Dose frequency: Comparison of language outcomes in preschool children with language impairment

Abstract

Data regarding treatment intensity provide evidence for service delivery decision-making in schools. In this pilot study, dose frequency effects (i.e. number of therapy sessions per week) on semantic and morphologic abilities in preschool children, ages 4;0 to 5;3 years of age, with language impairment were examined. Children enrolled in a concentrated treatment schedule (4 times per week for 6 weeks) were compared to children enrolled in a distributed treatment schedule (2 times per week for 12 weeks) using a pretest/posttest design. A concentrated schedule provides more therapy per week for fewer weeks than a distributed schedule. Therapy following a scaffolded-language structure with implicit strategies was used. Results indicate positive outcomes following both a concentrated and distributed schedules. No differences between groups were noted. Results suggest goal attainment of semantic and morphologic aspects of language within a school setting could occur in a variety of scheduling frameworks.

Keywords

language impairment language therapy preschool children service delivery methods treatment intensity

I Introduction

School-based speech and language therapists (SLTs) are required to consider evidence-based practice for service delivery decision-making (American Speech Hearing Association, 2010; Individuals with Disabilities Education Act, 2004). Among the decisions is level of therapy intensity (e.g. dose frequency, dose, total amount of therapy, dose form), which is key for determining therapy effects (Proctor-Williams, 2009; Warren et al., 2007). Few studies examine effects of therapy intensity in public school settings. Cirrin et al. (2010) and Schooling et al. (2010) concluded that there is insufficient evidence regarding differential effects of therapy intensity for preschool and elementary school-age children. Such data will provide evidence from which to clarify service delivery decision-making of school-based practitioners. The purpose of this pilot study was to compare weekly therapy scheduling (i.e. dose frequency) on semantic and morphologic abilities in preschool children with language impairment (LI) with other variables (i.e. dose, dose form, total amount of therapy) remaining constant.

1 Dosage of therapy

Dose frequency refers to the number of times therapy is delivered within a given time period (Warren et al., 2007). Some school-based speech and language therapists (SLTs) implement therapy that is distributed over a longer period of time, while others implement therapy within concentrated weekly schedules (Jellison, 2007; Ukrainetz et al., 2009). Dose frequency decisions are often determined by workload constraints rather than child outcomes (Brandel and Loeb, 2011). This may be a consequence of limited studies available to guide clinical decision-making for children with language impairment (LI). Only one study compares dose frequency with preschool children with LI. Barratt et al. (1992) compared expressive and receptive language outcomes following either a concentrated or distributed schedule of therapy. Within each condition children received 24 play-based therapy sessions across a 6-month time period. The concentrated schedule consisted of 4 sessions per week for 3 weeks per month. The distributed schedule consisted of 1 session per week. Significantly greater expressive language gains occurred following the concentrated schedule than the distributed schedule.

The time period of therapy and total amount of therapy may provide clarity into dosage effects. This is critical since therapy that continues for too long will require more time and money than may be necessary for producing an adequate effect. Nye et al. (1987) assessed effectiveness of language therapy continuing for an average of 38 weeks using meta-analysis. They found significant progress on some language abilities in children with LI. Therapy length of 4 to 12 weeks produced the greatest effect size. Lower effect sizes were noted following 13 weeks or more of therapy. Law et al. (2004) reached tentative conclusions following a meta-analysis of therapy effects. Expressive syntax therapy lasting longer than 8 weeks had larger effect sizes than those less than 8 weeks. Jacoby et al. (2002) examined effects of total therapy units (i.e. 15-minutes) on children age 3 to 6 years of age receiving speech and language services in a hospital setting. Functional communication in children with spoken language impairment improved as treatment units increased. The lower the language abilities at therapy initiation, the more units were needed for improvement.

Dosage effects may be related to frequency of application of clinical strategies during a single therapy session (i.e. dose). Dose in morphosyntax therapy is typically investigated (Proctor-Williams, 2009). Proctor-Williams et al. (2001) suggest children with specific language impairment do not benefit from a sparse number of clinical strategies. Leonard et al. (2004) suggest that .8 per minute may not be sufficient application of clinical strategies for meaningful change in children with specific language impairment. Proctor-Williams and Fey (2007) indicated sessions providing clinical strategies at .5 per minute were not enough to benefit children with specific language impairment. On the other hand, Nelson et al. (1996) reported 1.42–1.56 strategy per minute as a potentially appropriate level for change to occur. Results suggest higher dose may be better than lower dose. However, the relationship between dose and other dosage factors is not explicated.

Studies with children at risk for language and literacy difficulties lend insight into dose and dose frequency outcomes and interactions. Ukrainetz et al. (2009) exposed small groups of kindergarteners to an explicit phonological awareness skill instruction at differing levels of dose frequency. Effects on phonological awareness abilities were found following short, intense therapy (24 sessions across 8 weeks) and extended therapy (24 sessions across 24 weeks). McGinty et al. (2011) compared print knowledge outcomes of preschool children participating in a 30-week classroom-based print-referencing intervention at low (4 times per week) and high (2 times per week) dose frequency. Overall, high dose frequency was more effective than low dose frequency. However, dose frequency effects were mitigated by the frequency of instruction per session (i.e. dose). McGinty et al. (2011) concluded if the dose per session is high, low frequency of sessions per week is almost as effective as high frequency of session per week.

