Improving Narrative Production in Children With Language Disorders: An Early-Stage Efficacy Study of a Narrative Intervention Program Purpose As noted in this forum, more research is needed to support the work of school-based speech-language pathologists who are designing and implementing interventions for students with language disorders. This article presents the findings of a multiple-baseline, single-subject study that was conducted to assess the outcomes of an intervention designed ... Clinical Focus
Free
Clinical Focus  |   April 05, 2018
Improving Narrative Production in Children With Language Disorders: An Early-Stage Efficacy Study of a Narrative Intervention Program
 
Author Affiliations & Notes
  • Sandra L. Gillam
    Communicative Disorders and Deaf Education, Utah State University, Logan
  • Abbie Olszewski
    Speech Pathology and Audiology, University of Nevada, Reno
  • Katie Squires
    Communication Disorders, Central Michigan University, Mount Pleasant
  • Katie Wolfe
    Educational Studies, University of South Carolina, Columbia
  • Timothy Slocum
    Special Education and Rehabilitation, Utah State University, Logan
  • Ronald B. Gillam
    Communicative Disorders and Deaf Education, Utah State University, Logan
  • Disclosure: The Department of Communicative Disorders and Deaf Education at Utah State University and the Gillams receive royalties from sales of the narrative intervention program described in the study. Ronald Gillam has a financial interest in the Test of Narrative Language, which was administered to the participants.
    Disclosure: The Department of Communicative Disorders and Deaf Education at Utah State University and the Gillams receive royalties from sales of the narrative intervention program described in the study. Ronald Gillam has a financial interest in the Test of Narrative Language, which was administered to the participants.×
  • Correspondence to Sandra L. Gillam: sandi.gillam@usu.edu
  • Editor-in-Chief: Shelley Gray
    Editor-in-Chief: Shelley Gray×
  • Editor: Ashley Meaux
    Editor: Ashley Meaux×
  • Publisher Note: This article is part of the Clinical Forum: Exploring Curriculum-Based Language Assessment and Interventions.
    Publisher Note: This article is part of the Clinical Forum: Exploring Curriculum-Based Language Assessment and Interventions.×
Article Information
School-Based Settings / Language Disorders / Clinical Forum: Exploring Curriculum-Based Language Assessment and Interventions / Clinical Focus
Clinical Focus   |   April 05, 2018
Improving Narrative Production in Children With Language Disorders: An Early-Stage Efficacy Study of a Narrative Intervention Program
Language, Speech, and Hearing Services in Schools, April 2018, Vol. 49, 197-212. doi:10.1044/2017_LSHSS-17-0047
History: Received May 22, 2017 , Revised July 24, 2017 , Accepted August 29, 2017
 
Language, Speech, and Hearing Services in Schools, April 2018, Vol. 49, 197-212. doi:10.1044/2017_LSHSS-17-0047
History: Received May 22, 2017; Revised July 24, 2017; Accepted August 29, 2017
Web of Science® Times Cited: 1

Purpose As noted in this forum, more research is needed to support the work of school-based speech-language pathologists who are designing and implementing interventions for students with language disorders. This article presents the findings of a multiple-baseline, single-subject study that was conducted to assess the outcomes of an intervention designed to improve narrative discourse proficiency for children with language disorders.

Method Four school-age children with language disorders that included deficits in narration received an experimental version of a 3-phase narrative language intervention program called Supporting Knowledge in Language and Literacy (Gillam, Gillam, & Laing, 2014). Two additional children remained in baseline throughout the study and served as controls for history, testing, and maturation effects. Measures of story productivity (number of different words) and overall story complexity (Monitoring Indicators of Scholarly Language; Gillam, Gillam, Fargo, Olszewski, & Segura, 2016) were used to assess the children's self-generated narratives.

Results After the onset of treatment, all 4 children who received the narrative intervention made moderate-to-large improvements in narrative productivity (number of different words). Three of the 4 children also made moderate-to-large improvements in narrative complexity (Monitoring Indicators of Scholarly Language). The narrative abilities of the 2 children who did not receive intervention did not change over the course of the study.

Conclusion This study provides evidence for the feasibility of the Supporting Knowledge in Language and Literacy narrative instruction program for improving self-generated narratives by children with language disorders. Future research is needed to determine how gains in oral narration transfer to written narrative skills.

Approximately 7% of children have deficits in language comprehension or production that persist into the school-age years and interfere with social and academic functioning (Tomblin et al., 1997). Such children are often referred to as having a language disorder (Bishop, 2014). 1   Their language difficulties, which occur in the absence of sensory, motor, intellectual, and/or other medical or neurological conditions, often include problems with narrative discourse. With respect to narrative comprehension, children with language disorders are less likely to provide correct answers to literal or inferential questions about stories that have been read to them (Bishop & Adams, 1992; Gillam, Fargo, & Robertson, 2009; Laing & Kamhi, 2002; Wright & Newhoff, 2001). With respect to narrative production, their stories are often shorter and simpler in structure (Fey, Catts, Proctor-Williams, Tomblin, & Zhang, 2004; McFadden & Gillam, 1996; Newman & McGregor, 2006) than stories produced by typically developing children, and they tend to contain fewer complex sentences, less diverse vocabulary, and limited literate language features (Gillam & Johnston, 1992; Greenhalgh & Strong, 2001; Kaderavek & Sulzby, 2000; Scott & Windsor, 2000).
Narration is an important type of discourse that has clear ties to socialization (Fujiki, Spackman, Brinton, & Illig, 2008; Goencue & Klein, 2001), classroom discourse (Yawkey, Aronin, Street, & Hinojosa, 1974), and early literacy (Snow, Burns, & Griffin, 1998). Narrative proficiency is a crucial skill if students are to be successful in attaining grade-level listening and reading comprehension and oral and written composition skills as described in the Common Core State Standards (CCSS; National Governors Association Center for Best Practices and Council of Chief State School Officers, 2010). Unfortunately, elementary school–age children with language disorders who demonstrate poor narrative skills are disadvantaged during a large portion of the school day because a great deal of classroom instruction incorporates some degree of narrative discourse into the lessons. As a result, many scholars have suggested that functional language intervention for school-age children with language disorders should include lessons that target narrative proficiency (Boudreau, 2008; Naremore, Densmore, & Harman, 1995; Wellman et al., 2011).
As Dr. Powell (2018)  states in the lead article of this clinical forum, it is critical that speech-language pathologists (SLPs) implement interventions that will support students in meeting the demands of the curriculum. A focus on narrative proficiency is one way that SLPs can provide effective instructional strategies to aid in spoken and written language comprehension and production.
Studies of Narrative Intervention With School-Age Children
Although there have been several studies using narrative intervention procedures with preschool children and school-age children who are learning English or otherwise at risk for academic failure (Brown, Garzarek, & Donegan, 2014; Petersen & Spencer, 2016; Spencer, Petersen, Slocum, & Allen, 2015), there are relatively few studies that have examined the outcomes of intervention practices for improving narrative proficiency in school-age children with language disorders, and even fewer using narrative discourse not only as a target but as a context for intervention (Brown et al., 2014; Cirrin & Gillam, 2008; Gillam, Gillam, & Reece, 2012; Hoffman, 2009). Most of the studies that have been conducted targeting narrative proficiency have included explicit, direct instruction on specific story elements by asking children to identify and answer questions about them and/or to use them while retelling or creating stories that correspond to sequenced pictures. The studies reviewed in this article have contributed to an evidence base of narrative procedures that have been shown to improve students' ability to understand and compose in the narrative genre.
To illustrate, Hayward and Schneider (2000)  taught 13 children (aged 4–6 years) with moderate-to-severe speech and language impairments to identify story grammar elements, to determine when elements were missing in stories, and to sequence events in stories in chronological order. Phrases were written on color-coded cue cards to remind students to identify and use settings (when), characters (what), and actions (what doing), to include events in chronological order (what happens 1-2-3), and to include internal responses and reactions of characters (who + feelings + action). Narrative proficiency was measured by asking children to look at two different sets of five pictures from storybooks and compose stories for the examiner before and after instruction. Then, children were asked to tell their stories to an unfamiliar listener who was unable to see the pictures. More than half of the children who participated in the instruction demonstrated large improvements in the episodic complexity and content of their oral narratives.
In a later study with older children, Swanson, Fey, Mills, and Hood (2005)  included selected story elements in a narrative instruction program (i.e., setting, character, problem, resolution, complication, and ending) delivered to 10 school-age children with language disorders between the ages of 7 and 8 years. In addition to teaching story elements in the context of authentic stories, children were given multiple opportunities to draft and tell their own stories using pictographic planning (Ukrainetz, 1998). Narrative proficiency was measured by asking students to generate a story when provided sets of picture sequences after listening to the examiner tell a story using a different set of sequenced pictures. Eighty percent of the participants made clinically significant gains in their oral narratives after instruction. There was no control group; hence, it is not possible to demonstrate conclusively that the intervention, and not some other factor, caused the improvements that were noted. However, this study suggests that teaching story elements in context and giving students multiple opportunities to draft (using pictographic planning) and tell stories may be effective in improving narrative production.
A slightly different approach to narrative instruction was taken by Nathanson, Crank, Saywitz, and Ruegg (2007), who taught visual cues to signal story elements to 39 first-grade through fifth-grade children with learning disabilities. Half of the children participated in a narrative elaboration treatment and were taught to identify and use character and setting details as well as character actions and feelings in order to remember video vignettes. The children in the comparison group were asked to watch the same vignettes but were not given any instruction in the visual cuing system used to signal story elements. Children who participated in the narrative elaboration treatment training condition were shown to recall more information from a lesson on Mexican history that they had been exposed to 2 weeks earlier using the cuing system than children in the comparison condition. These results support earlier research findings that have suggested that knowledge of the story structure is associated with greater recall of information (van den Broek, Linzie, Fletcher, & Marsolek, 2000). This study is particularly important because it suggests that even delayed recall is improved when an organizational framework is used to assist in retrieval of information.
After reviewing several studies of narrative language intervention, it is clear that there are a number of procedures and activities that may be used to improve narrative proficiency. Most have targeted narrative by teaching students to identify and create simple, single-episode stories and utilize story-retelling tasks (many developed by researchers) or story-generation tasks using sequenced pictures to measure narrative proficiency. These are important skills and tasks that have their place among the curricular demands students must meet. However, there is also a need to test interventions that teach students to create longer, more complex, age- and content-appropriate narratives encountered in daily academic instruction (Lynch et al., 2008) and to evaluate performance using more rigorous outcome measures such as single-scene picture stimuli or verbal prompts. Toward that end, the intervention used in the current study incorporated many of the evidence-based procedures described in the literature and provided students with multiple opportunities to compose and evaluate the stories they create using single-scene prompts.
Some narrative interventions place heavy emphasis on teaching students about the story elements of narratives without much time devoted to microstructure features such as learning vocabulary from context, the use of complex syntax, and the causal and temporal words needed for detail, cohesion, and clarity in stories. The intervention used in this study incorporated all of these activities.
Finally, we could find no narrative intervention program that incorporated scaffolded lessons to teach students the metacognitive skills necessary for evaluating, comparing, and contrasting aspects of narrative discourse they hear or read (NGA, 2010; CCSS.ELA-Literacy.RL.2.9) or in editing the oral stories they compose (NGA, 2010; CCSS.ELA-Literacy.W.2.5). The narrative intervention used in this project incorporated metacognitive instruction.
Purpose and Theoretical Rationale
The purpose of this study was to determine whether a comprehensive narrative discourse program called Supporting Knowledge of Language and Literacy (SKILL; Gillam, Gillam, & Laing-Rogers, 2014) led to improvements in the self-generated stories produced by children with language disorders. SKILL was designed to target the psycholinguistic construction and integration processes necessary for understanding and composing spoken and written narratives. According to the construction–integration (C-I) model of text comprehension (Kintsch, 2013), narrative and expository discourse requires the construction of a textbase, which is a representation of what the oral or written discourse actually says, and integration, which is the reader or listener's interpretation of what the text means. According to the C-I model, the learner uses his or her knowledge of words and language structures to create the microstructure of the textbase (Perfetti & Stafura, 2014). For example, in the book Little Croc and Whale (Maddox, 2009), the first sentence says, “Little Croc was bored with swimming in the lake.” 2   The reader or the listener must first associate each word with a meaning and characterize it according to its syntactic role in the sentence.
  • Little Croc—a name for a character in the story, probably a crocodile (noun)

