domains across individuals. For instance, an undergraduate student reading
about Hawking and Grand Unification Theory or about quarks might well
perform differently on those two passages. However, when the data for all the
undergraduates were averaged and results correlated, predictable associa-
tions between knowledge, strategy use, and text learning emerged (Alexander
& Kulikowich, 1994). Those with demonstrated knowledge of the subject
matter reported more strategy use during reading and remembered more
from the passages than those with little relevant knowledge (e.g., Alexander,
Kulikowich, & Schulze, 1994).
Third, understanding the relation between knowledge and strategic proc-
essing was highly dependent on our ability to craft measures that are not
only statistically reliable and valid but also relevant to educational practice.
Our experience at creating reliable and valid subject-matter tests was duly
informed by the extant assessment literature (e.g., Linn & Gronlund, 2000).
However, the improvements we made in our knowledge measures over the
past 15 years have contributed to our modeling of domain expertise. For ex-
ample, when we devised multiple-choice tests of domain knowledge, we
identified a response-option model appropriate for the target population
and we followed that model consistently in item generation. In Alexander et
al. (1989), for instance, the human biology model for sixth grade was as fol-
lows: human biology (HB) correct, HB incorrect, science not HB incorrect,
and nonscience incorrect. Such hierarchical options permitted more sophis-
ticated analyses of students’ knowledge than is typical for multiple-choice
measures.
Knowledge and Interest Studies
As studies of knowledge and strategies were continuing, I was drawn into
questions about the relation between knowledge and interest. At the time,
there was debate in the literature as to whether students’ reported interests re-
flected their knowledge (Tobias, 1994). On one side was the argument that
these two constructs were not well-linked (Schiefele, 1991), a position sup-
ported by studies in which student knowledge was controlled or in which
readers processed texts on an unfamiliar topic. On the other side was the ar-
gument that knowledge and interest appear highly linked, especially for those
more competent in a subject (Renninger, 1992). Ultimately, our studies of
knowledge and interest helped clarify the knowledge–interest relation, and si-
multaneously taught my colleagues and me a few critical lessons about the
role in interest in developing expertise (e.g., Alexander & Jetton, 1996; Gar-
ner, Alexander, Gillingham, Kulikowich, & Brown, 1991).
First, it was evident that clarifying the relation between interest and
knowledge required a level of specificity to the constructs not common in the
282 ALEXANDER