The dictionary definition of "assessment" does not include the taking of high-stakes final exams. But that's how the word is being increasingly used. In the context of schools' legal requirements to meet Adequate Yearly Progress goals, assessment means accountability as defined by the reporting requirements of the No Child Left Behind Act.

For some people, this is a strategic opportunity to gain support for technology as an indispensable tool for test data collection, analysis, and reporting. And these people are correct. Leaving aside the question whether these tests actually measure what we need to know about creating future generations of well-informed, caring, and productive human beings, the nationwide pressure for accountability seems to be one of the few remaining drivers of technology investment that has survived the transfer of government money from education to tax cuts and war.

However, for those of us primarily motivated by the ability of technology to enrich teaching and learning, this is a much more ambiguous opportunity. Not only do we want assessment to mean something other than a way of revealing--and punishing--failure, we're not sure that assessment is the most important contribution technology can make in the first place.

I recently attended a forum on technology-based assessment at a national conference. The podium speakers correctly noted that technology makes testing more efficient. It cuts the turn-around time so teachers can get results more quickly. It facilitates disaggregation and submission of data. They talked about the coming use of technology for test distribution and test taking. These are exciting and powerful uses of technology that will stabilize its place in the education system.

In my own work as executive director of Mass Networks Education Partnership, I've learned that using technology to examine high-stakes test data can be a powerful stimulus for change. Every district we work with has a set of local "myths" that everyone believes, but not all of which are actually true. Often, there is a widespread belief that a particular group of students is dragging down the overall test scores--the special-education kids, the low-income kids, the kids who transfer into the district from neighboring towns, the non-white kids. It's always someone. Once the data are examined, however, the negative impact of the "problem group" often turns out to be much less significant than generally believed--or simply untrue. By dissecting aggregate numbers into subgroup trends, technology-based data analysis can undermine myths and dispel stereotypes.

We've also learned that data from high-stakes tests are best used to analyze group trends over time. This type of data is, as most teachers know, much less useful for revealing individual students' needs. These summative tests need to cover too much ground to delve into any particular area in the depth required to really expose a particular student's learning style, strengths, and weaknesses. In addition, any one student's scores will fluctuate on any one day based on numerous factors having little to do with what the student actually knows or is able to do. It is only by aggregating large numbers of separate scores that the distortions produced by arbitrary fluctuations are reduced enough for valid insights to emerge.