  • was—a verb indicating a state of being or existence

  • bored—an adjective indicating one lacks interest in an activity or is unoccupied

  • with—a preposition indicating that something or someone is accompanied by another person or thing

  • swimming—a verb indicating propulsion through the water

  • in—a preposition expressing the situation of something that is or appears to be enclosed or surrounded by something else

  • the—a determiner denoting one or more people or things that have already been mentioned or assumed to be common knowledge

  • lake—a noun representing a large body of water surrounded by land

The next sentence in the book says, “Wouldn't it be fun, he thought, to swim in the rapids, where the river runs fast and deep?” As the words in this sentence are associated with word meanings and the syntactic information that it is used to aid in understanding them, the listener or the reader begins to form links between the meanings and structures in the sentences and a microstructure (part of the textbase) of what is said or read begins to be constructed.
Kintsch believes storytellers and readers must also construct an overall model of the text called a macrostructure (another piece of the textbase) that represents the hierarchical relationships among key ideas. For example, consider the passage in Little Croc and Whale. As the listener or the reader constructs the microstructure of the textbase, he or she may begin to organize the information using an identifiable text structure, which, in this case, is a narrative text structure. As the story unfolds, the learner may realize that this is a fictional narrative that is predictably organized into common story elements (e.g., setting, initiating event, attempts, and consequence) and the causal relations between them (goal-directed actions taken by the character).
Integration involves the formation of a situation model. The situation model is the listener or reader's mental representation or interpretation of what is going on in the text. It is formed by comparing background knowledge and experiences to the textbase (microstructure and macrostructure) that has been constructed. This results in an interpretation that is unique to the listener or the reader. To complete the integration process, the new information in the textbase is linked to old information and stored in long-term memory through strategic and conscious effort. For example, The listener or the reader may think, “This story is about a little crocodile who is in a calm pond and is thinking about going over to swim in the rapids because he's tired of the little pond. I'm thinking the pond is safe and rapids are not. The rapids are probably fast like the river I fell in when I was little and Dad had to jump in and get me out. I was tired of sitting on the side of the river and that's why I jumped in. So I guess I was bored like Little Croc. If the Croc goes to the rapids, I wonder if he will be OK or if someone will need to pull him out? I will have to keep reading to find out.”
A failure to integrate the situation model into long-term memory through a process like this may result in encapsulated knowledge that is isolated from the contents of world knowledge. When the student encounters the word “rapids” in another situation, he or she may not recognize the word or be able to explain what it means unless he or she had actively and strategically linked it to the “fast river” that he or she had fallen into in the past.
The components of the C-I model informed each of the lessons in the SKILL program, which was designed to teach students to how to construct textbases and situation models and then integrate them in long-term memory for use in future encounters with similar discourse.
We used a single-subject, multiple-baseline, across-participants design with lagged intervention phases across two sets of children to obtain early-stage efficacy data on the feasibility of the SKILL program. The specific research questions were as follows:
  1. Does narrative intervention with the SKILL program lead to increases in language productivity in self-generated narratives produced by school-age children with language disorders?

  2. Does narrative intervention with the SKILL program lead to increases in the complexity of self-generated narratives produced by school-age children with language disorders?

Method
Participants
A principal from a local school in northern Utah identified monolingual, English-speaking children who were clinically identified as having a language impairment and were receiving speech language services in the school. SLPs working with the students reported that none were receiving narrative language intervention. The principal sent home institutional review board–approved recruitment letters and consents to all of the children. Signed consent forms were returned, and students were tested individually. Participants were excluded from the study if they demonstrated hearing or visual impairment, gross neurological impairment, oral–structural anomalies, or emotional/social disorders. Six participants ranging in age from 6;7 [years;months] to 10;4 qualified to participate in the study. Participant's cognitive skills were measured using the screening portion of the Universal Nonverbal Intelligence Test (Bracken & McCallum, 1998), language skills were measured with the Clinical Evaluation of Language Fundamentals–Fourth Edition (CELF-4; Wiig, Semel, & Secord, 2003), and narrative proficiency was measured using the Test of Narrative Language (Gillam & Pearson, 2004). All children who were enrolled in the study received standard scores within 1 SD from the mean on the Universal Nonverbal Intelligence Test and standard scores at or below 85 on both the CELF-4 and the Test of Narrative Language (see Table 1).
Table 1. Descriptive data on the six participants before treatment.
Descriptive data on the six participants before treatment.×
Variables Participants
Brock Sam Casey Paul Jenn Jack
Gender M M F M F M
Age [years;months] 7;9 6;7 10;4 8;1 6;7 7;8
CELF-4 73 58 84 60 82 81
TNL 73 58 76 76 85 76
UNIT 12 11 7 10 12 8
Note. CELF-4 = Core Language Standard Score from the Clinical Evaluation of Language Fundamentals–Fourth Edition (Wiig, Semel, & Secord, 2003); TNL = Narrative Language Ability Index from the Test of Narrative Language (Gillam & Pearson, 2004); UNIT = screening standard score (M = 10, SD = 3) from the Universal Nonverbal Intelligence Test (Bracken & McCallum, 1998).
Note. CELF-4 = Core Language Standard Score from the Clinical Evaluation of Language Fundamentals–Fourth Edition (Wiig, Semel, & Secord, 2003); TNL = Narrative Language Ability Index from the Test of Narrative Language (Gillam & Pearson, 2004); UNIT = screening standard score (M = 10, SD = 3) from the Universal Nonverbal Intelligence Test (Bracken & McCallum, 1998).×
Table 1. Descriptive data on the six participants before treatment.
Descriptive data on the six participants before treatment.×
Variables Participants
Brock Sam Casey Paul Jenn Jack
Gender M M F M F M
Age [years;months] 7;9 6;7 10;4 8;1 6;7 7;8
CELF-4 73 58 84 60 82 81
TNL 73 58 76 76 85 76
UNIT 12 11 7 10 12 8
Note. CELF-4 = Core Language Standard Score from the Clinical Evaluation of Language Fundamentals–Fourth Edition (Wiig, Semel, & Secord, 2003); TNL = Narrative Language Ability Index from the Test of Narrative Language (Gillam & Pearson, 2004); UNIT = screening standard score (M = 10, SD = 3) from the Universal Nonverbal Intelligence Test (Bracken & McCallum, 1998).
Note. CELF-4 = Core Language Standard Score from the Clinical Evaluation of Language Fundamentals–Fourth Edition (Wiig, Semel, & Secord, 2003); TNL = Narrative Language Ability Index from the Test of Narrative Language (Gillam & Pearson, 2004); UNIT = screening standard score (M = 10, SD = 3) from the Universal Nonverbal Intelligence Test (Bracken & McCallum, 1998).×
×
Narrative Intervention
Each participant who received intervention had one-on-one instruction from a certified SPL using the SKILL instruction program that consisted of three phases: Phase I: Teaching Story Structure and Causal Language, Phase II: Teaching Strategies for Creating a Situation Model, and Phase III: Teaching Strategies for Integration Into Long-Term Memory. The same clinician worked with the same participant throughout the study.
The three phases of the SKILL instructional program contained multiple lessons on story structure and organization and the specific linguistic markers used to signal the temporal and causal relationships between propositions (meaning units). These lessons were designed to support the development of a framework that may then be used to assist in the construction of an accurate and coherent representation of a story that is organized locally (phrase, sentence, paragraphs) and globally (discourse, text structure).
SKILL utilized icons and graphic organizers to assist students in constructing stories using traditional story grammar. Students participated in lessons that targeted the ability to create a coherent and cohesive situation model for later integration into long-term memory. SKILL also contained lessons designed to establish local and global coherence. Local coherence is established through the use of clear, accurate pronoun references, and global coherence is established and maintained by creating a stable situation model derived from an accurate representation of the textbase.
Furthermore, SKILL contained lessons that were designed to foster the use of strategies for creating a situation model of the text through practice with paraphrasing (Tell me what happened in this part of the story), sentence combining (use of vertical structuring to elicit two sentences and then combining them into one complex sentence), and the use of specific causal and temporal linguistic structures (because, so; then, next). Finally, SKILL lessons focused on the metacognitive skills needed to evaluate the hierarchical structures (macrostructure) in stories and to create coherent and cohesive hierarchical structures in the stories they create themselves.
Phase I contained 20 lessons that provided students with an understanding of the main story elements, including characters, setting, initiating event, internal response, plan, actions, and consequences in the context of wordless picture stories created by the authors. The wordless books were developed to contain single episodes containing all the key elements that were being taught. This allowed the researchers to control the difficulty of the stories used to teach the story elements and the causal relationships between them. Each story element was associated with a representative icon that served as a graphic organizer.
After each element was taught, students participated in lessons focused on identifying and answering questions related to the story elements, paraphrasing and retelling the wordless storybook, and creating and telling their own stories using the icons and graphic organizers with highly scaffolded assistance. Phase I ended with lessons that were designed to help students transfer their newly learned skills to comprehending, retelling, and composing stories using a popular children's literature book.
The explanations of story elements, icons, and storyboards were used to help the student connect the meanings of the words, phrases, and sentences in the discourse to word meanings, images, or experiences stored in working memory. These literate language components are used to form a microstructure (Perfetti & Stafura, 2014). The procedures in Phase I were also used to help the student construct a mental model of the discourse at a global level (macrostructure).
To move to Phase II, students had to demonstrate that they were able to (a) identify, define, and give examples for each story icon, (b) create and tell a story using a storyboard (with assistance), and (c) answer comprehension questions related to each story element after hearing a novel story. If students demonstrated difficulty in any of these skills, they completed additional practice lessons designed to teach them.
There were 18 lessons in Phase II (Teaching Strategies for Creating a Situation Model), which were designed to expose students to linguistic structures, concepts, and vocabulary in more elaborate, complex stories. In this phase, students were taught how to create longer stories using words to indicate temporal and causal relationships and how to make stories more interesting through the inclusion of more diverse metalinguistic verbs (e.g., yelled, growled, and whispered vs. said) and character dialogue. Wordless books were used initially for practicing newly learned skills to control for difficulty and to focus students on the specific aspects of macrostructure being taught.
Additional lessons targeted story comprehension skills. For example, children practiced answering questions about stories. In one lesson, students were read a story and asked to look at the pictures and retell the story to the clinician. If students added information that was not in the story, he or she was directed to look at the icons while the clinician reminded the student of the actual story. Children also practiced summarizing stories. For example, after retelling portions of a story about a hamster that got away from its owner, the student was asked, “What was this story about?” Leading questions were used to help the student arrive at the correct answers.
Some activities in Phase II focused on teaching students to create complex stories that include complicating events and a variety of adverbs, subordinated conjunctions, and adjectives. New icons (e.g., dialogue, plan again) and more elaborate storyboards were used to provide support as students learned new and increasingly more complex concepts in longer stories. Students also participated in mini-lessons targeting the use of the adverbial subordinating conjunction “because” and words to denote internal response (e.g., sad, frustrated, and afraid). Lessons to facilitate comprehension were incorporated throughout Phase II. In these lessons, children listen to new, unfamiliar stories, retell them, and answer questions about the story elements and causal relationships between events.
In order to move from Phase II to Phase III, students had to demonstrate the ability to (a) create and tell a story using a complex storyboard with minimal assistance; (b) incorporate the words because or so, two or more feeling words, two or more mental or linguistic verbs, one or more adverb, and one or more elaborated noun phrase into their self-generated stories; and (c) answer comprehension questions about and retell a novel story. If students demonstrated difficulty with any of these skills, they completed additional practice lessons that were included in the manual.
Phase III (Teaching Strategies for Integration Into Long-Term Memory) contained 12 lessons to give students multiple opportunities to retell and evaluate stories they hear and to create and edit their own spontaneously generated stories with and without icon and graphic organizer support. The major focus of Phase III was to develop metacognitive skills needed for independence in their understanding and use of narrative structure and complex oral language relevant to coherent stories (i.e., connecting terms, causal language, and mental state terms). Students were taught to use a rubric to critique children's literature books and to edit the stories they create. Lessons proceeded from highly scaffolded and supportive contexts such as creating and telling stories using sequenced pictures to more independent contexts such as creating and telling stories from single-scene picture prompts.
All of the participants who received intervention participated in all of the lessons in Phases I, II, and III. Students progressed through the program at their own pace, but all students who received instruction completed the entire program.
Treatment Fidelity
All intervention sessions were videotaped. For each participant who received intervention, a member of the research team who did not provide intervention observed the videotape of one randomly selected session per week. Treatment fidelity was monitored using an intervention observation checklist that contained elements that were specific to each day's lesson. For example, the fidelity items for Lesson 1 in Phase I included the following:
  • Introduce the lesson by telling the child what is expected: “We are going to learn how to tell great stories. Great stories have lots of parts to them. I'm going to tell you a story now that has lots of parts.”

