Efforts to extract the elements of LD are plagued by the well-known problem of
heterogeneity (see Keogh, 1987b). A candidate deficit may apply to a large
number of LD students, but fail to apply to an equally large number. For this
reason, a multiple-syndrome paradigm may be necessary to describe LD ( McKinney
, 1988).

In a multiple-syndrome paradigm, we must determine how many variables
should be included, how they should be chosen, and whether some variables are
more important than others. It is clear from large-scale investigations (e.g.,
Kavale &
Reese, 1992; Kirk & Elkins, 1975; Shepard, Smith, & Vojir, 1983)
that LD populations reveal a diverse assortment of problems. Consequently, the
selection of any subset of problems is likely to engender acrimonious debate
( Wong, 1986). In place of contentious debate, perhaps the selection of variables
could be based on an objective, empirical methodology. This was attempted by
Kavale and Nye ( 1985- 1986), in a quantitative synthesis of the LD research
literature.

The goal was to identify deficit areas where the performance differences
between LD and non-LD students were of sufficient magnitude to qualify as
primary components of LD functioning. Suppose, for example, that an LD
sample demonstrates performance levels significantly below that of an non-LD
comparison group on some variable. This variable will assume great importance
in any description of LD, because it is better able to differentiate LD functioning
from non-LD functioning. Those variables that demonstrate the greatest differen-
tiation between groups may then provide a foundation for speculating about the
nature of LD.

-221-