ABSTRACT
Researchers concerned with theoretical or applied aspects of language ability and disability typically ask questions of the form, “Do children with a particular developmental disability and children who are developing normally show the same or different developmental relations between particular aspects of language and/or cognitive development?” This question is then tested by matching the two groups on one aspect (control variable) and testing for group mean differences on the variable of interest (target variable). If the two groups do not differ significantly on the control variable, but do differ significantly on the target variable, then it is typically concluded that the group with the disability does not show the normal relation between the target and control variables. In most cases, the group with the disability performs significantly worse than the control group, leading the researchers to conclude that the disability group shows a deficit on the target variable. For example, suppose a researcher wishes to determine if children with specific language impairment (SLI) have more difficulty with grammar than typically developing children at the same nonverbal cognitive level. The standard way to address this question has been to conduct a t test comparing the scores of the two groups of children on the measure of nonverbal cognitive ability (e.g., Columbia Mental Maturity Scale; Burgemeister, Blum, & Lorge, 1972), confirm that the two groups do not differ significantly on this measure (p>.05), and then conduct another t test compar-ing the grammar scores (e.g., mean length of utterance, MLU) of the two groups. If the grammar scores are significantly different, then the typical conclusion would be that children with SLI have a deficit in grammatical ability.
