THE RESEARCH METHODOLOGY
6.1 Introduction
In the present chapter there will, firstly, be an attempt to justify the application of what are often considered to be divergent research methodologies. Secondly, there is a discussion of the samples used in the study. Thirdly, the hypotheses used to guide the study, and to which answers are sought, are discussed.
This chapter attempts to provide the reader with an appreciation of the issues that have come to provide a methodological focus for the present study. A researcher cannot fully appreciate the complexity of the methodological issues that must be resolved when conducting research without taking cognisance of writings in the area. What the researcher discovers as truth depends a great deal on the tools with which they go searching for the truth. What follows is a selective and idiosyncratic review of some of the pertinent literature.
The main focus of the research enterprise is to describe and explain the phenomena which it sets out to investigates. This task is complicated by the researchers' set of values and beliefs that channels their vision and explanation in a particular direction. Kuhn (1970: 18-19) viewed this channelling process in terms of paradigms which, while informing the research process, simultaneously constricted the research by setting up boundaries. Paradigms define what are acceptable research topics and theories. At the same time paradigms force the acceptance of a particular view of reality to the exclusion of other perspective's.
6.2 The Qualitative/Quantitative Dichotomy
In recent years there have been loud calls for the integration of quantitative and qualitative research from a number of different quarters (Schofield and Anderson, 1984). The main reason given for the acceptance of this integrated approach is that it opens up the possibility of integrating the positive elements of quantitative and qualitative methods, resulting in a better explanation of the research problem.
This scientific pluralism poses the problem of "how to establish an epistemology of complexity (Masulli, 1990: 4)" that will account for "those articulations that are destroyed by the separations between disciplines, ... and types of knowledge (Piaget, 1971: 362)." The renaissance and the start of the scientific revolution saw human beings as "seculari[s]ed object[s] for study" resulting in the development that viewed people as machines. This secularised objectivism lead to a quantitative science that was "functional, operative [and] geometrical (Masulli, 1990: 20)." This approach led to the formation of the quantitative methodologies in the social sciences.
Qualitative research (Spindler and Spindler, 1982; Wilcox, 1982), firstly, is conducted in natural settings; secondly, it relies on the researcher to both gather data and analyse it; thirdly, it concerns itself with "thick descriptions" that focus on events and their social meanings (Geertz, 1973); fourth, its primary concern is with social processes rather than outcomes (Schofield and Anderson, 1984: 4-5). This approach is usually characterised by the work of anthropologists (Bullivant, 1986; 1988; Moore, 1986a; 1986b; Ogbu, 1974; 1978; 1991) and sociologists (Gist and Wright, 1973; Lobo, 1988; 1994; Goffman, 1963).
While qualitative research is "holistic and contextual", quantitative research focuses on the testing of specific hypotheses. This approach uses the natural sciences as its model and while it does not always emphasise experimental design, it usually uses statistical methods to analyse its data. Quantitative research is characterised both by its focus on producing quantified data and by its emphasis on a research process which results in numbers which can be analysed using statistical packages and multivariate statistics (Schofield and Anderson, 1984: 5).
A number of writers have suggested (Goetz and Lecompte, 1981; Rist, 1977) that most researchers who use the terms qualitative and quantitative imply the existence of a dichotomy. However, it may be more accurate to think of these two approaches as different ends of a single continuum. As the awareness that qualitative and quantitative approaches are not antithetical to each other increases, there is a growing acceptance on the part of some quantitative researchers to use qualitative methods.
An example of the growing acceptance of qualitative research by quantitative researchers is that shown by Donald Cambell (Schofield and Anderson, 1984: 6). Campbell and Stanley's (1969:6) book dealing with quasi-experimental design contended that the "one-shot case study", which is typical of much qualitative research, has "such total absence of control as to be of almost no scientific value". More recently Campbell (1979: 52) in a complete turnaround wrote that when quantitative and qualitative results conflict "the qualitative results should be regarded as suspect until the reasons for the discrepancy are well understood".
