Furthermore, studies by Camine, Kameenui, & Coyle (1984) Beck, Mckeown &
McCaslin (1983) examined deliberate learning from context through the use of tasks that
explicitly asked readers to derive the meaning of an underlined word appearing in a text
passage. Carnine et. al. (1984) used experimental passages designed to present clues to
specific words and control for distance between the clue and the target word. They found
that fourth- and sixth-grade children were better able to identify meanings of words
presented in context than of words presented in isolation. However, even the best
performance, that of sixth graders, showed a success rate of only 40 percent, with type
and distance of clue making a difference. Synonym clues provided better results than clues
requiring an inference. Furthermore, clues closer to the target word were more helpful than
those far away. Clues that both required inferences and were distant from the target word
yielded correct outcomes only 17 percent of the time.
Beck et. al. (1983) used natural text to test their hypothesis of a continuum of effectiveness
of natural contexts for deriving word meaning. They identified four points along the
continuum: misdirective contexts, which seem to direct a reader toward an incorrect
meaning; nondirective contexts, which offer no direction for word meaning; general
directive contexts which offer correct but general clues; and directive contexts which
seem to lead to a correct meaning for a word. The researchers presented adults with stories
from fourth- and sixth-grade basal readers in which target words were blacked out and
asked them to supply the word or a synonym. The results varied greatly by context type.
Correct responses were given for 3 percent of the misdirective contexts, 27 percent of the
nondirective contexts, 49 percent of the general contexts, and 86 percent of the directive
contexts. Hence, Beck et. al.(1983) concluded that not all contexts are created equal.
The question of whether readers Ieam words incidentally from context through normal
reading was examined in studies by Jenkins, Stein, & Wysoski (1984) and by Nagy &
Herman and their colleagues (Herman, Anderson, Pearson & Nagy, 1987; Nagy, Anderson
& Herman, 1987; Nagy, Herman and Anderson, 1985). The tasks used in these situations
reflected a more natural reading situation in that readers were not told that the purpose was
related to word learning, and the target words were not identified.
60