Research in Science Education

, Volume 43, Issue 5, pp 2009–2034 | Cite as

Introducing Pre-university Students to Primary Scientific Literature Through Argumentation Analysis

  • Marcel Koeneman
  • Martin Goedhart
  • Miriam Ossevoort
Article

Abstract

Primary scientific literature is one of the most important means of communication in science, written for peers in the scientific community. Primary literature provides an authentic context for showing students how scientists support their claims. Several teaching strategies have been proposed using (adapted) scientific publications, some for secondary education, but none of these strategies focused specifically on scientific argumentation. The purpose of this study is to evaluate a strategy for teaching pre-university students to read unadapted primary scientific literature, translated into students’ native language, based on a new argumentation analysis framework. This framework encompasses seven types of argumentative elements: motive, objective, main conclusion, implication, support, counterargument and refutation. During the intervention, students studied two research articles. We monitored students’ reading comprehension and their opinion on the articles and activities. After the intervention, we measured students’ ability to identify the argumentative elements in a third unadapted and translated research article. The presented framework enabled students to analyse the article by identifying the motive, objective, main conclusion and implication and part of the supports. Students stated that they found these activities useful. Most students understood the text on paragraph level and were able to read the article with some help for its vocabulary. We suggest that primary scientific literature has the potential to show students important aspects of the scientific process and to learn scientific vocabulary in an authentic context.

Keywords

Primary literature Scientific literacy Science education Pre-university Argumentation analysis Authentic practice 

Introduction

Traditional pre-university science teaching has been marked by an emphasis on the delivery of scientific information. During the last decades, this has been complemented by an increasing attention to the process of science (Baggott la Velle and Erduran 2007; Kuhn 2010), whereby science is presented not as a static body of knowledge but as a dynamic endeavour in which new facts were actively searched for by means of field specific methods and techniques. Inquiry, as a fundament of modern science, became a major focus of school science in US secondary education (National Research Council 1996), often supported by a pedagogical approach of learning science by inquiry, as outlined by Tamir (1985). Analysis of this science education practice shows that students’ inquiry tasks are in fact simpler and more superficial than those in real science (Chinn and Malhotra 2002). Students’ ‘routines, roles and responsibilities’ in school science inquiry are usually much different from scientists’ (Ford and Wargo 2007). Falk and Yarden (2009) acknowledged this gap between school science and real science and analysed authentic scientific practices in contrast to school science in terms of complexity, coordination practices, argumentative rhetoric and reading and writing. They presented the use of adapted primary scientific literature (APL) as a means to engage high school students in coordination practices as a type of authentic epistemic practice, promoting learning science by inquiry (Falk and Yarden 2009). Primary scientific literature is defined here as reports of original observations, theories or opinions, written for peers in the scientific community. As virtually all attempts to use primary literature in pre-university science classrooms made use of adapted material, we wondered if non-adapted primary scientific literature (PSL) could also serve as authentic study material for pre-university students.

An important aim of present-day science education is providing students with a certain level of scientific literacy, according to a general definition given by the PISA Framework: ‘… the capacity to use scientific knowledge, to identify questions, to draw evidence-based conclusions in order to understand and help make decisions about the natural world and the changes made to it through human activity’ (The PISA 2003 Assessment Framework, OECD, p. 133). This rather general definition fits both the learning of scientific concepts and the learning of socio-scientific issues. In the context of this paper, we solely deal with the concepts of science and science writing. A dozen other definitions of scientific literacy are listed by Norris and Phillips (2003), forming a range that covers knowledge and skills as well as attitudes. However extensive, their list notably lacks the skill of retrieving information from PSL, one of the foremost sources of scientific information for the scientific community. For university students learning this skill it is rather obvious, but in pre-university education science knowledge is typically conveyed through textbooks and to a much lesser extent through media reports of science or review articles. For readers, it is a difficult transition from studying these secondary scientific information sources to PSL. Still, PSL could be employed both to teach content knowledge and to show the process of science (Muench 2000). In 1968 Epstein conducted his experiment (Epstein 1972) using research articles in an introductory biology course for 17-year-old students. Since then there have been many attempts to use adapted primary literature in classrooms (see Falk and Yarden 2009, for an overview), but these experiments have not led to a lasting implementation of the use of primary literature at secondary school level. Norris et al. (2011, p. 4) notice: ‘Like Falk et al. (2008), we have found no evidence of any high school curricula that use PSL to teach science despite the many advantages already described’. The most obvious reason for the absence of this type of literature in these classrooms is perhaps that research articles are notoriously hard to read for novice readers, for their content, vocabulary and compact style: ‘… [students’] early encounters can often be bewildering and humbling’ (Janick-Buckner 1997). Of equal importance is perhaps that teachers not necessarily have sufficient experience reading primary literature themselves. We asked ourselves whether and how PSL could be made accessible to pre-university students.

Among the aspects of the process of science that PSL could reveal to students are the nature of knowledge construction by which new scientific knowledge is deduced from collected data, and such argumentative traits as the way scientists support their claims, and, as a consequence, the tentative character of these claims. ‘Argumentation, … is a critically important epistemic task and discourse process in science’ (Osborne et al. 2004). If argumentation plays such an important role in science, how is it dealt with in pre-university science education? In the current Dutch pre-university examination programmes for biology (Boersma et al. 2007) as well as chemistry and physics various aspects of argumentation similar to the above mentioned are found: ‘drawing conclusions’, ‘describing how scientific knowledge is formed and its validity is judged’. Yet the examination programmes do not specify through which material or by what method these goals should be reached. One tempting possibility is employing PSL so that students can study a complete argumentation in an authentic scientific context.

We examined the possibility to make PSL from biology and the medical sciences accessible to pre-university students through a basic argumentation analysis done by the students themselves as their first approach to the text. In this way, reading and analysing are no subsequent steps but parallel and mutually reinforcing activities, resulting in a form of inquiry reading. This could offer students a new view on scientific inquiry, which is ‘essentially a dialectical process in which one grapples with the ideas, thoughts and reasoning of others often through the medium of written texts’ (Goldman and Bisanz 2002). We hypothesised that an appropriate argumentation analysis framework would provide students a workable tool to gain access to PSL. We designed a teaching strategy around three selected PSL articles to apply the argumentation analysis framework. The aim of the research presented here was to evaluate the use of this framework for introducing PSL to pre-university students.

