6.3.2 Possible Factors Behind Individual Variation...
The corpus used in the present study has been purposefully structured to enable the investigation of factors contributing to the individual differences in the use of moves:the influence of the former communities of thesis writers;the research traditions forming the academic environment for the theses;and the year in which the theses were submitted.
6.3.2.1 Influence of Community Conventions
As mentioned above,the texts in the New Zealand and America sub-corpora show more variation in the use of moves than the texts in the China sub-corpus(see Figure 11).It is suggested that this finding might well be related to the nature of the different communities,that is,whether they can be classified as mixed or uniform.The writers of the 30 texts in the China sub-corpus all completed their undergraduate and graduate studies in the same academic discourse community and from this perspective can be categorised as a uniform community.On the other hand,a significant proportion of the writers in the New Zealand and America sub-corpora undertook their undergraduate studies in local communities different from their current community and hence were likely to have been exposed to somewhat divergent styles of academic writing.This is what I mean by a mixed community.Thus,although the newcomers acquire the conventions of the new local community,I claim that the influence of the former community could well continue to have an effect on their writing.
Figure 11 The distribution of moves in individual texts(China,New Zealand &America)
Figure 11 (continue)The distribution of moves in individual texts(China,New Zealand &America)
Table 18 shows the presence of each move and metatexts in the five groups.The thesis writers are labelled as Chinese(CH)or non-Chinese(OTH).The percentage refers to the proportion of writers using the move.Thus,for example,for CHCH,87% of the writers in the Chinese in China group use Move 1 Familiarising readers with the field.In general,the percentages are high,with all moves are used by a considerable number of students in each group(over 85%).Move 1 is particular popular among students in New Zealand,whether their background is Chinese or not.Move 2a Preparing practical rationales is the least used move in all the five groups,appearing in 70% of the theses written by Chinese students in China and in America and reaching only 50%-60% in the other three groups.Move 2b Preparing empirical rationales is essentially an obligatory move that appears in nearly all theses(89/90)[2].The occurrence of Move 3 Establishing a theoretical framework is the most diverse,with 100% or 95% of theses from the China and New Zealand community exhibiting this move,compared with only 55% of theses in the OTHUS group(i.e.Non-Chinese in America).Move 4 Presenting the study is the second most widely used move found in 89 out of 90 theses.Metatexts,which serve the distinctive functions of stating intentions and providing an overview or summary of the content,also has a wide distribution among theses,occurring above 90% in all five groups.
Table 18 The number of theses containing each move and metatexts in the five groups
The number of occurrences of each move is,in fact,not the best measure when comparing the use of moves among the different groups.Close examination of the theses shows that the real difference in the use of moves lies in the proportion of the opening section assigned to fulfil each function.For example,a writer may use 3% or 30% of the opening section to familiarise readers with the field.Therefore,to reveal the influence of ethnicity and local community on genre practices,the inter-group comparisons are made based on the standardised length of the move.(See Footnote 1)
Table 19 shows the median of the standardised length of each move in the five groups.What is particularly interesting is that the median values for the Chinese overseas groups is typically between that of the Chinese students writing in China and that of other students in the same community.For example,the standardised length of moves of Chinese students writing in English in New Zealand(CHNZ)is between that of CHCH and non-Chinese students writing in English in New Zealand(OTHNZ).This indicates that the writing of Chinese overseas students is influenced both by the genre conventions in the new host community(NZ or US)and the former community(China).
Table 19 The median of standardised length of each move in the five groups
To follow up,I examine the performance of Chinese students in NZ and the US compared with the Chinese students in China and with the non-Chinese writers in NZ and the US.A Mann-Whitney U test reveals that the main differences between the writing of Chinese students in NZ and those in China is in Move 2b(U=286.00,z = - 2.43,p＜0.05,r = - 0.31)and Move 3(U=259.50,z = - 2.82,p＜0.01,r = - 0.36).As shown in Figure 12,the standardised length of Move 2b in texts written by CHNZ(Mdn=0.26)is also between that of CHCH(Mdn=0.21)and OTHNZ(Mdn=0.31),and the standardised length of Move 3 in texts written by CHNZ(Mdn=0.43)lies between that of CHCH(Mdn=0.47)and OTHNZ(Mdn=0.32).Again,this is taken as an indication that the writing of Chinese students in New Zealand is under the influence of both their former and current academic communities.The boxplots indicate the range of individual variation and it is clear that there are other factors involved,possibly relating to the content or simply resulting from individual preferences.I return to the issue of other factors below.