2 Dose form

Type of therapy approach (i.e. dose form) may play a role in dosage effects. In this pilot study, language therapy that follows a scaffolded-language structure with implicit clinical strategies (i.e. cloze procedures, expansions, models) was selected for two reasons. First, dose occurs with high frequency (i.e. 9 to 11 times per minute), which may influence dosage effects (Bellon-Harn and Harn, 2010; Bellon-Harn et al., 2012). This rate of delivery within a single session is higher than previous reports of dose within morphosyntax therapy. Differences may be due to the scaffolded-language structure. Therapy is characterized as an on-going interaction in which the clinician responds to the proposition of the preceding child utterance using cloze procedures, expansions, or models. Reciprocal dialogue builds. Adults adjust the complexity level of their language using expansions, cloze procedures, and models according to child initiations and responses (Bruner, 1978; Kirchner and Skarakis-Doyle, 1983; Nelson, 1985).

Second, the therapy addresses semantic and morphosyntactic abilities within a single treatment session. This is critical since preschool children with language impairment often exhibit deficits in morphosyntactic and semantic domains concomitantly (Brackenbury and Pye, 2005; McGregor et al., 2002; Sheng and McGregor, 2010). Strategies are not used to explicitly target a morphosyntactic form or semantic relationship. Instead, strategies implicitly provide an opportunity for a child to use more complex semantic or morphosyntactic language. Cloze procedures, expansions, and models provide a linguistic context for a child to use complex semantic and/or morphosyntactic language (Kouri, 2005; Norris and Hoffman, 1990; Snow et al., 1984; Yoder et al., 1995). Collective application of these strategies within a scaffolded-language structure resulted in positive change in semantic and morphosyntactic language in preschool children with LI.

Bradshaw et al. (1998) compared a scaffolded-language condition and a control condition using scripted wh-questions with preschool children with LI. Results indicated that number of utterances at high levels of semantic complexity (i.e. interpretive) was greater in the scaffolded-language than the control condition. Even though morphosyntactic forms were not targeted, participants in the scaffolded-language condition increased use of noun and verb phrases as well as past and future verb tenses. Bellon-Harn et al. (2004) noted positive outcomes following a scaffolded-language therapy with preschool children with language and phonological impairment. An association between child use of language at high levels of semantic complexity (i.e. interpretive) and cloze procedures and expansions was identified. Therapy in both studies occurred across 4 weeks, 3 times per week, totaling 12 sessions. Sessions were 20 minutes in duration.

Bellon-Harn and Harn (2010) compared language outcomes of 16 preschool children with LI speaking non-mainstream dialect (African-American English or Southern White English) enrolled in therapy following either a scaffolded-language structure with implicit strategies or an additive structure with explicit strategies. Verbal imitation and wh-questions related to picture cards representing semantic and morphosyntactic targets comprised the additive therapy (Connell and Stone, 1992; Gillum et al., 2003). Therapy occurred across 14 weeks, 3 times per week, totaling 42 sessions. Greater use of interpretations occurred at post- than pre-therapy with large effect sizes for the scaffolded (g = 2.78) and additive (g = 1.09) therapy groups. Effects were greater for the scaffolded therapy group than the additive therapy group (g = 2.78). Mean length of utterance in morphemes (MLU_m) increased from pre- to post-therapy with large effect sizes for both scaffolded (g = 1.82) and additive (g = 1.59) therapy groups. Copula (is/are), auxiliary (is/are), and third person singular were examined for frequency of use since these structures are identified as being used at lower rates in children with language impairment speaking nonmainstream dialect than their same dialect-speaking peers (Craig and Washington, 2000; Garrity and Oetting, 2010; Oetting and McDonald, 2001). As such, these forms may best typify outcomes. Results indicated greater frequency of each in post- than pre-therapy samples for both groups with large effect sizes (g = 1.45 to 3.30).

This pilot study compares dose frequency conditions using a high dose therapy. Children enrolled in a concentrated therapy schedule (4 times per week for 6 weeks) were compared to children enrolled in a distributed therapy schedule (2 times per week for 12 weeks). Total therapy amount, dose form, and dose were the same in each condition. Results will provide initial data for determining dosage in public school settings. A pretest/posttest design was used to answer the following questions:

Will children demonstrate increases in morphologic and semantic language from pre-therapy to post-therapy measures?

Will children participating in the concentrated versus distributed schedule differ on pre-therapy to post-therapy measures?

