The objectives of this research are as follows: (1) To identify whether there is a significant difference in vocabulary achievement between the students taught using Total Physical Response method and those taught using Grammar Translation Method or not. (2) To identify which group has a better achievement, the group taught using Total Physical Response or the group taught using Grammar Translation Method. (3) To identify which group has a better improvement, the group taught using Total Physical Response or the group taught using Grammar Translation Method.

This research used experimental method. The research was conducted at SD Negeri Karang, Klaten from October to Nevember 2014, in the academic year 2014/2015. The population of this research is the third grade students of SD Negeri Karang. The total number of population is 60 students. All of the population is taken as the sample of this research. The experimental group is the students from class III A which consists of 30 students, while the control group is the students from class III B which also consists of 30 students. The writer used t-test, normality test, and homogeneity test to check whether the two groups have the same vocabulary mastery, homogeneity and normal distribution or not. After analyzing the pre-test data, the writer finds that both groups are homogeneous and in normal distribution. Moreover, based on t-test of pre-test data analysis, the writer finds that to = 0.56 is less than tt(58,0.05) = 2.002 or to? tt. It means the Null hypothesis (Ho) is accepted and there is no significant difference in vocabulary mastery between the experimental group and control group.

The research design used is Quasi-experimental Design with Nonequivalent Control Group Design. In collecting data, the writer used tests as the instrument. Then, the data are analyzed using t-test formula. The result of t-test computation shows that to is 3.85 while the tt for the degree of freedom of 58 and at the level of significance of 0.05 is 2.002 or to ? tt. Therefore, the Null hypothesis (Ho) is rejected, while the alternative hypothesis (Ha) is accepted. Besides, the writer finds that the mean of the scores of the experimental group in post-test is 77.2, while the mean of the scores of the control group in post-test is 67.46. The difference between the two means is 9.74. . Statistically, µ1 > µ2, which means that Ho is rejected and Ha is accepted. In pre-test the mean score of experimental group is 53.27, while in post-test the mean score is 77.2. The difference between the experimental group mean score of pre-test and post-test is 23.93. In control group, the mean score in pre-test is 52.8 and the mean score in post-test is 67.46. The difference between them is 14.66. Statistically, it can be formulated as µ1 > µ2, which means that Ho is rejected and Ha is accepted.

The result of the research are as follows: (1) There is a significant difference in vocabulary achievement between the students taught using Total Physical Response method and those taught using Grammar Translation Method. (2) The group taught using Total Physical Response has a better achievement in vocabulary than those taught using Grammar Translation Method. (3) The group taught using Total Physical Response has a better improvement in vocabulary than those taught using Grammar Translation Method.

Keywords: total physical response, vocabulary