«Item type text; Electronic Dissertation Authors Goertler, Senta Publisher The University of Arizona. Rights Copyright © is held by the author. ...»
Although one might assume that the teacher would interact similarly in the two classes, due to the different levels of support, further investigation was necessary and a second analysis was done. The data were thus analyzed using a one factor betweensubjects ANOVA, with class as the factor and NSC, SSC and ESC as the levels. The dependent variable is the average words per minute of teacher output seen by the individual student. The main effect of class is significant (F(1, 43)=199.26, p.001) (see also table 4.5.). NSC students on average were exposed to 0.45 words per minute from the teacher, the SSC students 0.62, and the ESC 2.12. The significant difference means that the more support the teacher received the more words the students were exposed to by the teacher. One possible interpretation is that the teacher who is busy with assisting students with technological problems cannot produce as many words as the teacher who has a Lab Assistant to take care of such problems. This, in turn, may mean that the students in classes with less technological support are exposed to less target-like input.
Table 4.5. Test of Between-Subjects Class Effects of Teacher Input (Dependent Variable:
Words per Minute (Input))
Since there was a difference between both levels of support as a variable and between teachers as a variable, it is difficult to say which difference is more salient.
However, the difference may be greater between the teacher than between the levels of support based on the differences in means and also the difference in the F value. Since there was a significant difference between levels of support (see table 4.5.) and between teachers (see table 4.4.), an additional comparison (see table 4.6.) was done for further insight.
Table 4.6 Planned Comparison of Teacher Input between NSC and SSC
The comparison between the NSC and SSC revealed that there is a significant difference between the two classes (F(1,26)=4.84, p..05), both taught by MorningTeacher, suggesting that level of support has an impact on how many words the teacher can produce per minute due to the additional technological support duties. The SSC received significantly more input from the teacher than NSC (see also table 4.6).
in teacher output. While there was also a significant difference between teacher output for students taught by MorningTeacher and students taught by EveningTeacher, and since there was also a significant difference between groups, it cannot be determined whether the difference between the teachers is truly significant or simply due to varying levels of support. Looking at Illustration 4.1 and remembering that the NSC and SSC were taught by the same teacher, one can argue that the factor teacher had an important influence on the words per minute produced by the teacher. However, multiple explanations for these differences are possible.
Illustration 4.1. Teacher Output Comparison 2.50
Besides the explanation that support may be a differentiating factor between classes, one other possible explanation may be that MorningTeacher communicated more with the SSC because she had a better rapport with these students according to her own comments (during informal conversation which was not recorded). The difference between the combined NSC/SSC and the ESC in terms of input was clearly due to a teacher difference. Furthermore, the use of the “to all” function by EveningTeacher is a possible explanation for the difference between teachers. In conclusion, while a significant difference in teacher output according to teacher and group could be established, multiple explanations for such differences are possible. However, it can be concluded that in all cases but one, the teachers in this study produced over the course of the semester fewer words per minute than the students, suggesting that (a) students receive more input from peers than the teacher, and (b) that teacher dominance decreases in a context where chat is used as compared to what one might expect from a typical context (see for example Beauvois, 1998). It must be made clear, however, that “teacher dominance” was not investigated in a classroom context not using chat.
Besides understanding how much input the students received from the teacher, it is even more important to understand the nature of that input. For the qualitative comparison of the two teachers, the transcripts of the six case study subjects were analyzed establishing (a) teacher target language use, (b) teacher error rate, (c) teacher feedback moves, and (d) teacher moves in general.
The third semester German textbook’s approach strongly encourages exclusive use of the target language during teaching. As has been mentioned before, MorningTeacher did not introduce the activities, and handed out pieces of paper for group assignment during the observed lessons. When students called her over for explanations of the activities or to solve technological problems, she did so mostly in English. EveningTeacher let the Lab Assistant assign groups, which due to his lack of German knowledge, meant that group assignment was done in English. Furthermore, EveningTeacher often introduced the activities using English to explain unclear words and problematic grammar points. For the use of the target language in the virtual space, the total number of words excluding names was counted, and the number of German words excluding names was subtracted, to establish the number of English words (also excluding names) used. Then, the percentage of target language use was established (see table 4.7). In all of MorningTeacher’s case study subjects’ transcripts, her target language use was 100%. However, in other transcripts, she was occasionally observed using English, therefore it cannot be concluded that MorningTeacher used 100% German with all students at all times. EveningTeacher used an average of 93.65% German in the transcripts of the two case study subjects investigated. While this target-language use is lower than that of MorningTeacher, it is still high. Since there is no normal distribution among the six case study subjects, a non-parametric procedure had to be used for statistical analysis. The small number of subjects also lends itself to non-parametric procedures, therefore the chi squared analysis is used here instead of the ANOVAs previously used with the higher number of subjects and normal distributions. Using the chi squared statistical analysis, there was a significant difference between the teachers as illustrated in table 4.8.
Table 4.7 Teacher Target Language Use
The second measure of quality of teacher language was the percentage of errors (as illustrated in Table 4.9). Again for this comparison, only the transcripts from the case study subjects were analyzed, and generalizations to the teacher’s pattern with all students in the class cannot be drawn. In the transcripts analyzed, MorningTeacher’s average error rate was 1.13% (5 errors total), whereas EveningTeacher’s average error rate was 6.19% (48 errors). Since EveningTeacher liked using the “to all” function, most errors were probably counted twice, once in Amanda’s and once in Victoria’s transcripts.
Table 4.10 Difference in Teacher Error Rate MorningTeacher EveningTeacher Total Words Errors Total Degrees of freedom: 1 - Chi-square = 16.
