Assessment of Language and Literacy a Process of Hypothesis Testing for Individual Differences

LWW/TLD TLD200067 January 27, 2011 15:34 Char Count= 0

Top Lang Disorders Vol. 31, No. 1, pp. 24–39 Copyright c 2011 Wolters Kluwer Health | Lippincott Williams & Wilkins Assessment of Language and Literacy A Process of Hypothesis Testing for Individual Differences

Cheryl M. Scott, PhD

Purpose: Older school-aged children and adolescents with persistent language and literacy impairments vary in their individual profiles of linguistic strengths and weaknesses. Given the mul- tidimensional nature and complexity of language, designing an assessment protocol capable of uncovering linguistic variation is challenging. A process of clinical reasoning characterized as hypothesis testing is described to assist in language and literacy assessment. In-depth and explanatory assessment can contribute to establishing a common language across professionals and disciplines concerned with these impairments. Methods: Hypothesis testing relies on (a) review of research on language impairment variations found in the school-aged population and (b) careful analysis of content and form of commonly used assessment tools. A case study illustrates application of this material with a 10-year-old child. Results: Major sources of language variation include language and reading profiles, linguistic components and levels (word, sentence, text), gatekeeping factors, and explanations of language and literacy impairments. Three hypotheses and assessment tools that would be useful in attempts to confirm each hypothesis are described. Conclusions: Uncov- ering the “core” features of a persistent language and/or literacy impairment is difficult. Casting the process as one of testing hypotheses derived from research and analysis of testing tools has the potential to (a) assist language clinicians in determining best assessment practices and (b) improve communication among professionals. Key words: assessment, language comprehension, language impairments, language tests, reading disorders, school-aged children and adolescents with language impairment, specific language impairment

HILDREN and adolescents with persis- as psychology, neuropsychology, pediatric Ctent oral and written language impair- psychiatry, developmental behavioral pedi- ments have often seen several different pro- atrics, special education/learning disabilities, fessionals for purposes of assessment and/or speech-language pathology (SLP), school psy- intervention. Because their basic language im- chology, and audiology. Their parents, with a pairments can impact such critical domains as sense of urgency, arrive at each new appoint- literacy, academic achievement, and social re- ment with an expanding portfolio of profes- lationships, it is not uncommon to find that sional reports and individualized educational these children have “made the rounds”among plans. Frequently, they are frustrated, having several professionals from disciplines such heard several labels for their child’s problem (e.g., auditory processing disorder, dyslexia, learning disability, language impairment) and several (sometimes conflicting) recommen- Author Afﬁliation: Department of Communication dations for treatment. Although the indi- Disorders and Sciences, Rush University Medical vidualized educational planning process re- Center, Chicago, Illinois. quires different professionals to endorse a Corresponding Author: Cheryl M. Scott, PhD, Depart- plan of treatment, evidence of true collab- ment of Communication Disorders and Sciences, Rush University Medical Center, 600 S. Paulina St, 1015 A, oration across disciplines is sometimes hard Chicago, IL 60612 (cheryl m [email protected]). to discern in the record. When consultation DOI: 10.1097/TLD.0b013e31820a100d outside the school setting is sought, parents 24

Assessment of Language and Literacy 25

are often on their own to integrate find- commonly specialists, including psychiatry ings from several professionals. Sometimes and developmental behavioral pediatrics. Al- the only constant in this journey is that though the children differ on many dimen- the academic gaps between their child and sions (e.g., presence/absence of comorbid the child’s peers continue unabated (Catts, conditions such as attention-deficit hyperac- Bridges, Little, & Tomblin, 2008). tivity disorder), the common denominator for Like the other authors in this issue of Top- those referred is academic struggle. Some ics in Language Disorders, I am concerned have been “in the system” since they qual- about this scenario. To use the allegory of ified for early intervention as a late talker, the blind man and the elephant, it seems that then continued into a preschool early child- different professionals sometimes “see” only hood education program, and eventually met their part of the elephant. Would it not be eligibility criteria for special education ser- better if there was consensus on (a) how the vices in school in one or more categories (usu- elephant works (see Foorman, Arndt, & Craw- ally speech language impairment and/or spe- ford, 2011), (b) what we call the parts of the cific learning disability). Others have a shorter elephant, and (c) how different professionals history in special education, and some have can make complementary contributions to a never received services even though it seems more successful elephant. In this article, I de- clear that they could have benefited. The scribe a clinical model for assessing language questions that these physicians raise when re- and literacy impairments in older children ferring to us are (a) whether and how a lan- and adolescents as one type of contribution guage disorder is impacting school achieve- that has the potential to improve communi- ment, and (b) if so, what can be done about cation about, and understanding of, language it. impairments across disciplines. The model ad- The assessment model, characterized as a dresses multiple levels, domains, and modal- hypothesis-testing process, is covered in three ities of language and, importantly, draws on sections. First is an overview of several types research that helps explain connections be- of oral and written language impairment vari- tween them. The goal is to uncover individual ations found in the school-aged population. profiles of language and literacy impairment Awareness of these variations informs the at a level of explicit detail and integration hypothesis-testing process. The second sec- that can support cross-disciplinary communi- tion codifies this variation as several clinical cation. The desired outcome for the child is a hypotheses and presents a plan for obtaining more targeted intervention that all profession- the information necessary to confirm (or not als can endorse and reinforce. confirm) each hypothesis. Finally, a case study The model that I will describe is not spe- is presented that illustrates the hypothesis- cific to one work setting, but it has grown testing nature of language assessment. out of my work as a SLP working with older children and adolescents∗ with language and VARIATION IN LANGUAGE IMPAIRMENT literacy impairments. Because I work in an urban academic medical center, many of the Although heterogeneity is often asserted children who I see are referred by pedia- in literature on language impairments, there tricians, some in general practice, but more are fewer discussions about how assessment can be structured to reveal individual differences. A careful assessment that captures this variation would seem to be a first step in ∗ Older children and adolescents between the ages of ap- bridging the gap (or chasm) that often exists proximately 9 and 14 years with oral and written language impairments constitute the group of interest in this arti- between (a) the actual language knowledge cle. Use of the term child or children assumes this age and skills required to function in a classroom range. and (b) common language assessment and