  • Tell the story, “Camping Trouble with Dogs.”

  • Describe the story elements.

  • Introduce the concept of icons.

  • Ask children what an icon is.

  • Discuss the concept of a symbol.

To obtain a fidelity score of 100%, all six of the items above would have had to have been observed during the session. If fidelity fell below 85% for any lesson, the research staff held a short meeting with the interventionist to discuss what was omitted. There were sessions during which a clinician failed to meet the minimum fidelity requirement. In all cases, it was because the lesson had not been completed. When this happened, the omitted information was introduced in the following session. Interrater reliability was calculated on 50% of the fidelity data by two members of the research team. Point-by-point inter-rater reliability was 98%. Across the treatment sessions, 94.4% of the fidelity items were counted as being present. The excellent levels of fidelity occurred because the clinicians followed the protocol for each lesson in the SKILL manual closely.
Dependent Variables
The dependent variables were two narrative language sample measures: number of different words (NDW), which is a measure of story productivity, and Monitoring Indicators of Scholarly Language (MISL; Gillam, Gillam, Fargo, Olszewski, & Segura, 2016), which is a measure of story complexity. NDW is a valid and reliable indicator of story length that has been shown to predict narrative performance (Heilmann, Miller, Nockerts, & Dunaway, 2010). The MISL is an overall measure of narrative proficiency that was designed to capture the complexity of macrostructure and microstructure elements in stories. The MISL rubric used in this project contained seven macrostructure items and five microstructure (literate language) items. Each of the seven macrostructure elements (character, setting, initiating event, internal response, plan, action, consequence) received a score ranging from 0 to 3 (0 = absent, 1 = emerging, 2 = present, 3 = complex/elaborated use). For example, the use of unreferenced pronouns for characters, as in, “She is walking to the picnic table.” received a 0. General references such as, “The girl is walking to the picnic table.” received a 1. Specific names, as in, “Susan is walking to the picnic table.” received a 2. Character references receive a score of 3 when students explicitly stated the names of two or more characters. A child who told a complex story that contained multiple embedded episodes, complicating actions and clear connections between the initiating event, internal response, plan, actions, and consequences could receive a total macrostructure score of 21.
The five microstructure items (coordinating conjunctions, subordinating conjunctions, metacognitive/metalinguistic verbs, adverbs, and elaborated noun phrases) were also scored on a 0–3 scale (0 = absent, 1 = 1 exemplar present, 2 = 2 different exemplars present, 3 = 3 different exemplars present). For example, if a child used the words think, know, and said in their story, they would receive a score of 3 for use of metacognitive/metalinguistic verbs. Similarly, if a child's story contained multiple modifiers before a noun such as the big blackdog, a score of 3 would be given for elaborated noun phrases in contrast to a score of 2 for the utterance the bigdog.
Verb tense and grammaticality were also measured. The story was given a verb tense score of 0 if there were three or more tense changes and a 3 if the story was told in the same tense throughout. The story was given a score of 0 if there were three or more grammatical errors, and a score of 3 if there were no grammatical errors. The total possible score for this subscale including tense and grammaticality was 21. The scores for both subscales were combined to yield a total story structure score (maximum = 42).
Research assistants who were not involved in the intervention collected spontaneous stories from each participant at his or her school once each week during the baseline and intervention phases. The research assistants rotated in collecting stories from participants in order to remain relatively unfamiliar to the students. To collect spontaneous stories, the research assistants showed the children a picture of a scene and asked them to create a story about it. Each student was given the same picture prompt at each time point over the course of the study. For example, at baseline Time 1, all students were given a prompt that depicted a girl playing in leaves. At baseline Time 2, all students were given a prompt that depicted a beach scene. Once all of the scenes were used, the examiner started over with the first picture prompt. A list of the 32 picture scenes is shown in the Appendix.
The examiner elicited a story using the prompts saying, “I am going to show you a picture. I want you to make up a story that goes with the picture you see. Tell the best story you can. You can think about it for a minute. Start when you are ready.” None of the stimulus pictures were used during the intervention program. Examiners did not prompt the participants except to ask once if they were finished with their story.
Scoring Reliability
Audio recordings of the children's self-generated stories were transcribed by a primary research assistant using Systematic Analysis of Language Transcription (Miller, Andriacchi, & Nockerts, 2015) for transcription and coding procedures. A second research assistant independently transcribed each story. Discrepancies were resolved through consensus. The final transcript was scored by one or more members of a team of research assistants using the MISL. All members of the scoring team were trained on the MISL coding procedures and had demonstrated 90% or higher reliability on a set of training stories. After 10 stories were transcribed and scored, all of the members of the scoring team independently scored the last story. The team met to review their MISL scores throughout the study to ensure that point-by-point reliability remained at or above 90%. If there were any scoring discrepancies, they were resolved through consensus. If the point-by-point reliability was below 90%, an additional story was reviewed. This process was repeated throughout the study to control for potential scoring drift (i.e., the tendency to gradually move away from the intended target). In addition, a random selection of 20% of the stories were scored by an independent rater for a formal inter-rater reliability score for the MISL, which was 92%.
Design
We used a multiple-baseline, across-participants design in which the SKILL intervention was applied to two sets of three children in a public school setting during the school day. The children were assigned random numbers 1–6. The children assigned as 1, 2, and 3 constituted set 1, and the children assigned as 4, 5, and 6 constituted set 2. The first set included Brock, Sam, and Casey (Participants 1–3). The second set included Paul, Jenn, and Jack (Participants 4–6). During the baseline period, we collected self-generated stories produced in response to single-scene picture prompts. Three participants (Brock, Sam, and Paul) told stories that consisted of a simple episode without elaboration. The other three participants (Casey, Jenn, and Jack) produced descriptions of the pictures without any observable story elements or causal relationships between any of the actions or descriptions. All of the participants produced stories that contained very few literate language features, and none of the children produced stories that contained metacognitive or metalinguistic verbs.
Intervention began for the first participants in each set when they had participated in a minimum of three baseline sessions and had demonstrated a stable or downward trend in narrative performance as measured using the MISL. Subsequent intervention periods for the additional participants started once the first person in each set had completed all three phases of the intervention program. We were able to collect at least two follow-up stories beginning a week after intervention ended for three of the participants.
We had originally planned to complete three legs of the multiple-baseline study with the three participants in each set. Due to time constraints related to conducting the study in a public school setting, we were only able to complete two legs of the multiple-baseline study with each set of children before the school year ended. Therefore, the third child in each set remained in baseline throughout the study. The data from the continuous-baseline participants enabled us to further evaluate the degree to which repeated testing or maturation related to school instruction contributed to the changes that were observed in intervention.
Treatment Duration
The four children who received instruction participated in two sessions per week for 6–8 weeks of intervention (see Table 2). The number of treatment sessions ranged from 13 to 24 depending on the individual child's progress through the program. Children spent from 10 hr 8 min to 14 hr 6 min in instruction. Brock received the most hours of instruction (14 hr 6 min), followed by Paul (13 hr 44 min), Sam (10 hr 10 min), and then Jenn (10 hr 8 min). One reason for the difference in the time spent in therapy related to the length of the sessions, which varied between 34 and 48 min depending on the participant. Another reason for differences in the time spent in intervention related to individual differences in response to instruction. The program was designed to allow for children to progress at their own pace. Therefore, some participants completed the program earlier than others. All children completed all of the lessons in the program.
Table 2. Intervention duration statistics for the participants who received intervention.
Intervention duration statistics for the participants who received intervention.×
Intervention details Participants
Brock Sam Paul Jenn
Weeks in intervention 8 7 8 6
Number of sessions 24 18 22 13
Average minutes per session minutes of intervention 35 34 38 47
 Week 1 108 84 74 101
 Week 2 146 101 113 97
 Week 3 100 86 78 91
 Week 4 101 108 118 102
 Week 5 125 106 120 138
 Week 6 93 37 119 79
 Week 7 88 50 91
 Week 8 85 111
Total hours of therapy 14 hr 6 min 10 hr 10 min 13 hr 44 min 10 hr 8 min
Table 2. Intervention duration statistics for the participants who received intervention.
Intervention duration statistics for the participants who received intervention.×
Intervention details Participants
Brock Sam Paul Jenn
Weeks in intervention 8 7 8 6
Number of sessions 24 18 22 13
Average minutes per session minutes of intervention 35 34 38 47
 Week 1 108 84 74 101
 Week 2 146 101 113 97
 Week 3 100 86 78 91
 Week 4 101 108 118 102
 Week 5 125 106 120 138
 Week 6 93 37 119 79
 Week 7 88 50 91
 Week 8 85 111
Total hours of therapy 14 hr 6 min 10 hr 10 min 13 hr 44 min 10 hr 8 min
×
Results
There have been numerous debates in the literature about the use of statistical procedures for interpreting outcomes in single-subject studies (e.g., Scruggs & Mastropieri, 2000). Visual analysis is often preferred over numerical analyses because single-subject data do not meet many of the assumptions upon which many statistical methods depend, and statistical analyses may mask patterns in the data that a visual analysis would ascertain. However, visual analysis is subjective and can be highly influenced by the range of the scores that are reported. Also, there are no consistent standards for making judgments about intervention effects (Morgan & Morgan, 2009). The American Psychiatric Association (2013)  has recommended that all studies submitted for publication include measures of effect sizes to facilitate comparisons across studies. Therefore, we used qualitative and quantitative approaches to assess clinically significant improvement on the two dependent variables related to narrative ability: NDW, which was an index of productivity, and the total MISL score, which was an index of structural complexity.
Qualitative Visual Analysis
Figures 1 and 2 present the data for NDW (solid triangles with the scale on the left y-axis) and the total MISL score (open triangles with the scale on the right y-axis) for each day that spontaneous stories were collected (x-axis). The MISL score represented on the right y-axis ranged from 0 to 42, whereas the NDW represented on the left y-axis had no upper bound. The data for the first set of participants (see Figure 1) and those for the second set of participants (see Figure 2) were grouped together for ease of comparison. The scales for the two sets of participants differ to better represent the individual variation students demonstrated in response to the intervention and because children started and ended at different levels of narrative productivity and complexity. The vertical lines in the figures represent baseline, treatment, and follow-up phases of the study.
Figure 1.

