Previous studies examining withdrawal in children with language impairment (LI) have found that these children are more socially withdrawn than their typically developing peers. It seems reasonable to assume that a causal relationship exists between language deficits and withdrawal. However, there is growing evidence that different subtypes of withdrawal have varying social consequences and language may not be closely linked to each subtype. In the present study, subtests from the Comprehensive Assessment of Spoken Language (CASL; Carrow-Woolfolk, 1999) were used to evaluate specific language skills and the Teacher Behavior Rating Scale (TBRS; Hart & Robinson, 1996) was used to evaluate solitary-passive withdrawal behaviors of 34 children with LI and 34 children with typically developing language. These children were matched for age (ranging from 6;11 to 11;0) and gender. No significant difference in solitary-withdrawn behavior was found between groups of children with language impairment and children with typically developing language. There was also no significant relationship between the amount of solitary-passive withdrawal and the CASL subtest scores. The results suggest that low language ability may not be directly related to solitary-passive withdrawal. Rather, the relationship between language ability and solitary-passive withdrawal is complex and is likely influenced by other factors.

Children with language impairment (LI) often demonstrate increased levels of withdrawal, specifically the reticent and solitary-passive subtypes. Although it is recognized that there are several subtypes of withdrawal, the relationship between withdrawal and children with LI is complex and unclear. The aim of this study was to examine items on the Teacher Behavior Rating Scale (TBRS) to better understand the nature of withdrawal in children with LI. A factor analysis of TBRS scores of 355 participants was conducted to ensure that individual items on each subscale were measuring the same construct. An invariance analysis was also performed to ensure items were equally valid for both groups. Finally, a Pearson chi-square test was performed to see which items had the greatest power to separate typical and atypical children. Factor analysis confirmed that items on each subscale grouped together. All four reticence items were invariant, however only 3 of the 5 solitary-passive items were invariant across groups. Item analysis of the Reticence Subscale revealed that participants showed the greatest differences in items related to on-looking, unoccupied, and fearful behavior. Item analysis of the Solitary Passive Subscale revealed an increase in behavior related to a preference for solitary play. In accordance with previous work, children with LI displayed higher levels of both reticence and solitary-passive withdrawal than their typical peers. This study confirms the validity of using the TBRS as a tool to investigate the complex relationship between language and reticent behavior in future research.