LEARNING TO DO SCIENCE: IMPLICATI= ONS FOR SCIENCE TEACHER EDUCATION

&nbs= p;

Allan Feldman, University of Massa= chusetts Amherst

Kent Divoll, <= st1:PlaceType w:st=3D"on">University of Massa= chusetts Amherst

Allyson Rogan, University of Massa= chusetts Amherst

&nbs= p;

Abstract

This study examines an interdisciplinary scientific research project to understa= nd how people learn to be scientists. There are five sets of findings: 1) the configuration of research groups and laboratories varied according to the research area. 2) Professors had very different expectations for undergradu= ate, master's and doctoral students, which led to the development of a typology = of how students of science function as members of research groups. 3) The typo= logy is developmental, with students potentially being able to move from Novice Research to Proficient Technician to Knowledge Producer. 4) The education of new science researchers is informal and individual, and follows the structu= re of a traditional apprenticeship. 5) Research groups were both communities of practice and epistemic communities. Funded by NSF CHE-0221791.

Introduction

The National Science Education Standards (NSES) (NRC, = 1996) state that students should understand the nature of science as inquiry and "requires that students combine processes and scientific knowledge as = they use scientific reasoning and critical thinking to develop their understandi= ng of science" (p. 105). The NSES also state that for teachers to be able= to teach in this way, they should be "familiar enough with a science discipline to take part in research activities in that discipline" (p. 60). Few teachers have this knowledge or have had the opportunity to participate in scientific research activities. Accordingly, their students learn science as pre-packaged and delivered knowledge (Brickhouse, 1990; Fl= ick, Lederman, & Encohs, 1996; Lederman, 1992). To remedy this situation, the NSES call for science learning experiences that involve teachers as researc= hers in scientific inquiry along with working scientists. This suggests that for teachers to teach their students how to do science, they should know how to= do science. Given this, we ask, "How does someone learn to do science?&qu= ot; and in particular, "What beliefs do science and engineering professors have about the research education of their undergraduate and graduate stude= nts? "This is of importance to science teacher education because if we want= to know how to teach teachers how to engage in scientific research, we ought to know how scientists learn to be researchers.

Theoretical framework

Graduate study in science in the United States has two main components. The first consists of the accumulation of subject matter knowle= dge and the development of a deep conceptual understanding that occurs through coursework. This component occurs in the formal structure of an academic program in which the students are enrolled in courses. While the level of t= he content of the course, the number of students in the course, and the relationship between students and instructor are very different from undergraduate courses, the methods of instruction and the assessment of stu= dent learning continue to have a strong resemblance to what the students experie= nced as undergraduates. The second component is participation in research activities that leads to extensive knowledge of a subset of the subject domain, the learning of research skill= s, and the ability to frame and answer researchable questions. Except for those undergraduates who may have had the opportunity to do research – for example as part of an honors project – the research project is a new opportunity for learning for graduate students. This study focuses on the second component – learning to be a scientist while engaged in empiri= cal research.

In order to understand how people learn to be research= ers in the laboratory setting, we developed a theoretical framework that draws upon studies of apprenticeship learning and communities of practice. We use the concept of apprenticeship as legitimate peripheral participation in a community of practice that results in situated learning = of the skills and knowledge needed to be a working scientist (Lave & Wenge= r, 1991). Apprenticeship models of learning can be found in a wide variety of contexts and to a certain extent have characteristics unique to the context= in which they occur. However, apprenticeships also share some key commonalitie= s. An important characteristic of apprenticeships is the indistinguishable nat= ure of learning and the practice of work (Lave & Wenger, 1991). This is qui= te different from how students are taught in formal instructional settings in which they learn skills in isolation from their use in practice. Additional= key elements of an apprenticeship can be understood in terms of the role the teacher and learner play in the learning process in an apprenticeship. For = an apprentice, learning is demonstrated by performing tasks in a way that is analogous to the expert. For the instructor in an apprenticeship model, successful teaching is the ability to partition tasks into appropriate sizes that are useful for the developmental trajectory of the apprentice (Lave &a= mp; Wenger, 1991). There is some literature that suggests that the education of science researchers is done through apprenticeships (Fernández-Esqui= nas, 2003; LaPidus, 1997; Richmond, 1998). However, there is little research that supports this assumption.