6.3 Qualitative Research's Growing Popularity
The growing support for qualitative methods explicit in Campbell's (1979) statement seems to have many causes. First, there is a growing disenchantment with quantitative research due to factors such as the relatively trivial amount of variance, with typical r-squares of less than 0.4, often explained. Secondly, the highly technical nature of many quantitative multi-variate statistical techniques makes the research virtually incomprehensible even to well-educated readers (Van Maanen, 1982). Thirdly, there has been a re-evaluation of the traditionally prevailing view that qualitative and quantitative work are based on fundamentally different paradigms and are thus competing and irreconcilable ways of approaching research (Reichardt and Cook, 1979).
These methodological interactionists challenge the mono-methodological rubric that guides many social scientists (Campbell, 1979; Filstead, 1979; Spindler, 1982). According to these researchers qualitative and quantitative research differ in degree and these differences are far from being an "unbridgeable chasm" between the two schools of thought (Schofield and Anderson, 1984: 7). A lucid and persuasive argument for methodological interactionism is made by Reichardt and Cook (1979) who argue that method types are not irrevocably linked to different paradigms.
Reichardt and Cook (1979) argue that all social research has important subjective elements. The characterisation of quantitative research as objective ignores important aspects of subjectivity which enter the research at virtually every point. The whole quantitative research process contains subjective components:
from hypothesis formulation, to the selection of indices, to the interpretation of the data, all have their subjective aspects. Similarly, they point out that qualitative methodologies have no monopoly on validity, participant observation of a visual illusion is a case in which quantitative methods would lead to more valid conclusions about the stimuli than quantitative ones (Schofield and Anderson, 1984: 8).
The above perspective has been taken one step further by Gephart (1988) who has gone so far as to argue for "ethnostatistics". He defined this approach as the qualitative foundations upon which quantitative research is based.
6.4 Research Using Both Strategies
As Kuhn (1970) points out, no single paradigm, whether quantitative or qualitative can answer all the problems within its domain. Qualitative and quantitative methods "are not rooted in opposite and irreconcilable paradigms (Haase and Myers, 1988; Schofield and Anderson, 1984: 9)", there is no reason why they cannot be utilised simultaneously to complement each other.
Many authors continue to maintain (Smith, 1983: 13) that attempts to combine the two paradigms are based on the "unfounded assumption that the [two] methods are complementary". Fortunately, other authors are of the opinion that the two approaches can be combined quite successfully (Schofield and Anderson, 1984; Goodwin and Goodwin, 1984). The result, as perceived by many researchers, is that social research may profit from the different strengths of the two approaches (Bernard, 1994; Osborne, 1987; Campbell, 1979; Fetterman, 1982; Light and Pillemer, 1982; Spindler, 1982).
Whether a quantitative or qualitative approach is taken to the research problem either method is likely to generate its own particular problems. As Sechrest and Sidani (1995: 84) point out:
No such thing as an error-free method of inquiry exists. Moreover, it is difficult, sometimes impossible, to know even the direction of error, let alone its extent, from within the constraints of the application of the method itself, indicating the need for triangulation.
6.5 The Process of Triangulation
The process of "methodological triangulation" is seen as being the successful "merging" of the two methodologies (Duffy, 1987a). In this process qualitative and quantitative methodologies are used to reinforce each others findings rather than being "antagonistic" to each other (Duffy, 1987b). Using two or more different methods to study an area of interest or triangulating, is:
useful for confirming the direction of bias, for estimating its extent, for validating results (convergence), and for providing complementary (more complete) information (Sechrest and Sidani; 1995: 84).
Increasingly, it is not enough for the social scientist to just attempt to explain a particular social process. It is knowledge about causal relationships that has become vital to the scientific enterprise as well as to many policy decisions. The researcher is expected to make projections from data "maximising" the generation of information about causal relationships (Schofield and Anderson, 1984: 9).
Experimental and Quasi-experimental approaches, with their attendant statistical methodologies, provide the researcher with the opportunity to suggest causal relationships. At the same time these methods have their own particular drawbacks. They are far from ideal for exploring the impact of a whole variety of contextual factors which cannot be clearly specified prior to evolving the model. They tend to be rigid since change of course in midstream in response to new information is difficult when certain data has already been collected, through questionnaires for example (Schofield and Anderson, 1984: 9; Campbell, 1979).
