2 authors: Gina Biancarosa University of Oregon 49

Tools for Individualizing Supports Other articles in this issue explore how disparities in students’ skills and knowledge, combined with reading and learning impair- ments, complicate the task of improving literacy outcomes for all learners. Teachers
charged with delivering differentiated instruc- tion to meet the individualized needs of learners must often do so by trying to retrofit a one-size-fits-all curriculum to meet the needs of diverse learners—a cumbersome and time-consuming process.⁴⁶ Moreover, unless carefully designed, e-reading technology itself can replicate the problem, thus reproducing old barriers and generating new ones that marginalize diverse learners.
CAST (originally the Center for Applied Special Technology) uses an approach called Universal Design for Learning (UDL) to design e-reading technology that attempts to meet the needs of individual learners by assuming and taking into account their diverse needs.⁴⁷ A key aspect of UDL is to provide multiple ways both for students to gain knowledge and skills and also for them to express and apply that knowledge. In the case of e-reading technology, tools like text-to- speech, automated tutors, and individualized levels of support are built into e-reading applications from the beginning rather than being added later. Although the concept of UDL itself is not new, technological advances increase the feasibility of providing a wide range of supports to meet the needs of every learner. Research on matching students to technologies is still at an early stage.

Tools for Assessment

E-reading technology, particularly its instructional applications, often incorporates mechanisms for gathering data on students. The data may be restricted to use patterns, such as frequency and duration of use, or
it may extend to assessment of learning by incorporating placement and mastery assessments. Because studies of e-reading instructional tools have not examined
whether they are as effective with assessment as without it, we review briefly a few examples from the wide and increasing range of technological innovations for literacy assessment. Because space does not permit a full discussion of these innovations, we must overlook important ones such as clickers, automated scoring of written and spoken answers, and innovative assessments of higher-level comprehension skills.⁴⁸

One of the most popular tools for assess- ment in literacy (and beyond) has been

computer-adaptive testing (CAT). Regarded as an innovation a decade ago, CAT has become a mainstay of large testing firms. The Educational Testing Service regularly uses it
for online tests, and reading achievement tests, including the Computer Based Assessment System for Reading, Measures of Academic Progress, Scholastic Reading Inventory, and STAR Reading, are increasingly available in online CAT formats. Many states, including Florida, Maryland, and Oregon, have invested in online CAT systems for one or more state accountability tests. What CAT offers is an assessment that adapts to the test-taker.
Students who answer questions correctly are given questions of increasing difficulty, while students who respond incorrectly are given questions of decreasing difficulty. Each stu- dent thus completes a large number of items at her or his difficulty level, leading to a more precise estimate of the underlying ability being assessed. Although some observers have raised concerns that early careless errors may lead
to underestimates of student abilities, recent evidence suggests that such underestimation is rare and occurs primarily for students of very high or very low ability.⁴⁹

The turn to computerized delivery of assess- ments has raised concerns that such assess- ment, adaptive or not, might pose particular difficulties for anxious test-takers or those with less computer experience. Although evidence is limited, comparisons of adults taking the GRE suggest that anxiety is a strong predictor of performance and that computing confi- dence is a weak but significant predictor—but also that neither depends on the format in which a test is delivered.⁵⁰ Other research with adults suggests that older adults may compre- hend less and read less efficiently using computer screens than using paper, whereas younger adults show no difference.⁵¹ Studies with intermediate, middle, and high school