One solution to the problems of discrete-point testing are tests of reduced redundancy, that is, tests that introduce interference into the communication through actual introduction of noise or through deletion of text and that ask the learner to perform in spite of the noise. Three test types were originally developed to operationalize the technique of reduced redundancy: the noise test, in which varying levels of acoustic noise were introduced during dictation of short sentences; partial dictation based on extremely difficult tests, omitting segments of two to ten words; and cloze tests, with systematic deletions of every fifth to tenth word. While each test type is found to have problems of validity or viability, the original theory remains convincing. A fourth test type, called the C-Test, has been developed as a better operationalization of the theory. (MSE)