Academic journal article
By Jackson, Evelyn W.; McGlinn, Shirley
Journal of College Reading and Learning , Vol. 31, No. 1
Optimal performance on standardized tests is an important goal for many college students, including those aspiring to professional or graduate school or those seeking certification. Test performance is dependent on the examinee's specific knowledge and abilities measured by the test as well as test-taking skills. Numerous test preparation resources, including books, software and coaching courses, have been developed in response to the proliferation of standardized tests. Since preparation may be costly and time consuming, students should make sure that any preparation strategy includes attention to the knowledge and skills that can make a difference. The focus of this study is to demonstrate the effect of knowing the test as one aspect of a test preparation strategy.
Review of Literature
Test-training activities for standardized tests range from brief instruction on test taking tips and techniques to longer interventions aimed at teaching content and improving skills measured by the test. Many claims of coaching effectiveness are anecdotal and do not describe the specific interventions in detail. College Board Online (January 11, 1999) reported that there is no credible evidence to support large increases in SAT scores. In a meta-analysis of studies on coaching for the SAT, Powers (1993) reported short-term programs result in small gains, the effect of coaching was greater for math than for verbal scores and that longer programs yielded stronger effects. In an earlier study Powers, Clark and Grandy score changes of coached and uncoached examinees on standardized tests were related to other group characteristics. For instance, examinees that sought training also used more methods of preparation, expressed more anxiety and applied to more selective schools.
A meta-analysis examining the effects of test training programs on achievement scores in high school demonstrated that longer training programs had significant effects on scores. In addition, higher effect sizes were reported when training involved coaching for a specific test than for providing general test taking skills (Samson, 1985).
Since all test takers are not trained in test-wiseness, training may be considered a source of test score pollution (Haladyna, Nolen & Hass, 1991) or even blatantly dishonest (Mehrens & Kaminski, 1989). However, to counter the possible pollutant effect Berliner and Cassanova (1986) suggested blending test-wiseness training into the curriculum so that all students have the opportunity to learn and practice the principles.
While there are contradictory reports on the appropriateness and effectiveness of training for a standardized test, there is even less known about what components of training actually make a difference in the examinee's score when differences are reported. Studies of test-wiseness have examined the effects of answer changing behavior, number of options in a multiple choice question, and the interaction of test-wiseness and knowledge of the content tested (Geiger, 1997; Rogers & Bateson, 1991; Rogers & Harley, 1999). Hembree (1987), analyzed 120 studies of test-wiseness and concluded that test-wiseness is one factor that can affect performance.
Test scores of those who repeat a test can improve, remain the same or even be lower. Change may be accounted for by examinee growth or diminishing of abilities, skills or knowledge being tested, by intervening test-wiseness training, by practice on the first test or they may be due to characteristics of the test itself (Berliner & Cassanova, 1986; Powers, 1993; Powers, Clark & Grandy, 1985; Roznowski & Bassett, 1992; Samson, 1985). Even if the testing instrument is reliable, the change in scores may be due to measurement error (Alderman, 1981; Powers, 1993) or factors such as regression toward the mean (Powers, 1993; Wilson, 1988).
Performance patterns for examinees who repeat standardized tests indicate that measures of verbal ability tend to be stable, showing small variability in score changes. …