26 TOPICS IN LANGUAGE DISORDERS/JANUARY–MARCH 2011

intervention practices (as described recently lexical (vocabulary) skills are relatively higher by Wallach, 2010). The process of uncover- (Laws & Bishop, 2003)—a pattern that dif- ing a unique language profile is one of start- fers for children with Fragile X syndrome ing with one or more hypotheses about the (Abbeduto, Brady, & Kover, 2007). In stud- nature of oral and/or written language impair- ies involving children with autism who are ments and their variations and planning an as- verbal but language impaired, Kjelgaard and sessment protocol that allows for confirma- Tager-Flusberg (2001) report profiles resem- tion. An internal conversation about how a bling children with specific language impair- child’s oral and written language problems in- ment (SLI). Both groups are especially poor on teract might go something like the following: grammatical tasks and have difficulty repeat- ing long sentences and nonwords correctly. Everything I know about this child points to a spe- Grammatical impairment, in particular cific comprehension deficit that impacts both listening and reading. I don’t think the classic dyslexic tense marking, has long been considered as a profile of word reading inaccuracy explains her clinical marker of children with the primary problems. One way I can test this is to present com- condition known as specific language im- parable listening and reading comprehension tasks. pairment (Leonard, 1998; Rice, 2007)—one I wouldn’t expect listening performance to be bet- that “surfaces” again when these children are ter than reading. However, I anticipate that word asked to write, even though their speech reading accuracy will be within normal limits. is practically error free (Windsor, Scott, & Street, 2000). A large body of research on the Language impairment categories linguistic features of SLI and an emerging liter- and profiles ature on language characteristics of children Perhaps the most basic question about lan- with known syndromes provide a starting guage impairment is whether it stands alone point in the hypothesis-testing process by as a child’s only problem or, on the con- pointing to patterns of relative strength trary, there are additional problems—a dis- and weakness that could be encountered. tinction captured in the literature as pri- Refinements are expected as researchers mary or secondary (Law, Garret, & Nye, continue to make progress defining the 2004; Nelson, 2010). Children with primary phenotypic linguistic behaviors associated language impairments are otherwise typical with genetic findings in various groups (e.g., in cognitive, sensory, medical/neurological, Hayiou-Thomas, Harlaar, Dale, & Plomin, and social–emotional domains. Secondary lan- 2010; Rice, Smith, & Gay´an, 2009). guage impairments are those in which a lan- One of the most helpful tools in the process guage difficulty exists alongside problems in of testing hypotheses about the language pro- these other domains. Use of the term sec- file of any one child is to ask whether history ondary is unfortunate because it seems to im- and evidence point to a particular subtype of ply that the language impairment is of sec- language or reading impairment. The search ondary importance to the other condition, for for subtypes of impairments has a long his- example autism, when in fact the language im- tory (Aram & Nation, 1975; Conti-Ramsden, pairment is often one of the most problem- Crutchley, & Botting, 1997; van Daal, atic features of a child’s entire profile. Never- Verhoeven, & van Balkom, 2004). However, theless, the distinction is important because to date, the empirical evidence that would it channels our attention immediately to a validate one or another system as distinct body of research that addresses distinguish- subtypes with clinical utility is lacking and ing linguistic features of children in particu- use of the term subtype does not seem lar groups. For example, children with Down warranted. The only system to reach the syndrome are known to have syntactic diffi- level of official diagnostic coding thus far is culties that exceed predictions based on men- the distinction between expressive language tal age (Abbeduto & Chapman, 2005) whereas disorder and mixed receptive–expressive

Assessment of Language and Literacy 27

language disorder, as outlined in the current Another example of profiling is the well- Diagnostic and Statistical Manual of Mental known model of reading impairments based Disorders (fourth edition, text revision; on word recognition and linguistic compre- American Psychiatric Association, 2000). hension, the two components of the simple Indeed, this distinction is ingrained in the view of reading popularized by Gough and structure of most comprehensive language Tunmer (1986). In this model, how well tests that combine subtest scores into sepa- one reads (where reading is defined as com- rate expressive and receptive quotients. How- prehending what is read) depends on word ever, this system is currently under review in recognition accuracy and general linguistic the upcoming fifth edition. Further, on the ba- comprehension ability. The emphasis is on sis of empirical evidence (Tomblin & Zhang, general comprehension (listening compre- 2006) and critical analysis (Leonard, 2009), hension). A child with poor word recognition the validity of this distinction is questionable. and adequate general comprehension would Leonard (2009) argues that these categories meet the classic definition of dyslexia (Catts could be an artifact of the way language is & Kamhi, 2005). The opposite, a condition tested; certain structures that are easily tested where a child has poor general comprehen- in expressive tasks are less accessible in com- sion but adequate word-reading ability, has prehension tasks. He also reviews evidence been variously termed: specific comprehen- that language knowledge is “graded” rather sion deficit (Catts & Kamhi, 2005), poor than all or none and this graded level of knowl- comprehender (Cain & Oakhill, 2006), and edge may prove very difficult to measure hyperlexia (Aram, 1997). The child whose given current methods of testing in which reading (and writing) skills are limited by a response to any receptive test item is ei- poor general comprehension is an example of ther correct or incorrect (p. 117). Sometimes the importance of forming hypotheses about production tasks require greater knowledge how listening and reading comprehension of a particular structure than a comprehen- interact. A third possibility is the child with sion task where context and familiarity might both problems—poor word recognition and assist. poor general comprehension. This model For assessment planning, the obvious appli- of reading impairment has had a substantial cation is that comprehension and production impact on the way I conduct a language should be thoroughly tested in all children assessment and is one of the hypotheses that with confirmed or suspected language impair- I often explore. ments, and it should be tested beyond the One profile of SLI advanced over a num- level of single-word receptive vocabulary (i.e., ber of years is a form where the impair- in sentence- and text-level tasks). It is also im- ment is quite specific to select grammatical portant to carefully analyze both the content systems, namely, those that require complex and the task format of language tests/subtests. mapping and dependency relations. Two ma- A particular subtest of a comprehensive lan- jor syntactic groups have emerged. In one guage test may be relegated to the expressive account, the problem lies at the intersec- group mainly because the task format requires tion of morphology and syntax within the a spoken response when the task clearly re- verb system, blocking proper assignment of quires linguistic knowledge that is used in verb tense and agreement (Rice, 2002). Usu- both comprehending and producing language ally, the child omits the tense or agreement (sentence repetition tests are good examples). markers, resulting in sentences like “he bite The importance of carrying out careful con- him” (when telling a story about something tent and task analyses on norm-referenced happening in the past) or “he love his dog” tests and criterion-referenced tasks is a point (describing a general state of affairs). In an- explored in greater detail in the following other grammatical account of SLI, the focus section. is on structures described by long distance