Apprenticeships occur in communities of practice. A community of practice defines itself along three dimensions: mutual engagem= ent; a joint enterprise; and a shared repertoire. A community of practice involv= es more than the technical knowledge or skill associated with undertaking some task (Wenger, 1998) – its members are involved in a set of interperso= nal and professional relationships (Lave & Wenger, 1991; Wenger, 1998), whi= ch results in the community developing around things that matter to its members (Wenger, 1998). Members of communities of practice gain a sense of joint enterprise and identity because they are organized around a particular area= of knowledge and activity. Because they are communities of practice, th= eir members need to generate and appropriate a shared repertoire of ideas, commitments and memories; and various resources such as tools, documents, routines, vocabulary and symbols that in some way carry the knowledge of the community. In other words, it involves practice: ways of doing and approach= ing things that are shared to some significant extent among members (Capobianco & Feldman, 2006).

The community of practice in which graduate students i= n the sciences participate can be located in a particular location, such as a laboratory and/or it can be dispersed through space and time. Typically the community is a research group. A research group consists of at least one professor, a group of students and possibly one or more post-docs who engag= e in a joint research project or different, but related ones. Members of the research group meet regularly and report on and critique one another's rese= arch (Clark, 1997). Research groups can be as= small as one professor working with one or two students, or as large as those in high-energy physics, which can number in the hundreds (Knorr Cetina, 1999).=

Research groups as communities of practice are often associated with laboratories. A laboratory can be thought of as a place in which the natural world is manipulated and transformed through experimental work (Knorr Cetina, 1999). While the manipulation and transformation usually occurs to physical objects, it can also occur through the quantification of data and its subsequent numerical or statistical manipulation. However, research groups are not necessarily associated with laboratories. For examp= le, some geologists collect all their data in the field and do their data analy= sis in offices or computer centers. Of course, there are also research groups t= hat do theoretical studies and neither collect nor analyze data.

The research group working in the laboratory can be th= ought of as a community of practice in which new members learn how to maintain the laboratory and learn the skills needed for experimental work, such as the <= u>standard methods published for each field (e.g., Clescerl, Greenberg, & Eato= n, 1999). Even if there is no laboratory, the research group can be a communit= y in which new practices are developed and shared with a larger community of practicing scientists (Creplet, Dupouet, & Vaast, 2003).

However, to learn to be a scientist is more than to le= arn how to be a skilled practitioner in the laboratory. Scientists also have as their goal to create and warrant new knowledge. As a result, the research g= roup is not only a community of practice, it is also an epistemic community (Knorr Cetina, 1999), in which graduate students as legitimate peripheral participants attain the knowledge and skills needed to create and warrant new knowledge. Like a community of practice, an epistemic community= is a group of people with a shared repertoire, mutually engaged in a shared activity. However, while the community of practice has as its primary goal = the improvement of practice, the epistemic community has as its primary goal the creation and warranting of knowledge. Because epistemic communities have as their goal the creation of knowledge for the use by people who are not necessarily members of the local community, there is the need to convince t= he others that the knowledge is correct; that is, the knowledge must be warran= ted in some way. Epistemic communities must rely on some type of implicit or explicit procedural authority that plays a role in how the knowledge is warranted. Therefore, graduate education in the sciences has as its goal to teach new researchers how to warrant their knowledge by responding to the procedural authority that explicitly resides in guidelines for research and publication, and more implicitly in the review process for journal articles= , conference papers and funding proposals (Capobianco & Feldman, 2006).