II Method

1 Participants

Twelve African American English (AAE) speaking children with LI between the ages of 4;0 and 5;3 years participated. Children were referred by parents, teachers, and related service providers based on impressions of language status in comparison with peers following approval from the university institutional review board. Participants were recruited from Head Start preschool classes within a single school. All children were participating in full-day programs. All assessment and intervention activities were implemented in the public school milieu. Children were not previously enrolled in speech services. Criteria for enrollment in the treatment included the following:

English as the primary language;

no known oral-motor, hearing, psychiatric or neurological impairments;

no enrollment in additional language intervention outside of the current project;

no evidence of speech impairment based on performance above the 16th percentile on the phonology subtest of the Diagnostic Evaluation of Language Variation: Norm-referenced Test (DELV-NR; Seymour et al., 2005);

above 1 standard deviation on the Columbia Mental Maturity Scale (CMMS; Burgemeister et al., 1972);

below 1 standard deviation on the language composite score of the Diagnostic Evaluation of Language Variation: Norm-referenced Test;

below 1 standard deviation on the Structured Photographic Expressive Language Test: 3 (SPELT-3; Dawson et al., 2003); and

survey documentation from teachers and/or parents of impressions of language status in comparison with peers.

Children were randomly assigned to either the concentrated or distributed treatment group using random number assignment. Groups did not differ on language standard scores, nonverbal IQ standard scores, or phonology percentile scores (see Table 1). Phonological ability to use word-final grammatical morphemes as determined by use of word-final /s, z, t, d/ on sound-in-word subtest from the Goldman–Fristoe Test of Articulation: 2 (Goldman and Fristoe, 2000) was verified. Three females and three males were assigned to the concentrated group. Two females and four males were assigned to the distributed group.

Table 1

Selection criteria and pre-therapy measures of children in either condition

Measure	Concentrated (n = 6): M (SD)	Dispersed (n = 6): M (SD)	z	p
Age	4.61 (.48)	4.61 (.37)	–.08	.93
DELV-NR (standard scores) (language composite)	72.16 (8.86)	76.33 (4.13)	–.98	.32
DELV-NR (percentiles) (phonology subtest)	37.50 (10.36)	30.16 (12.08)	–.56	.51
SPELT-3 (standard scores)	68.33 (10.40)	70.16 (12.10)	–.16	.86
CMMS (standard scores)	100.00 (11.18)	95.33 (7.44)	–.88	.37
Interpretive utterances	12.16 (11.78)	23.83 (9.13)	–1.76	.10
MLU_m	4.19 (1.06)	5.03 (.54)	–1.76	.08
Auxiliary (probe)	.00 (.01)	.02 (.03)	–.73	.46
Auxiliary (sample)	.05 (.09)	.02 (.03)	–.38	.70
Copula (probe)	.09 (.10)	.03 (.03)	–1.35	.17
Copula (sample)	.32 (.30)	.20 (.31)	–.74	.45
Third person (probe)	.01 (.04)	.01 (.04)	.00	1.00

Notes: DELV-NR = Diagnostic Evaluation of Language Variation: Norm-referenced Test; SPELT-3 = Structured Photographic Expressive Language Test: 3; CMMS = Columbia Mental Maturity Scale

Five graduate student clinicians – hereafter referred to as clinician(s) – in a university program of speech-language pathology participated in the pre-therapy assessment, therapy, and pre- and post-therapy data collection under the supervision of the author. Each completed coursework and clinical practicum in child language assessment, treatment, and clinical processes prior to participation in this study.

2 Therapy procedures

Clinicians were trained in the therapy protocol via seminars and observations. Following training, each practiced using the therapy with children not participating in the study. The author evaluated application of clinical strategies. Each clinician met criteria of 90% for accurate use of clinical strategies prior to participation in this study. Each clinician provided individual therapy. Each was assigned two or three children. Assignment was based on availability of the child and clinician.

Therapy sessions were 20 minutes in length. A session was initiated with the introduction of the book without adult verbal prompts. If the child did not produce a gesture or verbal utterance relative to the book, the clinician asked a wh-question (see Appendix 1). After the child’s conversational turn, the clinician judged the child’s initiation as semantically appropriate or inappropriate. If judged appropriate, the clinician made an ‘on-line’ decision to provide more information, use an expansion with a cloze procedure, or use a cloze procedure only at a higher level of semantic complexity. If judged inappropriate, the clinician made an ‘on-line’ decision to provide more information, use a model with a cloze procedure, or use a cloze procedure only at a lower level of semantic complexity. Expansions typically were not used if the child’s utterance was judged as inappropriate. Models typically were not used if the child’s utterance is judged as appropriate.

Storybooks served as stimuli. Books used included narrative structures characterized as reactive or abbreviated sequences (Applebee, 1978). In a reactive sequence, each page contains one action or event; however, effects of one action on the next action are not planned. In an abbreviated sequence, each page contains one action or event; however, an underlying plan or intent is inherent in the narrative. Books that include relationships among characters and events provide opportunities to use and hear complex semantic and morphologic language. Storybooks provided stimulating pictures with limited text depicting familiar events. One storybook per three sessions was used. Each storybook was approximately 15 pages in length, including the cover and title page. At the first session the initial five pages were introduced. Each subsequent session included a review of the previous pages and introduced the next five pages. During the third session, the entire storybook was completed. For a list of the books, see Table 2.