Qualitative research, on the other hand, has strengths where quantitative designs are weak. Specifically, qualitative research has problems specifying causal connections with the degree of certainty or precision that many quantitative strategies can. However, it is ideally suited to exploring social processes and the context in which they occur. Further, qualitative research is particularly good at capturing "with both vividness and subtlety the perceptions of the individuals being studied" (Schofield and Anderson, 1984: 9-10).
With many researchers calling for simultaneous utilisation of quantitative and qualitative strategies, examples of research which do so are becoming more common. Increasingly, major studies use both approaches (Jiobu, 1988; Ransford, 1994; Fetterman, 1982; Wilcox, 1982; Osborne, 1979; Erickson & Mohatt, 1982; Verma and Ashworth, 1986).
The use of both strategies is, in general, "thought to produce more valid and reliable results" than simply using a single methodology (Sarantakos, 1993: 156). But not all researchers are in agreement with this position (Silverman, 1985). The strength of using both approaches is complicated by a number of problems. Firstly, the joint use of both strategies is likely to be expensive and time consuming; secondly, combining two different types of methodologies obviously requires a much broader range of research skills than using either one by itself (Reichardt and Cook, 1979).
The present study will use both quantitative and qualitative procedures to explore the research hypotheses. Logistic regression is used to explore the effect that a number of variables have on unemployment. Ordinary least squares regression models will be used to explore the issue of hourly income and how it is affected by a host of independent variables. Structured interviews are also conducted in an attempt to find support for the results of the regressions and to enrich the quantitative results with "thick descriptions" provided by the interviews. This two step approach is an attempt to overcome some of the shortcomings of different methodological approaches.
6.6 Methodological Difficulties of Quantitative Research
As the vast majority of quantitative studies are correlational, the possibility of the cause of a relationship being hidden by another variable is always present (Heath and Nielson, 1974). Moreover, randomisation, one of the key concepts in experimental research, is often not feasible for the assignment of subjects to particular test situations. Effects have been isolated almost universally on the basis of statistics which depend on randomisation (Heath and Nielson, 1974: 473-474). A whole set of other statistical assumptions, like the normality and homogeneity of variance of the data, are not checked in the bulk of research.
Analysis based on ordinary least squares regression methods assume "that relations among variables are, to a reasonable approximation, linear and additive (Kelley and McAllister, 1984: 390)." It also involves the quantification of a large number of variables concerning demographic characteristics personal and family background and socio-economic attainment. While it may be acceptable to present age, gender, income or years of education as interval numbers the quantification of class position or occupation is problematic.
For instance Kelley and McAllister (1983) dismissed the possibility of including capitalists in their study of power relations in Australia. In their opinion, a national sample has too few large capitalists to matter much. But the power and influence of these capitalists is not proportional to their numbers in Australian society (Jakubowicz and Castles, 1986). This issue of disproportionate influence is less critical in the case of the present study, given that an attempt is being made to examine the attainment of an "average" AI relative to an average UKD and AD.
The criteria of validity for the new causal models are that:
functions should be linear, with any interactive effects noted and countered. The model should be recursive, and avoid feedback loops. The causal direction of the variables can be specified, so that antecedents and outcomes can be identified. Error terms should be uncorrelated. The assumptions of multivariate analysis as to the interval scaling, homoscedasticity, and absence of multicolinearity, must hold (Menzies, 1982: 157).
Jakubowicz and Castles (1986: 17) argue that it is not possible to restrict social analysis to such data since it leads to "a reduction of social reality". Even more critically, they argue that, while the major attraction of the quantitative method is that data can be analysed using sophisticated multi-variate statistical procedures to generate complex output, the numbers that are analysed are arrived at in what is often a "subjective and arbitrary way". This output is then used to "convey an appearance of scientific objectivity" based on regression coefficients and goodness of fit indices such as the R-Square.
One of the main criticism of the quantitative approach to social processes is that it is primarily reductionist, in that it:
reduces the complexity of human interaction, the role of historically-mediated social institutions and the significance of values, norms and culture to a set of mechanistic and quantifiable relationships (Jakubowicz and Castles, 1986: 20).
The object of the quantitative study is not just the social process or outcome but the methodology of the researcher. Those social processes that are difficult to measure are often excluded, so that the research problem is defined by the methodology. Further, those processes that are measured are usually redefined by the process of measurement and statistical analysis. A given social relationship is described only in terms of quantifiable variables linked in a statistical model. While this model looks scientific and complex because of the statistics, in reality it has oversimplified a complex social process and replaced it with a complex statistical one (Jakubowicz and Castles, 1986).
