로그선형 회귀분석을 통한 발화문장의 차이 연구

This paper uses log-linear regression analysis to investigate whether there are any statistically significant interaction effects among the four selected categorical variables in Korean-to-English consecutive interpretation: S=Subject (student or interpreter); O=Overseas experience (less than two years or more than two years in majority English-speaking countries); L=Level of original text (easy or difficult); and T=Type of sentences (simple, complex, compound, or compound-complex). The present study also examines the uttered sentences by applying various readability indexes, such as the Flesch-Kincaid grade level, Gunning Fog, and Dale-Chall formula, to explore the complexity of the produced speeches. Our log-linear analysis demonstrates with statistical significance that the observed ratio of simple to complex sentences differs between students who have lived more than two years in majority English-speaking countries and those who have not. Furthermore, it is shown that such difference based on overseas experience does not exist among professional interpreters.