delays, and expressive language delays.
Conclusion
The results of this review indicate that parent-implemented
language interventions are an effective early language inter-
vention for young children with language impairments from
middle-class families whose parents agreed to participate in
research studies. Parents can be taught to implement both
general and specific language support strategies. Parent-
implemented interventions appear to be associated with sub-
stantial language growth in young children with varying
degrees of cognitive and language impairments. The largest
effects on language skills were for expressive morphosyntax
for both comparisons. Children whose parents received
training used more complex language than children whose
parents did not receive this training. These results indicate
that even a small amount of parent training can have substan-
tial effects on children’s language development. Although
the exact mechanisms linking changes in parental language
support to children’s outcomes were not examined in these
studies, it appears that children made receptive and expressive
language gains when compared to a control group, as a result of
parents learning to use specific language intervention strate-
gies. These conclusions must be considered in the context of
the limitations of this meta-analysis and the limitations of the
individual studies.
Acknowledgments
This research was supported in part by Department of Education
Doctoral Leadership Training in Early Childhood Special Educa-
tion Grant H325D070075 and Institute of Education Sciences
Training Grant R305B080025. We would like to acknowledge
Mark Lipsey for his guidance regarding the use of meta-analytic
techniques, as well as Ragan McLeod for her work on reliability
coding.
References
References marked with an asterisk indicate studies included
in the meta-analysis.
*Aldred, C., Green, J., & Adams, C. (2004). A new social
communication intervention for children with autism: A pilot
randomised controlled treatment study suggests effectiveness.
Journal of Child Psychology and Psychiatry, 45, 1420–1430.
Alpert, C., & Kaiser, A. (1992). Training parents as milieu
language teachers. Journal of Early Intervention, 16, 31–52.
Alston, E., & St. James-Roberts, I. (2005). Home environments
of 10-month-old infants selected by the WILSTAAR screen
for pre-language difficulties. International Journal of Language
& Communication Disorders, 40, 123–136.
*Baxendale, J., & Hesketh, A. (2003). Comparison of the effec-
tiveness of the Hanen Parent Programme and traditional clinic
therapy. International Journal of Language & Communication
Disorders, 38, 397–415.
*Buschmann, A., Jooss, B., Rupp, A., Feldhusen, F., Pietz, J.,
& Philippi, H. (2008). Parent based language intervention for
2-year-old children with specific expressive language delay:
A randomized controlled trial. Archives of Disease in Childhood,
94, 110–116.
Cheseldine, S., & McConkey, R. (1979). Parental speech to young
Down’s syndrome children: An intervention study. American
Journal of Mental Deficiency, 83, 612–620.
Comprehensive Meta-Analysis (Version 2) [Computer software].
Englewood, NJ: Biostat.
Conti-Ramsden, G., Hutcheson, G., & Grove, J. (1995). Con-
tingency and breakdown: Children with SLI and their con-
versations with mothers and fathers. Journal of Speech and
Hearing Research, 38, 1290–1302.
Cooper, L., & Hedges, L. (2009). Research synthesis as a scien-
tific process. In H. Cooper, L. Hedges, & J. Valentine (Eds.),
The handbook of research synthesis and meta-analysis (2nd ed.,
pp. 3–16). New York, NY: Russell Sage Foundation.
Cooper, L., Hedges, L., & Valentine, J. (Eds.). (2009). The
handbook of research synthesis and meta-analysis (2nd ed.).
New York, NY: Russell Sage Foundation.
Dawson, G., Rogers, S., Munson, J., Smith, M., Winter, J.,
Greenson, J., . . . Varley, J. (2010). Randomized, controlled
trial of an intervention for toddlers with autism: The Early Start
Denver Model. Pediatrics, 125, e17–e23.
Delaney, E., & Kaiser, A. (2001). The effects of teaching parents
blended communication and behavior support strategies.
Behavioral Disorders, 26, 93–116.
*Drew, A., Baird, G., Baron-Cohen, S., Cox, A., Slonims, V.,
Wheelwright, S., . .. Charman, T. (2002). A pilot randomised
control of a parent training intervention for pre-school chil-
dren with autism: Preliminary findings and methodological
challenges. European Child & Adolescent Psychiatry, 11,
266–272.