Set 1 of the multiple-baseline design for the outcome measures obtained from the Monitoring Indicators of Scholarly Language (MISL) rubric and the number of different words (NDW) at each time-point.

 Set 1 of the multiple-baseline design for the outcome measures obtained from the Monitoring Indicators of Scholarly Language (MISL) rubric and the number of different words (NDW) at each time-point.
Figure 1.

Set 1 of the multiple-baseline design for the outcome measures obtained from the Monitoring Indicators of Scholarly Language (MISL) rubric and the number of different words (NDW) at each time-point.

×
Figure 2.

Set 2 of the multiple-baseline design for the outcome variables obtained from the Monitoring Indicators of Scholarly Language (MISL) rubric and the number of different words (NDW) at each time-point.

 Set 2 of the multiple-baseline design for the outcome variables obtained from the Monitoring Indicators of Scholarly Language (MISL) rubric and the number of different words (NDW) at each time-point.
Figure 2.

Set 2 of the multiple-baseline design for the outcome variables obtained from the Monitoring Indicators of Scholarly Language (MISL) rubric and the number of different words (NDW) at each time-point.

×
Following Kratochwill et al. (2010), we looked for predictable patterns in the baseline data and evidence that data in the intervention phase indicated changes in the level of performance, trend (slope), variability, immediacy of effect, overlap, and consistency in comparison to the baseline phases for each participant. Clear evidence of a causal relationship between treatment and improvement on the dependent measures would be indicated by stable or decreasing baselines, consistent increases in the dependent measures during the treatment phase, stable baselines for the controls, minimal overlap between data collected in the baseline and treatment phases, and consistent results across the four participants who received treatment. Visual analysis of the data trends in the two figures provide support for the effectiveness of the treatment.
Set 1 (Brock, Sam, and Casey)
In the first set of participants (see Figure 1), both Brock (Participant 1) and Sam (Participant 2) had decreasing baselines for NDW and the total MISL score prior to the onset of intervention, with Sam presenting greater baseline variability. The initiation of treatment had an immediate positive effect on both participants, followed by continued improvements (with some degree of variability) during the intervention phase. Sam's trend is more precipitous than Brock's, with Brock showing more variability in both measures in the stories that were collected between the 6th and the 18th intervention sessions. Both Brock and Sam had minimal overlap between that data collected in the baseline and intervention phases.
Casey (Participant 3) remained in baseline throughout the study. There was greater variability in Casey's total MISL scores as compared to NDW, but both measures had a generally flat trend from the first to the last data points, with high levels of overlap between the first half and the last half of the data points.
Set 2 (Paul, Jenn, and Jack)
The data for the second set of participants (see Figure 2) differed somewhat from that of the first set of participants. Paul (Participant 4) had a flat baseline for the total MISL score and a decreasing baseline for NDW, whereas Jenn (Participant 5) had decreasing baselines (with greater variability in the total MISL score than NDW) prior to the onset of intervention. After treatment was initiated, both participants presented continuous improvements (with some degree of variability) across the intervention phase. Jenn's slope was greater than Paul's but it was also more variable. Paul appeared to have less nonoverlapping data between the baseline and intervention phases than Jenn. Thus, it appeared that Paul profited from the intervention but not to the same extent as Jenn.
Jack (Participant 6) remained in baseline throughout the study. There was less variability in Jack's NDW and MISL scores as compared to Casey's (the control child in Set 1). Like Casey, both measures had a generally level trend from the first to the last data points with high levels of data overlap between the first half and the last half of the data points.
Quantitative Analysis
We report quantitative analyses (see Tables 3 and 4) that were based on comparisons of data within and across phases. To assess the change in level, we computed the percent of change from the baseline to the end of the treatment period (% change) by dividing the mean score of the last three stories collected during treatment (representing the final outcome of treatment) by the baseline average and multiplying that proportion by 100. To assess the overlap in performance between baseline and treatment periods, we calculated the percent of all nonoverlapping data (PND) using the formula [1 − (number of overlapping data points in the treatment phase / total number of data points across baseline and treatment phases)] (Parker, Vannest, & Davis, 2011).
Table 3. Measures of level and trend for total number of different words (NDW) and the Monitoring Indicators of Scholarly Language (MISL) score for the participants who received treatment.
Measures of level and trend for total number of different words (NDW) and the Monitoring Indicators of Scholarly Language (MISL) score for the participants who received treatment.×
Participant Baseline avg Last 3 treatment session avg Percent change PND Tau-U
NDW
 Set 1
 1 – Brock 28.67 65.00 126% 95% .92**
 2 – Sam 45.00 147.67 228% 85% .71**
 Set 2
 4 – Paul 16.00 35.67 123% 84% .81**
 5 – Jenn 12.21 56.33 361% 90% .74**
MISL
 Set 1
 1 – Brock 11.67 23.00 97% 100% 1.00**
 2 – Sam 16.16 25.67 59% 79% .81**
 Set 2
 4 – Paul 10.33 19.00 84% 52% 0.26
 5 – Jenn 6.73 24.33 261% 87% .61**
Note. PND = percent of all nonoverlapping data [1− (number of overlapping data points in the treatment phase / total number of baseline and treatment data points)]; Tau-U = nonoverlap between intervention and baseline phases with baseline trend control (Parker, Vannest, & Davis, 2011).
Note. PND = percent of all nonoverlapping data [1− (number of overlapping data points in the treatment phase / total number of baseline and treatment data points)]; Tau-U = nonoverlap between intervention and baseline phases with baseline trend control (Parker, Vannest, & Davis, 2011).×
** p < .01.
p < .01.×
Table 3. Measures of level and trend for total number of different words (NDW) and the Monitoring Indicators of Scholarly Language (MISL) score for the participants who received treatment.
Measures of level and trend for total number of different words (NDW) and the Monitoring Indicators of Scholarly Language (MISL) score for the participants who received treatment.×
Participant Baseline avg Last 3 treatment session avg Percent change PND Tau-U
NDW
 Set 1
 1 – Brock 28.67 65.00 126% 95% .92**
 2 – Sam 45.00 147.67 228% 85% .71**
 Set 2
 4 – Paul 16.00 35.67 123% 84% .81**
 5 – Jenn 12.21 56.33 361% 90% .74**
MISL
 Set 1
 1 – Brock 11.67 23.00 97% 100% 1.00**
 2 – Sam 16.16 25.67 59% 79% .81**
 Set 2
 4 – Paul 10.33 19.00 84% 52% 0.26
 5 – Jenn 6.73 24.33 261% 87% .61**
Note. PND = percent of all nonoverlapping data [1− (number of overlapping data points in the treatment phase / total number of baseline and treatment data points)]; Tau-U = nonoverlap between intervention and baseline phases with baseline trend control (Parker, Vannest, & Davis, 2011).
Note. PND = percent of all nonoverlapping data [1− (number of overlapping data points in the treatment phase / total number of baseline and treatment data points)]; Tau-U = nonoverlap between intervention and baseline phases with baseline trend control (Parker, Vannest, & Davis, 2011).×
** p < .01.
p < .01.×
×
Table 4. Measures of level and trend for the two control participants.
Measures of level and trend for the two control participants.×
Participant Entire baseline
First half of data
Second half of data
Percent change PND Tau-U
M M M
NDW
 3 – Casey 38.9 40.0 37.7 −9% 0% −0.13
 6 – Jack 10.7 10.7 10.7 0% 0% −0.02
MISL
 3 – Casey 18.3 17.2 18.7 8% 0% 0.15
 6 – Jack 4.5 4.3 4.6 7% 0% 0.07
Note. PND = percent of all nonoverlapping data [1 − (number of overlapping data points in the treatment phase / total number of baseline and treatment data points)]; Tau-U = nonoverlap between intervention and baseline phases with baseline trend control (Parker, Vannest, & Davis, 2011); NDW = total number of different words; MISL = Monitoring Indicators of Scholarly Language.
Note. PND = percent of all nonoverlapping data [1 − (number of overlapping data points in the treatment phase / total number of baseline and treatment data points)]; Tau-U = nonoverlap between intervention and baseline phases with baseline trend control (Parker, Vannest, & Davis, 2011); NDW = total number of different words; MISL = Monitoring Indicators of Scholarly Language.×
Table 4. Measures of level and trend for the two control participants.
Measures of level and trend for the two control participants.×
Participant Entire baseline
First half of data
Second half of data
Percent change PND Tau-U
M M M
NDW
 3 – Casey 38.9 40.0 37.7 −9% 0% −0.13
 6 – Jack 10.7 10.7 10.7 0% 0% −0.02
MISL
 3 – Casey 18.3 17.2 18.7 8% 0% 0.15
 6 – Jack 4.5 4.3 4.6 7% 0% 0.07
Note. PND = percent of all nonoverlapping data [1 − (number of overlapping data points in the treatment phase / total number of baseline and treatment data points)]; Tau-U = nonoverlap between intervention and baseline phases with baseline trend control (Parker, Vannest, & Davis, 2011); NDW = total number of different words; MISL = Monitoring Indicators of Scholarly Language.
Note. PND = percent of all nonoverlapping data [1 − (number of overlapping data points in the treatment phase / total number of baseline and treatment data points)]; Tau-U = nonoverlap between intervention and baseline phases with baseline trend control (Parker, Vannest, & Davis, 2011); NDW = total number of different words; MISL = Monitoring Indicators of Scholarly Language.×
×
Percent of change from baseline to the end of treatment and PND are popular, are easy to calculate, and enable investigators to compare changes in level across studies. However, these measures are insensitive to baseline trend and autocorrelation, lack statistical power, do not discriminate well among successful interventions, and do not detect improving trend lines during intervention (Parker, Vannest, & Davis, 2011). Therefore, we also calculated Tau-U, which indexes the percentage of overlap minus nonoverlap for all pairwise data comparisons across baseline and treatment phases in a manner that controls for an undesirable positive baseline trend. Probability values and confidence intervals are available because Tau-U conforms to the “S” sampling distribution, similar to nonparametric analyses like Kendall's Rank Correlation and the Mann–Whitney U test. Tau-U is preferable to other effect size indices because it shows no ceiling or floor effects, is minimally influenced by autocorrelation, and discriminates well among samples of data series (Parker, Vannest, Davis, & Sauber, 2011). Tau-U has been summarized as a comprehensive measure of single-subject research effect size because it represents both the level change across phases and the baseline trend (Rakap, 2015). A Tau-U of 0–.65 is considered to be small, .66–.85 is medium, and .86–1.0 is large.
The effect size calculations for the four children who received treatment were based on within-child comparisons of data in the baseline and treatment periods (see Table 3). To assess the stability of the baseline data for the two control children who did not receive treatment, the effect size metrics were calculated on the first half and the last half of the stories that were collected (see Table 4).
Set 1: Participants 1–3
Brock (Participant 1)
Brock created three stories (one per day) on separate days during baseline. His baseline stories contained an average of 28.67 different words (range = 23–33) and earned an average MISL score of 11.67 (range = 10–13). There were decreasing baselines for both measures. With the introduction of the narrative intervention after baseline session 3, the NDW and MISL scores increased immediately with a continued increasing trend across the intervention period. Even though performance on both measures was somewhat variable, there was very little overlap between the baseline and intervention periods (PND = 95% for NDW and 100% for MISL). There was a large change in the level of Brock's narrative skills from the baseline average to measures taken after the last three sessions of the intervention period with a 126% improvement over the baseline average for NDW and a 97% improvement in MISL scores. Tau-U effect sizes of .92 for NDW and 1.0 for MISL were both statistically significant (p < .01). Finally, the level of improvement on both measures was maintained during follow-up testing.
Sam (Participant 2)
Sam remained in baseline while the narrative intervention was implemented with Brock. Participant 2 produced 19 stories on separate days during an extended baseline period while Brock was receiving intervention. Sam's baseline stories contained an average of 45 different words (range = 22–64) and earned an average MISL score of 16.16 (range = 7–24). There was a falling trend for NDW across the 19 baseline stories. The MISL scores increased somewhat over the first 13 stories collected during baseline and then fell sharply over the ensuing six stories. With the introduction of the narrative intervention, Sam's NDW and MISL scores increased immediately, with a continued increasing trend across the intervention period. There was 85% and 79% nonoverlap between the intervention and baseline period for the NDW and MISL scores, respectively. There was also a large change in the level of performance with a 228% improvement in NDW and a 59% improvement in MISL scores. Finally, Sam's Tau-U scores of .71 (NDW) and .81 (MISL) were both significant. We did not collect any follow-up stories from Sam because the study ended immediately after he completed the intervention phase.
Casey (Participant 3)
The third child in Set 1, Casey, remained in baseline throughout the study. The data for NDW and MISL scores were somewhat variable across the baseline period. For example, the NDW varied from 14 to 80 with an overall average of 40.03. There was a slight increase in NDW across the entire baseline period with an average NDW of 40 for the first half of the measurement sessions and an average of 37.7 for the last half of the measurement sessions (see Table 3). This represents a 9% decrease in NDW, which compares to the 127% and 228% increases in NDW for Brock and Sam, respectively. There was a slight increase (8%) in Casey's MISL scores from the first half (M = 17.23) to the second half (M = 18.7) of the baseline measures. However, this increase was much smaller than the increase in MISL scores for Brock (77%) and Sam (59%). Neither of the Tau-U values of −.13 for NDW and .15 for MISL were significant. The data for Casey, who did not receive narrative intervention, provide additional evidence that threats to internal validity related to testing effects and maturation did not influence the changes in the narrative scores that Brock and Sam obtained during the intervention period.
Set 2: Participants 4–6
Paul (Participant 4)
Paul had a falling baseline for NDW with an average of 16 different words (range = 13–18), but his MISL scores were fairly stable across baseline with an average of 10.33 (range = 10–11). Paul received a total of 16 intervention sessions during which he presented a steady, increasing trend for the NDW and MISL scores. There was 85% nonoverlap between the baseline and treatment periods for NDW and 52% nonoverlap between the baseline and treatment periods for the MISL scores. There was a substantial change in level for NDW, as indicated by a 123% improvement over the baseline average and a moderate change in level of the MISL scores (an 84% improvement over the baseline average). His Tau-U score of .81 for NDW was statistically significant (p < .01). However, his Tau-U score for the MISL was .26 (90% CI = −.03–.82) and was not significant, suggesting that the change in NDW was reliable while the change in the MISL score was not. His performance on both measures was maintained at follow-up.
Jenn (Participant 5)
Jenn had falling baselines for both NDW and MISL, with an average of 12.21 different words (range = 3–21) and an average MISL score of 6.73 (range = 2–15). There was 90% and 87% nonoverlap between the intervention and baseline periods for the NDW and MISL scores, respectively, with all overlap occurring during the first five intervention sessions. There were also large changes in the level of performance from the baseline to the intervention period with a 361% improvement in NDW over the baseline average and a 262% improvement in MISL scores over the baseline average. The Tau-U values of .74 and .61 were significant at the p < .01 level for NDW and MISL measures, respectively, indicating that both trends were highly reliable.
Jack (Participant 6)