Methods

The setting for this study is a National Science Found= ation (NSF) funded interdisciplinary collaboration among geologists, microbiologi= sts, environmental engineers, and science educators to study the natural remedia= tion of acid mine drainage (AMD) at an abandoned pyrite mine. The project has fi= ve principal investigators, all of whom are professors in a large, public rese= arch one university. Four of the professors are scientists or engineers. Each oversees a research group that can include undergraduate, masters', and doctoral students, as well as practicing middle or high school science teachers. The fifth professor is Allan Feldman, who does research in science education and is one of the authors of this paper. He, too, has a research group, which includes graduate students Kent Divoll and Allyson Rogan-Klyve, the co-authors of this paper. The focus of our research is how one learns t= o be a scientist while participating in a research group. In this paper, we focu= s on the ways in which the science and engineering professors conceptualize the research education of their undergraduate and graduate students.

We relied on two form= s of data collection for this study: interviews of the professors and participant observation in professor meetings, research seminars, and project meetings.= We recorded interview and observation data as notes and audiotapes, and select= ed meetings and field trips were videotaped. We developed an interview protoco= l to uncover the science and engineering professors' beliefs about how they educ= ate their undergraduate and graduate students to do research.=

We analyzed the data using the codi= ng of qualitative data (Miles & Huberman, 1994) and through the construction = of understanding inherent in the use of long and serious conversations as rese= arch (Feldman, 1999). Pre-conceived categories for coding were derived from the research literature on graduate education and apprenticeships, while emerge= nt categories were derived inductively from the data, following the methods of= the development of grounded theory (Strauss & Corbin, 1990). We used the qualitative analysis software HyperResearch to help us with our analysis. We also used Inspiration to graphically represent the relationships among the research group members (see Figure 1).

Findings

We organized our analysis along five dimensions: 1) the configuration of their research groups; 2) conceptualizations of students' expertise; 3) conceptualizations of the growth of expertise; 4) learning through apprenticeship; and 5) type of community.

Tightly- and loosely-organized research groups

The configuration of the research groups, their relati= onship to the laboratory, and the relationships of the individuals in the groups varied among the professors according to their research areas. The differen= ces among the type of group that the professors foster can be broken down into = two categories: a) a tightly-organized research group or b) a loosely-organized research group. Two of the professors, Karl, a microbiologist, and Sarah, an environmental engineer, established tightly-organized research groups by maintaining traditional laboratories in which all their graduate and undergraduate students associated with all of their funded projects work together. They each connected lab experiences to the members' outside lives= by holding community building events such as cookouts, dinner parties, baby showers, and birthday parties, to name a few. Both of these professors work= in the same type of physical setting, both have organized their students into research groups that meet on a weekly basis, both require their students to give presentations on their work, and both hold group discussions. In his interview, Karl stated:

I introd= uce them [new research group members] to everybody who is working in the lab. They k= now all the projects that are going on. They are socially integrated by ... we = have a birthday list. The person who has a birthday gets a cake from the person = who just had a birthday, so you don’t have to bring your own cake. We have outings: we have a summer picnic, we have a winter trip, and we have a Christmas party. The newest lab member brings the turkey for the entire Christmas party; it’s at my house. In the lab, people are encouraged = to talk to everybody and when a new person comes I talk to the senior grad students to talk to the new person, to stimulate that, to get that culture going with the new person also. (Interview 6/23/05)

In addition to holding social events, Sarah fosters her tightly-organized research group in the way that = she tries "to team students up together … one student with a more experienced student so that they’ll learn some basic skills in the lab" and meets with her graduate and undergraduate students "in t= eams with students working on common projects (Interview 7/5/05)."

The other two professors have loosely-organized resear= ch groups. Robert, a geologist, works with individual students. His students do field work, bring samples back to the laboratory where they are analyzed on communal instruments to quantify the data, and then they work individually = on their analysis. In his case, while there is a lab, the lab is not used as a place where the group congregates to do work. Douglas, who is a hydrologist, also works individually with students. However, since most of the work that they do is computer modeling, they infrequently make use of a traditional laboratory and therefore do not come together as a group often. Both Robert= and Douglas meet with their students by appointment. Robert’s view of the community that he fosters is very apparent: "I don’t pick studen= ts to work with me unless I see some self-starting initiative. I don’t t= hink that research is really appropriate for everyone to be doing (Interview 6/2= 0/05)."