Theoretical Framework

Reading PSL

In order to make students acquainted with PSL in an authentic context, we tried to shape our teaching strategy in such a way that the activities and the general atmosphere would resemble those in a team of scientists ‘reading the latest research papers’. We will describe our view on the latter shortly.

Whereas reading is often considered a solely individual activity, reading and evaluating PSL is probably not. PSL is supposed to be shared by members of the scientific community not just by convention but primarily because openly discussing the conclusions of a publication is an essential part of the process of knowledge construction. Research groups of scientists consist of highly individual professionals who will read publications in their personal way and form their own opinion on the claims put forward. Subsequent discussion at research group level and at international community level determines the status of the claims until eventually, if ever, consensus is established. During this process, scientists are well aware of the conventions that rule both publishing and discussing scientific findings. This enables them to effectively take part in the discussion. The conventions in a particular field of science encompass article structure, vocabulary, level of conciseness, evidence presentation, etc. The technical skills needed for reading and evaluating PSL are usually acquired by experience and support of expert readers, like senior scientists and teachers. These reading activities take place in every corner of the scientific community and their importance can hardly be underestimated: ‘Scientists, on average, spend well over 100 h (yearly) reading scholarly articles, an indicator that they recognise the importance and value of this activity’ (Tenopir and King 2001).

Reading PSL is naturally learned in an academic setting. Traditionally, a student would move from textbooks to primary literature, tutored by experienced teachers or colleagues, gradually developing their reading skills. PSL calls for an active reading approach since full understanding requires reading, re-reading and—vital to the non-expert—gathering necessary additional information elsewhere, such as the meaning of unknown words. This necessary approach makes reading PSL a form of inquiry reading, being not only a matter of reading skills, but also of reading attitude. In order to further promote the scientific authenticity of an intervention based on PSL, we would have to mimic this academic setting.

The last decades have seen the development of a number of courses for university students to acquire their PSL-reading competence (Janick-Buckner 1997; Kuldell 2003; Mulnix 2003). These approaches were not likely to meet our goals since most of them were designed for undergraduate university students and more often than not focus on the content of the articles rather than their form and general characteristics. A relatively small number of attempts were made to introduce (adapted) primary literature to pre-university students (Yarden et al. 2001; Yarden 2009, for an overview).

The Argumentative Structure of PSL

For research articles in the natural sciences, the so-called Introduction–Methods–Results–Discussion (IMRD) structure has become the norm. Intertwined with this surficial structure is another structure: the argumentative structure that is also subject to writing conventions but less easily recognisable for its dependence on text content rather than global text structure. Bazerman was one of the first to acknowledge the role of persuasion in written communication in science and pointed at the argumentative structure of scientific publications (Bazerman 1988), describing how a claim is presented, explained, supported and discussed in light of experimental data. Indeed research articles ‘seek to be persuasive as well as informative’ (Gillen 2006).

Suppe (1998) examined more than 1,000 empirical research articles from various disciplines and found that they share a number of organisational characteristics, resulting in a common argumentative structure. Phillips and Norris (2009) analysed the argumentative structure of a number of research articles in physics as a sequence of ‘speech acts’, such as motivating their study, explaining observations and challenging alternative interpretations. The authors quote Norris (1992) suggesting that ‘one goal of science education should be to teach students to see the ‘justificatory shape’ within science’ (p. 316). Interestingly, they conjecture that however the specifics of scientific justifications would be too difficult for most non-scientists, ‘they could be taught to grasp the general nature of scientific justifications so that the source of scientific findings would seem less of a mystery’ (Phillips and Norris 2009).

Close reading of a research letter or paper helps reveal the argumentation the author employs to explain and support his claim, but this might also work the other way around: identifying argumentative elements directly reveals the logical structure of the text which leads the way to further understanding of the message conveyed. In this way, the argumentative structure of a paper forms the backbone for the personal understanding and meaning that the reader gradually develops. Since PSL usually means highly structured texts that follow general writing conventions, readers would be able to analyse the argumentative structure by searching for different types of argumentative elements that we presumed to be present in the majority of PSL in the life sciences.

Argumentation Analysis Framework for PSL

The set of different types of argumentative elements together forming the argumentation analysis framework was designed especially for our line of research on teaching students to read, understand and analyse PSL. Our aim was to design a concise and accessible framework that would facilitate studying and analysing PSL without generating new problems. Ideally, students would use it intuitively. Furthermore, it would have to encompass only the most important argumentative types of elements in PSL. In applying the framework to an analysis of a PSL article, a student would only have to identify text fragments as argumentative elements in terms of the framework without further transcription.

We find this rather straightforward procedure of analysing texts also in genre analysis, of which Swales (1990) is the foremost pioneer. In genre analysis so-called rhetorical ‘moves’ are identified that serve a specific role in the text’s persuasive structure. For our goal, an argumentative framework required a somewhat more limited set of types of elements in order to fulfil its pedagogical task of guiding pre-university students into PSL. However limited in number of different types, the identified elements together should give a complete image of the article’s overall line of reasoning, from the very reason to do the research through to the implications. This is why we did not adopt an argumentation analysis model such as the one suggested by Kelly et al. (2008) that focuses on the epistemic levels from data to the final claims of a paper. This model enables the user to extensively evaluate the evidence provided by the authors to support their article’s claim by distinguishing several epistemic levels in various statements. This approach would probably be too difficult for pre-university students as we would rather help them ‘find their way’ in PSL.

