In this paper the advantages and disadvantages of qualitative and quantitative methodologies are discussed. It is proposed that the choice of methodology should be driven by the research questions or objectives, rather than any preference for a given methodology. Examination of the literature reveals that co-op researchers utilized qualitative means of inquiry when they wished to ascertain subjective views, such as perceptions of placement experiences, and when greater depth of inquiry was required.
the importance of research for co-op practitioners was highlighted in the 11th biennial World Association for Cooperative Education (WACE) conference held in Washington in July 1999, in which it was claimed there exists a great need for more research in co-op education, but ostensibly there is little evidence that much research is in fact being done (Rowe, Ricks, & Varty, 1999). Other authors have expressed similar sentiments (see, e.g., Ricks, Cutt, Branton, Loken, & Van Gyn, 1993). A number of reasons have been suggested for this somewhat depressing situation, with the lack of time for research considered as probably the major impediment (Rowe et al., 1999). Other reasons proffered include the suggestion that many co-op professionals, being placement coordinators rather than academic faculty with training in education, lack access to research funding, graduate students. and the support structures that make meaningful research practical in other academic disciplines (Ricks & Mark, 1997). Ricks and Mark (1997) observed that many co-op practitioners conduct inquiries. yet ostensibly do not equate their work with research, feeling that such inquires are markedly inferior to "orthodox scientific world views" (p. 47) as to what constitutes research. Consequently, with little time to undertake research, and lacking confidence and research skills, it is perhaps not surprising that co-op practitioners are not prolific publishers of research.
Interestingly. co-op practitioners arc well positioned to conduct action-research. Action-research comprises research initiated by teachers or other education practitioners, conducted within the environment of the practitioner, typically small-scale, and highly contextually based (Denzin & Lincoln, 1998). Such studies are commonly concerned with pragmatic outcomes such as improving learning, gaining a deeper understanding of classroom practice, or situated learning (Keeves, 1998). In fact, other than the focus and scale of research activity, action research does not differ significantly from other research. The researcher practitioner is faced with the same issues of research design: namely, definition of the research objective or question, design of the study. identification of sources of evidence. collection of evidence, and data analysis and interpretation (Burns, 1994; Keeves & Alagumalai, 1998). Though research design may prove problematic for many co-op practitioners. the authors believe that co-op practitioners should consider conducting action-research; in particular it is our belief that qualitative means of inquiry are well suited to co-op research inquiries. We came into co-op research from a background in the physical sciences and our research activity in our disciplines (chemistry and earth sciences) was based on an empirical positivist view of knowledge acquisition and a quantitative methodology. It is fair to say that we harbored some reservations about qualitative means of inquiry; chief amongst these were concerns about the reliability and validity of qualitative research and the ability to generalize findings to other settings, what is referred to as the trustworthiness of an inquiry (Guba & Lincoln, 1989), An analysis of research reports presented in the literature reveals that during the 1990s less than 5 percent of co-op research inquiries utilized qualitative means of inquiry; similarly, analysis of papers presented at the WACE conference in Washington in 1999, revealed that only 10 papers. or about seven percent, of the inquires used a qualitative or mixed methodological approach. Hence, it seems that co-op research has been dominated by quantitative based research studies. This is in marked contrast to, for example, science education research that is now dominated by qualitative or mixed methodology inquiries (White, 1997).
We believe that other co-op practitioners (or researchers) may share our early concerns about this methodology, resulting in a perception that qualitative studies are in some measure inferior to quantitative-based studies. Consequently, in this paper we describe qualitative means of inquiry and address the concerns identified above. We begin with an overview of the development of qualitative research in education and discussion of the merits of both quantitative and qualitative approaches. This is followed by a description of co-op qualitative studies that utilized qualitative or mixed methodological approaches. In view of the paucity of qualitative inquiries, we discuss the issue of trustworthiness, and provide some guidelines for interview technique.