Jack remained in baseline throughout the study. He missed 1 week of school (baseline sessions 32–38) when his family went on a vacation. Across the entire baseline period, the NDW in his stories varied from 4 to 21 with an average of 10.66. There was no change in NDW from the first half (M = 10.66) to the last half of the baseline measurement sessions (M = 10.66). This represents no change in NDW, compared to the 123% and 361% increases in NDW for Paul and Jenn, respectively. There was only a 7% change in his MISL scores from the first half of the baseline sessions to the last half. The Tau-U scores of −.02 and .07 were not significant. The data for Jack, who did not receive narrative intervention, suggested that repeated testing or maturation were not responsible for the positive changes in NDW and MISL scores obtained by Paul and Jenn during their intervention periods.
Discussion
Numerous authors have identified narrative discourse as an appropriate intervention target for school-age children because difficulties with narration can have negative consequences for socialization and academic success (Crais & Lorch, 1994; Hoffman, 2009; Naremore et al., 1995). This study was designed to test the degree to which a comprehensive, manualized narrative language program based on Kintch's (2013)  C-I model of textual discourse resulted in improvements in the production of self-generated stories by children with language disorders. SKILL incorporated narrative as a target first and then used narrative as a context for teaching multiple literate language features and metacognitive skills that are important for understanding, evaluating, and composing coherent, complex stories.
We used a multiple-baseline, single-subject design with six children with language disorders, four of whom received narrative discourse instruction. The baseline and treatment phases were lagged in two sets of children, with the third child in each set remaining in baseline for the duration of the study. This feature of the design added additional control for testing and maturation threats to internal validity.
Treatment Effects
This multiple-baseline study documented strong positive trends in language productivity and in story complexity that were associated with the onset of treatment. The four children who received treatment made moderate-to-large gains in language productivity as measured by NDW and small-to-large gains in story complexity, as measured by the total MISL score. Gains in NDW and MISL were maintained after intervention for the children we were able to elicit follow-up data from (3 of the 4). The changes in level and trend from the baseline to the intervention phases for the four participants who received intervention along with the observation that the students who remained in baseline did not demonstrate those changes suggest that intervention with SKILL was associated with the differences in the dependent measures, while controlling for threats to internal validity related to history, maturation, subject selection, instrumentation, testing, and statistical regression.
Recall that we reported three metrics of change in level from baseline to intervention phases: percent change from baseline, PND, and Tau-U, which indexes changes in level and the extent of nonoverlap for all pairwise data comparisons across baseline and treatment phases and controls for an undesirable positive baseline trend. For the four children who received intervention, there was greater percent change for NDW (range = 123%–361%) than MISL (range = 59%–261%), with one participant, Jenn, showing much greater change than the others. Similarly, there were somewhat higher levels of percent of nonoverlapping data for NDW (range = 85%–95%) than for MISL (range = 52%–100%). The Tau-U scores provide clues in interpreting these findings, as they show that all the comparisons between baseline and treatment scores were significant at the p < .01 level except for one, that is, Paul's Tau-U of .26 for MISL. The average Tau-U for NDW was .80, and the average Tau-U for MISL was .67, with both being significant at the p < .01 level. Taken together, these analyses suggest that the content and complexity of the children's stories improved considerably as a result of the SKILL intervention, with somewhat greater levels of improvement on the measure of language productivity (NDW) than our measure of story complexity (MISL).
With respect to language productivity, all four children told stories that were longer and contained more diverse vocabulary after intervention began as compared to their baseline stories. In fact, the students who received instruction doubled and tripled the diversity of their vocabulary used in their stories over the course of the study. This is particularly compelling since students were asked to create a spontaneous story from a picture without any support (e.g., story element icons) or extended time to plan their story.
With respect to story complexity, the students' pre-intervention stories were either descriptive in nature or contained one single, simple episode. For example, one student, when shown a picture of a family pointing their fingers at something in the distance created the following story: “They're looking off the cliff. The little boy's pointing. The dad's looking at the thing. That's all I know.” Later, students' self-generated stories contained multiple embedded episodes, complex characters, and multiple settings (i.e., in the morning, at 8:30 a.m., in Salt Lake City), as well as causal connections marked by linguistic terms such as “wanted to,” “decided to,” “because,” and “so.” Students who had received instruction also started including internal response terminology in their stories such as, “He feeled very scared because he didn't know what to do.” or “He was scared because he could not find it.” More elaborated noun phrases such as “a little tree,” “a little house,” or “an old tree” were noted as consistent use of coordinating and subordinating conjunctions and adverbs, such as in the following example from one student's story: “Now she has fifteen dollars but that wasn't enough money.”
Jenn (Participant 5) made the most gains (NDW 361%; MISL 261%), but spent the least number of hours in instruction (10 hr 8 min). She was one of the youngest participants in the study (6;7) and had the highest score on a measure of general language proficiency (CELF-4 = 82) before the study began. Unfortunately, due to similarities in treatment outcomes among the other three participants and small differences among their ages and language skills before entering the study, we are not able to draw any firm conclusions about the relationships among treatment duration, intervention outcomes, age, and pre-intervention language ability using the CELF-4.
Limitations
According to the What Works Clearinghouse Single Case Design Standards (Kratochwill et al., 2010), studies “must include three attempts to demonstrate an intervention effect at three different points in time or with three different phase repetitions” (p. 15). While the current study does not meet this criteria, the participants who remained in baseline address potential threats to internal validity such as history, maturation, instrumentation, and repeated testing by demonstrating that the behaviors measured did not change without the intervention. The results of this study do not confirm that narrative proficiency would only improve when intervention was implemented, but the data demonstrate that performance on two measures of narration did not improve unless the intervention was implemented.
The current study did not investigate the extent to which improvements in oral narrative discourse generalized to written discourse. This is an important area of inquiry that should inform school-based clinicians who work with oral and written language discourse. Future studies should include outcome measures for oral and written discourse skills.
One other potential limitation relevant to school-based settings is that all of the students received one-on-one instruction. This is not always possible in schools. It is possible that students who received instruction in groups may make less progress than the students who participated in the current project. Only future studies conducted using a group design will be able to explore this issue further.
Summary and Clinical Implications
As Dr. Powell (2018)  summarizes in her lead article in this forum, all states require that students are given the opportunity to receive effective, evidence-based instruction to meet their individual needs. The program described in this article is one approach to addressing narrative proficiency that is directly related to the curriculum.
The study described in this article was not conducted at the classroom level; however, SKILL has been implemented in a classroom setting with promising results. Gillam, Olszewski, Fargo, and Gillam (2014)  conducted a nonrandomized feasibility study to investigate whether the SKILL program could be implemented in a classroom and result in measurable gains in vocabulary and narration for first graders. Children in the experimental classroom made clinically significant improvements in narrative and vocabulary while the students in the control classroom did not. These findings provide some early evidence that it is feasible to implement this kind of narrative program in a classroom setting.
The SKILL narrative intervention program consists of three phases of activities that were designed to (a) increase knowledge and use of narrative text structure (e.g., story grammar elements) and the causal and temporal relationships between them for use in creating an accurate textbase, (b) increase knowledge and use of strategies for creating an accurate situation model of the textbase (paraphrasing, sentence combining, linguistic structures for cohesion, coherence), and (c) increase knowledge and use of strategies for integrating information into long-term memory (retelling, summarizing, and composing stories). These phases were designed to be consistent with the basic tenets of Kintsch's (2013)  C-I model. In this case, the program yielded positive and consistent changes in measures of language productivity and story complexity in self-generated stories composed by participants with language disorders.
All four participants who received intervention presented patterns in which stable or decreasing baselines were followed by consistent increases in NDW, and three of the four participants presented this pattern in their MISL scores. In addition, there were small-to-large and statistically significant amounts of nonoverlap between data collected for both measures in the baseline and treatment phases for students who received the narrative discourse intervention.
We believe the manualization of the lesson plans may have contributed to the systematic, consistent outcomes that were obtained by the students who received treatment from four different clinicians with differing levels of experience. The standardization of the instruction within and between the different clinical providers decreases both error variance and the chances of detecting change when no change has been made. This is important for school-based clinicians who may not have the time or resources to spend in the development of curricula for improving narrative language proficiency. In addition, the high level of fidelity of the implementation of the intervention lends to the credibility of the findings and to the likelihood that school-based clinicians may achieve similar outcomes with the students on their caseloads. SLPs in schools do not have time to reinvent the wheel and would benefit by having access to curricular materials that have research evidence to support their use. This study along with others suggest that we might expect positive outcomes in narrative production when our narrative instruction involves (a) the use of activities that strengthen knowledge of story structure and metacognitive skills, (b) multiple opportunities to compose and evaluate aspects of complex narratives, and (c) systematically decreasing the amount of support given to students as they edit and create their own stories.
Acknowledgments
This research was supported by a grant from the Institute on Educational Sciences, National Center on Special Education Research (R324A100063). The authors thank Chad Bingham and Tami Pyfer for their assistance in carrying out this study and the Logan School District for allowing us access to the participants. Special thanks to Allison Hancock, Natalie Nelson Buttars, Julise Nelson, Casey Ragan, Brittany Martinez, Shannon Davenport, MaryAnn Hammond-Stenquist, Andrea Deakin, Sara Hicken, Ariel Hendrix, Sara Hegsted, Rebecca Mortensen, and Kelli Reese for collecting, transcribing, and coding the stories.
References
American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.). Washington, DC: Author.
American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.). Washington, DC: Author.×
Bishop, D. V. M. (2014). Ten questions about terminology for children with unexplained language problems. International Journal of Language and Communication Disorders, 49(4), 381–415. https://doi.org/10.1111/1460-6984.12101 [Article] [PubMed]
Bishop, D. V. M. (2014). Ten questions about terminology for children with unexplained language problems. International Journal of Language and Communication Disorders, 49(4), 381–415. https://doi.org/10.1111/1460-6984.12101 [Article] [PubMed]×
Bishop, D. V. M., & Adams, C. (1992). Comprehension problems in children with specific language impairment: Literal and inferential meaning. Journal of Speech and Hearing Research, 35(1), 119–129. [Article] [PubMed]
Bishop, D. V. M., & Adams, C. (1992). Comprehension problems in children with specific language impairment: Literal and inferential meaning. Journal of Speech and Hearing Research, 35(1), 119–129. [Article] [PubMed]×
Boudreau, D. (2008). Narrative abilities: Advances in research and implications for clinical practice. Topics in Language Disorders, 28(2), 99–114. [Article]
Boudreau, D. (2008). Narrative abilities: Advances in research and implications for clinical practice. Topics in Language Disorders, 28(2), 99–114. [Article] ×
Bracken, B. A., & McCallum, R. S. (1998). The Universal Nonverbal Intelligence Test. Itasca, IL: Riverside.
Bracken, B. A., & McCallum, R. S. (1998). The Universal Nonverbal Intelligence Test. Itasca, IL: Riverside.×
Brown, J., Garzarek, J., & Donegan, K. (2014). Effects of a narrative intervention on story retelling in at-risk young children. Topics in Early Childhood Special Education, 34(3), 154–164. https://doi.org/10.1177/0271121414536447 [Article]
Brown, J., Garzarek, J., & Donegan, K. (2014). Effects of a narrative intervention on story retelling in at-risk young children. Topics in Early Childhood Special Education, 34(3), 154–164. https://doi.org/10.1177/0271121414536447 [Article] ×
Cirrin, F. M., & Gillam, R. B. (2008). Language intervention practices for school-age children with spoken language disorders: A systematic review. Language, Speech, and Hearing Services in Schools, 39(1), S110–S137. [Article] [PubMed]
Cirrin, F. M., & Gillam, R. B. (2008). Language intervention practices for school-age children with spoken language disorders: A systematic review. Language, Speech, and Hearing Services in Schools, 39(1), S110–S137. [Article] [PubMed]×
Crais, E. R., & Lorch, N. (1994). Oral narratives in school-age children. Topics in Language Disorders, 14(3), 13–28. [Article]
Crais, E. R., & Lorch, N. (1994). Oral narratives in school-age children. Topics in Language Disorders, 14(3), 13–28. [Article] ×
Fey, M. E., Catts, H. W., Proctor-Williams, K., Tomblin, J. B., & Zhang, X. (2004). Oral and written story composition skills of children with language impairment. Journal of Speech, Language, and Hearing Research, 47(6), 1301–1318. [Article]
Fey, M. E., Catts, H. W., Proctor-Williams, K., Tomblin, J. B., & Zhang, X. (2004). Oral and written story composition skills of children with language impairment. Journal of Speech, Language, and Hearing Research, 47(6), 1301–1318. [Article] ×
Fujiki, M., Spackman, M. P., Brinton, B., & Illig, T. (2008). Ability of children with language impairment to understand emotion conveyed by prosody in a narrative passage. International Journal of Language & Communication Disorders, 43(3), 330–345. [Article] [PubMed]
Fujiki, M., Spackman, M. P., Brinton, B., & Illig, T. (2008). Ability of children with language impairment to understand emotion conveyed by prosody in a narrative passage. International Journal of Language & Communication Disorders, 43(3), 330–345. [Article] [PubMed]×
Gillam, R. B., & Johnston, J. R. (1992). Spoken and written language relationships in language/learning-impaired and normally achieving school-age children. Journal of Speech and Hearing Research, 35(6), 1303–1315. [Article] [PubMed]
Gillam, R. B., & Johnston, J. R. (1992). Spoken and written language relationships in language/learning-impaired and normally achieving school-age children. Journal of Speech and Hearing Research, 35(6), 1303–1315. [Article] [PubMed]×
Gillam, R. B., & Pearson, N. (2004). Test of Narrative Language. Austin, TX: Pro-Ed.
Gillam, R. B., & Pearson, N. (2004). Test of Narrative Language. Austin, TX: Pro-Ed.×
Gillam, S., Olszewski, A., Fargo, J., & Gillam, R. (2014). Classroom-based narrative and vocabulary instruction: Results of an early-stage, nonrandomized comparison study. Language, Speech, and Hearing Services in Schools, 45, 127–136. [Article] [PubMed]
Gillam, S., Olszewski, A., Fargo, J., & Gillam, R. (2014). Classroom-based narrative and vocabulary instruction: Results of an early-stage, nonrandomized comparison study. Language, Speech, and Hearing Services in Schools, 45, 127–136. [Article] [PubMed]×