Rather than create a tightly-organized research group,= " Douglas attempts to build relationships = on an individual basis:

[In reference to projects by undergraduate students] W= ell, normally, it’s you have to go out on a certain date and drill these h= oles or take these measurements or whatever. So it’s actually more relationship building, in other words conversations about sports or whatever takes us to the work at hand. (Interview 7/6/05).

Because their research is not laboratory based, neither Robert nor Douglas has a research group that can = be considered tightly-organized research groups.

One of the important differences between the two types of research groups is th= e center for action. In the tightly-organized research groups in our study, the labo= ratory served as a center of action. While the professors were important in develo= ping the groups and facilitating interactions among the students, student-student interactions took place on a continuing basis as a result of the sharing of= a common space. In the loosely-organized research groups, the professors were= the centers of action. There were few student-student interactions because of little time that students spent together and the fact that students' connections to the group were through the professor.

Figure 1: Web of connections= among participants in research groups

The differences between the tightly-connected and loosely-connected research gr= oups are illustrated in the web of connections shown in Figure 1.The four science and engineering professors form a square in the center of the figure. They = are identified by the letter that begins their names. The lines emanating from = the professors go to their advisees. The lines connecting the students (unlabel= ed ovals) signify connections that were described to us by the students. The figure clearly shows that there are many more connections among students in= the two tightly-connected research groups (Sarah and Karl) than in the loosely-connected groups (Robert and Douglas).

Conceptualizations of students' expertise

While the there were physical differences among the se= ttings of the research groups, and differences in the way the professors and the students interacted with one another, we found no differences in how the professors conceptualized the growth of their students' expertise. All made similar distinctions among undergraduate, master's, and doctoral students. Based on our analysis of our data, we have organized those distinctions in a typology of how students of science function as members of scientific communities. Novice Researchers have little or no experience with scientific research, such as many if not most undergraduates. Our data sugg= est that when undergraduates are allowed to be legitimate peripheral participan= ts in a research group, they are generally seen as temporary members who can, = for example, develop the skills to help maintain the laboratory and collect dat= a, but are not expected to contribute much if anything to the analysis of data= or the creation of new knowledge. In the interviews, the professors talked abo= ut the inability of students at this level to formulate research questions, th= eir lack of laboratory or research skills, and the difficulty the students have= in drawing defensible conclusions from data.

The professors saw graduate students at the master's l= evel quite differently from the undergraduates. In the interviews the professors noted that they did not expect their master's students to be ade= pt at developing research questions. As Sarah told us,

When they [master’s d= egree students] come to work with me I pretty much hand them a project. It’= s a proposal that has been written and that’s funded and that has particu= lar objectives and tasks that need to be accomplished. (Interview 7/05/05)

The professors do expect the= m to have the research skills that allow them to do what is necessary for the research project. The students can also apply the methods that they have learned to new situations. From our analyses of the interviews it became cl= ear that the professors had the expectation that at the end of their studies, master's students would be at least Proficient Technicians. That is, they would have attained the knowledge and skills necessary to become skill= ed practitioners in their field.

All of the professors supervise doctoral students. Karl noted that a PhD indicates that a researcher has "intellectual proficiency" as well as technical proficiency. To Sarah this means bei= ng able to make "a significant contribution to the science and to the engineering." (Interview 7/05/05) Douglas and Robert shared this expectation that doctoral students should be able to formulate their own research questions, to develop new research methods, and to add to the literature. In short, all the professors expected their doctoral students to become Knowledge Producers.=

Conceptualizations of the growth of expertise

The science and engineering professors perceived the typology that we described above as developmental: with appropriate experie= nce and guidance an individual can move along a continuum from Novice Researche= r to Knowledge Producer. An important aspect of the professor's awareness of the developmental nature of participation in the research group is that they ke= ep a watch out for likely prospects whom they nurture along the developmental pa= th from Novice Researcher to Proficient Technician to Knowledge Producer.