Research in education since the turn of the century has been dominated by the use of a quantitative methodological approach (Denzm & Lincoln, 1998; Rossman & Wilson, 1985). This approach, based on scientific empiricism, uses the power of mathematical analysis to establish general laws and principles. Gradual recognition of the differences between scientific and educational research led to increasing interest in a qualitative methodological approach with its foundations in ethnography using data gathering tools such as unobtrusive observation, participant observation, and in-depth interviews. Changes in the methodology of education research can be traced to the development of the constructivist paradigm, which in turn had its origins in changes to scientific thinking that occurred in the twentieth century (Nussbaum, 1989). Chief amongst these were drastic changes in the ideas of modem physics that started to undermine the prevailing empiricist belief that science could make absolutist claims, in other words, that once knowledge is acquired it can be described in absolutist terms such as true or proven. Work by Karl Popper (1959) raised questions about the possibility of ever proving or confirming knowledge. This led to the proposal that knowledge was not discovered but was instead the result of construction by the human mind. Unlike empiricism, which claimed that observation came before theory. constructivism presupposes that theory precedes observation and that observations can be selected and conducted only through theoretical expectations. Therefore, our own constructed theories determine how we perceive the world (Driver, 1989). A similar paradigm shift took place in psychology where for decades following the 1920s behaviorism dominated. Behaviorism is a school of thought that emphasizes the passivity of the mind, with information from the environment providing an input that is directly transmitted to, and accumulated by, the individual (Gilbert & Watts, 1983). This resulted in the so-called cultural transmissive teaching approach or conduit model for instruction that became dominant in Western education (Pope & Gilbert, 1983; Tobin, Briscoe, & Holman, 1990). In summary, the empirical-positivist philosophy has dominated approaches to teaching and learning as well as methodology in education research.
Quantitative and qualitative means of inquiry each possesses advantages and disadvantages. As Patton ( 1990) points out "the advantage of the quantitative approach is that it is possible to measure the reactions of many subjects to a limited set of questions, thus facilitating comparison and statistical aggregation of data" (p. 165). In contrast to quantitative inquiries, qualitative methods typically produce a wealth of detailed data about a much smaller number of people and cases. Hence, depending on the issue of interest, co-op researchers may find it more useful to carry out an in-depth investigation using fewer subjects, in order to gain greater detail about specific issues. In contrast, the purpose of most quantitative inquires is to afford greater confidence in the generalization of the research findings. The validity of such generalizations is highly dependent on the sampling process employed. Because co-op programs are highly variable in nature (see, e.g., Eames & Rowe, 1996), this may mean the generalization of the findings of an inquiry based on one institution, or in one country, to an international audience is tenuous. We discuss this issue in greater detail later when we describe the means necessary to confirm the trustworthiness of qualitative inquiries. The choice of methodology often comes down to a trade-off between breadth and depth.
Qualitative methods permit the evaluation researcher to study selected issues in depth and detail: the fact that data collection is not constrained by predetermined categories of analysis contributes to the depth and detail of qualitative data.
Quantitative methods, on the other hand, require the use of a standardised approach so that the experiences of people arc limited to certain predetermined response categories.
Researchers are not necessarily required to choose between the two approaches. Indeed, Guba and Lincoln ( 1989) believe "both quantitative and qualitative methods may be used appropriately with any research paradigm" (p. 105) and Shulman ( 1998) points out qualitative inquiries typically utilize a variety of data gathering tools, and can include quantitative tools such as survey instruments. The principal difference between a qualitative or phenomenological approach and a purely quantitative approach is that the former, with its roots in constructivism, recognizes the significance of subjective experience and in general is characterized by greater depth (Wolcott, 1988b). In recent times there have been a number of calls for the use of a combined methodological approach to research in science education (Denzin, 1970; Fraser, 1991, 1994; Reichardt & Cook, 1979). Because the two methodological approaches have their origins in different paradigms and are derived from opposing theoretical and philosophical perspectives. some authors argue that they are incompatible and should not be combined (e.g .. Burrell & Morgan, 1979; Smith, 1983 ). However. other authors have argued against this stance, maintaining that the paradigm argument is essentially irrelevant. Rather, qualitative and quantitative methodological approaches are investigating the same thing, and the difference is essentially one of choice of data collection tools (Cizek, 1995; Goodwin & Goodwin,1984).
