The first important finding was the confirmation of the previously estab- lished two-factor model of judgement. Second, relationships were found between teacher ratings and the status and ethnicity of the children; thus, low-status children were seen as more ethnic/non-standard, and were rated less favourably on the confidence/eagerness dimension. Within this low- status group, white children were seen most favourably. For middle-status children, results were similar with regard to the ethnicity/non-standardness dimension. However, on the confidence/eagerness dimension, black middle- class children were rated slightly more favourably than were white children (see Giles and Bourhis, 1976a), and both black and white children were rated more positively than were the Mexican-American children. A third finding was that the teachers' general stereotypes (of the groups overall) correlated reasonably well with their evaluations of actual group members. Williams suggests, then, that the tnore general teacher stereotypes may have affected judgements made of individual children (and this, of course, is exactly the problem referred to above). Fourth, predictions of children's scholastic achievement were related to the speech evaluations. It is of interest here that the ratings made by black and white teachers were remarkably similar. This work by Williams has been discussed in some detail here since it represents a series of investigations in which similar results have emerged, and since it underlines once again the power of speech patterns to influence teacher ratings specifically.