The first move along this continuum is from Novice Researcher to Proficient Technician. Although each of the four professors s= aw the first transformation in different terms, they all recognized the transformation. Karl saw this transition as a growth in confidence and the ability to do more independent thinking:

The confidence level …= ; Up front they’re not confident. They’re very conservative and sometimes reactionary where they try to throw everything away and sometimes don’t trust their own data, question everything. (Interview 6/23/05) =

Dougl= as suggested that this transition could take place for some undergraduate students:

Peter is a good example of an undergraduate who became almost like a Master’s student in the sense = of problem solving, and we developed a relationship in which I trusted him, wh= ere I could say, here’s the problem go survey the site, you figure out the details, you solve the little problems, here’s the big issue that nee= ds to be resolved and he went and did it. (Interview 7/6/05)

Sarah recognized the shift f= rom Novice Researcher to Proficient Technician in her master’s students a= nd categorized it in terms of her students' ability to design their own experiments.

The second move along this continuum is from Proficient Technician to Knowledge Producer. Typically, only the doctoral candidates m= ade the transformation to a Knowledge Producer. As was the case with the first transition, each of the professors recognized the second transformation, bu= t in different terms. Three of the professors (Douglas, Sarah, and Robert) prese= nted this transformation in terms of product. Douglas expected his doctoral students:

to demonstrate an ability t= o add something to the literature, that’s sort of the test. If it’s a= dded to the literature, it’s got to be new and have some element of uniqueness. (Interview 7/6/05)

Sarah recognized this transf= ormation in one of her students:

I could see in the journal = articles that she’s written that there was much more discussion, that there’s much more analysis of her data. (Interview 7/05/05)

Although Robert suggested th= at a Knowledge Producer should publish, he elevated this notion:

For the PhD project, I expect that = it will have global implications in their subject. That whatever the finished product is can be put into a context that can be referred to in another par= t of the world, another part of the country, lots of references to parallel problems, parallel sites, here’s what I’ve found and this is wh= at it means in terms of the context of geology or geochemistry. (Interview 6/2= 0/05)

Sarah and Karl also spoke ab= out the intellectual relationship that they expected to have with their students as they became Knowledge Producers. For example, Sarah told us that "Karen used to ask me questions all the time and now I ask her for advice" (Interview 5/5/05) and "Ash has a much better grasp of the literature = than I do" (Interview 5/5/05). Karl expects his students to engage him like= a colleague:

In the end, they should cri= ticize me; they should correct me in what I’m saying because they then become the experts in their niche in their field. (Interview 6/23/05)

All four professors acknowle= dged that they have lofty expectations of Knowledge Producers and that in most c= ases it is only achieved at the end of a doctoral program.

Learning through apprenticeship

It is important to note that none of the doctoral prog= rams with which the professors are associated -- microbiology, environmental engineering, or geosciences -- have courses that explicitly teach students = how to conduct research in that field. As a result, while students did take cou= rses like statistics or laboratory methods, the remainder of their education as researchers was informal and "on the job." Given the informal nat= ure of this education, it is not surprising that none of the professors had giv= en it much thought before we interviewed them. However, as they responded to o= ur questions, it became clear that they all went about it the same way using a= n apprenticeship model, which was how their mentors trained them. The methods that they use = are similar to what you would expect in any apprenticeship – in the early stages the students were heavily supervised and given specific tasks to accomplish, including review and critique of the literature. As students progressed in the programs, the professors became less directive and turned more to questioning students. In the final stages of dissertation research,= the professors expected to engage in collegial conversations with the students about the students' research. Throughout the process the professors tailored their methods to the backgrounds and needs of individual students.