In fact there are a number of potential advantages of combining quantitative and qualitative methodologies (Fraser, 1991 . 1995; Reichardt & Cook. 1979). Educational research differs from science and discipline-based research in that it has multiple purposes, being interested in both processes and outcomes. Thus, analysis of monitoring, impact assessment. and causal explanation represent a broad range of tasks most efficiently achieved by a combination of methods. Furthermore, quantitative and qualitative approaches build upon each other. For example, choosing a statistical model to fit the data, interpreting the output results, and generalizing the findings to their settings. all rely on qualitative knowledge; "quite simply, researchers cannot benefit from the use of numbers if they do not know, in common sense terms, what the numbers mean" (Reichardt & Cook. 1979. p. 23). In addition, data triangulation, the gathering of data from different sources through convergence. affords more confidence in the interpretation of the data (Denzin & Lincoln. 1998). The use of a combined methodological approach is clearly challenging for researchers, including co-op researchers, but this strategy is gradually gaining momentum in education (e.g., Blumenfeld & Meece, 1988; Fraser, 1991, 1994; Gogolin & Swartz, 1992; Shulman, 1998) and science education research (White 1997).
The issue of choice of methodology for research in co-op, like other areas of inquiry, is best resolved by considering the more fundamental issue of research design (Shulman, 1998; White, 1997). Patton ( 1990) asserts that it is the research question or goal that should dictate which methodology, or if a combination of methodologies, is appropriate:
Which research design is best? Which strategy will provide the most useful information to decision makers? There is no simple, immediate, and universal answer to that question. The answer in each case will depend on what intended users want to know, the purpose of the study, the funds available, the political context, and the interest/abilities/biases of the researchers. (pp. 95-96)
Thus, some research questions will be readily answered using qualitative means, others quantitative, and some will be best addressed using a combination of the two. What is necessary is appropriate research design. That is, researcher need to decide what information is most needed and most useful in a given inquiry, and then to employ those methods best suited to producing the needed information.
Co-op research utilizing qualitative means of inquiry has focused on obtaining an increased depth of understanding of subjective perceptions that students and employers hold of work experience. Studies involving students' perceptions include perceptions of the advantages and disadvantages of international placements (Coll & Chapman, 1999, in press), perceptions on whether the qualities needed for meaningful employment is provided by co-op (Huber & Tegart, 1996), articulated beliefs of changes accrued during co-op placements (Van Gyn & Hendy. 1999), employers and students' perceptions of workplace performance (Hoskyn, 1999), perceptions of career direction (Comerford, 1999), perceptions of proteges' perceptions of characteristics of the mentoring process in co-op (Van Gyn & Ricks, 1997), mentors value in teacher training as a result of pre-service work experience (Thuynsma, 1999), and desired learning outcomes from work placements (Langford & Lockart, 1999). These studies utilized interviews as a key data gathering tool, allowing the researcher to probe stakeholders' views in depth. For example, Coll and Chapman's (1999, in press) study of students' perceptions of international placements revealed a large number of concerns that placement coordinators were unaware of. The use of an interview approach, however, established that whilst such issues were of concern, the benefits accrued from an international placement, although numerically fewer, far outweighed the difficulties. A survey instrument may thus have created a misconception that students were unhappy with overseas placements.
A number of studies employed a mixed methodological approach, typically questionnaires accompanied by in-depth semi-structured interviews. These studies drew on qualitative tools to prove deeper understanding of trends or issues identified from quantitative data obtained from survey instruments: for example, Eames's (1999) study of learning as a social process based on situated social-cognition, co-op learning processes as a result of self-efficacy for graduates (Linn, 1999), and the identification of the crucial factors required to build a quality workforce from an examination of the important relationships between the students' industry experience, employer satisfaction and career progression following graduation (Comerford, 1999). Likewise to the qualitative studies reported above, the use of interviews in conjunction with surveys provided greater depth of understanding. For instance, Eames (1999) study revealed that many students held ill-formed views as to what constitutes science, or technology, and were unsure of how their enculturation into the workforce might occur.