Quantitative research undoubtedly has its problems. However its strength is its ability to quantify a complex social problem in the form of a statistical model. This model can then be used to make predictions about future events given changes within the model (Hope, 1992: 42). In the present study, where discrimination is being investigated, the quantitative approach is especially useful. Victims of discrimination are often not aware of it. Conversely, other individuals may quite possibly imagine discrimination when there is none. The use of a quantitative approach allows the researcher to overcome some of the above problems by using quantifiable indicators such as hourly income and participation and unemployment rates to deal with a particularly subjective issue (Jones, 1992a).
6.7 Methodological Problems Associated with Qualitative Research
A major problem associated with the qualitative approach is that of reliability (LeCompte & Goetz, 1982:35). Unlike quantitative studies, the main instrument of research in qualitative studies is the researcher: a person, not an observation scheme, a standardised test or a questionnaire form (Pelto, 1970: 140; Wolcott, 1976: 27). Furthermore, because social settings are not uniform, it is difficult if not impossible to replicate qualitative studies. As LeCompte and Goetz (1982: 35) stress, this problem exists regardless of the approach used because both quantitative and qualitative researchers have an equally difficult task in selecting identical replication sites.
6.8 Obstacles to Reliability and Validity
There are two types of reliability in ethnographic studies, external and internal reliability (LeCompte & Goetz 1982:37-43). External reliability is the degree to which the study can be replicated. Internal reliability is the degree to which multiple observers within a single setting agree about what they see and hear. In addition to the threats to reliability there are other threats to the validity of ethnographic studies. Internal validity concerns whether researchers actually observe and measure what they think they are observing and measuring. External validity is the extent to which the theoretical constructs and postulates generated or tested are applicable across groups (LeCompte and Goetz, 1982:43-53).
There are several weaknesses associated with ethnography which stem from the researcher's long-term, personal involvement in the research setting. There are problems of, firstly, ethnographer fatigue (Wolcott, 1976:32); secondly, personal involvement and its effects on both the social setting and the researcher (Hord, 1980; Wolcott, 1976: 32-33; Zigarmi & Zigarmi, 1980); thirdly, ethical issues of what to do about personal confidences shared by informants during the data collection process (Rynkiewich & Spradley, 1976; Spradley, 1979: 34-39); fourthly, interpersonal skills and sensitivity (Wolcott, 1976: 28); and fifth, the substantial time required to write up the data which have been collected (Wolcott, 1976: 31).
Just as the quantitative approach has its share of problems, similarly the qualitative approach has its own weaknesses. The main strengths of qualitative research are its "flexibility" and its ability to "develop a deeper understanding" of the subjects social environment (Sarantakos, 1993: 52).
6.9 Methodological Interactionism
What the researcher discovers is heavily dependent upon the research tools and methodology used to explore social reality. Researchers are always guided by theory, whether implicit or explicit, that by its very nature can provide only a partial explanation of reality (Kuhn, 1970). In recent years there have been increasing calls for the integration of qualitative and quantitative approaches (Bernard, 1994; Sarantakos, 1993; Schofield and Anderson, 1984). An eclectic methodological approach can often give the social researcher a more rounded picture. In contrast to the holistic and contextual approach taken by qualitative research, quantitative research emphasises experimental design and statistical analysis. Increasingly researchers are beginning to view these two approaches as different ends of a single continuum (Goetz and Lecompte, 1981) rather than as irreconcilably different and antagonistic methodologies.
"Methodological interactionism" is finding increasing support among a wide range of researchers. These researchers consider all methodologies to have limitations and as such the researcher should not be constrained by the limitations of a particular methodology. To do the research topic justice all approaches should be drawn upon freely so as to conduct effective research.
6.10 Triangulation and the Two Stages of Analysis
Consistent with concept of triangulation, there are two stages of analysis used in the present study. The first stage takes a highly quantitative approach where regression models are built and then used to make predictions about hourly income and unemployment. In the second stage, AIs were interviewed so as to get their "subjective impressions" about life in Australia, Canada, the U.K and India.