All four of the professors recognized that students in= the early stages of an apprenticeship need close support and direction. For example, Karl told us how doctoral candidates need close support early in t= heir programs:

In the first year [doctoral student= s] need to have their hand held, even if they come with a master’s from someplace else, from a different lab. It should be at least the first half-= year but preferably the first full year that they should be intensively advised = and mentored. (Interview 6/23/05)

Because Sarah, Robert, and D= ouglas work with undergraduate honors students and master's students in addition to doctoral students, they also talked about the close support that those stud= ents need as they begin their research education.

As the students gain expertise as researchers the prof= essors become less directive and more focused on the students taking on a more act= ive role. For example, Douglas described his= work with master’s students and beginning doctoral candidates in this way:=

I think that both masters' and PhD students shouldn't just take up what I might teach them in terms of skills = but to develop on their own skills to search the literature about how to do something, and to come up with a new idea on how to perform some analysis. (Interview 7/06/05)

Toward the end of the apprenticeship period the profes= sors expected their students to develop the ability to develop and guide their o= wn research. Sarah spoke about how she expected to see her students show independence as they begin to work on their dissertations:

When [doctoral students] see the pr= oject in relation to their dissertation they will take the next step of actually going on to develop their knowledge by starting to collect literature and to organize it and make sense out of it. (Interview 7/05/05)

Dougl= as told us as students advanced in their research education he "would exp= ect them to be able to do analysis and interpretation or learn how to do it or figure out how to do it or devise a new way to do it" (Interview 7/06/= 05). Robert told us that students at this level have collegial conversations that are “about the bigger picture” and about “the meaning of = the results that you’ve gotten” (Interview 6/20/05). Karl described= his expectations of advanced students in this way:

They should be independent. If new = grad students come in to the lab they shouldn’t wait for me to tell them to help them, they should approach them and say, "Here, look this is how it’s done," they should want to help others. If there is a meeti= ng in Boston or at Yale or another university they should suggest to me that they would like t= o go there and present. When we go to a meeting together they should stay with me and I'll introduce them to other people because they know by that time how important networking is to all these things. (Interview 6/23/05)

Our interview data shows that while the professors had= not explicitly thought about how they educated their students to be researchers, they were in fact behaving in ways that Lave and Wenger (1991) would call successful apprenticeship teaching. The graduate students learn to do resea= rch by performing tasks in ways that are analogous to how their professors perf= orm them. That is, they are legitimate participants in the research process. In addition, the professors structure the ways that the students participate by assigning them tasks appropriate to their development as researchers.

Communities of practice and epistemic communities

The research groups that we studied were more than a community of practice because of their objective to create and warrant new knowledge. While all group members participated legitimately in the research community, some did so only in the community of practice, while others participated in that community and the epistemic community. We illustrate t= his in Figure 2.

Figure 2: Members of communi= ties of practice and epistemic communities

There were many examples described by the four profess= ors of their students' participation in a community of practice. This included the collection of samples, the analysis of samples using ion chromatography, and the analysis of data using statistical and graphical methods. There was also the acknowledgement among the professors that they, as scientists and engineers, are involved in epistemic communities and that they expected the= ir doctoral students to become part of those communities by learning how to produce and warrant new knowledge; by showing how it relates to the field; = and demonstrating that it has implications that go beyond the research at hand.= To some of the professors, this was indicated among the students by a transiti= on in which the doctoral student becomes the expert, as can be seen in these two comments by Sarah: "Karen used to ask me questions all the time and no= w I ask her for advice (Interview 5/5/05)" and "Ash has a much better grasp of the literature on perchlorate than I do" (Interview 5/5/05). Robert also alluded to the knowledge generation capabilities of doctoral students when he told us of his disappointment of not having a doctoral stu= dent to work closely with, because he found that with his master's students, "the knowledge always flowed from him to them and never vice versa (Interview 6/20/05)."

