Finally, investigators should develop and validate integrated writing assessment systems
that provide immediate, instructionally relevant multi-vector data to teachers so that they are
better equipped for pinpointing writing problems and responding accordingly. Thus far, no
written language measurement approach appears to be adequate for this demanding task.
Portfolios lack sufficient reliability and immediacy, though they do offer teachers and students a
mechanism for deep reflection about writing processes, performances, and beliefs. CBM
provides reliable and immediate information, but its relevance to teachers’ instructional choices
is questionable because the measures themselves do not reflect the complexity inherent in most
writing tasks. Standardized tests provide reliable data, some of which may help pinpoint specific
deficiencies, but these data are summative rather than formative and thus are too far removed
from daily writing instruction. An integrated combination of these approaches, perhaps coupled
with computer-assisted delivery and interpretation (see Shermis, Burstein, & Leacock, 2006),will likely confer greater advantages to instructional design and student achievement than any
one alone.