Figure 12 A comparison of CHCH,CHNZ and OTHNZ
This dual influence is also true for Chinese students in America whose writing can be seen to be situated between that of CHCH and OTHUS.A Mann-Whitney U test reveals that the differences between the China and America sub-corpora lie in Move 2b Preparing empirical rationales(U=212.00,z=-3.52,p＜0.001,r=-0.45),Move 3 Establishing a theoretical framework(U=187.50,z=-3.89L,p＜0.001,r=-0.50),and Move 4 Presenting the study(U=314.50,z=-2.02,p＜0.05,r=-0.26).As shown in Figure 13,the standardised length of Move 2b in texts written by CHUS(Mdn=0.32)is between CHCH(Mdn=0.21)and OTHUS(Mdn=0.56),and the standardised length of Move 3 Establishing a theoretical framework in texts written by CHUS(Mdn=0.33)is also between CHCH(Mdn=0.47)and OTHUS(Mdn=0.15).For Move 4 Presenting the study,though the median of the standardised length of the texts written by CHUS(Mdn=0.03)equals that of CHCH(Mdn=0.03),the range of CHUS(0.00,0.12)resembles that of OTHUS(0.01,0.12),suggesting a similar pattern of variation.
These results consistently show an interesting pattern.The genre practices of Chinese students in overseas communities resemble those of Chinese students in China.It is reasonable to suggest that this is due to the continuing influence of the genre conventions of the former community.One consequence is the diversification of move patterns seen in the New Zealand and America sub-corpora(Figure 11),arising out of differences in the writers’undergraduate community.
Figure 13 A comparison of CHCH,CHUS and OTHUS
I have seen that the overseas Chinese students are influenced by their former and current communities.In the remainder of this section,I examine the influence from current and former communities separately.
6.3.2.1 .1 Influence from the current community
The first comparison is among the Chinese students located in three local communities(i.e.China,NZ and US).All these students share the same former community(i.e.China)but have different current communities.Therefore,we can assume that any difference identified is likely to be associated with the influence of the current community.A MANOVA test was performed.Based on Pillai’s Trace[3],there is a significant effect of the difference in current communities on the standardised length of moves in the writing of the students,V=0.58,F(12,86)=2.98,p＜0.01.Significant differences were revealed between CHCH and CHNZ(V=0.45,F(6,33)=4.53,p＜0.01)as well as CHCH and CHUS(V=0.30,F(6,33)=2.43,p＜0.05).A post-hoc test of Fisher’s least significant difference(LSD)reveals significant differences,mainly between CHCH and the two overseas groups(see Table 20).
Table 20 Significant differences revealed by the post-hoc test(LSD)
6.3.2.1 .2 Influence from the former community
The second comparison is among the groups in the same current community but a different former community(i.e.CHNZ vs.OTHNZ;CHUS vs.OTHUS).Since the comparison is between students belong to the same current community,any difference identified is likely be related to the influence of the former community.As mentioned above,since there are few international students studying Applied Linguistics in China,this comparison is only conducted for students located in the NZ and US community.Based on Pillai’s Trace,there was no significant effect of former community on the standardised length of moves in the writing of students in the same local community(for CHNZ vs.OTHNZ,V=0.18,F(5,24)=1.06,p=0.40;for CHUS vs.OTHUS,V=0.35,F(6,23)=2.14,p=0.08),indicating the influence of the genre conventions of the former community,while present,is not strong enough to show a statistically significant difference.
Combining the results of the two comparisons,we can see that students of the same current community form a more cohesive group in terms of genre practice than students of the same former community.Obviously,the genre practices of overseas Chinese students incline to the conventions of their current community rather than the former.Although individual differences were observed between students in the same local community,they still form a relatively cohesive group.Figure 14 shows the result of a discriminant analysis which plotted the 90 opening sections using the standardised length of the four moves and the metatext of each text.As shown in the figure,although the plots of the three groups are dispersed,each group clusters around a distinct position,indicated by the group centroid,which is shown as a square.
Figure 14 Combined groups plot
Two main discriminant functions contributed to the variation between groups.The first accounted for 66.2% of the variance,canonical R2=0.29,and the second covered 33.8% of the variance,canonical R2=0.18.In combination these discriminant functions differentiated the local community groups,Λ=0.58,X2(12)=45.73,p＜0.001.After removing the first function,the second function alone also significantly discriminates the three groups,Λ=0.83,X2(5)=16.28,p＜0.01.The correlations between outcomes and the discriminant functions revealed that Move 2b Preparing empirical rationales,Move 3 Establishing a theoretical framework and Move 4 Presenting the study loaded more highly on the first function,which distinguished the China group from the America group(Move 2b:r=0.78,Move 3:r = -0.77,Move 4:r=0.43);Metatext loaded fairly highly on the second function,which differentiated the New Zealand group from the other two groups(r = -0.63,for the first function:r=0.10).