One of the most frequently applied models in science education research is Toulmin’s argumentation model (Toulmin 1958). It has mostly been used to evaluate the argumentative structure of students’ writing. See Sampson and Clark (2008), for an overview, and Venville and Dawson (2010) for a recent example. One of our reasons not to use Toulmin’s model is that it does not provide a type of element that could directly serve as a research question or research goal, which we see as the main directive entity of the typical IMRD article. A research question in itself may not be strictly argumentative in nature, but as the answer—the main conclusion—calls for ample justification the question is an indissoluble part of the article’s argumentation as a whole. Importantly, the research question guides readers towards the conclusion of the article and its various supports (Brill and Yarden 2003). For novice readers, this may provide a major advantage. Another problem concerning Toulmin’s model is that the warrant is very demonstrable in for instance legal matters, but hardly ever in PSL. In science we often deal with ‘tacit’ warrants that are not easily discernible for the novice reader. Also, the conciseness of Toulmin’s framework renders it difficult to describe a complex argumentation structure in a single scheme. One could link or embed several schematic ‘Toulmin units’ so that the claim of one unit acts as a ground in another. Thompson (1993), for instance, used two coupled units to represent the proof lines in just the Results section of an article by Nobel laureate chemist Kornberg. For our pre-university target group we preferred a less general framework, tailored to fit PSL so that in most cases no multiple units would be needed. Finally, while Toulmin’s model offers an analytical tool for describing argumentations in general, our framework should also act as a heuristic for students, a pedagogical tool that would help them to fathom specifically scientific argumentation.

We now describe the most important features of the framework used in this study. We distinguish seven types of elements. The motive explains the reason for conducting the investigations and frames the research for the non-expert. The motive is used to introduce the objective, which explains in what direction the efforts of the researchers will be leading. The objective is formulated as a research goal or research question and is eventually addressed by the main conclusion. The main conclusion is the most important outcome of the research and serves as an answer to the objective. The main conclusion leads to one or more implications, as its possible consequences. An implication explains new possibilities or what future research needs to be undertaken.

These first four elements, identified in an article, together form a single line of reasoning within the argumentative structure of the article. Yet the main conclusion still has to be justified by the researchers. The acknowledgement of the vital relation between these justifications and the main conclusion is central to scientific literacy. The researchers state these justifications from their own empirical data but they may also refer to relevant findings in other research, common knowledge or straightforward logic. We call all ways of justification of the main conclusion simply supports. The main conclusion may be challenged by counterarguments. Although these argue against the main conclusion, they are advanced by the authors themselves. Counterarguments take the form of self-criticism; they apparently weaken the authors’ case, as they cast doubt upon the main conclusion of the article. Usually counterarguments are explicitly refuted by refutations, by which the authors cope in advance with potential criticism and thus strengthen their case. Supports, counterarguments and refutations are logically positioned in between the objective and the main conclusion, indirectly connecting the two. Each element in these three categories may be supported in an article by underlying, or secondary, supports. All supports, counterarguments and refutations together form what we call the underpinning of the main conclusion.

The first two element types, motive and objective, are, as a rule, located in the Introduction section of the article. Two other element types, the main conclusion and the implication(s), are found in the Discussion section, although sometimes authors may give their main conclusion away in the Introduction or even in the title of the article. Supports, counterarguments and refutations can be found in both the Discussion and Results sections. Whereas a specialised scientist may first search the Results for data that point towards the main conclusion and proceed to the Discussion to see how the authors use them to justify their main conclusion, we teach our students to first find those supports that are (logically) most proximate to the main conclusion, the primary supports. These in turn may refer to the underlying secondary supports, forming support chains ending with the basic supports formed by the data in the Results section, such as tables or figures. The same goes for counterarguments and refutations. This approach of searching backwards is easier for students since the more basic the data, the more technical they tend to be.

The reason we used this rather formal structure and explicit set of terms and relations is that an adequate vocabulary would greatly facilitate communication with and between students. Also, students fitted with some theory would be able to receive instructions in a nuanced vocabulary and to understand explicit feedback from their teacher (Van Gelder 2005).

Research Questions

We chose a design-based research approach with the argumentation analysis framework employed in an intervention, embedded in the curriculum of an existing school subject, following the guidelines given by Collins et al. (2004). This would enable us to test the validity of our approach in a natural educational setting. Wherever relevant we would distinguish between science and non-science major students. The intervention design will be described in more detail below. Since one of the major educational aims of the intervention was to teach students to analyse the argumentative structure of PSL, we were primarily interested in how well, after the intervention, students could analyse the argumentation in an article that was completely new to them. This led to our main research question (1): To what degree are students able to identify the motive, objective, main conclusion and its underpinning and implication in a PSL article after the intervention?

The argumentation analysis was primarily meant to be a means of access to the article as a whole but would probably influence the students’ ideas on what is important and what not in an article, because much attention would be focused on the identified elements featured in the analysis. We hypothesised that if students were to compose a summary of the paper they would choose what they felt was most important in the text. A detailed analysis of the differences between their summaries before and after the intervention would show how students’ ideas of what is important in an article changed due to the intervention. Moreover, a detailed analysis of the composition of their summaries could reveal their level of understanding. This led to research question (2): What argumentative elements are present in students’ summaries of a PSL article before and after the intervention?

During the intervention, we would closely monitor what difficulties students encountered reading and studying one of the articles in order to establish its accessibility with respect to textual content and vocabulary. For content we let students match each paragraph of the Introduction section of a PSL article with a given summary, to answer research question (3a): To what extent are students able to understand the content of the Introduction section of a PSL article? We chose the Introduction section because here the context, the theoretical framework and the aims of the research are presented, with most of the potential problematic concepts and vocabulary. The Methods section would take much longer to understand in detail. The Discussion section and most of the Results section were extensively studied by participants during the teacher-led analysing activities.

As for vocabulary, we wanted to know to what extent students’ vocabularies are matched for the demands of the article since the text would probably contain terms students were unfamiliar with. We checked this by letting students write down all the terms in a PSL article they did not know, in order to answer research question (3b): To what extent are students’ vocabularies sufficient for the vocabulary of a PSL article?