Table 2

Therapy storybooks

Storybook title
Mrs. Wishy Washy (Cowley, 1990a)
Wishy Washy Day (Cowley, 1998a)
Splishy Sploshy (Cowley, 1998b)
Mr. Wishy Washy (Cowley, 1998c)
The Hungry Giant (Cowley, 1990b)
The Scrubbing Machine (Cowley, 1998d)
The Red Rose (Cowley, 1990c)
The Grumpy Elephant (Cowley, 1998e)

3 Therapy fidelity

Five-minute samples from 25% of the therapy sessions per child were randomly selected (6 sessions per child; total of 72 transcripts). Clinician and child utterances from each session were transcribed and analysed using the Codes for Human Analysis of Transcripts (CHAT) (MacWhinney, 2000a) and Child Language Analysis (CLAN) computer programs (MacWhinney, 2000b). Clinician utterances were coded according to implicit strategy type. Frequency of cloze procedure, expansion, and model per 5-minute sample was calculated using the FREQ program from CLAN. FREQ is a program that provides a frequency count. Total number of each strategy within 5-minute samples per clinician–child dyad was divided by the total number of minutes (i.e. 5). Average number of cloze procedure, expansion, or model used per minute ranged between 7 and 13 during each sampled session for all children (see Figure 1). Total number of each strategy type used for each child was calculated and divided by the total number of strategies. Percent occurrence of cloze procedure, expansion, or model was calculated (see Figure 2). Cloze procedures and expansions were used more frequently than models for 7 of the 12 children. Models were used more frequently than cloze procedures or expansions for two. Of the remaining children, scaffolds were used at comparable amounts for one child, and models were used slightly more often than expansion for two children. Although minor differences in frequency of each strategy type were noted, all strategies were delivered collectively with consistent rates.

Figure 1

Average number of strategies per minute per sampled session for each child

Figure 2

Average number of cloze procedure, expansion, or model used in 25% of sessions for each child

4 Outcome measures

Percent occurrence of interpretive utterances used in a language sample provided the semantic outcome measure. An utterance was considered an interpretation if the child expressed concepts that were not clearly represented; rather the utterance included abstract concepts such as goals, states, qualities, or changes (e.g. The mom is mad because the children are dirty). The following criteria were used:

The utterance served to predict an outcome or related two events (e.g. He was still making noise while everybody was asleep; So the cow came home and he was (a)bout to cry.).

The utterance expressed an emotional state of the character (e.g. Then the mouse ran and the cat was scared).

The child stated what a character would say from the perspective of character (e.g. The big hungry giant say the cake was ready). If the child stated that an animal produced an animal sound the utterance was identified as descriptive.

The child described a character’s intent or mental state (e.g. The cow is trying to look at him in the window; She probably do-not want her clothes wet).

Number of interpretive utterances was calculated using the FREQ program from CLAN. Percent occurrence was calculated.

MLU_m in a language sample provided an overall morphologic measure. MLU_m was calculated using the MLU program from CLAN. Other morphologic measures included frequency of copula (is/are), auxiliary (is/are), and third person singular in language samples and controlled elicited probes. Total number of each in a language sample was calculated and divided by total number of obligatory contexts. Correct responses to probes were counted and percent correct was calculated.

5 Data collection

Pre- and post-therapy outcome measures were obtained from language samples and controlled elicited probes. Clinicians were trained to collect language samples and administer probes. Clinicians collected data on children that were not part of their therapy assignment. Probes and samples were collected in rooms used for assessment and therapy at the child’s school. Samples were obtained by presenting the child with a storybook not used in treatment. Different storybooks were used in pre- and post-therapy sample collection. The clinician instructed the child to ‘Tell the story in the book.’ The clinician used conversational markers (i.e. mm-hm, oh, really). If the child did not initiate the task, the clinician prompted the child with a broad wh-question (e.g. What is happening in this story?) Digital flash audio recorders (Marantz Model PMD 660) were used to record the samples. Probe responses were obtained by presenting children with photographic cards depicting actions and attributes representing grammatical morphemes of interest. Probes consisted of five sets of 12 photographic cards per morpheme. Clinicians asked wh-questions designed to elicit the grammatical morpheme of interest.

6 Data transcription and coding

Language samples and probes were transcribed and analysed using CHAT and CLAN. Language sample length ranged from 65 to 75 utterances. Fifty to 100 child utterances are considered adequate when there are other data collection methods (Cole et al., 1989; Owens, 2009). Only story-related utterances were included to control for types of utterances analysed. An utterance comprised a complete thought. Words that were divided by a pause of approximately 2 seconds, but had a complete thought were segmented as separate utterances. Fillers and word or phrase repetitions were omitted from the utterance for final analysis. Long strings of phrases or clauses with the word and were segmented as separate utterances.