28 TOPICS IN LANGUAGE DISORDERS/JANUARY–MARCH 2011

dependency relations found in structures use). Another system is to speak of the such as passive voice, relative clauses, and WH level of language—word, sentence, or dis- questions (van der Lely, 2005; van der Lely, course/text. Some overlap between compo- Rosen, & McClelland, 1998). To illustrate, in nents and levels is evident; for example, the following sentence there is a dependency syntax plays out at the sentence level. Seman- relationship between bill and the “trace” that tics, however, is seen at multiple levels includ- it leaves as a grammatical object in the embed- ing: word meaning (the lexicon), sentence ded relative clause (shown at the point of the meaning (propositions conveyed by a simple arrow) that must be represented before the sentences and meaning relationships between sentence as a whole is comprehensible: clauses in complex sentences), and text meaning (organization of a text, and overall gist The bill that the House passed ↑ recently is un- or core meaning, as conveyed by a succinct likely to make it through the Senate in the same form. summary). Children and adolescents with language impairments can show various patterns Both of these specific, grammatically based of strengths and weaknesses across these lev- varieties (morphosyntax and dependency re- els. A common area of difficulty is at the lations) are the phenotypic outcomes of sentence level—producing and understand- a domain-specific linguistic representational ing complex sentences that involve main and deficit. This deficit, in turn, is most likely ge- subordinate clauses characteristic of higher netically determined and resides in special- level language (Scott, 2009a; Silliman & Scott, ized cognitive processes that serve grammar 2009). (Silliman & Scott, 2006, p. 3). Even though Children with sentence-level problems the theoretical significance of these grammat- might post scores within normal limits on ical varieties of SLI is not universally accepted word-level (vocabulary) tests and could also (Bishop, Adams, & Rosen, 2006), no one show reasonable abilities in conveying an doubts that children with SLI find these spe- organized narrative. Sometimes the amount cific structures to be problematic. Whether of contextualization in a task impacts per- one adheres to a representational account or a formance at any one level. For example, a domain-general cognitive-processing account child could do poorly on a decontextualized (e.g., limited capacity processing) as the ex- sentence task but produce longer, complex planation of SLI, there would appear to be sentences when the task is under his/her clinical utility in assessing these structures in control, as when generating a story. Language comprehension and production tasks. tasks at the text level can either help or hurt language performance. Some children show Linguistic components, levels, improved performance on reading compre- and modalities hension tasks when the broader context of Another source of variation in oral and writ- the text or background knowledge boosts ten language impairments is the matter of comprehension, but falter when filling in a where, in the vast language domain, prob- missing word in a sentence or short text on lems are found. It is not very helpful to a different type of comprehension task. Con- say that a child has a language problem and versely, a child could perform within normal leave it at that. Because of the complexity limits on sentence-level tasks or tests, but find of language, SLPs are accustomed to break- the prospect of organizing a text, particularly ing the topic into more manageable sub- an informational text, totally beyond his or components. One system is to categorize her capability (Nippold & Scott, 2010). At language according a typology of linguistic any one level, modality (whether spoken or knowledge—phonology (sound knowledge), written language is required) can also impact semantics (word and word relations), syn- performance. In this scenario, the usual tax (grammar), and pragmatics (function and pattern would be that a child might perform

Assessment of Language and Literacy 29

better in oral tasks compared with written to explaining language profiles. I find it help- tasks, but again, the opposite can occur, as it ful to think about explanations of language im- did in the case discussed later in this article. pairments using an analogy of peeling away For this child, reading comprehension was the layers of an onion. At the outermost superior to listening comprehension; this level are the observable language behaviors child seemed to benefit from having a per- that we capture in our measures, for exam- manent visual representation of information ple, reduced utterance length, an absence of rather than the fleeting input of auditory-only higher level vocabulary and sentence struc- stimuli. ture, inability to tell a well-organized narrative, poor single-word receptive vocabu- Linguistic gatekeepers lary, and so forth. At the next layer into the Another perspective on language impair- onion are the cognitive processes that ac- ments that feeds directly to testing hypothe- count for such language behaviors. As noted ses is the matter of uncovering linguistic above, grammatical accounts of SLI posit “gatekeepers.” When one language domain that a domain-specific (but nevertheless brain- is severely impaired, it can serve as a gate- based) representational difference underlies keeper, limiting the development of another the language difficulties. For more domain- language skill. An obvious example is spelling. general accounts of the problem, the process- To illustrate, the developmental spelling level ing of auditory information is of particular of a 10-year-old child evaluated in our clinic interest, with recent focus on the speed of resembled that of a beginning speller. This processing (Leonard, et al., 2007), the pro- child had some knowledge of initial and fi- cessing of temporal and frequency informa- nal singleton consonants and short vowels in tion (Hill, Hogben, & Bishop, 2005), and the consonant-vowel-consonant words but very concurrent retention and processing of audi- little knowledge beyond that (e.g., long vow- tory information (auditory working memory) els, consonant blends and digraphs, and sylla- (for a recent review of working memory as- ble juncture patterns were misspelled). This sociations with SLI, see Montgomery, Magi- child could tell a story with basic text struc- maira, & Finney, 2010; for working memory ture and content, but spelling was so diffi- and reading/writing associations, see Swan- cult that his attempt to write the same ma- son & Berninger, 1996a, 1996b). The ties be- terial quickly deteriorated into a few simple tween language ability, working memory, and sentences and he was unwilling to persevere. academic achievement are of great interest Another example of gatekeeping is the in- because of the sentence- and text-processing teraction between lexical and sentence/text requirements encountered in academic texts. knowledge. The child whose lexical options Many children with language impairments for describing character emotions in a story perform poorly on language measures that tax are limited to happy, mad, and sad is likely working memory and on independent tests of to express these in simple sentences (e.g., working memory (Montgomery et al., 2010); he was mad vs. he was jealous because he however, there are exceptions (Archibald & wanted to be the only pet). Joanisse, 2009). As diagnosticians, we should attempt to answer the question of whether Explaining language impairments working memory constraints impact a child’s As is the case for children with broader ability to process language. developmental delays (C. Ochoa, personal At the next level of the onion, we en- communication, October 19, 2009), succinct counter neurobiological differences in the explanations (medical, cognitive, and/or ge- brain. These are not observable in our be- netic) for developmental language impair- havioral testing protocols, but an increas- ments are often elusive. Nevertheless, there ing number of neural imaging studies are are hypotheses to pursue that can contribute contributing information about brain-based