Gillam, S. L., Fargo, J. D., & Robertson, K. S. C. (2009). Comprehension of expository text: Insights gained from think-aloud data. American Journal of Speech-Language Pathology, 18(1), 82–94. [Article] [PubMed]
Gillam, S. L., Fargo, J. D., & Robertson, K. S. C. (2009). Comprehension of expository text: Insights gained from think-aloud data. American Journal of Speech-Language Pathology, 18(1), 82–94. [Article] [PubMed]×
Gillam, S. L., Gillam, R. B., Fargo, J. D., Olszewski, A., & Segura, H. (2016). Monitoring indicators of scholarly language: A progress-monitoring instrument for measuring narrative discourse skills. Communication Disorders Quarterly, 38(2), 96–106. https://doi.org/10.1177/1525740116651442 [Article]
Gillam, S. L., Gillam, R. B., Fargo, J. D., Olszewski, A., & Segura, H. (2016). Monitoring indicators of scholarly language: A progress-monitoring instrument for measuring narrative discourse skills. Communication Disorders Quarterly, 38(2), 96–106. https://doi.org/10.1177/1525740116651442 [Article] ×
Gillam, S. L., Gillam, R. B., & Laing, C. E. (2014). Supporting knowledge in language and literacy. Logan, UT: Utah State University.
Gillam, S. L., Gillam, R. B., & Laing, C. E. (2014). Supporting knowledge in language and literacy. Logan, UT: Utah State University.×
Gillam, S. L., Gillam, R. B., & Reece, K. (2012). Language outcomes of contextualized and decontextualized language intervention: Results of an early efficacy study. Language, Speech, and Hearing Services in Schools, 43(3), 276–291. https://doi.org/10.1044/0161-1461(2011/11-0022) [Article] [PubMed]
Gillam, S. L., Gillam, R. B., & Reece, K. (2012). Language outcomes of contextualized and decontextualized language intervention: Results of an early efficacy study. Language, Speech, and Hearing Services in Schools, 43(3), 276–291. https://doi.org/10.1044/0161-1461(2011/11-0022) [Article] [PubMed]×
Goencue, A., & Klein, E. L. (2001). Children in play, story, and school. New York, NY: Guilford.
Goencue, A., & Klein, E. L. (2001). Children in play, story, and school. New York, NY: Guilford.×
Greenhalgh, K. S., & Strong, C. J. (2001). Literate language features in spoken narratives of children with typical language and children with language impairments. Language, Speech, and Hearing Services in Schools, 32(2), 114–125. [Article] [PubMed]
Greenhalgh, K. S., & Strong, C. J. (2001). Literate language features in spoken narratives of children with typical language and children with language impairments. Language, Speech, and Hearing Services in Schools, 32(2), 114–125. [Article] [PubMed]×
Hayward, D., & Schneider, P. (2000). Effectiveness of teaching story grammar knowledge to pre-school children with language impairment: An exploratory study. Child Language Teaching and Therapy, 16(3), 255–284. https://doi.org/10.1177/026565900001600303 [Article]
Hayward, D., & Schneider, P. (2000). Effectiveness of teaching story grammar knowledge to pre-school children with language impairment: An exploratory study. Child Language Teaching and Therapy, 16(3), 255–284. https://doi.org/10.1177/026565900001600303 [Article] ×
Heilmann, J., Miller, J., Nockerts, A., & Dunaway, C. (2010). Properties of the narrative scoring scheme using narrative retells in young school-age children. American Journal of Speech-Language Pathology, 19, 154–166. [Article] [PubMed]
Heilmann, J., Miller, J., Nockerts, A., & Dunaway, C. (2010). Properties of the narrative scoring scheme using narrative retells in young school-age children. American Journal of Speech-Language Pathology, 19, 154–166. [Article] [PubMed]×
Hoffman, L. M. (2009). Narrative language intervention intensity and dosage: Telling the whole story. Topics in Language Disorders, 29(4), 329–343. [Article]
Hoffman, L. M. (2009). Narrative language intervention intensity and dosage: Telling the whole story. Topics in Language Disorders, 29(4), 329–343. [Article] ×
Kaderavek, J. N., & Sulzby, E. (2000). Narrative production by children with and without specific language impairment: Oral narratives and emergent readings. Journal of Speech, Language, and Hearing Research, 43(1), 34–49. [Article]
Kaderavek, J. N., & Sulzby, E. (2000). Narrative production by children with and without specific language impairment: Oral narratives and emergent readings. Journal of Speech, Language, and Hearing Research, 43(1), 34–49. [Article] ×
Kintsch, W. (2013). Revisiting the construction-integration model of text comprehension and its implications for instruction. In Alvermann, D. E., Unrau, N. J., & Ruddell, R. B. (Eds.), Theoretical models and processes of reading, 6th ed. (pp. 807–839). Newark, DE: International Reading Association. [Article]
Kintsch, W. (2013). Revisiting the construction-integration model of text comprehension and its implications for instruction. In Alvermann, D. E., Unrau, N. J., & Ruddell, R. B. (Eds.), Theoretical models and processes of reading, 6th ed. (pp. 807–839). Newark, DE: International Reading Association. [Article] ×
Kratochwill, T. R., Hitchcock, J., Horner, R. H., Lvein, J. R., Odo, S. L., Rindskopf, D., & Shadish, W. R. M. (2010). Single case designs technical documentation. Retrieved from http://ies.ed.gov/
Kratochwill, T. R., Hitchcock, J., Horner, R. H., Lvein, J. R., Odo, S. L., Rindskopf, D., & Shadish, W. R. M. (2010). Single case designs technical documentation. Retrieved from http://ies.ed.gov/ ×
Laing, S., & Kamhi, A. (2002). The use of think-aloud protocols to compare inferencing abilities of average and below-average readers. Journal of Learning Disabilities, 35, 437–448. https://doi.org/10.1177/00222194020350050401 [Article]
Laing, S., & Kamhi, A. (2002). The use of think-aloud protocols to compare inferencing abilities of average and below-average readers. Journal of Learning Disabilities, 35, 437–448. https://doi.org/10.1177/00222194020350050401 [Article] ×
Lynch, J. S., van den Broek, P., Kremer, K. E., Kendeou, P., White, M. J., & Lorch, E. P. (2008). The development of narrative comprehension and its relation to other early reading skills. Reading Psychology, 29(4), 327–333. [Article]
Lynch, J. S., van den Broek, P., Kremer, K. E., Kendeou, P., White, M. J., & Lorch, E. P. (2008). The development of narrative comprehension and its relation to other early reading skills. Reading Psychology, 29(4), 327–333. [Article] ×
Maddox, T. (2009). Little croc and whale . Marylebone, England: Piccadilly Press.
Maddox, T. (2009). Little croc and whale . Marylebone, England: Piccadilly Press.×
McFadden, T., & Gillam, R. (1996). An examination of the quality of narratives produced by children with language disorders. Language, Speech, and Hearing Services in Schools, 27(1), 48–56. [Article]
McFadden, T., & Gillam, R. (1996). An examination of the quality of narratives produced by children with language disorders. Language, Speech, and Hearing Services in Schools, 27(1), 48–56. [Article] ×
Miller, J., Andriacchi, K., & Nockerts, A. (2015). Assessing language production using SALT software: A clinician's guide to language sample analysis (2nd ed.). Madison, WI: University of Wisconsin, Waisman Center, Language Analysis Laboratory.
Miller, J., Andriacchi, K., & Nockerts, A. (2015). Assessing language production using SALT software: A clinician's guide to language sample analysis (2nd ed.). Madison, WI: University of Wisconsin, Waisman Center, Language Analysis Laboratory.×
Morgan, D. L., & Morgan, R. K. (2009). Single-case research methods for the behavioral and health sciences. Los Angeles, CA: Sage. [Article]
Morgan, D. L., & Morgan, R. K. (2009). Single-case research methods for the behavioral and health sciences. Los Angeles, CA: Sage. [Article] ×
Naremore, R. C., Densmore, A. E., & Harman, D. R. (1995). Language intervention with school-aged children: Conversation, narrative, and text. San Diego, CA: Singular.
Naremore, R. C., Densmore, A. E., & Harman, D. R. (1995). Language intervention with school-aged children: Conversation, narrative, and text. San Diego, CA: Singular.×
Nathanson, R., Crank, J. N., Saywitz, K. J., & Ruegg, E. (2007). Enhancing the oral narratives of children with learning disabilities. Reading & Writing Quarterly, 23(4), 315–331. https://doi.org/10.1080/10573560701501610 [Article]
Nathanson, R., Crank, J. N., Saywitz, K. J., & Ruegg, E. (2007). Enhancing the oral narratives of children with learning disabilities. Reading & Writing Quarterly, 23(4), 315–331. https://doi.org/10.1080/10573560701501610 [Article] ×
National Governors Association Center for Best Practices and Council of Chief State School Officers. (2010). Common Core State Standards for English language arts and literacy in history/social studies, science, and technical subjects. Washington, DC: Author.
National Governors Association Center for Best Practices and Council of Chief State School Officers. (2010). Common Core State Standards for English language arts and literacy in history/social studies, science, and technical subjects. Washington, DC: Author.×
Newman, R. M., & McGregor, K. K. (2006). Teachers and laypersons discern quality differences between narratives produced by children with or without specific language impairment. Journal of Speech, Language, and Hearing Research, 49(5), 1022–1036. [Article]
Newman, R. M., & McGregor, K. K. (2006). Teachers and laypersons discern quality differences between narratives produced by children with or without specific language impairment. Journal of Speech, Language, and Hearing Research, 49(5), 1022–1036. [Article] ×
Parker, R. I., Vannest, K. J., & Davis, J. L. (2011a). Effect size in single-case research: A review of nine non-overlap techniques. Behavior Modification, 35(4), 303–322. [Article]
Parker, R. I., Vannest, K. J., & Davis, J. L. (2011a). Effect size in single-case research: A review of nine non-overlap techniques. Behavior Modification, 35(4), 303–322. [Article] ×
Parker, R. I., Vannest, K. J., Davis, J. L., & Sauber, S. B. (2011b). Combining non-overlap and trend for single-case research: Tau-U. Behavior Therapy, 42, 284–299. [Article]
Parker, R. I., Vannest, K. J., Davis, J. L., & Sauber, S. B. (2011b). Combining non-overlap and trend for single-case research: Tau-U. Behavior Therapy, 42, 284–299. [Article] ×
Perfetti, C., & Stafura, J. (2014). Work knowledge in a theory of reading comprehension. Scientific Studies of Reading, 18(1), 22–37. https://doi.org/10.1080/10888438.2013.827687 [Article]
Perfetti, C., & Stafura, J. (2014). Work knowledge in a theory of reading comprehension. Scientific Studies of Reading, 18(1), 22–37. https://doi.org/10.1080/10888438.2013.827687 [Article] ×
Petersen, D., & Spencer, T. (2016). Using narrative intervention to accelerate canonical story grammar and complex language growth in culturally diverse preschoolers. Topics in Language Disorders, 36(1), 6–19. [Article]
Petersen, D., & Spencer, T. (2016). Using narrative intervention to accelerate canonical story grammar and complex language growth in culturally diverse preschoolers. Topics in Language Disorders, 36(1), 6–19. [Article] ×
Powell, R. K. (2018). Unique contributors to the curriculum: From research to practice for speech-language pathologists in schools. Language, Speech, and Hearing Services in Schools, 49, 140–147.
Powell, R. K. (2018). Unique contributors to the curriculum: From research to practice for speech-language pathologists in schools. Language, Speech, and Hearing Services in Schools, 49, 140–147.×
Rakap, S. (2015). Effect sizes as result interpretation aids in single-subject experimental research: Description and application of four nonoverlap methods. British Journal of Special Education, 42(1), 11–33. https://doi.org/10.1111?1467-8578.12091 [Article]
Rakap, S. (2015). Effect sizes as result interpretation aids in single-subject experimental research: Description and application of four nonoverlap methods. British Journal of Special Education, 42(1), 11–33. https://doi.org/10.1111?1467-8578.12091 [Article] ×
Scott, C. M., & Windsor, J. (2000). General language performance measures in spoken and written narrative and expository discourse of school-age children with language learning disabilities. Journal of Speech, Language, and Hearing Research, 43(2), 324–340. [Article]
Scott, C. M., & Windsor, J. (2000). General language performance measures in spoken and written narrative and expository discourse of school-age children with language learning disabilities. Journal of Speech, Language, and Hearing Research, 43(2), 324–340. [Article] ×
Scruggs, T. E., & Mastropieri, M. A. (2000). The effectiveness of mnemonic instruction for students with learning and behavior problems: An update and research synthesis. Journal of Behavioral Education, 10(2), 163–173. https://doi.org/10.1023/A:1016640214368 [Article]
Scruggs, T. E., & Mastropieri, M. A. (2000). The effectiveness of mnemonic instruction for students with learning and behavior problems: An update and research synthesis. Journal of Behavioral Education, 10(2), 163–173. https://doi.org/10.1023/A:1016640214368 [Article] ×
Snow, C. E., Burns, M. S., & Griffin, P. (1998). Preventing reading difficulties in young children. Washington, DC: National Academy Press.
Snow, C. E., Burns, M. S., & Griffin, P. (1998). Preventing reading difficulties in young children. Washington, DC: National Academy Press.×
Spencer, T., Petersen, D., Slocum, T., & Allen, M. (2015). Large group narrative intervention in Head Start preschools: Implications for response to intervention. Journal of Early Childhood Research, 13(2), 196–217. [Article]
Spencer, T., Petersen, D., Slocum, T., & Allen, M. (2015). Large group narrative intervention in Head Start preschools: Implications for response to intervention. Journal of Early Childhood Research, 13(2), 196–217. [Article] ×
Swanson, L. A., Fey, M. E., Mills, C. E., & Hood, L. S. (2005). Use of narrative-based language intervention with children who have specific language impairment. American Journal of Speech-Language Pathology, 14(2), 131–143. https://doi.org/10.1044/1058-0360(2005/014) [Article] [PubMed]
Swanson, L. A., Fey, M. E., Mills, C. E., & Hood, L. S. (2005). Use of narrative-based language intervention with children who have specific language impairment. American Journal of Speech-Language Pathology, 14(2), 131–143. https://doi.org/10.1044/1058-0360(2005/014) [Article] [PubMed]×
Tomblin, J. B., Records, N. L., Buckwalter, P., Zhang, X., Smith, E., & O'Brien, M. (1997). Prevalence of specific language impairment in kindergarten children. Journal of Speech, Language, and Hearing Research, 40(6), 1245–1260. https://doi.org/10.1044/jslhr.4006.1245 [Article]
Tomblin, J. B., Records, N. L., Buckwalter, P., Zhang, X., Smith, E., & O'Brien, M. (1997). Prevalence of specific language impairment in kindergarten children. Journal of Speech, Language, and Hearing Research, 40(6), 1245–1260. https://doi.org/10.1044/jslhr.4006.1245 [Article] ×
Ukrainetz, T. (1998). Stickwriting stories: A quick and easy narrative representation strategy. Language, Speech, and Hearing Services in Schools, 29, 197–206. [Article] [PubMed]
Ukrainetz, T. (1998). Stickwriting stories: A quick and easy narrative representation strategy. Language, Speech, and Hearing Services in Schools, 29, 197–206. [Article] [PubMed]×
van den Broek, P., Linzie, B., Fletcher, C., & Marsolek, C. J. (2000). The role of causal discourse structure in narrative writing. Memory & Cognition, 28(5), 711–721. https://doi.org/10.3758/BF03198405 [Article] [PubMed]
van den Broek, P., Linzie, B., Fletcher, C., & Marsolek, C. J. (2000). The role of causal discourse structure in narrative writing. Memory & Cognition, 28(5), 711–721. https://doi.org/10.3758/BF03198405 [Article] [PubMed]×
Wellman, R. L., Lewis, B. A., Freebairn, L. A., Avrich, A. A., Hansen, A. J., & Stein, C. M. (2011). Narrative ability of children with speech sound disorders and the prediction of later literacy skills. Language, Speech, and Hearing Services in Schools, 42, 561–579. https://doi.org/10.1044/0161-1461(2011/10-0038) [Article] [PubMed]
Wellman, R. L., Lewis, B. A., Freebairn, L. A., Avrich, A. A., Hansen, A. J., & Stein, C. M. (2011). Narrative ability of children with speech sound disorders and the prediction of later literacy skills. Language, Speech, and Hearing Services in Schools, 42, 561–579. https://doi.org/10.1044/0161-1461(2011/10-0038) [Article] [PubMed]×
Wiig, E., Semel, E., & Secord, W. (2003). Clinical Evaluation of Language Fundamentals–Fourth Edition. San Antonio, TX: Psychological Corporation.
Wiig, E., Semel, E., & Secord, W. (2003). Clinical Evaluation of Language Fundamentals–Fourth Edition. San Antonio, TX: Psychological Corporation.×
Wright, H. H., & Newhoff, M. (2001). Narration abilities of children with language-learning disabilities in response to oral and written stimuli. American Journal of Speech-Language Pathology, (10), 432–441. [Article]
Wright, H. H., & Newhoff, M. (2001). Narration abilities of children with language-learning disabilities in response to oral and written stimuli. American Journal of Speech-Language Pathology, (10), 432–441. [Article] ×
Yawkey, T. D., Aronin, E. L., Streett, M. A., & Hinojosa, O. M. (1974). Teaching oral language to young Mexican-Americans. Elementary English, 51(2), 198–202.
Yawkey, T. D., Aronin, E. L., Streett, M. A., & Hinojosa, O. M. (1974). Teaching oral language to young Mexican-Americans. Elementary English, 51(2), 198–202.×
Appendix
List of Single-Scene Pictures Used to Elicit Narratives (In Order)
  1. Girl playing in leaves