Three teams of two clinicians did transcription and coding. Each participated in training and practiced on samples of children not participating in the study. Agreement of 90% or above between the clinician’s and author’s transcription and coding of practice samples was required prior to participation in this study. Clinicians evaluated data from children that were not in their therapy assignment. First, each member of the team transcribed or coded individually. Then, they compared samples and probes. If agreement could not be reached on an utterance, a third judge evaluated the utterance for a final decision.

7 Transcription and coding reliability

Utterance by utterance transcription and coding of the team and a third individual not participating in this study were compared on 20% of the utterances in the language samples and probes. Reliability was determined using the following formula: agreements divided by agreements plus disagreements multiplied by 100. Inter-observer agreement on transcription ranged from 89%–98% for language samples and 99%–100% for probes. Inter-observer agreement ranged between 94%–100% for interpretive utterances, 97%–99% for each morphologic structure in the sample, 93%–96% for identification of obligatory contexts, and 99%–100% for each morphologic structure in the probe.

8 Data analysis

Given the small number of participants in the present study, nonparametric statistics were used. Pre- and post-therapy outcomes from children in both groups combined were compared using the Wilcoxon nonparametric test. Differences in pre- and post-therapy gains from the concentrated and distributed groups were compared using the nonparametric Mann–Whitney U test. A standardized effect size measurement was used. Hedge’s g was calculated because it accounts for the small sample size and approximates the value of Cohen’s d statistic (Devilly, 2004; Meline, 2010). Effect size values of small, medium, and large are given as 0.2, 0.5, and 0.8, respectively (Cohen, 1988).

No differences between the concentrated and distributed groups were indicated on pre-therapy outcome measures (see Table 1). Only 7 of the 12 children produced contexts during both pre- and post-therapy samples in which third person singular was obligatory. Therefore, no comparison between groups was completed on this measure.

III Results

In this pilot study, preschool children with LI participated in a scaffolded-language intervention with implicit strategies following either a concentrated treatment schedule or a distributed treatment schedule. Semantic and morphosyntactic outcome measures were compared. Differences between pre- and post-therapy measures for all children participating were analysed. Differences in gains for children in each dose frequency group were analysed.

Children used a significantly greater percent of utterances with interpretive language and greater MLU_m during post- than pre-therapy language sample times (see Table 3). Pre- and post-therapy comparison of copula and auxiliary use in language samples indicates a significant difference. Of the samples analysed, results indicate children did not differ in frequency of use of third person singular from pre- to post-therapy. Percent occurrence of auxiliary, copula, and third person singular during the probe indicate significant differences between pre- and post-therapy levels. No differences were noted in gains between the concentrated and distributed groups on any outcome measure (see Table 4).

Table 3

Means and standard deviations of percent occurrence of interpretive utterances, MLUm, frequency of copula, auxiliary, third person singular during language samples and probes from pre-therapy to post-therapy for all children

Measure	Pre-treatment (n = 12): M (SD)	Post-treatment (n = 12): M (SD)	z	p	g
Interpretive utterances	18.00 (11.75)	42.33 (19.70)	–2.66	.008	1.38
MLU_m	4.61 (.93)	5.26 (.87)	–2.4	.01	.66
Auxiliary (probe)	.01 (.02)	.23 (.26)	–2.43	.01	1.10
Auxiliary (sample)	.03 (.07)	.34 (.23)	–2.66	.008	1.68
Copula (probe)	.06 (.07)	.44 (.31)	–3.06	.002	1.56
Copula (sample)	.26 (.30)	.61 (.30)	–1.96	.05	1.07
Third person (probe)	.01 (.03)	.07 (.09)	–2.22	.02	.82
Third person (sample)	.00 (.00)	.21 (.33)	–1.82	.07

Notes: Nonparametric Wilcoxon test was used. Significant at the p < .05; g < .2 = small; g < .5 = medium; g < .8 = large

Table 4

Gains of concentrated and distributed groups for percent occurrence of interpretive utterances, MLUm, frequency of copula, auxiliary, third person singular during language samples and probes from pre-therapy to post-therapy

Measure	Concentrated (n = 6)	Distributed (n = 6)	z	p
Interpretive utterance	29.28 (25.85)	20.50 (25.80)	–.80	.42
MLU_m	.94 (.92)	.37 (.44)	–1.44	.15
Auxiliary probe gain	.13 (.22)	.30 (.27)	–1.37	.17
Auxiliary sample gain	.26 (.23)	.33 (.30)	–.64	.52
Copula probe gain	.29 (.22)	.41 (.24)	–.72	.48
Copula sample gain	.21 (.49)	.51 (.36)	–1.20	.23
Third person probe gain	.06 (.08)	.04 (.06)	–.25	.79

Notes: Mann–Whitney U test was used. Significant at the p < .05

IV Discussion

Data regarding dosage effects provide evidence for service delivery decision-making. In this pilot study, dose frequency effects on semantic and morphologic abilities in preschool children with LI were examined. Therapy following a scaffolded-language structure with implicit strategies was used. Results indicate positive outcomes following both a concentrated and distributed schedules. No group differences were identified.