30 TOPICS IN LANGUAGE DISORDERS/JANUARY–MARCH 2011

differences in children with and without lan- in each hypothesis is Matthew. An important guage and reading disorders (Berninger, 2008; caveat to consider before starting this discus- Shaywitz, Gruen, & Shaywitz, 2008; Silliman sion is that norm-referenced tests seldom have & Berninger, 2011). And finally, at the cen- enough items of any one type to allow ex- ter of the onion are genetic and chromosomal aminers to draw absolute conclusions about differences. The chances that a child we are a linguistic trait. The tools are used to test testing has a sibling, parent, grandparent, our working hypotheses. The fact that clini- or extended family member with a devel- cal hypotheses (in many medical and psycho- opmental language/reading disorder is much logical practices) are working hypotheses that higher than in the general population (Hayiou- gather strength with time and accumulating Thomas, Harlaar, Dale, & Plomin, 2010; Rice, evidence is not new. In the numbered points Smith, & Gay´an, 2009). Confirming a family that follow, working hypotheses are listed, fol- history of language/reading impairment does lowed by related questions that will be used to not “condemn”a child to a poor outcome be- test the hypotheses. cause of the enormous complexity of genetic 1. Matthew has a primary language impair- and environmental contributions to language ment. learning, which differ across children. But, Is this a primary language impairment, isolated to as mentioned previously, future research does language, or are there additional cognitive, social– hold the promise that certain genetic histories emotional, medical/neurological factors that play a link to certain phenotypes of linguistic behav- significant role? ior, and this information should be very useful to us. This very basic question seems obvious but is nevertheless important to confirm (note TESTING HYPOTHESES ABOUT that the hypothesis could have been stated LANGUAGE DISORDERS in its dichotomous form—Matthew has a secondary language impairment). A careful de- The intent of this section is to opera- velopmental, medical, and educational his- tionalize variation in language impairments tory from the parent/caregiver usually pro- as testable hypotheses that shape the assess- vides the answer to this question. Obtaining ment process. The tests, tasks, and measures copies of all relevant documents (e.g., med- used by the diagnostician are chosen specif- ical records, psychoeducational evaluations, ically to allow a hypothesis to be confirmed individualized educational plan documents) is (or not). The process described is shaped by critical. It is particularly important to verify my own work setting—a hospital-based out- cognitive status. Although researchers study- patient clinic, where norm-referenced tests ing children with SLI typically invoke the cri- and criterion-referenced tasks are the most terion that participants score within 1 SD on commonly used tools, but there is no question a test of nonverbal intelligence, there is a lit- that the process would be enriched by direct erature on the longitudinal course of oral and observations of classroom performance and written primary language impairment that has other models, including curriculum-based as- used a criterion of ±2 SDs (e.g., Ellis-Weismer sessment and dynamic assessment. et al., 2000; Rice, Tomblin, Hoffman, Rich- Drawing from the previous discussion, I man, & Marquis, 2004). In these investiga- present three hypotheses as examples of the tions, participants with a cognitive standard types of hypotheses that could be formu- score of 85 or more are grouped separately lated. Each one is also restated as one or from those whose scores fall between 70 and more questions to clarify the intended mean- 85. Not uncommonly, children in the latter ing. This is followed by a discussion of ap- group (termed nonspecific language impair- plicable assessment tools and decision path- ment, or NLI) post lower scores on a variety ways. The hypothetical child (or adolescent) of oral and written language tests and tasks. If

Assessment of Language and Literacy 31

cognitive test scores are not available in acces- Examining oral comprehension at the sen- sible records or historical information, a non- tence level is critical. Academic writing, while verbal cognitive measure, such as the Test of undeniably a text, unfolds as individual sen- Nonverbal Intelligence, third edition (TONI-3) tences, coded by the use of capital letters (Brown, Sherbenau, & Johnsen, 1997), can be and periods or other end-sentence punctua- administered. Determination of primary ver- tion. To understand these texts, students must sus secondary status of a language impairment comprehend sentences. Table 1 provides ex- has implications not only for the linguistic amples of norm-referenced tests that can be phenotype (pattern of language strengths and used as measures of sentence comprehension weaknesses), as discussed previously, but also with children in the age range of interest here for prognosis because factors in addition to along with examples of items and a brief anal- language per se can affect a child’s ability to ysis of their content and the nature of the task learn. (what the child is asked to do). As discussed 2. Poor general language comprehension, in more detail later, content and task analy- not decoding (word recognition), appears to sis reveals that these sentence-level tools tap explain Matthew’s reading disorder. different aspects of sentence knowledge and are not necessarily interchangeable. For pur- Is this reading problem a case of dyslexia, specific poses of testing the present hypothesis, how- comprehension deficit, or is it mixed? ever, it is enough to establish that sentence- This hypothesis assumes that tests of read- level comprehension is compromised. Com- ing have determined that reading compre- pared with word- and sentence-level tests of hension is below age expectations. Now the listening comprehension, there are very few question becomes whether a profile of read- text-level tests of comprehension. An exam- ing disorder can be established. A child ple of one tool is the Listening to Paragraphs with a specific comprehension deficit should subtest of the Clinical Evaluation of Language do poorly on oral (listening) comprehension Fundamentals, fourth edition (CELF-4; Semel, tests (as well as reading comprehension tests) Wiig, & Secord, 2003), where children hear but perform within normal limits on tests of short texts (5–7 sentences) and then answer word recognition (accuracy). The outermost factual and inferential questions about them. level in the onion analogy discussed earlier, Several other tools that are not norm ref- the linguistic features that are observable, is erenced are also helpful in exploring listen- the focus of this hypothesis. For SLPs, this link ing comprehension. One of these is to ob- between general language comprehension serve a child retelling or generating stories. ability and reading is critical to explore; Because the format requires production of lan- the hypothesis, if confirmed, should have guage, narrative retells are sometimes over- an obvious impact on the direction of looked as sources of information about text intervention. comprehension. If a child is asked to retell a Careful testing of oral/listening comprehen- story read by the examiner, that child must sion is also required to confirm this hypoth- understand the story to retell it well. A poor esis. Word-level tests tap either vocabulary performance on this task could be attributed breadth or depth. An example of the for- more to a production constraint if follow-up mer is asking a child to point to the picture probe questions revealed good understanding named by an examiner (from a field of sev- of the story content. On the other hand, if eral pictures). Tests of vocabulary depth ex- both production and answers to probe ques- amine how well a word is known in relations were problematic, text comprehension tion to other words (e.g., picking two words difficulties are more likely. This distinction be- that are related from a field of four or testing tween production and comprehension of nar- knowledge of synonyms, antonyms, or super- ratives is built into the Test of Narrative Lan- ordinate/subordinate relationships). guage (TNL) (Gillam & Pearson, 2004).

32 TOPICS IN LANGUAGE DISORDERS/JANUARY–MARCH 2011

Table 1. A content/task analysis of four tests of language comprehension at the sentence level

Task: What the child Content: Linguistic Test does characteristics of input sentences Clinical Evaluation of Points to objects (e.g., Vocabulary is limited to simple object Language Fundamentals, house, shoe, fish, car) names and modifiers that signal Fourth Edition (Semel, in accordance with an sequence, order, spatial location, Wig, & Secord, 2003): instruction exclusion, and temporal order. Concepts and Following Grammatical requirements focus on Directions subtest adverbial subordinate clauses (e.g., Point to the second little house after you point to the last big apple and the last little apple), noun phrase postmodifying constructions (e.g., relative clauses such as, Point to the two cars that are to the right of a house, then point to the last house) and several other higher level forms. Some directions are 18–20 words long. Comprehensive Test of Hears a target sentence Requires recognition of syntactic Spoken Language and is asked whether a paraphrase, for example, the pretty cat (Carrow-Woolfolk, 1999): second sentence, then plays with the girl means the same Sentence Comprehension a third sentence “means thing as the cat that is playing with the Subtest the same thing” girl is pretty. The majority target (yes/no) and/or comparison sentences contain subordinate clauses including adverbial, relative, and object-complement clauses. Test for Reception of Points to the picture Twenty grammatical constructs are each Grammar-Version 2 (from a field of four) tested four times. Higher level language (Bishop, 2003) that illustrates the structures tested include: passive; sentence spoken by the center-embedded object relatives (e.g., examiner The sheep the girl looks at is running). Token Test for Touches or manipulates Parts 1, 2, and 3 require understanding of Children–Second Edition colored circles and color/size premodification of objects (McGhee, Ehrler, & squares in accordance (circle, square). The number of DiSimoni, 2007) with an instruction premodifiers and objects increase: Part 1: Touch the large red square. Part 2: Touch the white square and the red square. Part 3. Touch the small white square and the large red square. Part 4: Nine (of 16) items require comprehension of complex sentences that feature temporal or conditional adverbial clauses (e.g., While touching the red circle, touch the red square, but first touch the green circle).

