research. In particular, whether knowledge and interest were related de-
pended on the type of knowledge and the form of interest being measured.
When researchers examined domain or topic knowledge in relation to indi-
viduals’ abiding interest in that domain or topic, the positive association be-
tween knowledge and interest emerged. Moreover, the strength of that asso-
ciation became stronger as knowledge in the domain rose.
Second, motivational and affective variables, not often studied in the read-
ing literature, could be potent forces in students’ text-based learning (Hidi,
1990; Wigfield & Guthrie, 1997). Students who reported higher interest in a
particular topic (e.g., frogs) or in reading performed better on text-based
learning tasks than their less interested classmates. Third, there were often no
pre-reading measures of knowledge or interest that allowed post-reading
data to be more adequately judged (Alexander, Kulikowich, & Jetton, 1994).
Consequently, individual or group differences that might preexist an inter-
vention or experiment could not be eliminated as factors in outcome effects.
In hindsight, it seems evident that the more researchers knew about partic-
ipants prior to any reading activity or experimental intervention, the better
they could determine the degree to which knowledge or interest interacted as
consequence of that activity or intervention.
As the picture of text-based learning became more complicated, I found
that a model that captured the interplay among critical forces over time was
needed. Once the MDL was initially formulated and predictions forwarded,
studies were undertaken to test the hypothesized relations within-indi-
viduals over time and across domains, as well as between individuals for the
same domain.
THE MDL: EXPERTISE AS ACADEMIC
DEVELOPMENT
While I refer to the MDL (Fig. 10.1) as a model of developing expertise in ac-
ademic domains, there have been evolving versions of the model. Since 1993
research has resulted in a multidimensional and multistage representation of
domain learning. The earliest prototype for the MDL, which appeared in Al-
exander, Kulikowich, and Schulze (1994), was quite stark by comparison, of-
fering simple linear contrasts between subject-matter knowledge and two
forms of interest. Before discussing the findings and implications of MDL
studies, I want to overview the central dimensions and stages of the model.
Multiple Dimensions of the MDL
Three dimensions frame the MDL—knowledge, interest, and strategic proc-
essing—with subcomponents to each (Alexander, 1997). As the traditional ex-
pert–novice research strongly established, knowledge and strategic processing
- MODEL OF DOMAIN LEARNING 283