William O. Haynes

Differentiating Dialect from Disorder: A Comparison of Two Processing Tasks and a Standardized Language Test.

UNLABELLED Previous research has indicated that norm-referenced language assessment protocols are often biased against dialectal speakers. Recently, the use of processing tasks has emerged as one possible means of reducing this bias in language testing. Processing tasks measure a childs ability to process and manipulate language rather than tap previous linguistic knowledge. The present study utilized 40 subjects between the ages of 7;0 and 7;3 in the following equal groupings: White normal language, White language impaired, African American normal language, African American language impaired. The subjects were administered the Test of Language Development-2P (TOLD-2P), the Nonword Repetition Task (NRT), and the Competing Language Processing Task (CLPT). Results indicated that all three measures differentiated normal-language and language-impaired subjects from one another. With regard to cultural group, confirmed speakers of African American English (AAE) with normal language scored significantly lower on the TOLD-2P compared to White normal-language subjects. Scores of the AAE-speaking subjects with normal language on the NRT and CLPT, however, did not differ significantly from the White normal-language subjects. These results suggest that AAE speakers with normally developing language (LN) may be at a disadvantage on tests of prior language knowledge and that processing tasks may be a useful tool in combination with other assessment measures to make less biased clinical decisions. EDUCATIONAL OBJECTIVES As a result of this activity, the reader will (1) be able to determine the utility of processing tasks in culturally unbiased language assessment. (2) The reader will be able to discriminate the difference between the results of a standardized language test and processing tasks on speakers of AAE.

Task Variability Effects on the Language Test Performance of Southern Lower Socioeconomic Class African American and Caucasian Five-Year-Olds

Twelve African American and twelve Caucasian preschool children were administered items from the Preschool Language Assessment Instrument (PLAI) under standard conditions and in thematic interactions (PLAI-T) to determine if task variability had an effect on language test scores. The African American group earned significantly higher test scores when the items were administered in the thematic mode as compared to the standardized test format, with the major score increases tending to occur on the more complex and difficult items. Clinical implications of considering task effects and dynamic assessment in multicultural assessment are discussed.

A Cross-Sectional Developmental Study of Final Consonant Production in Southern Black Children from Preschool through Third Grade

The sounds-in-words subtest of the Goldman Fristoe Test of Articulation (GFTA) was administered to 222 Black children in preschool through third grade. The children resided in rural east central Alabama, and used the Black English dialect common to that region. The childrens responses were analyzed using the PROPH computer program for analysis of phonological processes. The analysis revealed phonological process patterns similar to those reported in the developmental literature with the exception of final consonant deletion. The data suggest that southern Black children continue to delete final consonants well beyond the age indicated by norms gathered on predominantly White subjects. Clinical implications are discussed.

A Pragmatic Aspect of Language Sampling

This investigation studied language sampling with normal four- and six-year-old children in three conditions: (1) conversation; (2) picture description with the experimenter and child looking at the stimuli; and (3) picture description with the experimenter unable to view the stimuli with the child. The results show statistically significant differences in Developmental Sentence Score and MLU between the age groups and among the experimental conditions. They suggest that children of these ages are able to alter their linguistic behavior as a result of the listener’s perspective. Implications for language sampling and future research are discussed.

Topic manipulation and cohesive adequacy in conversations of normal adults between the ages of 30 and 90

Few investigations have studied changes in normal conversational performance throughout the lifespan. The present study gathered preliminary data on 40 normal-language subjects between the ages of 30 and 90 years. The subjects were matched for age and sex in dyads and engaged in conversations. The conversations were analyzed for topic manipulation skills (introduction, reintroduction, maintenance, shading) and cohesive adequacy (complete ties, erroneous ties, incomplete ties). Significant differences were found between older and younger dyads on the topic and cohesion variables.

Joint Book-Reading Strategies in Middle-Class African American and White Mother-Toddler Dyads: Research Note

Twenty middle-socioeconomic class mother-toddler dyads were video recorded during three joint book-reading activities. Ten of the dyads were White, and 10 were African American, who were balanced for family income and parental occupation. The children ranged in age from 18 to 30 months and were normally developing. The parents read an experimental book to their child two times and a favorite book they brought from home one time. Videotapes of the joint book-readings were analyzed to determine cultural differences and the effects of book familiarity on the occurrence of maternal and child communication behaviors. The results do not support the findings of an earlier study in which differences in caretaker questioning behaviors were found between working-class African American and White participants. In contrast, the present study showed many similarities between the cultural groups in joint book-reading behaviors. However, the White group used significantly more labeling as compared to the African American group. These findings suggest the importance of considering socioeconomic level in multicultural research. Several effects of familiarity were also found. The findings are compared to anthropological reports on caretaker-child interaction in African American families, and implications are discussed.

Semantic and phonological performance in adults learning novel object and action words

Twenty normal adults participated in two training sessions designed to teach eight novel object words and eight novel action words. Each subject received comprehension and production training and was probed at the end of each session to determine the ability to retrieve lexical items in a naming task. The probe data were also analyzed for phonological accuracy. Reaction time measures for the production probes were recorded to determine differences between action and object words in terms of retrieval difficulty. The results demonstrated a significant difference in early learning of action and object words during comprehension training and production training and in production probes. Object words tended to be learned more easily in both comprehension and production training, and they were retrieved correctly more often in production probes. There were no significant differences in the phonological accuracy between the word types. Reaction time data indicated significant differences in early as well as later learning between action and object words, with object words having significantly shorter rection times as compared with action words. Results are discussed in the context of prior research.

Topic Manipulation and Conversational Participation as a Function of Familiarity in School-Age Language-Impaired and Normal Language Peers.

This study investigated the topic manipulation skills and conversational participation of school-age language-impaired children (LI) in interactions with normal language peers. The subjects consisted of eight dyads of a normal language child and a language-impaired child balanced for approximate age and sex. They ranged in age from 5.11 to 7.11 years of age. The subjects participated in two conversations in order to explore the effects of familiarity on conversational participation. The subjects were unfamiliar with each other before the study began. Fifteen to twenty minutes of discourse were taped, transcribed, and analyzed for the number and proportion of topics maintained, topics introduced, topics reintroduced, topics shaded, and the number and proportion of back channel responses produced in Session 1 and Session 2. A behavioral measure was included to clarify whether conversational nonassertiveness was related to a behavioral style of interaction. No significant differences were found between the LI children and the normal children for the number and proportion of topics maintained, topics introduced, or topics shaded. However, the LI subjects did produce significantly more topic reintroductions than the normal subjects. The number and proportion of back channel responses was not found to differ significantly between the two groups. A familiarity effect was demonstrated by a reduction in the number of utterances produced in Session 2 as compared to Session 1 as the children interacted less in the second conversation. Several interpretations of the results are discussed.

Language Performance in Low-Achieving Elementary School Students.

The purpose of the present study was to compare normally achieving and low-achieving (LA) elementary school-age children on a wide range of linguistic tasks. Over half of the LA group earned scores on the language measures that were low enough to suggest consideration for evaluation and/or treatment. Implications for referral and treatment of the LA population are discussed.

Profiling teacher/child classroom communication: reliability of an alternating time sampling procedure

Reliability of an on-line, alternating time sampling (A TS) procedure was examined for evaluating classroom communication acts. Fifteen limited-language children were videotaped during classroom activities while two trained judges simultaneously coded teacher/child com munication acts on-line. Judges systematically alternated attention from child to teacher each minute during sampling. Analyses revealed high interjudge reliability for live coding and high agreement between A TS and continuous coding on all measurements. ATS appears practical for classroom assessment and measurement of treatment effects.