Assessment of Language and Literacy 33

Another strategy for determining whether prehension across levels of language (this oral comprehension difficulties explain read- analysis is still situated at the outermost layer ing comprehension difficulties is to compare of the onion). The hypothesis gains strength if the two directly. Two passages (same level) the lowest scores posted are on sentence-level from an informal reading inventory can be tasks. As noted, a child’s factual and experi- used for this purpose; one is read to the child ential knowledge can boost comprehension and the other is read by the child, and compre- of text, particularly narrative text. Compre- hension questions are asked following both. hending decontextualized sentences of the When listening comprehension accounts for types shown in Table 1 removes that redun- reading comprehension difficulties, the ex- dancy and forces the child to rely more di- pectation is that performance levels would be rectly on grammatical knowledge. Second, a similar across modalities (difficulty in both). content/task analysis reveals that the various Carlisle and Rice (2002) caution that there sentence comprehension tests tap different is a developmental window, surrounding the types of knowledge and have different work- fifth grade, when this comparison has the ing memory requirements. Using the exam- best validity. In children at the beginning ples shown in Table 1, the two items for stages of reading, listening comprehension is Concepts and Following Directions require usually better than reading comprehension sophisticated grammatical analysis (adverbial (for equivalent texts) because of limited word subordination and relative clause embedding) recognition; conversely, in older students, but also feature specific semantic premodi- reading comprehension has the advantage fiers of order and direction (second, last, to because the reader has more control over the the right of). If a child missed these two items, pace of information. Another caveat is that it would not be clear whether it was the com- the texts that the child hears and reads should plex grammar or the concepts that caused the be comparable in terms of difficulty level and problem. But, if the child was correct on the genre. Hearing an expository passage but earlier and grammatically simpler items that reading a narrative passage could be a disad- test the same concepts (e.g., point to the sec- vantage for the listening comprehension per- ond house, point to the last ball in the row), formance because informational text is more we could more confidently attribute error re- challenging to many school-aged students, sponses on the longer items to a grammatical particularly those with language impairments and/or working memory constraint. (Ward-Lonergan, 2010). Ideally, clinicians To pursue and untangle the grammar and would use several inputs for both reading and working memory explanations, analysis of listening comprehension and establish that test results on the Test for Reception of there is a pattern whereby listening compre- Grammar–Version 2 (TROG-2; Bishop, 2003) hension is (or is not) the superior modality. and Token Test for Children–Second Edition 3. The core feature of Matthew’s language (TTFC-2; McGhee, Ehrler, & DiSimoni, 2007) disorder is difficulty comprehending and pro- would be helpful. Examination of the exam- ducing higher-level meanings/syntax found in ples for Parts 2 and 3 on the TTFC-2 shows complex sentences; working memory limita- that the grammar is simple and unchanged tions exacerbate his difficulties. going from Part 2 to Part 3. For Part 2, the child must hold the input direction in the au- Does Matthew produce higher-level sentence ditory “loop” long enough to discern the two structures at rates expected for his age? Does colors mentioned from five choices. Part 3 Matthew still omit obligatory tense and agreement requires the same calculation but adds a re- errors when speaking and/or writing? How is sentence comprehension affected by sentence com- quirement of size (large and small). In my ex- plexity and length? perience, some children do well on Part 2 but have a precipitous drop in correct responses Exploring this hypothesis involves several on Part 3. When this occurs, I interpret this analyses. First, I would need to look at com- as a sign that working memory may play a

34 TOPICS IN LANGUAGE DISORDERS/JANUARY–MARCH 2011

significant role in sentence comprehension phosyntactic (tense and agreement) knowl- difficulties. The TROG-2, on the other hand, edge. Because the amount and type of subor- is helpful in confirming grammatical con- dination that a child uses in a language sam- tributions to sentence comprehension prob- ple is optional (i.e., under the child’s con- lems because sentence length is shorter and trol), we do not have “exact” developmental specific grammatical structures are targeted norms, but it is possible to use the extant with four items for each structure. It is not literature and clinical experience to develop uncommon to find that a child has clear dif- reasonable expectations. With children aged ficulty with a small set of structures, usually 11 years and younger, I use the frog story the more syntactically complex items such as wordless picture book One Frog Too Many the center-embedded object-relative clause as (Mayer & Mayer, 1975) to elicit a spoken shown in the example in Table 1. Working narrative. After using the same book count- memory should play less of a role in perfor- less times and with the help of the System- mance on this test because sentences are gen- atic Analysis of Language Transcripts (SALT) erally shorter. The longest sentence on the reference database for narratives elicited in TROG-2 has 10 words and average sentence response to wordless picture books (Miller, length across all items is seven words. This 2009), I expect children with typical language compares with Concepts and Following Di- development to generate a story using at least rections (CELF-4) where the five longest sen- a few exemplars of a variety of subordinate tences have 19 or 20 words. Part 4 on the clauses in complex sentences (those with 2 TTFC-2 contains sentences of 15 words (3), or more clauses). Types I look for include ad- 18 words (1), 23 words (1), and 30 words verbial clauses (particularly those with later- (1). Of note, vocabulary is controlled on all developing conjunctions such as until, un- three sentence comprehension tests—simple less, while), relative clauses, and object com- objects (e.g., house, fish, shoe, pencil, cup), plements. In children with language impair- shapes (square, circle), people (e.g., girl, ments, it is not uncommon to find a restricted man, boy), and animals (sheep, elephant)— set of complex sentences, if any. When sub- and therefore should not affect comprehen- ordinate conjunctions are used, they tend to sion performance. Suggestions for dedicated be limited to early-developing forms such as measures of working memory can be found because, when, and so. Relative clauses are in Montgomery et al. (2010). rarely seen. Scott recently reviewed the types A third source of information about gram- of information about word, sentence, and text matical contributions to language impairment complexity available from the analysis of writ- is found in language tests that are typically ten language samples (Scott, 2009b) and ex- considered to be “expressive” tests (see the pository (informational) text (Scott, 2010). discussion given earlier). Responses on the To summarize, three hypotheses that can Recalling Sentences subtest of the CELF-4 can be used to structure an assessment provide a be analyzed to see whether content and/or pathway for uncovering possible profiles that grammar is preserved on items with two or have surfaced in research on the constructs of more clauses. A majority of items at higher lev- oral and written language impairments. Each els of the Formulated Sentences subtest of the hypothesis narrows the search. The first hy- CELF-4 require the child to generate a com- pothesis establishes whether the impairment plex sentence using adverbial conjunctions is restricted to language or not. The second es- (e.g., if, unless, until, as soon as) or conjuncts tablishes the manner in which oral and writ- (e.g., otherwise). ten components of the problem relate and Spontaneous (naturalistic) spoken and writ- the third pursues the precise nature of the ten language samples also provide a wealth of basic problem in general comprehension and information about a child’s ability to generate asks whether an underlying processing deficit sentence complexity and the status of mor- (verbal working memory) plays a role. In the