  2. Beach scene

  3. People on beach fishing

  4. People on beach running to waves

  5. Men in a bike race

  6. People ice skating on a lake

  7. A cat and dog looking curious

  8. A boy pointing at a goat

  9. A man underwater diving with fish

  10. A baby reaching to pet a dog

  11. A trainer petting two dolphins

  12. A duck and her babies next to a pond

  13. A mom and baby elephant touching

  14. A family hiking and pointing to something

  15. A brush fire on a mountain next to houses

  16. A man standing in a flooded area with houses around

  17. A girl wakeboarding on her knees

  18. A girl skipping on the beach

  19. A helicopter flying

  20. A girl eating a cookie and making a “yucky” face

  21. A man kayaking in rapids

  22. A cartoon character walking into a crocodile's mouth

  23. Three boys playing football in the mud

  24. Kids looking out of windows in a school bus

  25. Mountain rescuers on skies taking someone down mountain

  26. A couple posing while skiing

  27. A boy, face down in snow after falling

  28. A boy rolling a snow ball as if to make a snowman

  29. Girls playing soccer

  30. A girl with head and arms in stocks

  31. Kids playing on a rope swing

  32. A big and little dog looking at each other

Footnotes
1 The diagnostic label “developmental language disorder” is used in the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (American Psychiatric Association, 2013). In the scientific literature, these children are often referred to as having specific language impairment.
The diagnostic label “developmental language disorder” is used in the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (American Psychiatric Association, 2013). In the scientific literature, these children are often referred to as having specific language impairment.×
2 Quoted material appears with permission of the publisher. Copyright © Picadilly Press Ltd.
Quoted material appears with permission of the publisher. Copyright © Picadilly Press Ltd.×
Figure 1.