Intervention Design

The intervention was planned as a means of testing the applicability of the argumentation analysis framework described above to enable pre-university students to access PSL. We composed a curriculum unit especially for this purpose, consisting of a series of initially teacher-centered lessons based on three PSL articles. We chose a teacher-centered approach to safeguard the first reading and analysing activities. As outlined above we had few precedents to follow. A number of studies have been published that describe good practices at university level, such as Janick-Buckner (1997), Kuldell (2003), Peck (2004), Almeida and Liotta (2005) and Hoskins et al. (2011). These practices involved guided reading followed by some kind of group discussion. We adopted these ideas with some modifications. In our study, the guided reading would take place in the classroom, so that help was always at hand. Also, the group discussions were partially substituted by individual or team assignments to facilitate data collection since the abovementioned studies mainly provided classroom observations and students’ self-evaluations.

As can be seen in the intervention outline (Fig. 1) below, the intervention moves from highly structured assignments to increasingly teacher independent activities. The implementation of this gradual withdrawal of the teacher’s help and other aids, was based on Cognitive Apprenticeship as described by Collins et al. (1991): ‘Apprenticeship was the vehicle for transmitting the knowledge required for expert practice … (cognitive apprenticeship) … is a model of instruction that goes back to apprenticeship but incorporates elements of schooling’ (p. 6). As stated above, we see reading PSL as an expert practice that is usually learned from experienced university teachers or colleagues. In Collins et al.’s view, this works well as the instructor or mentor not only conveys technical information but also actually shows their thoughts performing the analysis of an article. For this reason, we chose for mainly oral teaching in which the teacher enacts the analysis, leaving more and more to the students as the intervention develops.
Fig. 1

Schematic overview of the intervention. Each box represents an activity or set of related activities. Learning activities are left aligned; other activities such as the questionnaire are right aligned. Activities that serve as data sources for this paper are in bold print

During the beginning of the intervention, after a pre-test, students are closely guided through the text of the first article, assisted by introductory texts, a glossary and a detailed explanation of the text (scaffolding). The analysis of the article is partly shown (modelling) and partly carried out together. In the following part, students receive less detailed instructions for studying and analysing the second article (fading) but are invited to inquiry reading by purposeful assignments. The teacher would always be available during any activity, responding to questions from students by referring them to each other or to the available materials to help them find an answer to their question themselves. When they were at a loss, the teacher would provide them with slightly more substantial hints to get them going again (coaching). In the last part, students are expected to study and analyse the third article without any guiding materials or teacher assistance (last stage of fading). This final assignment breaks down into an individual part and a team part, the latter mainly spent on evaluating team members’ analyses and preparing a presentation of their combined team analysis.

The curriculum unit is composed of 30 activities, with a total duration of 10 h. Activities take 5–30 min each, except for the final assignment in two parts, that is planned to take 2 h in all. An outline of the curriculum unit is given in Fig. 1. Each of the 19 boxes represents an activity or a set of related activities. Boxes on the left side of the outline indicate learning activities. Additional activities were introduced for research purposes only; they are on the right side in bold print. Those learning activities that provide research data for this study as well are also in bold print.

Before each activity students receive teaching materials, such as copies of the PSL articles to be studied, theory sheets, instructions and assignments and add them to their binder. All binders are collected by the teacher at the end of each lesson and no homework is given.

Research Articles

Muench (2000) proposed a number of valuable criteria for primary literature to make useful material for undergraduate courses. In choosing the three articles, we followed a number of Muench’ criteria, adapted and completed for our purposes: (1) the connection between data and main conclusion should not contain too many consecutive steps, (2) the subject of the paper should be thematically appealing to students, and (3) the article should be reasonably concise, so that students could quickly get a picture of the text as a whole. To avoid any comprehension difficulties by the language barrier, the papers were translated into the native language of the students. We made no other alterations, carefully preserving all other characteristics including style and layout. The papers used are:
  1. 1.

    Warren (1983). A research letter that for the first time shows the existence of colonies of possibly pathogenic bacteria in the human stomach. Its argumentative structure is straightforward. The absence of Methods and Results sections leave the reader undistracted from the argumentation, which runs through the full length of the letter.

     
  2. 2.

    Campbell et al. (2009). A research paper which shows that chimpanzees actually empathise with animated conspecifics, just like humans do. Here, the argumentative structure is more complex and accompanied by an extensive Results section but carefully expounded.

     
  3. 3.

    Vinther et al. (2008). A research paper that argues how fossilised feathers contain clues for their original colour. Its argumentative structure is clear in that it strongly relies on the apparently decisive experimental results provided by the authors.

     

Each of the three publications features a subject that is most likely completely new to our students but at the same time easily conceived or visualised, being a combination of two well-known concepts: stomach and bacteria, apes and yawning and fossils and feather colours, respectively.

Implementation

Participants

The participating students were 83 year-11 pre-university students with an age range of 16–19 years. There were 37 female (45 %) and 46 male students (55 %). Of the students, 63 % were science majors and 37 % non-science majors. Both streams had shared the same curriculum in years 7–9. All science majors had had 2.5 years of chemistry education and 2–3.5 years of physics and biology. Non-science majors had 1 year of chemistry and 2 years of biology and physics. Participants were not selected but grouped by the school timetable in four mixed (gender and stream) groups, as shown in Table 1. In both the science and the non-science streams, all students are prepared specifically for further education at university level. We included the non-science students in our study since differences in performance or attitude could be indicators of how students deal with PSL. Students had no prior experience with PSL, except for three students who reported having read a number of articles from the field of economics.
Table 1

Characteristics of the participants (numbers (n) per characteristic)

Group

Number

Gender

Major

Teacher

Male

Female

Science

Non-science

1

26

17

9

21

5

A

2

23

10

13

12

11

A

3

17

13

4

11

6

B

4

17

6

11

8

9

B

Total

83

46

37

52

31

 

The intervention was executed by two male teachers. Teacher A, who had 24 years of teaching experience, is also the researcher. He designed the curriculum unit and is first author of this paper. He has an academic background in biology. Teacher B, with 12 years of teaching experience, has an academic background in chemistry. Both teachers had ample experience in reading PSL in their respective disciplines during university training. Students received uniform instructions, in accordance with a detailed teachers’ guide. During the intervention the teachers frequently met to discuss the progression of the intervention.