6.3.2.2 Research Traditions
In this subsection,I explore the question of whether the differences in move structures are due to differences in the research paradigms used in the three academic communities.Lim(2006)suggests that the type of research tradition(qualitative,quantitative and mixed method)may be associated with thes variation in move structure of RA method sections,although in that study statistical tests on studies following the different research traditions did not produce any significant differences.The present study also pursues the influence of different research tradition on the use of moves.There are different ways in which the classification of research traditions can be approached.Since the present study examines the Introduction and Literature Review chapters,which are mainly concerned with the theoretical foundations of a study rather than methodological issues,such as the types of data or analysis method,the type of research is used as the basis for the classification.In particular,the definition of the three research approaches provided by Ellis(2012)is used to categorise the theses as“confirmatory”,“descriptive”or“hybrid”.It was found that some students who did confirmatory research,rather than formulating hypotheses based on theories,made predictions about relationships between variables according to their own learning or teaching experience(e.g.Guinn-Collins,2011).Since these researchers had a clear idea about the variables that they were going to examine and a prediction about the relationship between variables prior to conducting the study,their theses were also labelled as confirmatory research.The theses were sampled randomly and thus are unbalanced in terms of research type.As shown in Table 5,41 theses out of 90 are confirmatory research,39 are descriptive research and 10 are hybrid[4].Two thirds of theses in the China sub-corpus are confirmatory research,which is the most popular research type in the Chinese community,while descriptive research is the most popular research type in the New Zealand and America communities with half of the theses in each of the two communities falling under this category.Thus,the researcher does find some differences in research paradigms in the difference communities.
Because of the unbalanced distribution of texts among the three research traditions in each local community and the sparseness of data,a statistical test on the three research traditions was conducted without distinguishing the community in which the texts were written.A Kruskal-Wallis test shows no significant difference in any of the five moves or metatext.In other words,there is no substantial difference in the move structures between the opening section of theses reporting confirmatory and descriptive research,or confirmatory and hybrid research,or descriptive and hybrid research.Thus we can conclude that the research tradition is thus not a major contributor to the individual variation shown in Table 21.
Table 21 The distribution of texts over research traditions
6.3.2.3 Time
The three discourse communities are,of course,not isolated and it is interesting to consider changes in move structure over time.In the China sub-corpus,a Kruskal-Wallis test reveals a significant difference in the standardised length of Move 3 Establishing a theoretical framework over the 3 years in the China Applied Linguistics community(X2=9.30,df=2,p＜0.05).As shown in Figure 15,on average,the standardised length of Move 3 is the longest in the theses submitted in 2009(Mdn=0.58,Interquartile range:0.49,0.68)and reduces considerably in 2011(Mdn=0.39,Interquartile:0.29,0.53)and 2013(Mdn=0.43,Interquartile range:0.40,0.51).Although the median of standardised length in 2011 is shorter than that in 2013,the range of the length of Move 3 in the theses of 2013 is much smaller than that of 2011,suggesting that the practice of thesis writers of 2013 is more congruent,apart from the outlier.Because of the small sample size,the finding can be only seen as preliminary,but it leads to the interesting speculation that the thesis writers in the China community are reducing the proportion of theoretical foundations in their thesis,and writing more like their peers in the English-medium communities.
Figure 15 The boxplot of the distribution of Move 3 over time in the China Applied Linguistics community
In the America sub-corpus,a Kruskal-Wallis test reveals significant differences in Move 1 Familiarising readers with the field(X2=6.39,df=2,p＜0.05)and Move 2a Preparing practical rationales(X2=6.92,df=2,p＜0.05)over the 3 years in the America sub-corpora.As shown in Figure 16,the standardised length of Move 1 in theses submitted in 2013(Mdn=0.15)is longer than that in 2009(Mdn=0.04)and 2011(Mdn=0.04).The MA students who submitted in 2011(Mdn=0.00)give less practical rationale than those in 2009(Mdn=0.06)and 2013(Mdn=0.04).Again,the sample size is quite small and the significant differences may be influenced by sampling errors.However,it may indicate the tendency of American students to pay more attention to introducing the basics to their expected audience and putting in more effort to advertise their research.As suggested by Ayers(2008),this may be related to the change of target audience from experts and peers in their own community to general readers such as researchers from other discipline,language learners,or language teaching practitioners.These results are suggestive and interesting,but,clearly,further research using a larger sample and temporal range is needed to investigate these types of changes over time.
Figure 16 The boxplot of the distribution of Move 1 &2a over time in the America Applied Linguistics community
The results of present study reveal the extent of individual variation within each local disciplinary community and explore some possible factors to account for that variation.In the following section,the consequences of these findings for research on genre analysis are discussed.