1 Dosage of therapy

Outcomes are not consistent with Barratt et al. (1992) in which expressive language outcomes were greater following a concentrated schedule. Both this study and Barratt et al. (1992) included 24 sessions; however, therapy in Barratt et al. (1992) spanned 6 months as compared to 6 or 12 weeks in this study. Additionally, Barratt et al. (1992) included a sample size of 42 children. Subtle group differences in this study may have emerged with a larger sample size.

Studies examining dose in morphosyntax therapy indicate higher frequency of clinical strategy application may be better than lower frequency. The therapy used in this study applied implicit clinical strategies to facilitate semantic and morphosyntactic language at a higher dose than reported in morphosyntax therapy studies. High frequency of implicit strategies within a scaffolded-language structure may play a unique role in facilitating semantic content and morphologic structures in children with LI. However, interpretation of the influence of dose is limited since comparison with a low dose therapy and interaction effects between dose and dose frequency was not included. Further, in a scaffolded-language therapy, there is no predetermined script or target. As such, questions remain about the frequency of linguistic forms within each cloze procedure, expansion, and model.

Studies including children at risk for language and literacy difficulties indicate greater outcomes following high than low weekly frequency; however, low weekly frequency may be as effective as high if the dose per session is high. It is difficult to generalize among Ukrainetz et al., McGinty et al., and this study due to different participant characteristics (impaired vs. at risk), dose form (intervention vs. instruction), and service-delivery of instruction (individual vs. group). Some inferences can be made, albeit cautiously. The therapy in this study was high dose per session. As such, dose per session may have contributed to positive language outcomes regardless of weekly intensity. However, detecting differences in dose frequency may have been limited since both groups received high quality and intensive support.

Nye et al. (1987) found effects on some language abilities following 38 weeks of therapy. Greater effect sizes were noted following 4 to 12 weeks of therapy, and lower effect sizes following 13 weeks or more of therapy. Law et al. (2004) found larger effect sizes following 8 weeks of therapy. Based on these studies an optimal duration of therapy for goal attainment may range between 4 and 13 weeks. In this study significantly positive outcomes occurred following both the concentrated (6 weeks) and distributed (12 weeks) groups. Bellon-Harn and Harn (2010) identified similar outcomes following 42 scaffolded-language therapy sessions distributed across 14 weeks, 3 times per week. Although effect sizes from this and Bellon-Harn and Harn (2010) should be interpreted cautiously due to small sample size, results suggest similar outcomes following a high dose therapy in the form of a scaffolded-language structure in 24 sessions across 6 or 12 weeks of therapy.

2 Dose form

Outcomes of this study corroborate previous studies using a scaffolded-language structure with implicit strategies. Greatest effects were noted in child production of interpretive language. Adults consistently responded to child initiations and responses at higher levels of semantic complexity. Although effects were not as large as other outcome measures, pre- and post-therapy MLU_m was significantly different. Expressing more complex semantic language may require the production of linguistic structures to instantiate the content.

Copula, auxiliary, and third person singular were examined since use of these forms may best demonstrate treatment effects in children with LI speaking AAE. Further, these forms occur early in morphologic development and are identified as areas of difficulty in preschool children with language impairment. Bellon-Harn and Harn (2010) noted increased production of auxiliary (is/are), copula (is/are), and third person singular during spontaneous language samples following therapy. Results of this study identified significant differences in pre- and post-therapy samples for auxiliary and copula only. Use of dialect is context specific, which may have influenced spontaneous production of third person singular in language samples. Probes were used to control for opportunity of production of auxiliary, copula, and third person singular. Moderate to large effect sizes were noted for percent occurrence of each pre- and post-therapy probe comparison.

In sum, dose frequency, dose, total treatment time, and dose form are interrelated. This study contributes some information regarding dosage. With a high dose therapy following a scaffolded-language structure using implicit strategies, 24 sessions within either concentrated or distributed weekly frequency ranging from 6 to 12 weeks may effect change. Overall, results suggest goal attainment within a school setting on semantic and morphologic aspects of language could occur in variety of scheduling frameworks.

3 Limitations and future research

More therapy may not always be better than less therapy (Warren et al., 2007). What is the dosage needed to reach a threshold of performance when applying a high dose intervention? Extending this study to examine relationships between dose, dose frequency, and total therapy time will clarify the service delivery options for high dose interventions in public school settings. A limitation of this study is post-therapy data was collected immediately following either the 6- or 12-week schedule. Stronger conclusions could have emerged with follow-up data collection (e.g. 6 weeks following each schedule). A larger sample size with children at various levels of language severity will provide more information regarding dosage. Additionally, various therapy forms (e.g. therapy with distributed trials across a session) should be considered. Dose frequency needed for optimal results may differ, which would inform practicing school-based SLTs.

Footnotes

Appendix 1

References

American Speech-Language-Hearing Association (2010) Roles and responsibilities of speech-language pathologists in schools. Professional issues statement. Retrieved from: http://www.asha.org/docs/pdf/PI2010-00317.pdf (March 2012).