Set 1 of the multiple-baseline design for the outcome measures obtained from the Monitoring Indicators of Scholarly Language (MISL) rubric and the number of different words (NDW) at each time-point.

 Set 1 of the multiple-baseline design for the outcome measures obtained from the Monitoring Indicators of Scholarly Language (MISL) rubric and the number of different words (NDW) at each time-point.
Figure 1.

Set 1 of the multiple-baseline design for the outcome measures obtained from the Monitoring Indicators of Scholarly Language (MISL) rubric and the number of different words (NDW) at each time-point.

×
Figure 2.

Set 2 of the multiple-baseline design for the outcome variables obtained from the Monitoring Indicators of Scholarly Language (MISL) rubric and the number of different words (NDW) at each time-point.

 Set 2 of the multiple-baseline design for the outcome variables obtained from the Monitoring Indicators of Scholarly Language (MISL) rubric and the number of different words (NDW) at each time-point.
Figure 2.

Set 2 of the multiple-baseline design for the outcome variables obtained from the Monitoring Indicators of Scholarly Language (MISL) rubric and the number of different words (NDW) at each time-point.

×
Table 1. Descriptive data on the six participants before treatment.
Descriptive data on the six participants before treatment.×
Variables Participants
Brock Sam Casey Paul Jenn Jack
Gender M M F M F M
Age [years;months] 7;9 6;7 10;4 8;1 6;7 7;8
CELF-4 73 58 84 60 82 81
TNL 73 58 76 76 85 76
UNIT 12 11 7 10 12 8
Note. CELF-4 = Core Language Standard Score from the Clinical Evaluation of Language Fundamentals–Fourth Edition (Wiig, Semel, & Secord, 2003); TNL = Narrative Language Ability Index from the Test of Narrative Language (Gillam & Pearson, 2004); UNIT = screening standard score (M = 10, SD = 3) from the Universal Nonverbal Intelligence Test (Bracken & McCallum, 1998).
Note. CELF-4 = Core Language Standard Score from the Clinical Evaluation of Language Fundamentals–Fourth Edition (Wiig, Semel, & Secord, 2003); TNL = Narrative Language Ability Index from the Test of Narrative Language (Gillam & Pearson, 2004); UNIT = screening standard score (M = 10, SD = 3) from the Universal Nonverbal Intelligence Test (Bracken & McCallum, 1998).×
Table 1. Descriptive data on the six participants before treatment.
Descriptive data on the six participants before treatment.×
Variables Participants
Brock Sam Casey Paul Jenn Jack
Gender M M F M F M
Age [years;months] 7;9 6;7 10;4 8;1 6;7 7;8
CELF-4 73 58 84 60 82 81
TNL 73 58 76 76 85 76
UNIT 12 11 7 10 12 8
Note. CELF-4 = Core Language Standard Score from the Clinical Evaluation of Language Fundamentals–Fourth Edition (Wiig, Semel, & Secord, 2003); TNL = Narrative Language Ability Index from the Test of Narrative Language (Gillam & Pearson, 2004); UNIT = screening standard score (M = 10, SD = 3) from the Universal Nonverbal Intelligence Test (Bracken & McCallum, 1998).
Note. CELF-4 = Core Language Standard Score from the Clinical Evaluation of Language Fundamentals–Fourth Edition (Wiig, Semel, & Secord, 2003); TNL = Narrative Language Ability Index from the Test of Narrative Language (Gillam & Pearson, 2004); UNIT = screening standard score (M = 10, SD = 3) from the Universal Nonverbal Intelligence Test (Bracken & McCallum, 1998).×
×
Table 2. Intervention duration statistics for the participants who received intervention.
Intervention duration statistics for the participants who received intervention.×
Intervention details Participants
Brock Sam Paul Jenn
Weeks in intervention 8 7 8 6
Number of sessions 24 18 22 13
Average minutes per session minutes of intervention 35 34 38 47
 Week 1 108 84 74 101
 Week 2 146 101 113 97
 Week 3 100 86 78 91
 Week 4 101 108 118 102
 Week 5 125 106 120 138
 Week 6 93 37 119 79
 Week 7 88 50 91
 Week 8 85 111
Total hours of therapy 14 hr 6 min 10 hr 10 min 13 hr 44 min 10 hr 8 min
Table 2. Intervention duration statistics for the participants who received intervention.
Intervention duration statistics for the participants who received intervention.×
Intervention details Participants
Brock Sam Paul Jenn
Weeks in intervention 8 7 8 6
Number of sessions 24 18 22 13
Average minutes per session minutes of intervention 35 34 38 47
 Week 1 108 84 74 101
 Week 2 146 101 113 97
 Week 3 100 86 78 91
 Week 4 101 108 118 102
 Week 5 125 106 120 138
 Week 6 93 37 119 79
 Week 7 88 50 91
 Week 8 85 111
Total hours of therapy 14 hr 6 min 10 hr 10 min 13 hr 44 min 10 hr 8 min
×
Table 3. Measures of level and trend for total number of different words (NDW) and the Monitoring Indicators of Scholarly Language (MISL) score for the participants who received treatment.
Measures of level and trend for total number of different words (NDW) and the Monitoring Indicators of Scholarly Language (MISL) score for the participants who received treatment.×
Participant Baseline avg Last 3 treatment session avg Percent change PND Tau-U
NDW
 Set 1
 1 – Brock 28.67 65.00 126% 95% .92**
 2 – Sam 45.00 147.67 228% 85% .71**
 Set 2
 4 – Paul 16.00 35.67 123% 84% .81**
 5 – Jenn 12.21 56.33 361% 90% .74**
MISL
 Set 1
 1 – Brock 11.67 23.00 97% 100% 1.00**
 2 – Sam 16.16 25.67 59% 79% .81**
 Set 2
 4 – Paul 10.33 19.00 84% 52% 0.26
 5 – Jenn 6.73 24.33 261% 87% .61**
Note. PND = percent of all nonoverlapping data [1− (number of overlapping data points in the treatment phase / total number of baseline and treatment data points)]; Tau-U = nonoverlap between intervention and baseline phases with baseline trend control (Parker, Vannest, & Davis, 2011).
Note. PND = percent of all nonoverlapping data [1− (number of overlapping data points in the treatment phase / total number of baseline and treatment data points)]; Tau-U = nonoverlap between intervention and baseline phases with baseline trend control (Parker, Vannest, & Davis, 2011).×
** p < .01.
p < .01.×
Table 3. Measures of level and trend for total number of different words (NDW) and the Monitoring Indicators of Scholarly Language (MISL) score for the participants who received treatment.
Measures of level and trend for total number of different words (NDW) and the Monitoring Indicators of Scholarly Language (MISL) score for the participants who received treatment.×
Participant Baseline avg Last 3 treatment session avg Percent change PND Tau-U
NDW
 Set 1
 1 – Brock 28.67 65.00 126% 95% .92**
 2 – Sam 45.00 147.67 228% 85% .71**
 Set 2
 4 – Paul 16.00 35.67 123% 84% .81**
 5 – Jenn 12.21 56.33 361% 90% .74**
MISL
 Set 1
 1 – Brock 11.67 23.00 97% 100% 1.00**
 2 – Sam 16.16 25.67 59% 79% .81**
 Set 2
 4 – Paul 10.33 19.00 84% 52% 0.26
 5 – Jenn 6.73 24.33 261% 87% .61**
Note. PND = percent of all nonoverlapping data [1− (number of overlapping data points in the treatment phase / total number of baseline and treatment data points)]; Tau-U = nonoverlap between intervention and baseline phases with baseline trend control (Parker, Vannest, & Davis, 2011).
Note. PND = percent of all nonoverlapping data [1− (number of overlapping data points in the treatment phase / total number of baseline and treatment data points)]; Tau-U = nonoverlap between intervention and baseline phases with baseline trend control (Parker, Vannest, & Davis, 2011).×
** p < .01.
p < .01.×
×
Table 4. Measures of level and trend for the two control participants.
Measures of level and trend for the two control participants.×
Participant Entire baseline
First half of data
Second half of data
Percent change PND Tau-U
M M M
NDW
 3 – Casey 38.9 40.0 37.7 −9% 0% −0.13
 6 – Jack 10.7 10.7 10.7 0% 0% −0.02
MISL
 3 – Casey 18.3 17.2 18.7 8% 0% 0.15
 6 – Jack 4.5 4.3 4.6 7% 0% 0.07
Note. PND = percent of all nonoverlapping data [1 − (number of overlapping data points in the treatment phase / total number of baseline and treatment data points)]; Tau-U = nonoverlap between intervention and baseline phases with baseline trend control (Parker, Vannest, & Davis, 2011); NDW = total number of different words; MISL = Monitoring Indicators of Scholarly Language.
Note. PND = percent of all nonoverlapping data [1 − (number of overlapping data points in the treatment phase / total number of baseline and treatment data points)]; Tau-U = nonoverlap between intervention and baseline phases with baseline trend control (Parker, Vannest, & Davis, 2011); NDW = total number of different words; MISL = Monitoring Indicators of Scholarly Language.×
Table 4. Measures of level and trend for the two control participants.
Measures of level and trend for the two control participants.×
Participant Entire baseline
First half of data
Second half of data
Percent change PND Tau-U
M M M
NDW
 3 – Casey 38.9 40.0 37.7 −9% 0% −0.13
 6 – Jack 10.7 10.7 10.7 0% 0% −0.02
MISL
 3 – Casey 18.3 17.2 18.7 8% 0% 0.15
 6 – Jack 4.5 4.3 4.6 7% 0% 0.07
Note. PND = percent of all nonoverlapping data [1 − (number of overlapping data points in the treatment phase / total number of baseline and treatment data points)]; Tau-U = nonoverlap between intervention and baseline phases with baseline trend control (Parker, Vannest, & Davis, 2011); NDW = total number of different words; MISL = Monitoring Indicators of Scholarly Language.
Note. PND = percent of all nonoverlapping data [1 − (number of overlapping data points in the treatment phase / total number of baseline and treatment data points)]; Tau-U = nonoverlap between intervention and baseline phases with baseline trend control (Parker, Vannest, & Davis, 2011); NDW = total number of different words; MISL = Monitoring Indicators of Scholarly Language.×
×