Co-op researchers unfamiliar with qualitative research may find it difficult to understand how researchers can be confident about the quality of the data, interpretation of data, and ability to generalize research findings. Four criteria have been proposed for qualitative inquiries that subscribe to a constructivist paradigm, in order to judge the trustworthiness of the inquiry: credibility, dependability, confirmability, and transferability (Altheide & Johnson, 1988; Guba & Lincoln, 1989).
Credibility represents an assessment of the isomorphism between constructed realities, that is, a match between the constructed realities (Guba & Lincoln, 1994). Matching is enhanced by prolonged engagement, persistent observation, peer debriefing, negative case analysis, member checks, and progressive subjectivity. Prolonged engagement provides an opportunity to establish good rapport and trust with participants, allowing the researcher to overcome the effects of misinformation or presented fronts. Persistent observation allows the researcher to identify characteristics and elements in the inquiry that are most relevant to the issue under inquiry. Peer debriefing helps the investigator to understand his or her own posture and values and their role in the inquiry. Negative case analysis involves revising the hypothesis in the light of hindsight; in other words, an inquiry is seen as a dynamic rather than static process. Member checks. that is, the process of negotiation with stakeholders (i.e., all the participants), provides participants with the opportunity to offer additional information to that gained from, for example, interviews, and allows participants the opportunity to confirm individual data. Progressive subjectivity, the process of monitoring the researchers' own developing construction, serves to remind the inquirer that inquiries are best considered joint investigations between researcher and participants (Bell, J. 1993; Bums, 1994; Denzin & Lincoln, 1998). Dependability is concerned with the stability of data over time (Cronbach & Suppes, 1969).
Methodological changes and shifts in construction are common in qualitative inquiries, and are not considered to impact adversely on dependability. In contrast they are seen as an integral part of the inquiry process indicating increasing maturity of the inquiry. What is critical is that the changes and shifts in constructions are identified and described fully. Similarly. confirmability seeks to ensure that the results of an inquiry have not been subject to influence by the investigator and is enhanced when the raw data and process used to compress them are made available to scrutiny by the reader, thus providing an audit trail.
Transferability considers the extent to which the findings of one study can be applied to other situations (Lemke, 1998). In quantitative inquiries, generalizability is strongly dependent on sample selection, specifically on the randomness of the sample (Carver, 1978, 1993). Typically a target population is identified and a selection of participants made via a random sampling procedure or some more efficient variation such as stratified random sampling (Saikali & Jain, 1997; Wiersma, 1991). However, for qualitative inquiries the onus is shifted from the inquirer to receiver (Bell, J. 1993; Geertz, 1973) Thus the qualitative inquirer describes the context of the inquiry and provides detailed, or the so-called thick description of methodology and interpretation: it is then up to the reader to decide if the findings are relevant or pertinent to their own situation (Burns, 1994; Merriam, 1988).
Triangulation involves the gathering of information from a number of sources to crosscheck and then to assess the authenticity of individual accounts. By comparing and contrasting one source of information with another, the researcher is able to produce a more comprehensive and balanced study in response to the research question or objective. This approach not only provides more information about the validity of a study but aids in interpretation in that it affords the investigators an opportunity to "solve the problem of rival causal factors" (Denzin. 1970, p. 26). There are three possible outcomes of triangulation in educational research - convergence. inconsistency. and contradiction (Mathison. 1988). The historical goal of triangulation has been convergence. that is, the data provide consistent evidence for a particular conclusion. Inconsistency occurs when one data set does not confirm the findings of another set, but does not completely contradict it either. Contradiction occurs when the data disagree to such an extent that it is not possible to provide a plausible explanation. These latter two outcomes are often not perceived favorably since sometimes researchers feel that they cast doubt on the validity of the study or require greater imagination to devise plausible explanations for contrary data. However, lack of convergence provides different perspectives that may ultimately produce a deeper understanding of the research questions or goals. Because of this. each of the three outcomes of triangulation can have real value in education and co-op research (Denzin & Lincoln, 1998; Guba & Lincoln, 1994).
