The primary aims of this study were to derive a taxonomy of the constr
ucts used by primary teachers in assessing children's work in visual a
rt, music and creative writing, and to investigate the extent to which
these constructs can be used consistently by different teachers. They
were operationalized in the form of a series of rating scales derived
from repertory grid type analyses of teachers' descriptions of arts a
ctivities and artistic products. In two pilot studies and a main study
, which included two in-service day conferences, teachers from 17 prim
ary schools evaluated children's products from activities which were o
perationally defined as either structured or unstructured in each of t
he three artforms. Statistical analysis of the ratings revealed three
clear findings, namely (1) that there was a high level of intercorrela
tion between individual teachers' ratings of different pieces of work
across all scales, (2) that there was a high level of intercorrelation
between the different rating scales employed in each of the artforms
and (3) that the products of unstructured activities received signific
antly higher overall ratings on those of the scales with an evaluative
component than those from structured activities. These results have i
mportant practical and theoretical implications for further research o
n assessment in the arts.