Dunn, L. M., & Dunn, D. M. (1997). Peabody Picture Vocab-
ulary Test—III. Circle Pines, MN: AGS.
Dunst, C. (2002). Family-centered practices: Birth through high
school. Journal of Special Education, 26, 139–147.
Fenson, L., Dale, P. S., Reznick, J. S., Thal, D., Bates, E.,
Hartung, J. P., . . . Reilly, J. S. (1993). The MacArthur Com-
municative Development Inventories. San Diego, CA: Singular.
*Fey, M., Cleave, P., Long, S., & Hughes, D. (1993). Two
approaches to the facilitation of grammar in children with lan-
guage impairment. Journal of Speech and Hearing Research,
36, 141–157.
*Gibbard, D. (1994). Parental-based intervention with pre-school
language delayed children. European Journal of Disorders of
Communication, 29, 131–150.
*Gibbard, D., Coglan, L., & McDonald, J. (2004). Cost-
effectiveness analysis of current practice and parent intervention
for children under 3 years presenting with expressive language
delay. International Journal of Communication Disorders, 39,
229–244.
Gillet, J., & LeBlanc, L. (2007). Parent-implemented natural
language paradigm to increase language and play in children
with autism. Research in Autism Spectrum Disorders, 1,
247–255.
*Girolametto, L. (1988). Improving the social-conversational
skills of developmentally delayed children: An intervention
study. Journal of Speech and Hearing Disorders, 53, 156–167.
*Girolametto, L., Pearce, P., & Weitzman, E. (1996a). The
effects of focused stimulation for promoting vocabulary in
young children with delays: A pilot study. Journal of Children’s
Communication Development, 17, 39–49.
*Girolametto, L., Pearce, P., & Weitzman, E. (1996b). Inter-
active focused stimulation for toddlers with expressive vocab-
ulary delays. Journal of Speech and Hearing Research, 39,
1274–1283.
Roberts & Kaiser: Parent-Implemented Language Interventions 197

*Girolametto, L., Weitzman, E., & Clements-Baartman, J.
(1998). Vocabulary intervention for children with Down syndrome:
Parent training using focused stimulation. Infant-Toddler Inter-
vention: The Transdisciplinary Journal, 8, 109–125.
Girolametto, L., Weitzman, E., Wiigs, M., & Pearce, P. (1999).
The relationship between maternal language measures and
language development in toddlers with expressive vocabulary
delays. American Journal of Speech-Language Pathology, 8,
364–374.
*Green, J., Charman,T.,McConachie, H., Aldred, C., Slonims,
V., Howlin, P., .. . Pickles, A. (2010). Parent-mediated
communication-focused treatment in children with autism (PACT):
A randomized controlled trial. The Lancet, 375, 2152–2160.
Grimm, H. (2000). Sprachentwicklungstest für zweijährige Kinder
(SETK-2) [Developmental language test for 2-year-old chil-
dren]. Göttingen, Germany: Hogrefe.
Hammer, C., Tomblin, B., Zhang, X., & Weiss, A. (2001).
Relationships between parenting behaviors and specific lan-
guage impairment in children. International Journal of
Language & Communication Disorders, 36, 185–205.
Hancock, T., Kaiser, A., & Delaney, E. (2002). Teaching parents
of preschoolers at high risk: Strategies to support language and
positive behavior. Topics in Early Childhood Special Education,
22, 191–212.
Harris, M., Jones, D., Brookes, S., & Grant, J. (1986). Rela-
tions between the non-verbal context of maternal speech and rate
of language development. British Journal of Developmental
Psychology, 4, 261–268.
Hart, B., & Risley, T. (1995). Meaningful differences in the every-
day experience of young American children. Baltimore, MD:
Brookes.
Hedges, L. (1981). Distribution theory for Glass’s estimator of
effect size and related estimators. Journal of Educational
Statistics, 6, 107–128.
Hedges, L. (2009). Statistical considerations. In H. Cooper, L.