The interview is a key tool in the armory of the qualitative researcher. However, conducting interviews is a complex and demanding task. There is a large literature on the conducting of interviews (see. e.g .. Denzin & Lincoln. 1998; Jaeger. 1998). Here we provide a brief description of the most common means of conducting interviews. There are three approaches to conducting interviews; the informal conversational interview, the general interview guide approach, and the standardised open-ended interview (Fontana & Frey, 1998; Erickson, 1998). These approaches differ in the extent to which the interview questions are determined and are standardised before the interview occurs.
The informal conversational interview. as the name implies, is relaxed in nature, and the generation of questions is spontaneous arising from natural conversational flow. This approach is common in ethnographic inquiries where it is usually combined with participant-observation and fieldwork (Wolcott, 1988b). Such inquiries frequently involve multiple interviews and are of long duration, months or even years (e.g ., Wolcott, 1988a). The same participant may be interviewed on many different occasions, and for multiple participant inquiries the questions asked and data gathered will be different for each individual interviewed. The interviewer has the advantage of maximum flexibility, and can modify questions depending on the context of the investigation. The main advantage of the use of an informal interview approach is the depth of information gathered compared with other more structured approaches. There are, however, a number of disadvantages to the approach. Since data collection is less systematic, analysis may prove problematic. In addition, because interviews are often conducted in the field, audio-taping may not be practical - thus it is usually necessary to take extensive field notes. The general interview guide approach, also referred to as partial or semi-structured interview, is more structured in nature than the informal conversational interview and involves outlining a set of issues that are to be explored before interviewing begins (Wiersma, 1991). There is not necessarily a set order to the questions, and the specific wording used varies from participant to participant.
The interview guide serves as a form of checklist to ensure that all relevant topics are covered. There is still a considerable degree of flexibility retained in this approach, but the use of the interview guide helps to make the data gathering more systematic, facilitating analysis.
The standardized open-ended interview is the most structured of the three interview approaches and resembles a verbal questionnaire. There is little flexibility. the emphasis is on minimizing interviewer influence, thus data analysis is more straightforward. Variation among interviewers is minimised allowing for multiple interviewers, and because the interview is so highly focused, interview time is minimised. Such advantages are also realised by the use of written questionnaires (Fraser, 1991). The advantage a standardised open-ended interview holds over a self-completion questionnaire is that despite its relatively structured approach, some flexibility is retained, for example, participants have the opportunity to clarify ambiguity in questions.
A common characteristic of all three interview approaches is that they are more flexible than written questionnaires and afford the participants the opportunity to express their own views and perceptions in their own words, that is, the response is open-ended, and not confined to set predetermined categories. In practice, a given inquiry may employ a number of interview approaches.
In this paper we described historical developments in educational research and methodological options for co-op researchers. Research inquiries in co-op are dominated by the use of a quantitative methodology with few qualitative or mixed-methodological approaches reported. White ( 1997) reported that research in science education has moved from "controlled laboratory-style experiments to lengthy observations and descriptions" (p. 215), pointing out that education research has moved to become more relevant to practice. Recent co-op studies have illustrated the benefits qualitative research can provide for researchers. It is our view that co-op researchers and practitioners should likewise consider the use of a qualitative methodology, or if their research goals dictate, a mixed-methodology approach. In order to facilitate this objective, we have described some facets of interviewing and the necessary measures to ensure qualitative inquiries remain trustworthy in nature. In discussing these issues, it is our hope that co-op practitioners, who may have felt inhibited in conducting qualitative research, may cast aside these inhibitions and reap the benefits of being involved in what we have found to be a valuable means of inquiry.