Educational Setting

The intervention was implemented in a school setting, as a teaching sequence within participants’ compulsory school subject ANW, comparable to the UK school subject Science in Society. This school subject provides background for—and links between the traditional science subjects astronomy, physics, chemistry, biology, earth science and medical science. The themes are usually presented from a historical and/or sociological point of view and include socio-scientific issues. We chose this subject as the context for our intervention because of its flexible curriculum and because it is the only science subject at this age level that is followed by both science and non-science majors. The school is an average Dutch comprehensive school of approximately 1,400 students situated in a town of 160,000 inhabitants.

Data Collection and Analysis

For answering our research questions, we used data from a number of sources (see Fig. 1): the Summary Assignment (pre- and post-test), the Matching Assignment, the Glossary Assignment and the Final Assignment. Some additional data were acquired from the questionnaire. We will describe the data collection and analysis below, following the order of the research questions.

Final Assignment: Identifying Argumentative Elements

At the end of the intervention, students made the Final Assignment (individual part).

After reading Vinther et al.’s (2008), students individually answered the following questions: (a) What was the motive for doing this research? (b) What was the objective of this research? (c) What was the main conclusion the researchers drew on the basis of their results? (d) Indicate how the researchers underpin the main conclusion. (e) What are the implications of this research, according to the researchers?

The students were instructed to quote the identified text fragments either literally or by writing down the beginning and ending of the text fragment. For analysis, we (authors 1 and 3) compared students’ answers with the agreed answers of authors 1 and 3, in accordance with our analysis given in Table 2. The interrater agreement for the analysis of students’ answers was 0.90 (Cohen’s kappa). Each fragment stated correctly was scored. We accepted the statement of an element as a correct identification in case of a verbatim quote but also in paraphrased form as long as the original length and location were recognisable. Quotes from the Abstract instead of the main body of text, albeit correct with respect to content, or those with an unclear original location were rejected. The term ‘underpin’ in question (d) referred to any support, counterargument or refutation. The data source were students’ answers, written on a two page question form. During the assignment students had access to all materials received and used during the intervention.
Table 2

Percentage of pre-university students (total n = 79, of which 50 science and 29 non-science majors) who successfully identified a text fragment as a certain argumentative element in Vinther et al. (2008)

Text fragment (location, page; I = introduction, R = results, and D = discussion)

Argumentative element

Science (n = 50; %)

Non-science (n = 29; %)

Total (n = 79; %)

These bodies, which delineate the feather outline, were interpreted as fossilised feather-degrading bacteria (I, 522)

Motive

62

52

58

We decided to reinvestigate these structures to determine whether they might represent melanosomes rather than bacteria (I, 522)

Objective

88

97

91

… based primarily on their similarity to these structures in modern feathers (D, 524)

Support 1

38

31

35

The oblate structures from the dark bands in the fossil Crato feather are strikingly similar in size, shape and orientation to eumelanosomes from the barbules of … a Redwinged Blackbird … and … other modern bird feathers (R, 524)

Support 1.1

20

17

19

(Reference to) Fig. 1 (R, 522–524; D, 524)

Support 1.1.1

20

14

18

There is no reason for bacteria to be preserved on the dark bands alone and not on the light bands of the fossil feather (D, 524)

Support 2

46

34

42

Rather, … the presence of eumelanin in dark feathers makes them less prone to bacterial decay than white feathers (D, 524)

Support 2.1

54

34

47

Many melanin-bearing structures are preserved in fossils (D, 524)

Support 3

8

3

6

The eye of the Eocene bird…appears to preserve retinal eumelanosomes based on a comparison with a modern bird … (D, 524)

Support 3.1

12

3

9

(Reference to) Fig. 2 (R, 523–524; D, 524)

Support 3.1.1

10

3

8

Similar structures have been reported in the eyes of fossil fish from the Cretaceous of Spain … (D, 524)

Support 3.2

6

14

9

(Similar structures) … from the Eocene of Grube Messel … (D, 524)

Support 3.3

2

14

6

Structures preserving the fur of mammals from Messel are strikingly similar to melanosomes … (D, 524)

Support 3.4

6

7

6

It has also been suggested that the organic imprint of the integument of ichthyosaurs could be composed of melanocytes … (D, 524)

Support 3.5

0

0

0

Fossil squids preserve the ink sac as an organic mass composed of fossilised eumelanin granules … (D, 524)

Support 3.6

2

0

1

It is likely that melanosomes are the source of the carbon in fossil feathers, as the eumelanin inside the melanosomes remains even after the keratin that encloses them has degraded (D, 524)

Support 4

26

34

29

… indicating that these colour bands are [not] preservation artefacts (R, 522)

Counterargument

2

3

3

The margins of the bands match isochronic sections in melanin pigment patterning in modern feathers … (R, 522)

Refutation

4

3

4

We interpret the oblate structures in the fossil feathers as fossilised eumelanosomes … (D, 524)

Main conclusion A

36

28

33

Our observations indicate that the structures on fossil feathers previously reported as bacteria are melanosomes … (D, 524)

Main conclusion B

46

41

44

Our discovery of melanosomes in a fossil feather therefore opens up the possibility of predicting feather colour in ancient birds and perhaps in other theropod dinosaurs, with obvious implications for understanding their ecology and behaviour (D, 524)

Main implication

88

66

80

Feathers in dinosaurs may also preserve melanosomes, although they have yet to be investigated (D, 524)

Minor implication 1

10

7

9

… their [phaeomelanosomes] preservation potential remains to be assessed. Their preservation in fossil feathers would also be of interest as they produce rusty red to buff yellow colours (D, 524)

Minor implication 2

18

14

16

Thus, the preservation of eumelanin may be important in the preservation of soft tissue outlines in animal fossils from a number of localities (D, 524)

Minor implication 3

4

3

4

The first two columns depict authors’ analysis of the argumentative structure of the article. In the second column, the numbers show how support relate to each other: support 1, for instance, is supported by (secondary support) 1.1, which is in turn supported by 1.1.1. Together these three supports form a support chain

Students’ Summaries Before and After the Intervention

To measure students’ progression we determined which argumentative elements are present in students’ summaries of Warren (1983), at the beginning of the intervention (Summary Assignment pre-test) and right before the Final Assignment (post-test), respectively. We used a summary assignment as a pre- and post-test because we could not give a genuine argumentative analysis assignment before the intervention. It would be almost impossible to do a baseline assessment of students’ analysing skills since participants first had to learn the vocabulary of the framework and the constituent concepts. Instead, we checked which argumentative elements were present in participants’ summaries before and after the intervention, as an indicator of their changed ideas on what is important in the publication.

