PLACEMENT TEST VALIDATION: WAYS OF IMPROVEMENT

Nowadays Russian universities tend to provide foreign language learners with programs depending on the students’ level of foreign language proficiency. To determine their levels and readiness students have to do placement tests. But the results of the tests may direct students into the wrong level course. The fact is stated on the base of misplaced students’ requests to send them to lower or higher level courses as well as instructors’ identification that they have students erroneously directed into their sections. Having analyzed the data the authors found out that placement test results may fail to place students correctly. The results indicate that misplacement to a higher level takes place more often than to a lower one. The purpose of the paper is to study placement test structure and accuracy to make it more valid for our educational environment and consequently to minimize the number of misplaced students. To achieve this, the authors studied different models of placement tests to find out the most accurate one. The factors influencing the accuracy of placement test results were also determined. The misplaced students were questioned and the results of the questionnaire helped to define the reasons of the placement test failures. The present study applies mixed methods approach (traditional research methods of observation, product analysis, comparing and questionnaire). The resulting analysis of different placement test models discovered the one that seems to be the most accurate for our educational environment as it has a wide range of marks and covers all types of students’ skills. It measures both receptive (listening, reading and grammar recognition) and productive skills (speaking and writing) to provide comprehensive picture of students’ language ability. The good viability of the model was confirmed by the forthcoming analysis which also allowed the authors to measure the inaccuracy level of the test.