Abstrak
This study was motivated by the findings of researcher during observations and interviews at the tenth grade of MAN 3 Agam. The researcher found that students' pronunciation was not good and difficult to understand. Furthermore, students' mastery of vocabulary is a little, making it difficult for students to convey their ideas so that it hinders students' fluency in speaking and makes students prefer to use local languages rather than the target language in interacting in class. This study aims to determine the effect of Think Pair Share technique on students' speaking ability. This study uses a type of experimental research, namely Quasi experiment with a nonequivalent control group design. The sample in this study was selected using cluster random sampling technique and the selected class was class XE. 3 consisting of 26 students as the control class and XE.2 consisting of 25 students as the experimental class. The instrument used in this study is a speaking test with the Extensive speaking test type (monologue) which has been validated with content validity and assessed with inter-rater reliability. Data analysis used includes prerequisite tests of normality and homogeneity, paired t-tests and independent paired t-tests to test the hypothesis of this study. Based on the results of data analysis, it was concluded that the Think Pair Share technique has a significant effect on improving students' speaking ability.