During the pre-test, students received a copy of the research letter by Warren (1983) accompanied by a short introductory text and a glossary. Students were told to read the introductory text and Warren’s letter, using the provided glossary, and make a summary of the letter in about 50 words. For this assignment, they got 30 min. During the post-test, on average 10 weeks later, the publication and the procedure were the same, except for that reading the introduction beforehand was not obligatory. As a data source, we used the summaries that the students handed in.

Authors 1 and 3 agreed upon which text fragment represents a certain argumentative element. The Warren paper features—in our analysis—two main conclusions, an implication, and 20 other underpinnings (supports, counterarguments and refutations) that relate to one of both conclusions (Table 3). We counted the frequency of each argumentative element in students’ summaries in both the pre- and post-test. If a specific fragment that was not included in our analysis, was present in more than 10 % of participants’ summaries, we mention these in our results. The interrater agreement for the analysis of students’ summaries was 0.92 (Cohen’s kappa).
Table 3

Percentage of students using each text fragment in their summaries of Warren (1983)

Text fragment analysed as argumentative elements

Percentage of students choosing the element

Pre-test (n = 75)

Post-test (n = 76)

Elements supporting or otherwise associated with main conclusion 1

A … show bacterial colonisation of their stomachs

15

16

Ba… a colonisation remarkable for the constancy of both the bacteria involved and the associated histological changes

4

28

Ca … during the past three years I have observed small curved and S-shaped bacilli in 135 gastric biopsy specimens

52

75

Da … the bacteria were closely associated with the surface epithelium, both within and between the gastric pits

16

32

Ea … they were difficult to see with haematoxylin and eosin stain

0

8

Fa … but stained well by the Warthin-Starry silver method …

17

25

Ga (figure)

0

0

H … the constant morphology of these bacteria and their intimate relationship with the mucosal architecture contrasted with the heterogeneous bacteria often seen in the surface debris

3

0

I … while the curved bacilli were on the epithelium beneath, closely spread over the surface …

3

1

J … that they are present in about half of our routine gastric biopsy specimens in numbers large enough to see on routine histology

11

5

Ka … these bacteria were not mentioned in two major studies of gastrointestinal microbiology …

3

18

La … their unusual atmospheric requirements and slow growth in culture …

5

14

M … the bacteria grow in close contact with the epithelium

7

9

Na … presumably near the neutral end of this gradient

12

18

Oa … are protected by the overlying mucus

27

30

Main conclusion 1a … the stomach must not be viewed as a sterile organ with no permanent flora

20

55

Elements that most directly connect the bacteria to gastritis, supporting main conclusion 2

P … when there was no inflammation bacteria were rare

39

21

Q … the curved bacilli were almost always present in ACG, often in large numbers and often growing between the cells of the surface epithelium …

51

28

Main conclusion 2a … bacteria in numbers sufficient to see by light microscopy are closely associated with an active form of gastritis, a cause of considerable morbidity (dyspeptic disease)

40

53

Elements that correspond, with respect to content, with the implication

R … that they are almost unknown to clinicians and pathologists alike

31

22

Sa … how the bacteria survive is uncertain

32

32

T … the identification and clinical significance of this bacterium remain uncertain

19

13

Implicationa … these organisms should be recognised and their significance investigated

29

38

Text fragments not considered vital argumentative elements in our analysis but frequently mentioned by participants

… found actively growing in the stomach is Candida

16

13

… they were seen most consistently in the gastric antrum

39

24

aFragments that were highlighted during the intervention. No motive nor objective were present in the text

Matching Assignment: Accessibility

To elucidate to which extent students understand the content of the Introduction section of the research paper by Campbell et al. (2009), students were given the Matching Assignment. They read the original Introduction section of the article and were assigned to match each of the six paragraphs of this section with one of six short summaries of the paragraphs. The summaries had been made by the first author in a slightly colloquial style to avoid new reading difficulties. Quotes were kept to a minimum to prevent students matching by choice of trigger-words instead of content. Their length was not proportional to the length of the corresponding paragraph. The summaries were presented in semi-random order.

For comparison, here, we show as an example two of the six paragraphs of the Introduction section with their summaries.

Paragraph 4 (Campbell et al. 2009, p. 2):

The next step is to determine whether chimpanzees identify with animations, thus addressing the second question above. We tested whether chimpanzees show contagious yawning in response to animated chimpanzee yawns. There are both theoretical and empirical links between contagious yawning and empathy. Lehmann (1979) considered yawning an ‘affective expression’ dependent upon empathy. According to the PAM, contagious yawning is controlled by the same mechanism that makes emotions contagious (Preston and de Waal 2002). Empirical evidence comes from the findings that individuals who possessed more schizotypal personality traits performed less contagious yawning (Platek et al. 2003), and contagious yawning was greatly reduced, and may even have been absent, in children with ASD (Senju et al. 2007; Giganti and Esposito Ziello 2009). In both schizotypy and ASD, empathy may be impaired, although Senju et al. (in press) suggest that attention may also be an issue for children with ASD.

Our summary of praragraph 4:

In this study, we tested whether chimpanzees not only recognise animated faces as faces but also empathise with them. Other research mentioned here shows that contagious yawning is a measure of empathy in humans.

