The much discussed 'whole language' or 'real books' approach to children's
reading is based on a theory which makes two quite separate predictions: (1
) that children will read difficult words more successfully in context than
in isolation, and (2) that contextual experience of words will lead to gre
ater improvements in word reading than isolated experience of words. We rep
ort a study which tests both predictions and supports the first but not the
second.