A corpus-based study on noun and verb agreement errors by Turkish learners of English as a foreign language

In this study, Turkish EFL learner`s interlanguage noun and verb agreement errors has been investigated. The study has made used of Error Analysis in analyzing the data for noun and verb agreement errors. The present study has made used of the corpus-based techniques to analyze the use of noun and verb agreement in the Turkish EFL learner`s written performance across A1 – A2, B1 – B2, and C1 – C2 CEFR levels of proficiency available in the related corpora of the Cambridge Learner Corpora. Two different research methods were used in this study, both descriptive and experimental method. Two different groups were involved in this study, experimental and the control group. Experimental group were taught noun and verb using the corpus data-driventechniques; and the control group were taught in a traditional way of teaching noun and verb agreement. The result of the experiment was determined by the test on noun and verb agreement taken by the participants after the experiment. In addition, the overuse, underuse and the errors made in using noun and verb agreement by the Turkish EFL learners were analyzed using Log-Likelihood and Log-Dice in investigating and finding out the statistic evidence upon the possible errors committed by the Turkish EFL leaners. In thıs study Turkısh EFL learner's use of noun and verb agreement in their academic writing has been checked and comparison of the result have been made between the control and the experimental group, however the result has revealed that Turkish EFL learner overuse noun and verb agreement in their writings. In addition corpus-based implantation on noun and verb agreement errors has been done to see if the participants could master the rules or not. the result have shown that the both group have improved, but the experimental group overused noun and verb agreement more than their counterpart control group. Firstly Overall frequencies have been identified and analyzed in a tabular form both CLC and the Corpora compiled by the researcher across A1 – C2 proficiency band of CEF. The data were analyzed using the soft wares called Sketch Engine, AntConc, Log-dice and Log-likelihood in order to compare the result. Secondly, the frequencies of all the categories A1 – C2 of the CLC data as well as the corpus compiled by the researcher during his implementation were analyzed respectively, and each proficiency level frequency is shown respectively.