Implications

In this study we interviewed four professors about the= way they prepare their students to become researchers. We grouped our findings = from our analysis of the interviews into five areas: the configuration of their research groups; conceptualizations of students' expertise; conceptualizati= ons of the growth of expertise; learning through apprenticeship; and type of community. Returning to the introduction to this paper, we believe that the= se findings have implications for the type of teacher learning activities that= are called for by the NSES.

First, if science teachers engage in scientific inquiry along with scientists, the type of research group that they join can affect their learning. As we saw in the tightly-organized research groups students benefited from having several mentors beside their professor. The research groups were made up of students who were at different levels of expertise a= nd doing different types of tasks. Novice researchers benefited from being in frequent contact with Proficient Technicians, and Proficient Technicians we= re mentored by the growing expertise of the advanced doctoral students as they became Knowledge Producers. This was not at all the case for the students in the loosely-organized research groups. They infrequently had contact with m= ore advanced students and so for the most part they relied solely on their professors as their mentors.

Teachers who work in these groups would most likely ha= ve experiences similar to that of graduate students. In the tightly-organized groups they would interact with research group members on a daily basis, an= d be mentored by Proficient Technicians and Knowledge Producers, and in turn, me= ntor other members of the group. Teachers in the loosely-organized groups would = not have close mentorship and therefore depended more on supervision from the professors. Given the time constraints on professors, they would most likely interact infrequently with the teachers. Therefore, teachers in the loosely-organized groups would have less support than those in the tightly-organized research groups.

A second implication comes from the typology of levels= of expertise and roles that people play in research groups. When we say that we want teachers to have knowledge of doing science so that they can teach the= ir students how to do science, are we expecting them to have the expertise of a Novice Researcher, Proficient Technician, or Knowledge Producer? That is, i= f the typology is developmental, what level is the best fit for a K-12 teacher and what level do they need to be at to adequately teach their students how to = do scientific research?

The level of expertise of the teacher also determines = the level that their students would be able to reach. While it seems clear from= the apprenticeship model that a Knowledge Producer can have apprentices who are Novice Researchers, Proficient Technicians, and new Knowledge Producers, a Proficient Technician would only be able to train Novice Researchers and new Proficient Technicians, and Novice Researchers may not have the expertise to even produce others at their own level.

A third implication relates to the time and resources = needed to educate Knowledge Producers. Scientists and engineers are taught to be researchers through apprenticeship programs that can last for five or more years. It would therefore require a tremendous investment in time, money, a= nd other resources to train all science teachers to be producers of scientific= knowledge. It seems unlikely that society would be willing to make such an investment.= In addition, at some level it seems wasteful to invest so much to prepare teac= hers to do work that is outside their practice. If teachers are to be Knowledge Producers it seems more reasonable for them to become members of epistemic communities that generate knowledge about teaching and learning.

We also believe that our research has important implic= ations for the education of new researchers in the sciences. One is that, as we no= ted above, the scientists and engineers who we interviewed had given little tho= ught to how they educate new researchers. Instead they go about doing what is familiar to them, which is based on the experiences they had as new researchers. This suggests that even in the informal learning situations of apprenticeships, teachers teach the way that they were taught. On the surfa= ce this appears to work: new researchers who can contribute to the knowledge b= ase of the field are produced. However, we have little data about those who do = not come through the process as "Knowledge Producers." They not only include the students who do not pass their comprehensive examinations, but = also those who for one reason or another choose to stop at the master's degree l= evel or choose to leave doctoral programs without completing their dissertations. Therefore, we believe that it is important to continue this research line to gather data about students' experiences as well as the perceptions of their professors.

Another implication arises from our observation is tha= t in those fields where there are tightly-organized research groups, students are teaching other students how to be researchers. It may even be the case that advanced students do the majority of instruction in research methods. While interviews indicate that Karl provides his students with some guidance on h= ow to do this, there is the possibility that in most situations the more advan= ced students have little or no guidance on how to be mentors. If we are going t= o continue to use apprenticeships as the primary model for teaching people to be researchers, it may behoove us to provide the mentors, both faculty and advanced students, with instruction in how to be successful mentors.

