An experiment was conducted to explore the effect of placing test items in differentrnpositions in listening comprehension tests. In this experiment three groups were made tornview listening comprehension test items in three different positions. The first grouprnviewed these items before two listenings while the second one viewed them after twornIistenings. The third group, on the other hand, viewed the questions between twornlistenings.rnFirst, Test of English as a Foreign Language (TOEFL) was used to divide the testees in tornthree groups which had comparable listening abilities. The data obtained from thisrnstandardized test was analysed using descriptive and inferential statistics. The result thenrnshowed that the three gro ups were comparable in their listening abilities.rnAfter this was done, a test prepared for College English students by the Department ofrnForeign Languages and Literature at Addis Ababa University was administered torninvestigate the mentioned treatments.rnTo accept or reject the null hypothesis which stated that placing test items in differentrnpositions in a listening comprehension test has no effect on the testees' performance, arnone-way ANOVA was employed. The AN OVA, then detected a significant mean ditlerencernin scores across the three approaches. The calculated value of F (F=8-440 with degrees ofrnfreedom 2 and 132) exceeded the value of F tabulated at .05 probability level set for thisrnstudy. As a result, the null hypothesis mentioned above was rejected. The Scheffs posthocrncomparison of means also helped to locate the direction of the difference. The resultrnof this test, thus, showed that the 'Sandwich group' produced significantly higher scoresrnin the listening comprehension test than did the 'After reading group' and the 'Beforernreading group'. On the other hand, these latter groups did not show a significantrndifference in their performances.