6.11 The Quantitative Analysis
The quantitative stage itself consisted of two sections. In the first section an attempt was made to gain an accurate impression about AI progress in Australia. The ABS in Canberra was approached to produce tables of hourly income for AIs depending upon their occupation, education and industry they were employed in. In the second stage the 1991 one percent population sub-sample was used to build regression equations and conduct comparisons between AIs, UKDs and ADs.
6.12 Analysis of the 1986 Census Data
It was deemed to be vitally important to gain an accurate view of the number of AIs in Australia. Towards this end the ABS was requested to scan their 1986 census data file. The 1986 census was unusual in that it requested respondents to answer a question about their ancestry. The AIs were separated out from other "ethnic" Indians by selecting only those Indians with a European ancestry. The ABS then provided detailed figures for AI hourly income in the areas of qualifications, occupation and industry. Once the data regarding hourly income for AIs was available, the 1986 one percent sub-sample was used to calculate hourly income for UKDs and ADs.
6.13 Analysis of the 1991 Census Data
The 1991 one percent population sample was used to build regression models. While the 1986 analysis relied on having a large number of AIs for analysis, the 1991 analysis relied on a selection of Eurasians. A Eurasian was defined as being a person who had both parents born on the Indian sub-continent and Sri Lanka, was Christian and spoke only English at home. Undoubtedly, some of the respondents selected would not have European ancestors, but they would have still been highly Westernised. The 1991 analysis assumed that these Eurasians could be used as a proxy for AIs.
Once the proxy AI sample was selected, the two comparison groups were then selected. Respondents of United Kingdom Descent (UKDs) were defined as those who had both parents born in the U.K and spoke only English at home. Those of Australian Descent (ADs) had both parents born in Australia and spoke only English at home.
The three sample groups had their average hourly incomes compared using One-Way-ANOVA's. Then regression models were built for the AIs, UKDs and ADs, so that their hourly income and unemployment rates could be compared. Further, the effects of overseas and Australian work experience, overseas and Australian education were modelled for the three groups. These statistical models will be explored in detail in the next chapter.
6.14 The Qualitative Analysis
Consistent with the theory of triangulation, the statistical analyses will be followed by a chapter dealing with a series of structured interviews. A total of 15 AIs from Australia, Canada, the U.K and India were interviewed. These interviews were an attempt to gain a deeper understanding of the issues than could be gained through statistical analysis.
6.15 Fieldwork procedure
Data gathering, in the form of interviews, began early in 1995. Those people who were interviewed were contacted before the interview and a mutually agreed time for the meeting was settled on. In most cases the interviews were held at the subjects house for about one hour. Tape recordings were made of the meeting with the approval of the subject, these tapes have been made available to my supervisor to corroborate my findings.
6.16 The structured interviews
The people interviewed represented people with both tertiary qualifications and those without. The group had representatives from both the middle and working classes as indicated by occupation. Further, people who achieved their highest qualification within the Australian educational system and those who achieved their highest qualifications in India were also interviewed.
Table 6.2
Country Highest Qualification was Gained in for Those Interviewed Country Highest Tertiary Qualifications Qualifications Gained Yes No Total Australia and Other 3 3 6 India 1 8 9 Total 4 11 15
Birthplace by Socio Economic Status for those Interviewed Socio-Economic Status Birth Place Middle Working Total Australia - 2 2 India 5 8 13 Total 5 10 15
The interviewed group lacked middle class AIs who had been born in Australia. This is unlikely to be a serious shortcoming given that Australian born AIs are likely to be mostly young. Six of the respondents who were interviewed were women, and nine were men. A more detailed description of the respondents appears in an appendix.
6.17 Selecting the Quantitative Sample
According to Moser and Kalton (1986: 147) the following formula can be used to select the sample required for a given amount of error.
PP' (1 - PP') n = -------------------- [S.E.(p)]2
n = simple random sample of size n
PP' = proportion in the population with some attribute [S.E.(p)]2 = the standard error of the estimator If we estimate the proportion of AIs who meet our sampling criteria of, having worked at least 35 hours a week in Australia, at 7,692/18,000 about 43 per cent and we assume that a standard error of more than 5 per cent would be undesirable, we can use the above formula to solve for n.