Assessment of Language and Literacy 35

last section, a case is presented that illustrates SLI (Kjelgaard & Tager-Flusberg, 2001). With the process of hypothesis testing. The case is this in mind, it was particularly important to actually a child with a language impairment explore grammatical knowledge in both sen- accompanied by broader behavioral difficul- tence comprehension tasks and production ties. The process of hypothesis testing would tasks. Literacy was also a concern. Jonathan not differ in a case of secondary language was struggling with reading comprehension impairment. More specifically, the three hy- according to school reports and his parents potheses explored in this section (or their were concerned about writing skills. It will be counterhypotheses) can be discerned. They important to see whether problems in word include (a) predictions that follow from the recognition and general linguistic comprehen- secondary nature of the language impairment, sion contribute to read difficulties or, alter- (b) the reading profile and the relative contri- natively, whether a specific comprehension bution of word recognition and general com- deficit is the main contributor to poor reading prehension, and (c) the specific nature of sen- comprehension. The latter possibility would tence comprehension difficulties and their re- be the favored hypothesis for a child with lation to working memory. Jonathan’s history (Nation, Clarke, Wright, & Williams, 2006). Furthermore, are there any gatekeepers in Jonathan’s linguistic profile? A CASE STUDY What factors account for his struggles with writing composition? We might hypothesize Jonathan∗ (age 10 years 7 months) was re- that higher level semantic/syntactic difficul- ferred to our outpatient clinic by a pediatri- ties at both sentence and text levels account cian who suspected that a language disorder for writing problems and one approach would was interfering with academic achievement be to compare spoken and written output. and literacy. This physician was interested Spelling as a gatekeeper to writing is less likely in an in-depth analysis of language strengths if decoding/spelling are relative strengths. and weaknesses that could inform interven- With these questions in mind, we con- tion. Jonathan’s history included both med- firmed that Jonathan has significant problems ical and educational information that would with oral language, scoring more than 2 SD place him in the group of secondary language below the mean on the CELF-4 core language disorder (seizure disorder as a younger child quotient. Looking more closely at his perfor- and special education eligibility in the cate- mance on the subtests that comprise this quo- gory of autism). His social–emotional status tient, there was a substantial discrepancy be- was not a primary focus of the assessment; tween sentence- and word-level tasks (word- he could easily participate in a casual con- level performance was better), pointing to versation, and he tolerated both informal and the possibility that sentences are difficult for formal testing well. Rather, we were inter- Jonathan to comprehend and produce. To ested in describing spoken and written lan- observe sentence comprehension in more guage status. The fact that Jonathan is a verbal detail, variation in performance across three child and is being educated in a general ed- different tests was examined. Jonathan’s per- ucation classroom with various support ser- formance on Concepts and Following Direc- vices pointed to the hypothesis that “clini- tions (CELF4) was very poor (3 SD below the cal marker”features of a language impairment mean). He reached a ceiling quickly before might resemble those seen in a child with progressing to items that required comprehension of complex sentences. On the TROG-2, with shorter sentences and grammar-specific items, Jonathan’s score ∗A pseudonym; several details have been altered to pro- was 1 SD below the mean. He had difficulty tect identity. with center-embedded object-relative clauses

36 TOPICS IN LANGUAGE DISORDERS/JANUARY–MARCH 2011

and neither/nor constructions. His perfor- A comparison of Jonathan’s oral and written mance on the TTFC-2, greater than 1 SD be- narratives is revealing. Although he produced low the mean, was revealing in terms of work- short and simple sentences when telling a ing memory. When an item required just one story, he was relatively fluent (i.e., he pro- modifier (see Part 2 of the example in Table duced a monologue of 48 utterances). He 1) he was accurate, but performance deterio- could not even start a written story, produc- rated significantly with two modifiers (Part 3 ing instead the list described earlier. Although of the example in Table 1). He also had great making up a story about one picture (a scene) difficulty with Part 4 sentences, missing most is more difficult than telling a story to accom- items; this portion of the test requires com- pany a wordless picture book, it is doubtful prehension of temporal and conditional sub- that the task difference could explain the en- ordinate clauses and many of these items are tire lack of written composition ability. Exam- long sentences. ination of his spelling skills showed accuracy Taken together, these results on sen- on singleton consonants and short vowels in tence comprehension tests demonstrate that consonant-vowel-consonant words but errors Jonathan will have a difficult time understand- on all higher level spelling patterns including ing complex sentences (e.g., those that con- long vowels, syllable juncture patterns, and tain relative and adverbial clauses) and that derived words. It is difficult not to conclude his performance can be expected to decline that Jonathan’s poor spelling is a gatekeeper as sentences become longer and tax working for written composition—the gap is simply memory. In addition, he does not understand too wide between his transcription skills (try- the meaning of particular relational modifiers ing to spell words) and content needs (what like those tested on Concepts and Following he is trying to communicate in writing). He Directions. His poor understanding of higher is reduced to filling a page with simple “label- level language encoded in complex sentences ing”sentences that repeat. Although persever- presented difficulties on production tasks as ation is a common characteristic of children well. He could not make up sentences using with autism, we should note that repetition words like unless and until (these would have was not seen anywhere else in his language required a complex sentence with a subordi- profile. nate adverbial clause), and he spontaneously Finally, examining Jonathan’s reading produced two-clause sentences in only 2 of scores, the hypothesis of a specific compre- the 48 utterances when he generated a story. hension deficit as the sole explanation of Additionally, he wrote only simple, list-like, Jonathan’s reading comprehension difficul- object-naming simple sentences when asked ties is not confirmed. His general linguistic to write a story in response to a picture comprehension difficulties certainly con- scene. Considering the types of sentences that tribute to reading problems, but Jonathan’s Jonathan will be asked to comprehend when reading also suffers from poor word recog- he enters a fourth-grade general education nition, decoding, and fluency. Nation and classroom in the coming year, either when colleagues (2006), while confirming poor listening to his teacher or when reading text- comprehension as the major contributor book assignments, he will be at a consider- to poor reading in a group design with 41 able disadvantage. Likewise, whether he is children with autism, reported that variability responding orally to questions about what he among children was high and some children has learned, or when asked to summarize ma- were poor at word and nonword reading. terial in writing, Jonathan will struggle. The Jonathan’s relatively better comprehension hypothesis of a significant grammatical com- scores (within 1 SD on two different reading ponent in his language profile is strengthened, comprehension tests) are, in fact, unexpected and working memory appears to impact per- given his difficulty with oral comprehension. formance. There are several possible explanations.