Paragraph 6 (Campbell et al. 2009, p. 2):

Anderson et al. (2004) found that two of six chimpanzees yawned more in response to videos of chimpanzees yawning than to control videos. The population-level statistic was non-significant, which is not surprising given the small sample size. We presented 24 chimpanzees with three-dimensional computer-animated chimpanzees yawning or displaying control mouth movements. We hypothesised that if the chimpanzees identified with the animations, then they would yawn more in response to the yawn animations than the control animations.

Our summary of paragraph 6:

‘There has been done similar research, but we will take it on a bit more extensively’.

The assignment took 10 min. For analysis, we determined how frequently students matched paragraphs correctly with their summaries. To reveal why students made a particular one-to-one swap, we interviewed three students. We did not tell them they had made this particular mismatch and gave them the two paragraphs and the two summaries without telling the key and asked them to match them again and explain what their reasons were. After that we asked them to try and remember how they had reasoned when they had done the assignment the first time.

Glossary Assignment: Vocabulary

To reveal students’ prior knowledge of the vocabulary used in the research paper of Campbell et al. (2009) we gave students the Glossary Assignment in which they made their own individual glossary.

Students read the whole text of Campbell et al. (2009) and individually made a list of terms they did not know or understand. Later, they had to look up the meaning of the terms with the help of a dictionary and, if necessary, of their teacher. As a data source the list of unknown terms of each student was collected.

For analysis, we counted for each student the number of unknown terms and for each term how many students reported it as unknown. We also determined the frequency of these terms in the article’s text. For a valid analysis of differences between science and non-science majors, we focused on the terms listed by at least ten students.

Questionnaire

To evaluate the intervention a validated Likert-scaled questionnaire was administered to participants: part A after the post-test, to assess students’ views on the intervention so far, and part B at the very end of the intervention, to assess their views on the final assignment and on the intervention as a whole. The questionnaire posed statements on how difficult, positive (e.g. pleasurable or fun) and informative students found the activities of the intervention. Other items concerned the applicability of what they had learned and how useful they found the intervention as a whole. The questionnaire was printed on A4 paper with a key on top of each page, stating: 1 = totally agree, 2 = agree, 3 = neutral, 4 = disagree, 5 = totally disagree and NA = not applicable. The last choice was to prevent students that had been absent during an activity from still giving their opinion. Examples of statements in the questionnaire:
  • For the Glossary assignment: I found looking up and writing down the meaning of the unknown words in Campbell et al. (2009) instructive.

  • What I have learned about the various types of argumentative elements (Main conclusion, Objective, Motive, Implications, Supports, Counterarguments and Refutations) I will be able to use after this lesson series.

The questionnaire had been validated with respect to textual and organisational aspects by two students from the group; these did not take part in the actual questionnaire administration.

Results

The intervention was executed from January to June 2010. The total duration of the activities as planned in the teacher’s guide was nearly 11 h. The actual execution varied from 16 to 17 1-h lessons due to organisational issues and to the fact that for data validity some activities could not be split in two and had to be postponed to the next lesson.

Students reported never to have heard of any content of the articles, although by the time of the intervention the findings of Vinther et al. (2008) and, to a lesser extent, also Campbell et al. (2009) already had made their way into popular scientific magazines, newspapers and internet reports.

Final Assignment: Identifying Argumentative Elements

The percentages of students that identified a certain text fragment in Vinther et al. (2008) as a specific argumentative element type during the Final Assignment are presented in Table 2. The elements motive, objective, one or both instances of the main conclusion and the main implication were recognised as such by 58, 91, 73 (three students identified both) and 80 %, respectively. These types of element can often be found by certain signal words, as had been explained to the students in their handout with identifiers. The signal words in this article were: ‘… decided to reinvestigate …’ (objective), ‘We interpret …’ (main conclusion A), ‘Our observations indicate …’ (main conclusion B), and ‘Our discovery … opens up the possibility …’ (main implication).

Of those students that could not find the objective (9 %), the main conclusion (27 %) or the main implication (20 %), a majority still produced answers with a content or wording related to the type of element they were trying to identify. In contrast, of the students unable to find the motive (42 %), a large majority mainly mentioned implications instead or failed to clearly refer to any text fragment. Apparently, of these four types of element, the motive was not only the most difficult to correctly identify but also the most easily confused with another type of element, the implication.

Compared with the elements mentioned above, the supports and other elements were identified to a far lesser extent. The highest frequencies among these were found for supports 1 (35 %), 2 (42 %), 2.1 (47 %) and 4 (29 %) (see Table 2). They are located close behind one of the two instances of the main conclusion. Support 1 was in fact regarded by another 14 % of all students as part of main conclusion A. Support 4, located right after main conclusion B, was mentioned by 44 % of the students that found main conclusion B but not A, while it was mentioned by only 12 % of the students that found main conclusion A but not B. The only counterargument and its refutation were somewhat concealed in this article, being in reverse order and located in the ‘Results’ section.

Science majors found most of the elements more frequently than non-science majors. On average the frequencies for the science majors were 16 % higher than for the non-science majors, a significant difference (Wilcoxon signed-rank test; Z = 1.979, p = 0.048.). This difference varied somewhat with the type of element: the average frequency for the science majors of the element types of motive, objective, main conclusion and implication together was 13 % higher than for the non-science majors. Of the primary supports 1, 2, 3 and 4, the average difference was 16 %. Secondary supports, supporting the primary ones, were found by on average 25 % more science majors than non-science majors.

In the questionnaire, 45 % of the science majors reported this assignment as a positive activity against only 18 % of the non-science majors.

Students’ Summaries Before and After the Intervention

To elucidate students’ progress during the invention, students had to read the paper by Warren (1983) and make a summary both before and after the intervention. We analysed which argumentative elements were written down into students’ summaries. Our analysis and the frequency of these elements in students’ summaries are given in Table 3. The argumentative elements that were identified and discussed during the lessons (between pre- and post-test) are indicated by an asterisk in the first column.