Applebee

(1978) A child’s concept of story: Ages 2–17. Chicago, IL: University of Chicago Press.

Barratt

Littlejohns

Thompson

(1992) Trial of intensive compared with weekly speech therapy in preschool children Archives of Disease in Childhood 67: 106–08.

Bellon-Harn

Harn

(2010) Addressing multiple aspects of language simultaneously. Research presented at the annual meeting of the American Speech Language and Hearing Association, Philadelphia, PA, USA.

Bellon-Harn

Hoffman

Harn

(2004) Use of cloze and contrast word procedures on in repeated storybook reading: Targeting multiple domains. Journal of Communication Disorders 37: 53–75.

Bellon-Harn

Credeur-Pampolina

LeBoeuf

(2012) Scaffolded-language intervention: Speech production outcomes. Communication Disorders Quarterly (advanced online publication).

Brackenbury

Pye

(2005) Semantic deficits in children with language impairments. Language, Speech, and Hearing Services in Schools 36: 5–16.

Bradshaw

Hoffman

Norris

(1998) Efficacy of expansions and cloze procedures in the development of interpretations by preschool children exhibiting delayed language development. Language, Speech, and Hearing Services in Schools 29: 85–95.

Brandel

Loeb

(2011) Program intensity and service delivery modes in the schools: Survey results. Language, Speech, and Hearing Services in Schools 42: 461–90.

10.

Bruner

(1978) The role of dialogue in language acquisition. In: Sinclair

Jarvella

Levelt

WJM

(eds) The child’s conception of language: Springer series in language and communication. New York: Springer, 242–56.

11.

Burgemeister

Blum

Lorge

(1972) Columbia mental maturity scale. 3rd edition. Los Angeles, CA: Harcourt Brace Jovanovich.

12.

Cirrin

Schooling

Nelson

. (2010) Evidence-based systematic review: Effects of different service delivery outcomes for elementary school-age children. Language, Speech, and Hearing Services in Schools 41: 233–64.

13.

Cohen

(1988) Statistical power analysis for the behavioral sciences. 2nd edition. Hillsdale, NJ: Erlbaum, 191–96.

14.

Cole

Mills

Dale

(1989) Examination of test-retest and split-half reliability for measures derived from language samples of young handicapped children. Language, Speech, and Hearing Services in Schools 20: 259–68.

15.

Connell

Stone

(1992) Morpheme learning of children with specific language impairment under controlled instructional conditions. Journal of Speech and Hearing Research 35: 844–52.

16.

Cowley

(1990a) Mrs Wishy Washy. Bothell, WA: The Wright Group.

17.

Cowley

(1990b) The hungry giant. Bothell, WA: The Wright Group.

18.

Cowley

(1990c) The red rose. Bothell, WA: The Wright Group.

19.

Cowley

(1998a) Wishy Washy Day. Bothell, WA: The Wright Group.

20.

Cowley

(1998b) Splishy sploshy. Bothell, WA: The Wright Group.

21.

Cowley

(1998c) Mr Wishy Washy. Bothell, WA: The Wright Group.

22.

Cowley

(1998d) The scrubbing machine. Bothell, WA: The Wright Group.

23.

Cowley

(1998e) The grumpy elephant. Bothell, WA: The Wright Group.

24.

Craig

Washington

(2000) An assessment battery for identifying language impairments in African American children. Journal of Speech, Language, and Hearing Research 43: 366–79.

25.

Dawson

Stout

Eyer

(2003) Structured Photographic Expressive Language Test: Third edition (SPELT-3). DeKalb, IL: Janelle Publications.

26.

Devilly

(2004) The effect size generator for Windows: Version 2.3 [computer programme]. Melbourne: Centre for Neuropsychology, Swinburne University.

27.

Garrity

Oetting

(2010) Auxiliary BE production by AAE-speaking children with and without specific language impairment. Journal of Speech and Hearing Research 53: 1307–20.

28.

Gillum

Camarata

Nelson

Camarata

(2003) A comparison of naturalistic and analog treatment effects in children with expressive language disorder and poor preintervention imitation skills. Positive Behavior Interventions 5: 171–78.

29.

Goldman

Fristoe

(2000) Goldman–Fristoe test of articulation: 2. Bloomington, MN: Pearson Assessments.

30.

Individuals With Disabilities Education Act (IDEA) (2004), Public Law 108–446.

31.

United States Government (2004) Individuals with Disabilities Education Act (IDEA). In: Public Law. Washington, DC: 108th Congress, 108–446. Retrieved from: http://www.copyright.gov/legislation/pl108-446.pdf (March 2012).

32.

Jacoby

Lee

Kummer

Levin

Creaghead

(2002) The number of individual treatment units necessary to facilitate functional communication improvements in the speech and language of young children. American Journal of Speech-Language Pathology 11: 370–80.

33.

Jellison

(2007) Managing workloads. Perspectives on School-Based Issues 8: 14–16.