Assessment of Language and Literacy 37

Unlike the oral comprehension tests, which tic strengths and weaknesses. Unfortunately, are highly decontextualized and almost however, academic success requires “all lan- every word contributes critical meaning, guage hands on deck” and suffers from any the reading tests allow for some benefit of combination of language deficiencies. As chil- general knowledge and textual knowledge— dren reach the upper elementary and mid- it is not always necessary to understand dle school years with a history of successive every word in the sentence to answer cor- years of academic struggle, the core linguis- rectly. Also, answers to many of the reading tic culprits are far from obvious, and there comprehension items are encoded in simpler are many candidates. A process of careful lan- sentences than those found on oral sentence guage assessment in such cases is described. comprehension tests. It is also possible that The process can be conceptualized as one he benefits considerably from being able to of drawing on the research literature in oral see a permanent visual record of the material and written language impairments to assist that he is expected to comprehend. in forming possible hypotheses that can be Jonathan will need an intervention program reasonably tested in the clinic for individual that addresses his lack of knowledge of higher cases. Careful content and task analysis of level sentence complexity and the meanings what we ask children to do during an assess- and uses of such sentences within wider texts ment (both norm-referenced and criterion- (Balthazar & Scott, 2007; Scott & Balthazar, referenced tasks) is also required. In the case 2010). He also needs guided word study that of Jonathan, the process was helpful in arriv- is individualized to his specific developmental ing at core features of a language profile that level of decoding and encoding orthographic can be used to inform intervention. The de- patterns, with immediate applications to gree of detail and integration across levels of develop fluency in text-level reading and language and attempts to explain modality in- writing. teractions and processing components should To conclude, the premise of this article stimulate and improve cross-disciplinary com- is that older children and adolescents with munication among those who share responsi- language impairments are a heterogeneous bilities for improving the academic prospects group in terms of specific patterns of linguis- for these children.

REFERENCES

Abbeduto, L., Brady, N., & Kover, S. T. (2007). Lan- Archibald, L. M. D., & Joanisse, M. F. (2009). On the sen- guage development and Fragile X syndrome: Pro- sitivity and specificity of nonword repetition and sen- files, syndrome-specificity, and within-syndrome dif- tence recall to language and memory impairments in ferences. Mental Retardation and Developmental children. Journal of Speech, Language, and Hearing Disabilities Research Reviews, 13, 36–46. Research, 52, 899–914. Abbeduto, L., & Chapman, R. S. (2005). Language devel- Balthazar, C., & Scott, C. (2007). Intervention for syntax opment in Down syndrome and fragile X syndrome: and morphology. In A. G. Kamhi, J. J. Masterson, & K. Current research and implications for theory and prac- Apel (Eds.), Clinical decision making in developmen- tice. In P. Fletcher & J. F. Miller (Eds.), Developmental tal language disorders (pp. 143–165). Baltimore: Paul theory and language impairments (pp. 53–72). Am- H. Brookes. sterdam: John Benjamins. Berninger, V. W. (2008). Defining and differentiating dys- American Psychiatric Association. (2000). Diagnostic graphia, dyslexia, and language learning disability and statistical manual of mental disorders (4th ed., within a working memory model (pp. 103–134). In text revision). Washington, DC: Author. M. Mody & E. R. Silliman, E.R. (Eds.). Brain, behav- Aram, D. (1997). Reading without meaning in young chil- ior and learning in language and reading disorders. dren. Topics in Language Disorders, 17(3), 1–13. New York: Guilford. Aram, D., & Nation, J. (1975). Patterns of language behav- Bishop, D. V.M. (2003). Test for reception of grammar— ior in children with developmental language impair- Version 2. San Antonio, TX: Pearson. ments. Journal of Speech and Hearing Research, 18, Bishop, D. V. M. (2006). Developmental cognitive genet- 229–241. ics: How psychology can inform genetics and vice

38 TOPICS IN LANGUAGE DISORDERS/JANUARY–MARCH 2011

versa. Quarterly Journal of Experimental Psychol- for genetic subgroups. Language and Cognitive Pro- ogy, 59, 1153–1168. cesses, 16, 287–308. Bishop, D. V.M. (2009). Specific language impairment as a Law, J., Garrett. Z., & Nye, C. (2004). The efficacy of treat- language learning disability [Commentary]. Child Lan- ment for children with developmental speech and guage Teaching and Therapy, 25, 163–165. language delay/disorder: A meta-analysis. Journal of Bishop, D. V. M., Adams, C. V., & Rosen, S. (2006). Speech, Language, and Hearing Research, 47, 9242– Resistance of grammatical impairment to computer- 943. ized comprehension training in children with spe- Laws, G. L., & Bishop, D. V. M. (2003). A comparison of cific and non-specific language impairments. Interna- language abilities in adolescents with Down syndrome tional Journal of Language and Communication and children with specific language impairment. Jour- Disorders, 41, 19–40. nal of Speech, Language and Hearing Research, 46, Brown, L., Sherbenou, R. J., & Johnsen, S. K. (1997). 1324–1339. Test of nonverbal intelligence (3rd ed.). Austin, TX: Leonard, L. (1998). Children with specific language im- Pro-Ed. pairment. Cambridge, MA: The MIT Press. Cain, K., & Oakhill, J. (2006). Profiles of children with spe- Leonard, L. (2009). Is expressive language disorder and cific reading comprehension difficulties. British Jour- accurate diagnostic category? American Journal of nal of Educational Psychology, 76, 683–696. Speech-Language Pathology, 18, 115–123. Catts, H. W., Bridges, M. S., Little, T. D., & Tomblin, J. B. Leonard, L., Ellis-Weismer, S., Miller, C., Francis, D., (2008). Reading achievement growth in children with Tomblin, J. B., & Kail, R. (2007). Speed of processing, language impairments. Journal of Speech, Language, working memory, and language impairment in chil- & Hearing Research, 51, 1569–1579. dren. Journal of Speech, Language, and Hearing Re- Carlisle, J. F., & Rice, M. S. (2002). Improving reading search, 50, 404–428. comprehension: Research-based principles and prac- Mayer, M., & Mayer, M. (1975). One frog too many.New tices. Baltimore: York Press. York: Dial Books for Young Readers. Carrow-Woolfolk, E. (1999). Comprehensive assessment McGhee, R. L., Ehrler, D., & DiSimoni, F. (2007). Token of spoken language. Circle Pines, MN: AGS Publish- test for children (2nd ed). Austin, TX: Pro-Ed. ing. Miller, J.F. (2009). SALT: Systematic Analysis of Lan- Catts, H. W., & Kamhi, A. G. (2005). Language and read- guage Transcripts [Computer software]. Madison, ing disabilities (2nd ed.). Boston: Pearson. WI: University of Wisconsin-Madison, Waisman Conti-Ramsden, G., Crutchley, A., & Botting, N. (1997). Center, Language Analysis Laboratory. http://www. The extent to which psychometric tests differentiates languageanalysislab.com/ subgroups of children with SLI. Journal of Speech, Montgomery, J. W., Magimaira, B. M., & Finney, M. C. Language, and Hearing Research, 40, 765–777. (2010). Working memory and specific language im- Ellis-Weismer, S., Tomblin, J. B., Zhang, X., Buckwalter, pairment: An update on the relation and perspectives P., Chynoweth, J. G., & Jones, M. (2000). Nonword on assessment and treatment. American Journal of repetition performance in school-age children with Speech-Language Pathology, 19, 78–94. and without language impairment. Journal of Speech, Nation, K., Clarke, P., Wright, B., & Williams, C. (2006). Language, and Hearing Research, 43, 865–878. Patterns of reading ability in children with autism Foorman, B. R., Arndt, E. J., & Crawford, E. C. (2011). spectrum disorder. Journal of Autism and Develop- Important constructs in literacy learning across disci- mental Disorders, 36(7), 911–919. plines. Topics in Language Disorders, 31(1), 73–83. Nelson, N. W. (2010). Language and literacy disorders: Gillam, R., & Pearson, N. (2004). Test of narrative ability. Infancy through adolescence. Boston: Allyn & Bacon. Austin, TX: Pro-Ed. Nippold, M., & Scott, C. M. (2010). Expository discourse Gough, P., & Tunmer, W. (1986). Decoding, reading, and in children, adolescents, and adults: Development reading disability. Remedial and Special Education, and Disorders. New York: Psychology Press. 7, 6–10. Rice, M. L. (2002). Grammatical symptoms of specific lan- Hayiou-Thomas, M. E., Harlaar, N., Dale, P. S., & Plomin, guage impairment. In D.V. M. Bishop & L. B. Leonard R. (2010). Preschool speech, language skills, and read- (Eds.), Speech and language impairments in chil- ing at 7, 9, and 10 years: Etiology of the relationship. dren: Causes, characteristics, intervention and out- Journal of Speech, Language, and Hearing Research, come (pp. 17–33). Philadelphia: Psychology Press. 53, 311–332. Rice, M. L. (2007). Children with specific language im- Hill, P. R., Hogben, J. H., & Bishop, D. M. V. (2005). Audi- pairment: Bridging the genetic and developmental tory frequency information in children with specific perspectives. In E. Hoff & M. Shatz (Eds.), Blackwell language impairment: A longitudinal study. Journal of handbook of language development (pp. 411–431). Speech, Language, and Hearing Research, 48, 1136– Malden, MA: Blackwell. 1146. Rice, M. L., Smith, S. D., & Gay´an, J. (2009). Convergent Kjelgaard, M., & Tager-Flusberg, H. (2001). An investiga- genetic linkage and associations to language, speech tion of language impairment in autism: Implications and reading measures in families of probands with

