Extended Time Testing Accommodations for Students with Disabilities: Impact on Score Meaning and Construct Representation

By their very nature, standardized tests are designed to be administered to all examinees under similar conditions. Indeed, variability in administration across examinees is a clear threat to test fairness (Wollack & Case, 2016). That being said, testing accommodations must be provided when a disability condition keeps an examinee from accessing the test under standard administration conditions. Two examples of such accommodations—which change the administration conditions of the test in some way without changing the actual test content (Lovett & Lewandowski, 2015)—are reading the test items aloud to a visually impaired examinee and providing an examinee in a wheelchair with a wheelchair-accessible desk for the testing session. Accommodations are required under disability discrimination and special education laws when needed for examinees to access tests, and the Standards for Educational and Psychological Testing (American Educational Research Association, American Psychological Association, & National Council on Measurement in Education, 2014) recognizes the responsibility of test developers and users to develop and provide appropriate accommodations.

Testing accommodations are common; during the 2015-16 test cycle, the College Board (which administers the SAT, among other tests) received 160,000 requests for accommodations—double the number received just 5 years prior to that (Yellin, 2017). Accommodations often modify the examinee’s response mode (e.g., allowing dictation of answers), the presentation format of the test (e.g., an audio recording of test items), or the scheduling of the test (e.g., in the afternoon, due to health problems that flare in the morning). The most common accommodation on virtually every test is extended time, and it is exactly what it sounds like: giving an examinee additional time to work on the test (see, e.g., U.S. Government Accountability Office, 2011). The present chapter focuses on this accommodation, with particular attention to its impact on the meaning of resulting test scores and more broadly on the ability of the (accommodated) test to measure all aspects of the intended construct.1

The Popularity of Extended Time Accommodations

Many students receiving extended time do not have any sensory or physical handicaps; instead, they have been diagnosed with learning, cognitive, or psychiatric disabilities—conditions that together are often called “hidden disabilities.” On admission, certification, and licensing tests, learning disabilities and attention-deficit/hyperactivity disorder (ADHD) are the most commonly accommodated conditions (U.S. Government Accountability Office, 2011). Similarly, in school settings, the highest-incidence special education classification is learning disabilities (Heward, 2013). Many testing agencies also have reported an increase in the number of students requesting extended time for psychiatric conditions, particularly anxiety. The high incidence of hidden disabilities is one reason why extended time is such a common accommodation; another reason is that students who use auxiliary aids or human assistants (e.g., a live test reader) often need additional time to utilize these other accommodations.

Of course, time pressure during tests is an experience that is common to examinees both with and without disabilities. Time limits are a known contributor to test anxiety, and so it is not surprising that time extensions would be viewed as desirable by many examinees. In their review of the literature on students’ perceptions of testing accommodations, Lovett and Leja (2013) concluded that accommodations reduce anxiety and discomfort for students with and without disabilities. Extended time specifically has been found to do this; for instance, Elliott and Marquart (2004) found that giving fourth-grade students 40 minutes instead of 20 minutes to complete a math test (i.e., 100% extended time) made the majority of students with and without disabilities feel more relaxed (78% and 75%, respectively). The reason for this relaxation may be an expectation of score improvement; in one large-scale survey, Lewandowski, Lambert, Lovett, Panahon, and Sytsma (2014) found that the vast majority of college students with and without disabilities (approximately 87% of each group) felt that 50% extended time would lead to at least somewhat higher scores for them on a high-stakes test.

Students’ perceptions of extended time correspond to the actual effects of the accommodation; extended time consistently has been found to improve scores on standardized tests for students with and without disabilities. Although some disability advocates (e.g., Shaywitz, 2003) claim otherwise, three systematic literature reviews (Cahan, Nirel, 8c Alkoby, 2016; Lovett, 2010; Sireci, Scarpati, 8c Li, 2005) all have found this nonspecific effect of extended time. Whether students with disabilities benefit more from extended time than do nondisabled students is a more complex issue that appears to depend on just how speeded the test is. On moderately speeded tests, some nondisabled students are more likely to finish within the standard time limits, making the extended time of negligible value to that subgroup; on highly speeded tests (cf. Lewandowski, Lovett, Parolin, Gordon, 8c Codding, 2007), most or all nondisabled examinees are working throughout the full (extended) time allotment and complete more work during that allotment than do examinees with disabilities, thus benefiting more. Finally, some teacher-made classroom tests have such liberal time limits that extended time is of little value to anyone with or without a disability. On these tests, most students who receive extended time accommodations actually complete their exams within the standard time allotment (see, e.g., Spenceley 8c Wheeler, 2016). However, when there is at least some time pressure, students with and without disabilities tend to benefit from extended time.

Because extended time is so widely desired (and so beneficial), it is especially important to develop guidelines for when it should be given. Although no consensus currently exists on detailed decision-making algorithms for extended time, Phillips (1994) offered five questions that should be asked before providing any accommodation. The questions, which follow, raise issues that remain salient a quarter of a century later;

  • 1. Will format changes or alterations in testing conditions change the skill being measured?
  • 2. Will the scores of examinees tested under standard conditions have a different meaning than scores for examinees tested with the requested accommodation?
  • 3. Would nondisabled examinees benefit if allowed the same accommodation?
  • 4. Does the disabled examinee have any capability for adapting to standard test administration conditions?
  • 5. Is the disability evidence or testing accommodations policy based on procedures with doubtful validity and reliability? (Phillips, 1994, p. 104)

The present chapter will focus on the first two of these questions, presenting recent research results and analysis (for an earlier review of research on all five questions, see Lovett, 2010). Phillips’s second—and narrower—question is addressed first: do test scores obtained under extended time conditions have a different meaning than those obtained under standard time conditions? The focus here is on studies examining the predictive validity of test scores obtained with and without extended time accommodations. The chapter then turns to Phillips’s first question: do time extensions change the nature of the skill(s) being measured by a test in ways that are problematic, leading to construct underrepresentation for students receiving accommodations and to construct-irrelevant variance given the varying time allotments for different examinees? Even if the meaning of the score changes, does that imply that the test is no longer measuring the construct adequately? This broader question is examined by sampling results from a variety of different types of relevant research. The chapter concludes with recommendations for making decisions about extended time accommodations based on the research reviewed.

 
Source
< Prev   CONTENTS   Source   Next >