10 - p 0.001.
For the same reason as discussed in the previous section, a chi squared analysis was used to identify differences between teachers in regards to teacher’s error rate. There was a statistically significant difference between EveningTeacher’s and MorningTeacher’s error rate according to a chi squared analysis as illustrated in table 4.10.
The error rate only paints a partial picture about the non-target like language produced by the teachers. Hence, the nature of the errors should be described here.
MorningTeacher’s errors, in the transcripts analyzed, appeared to be mistakes, rather than systematic errors. For example, one activity asked students to ask and answer questions about historic events in the passive voice. All examples except for one used past tense passive voice. In SSCEmily’s transcript MorningTeacher used past tense passive to ask about the upcoming soccer world cup. The asterisk indicates the errors in the examples.
This was also the only incident found by MorningTeacher of an incorrect correction.
SSCEmily: Was wird 2006 gemacht werden?
MorningTeacher: Ja, was *wurde 2006 gemacht? (yes, what was done in 2006) SSCEmily: *In 2006 wird die Fußball-WM in Deutschland ausgetragen werden.
NSCDanielle: 2006 wurde die Fussball-WM in Deutchland *austragen. ….
MorningTeacher: Jawohl, im Jahre 2006 wird die FWM in Deutschland ausgetragen.
This incorrect use can be argued to be a mistake rather than a systematic error, because as you can see above, MorningTeacher used the tense correctly in other transcripts. EveningTeacher, however, made several systematic errors. One of those systematic errors was the repeated use of accusative pronouns instead of nominative pronouns.
EveningTeacher: Was musstet *euch *für Hausaufgaben machen? Hattet *euch viele *Hausaudgaben? Waren die *Lehreren böse oder nett?, interessant oder langweilig? (What did you (accusative) have to do for homework? Did you (accusative) have a lot of homework (spelling error)? Were the teachers (wrong plural ending) mean or nice?, interesting or boring?) In conclusion, both the quantity and the quality of errors differ between the two teachers. EveningTeacher made significantly more mistakes in the transcripts analyzed.
In addition, those mistakes were systematic errors, rather than unsystematic mistakes that even native-speakers make.
The next qualitative and quantitative comparison between the participation of the two teachers was the number and kind of feedback moves used by the teachers in response to errors made by the case study subjects or any other subject who was chatting with the case study subject. First, I will provide examples of the types of feedback used by the different teachers and the number of occurrences (see also table 4.11.). Next, I will discuss the rate of corrective feedback to the case study subject’s own and observed errors (see table 4.12). Finally, I will discuss quantitative (see table 4.13.) and qualitative differences in teacher feedback between the two teachers.
The MorningTeacher mostly used repetitions with correction (13) and models (9) as corrective feedback. She also used translation requests, explicit error correction, marked partial repetition, marked models, and clarification requests. The EveningTeacher used a variety of different feedback forms, with all forms only once: repetition with correction, marked partial repetition, clarification request, translation, marked repetition with correction, and partial repetition with correction.
Table 4.11 Teacher Feedback Styles
illustrates the percentage of errors that received feedback in the case study subjects’ transcripts. The errors were divided into own errors and observed errors.
Corrective feedback by the teacher in relation to the errors made was generally low, not to exceed 3%. The only difference were the transcripts of SSCEmily who received teacher feedback on 10.42% of her errors, and observed feedback (feedback given to a peer of the case study subject in the case study subject’s transcript) to the observed errors at a rate of 5.04%. Overall, MorningTeacher provided feedback to 26 of the 817 errors made in the case study subjects’ transcripts (3.18%). EveningTeacher provided corrective feedback to 6 of the 800 errors made in the case study subjects’ errors (0.75%).
Table 4.12 Corrections Made by Teacher
Again, due to the low number of subjects, a chi squared analysis, i.e., a nonparametric procedure, was used to analyze differences between the amount of feedback received by the students from the teacher. In a chi squared analysis (see table 4.13), it was found that the difference between the amount of feedback given by the two teachers in the transcripts analyzed was significant, with the MorningTeacher providing more feedback than the EveningTeacher. Since SSCEmily’s transcripts set showed more feedback by the teacher than the others, the same statistics were also run without her transcript set and the difference between the teachers remained significant ( p 0.025).
Table 4.13 Differences in Errors Receiving Teacher Feedback MorningTeacher EveningTeacher Total Errors Corrections Total Degrees of freedom: 1; Chi-square = 11.
86; p 0.001.
In conclusion, it was found that the two teachers employed different feedback styles in the transcripts of the case study subjects. MorningTeacher used more feedback and with a higher frequency rate for repetitions with correction and models.
It could be guessed that given the fact that EveningTeacher used significantly more words per minute, yet used significantly less corrective feedback, that in general teacher moves differ between the two teachers. In the following section, the teacher moves will be described and examples will be given (see also table 4.14.).
As can be seen in the table, MorningTeacher utilized the following teacher moves in the case study subjects’ transcripts: corrective feedback moves, modeling of activity and language, language policing (i.e., she told students to use the target language when they were code-switching), bringing students back on task, praise, conversing, topic policing, and providing words. EveningTeacher used some of the same and some different teacher moves in the case study subjects’ transcripts: error correction, modeling of language and activity, conversing, providing words, procedural help, sharing, and expanding the topic. In contrast to MorningTeacher, EveningTeacher provided procedural help in the chat transcripts rather than just in the physical environment.
MorningTeacher kept the conversations more limited to demands of the task by commenting on the language, the topics of discussion, and off-task behavior.