Cheating on the Rise

K-12 Testing

As more and more states and districts impose high-stakes tests for promotion and graduation, the number of reported cheating instances by administrators, teachers and students is on the rise. For example, Georgia's Professional Standards Commission saw one case of improper administration of standardized tests three years ago, but 25 cases from July 1998 to March 1999. The violations ranged from giving students the answers to using practice tests based on previous years' exams that had not been publicly released.


Similarly, in Texas, which has been highly touted for the positive consequences of its state testing program, criminal charges for altering students' records to improve school-level test scores have been lodged against one administrator in Austin, and the entire district faces civil charges and a fine. Investigations of erasures on test forms were ordered in eleven districts, including Dallas and Houston, where a principal and three teachers have been asked to resign.


In Arizona, the state is seeking to revoke the licenses of an administrator who made copies of AIMS exams and gave them to consultants hired to help teachers raise test scores. Cheating cases have started to sprout like mushrooms in Alabama as well, where five schools are under investigation. And in Virginia, a principal has been reassigned for copying questions from the state test and providing teachers with similar questions.


While erasing mistakes and changing answers are fairly blatant forms of cheating, teaching to the test, an increasingly common practice, can also raise ethical questions. William Mehrens and John Kaminski, writing in Educational Measurement (Spring 1989), outlined a 7-point continuum of test preparation activities leading from general instruction on test objectives to practice on the actual exam. Teaching to the test is a borderline issue, they concluded, while teaching the exact items is clearly cheating. The authors argued that even the mid-point on their scale, "instruction based on objectives... that specifically match those on the ...test," may well be illegitimate. By pressuring schools to teach to the test, states and districts themselves encourage this form of questionable behavior.


Major publishers and test-coaching companies feed the frenzy by selling materials clearly linked to specific exams. Mehrens and Kaminski argued that the Random House Scoring High series was so similar to the tests that their use was a form of cheating.


Cheating -- or simply getting the high scores you need by any means -- is the easy way out, and an all-too-human reaction, when high-stakes decisions are based on tests. This is particularly true of multiple-choice tests, where all it takes is an eraser and a No. 2 pencil to boost scores. This sort of cheating would be less likely with performance assessments, but with high stakes educators often feel pressured to figure out how to "pick the lock" of any one-shot test -- by coaching, teaching to the test, teaching the actual test, helping students answer questions, or changing students' answers.


For some experts, the solution to cheating is greater vigilance, such as having tests administered the way college entrance exams are, using independent proctors. Research, however, has shown that scores of some minority children decline when tests are administered by strangers. Thus, the "solution" creates even more problems.


The push for increased test policing is an example of a dangerous prescription based on a faulty diagnosis. It is the policy of high-stakes testing that produces the unwanted side-effect of cheating. Rather than intensify policing, it is time to rethink the basic policy of high-stakes testing.