2.2　Results and Discussions froMListening Tests

A pretest and a posttest are carried out in order to find out whether there is anyimprovement of students’performance in news listening comprehension between the controlled group and the experimental group after discourse markers’training.Both the tests are carried out under the same environment with the same time and the same listening equipment.The results of both the pretest and the posttest are collected and marked by the same teacher and the samemarking standard.And the results of the pretest and the posttest were input into the computer.They were analyzed by a statistic analysis system(SPSS16.0)

2.2.1　Results and Analysis of the Topic Familiarity Survey

Researches，including Dunkel，1991，Nam，1997，and Jung，2003 has proved that background knowledge does play an important role in the performance of listening process.An experimental result froMYe(2001:23)shows that about 95％of the students believe that topic familiarity influences their discourse listening comprehension.Consequently，whether the participants are familiar with the background knowledge on the content of experiment or not contributes a lot to the validity of the experiment.Therefore，in the designing stage of current study，the control of background knowledge has been taken into consideration.In the third chapter the author has mentioned about a survey concerning the participants’background knowledge of the listening materials—Before the analysis of the data froMtwo tests，now it is necessary to present the results froMthe survey.According to the results froMthe survey，among the 104 participants froMboth the two groups，there are four subjects who are familiar with at least one iteMof the listening materials，for one froMthe experimental group and three froMcontrolled group.All of their test scores will not be included in the collected data，which ensures the validity of the analyzing result in the following section.

2.2.2　Results and Analysis of the Pretest

Comparative analysis of the pretest results betweencontrolled group and raining group

The pretest taken by the students after questionnaire mainly intends to investigate the students’news listening proficiency of the two groups and make an analysis of the outcome of the two groups，froMwhich consultancy can be provided for the discourse markers training.Table 4.6 shows the results of independent samples t-test between the controlled group and experimental group.

According to table 4.5，we can see that the mean and standard deviation of experimental gsroup are 62.2353 and 10.8047，and the controlled group 62.6122 and 11.4906.The difference of mean is not prominent.Even the mean of the controlled group is slightly higher than the experimental group.The figures shown in table 4.6 prove that there is no significant difference between the two groups.Evidences can be seen in above table:F=.177，sig.=.674，so we should analyze the results in terms of Equal variances assumed in t－test Equality of Means.FroMfigures as following:t=－.169，sig.(2－tailed)=.866.We can clearly say that the difference between experimental group and controlled group is not significant before training，which means that the two groups were approximately at the same level in news listening comprehension before the training.In another word，the students in the two classes are homogeneous in their English listening scores;the students’English listening performances are quite similar before the discourse markers instruction.

What’s more，froMthe mean scores of the two groupswe can also get that the subjects’overall level of news listening comprehension is not high.Possible reasons could account for the above results.First of all，many students have been taught some other listening skills，such as grammar，vocabulary，sentence analysis and so on in the previous English listening classes.Besides，many students take these skills as the direct way to improve their news listening achievements in a short period;hence they have practiced some of the skills consciously to improve their listening proficiency.Furthermore，many students are accustomed to the short conversations and some other types of discourse，when it comes to the type of news discourse，few of theMcan find an effective way to deal with such kind of discourse due to the particular features of news discourse.Finally，in the listening class few English teachers can lay emphasis on discourse analysis，not to mention discourse marker which is only a small section of it，hence，it is no wonder few students can aware of discourse markers，especially their functions and application in news listening discourse.Therefore，it’s possible to conduct the discourse markers training to further explore the effectiveness of discourse markers training.

2.2.3　Results and Analysis of the Posttest

2.2.3.1　Comparative Analysis of the Posttest Results between the Controlled Group and the Experimental Group

The posttest，which is conducted right after the training，intends to examine the effect of the discourse markers’training on news listening comprehension according to the data collected froMit.

According to table 4.7，we can get results as follows:Both the two groups have some improvement，but the experimental group rises by a larger degree.The mean score of the experimental group rises froM62.2353 to 72.1176，while the scores of the controlled group ranges froM62.6122 to 63.3061.The mean of the experimental group is 72.1176 and the mean of the controlled group is 63.3061.The standard deviations of the two groups are respectively 12.55013 and 11.54376.What can be seen froMthese figures is the mean of the experimental group is higher than the controlled group apparently，which is different froMthe results in pretest.What can be seen froMtable 4.8 is as following:F=.712，sig=.401 which shows that the results of t－test for Equality of Means should be analyzed in terms of Equal variances assumed.The figures that t=sig.(2－tailed)=.000 can be seen in the above table.Sowe can say that the difference between the two groups is significant and the comprehension of news listening of the experimental group is higher than the controlled group.Considering the obvious difference between the two groups，certainly we should attribute the reason to the discourse markers training of the experimental group，because of which can the experimental group make greater progress than the controlled group.Therefore，there exists a close relationship between discourse markers and news listening comprehension and based on this research question 2 has be addressed.

