To Test (or Not) in Arena Settings
Kiger, Derick M., Florida Journal of Educational Administration and Policy
Public schools face raised performance expectations on group-administered achievement tests. In response, educators have implemented various preparation strategies to ready students. However, the research community largely ignores the impact of setting on test performance. This randomized field trial assesses whether or not there were significant performance differences between two group test-taking conditions (classroom vs. arena) on the Wisconsin Knowledge and Concepts Exams. Participants included tenth-grade students from a medium-sized Wisconsin high school. Results showed no overall achievement advantage for classroom test takers. However, students in the top GPA quartile performed better in the classroom setting while students in the bottom GPA quartile performed better in the arena setting. Implications for school officials are presented.
Public schools face raised performance expectations on group-administered achievement tests due to the accountability requirements of No Child Left Behind (NCLB). According to this federal legislation, public schools must demonstrate adequate yearly progress on statewide reading and math tests to avoid sanctions. In response, educators have implemented various strategies (e.g., curriculum and assessment alignment; practice tests, performance incentives, etc.) to ready students for testing. However, the research community has largely ignored the impact of group setting on student test performance, a potentially significant and controllable test score contaminant. This study helps bridge this gap.
Optimal Learning Environments: Let's Get Small!
Most educators and parents prefer small schools and class sizes for their children due to individualization of the teaching and learning environment (Finn, 2002; Graue, Hatch, Rao, and Oen, 2007; Kiger, 2002). Teachers have more time to know and connect with students, manage student behavior, and use a variety of instructional and assessment techniques resulting in positive student outcomes (Archibald, 2006; Howley and Howley, 2006; Nye, Hedges, and Konstantopoulos, 2004; Smith, Molnar, and Zahorik, 2003; Wasley, 2002). The down side of small-scale learning environments is high programming costs, limited space, and availability of qualified teachers (Brewer, Krop, Gill, and Reichardt, 1999; Hanushek, 1999; Hruz, 2000).
Optimal Test Setting: Let's Get Small?
The belief in "small learning" is evident in testing practices, too. For example, The Wisconsin Department of Public Instruction (2005) advises schools to administer its standardized achievement tests " ... in small groups of classroom size, rather than in a large group and auditorium-type hall." Presumably, smaller settings are familiar and authentic for students while affording proctors more control over testing procedures. The expected result: students feel more comfortable, and hence, less anxious, distracted and stressed during testing.
Unfortunately, there is little research available to recommend a classroom-like test setting over other group configurations. Ingle and Deamico (1969) compared the scores of high school students taking the Stanford Achievement Test in a 1,000-seat auditorium under relatively poor physical conditions (e.g., inadequate lighting, portable writing surfaces, etc.) with students taking the test in classrooms under normal conditions. The researchers found no significant total test score differences between the groups (d = +0.03) and concluded that large-group testing may be a viable option for schools.
Hembree (1987) conducted a meta-analysis of 120 studies to determine the effect of noncontent variables on math scores. Eight of the studies considered the impact of group size (large > 70 test takers vs. small < 30 test takers) on the test performance of fifth- to eighth-grade students. The mean effect size was not significant suggesting test performance is insensitive to group size. …