Hedges, & J. Valentine (Eds.), The handbook of research syn-
thesis and meta-analysis (2nd ed., pp. 37–47). New York, NY:
Russell Sage Foundation.
Hedrick, D.L., Prather, E. M., & Tobin, A. R. (1984). Sequenced
Inventory of Communication Development—Revised.Seattle:
University of Washington Press.
Hemmeter, M., & Kaiser, A. (1994). Enhanced milieu teaching:
Effects of parent-implemented language intervention. Journal
of Early Intervention, 18, 269–289.
Hoff, E., & Naigles, L. (2002). How children use input to acquire
a lexicon. Child Development, 73, 418–433.
Huttenlocher, J., Haight, W., Bryk, A., Seltzer, M., & Lyons, T.
(1991). Early vocabulary growth: Relation to language input and
gender. Developmental Psychology, 27, 236–248.
Individuals with Disabilities Education Improvement Act of
2004, Pub. L. No. 108-446, §118 Stat. 2647 (2004).
Kaiser, A. (1993). Parent-implemented language intervention:
An environmental system perspective. In A. Kaiser & D. Gray
(Eds.), Enhancing children’s communication: Research foun-
dations for intervention (Vol. 2, pp. 63–84). Baltimore, MD:
Brookes.
*Kaiser, A., & Hancock, T. (1998). The effects of parent-
implemented language intervention on mentally retarded
children’s communication development. Unpublished manu-
script, Department of Special Education, Vanderbilt University,
Nashville, TN.
Kaiser, A., Hancock, T., & Hester, P. (1998). Parents as cointer-
ventionists: Research on applications of naturalistic language
teaching procedures. Infants and Young Children, 10(4), 46–55.
Kashinath, S., Woods, J., & Goldstein, H. (2006). Enhancing
generalized teaching strategy use in daily routines by parents
of children with autism. Journal of Speech, Language, and
Hearing Research, 49, 466–485.
Law, J., Garrett, Z., & Nye, C. (2004). The efficacy of treatment
for children with developmental speech and language delay/
disorder: A meta-analysis. Journal of Speech, Language, and
Hearing Research, 47, 924–943.
*Law, J., Kot, A., & Barnett, G. (1999). A comparison of two
methods of providing intervention to three year old children with
expressive/receptive language impairment. London, England:
City University London.
Lipsey, M., & Wilson, D. (2001). Practical meta-analysis.
Thousand Oaks, CA: Sage.
Mahoney, G., & Powell, A. (1988). Modifying parent–child inter-
action: Enhancing the development of handicapped children.
Journal of Special Education, 22, 82–96.
Manolson, A. (1992). It takes two to talk: A parent’s guide to
helping children communicate (3rd ed.). Toronto, Ontario,
Canada: The Hanen Centre.
McConachie, H., & Diggle, T. (2007). Parent implemented early
intervention for young children with autism spectrum disorder: A
systematic review. Journal of Evaluation in Clinical Practice,
13, 120–129.
*McDade, A., & McCartan, P. (1998). Partnership with parents:
A pilot project. International Journal of Language & Commu-
nication Disorders, 33, 556–561.
Meadan, H., Ostrosky, M., Zaghlawan, H., & Yu, S. (2009).
Promoting the social and communicative behavior of young
children with autism spectrum disorders. Topics in Early Child-
hood Special Education, 29, 90–104.
Morris, S. (2008). Estimating effect sizes from pretest-posttest-
control group designs. Organizational Research Methods, 11,
364–386.
Mundy, P., Kasari, C., Sigman, M., & Ruskin, E. (1995). Non-
verbal communication and early language acquisition in
children with Down syndrome and in normally developing
children. Journal of Speech and Hearing Research, 38,
157–167.
Price, G. (1984). Mnemonic support and curriculum selection in
teaching by mothers: A conjoint effect. Child Development, 55,
659–668.
Reynell, J., & Gruber, C. (1990). Reynell Developmental Lan-
guage Scales, U.S. Edition. Los Angeles, CA: Western Psycho-
logical Services.
Rice, M., Sell, M., & Hadley, P. (1991). Social interactions of
speech- and language-impaired children. Journal of Speech and
Hearing Research, 34, 1299–1307.
