Conti-Ramsden 4 Corpus

Gina Conti-Ramsden
Human Communication and Deafness
The University of Manchester

Nicola Botting
Language and Communication Science
City University of London


Participants: 19 SLI + 99 control
Type of Study: clinical
Location: UK
Media type: no longer available
DOI: doi:10.21415/T56019

Browsable transcripts

Download transcripts

Citation information

In accordance with TalkBank rules, any use of data from this corpus must be accompanied by at least one of the above references.

Project Description

Narrative study of British SLI young adolescents and a large comparison group of typically developing peers. 19 adolescents with specific language impairment (SLI) and 99 typically developing peers (TD) were given two different types of narrative task: a story telling condition and a conversational condition.

Typically developing adolescents
A large group of 99 typically developing adolescents (61 = female, 38 = male) were recruited through two secondary schools in central England. The adolescents had no history of speech and language therapy or special needs educational support as reported by school or parents. English was their first and only language. The young people were recruited from 3 age groups: 13 yr olds, 14 yr olds and 15 yr olds. Table 1 shows the mean age and gender profiles. A large group of comparison children was felt necessary to provide age appropriate normative information on the tasks.

Adolescents with Specific Language Impairment
The group consisted of 19 adolescents recruited from a wider study (Conti-Ramsden et al., 1997, Conti-Ramsden and Botting, 1999, Conti-Ramsden et al., 2001). All adolescents had a classic SLI profile at least at one time point in the study (i.e. Each child had a nonverbal IQ of ≥ 80 and scores of at least one standard deviation below the normative mean on one or more standard language assessment tests at either 7, 8 or 11-years-old). However at the point of testing, 4 children had a non-verbal IQ below this threshold. The mean age and gender profile of the group of adolescents with SLI can also be found in table 1.

The current language profiles of the group were mixed, but the majority still scored below 1SD on at least one part (i.e. expressive or receptive) of the Clinical Evaluation of Language Functions (CELF 3; Semel, Wiig, Secord, 1995). The main shift in profile since recruitment was towards lower performance IQ scores. Eight of the 19 participants with SLI had a lower than normal performance IQ (< 85) when assessed for the present study at 14 years of age. Of the 19 adolescents recruited with a history of SLI, 9 adolescents still fitted the SLI profile. Information regarding educational placement was unavailable for 2 adolescents with a history of SLI however the remaining 17 all attended mainstream schools at the time of the current study. Of the 17 adolescents, 10 adolescents (58.8%) had some educational support within the school environment (varying in degree from 1 hour a week to every lesson). The group profiles can be examined in table 2.

There were two semi-structured naturalistic oral narrative tasks: a story telling task and a spontaneous personal narrative. These tasks were chosen because they were thought to well better represent the everyday language skills of older children than some standardized language measures. Personal narratives form a large proportion of all language (Beals and Snow, 2002) Whilst story-telling, may not be an everyday activity, it is certainly evident in educational curriculums and social communication. Moreover, the ability to construct a cohesive structured ‘story’ relates directly to the essay skills required of adolescents in the UK school examination system. Formal standardized tests are known to be less specific as children get older and often have inadequate normative data for adolescent populations. Both tasks have previously been used with this age of participant and have yielded interesting results (see below for details). Furthermore, they represent complementary paradigms in a number of ways: One has picture prompts, while the other does not; one is based on a fictional scenario, while the other is a real-life description; one encourages past tense use, while the other is more likely to elicit present tense structures. In this way, the story task is likely to draw more heavily on working memory resources as the participant mentally holds the story elements together at the same time as constructing linguistically accurate sentences. The conversational task on the other hand, has less externally guided structure, which may mean that individuals choose to use less complex language or need more guidance from the researcher. The two tasks are described in detail below.

Story telling narrative task – Frog, where are you? (Mayer, 1969)

'Frog, where are you?' by Mercer Mayer is a wordless 24-picture storybook telling the adventures of a boy and his dog who are in search of their frog that has escaped from a jar in the boy's bedroom. It provides an excellent prompt for looking at structural language ability. The ‘Frog Story’ as it has become known, has been widely used as a narrative tool in the literature as it provides a series of events that give structure to the narratives and can be used to provide a semantic score as a measure of information included. Data have been collected using this task from typically developing children and adults (Wigglesworth, 1997, Berman and Slobin, 1994, Reilly, 1992, Bamberg and Damrad-Frye, 1991) and special developmental populations including children with autism (Tager-Flusberg, 1995) and children with focal brain injury (Reilly et al., 1998) as well as those with language impairment (Norbury and Bishop, 2003, Reilly et al., 2003, Botting, 2001, Van der Lely, 1997). In this study the protocol used by Van der Lely (1997) was followed to collect the narratives (see figure 1a). Adolescents are asked to choose a story from an envelope so that the researcher apparently doesn’t know which story has been picked. This encourages thorough and complete narratives with no ‘assumption of knowledge’. Note that while in this procedure adolescents were told that there were 4 different but similar stories, in fact each envelope contained the same ‘Frog Story’ as described above. This procedure was used because following van der Lely (1997), it was thought to be an appropriate protocol to encourage complete narratives from participants in the adolescent age range. It also meant that data collected would be comparable in kind to previously reported studies. The present study also encouraged the adolescents to tell the narratives in the past tense. This was done in order to give the adolescents the opportunity to use past tense forms, which have been reported in the literature to cause specific difficulties for children with SLI (see Leonard, 1998 for a review). However, the Frog Story itself also lends itself to past tense narration, since the participants are asked to ‘tell the story’ rather than describe what is happening. Adolescents in this study had no difficulty responding to the prompt to complete the task in past tense.

Story telling narrative protocol:

The participant was encouraged to tell the story in the past-tense thus if the participant started in the present tense, a prompt like “what happened then?” was used. If the participant continued in the present after two prompts, no further prompts were made.

Spontaneous narrative task, (Ingham, personal communication).

This task was used to elicit conversational style spontaneous narratives and encourage the adolescents to use verbal 3rd person singular –s. Agreement errors have also been widely reported for children with SLI (for example see Clahsen et al., 1997, Clahsen, 1999, Schütze and Wexler, 1996). Figure 1b details the protocol for this narrative. This narrative does not constitute a story in order to provide contrast with the story-telling task above. However it does ask the participant to recount ‘actions’ which are sometimes felt to be a key definition of narrative. Furthermore, although this task was about ‘annoying features’ it did not result in list of attributes but rather personal narratives about another person. This type of verbal account is also felt to be more characteristic of everyday language for teenagers than story telling.

Spontaneous narrative protocol:

Age and gender profiles for the two groups
GroupsMean age (s.d.)Minimum ageMaximum ageN and % male
Typically developing
14.5 (.84)13.115.938 (38.4%)
14.3 (.64)13.315.314 (73.7%)

Language profile for adolescents with a history of SLI
SLI (n=19)CELF expressive
language score
CELF receptive
language score
CELF total
language score
Perform. IQ
Verbal IQ
Full IQ
Mean (s.d.)70.05 (10.07)84.79 (17.26)75.84 (13.23)88.95 (13.14)82.53 (13.98)84.63 (12.25)

Part of the data collection was used in Danielle Wetherell's Ph.D. dissertation.