0.43(1 - 0.43) 0.244 n = ------------------- = ------------------ = 98 (0.05)2 0.0025
The solution is n = 98, which is the required sample size if the population the subjects are selected from is random. The samples that have been drawn are substantially larger than that suggested by the sampling formula. This is likely to assist in avoiding serious statistical criticisms.
A sample of 7,692 AIs was drawn from the 1986 census. These were AIs who were working 35 or more hours per week. The total number of people who had two parents born on the Indian subcontinent and were of European ancestry was 18,000. With regard to the 1991 one percent sample, a total of 211 people were selected. These respondents had both parents born on the Indian subcontinent and Sri Lanka, were Christian and spoke only English at home. While some of these people, perhaps 5 percent, are undoubtedly ethnic Indians or Sri Lankans, they are obviously highly Westernised and are quite similar to many AIs.
6.18 Ethical Considerations
A convenient way of thinking about ethical issues is to follow through the logical sequence of the research. Ethical issues arise at four key phases: firstly, entering the field, one can misrepresent oneself; secondly, working in the field, when one can choose to exploit one's subjects; thirdly, exiting from the field, when it is possible to leave in bad grace; and fourthly, publication of results, when one of the key problems is maintaining confidentiality. Each of the phases is taken in turn and briefly discussed.
6.19 Entering into the Field
The process of conducting the research raised complex ethical and legal issues for the researcher. In an attempt to anticipate and resolve these issues, the procedures laid down in a number of recognised texts were followed (Sarantakos, 1993: 21-26; Homan, 1991; Sieber, 1992). At no point were the intentions of the research concealed, nor were promises made that could not be kept. The researchers affiliation with Monash University was declared and validated by the presentation of my Student Identification card.
6.20 Working in the Field
Efforts were made to minimise any suggestion of exploitation of the respondents. The demand on the respondents' time was lessened as far as possible. Respondents were interviewed after they had given their "consent", in their own time in their homes so that they would feel comfortable (Sarantakos, 1993: 25; Sieber, 1992: 26-43).
The respondents' participation was secured without coercion and their right not to be involved was respected. They knew of the purpose of the research and no deliberate attempts were made to deceive them. Respondents were informed that the primary purpose of the research was to gather information about the socio-economic position of the AI community both in Australia and overseas. It was hoped that if there was a problem in this area, then it could be identified and solved. Assurances were given that absolute confidentiality would be maintained regarding the information provided and the identity of the respondent (Homan, 1991: 41-68; Sieber, 1991).
6.21 Exiting from the Field
Leaving the field in good standing is often a difficult process, given that some people can be quite sensitive about giving interviews. Every endeavour was made to leave the field as harmoniously as possible given that the AI community is a small one. If further research is to be conducted then it becomes vitally important to keep people on side. Otherwise any future research would be compromised by lack of interest and cooperation.
The interviews were conducted in two stages. The first stage involved obtaining the respondents opinions and attitudes about the socio-economic position of AIs. In the second stage of the interview, the respondent was "debriefed" by being told about the research study and the findings from the quantitative part of the study (Sieber, 1992: 39-43).
6.22 Publication of Results
As the first part of the study is highly quantitative and is based on data provided by the ABS there appears to be little likelihood of any problems with confidentiality or disclosure. The results will be presented in grouped form and not as individual data. Hence, there is no danger of disclosing the subjects' identities. In the second part of the study, which involves interviews, no names will be used. As a result there appears to be little possibility of a breach of confidentiality (Homans, 1991: 140-159; Sieber, 1992: 52-63).
6.23 Statement of Research Problem
On a general level the present research attempts to remedy a gap in the literature. The intellectual rubric that guides most Australian researchers involves the categorisation of ethnic groups from Asian countries simply as Asians, or, even more surprisingly, the aggregation of all "non-English speaking background" groups from Europe and Asia as "NESB". This practice ignores the enormous diversity that characterises people from apparently non-English speaking backgrounds.
Most researchers would simply categorise AIs as Asian. They would be labelled as non-English-speaking with a culture that is quite different from the Anglo-Celtic culture that predominates in Australia. The assumption is that ethnic groups are monolithic entities with homogeneous needs and characteristics. This is simply not true.
