The function is to edit the term, which means to correct the speech error of the speakers. Thus, the speaker is aware that s/he has made a speech error so s/he wants to correct it Baalen (2001, par.8).
      Due to no further explanation on how to classify these functions of fillers exactly, I had to consider some factors that helped me a lot in deciding these functions, that is the speaker's smoothness, length of time needed, intonation, certainty, and intention or the whole meaning of speaking. By noticing and counting these factors, I could classify easier these functions. It cannot be avoided that one filler can have more than one function due to some intention from the speaker, but only the most obvious function was used in this research.
      For an easier way of classifying the data, I made a table, consisting of four columns, that is number of turn, utterances, type of fillers, and functions of fillers. Then, I put the utterances into the table based on the number of turn. After that, beside the column of Utterances, I wrote the type and the function of fillers in the provided columns, based on the number of fillers marked or referred. Here is the example of the table.
3.3.3 Â Â Analyzing the Data
      There were some steps that I followed in order to do the analysis:
a. Â Regarding the types of fillers used by the first and the seventh semester students:
After classifying those fillers into the comparison table of Types of Fillers between the first and the seventh semester students, I counted the number and the percentage of each type of fillers, that is silent pauses and filled pauses (unlexicalized filled pauses and lexicalized filled pauses). After that, I counted the total number and the percentage of each type of fillers. For an easier way of analyzing the data, I put the total number of each type of fillers in the Table 3.2 below. Then, I analyzed the number of each type of fillers and started to interpret them.
Regarding the functions of fillers used by the first and the seventh semester students:
After classifying those functions of fillers into the comparison table of Functions of Fillers between the first and the seventh semester students, I counted the number and the percentage of each function of fillers, that is filling the pause, hesitating, interrupting, holding a conversation turn, empathizing, mitigating, and editing the term. After that, I counted the total number and the percentage of each function of fillers. For an easier way of analyzing the data, I put the total number of each function of fillers in Table 3.3. Then, I analyzed the number of each function of fillers and started to interpret them.
FINDINGS AND DISCUSSION
This chapter is divided into several parts, that is the findings and the interpretations on the findings.