2.2.3.2　Comparative Analysis of the Controlled Group before and after the Training

The figures in table 4.9 and table 4.10 show that before the training，the mean score of controlled group is 62.21;after threemonths，themean score of controlled group is 63.31.The improvement of the mean scores is not so prominent.In order to find out whether there are significant changes of the controlled group between before and after the training，then the author made a Paired Samples Test of controlled group getting such figures:df=48 t=－1.916 P=.061＞0.05.The results froMthe two tables above indicate that after three months，the level of the overall news listening comprehension of the students in the controlled group are basically consistentwith the performance in pretest.So there is no significant difference in the controlled group's capability of news listening comprehension between before and after the threemonths.

2.2.3.3　Comparative Analysis of the Experimental Group before and after the Training

The performance of experimental group between pretest and post－test can be seen froMthe results shown in the table4.11 and table 4.12.What we can see froMthe two tables above are as follows:Before the training，themean score of the experimental group is62.24;after the training，themean score is72.12.It is obvious that themean score of experimental group in pretest is higher than that of experimental group in posttest.The subjects in the experimental group get obvious improvement through the training of discourse markers.Then the authormakes a Paired Samples Test of experimental group posttest and get such figures df=50 t=－24.850 P=.000＜0.05.These figures indicate that the differences between the pretest and the posttest are significant in experimental group.FroMthe great improvement of themean score between the pretest and posttest in the experimental group，we can get that the discourse marker instruction in news listening course can help improve and facilitate students’listening comprehension and the instruction of discourse to a large extent affects English news listening comprehension.Accordingly，the conclusion above gives a clear answer to research question 3.

2.2.3.4　Comparison Analysis of High－proficiency and Lowproficiency Students in Experimental Group before and after the Training

Owing to differences lying innews listening comprehension，the author divided the subjects in to two groups:High－proficiency group(including the first ten subjectswho have higher scores)and low－proficiency group(including the last ten subjects who have lower scores)according to their performance in pretest，and to find outwhether there is any difference between the two groups’performance after the discourse markers’training.The author will give some general analysis according to the two groups’mean scores of the two tests.

As table 4.13 indicated，before the training themean score of the highproficiency group is 75.8000，while the mean score of the high－proficiency group reaches 87.8000 after the training，so we can easily calculate that the mean score of the high－proficiency group subjects increases by 12 points.Table 4.14 is concerned with the results of the mean score of the lowproficiency group，before the training the mean score of this group is 45.6000 and it achieves to 53.6000 after the training，thus we can alsoclearly present that themean score of the low－proficiency subjects improves by 8 points after the training.Accordingly，we can easily demonstrate that the high－proficiency group students tend to make greater progress than the low－proficiency ones，because there indeed exists difference in the performance of the two tests between the two groups.In fact，the result presented echoed with the findings of Wen Qiufang’s research.According to Wen(1996)，the main difference between successful and unsuccessful learners lies in whether they are good at managing their studies.Considering this point，it is significant and necessary for teachers to pay more attention to the low－proficiency students and give theMmore focus during the instruction of discourse markers.

However，the difference is not so significant，and the most important point is that both of the high－proficiency and the low－proficiency students have made dramatic progress after discourse markers training.Thus，all the students welcomed the discourse markers training.Meanwhile，such fact offers the foundation for our strategy training research.

2.2.4　Discussions of the Results froMTests

Through analysis of the pretest，we can get that before the treatment of discourse markers training，the experimental group and controlled group shows little difference in the news listening comprehension with the samemastery level of discourses markers.However，after receiving three months＇training of discourse markers，the performance of experimental group in news listening comprehension gets improvement.Since the only difference between controlled group and experimental group is that the experimental group received discourse markers’training，it can be safely concluded that awareness of discourse markers have a close relationship with news listening comprehension and discourse markers’training is effective in the process of news listening comprehension.Also，through analysis of the results froMthe experimental group between its pretest and posttest，we can see the effect of discourse markers on students＇news listening comprehension is significant.What’smore，in order to figure out whether there exists difference between the high proficiency and low proficiency students in their performance after the training of discourse markers，the author has also made comparison of the posttest’s results between high－proficiency subjects and low－proficiency subjects.According to the analysis，it has proved that the high－proficiency subjects can make larger progress than the low－proficiency ones after the same training.Thus，there is no doubt to get such conclusion that there exists a close relationship between discourse markers and news listening comprehension，the news listening comprehension will rise to a large extent through getting systematic training of discourse markers and high－proficiency subjects are more effective in applying discourse markers in news listening comprehension than the low－proficiency ones.

In fact，the above conclusion we have drawn is just in accordance with relevance theory:“Discourse markers are important tools for language communication and the indicators of discourse intentions.In addition，discourse markers can make the listener obtain the expected contextual assumptions and effects，thus providing the listeners with obvious discourse marks”.To some extent，there are some contemporary views about discourse markers.Discourse markers are useful and common in daily life，not just in short conversations but also in news listening discourse.However discourse markers are not seen as an isolated phenomenon suspended in a vacuuMbut as an integral part of discourse.Therefore，their features and main functions in various kinds of discourse(reading，writing，listening and so on)should be taken into account.Among these，discourse markers’function in news listening discourse is worth studying.Therefore，to acquire the main functions of discourse markers in the process of news listening，learners can not focus solely on the forMof discourse markers or just study the knowledge of discourse markers insolently and mechanically.The correctmethod is to receive the knowledge of their practical application in discourse and to feel how do these words create and function in the process of news listening.