The Mixed Methodology
混合的方法
In this chapter the strengths of the parallel mixed method will be dis-cussed and a critical phase, the results triangulation, will be addressed ex-plaining how the researcher has dealt with it. It will be then exposed in detail how ethical issues have been addressed in the best interest of the research participants.
在这一章中的并行混合方法的优势将显示讨论的关键阶段，结果三角，解释研究者如何处理将得到解决。然后将暴露在伦理问题如何得到解决，在研究参与者的最佳利益。
Mixed Methods and Pragmatism
混合方法和实用主义
The philosophical paradigm at the light of which this research should be read is the pragmatist paradigm. This paradigm seems to be the most suit-able paradigm in mixed methods designs where ‘the research word view is composed of several word views’ (Creswell and Clark, 2007, p.5). The coex-istence of antagonist philosophical perspectives on the nature of reality andon how knowledge can be generated are fully legitimate because they become instruments to answer the research questions. This however will not exempt us from discussing the pragmatist perspective and the main philosophical
应读的光，而本研究的哲学范式是实用主义范式。这的范例似乎是最适合能够混合方法设计范式“的研究字视图由几个单词的意见”（克雷斯韦尔和克拉克，2007年，第5页）。COEX的存在的拮抗剂哲学观点对现实的安灯的性质如何产生的知识可以是完全合法的，因为他们成为工具来回答研究问题。然而，这不会免除我们讨论实用主义的角度来看主要的哲学。
Table1.1
paradigms adopted in this research, these in fact will be the general tent sunder which the research process and results can be evaluated.
本研究采用的范式，这些其实可以评估的研究过程和结果将是一般的帐篷破甲。
The association of mixed method designs with the pragmatist paradigmis supported by most researchers (Sounders et al., 2009; Creswell and Clark,2007; Tashakkori and Teddie, 2009). However while Tashakkori and Teddie(2009) see pragmatism as the most suitable paradigm for mixed methodsdesigns others welcome the application of multiple views and accept thatthe dispute between contradictory ideas cannot be reconciled as it reflectsdifferent ways of knowing (Green and Ceracelli, 2003 in Creswell and Clark,2007, p.27).
A third perspective advocates that ‘the philosophical paradigms may dif-fer depending on the type of mixed method design’ (Creswell and Clark,2007). In particular in the case of the triangulation designs, where thedata are merged together into one interpretation, the authors recommendthe pragmatist approach. In experimental or exploratory mixed methods
Table1.2
designs they suggest other paradigms, such as the post-positivist or the con-structivist paradigms, which fit better the design purpose and maintain amore critical position on qualitative and quantitative findings.
At the light of Creswell and Clark (2007) conclusions the most suitableparadigm for our research, characterised by a triangulation design with dataconvergence (Figure 7.1) seems to be the pragmatist one. The pragmatistposition is that the research question informs the nature of reality ontology,how we gain knowledge of what we know epistemology and the role valuesplay in research axiology.
Table 7.1 specify which elements belonging todifferent paradigms have been utilised in the pragmatic approach.
From a strategic point of view the research is based on a grounded theorydesign since the objective is that of generating a theory which explains how aprocess, the student experience with the school of management, takes place.‘Grounded theory is generally considered supported only by an inductive ap-proach’ (Glaser and Strauss, 1967 in Sounders et al., 2009) however morerecently Goulding, (2002, in Sounders et al., 2009) defined it a deductive-inductive design. The researcher in fact starts by collecting and analysingdata and literature on the research object which enables him to formulate aconceptual model, he then collects first hand data on the research object, heanalyses and interpret it and either confirms the original conceptual model,amends it or proposes a new descriptive model which represent better theobject. The research therefore will present a conceptual model discussed to-gether with the relevant supporting literature and the researcher’s hypothesis(Section 3.4).
The qualitative and quantitative conclusions will be comapred and con-trasted in a synthetic table from which the final research conclusion will bederived.