Assessment of Language and Literacy 39

specific language impairment. Journal of Neu- Silliman, E. R. & Scott, C. M. (2009). Research-based rodevelopmental Disorders, 1 (4), 264–282. doi: oral language intervention routes to the academic lan- 10.1007/s11689-009-9031-x. Retrieved December 29, guage of literacy: Finding the right road. In S. Rosen- 2010, from http//:www.springerlink.com/content/ field & V. Berninger (Eds.), Implementing evidence- 6808305i34217466/ based academic interventions in school settings (pp. Rice, M. L., Tomblin, J. B., Hoffman, L., Richman, W. A., & 107–146). New York: Oxford University Press. Marquis, J. (2004). Grammatical tense deficits in chil- Swanson, H. L., & Berninger, V. (1996a). Individual dif- dren with SLI and nonspecific language impairment. ferences in children’s working memory and writing Journal of Speech, Language, & Hearing Research, skills. Journal of Experimental Child Psychology, 63, 47, 816–834. 358–385. Scott, C. M. (2009a). A case for the sentence in reading Swanson, H. L., & Berninger, V. (1996b). Individual dif- comprehension. Language, Speech, and Hearing Ser- ferences in children’s writing: A function of work- vices in Schools, 40, 184–191. ing memory or reading or both processes? Read- Scott, C. (2009b). Language-based assessment of written ing and Writing: An Interdisciplinary Journal, 8, expression. In G. Troia (Ed.), Writing instruction and 357–383. assessment for struggling writers: From theory to Tomblin, J. B., & Zhang, X. (2006). The dimensionality evidenced-based principles (pp. 358–385). New York: of language ability in school-age children. Journal of Guilford. Speech, Language, and Hearing Research, 49, 1193– Scott, C. (2010). Assessing expository texts produced by 1208. children and adolescents. In M. Nippold & C. Scott van Daal, J., Verhoeven, L., & van Balkom, H. (2004). Sub- (Eds.), Expository discourse in children, adolescents, types of severe speech and language impairments: Psy- and adults: Development and disorders (pp.195– chometric evidence from 4-year-old children in the 217). Psychology Press/Taylor & Francis. Netherlands. Journal of Speech, Language, and Hear- Scott, C., & Balthazar, C. (2010). The grammar of infor- ing Research, 47, 1411–1423. mation: Challenges for older students with language van der Lely, H. K. J. (2005). Domain-specific cognitive impairments. Topics in Language Disorders, 30(4), systems: Insight from grammatical SLI. Trends in Cog- 288–307. nitive Science, 9, 53–59. Semel, E., Wiig, E., & Secord, W. (2003). Clinical evalu- van der Lely, H. K. J., Rosen, S., & McClelland, A. (1998). ation of language fundamentals—4 (pp. 209–239). Evidence for a grammar-specific deficit in children. San Antonio, TX: Harcourt Assessment. Current Biology, 8, 1253–1258. Shaywitz, S. E., Gruen, J. R., & Shaywitz, B. A. (2008). Wallach, G. P.(2010). It was a dark and stormy night. Top- Dyslexia: A new look at neural substrates. In Mody, ics in Language Disorders, 30(1), 6–14. M. & Silliman, E.R. (Eds.), Brain, behavior and learn- Ward-Lonergan, J. M. (2010). Expository discourse in ing in language and reading disorders. New York: school-age children and adolescents with language Guilford. impairments: Nature of the problem. In Nippold, Silliman, E. R., & Berninger, V. (2011). Cross-disciplinary M.A. & Scott C. M. (Eds.), Expository discourse dialogue about the nature of oral and written language in children, adolescents, and adults: Development problems in the context of developmental, academic, and disorders (pp. 155–190). New York: Psychology and phenotypic profiles. Topics in Language Disor- Press. ders, 31(1), 6–23. Windsor, J., Scott, C., & Street, C., (2000). Verb and noun Silliman, E. R., & Scott, C. M. (2006). Language impair- morphology in the spoken and written language of ment and reading disability: Connections and com- children with language learning disabilities. Journal plexities. Learning Disabilities Research & Practice, of Speech, Language, and Hearing Research, 43, 21(1), 1–7. 1322–1336.