34.

Kirchner

Skarakis-Doyle

(1983) Developmental language disorders: A theoretical perspective. In: Gallagher

Prutting

(eds) Pragmatic assessment and intervention issues in language. San Diego: College-Hill Press, 215–46.

35.

Kouri

(2005) Lexical training through modeling and elicitation procedures with late talkers who have specific language impairment and developmental delays. Journal of Speech, Language, and Hearing Research 48: 157–71.

36.

Law

Garrett

Nye

(2004) The efficacy of treatment for children with developmental speech and language delay/disorder: A meta-analysis. Journal of Speech and Hearing Research 47: 924–43.

37.

Leonard

Camarata

Brown

Camarata

(2004) Tense and agreement in the speech of children with specific language impairment: Patterns of generalization through intervention. Journal of Speech and Hearing Research 47: 1363–79.

38.

McGinty

Breit-Smith

Fan

Justice

Kaderavik

(2011) Does intensity matter? Preschoolers’ print knowledge development within a classroom-based intervention. Early Childhood Research Quarterly 26: 255–67.

39.

McGregor

Newman

Reilly

Capone

(2002) Semantic representation and naming in children with specific language impairment. Journal of Speech and Hearing Research 45: 998–1014.

40.

MacWhinney

(2000a) The CHILDES Project: Tools for Analyzing Talk. Electronic edition. Part 1: The CHAT Transcription Format. Mahwah, NJ: Lawrence Erlbaum Associates. Retrieved from: http://childes.psy.cmu.edu/manuals/CHAT.pdf (March 2012).

41.

MacWhinney

(2000b) The CHILDES Project: Tools for Analyzing Talk. Electronic edition. Part 2: The CLAN Programs. Mahwah, NJ: Lawrence Erlbaum Associates. Retrieved from: http://childes.psy.cmu.edu/manuals/CLAN.pdf (March 2012).

42.

Meline

(2010) Research in communication sciences and disorders. 2nd edition. Upper Saddle River, NJ: Pearson Education.

43.

Nelson

(1985) Making sense: The acquisition of shared meaning. New York: Academic Press.

44.

Nelson

Camarata

Welsh

Butkovsky

Camarata

(1996) Effects of imitative and conversational recasting treatment on the acquisition of grammar in children with specific language impairment and younger language-normal children. Journal of Speech, Language and Hearing Research 39: 850–59.

45.

Norris

Hoffman

(1990) Language intervention within naturalistic environments. Language, Speech, Hearing, Services in Schools 21: 72–84.

46.

Nye

Foster

Seaman

(1987) Effectiveness of language intervention with the language/learning disabled. Journal of Speech and Hearing Disorders 52: 348–57.

47.

Oetting

McDonald

(2001) Nonmainstream dialect use and specific language impairment. Journal of Speech, Language and Hearing Research 44: 207–23.

48.

Owens

(2009) Language disorders: A functional approach to assessment and intervention. 5th edition. Boston, MA: Allyn and Bacon.

49.

Proctor-Williams

(2009) Dosage and distribution in morphosyntax intervention: Current evidence and future needs. Topics in Language Disorders 29: 294–311.

50.

Proctor-Williams

Fey

(2007) Recast density and acquisition of novel irregular past tense verbs. Journal of Speech, Language and Hearing Research 50: 1029–47.

51.

Proctor-Williams

Fey

Loeb

(2001) Parental recasts and production of copulas and articles by children with specific language impairment and typical language. American Journal of Speech and Language Pathology 10: 155–68.

52.

Schooling

Venediktov

Leech

(2010) Evidence-based systematic review: Effects of service delivery on the speech and language skills of children from birth to 5 years of age. Available at: http://www.asha.org/uploadedFiles/EBSR-Service-Delivery.pdf#search=%22Effects%22 (March 2012).

53.

Seymour

Roeper

de Villiers

(2005) Diagnostic Evaluation of Language Variation (DELV): Norm-referenced. San Antonio, TX: The Psychological Corporation.

54.

Sheng

McGregor

(2010) Lexical-semantic organization in children with specific language impairment. Journal of Speech and Hearing Research 53: 146–59.

55.

Snow

Midkiff-Borunda

Small

Proctor

(1984) Therapy as social interaction: Analyzing the contexts for language remediation. Topics in Language Disorders 4: 72–85.

56.

Ukrainetz

Ross

Harm

(2009) An investigation of treatment scheduling for phonemic awareness with kindergarteners who are at risk for reading difficulties. Language, Speech, and Hearing Services in Schools 40: 86–100.

57.

Warren

Fey

Yoder

(2007) Differential treatment intensity research: A missing link to creating optimally effective communication interventions. Mental Retardation and Developmental Disabilities Research Reviews 13: 70–77.

58.

Yoder

Spruytenburg

Edwards

Davies

(1995) Effect of verbal routine contexts and expansions on gains in mean length of utterance in children with developmental delays. Language, Speech, and Hearing Services in Schools 26: 21–32.