Rowe, M. (2008). Child-directed speech: Relation to socioeconomic
status, knowledge of child development and child vocabulary
skill. Journal of Child Language, 35, 185–205.
Schreibman, L., & Koegel, R. L. (2005). Training for parents
of children with autism: Pivotal responses, generalization,
and individualization of interventions. In E. Hibbs & P. Jensen
(Eds.), Psychosocial treatment for child and adolescent disor-
ders: Empirically based strategies for clinical practice (2nd ed.,
pp. 605–631). Washington, DC: American Psychological
Association.
Smith, K., Landry, S., & Swank, P. (2000). Does the content
of mothers’verbal stimulation explain differences in children’s
development of verbal and nonverbal cognitive skills? Journal
of School Psychology, 38, 27–49.
Sparrow, S., Balla, D., & Cicchetti, D. (1984). Vineland Adap-
tive Behavior Scales: Expanded manual. Circle Pines, MN:
AGS.
Tamis-LeMonda, C., Bornstein, M., & Baumwell, L. (2001).
Maternal responsiveness and children’s achievement of lan-
guage milestones. Child Development, 72, 748–767.
198 American Journal of Speech-Language Pathology •Vol. 20 •180–199 •August 2011

*Tannock, R., Girolametto, L., & Siegel, L. (1992). Language
intervention with children who have developmental delays:
Effects of an interactive approach. American Journal on Mental
Retardation, 97, 145–160.
Tomasello, M., & Todd, J. (1983). Joint attention and lexical
acquisition style. First Language, 4, 197–212.
*Van Balkom, H., Verhoeven, L., Van Weerdenburg, M., &
Stoep, J. (2010). Effects of parent-based video home training in
children with developmental language delay. Child Language
Teaching and Therapy, 26, 221–237.
Vigil, D., Hodges, J., & Klee, T. (2005). Quantity and quality
of parental language input to late-talking toddlers during play.
Child Language Teaching and Therapy, 21, 107–122.
Weistuch, L., & Lewis, M. (1985). The language interaction inter-
vention project. Analysis and Intervention in Developmental
Disabilities, 5, 97–106.
Weizman, Z., & Snow, C. (2001). Lexical input as related to
children’s vocabulary acquisition: Effects of sophisticated
exposure and support for meaning. Developmental Psychology,
37, 265–279.
Wetherby, A., Prizant, B., & Hutchinson, T. (1998). Commu-
nicative, social/affective, and symbolic profiles of young
children with autism and pervasive developmental disorders.
American Journal of Speech-Language Pathology, 7(2), 79–91.
Wilson, S. (2009). Systematic coding. In H. Cooper, L. Hedges, &
J. Valentine (Eds.), The handbook of research synthesis and
meta-analysis (2nd ed., pp. 159–176). New York, NY: Russell
Sage Foundation.
Wulbert, M., Inglis, S., Kriegsmann, E., & Mills, B. (1975).
Language delay and associated mother-child interactions.
Developmental Psychology, 11, 61–67.
Yoder, P., McCathren, R., Warren, S., & Watson, A. (2001).
Important distinctions in measuring maternal responses to
communication in prelinguistic children with disabilities.
Communication Disorders Quarterly, 22, 135–147.
Zimmerman, I. L., Steiner, V. G., & Pond, R. E. (1992). Pre-
school Language Scale–3. San Antonio, TX: The Psychological
Corporation.
Roberts & Kaiser: Parent-Implemented Language Interventions 199

DOI: 10.1044/1058-0360(2011/10-0055)
2011;20;180-199; originally published online Apr 8, 2011; Am J Speech Lang Pathol
Megan Y. Roberts, and Ann P. Kaiser
Meta-Analysis
The Effectiveness of Parent-Implemented Language Interventions: A
http://ajslp.asha.org/cgi/content/full/20/3/180#BIBLaccess for free at: The references for this article include 16 HighWire-hosted articles which you can
This information is current as of August 19, 2011
http://ajslp.asha.org/cgi/content/full/20/3/180
located on the World Wide Web at:
This article, along with updated information and services, is