Mixed Methods and Results Triangulation
混合方法和结果的三角网
The mixed method design with triangulation and convergence also calledin Tashakkori and Teddie (2009) parallel mixed method with data conver-gence presents several strengths. Firstly, the fact that the inferences formthe qualitative and quantitative independent stands are compared and con-trasted allows the researcher to cross validate the results and obtain a morecomplete answer to the research question. Since the qualitative and quantita-7.2 Mixed Methods and Results Triangulation 67tive perspectives have the same weight the comparison of the meta inferencesderived from each phase is more balanced and is not biased towards one ofthe perspectives. As Creswell and Clark (2007) state ‘the advantage of thisdesign is that it brings together the differing strengths and nonoverlappingweaknesses of quantitative methods with those of qualitative methods’. Theparallel mixed method design is also very efficient since data are collected atthe same time and during data collection one design may inform the other,ex. the results gained form the first half of the respondents in the ques-tionnaire may suggest to focus more on a particular topic in the qualitativeinterviews.
The mixed method design will be very useful to answer the research ques-tion because it will allow the researcher to use all the tools of qualitativeanalysis (openhanded questions, one to one contact, high discretionality andinsider point of view) to explore and gain an understanding of the studentphysical and emotional experience during their journey into postgraduate ed-ucation. The quantitative phase will allow to further generalise the matchingfindings or to corroborate them by identify the tenets and limitations of thequalitative findings.
The main critical points arise from the selection of the two samples qual-itative and quantitative and from the integration of the different set of data.According to the quantitative method the sample should be a good, possiblystatistical, representation of the nucleus. In the research design the nucleus,the School of Management students, is composed of about 450 students there-fore to obtain a statistically reliable sample with an error of 5% more than207 students should have been surveyed. Obvious practical and data qualityconsiderations instead suggests that the qualitative sample should not be toowide but rather proportionate to the research time horizon, in this research7.2 Mixed Methods and Results Triangulation 683 months. It is fundamental in fact that the researcher can have the time toconduct and analyse carefully the results of the interviews. The statisticalrelevance of the population in qualitative interviews is of secondary impor-tance and purpose samples are very frequent. the most important qualitythat the sample should have is the participants profile relevance to the re-search and their information richness. The problem with the parallel mixedmethod is therefore how the meta-inferences derived form the analysis of thedata collected form two samples can be integrated and how comparison ma-trices can be built to synthesise the research findings (Creswell and Clark,2007).
The researcher has addressed the sampling problem by delimiting thescope of the research. The quantitative sample was based on students self-selection and only 42 students were surveyed (9% of the population) there-fore precluding any statistically based inference on the overall population.Although the comparison on quantitative basis between the two databases islimited because of the discrepancy in the sample sizes qualitative inferencesare legitimate.
Where quantitative comparisons were necessary the researcher has weightedthe qualitative cases so that the proportion between the two samples wasmaintained as suggested by Creswell and Clark (2007). Since the two qual-itative and quantitative designs were based on the same research questionsthe researcher has been able to synthesise the results in a matrix where theoutcome of the two designs can be contrasted and compared.
Ethical Issues Involved in the Research
参与这项研究的伦理问题
There were a number of possible ethical issues linked to the differentstages of this research. Due to the fact that during the interviews and surveysensitive information was collected it was be necessary, before starting tocollect any kind of data, to inform the students briefly on the nature of theresearch subject, where and on how the data would be collected, stored andmanipulated.
In this particular research project sharing the research objectives and theresearch questions with the students was beneficial, this in fact raised genuineinterest and commitment while sharing the research purpose did little harmto the quality of the data collected. The students were given the choiceto determine whether the researcher was able to share the data with otherparties by signing an informed consent module (Silverman, 2005). Duringthe course of the interviewing process the fact that participants’ interventionswere recorded made necessary, also after the interview had taken place, tosign another consent form as potentially sensitive information, which theparticipants were not aware of they would have been asked, might have beenshared.
In displaying the data collected confidentiality and anonymity was re-spected as far as possible, in the transcripts the students were identified bya code and no link appeared between the data collected and the students’identity. Since fully informed consent is always difficult to achieve the re-searcher handled with maximum responsibility the powerful position that heheld and its associated moral duty.