One way to manage marking error, in a large-scale educational testing context, is to establish a mechanism through which appeals can be lodged. While, at one level, this seems to offer a straightforward technical solution to the problem of marking error, it can also result in unintended consequences, with political, social or educational ramifications. It is therefore important to monitor the operation of any appeal system, to determine how effectively it meets its objectives. The present paper was based on an evaluation of the system which operates for National Curriculum testing in England. Four underlying objectives were identified: the measurement objective, the political objective, the educational objective and the psychological objective. Although there is reason to believe that such goals can be achieved through appeal systems, there are major threats to achieving them, many of which appear to be inevitable. These threats are examined within the paper and implications for policy and practice are explored.