Conclusion

What does it mean that science teachers ought to know = how to do science? Does it mean that they should be at the level of a Novice Researcher who has been exposed to a community of practice but has developed little of the skills needed to develop and carry out a research project? Or does it mean that a teacher should be at least at the level of the Proficie= nt Technician, who is a skilled member of the community of practice, but does = not participate in the creation or warranting of new knowledge? Or does it mean that for a teacher to adequately teach children how to do science, they mus= t be Knowledge Producers? The answers to these questions would determine the tim= e, money and other resources needed to sufficiently educate science teachers so that they can serve as research mentors to their students. As our research suggests, there is much more to becoming a scientist than can be accomplish= ed in the professional development models currently used, including the Nation= al Science Foundation's Research Experiences for Teachers, if the traditional apprenticeship model is used. This suggests that either we must change the = goal that we have that K12 students ought to learn how to do science so that it = does not require teachers who are at the level of Knowledge Producers or we need= to find new and more efficient ways to help teachers learn how to do science.<= /p>

References

Brickhouse, N. W. (1990). Teachers’ Beliefs about the Nature of Science and Their Relationship to Classroom Practice. Journal of Teacher Education, 41(3), 53-62.

Capobianco, B., & Feldman, A. (2006). Promoting quality for teacher action research: Lessons learned from science teachers’ action research. Educational action research, 14(4), 497-512

Clark, B. R. (1997). The modern integration of research activities with teaching a= nd learning. The Journal of Higher Education, 68(3), 241-255.

Clescerl, L. S., Greenberg, A. E., & Eaton, A. D. (Eds.). (1999). Standard Methods for Examination of Wa= ter & Wastewater (20th ed.). NY: American Public Health Association.

Creplet, F., Dupouet, O., & Vaast, E. (2003). Episteme or practice? Differentiated communitarian structures in a biology laborator= y. Paper presented at the Communities and technologies: proceedings of the fir= st International Conference on Communities and Technologies.

Feldman, A. (1999). The Role of Conversation in Collaborative Action Research. Educational Action Research, 7(1), 125-144.

Fernández-Esquinas, M. (2003). From 'Apprenticeship' to 'Training': An empirical Enquiry into the Preparation of Scientists in Span= ish Academic Science (No. IESA Working Papers Series 0305). Cordoba, Spain: Institute for S= ocial Studies of Andalusia - Higher Council for Scientific Research.

Flick, L. B., Lederman, N. G., & Encohs, L. G. (1996). Relationship between teacher and student perspectives on inquiry-or= iented teaching practice and the nature of science. Paper presented at the Ann= ual Meeting of the National Association for Research in Science Teaching, St. Louis, MO.

Knorr Cetina, K. (1999). Epistemic Cultur= es: How the Sciences Make Knowledge. Cambridge, MA: Harvard University Press.

LaPidus, J. B. (1997). Doctoral education: Preparing for the future. = Washington, DC: Council of Graduate Sch= ools.

Lave, J., & Wenger, E. (1991). Situat= ed learning: Legitimate peripheral participation. Ca= mbridge: Cambridge= University Press.

Lederman, N. G. (1992). Students’ and teachers’ conceptions of the nature= of science: A review of the research. = Journal of research in science teaching, 29(4), 331-359.

Miles, M., & Huberman, M. (1994). Qual= itative data analysis (2nd ed.). T= housand Oaks, CA: Sage Publications.

National Research Council. (1996). National science education standards. Washington, DC: National Association Pr= ess.

Richmond, G. (199= 8). Scientific apprenticeship and the role of public schools: General education= of a better kind. Journal of research = in science teaching, 35(6), 583-587.

Strauss, A., & Corbin, J. (1990). Basics= of qualitative research: Grounded theory procedures and techniques. Newbury Park, CA: Sage Publications.

Wenger, E. (1998). Communities of practice: learning, meaning, and identity. Cambridge, UK: Cambridge University Press.