There have been many research studies conducted dealing with the AIs in India. These studies have had two threads in common. Firstly, they have used qualitative methodologies to investigate the AI situation. Secondly, these studies have concluded, without exception, that AIs attain poorly at an academic and socio-economic level.
Part of the justification for the present research lies, firstly, with filling a gap about Westernised Asians in the literature. Secondly, it provides for the AIs of Australia an accurate indication of how well the people of the community are performing socio-economically and academically. Thirdly, it aims to provide at least a partial explanation of their socio-economic and academic attainment. A fourth reason for conducting the research is to fill the need for studies dealing with "mixed-race" minorities in the countries to which they have emigrated.
The AIs have been painted as being primarily a poorly educated working and lower class grouping in India. Using Ogbu's caste theory results in the general perception that under certain conditions the AIs will not attain. The main characteristics of a caste group is that it fails to attain academically and socio-economically. Members of the caste group tend to participate less in the work-force, have higher unemployment levels and earn less than the non-caste groups. Further, the members of the caste group tend to be poorly qualified and occupy lower status occupations. If the members of the caste group emigrate to a new country, they often begin to achieve both educational and socio-economic success. Further, those members of the caste group most similar to the majority grouping in terms of human capital and appearance usually outperform other caste members
6.24 The Hypotheses
In this section a number of specific hypotheses will be stated prior to their being tested. The hypotheses will be examined in the following chapters in an attempt firstly to gain an appreciation of AI attainment in Australia, and secondly, to describe the AIs' subjective perceptions about their attainment in Australia. The following specific hypotheses will be examined.
6.24.1 Unemployment Rates
H1: AIs who are Anglican will have lower unemployment rates than AIs who are Catholic.
H2: AIs with overseas work experience will have higher unemployment rates than UKDs and ADs with overseas work experience.
H3: AIs with overseas work experience will have higher unemployment rates than AIs with both overseas work experience and post-migration human capital.
H4: AIs with overseas education will have higher unemployment rates than UKDs and ADs with overseas education.
H5: AIs with overseas education will have higher unemployment rates than AIs with overseas education and post-migration human capital.
H6: AIs with Australian work experience will have higher unemployment rates than UKDs and ADs with Australian work experience.
H7: AIs with Australian work experience will have higher unemployment rates than AIs with Australian work experience and AD human capital.
H8: AIs with Australian education will have higher unemployment rates than UKDs and ADs with Australian education.
H9: AIs with Australian education will have higher unemployment rates than AIs with Australian education and AD human capital.
6.24.2 Hourly Income
H10: AIs who are Anglican will have higher hourly earnings than AIs who are Catholic.
H11: UKDs and ADs who are Anglican will have hourly earnings that are similar to UKDs and ADs who are Catholic.
H12: AIs with Australian education will have lower hourly incomes than UKDs and ADs with Australian education.
H13: AIs with Australian education would have lower hourly incomes than AIs with Australian education and AD human capital.
H14: AIs with overseas education will have lower hourly incomes than UKDs and ADs with overseas education.
H15: AIs with overseas education will have lower hourly incomes than AIs with overseas education and AD human capital.
H16: AIs with Australian work experience will have lower hourly income than UKDs and ADs with Australian work experience.
H17: AIs with Australian work experience will have lower hourly incomes than AIs with Australian work experience and AD human capital.
H18: AIs with overseas work experience will have lower hourly incomes than UKDs and ADs with overseas work experience.
H19: AIs with overseas work experience will have lower hourly incomes than AIs with overseas work experience and AD human capital.
6.24.3 Qualitative Hypotheses
H20: AIs in Australia are less likely to believe in a job ceiling than in India.
H21: AIs in Australia are likely to manifest effort optimism.
H22: AIs in Australia are likely to manifest a third world yardstick.
H23: AIs in Australia of both the first and second generations will manifest the belief that they could succeed.
H24: AI childrens' attitudes toward education in Australia will be achievement oriented.
H25: AI parents' attitudes toward education in Australia will be achievement oriented.
H26: AIs in Australia are more likely to view education as helping to achieve socio-economic success.
H27: AIs in Australia are unlikely to view the maintenance of their language and western culture as a marker of caste.
H28: AIs in Australia will strive for higher education.
H29: AIs in Australia are less likely to believe that they are being discriminated against.