How manipulating test scores affects school accountability and student achievement

Standardized testing can create incentives to manipulate test results and generate misleading indicators for public policy

Queen Mary University of London, UK, and IZA, Germany

one-pager full article

Elevator pitch

Standardized testing has become the accepted means of measuring a school’s quality. However, the associated rise in test-based accountability creates incentives for schools, teachers, and students to manipulate test scores. Illicit behavior may also occur in institutional settings where performance standards are weak. These issues are important because inaccurate measurement of student achievement leads to poor or ineffective policy conclusions. The consequences of mismeasured student achievement for policy conclusions have been documented in many institutional contexts in Europe and North America, and guidelines can be devised for the future.

Effects of “borderlining” on test results
                        in the UK

Key findings

Pros

Each investigation of test score manipulation must be done in context and requires qualitative indicators to assess the extent of the problem.

Simple indicators can often characterize the nature of the test score manipulation.

Controlled retesting of students, grading from independent markers, or random assignment of external monitors on the test day should be part of any testing protocol.

Cons

Testing manipulation is a pervasive problem that may follow from accountability pressures, ineffective implementation of testing protocols, or student cheating.

Manipulation of test results distorts student performance indicators leading to misleading evaluations of the effectiveness of teachers and school programs.

The manipulation of test results is intentional; as such, performance indicators obtained by dropping corrupted data are not reliable.

Author's main message

Manipulation distorts the accuracy of student achievement indicators, calling into question their validity as a tool for evaluating teacher performance and to enforce school accountability policies. Evidence of compromised scores can be obtained using simple indicators, but these rarely reveal (with certainty) who manipulated test scores or the reasons for this behavior. The most compelling evidence on manipulation comes from controlled retesting of students and from random assignment of monitors to classrooms. These procedures should be part of any testing protocol in contexts where manipulation is a serious threat to the fidelity of results.

Full citation

Full citation

Data source(s)

Data type(s)

Method(s)

Countries