Rollins Corpus

Pamela Rosenthal Rollins
Callier Center for Communication Disorders
University of Texas at Dallas


Participants: 5
Type of Study: clinical
Location: USA
Media type: waiting for answer from Rollins
DOI: doi:10.21415/T5KK53

Browsable transcripts

Download transcripts

Citation information

Publications using these data should cite:
Rollins P. R. (1999). Pragmatic accomplishments and vocabulary development in pre-school children with autism. American Journal of Speech-Language Pathology: A Journal of Clinical Practice, 8, 85–94.

Additional references are:

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

This corpus consists of transcripts of video recordings of 5 boys with autism who at-tended a preschool program for children on the autistic spectrum at the University of Texas at Dallas. To be included in this corpus, a child had to meet the following criteria: (a) have an initial diagnosis of autism by a psychologist or a neurologist; (b) have been preverbal at the time of intake; (c) have attended the preschool program for at least 1 year; and (d) have some conventional expressive vocabulary skills upon completion of the program. The preschool program routinely videotapes each participating child for the entire morning session several times during the school year. For each child, four videotapes were selected for later transcription and analysis (the first, last, and two intermediate tapes). The transcripts for each child are numbered 1 to 4 or 5 corresponding to the tape number. The header file indicates the date of the video recording as well as the child’s age.

To capture each child's optimal level of on-task communicative functioning, only intervals where the child was interacting one-on-one with his clinician were transcribed and coded for analyses. Therefore, activities such as small group, music, and snack time were, by definition, excluded from the analyses. This criterion was used because the language skills of children on the autistic spectrum are influenced by both the setting and the participants. Furthermore, efforts to capture the child's optimal level of on-task communicative behaviors were made by excluding from the total number of usable minutes the following intervals: (a) when the clinician or child was out of the room, (b) when another child or teacher talked with the target child and clinician, (c) when the clinician attempted to engage the target child in an activity but where the target child refused to cooperate for longer than 30 seconds, (d) when the target child actively avoided an activity or interactions with the clinician for longer than 30 seconds, and (e) when the clinician and target child negotiated the next activity for longer than 60 seconds. This substantially reduced the total number of usable minutes available for transcription. Videotapes were viewed and cataloged. The catalog included a time record for each activity so that the total number of usable minutes for coding could be calculated. Activity header lines were used to mark each new activity on the transcript. Twenty minutes was the maximum number of usable minutes that was available for all children in the study at each time point. To ensure that the sample of 20 minutes was representative for each child the videotaped interactions were reviewed by persons fa-miliar with each child.

Because the corpus was originally collected to describe pragmatic skills in children with autism from the prelinguistic to early one-word stage, a good deal of nonverbal information is transcribed. The transcripts include %spa codes using the Ninio, Snow, Pan, & Rollins (1994) INCA system described in the CHAT manual. In order to be coded as communicative, each communicative act had to supported by behavioral evidence that the child had a plan/intention to achieve a goal with awareness that another person can be a means to that end. This behavioral evidence has been outlined by Prizant and Wetherby (1988) and includes the following: (a) alternating eye gaze between a goal and the listener, (b) persistent signaling until the goal has been met, (c) changing the quality of the signal until the goal has been met, (d) ritualizing or conventionalizing the form of signal within specific communicative contexts, (e) awaiting a response from the listener, (f) terminating the signal when the goal is met, (g) displaying satisfaction when the goal is attained or dissatisfaction when it is not. Communicative means is indicated on the third level of the speech act tier.

The gesture codes are adapted from McLean, McLean, Brady, and Etter(1991). The full set of codes is as follows:

Word-like vocalizations were transcribed on the main line and appended with @ap when there was sufficient contextual information to identify the target lexical item (e.g., saying "ba" when holding a ball was transcribed as ball@ap). Individual participant char-acteristics are presented in the table that follows. All of the children were relatively young at the start of the study (mean age of 2;7) and were severely delayed in language as measured by the SICD (mean receptive language age 1;2, and mean expressive language age 0;10). The delays in language skills were corroborated by the Vineland Adaptive Behavior Scales (Sparrow, Balla, & Cicchetti, 1984), as was the social impairment, as indicated in the following table.
The mean of each Vineland subtest in this table is 100 and the standard deviation is 15.

This research was supported by an American Speech and Hearing Foundation, First Investigators award.