The average length of the summaries went down from 68 words for the pre-test to 59 for the post-test, nearing the proposed length of ‘about 50 words’ in the assignment. In the mean time, the number of elements in these summaries went up from 4.9 to 5.8 on average. Of the 14 elements identified and discussed during the intervention, 12 went up in frequency and 2 did not change. Thirteen of these were explicitly highlighted in an analysis of the underpinning of the first main conclusion of the paper. The 14th element, main conclusion 2, which received relatively little attention during the intervention also rose in frequency from 40 (pre) to 53 % (post). Two text fragments not considered to be argumentative elements connected to either of the main conclusions frequently appeared in students’ summaries: the notion that the observed bacteria were ‘most consistently found’ in the antrum of the stomach, and the remark that the yeast Candida is sometimes found dwelling in the stomach. The text fragment of the antrum as a preferred location for the observed bacteria was not connected to any other information in the text and could as well be left out of the summary. The presence of the Candida however, can be interpreted as indirectly supporting main conclusion 1, namely that permanent flora in the stomach is indeed possible. This connection was never explicitly explained, nor in the text, nor by the teachers. While the antrum text fragment went significantly down in frequency, the relevant Candida fragment remained nearly constant.

Additional information on choices made by students can be found in Table 4. The number of students that did not include any of both main conclusions in their summary shrunk from 51 to 28 %. The number of students that included both main conclusions went up from 11 to 36 %. Main conclusion 1, of which the underpinning was studied extensively during the intervention, was accompanied by at least one of its associated argumentative elements in 87 % of the cases in the pre-test and in 90 % in the post-test. For main conclusion 2, these figures went down from 60 to 25 %. In the pre-test, 81 % presented supports for a main conclusion in their summary without incorporating this main conclusion itself; in the post-test, 49 % did so. In contrast, 17 % put a main conclusion in their summary without including any of its associated supports; in the post-test, this frequency had risen to 42 %. All in all the results from this summary assignment show an increased focus on the main conclusions and an increased association between the conclusions and their respective supports.
Table 4

Percentages and numbers of students that mentioned main conclusions 1 and/or 2 in their summaries of Warren (1983) (pre and post), subdivided (numbers only) according to whether they included at least one support, counterargument or refutation associated with the mentioned main conclusions

Main conclusion(s) present in students’ summary

Percent

Number

Pre (total n = 75)

Percent

Number

Post (total n = 76)

Numbers of students that give at least one support, counterargument or refutation associated with main conclusion, number(s) …

Numbers of students that give at least one support, counterargument or refutation associated with main conclusion, number(s) …

1

1 and 2

2

None

1

1 and 2

2

None

Only main conclusion 1

9

7

3

4

0

0

20

15

8

5

2

0

(In any case) main conclusion 1

20

15

4

9

1

1

55

42

27

11

2

2

Only main conclusion 2

29

22

7

12

0

3

17

13

8

3

1

1

(In any case) main conclusion 2

40

30

8

17

1

4

53

40

27

9

1

3

Both main conclusions 1 and 2

11

8

1

5

1

1

36

27

19

6

0

2

Neither main conclusion 1 nor 2

51

38

15

19

4

0

28

21

10

9

0

2

Matching Assignment: Accessibility

Of all 75 participating students, 45 correctly matched all paragraphs with their summaries in the Matching Assignment; 37 % of the students stated in the questionnaire that reading the Introduction section of Campbell et al. (2009) was (very) easy while only 21 % found it (very) difficult. The paragraphs most difficult to match appeared to be paragraphs 4 and 6, with 20 and 27 incorrect links, respectively. Ten students directly swapped these two, i.e. they linked paragraphs 4 and 6 to each other’s summaries. Paragraph 4 presents the objective of the article and explains the validity of the chosen approach by referring to other research from another field. Paragraph 6 explains how the investigation is related to other research within the same field. Some overlap in the content of both summaries may have misled these students. This was further investigated during an interview with three students, in which they all mentioned ‘similarity’ between paragraph and summary as their main reason to choose this incorrect match. Confronted with the same assignment again, two of the three elaborated on this, giving examples that showed a considerate level of detail in their observations: ‘… that results were not significant.’, ‘… made a comparison between their own and the previous research.’, and ‘In paragraph 4 they rather deal with the cause of yawning …’, indicating that their incorrect choice was a result of inaccurate reading rather than misunderstanding. In general, the message conveyed in each paragraph as described in the summary, was clearly recognised by a large majority of students.

Glossary Assignment: Vocabulary

In the Glossary Assignment, students (n = 70) individually listed 2 to 31 terms as unknown to them, with an average of 11.7 terms per student. Science major students listed fewer terms than non-science major students: on average 10.7 (n = 43, range 2–20) for science majors and 12.9 (n = 27, range 2–31) for non-science majors. From all 2,756 terms in the (translated) article, a total of 59 different terms were listed by the students. Only 23 of these terms were listed by more than 10 students, the other 36 apparently being known to at least 60 (86 %) of their classmates. The 23 terms most frequently mentioned are listed in Table 5.
Table 5

Frequency of terms listed by more than ten students as unknown to them, as percentage of students, for science majors, non-science majors and all students; frequency of occurrence of each term in the text of Campbell et al. (2009)

Term

Frequency as percentage of students

Frequency occurrence term in article

Science majors (n = 43)

Non-science majors (n = 27)

All students (n = 70)

Cognition

74

81

77

2

Configurative

74

81

77

1

Empathise

72

74

73

12

Binomials

65

63

64

1

Pseudo-random order

60

67

63

1

Psychophysiological

51

59

54

1

Empirical

47

56

50

2

Perception-action model

49

48

49

1

Schizotypical

49

44

47

2

Correlations

40

56

46

1

Affective

42

48

44

2

Correlated

47

33

41

2

Compatibility

37

48

41

1

Induced

28

52

37

1

Baseline

33

37

34

1

Physiological

23

48

33

1

Graded

28

33

30

1

Stimuli

33

22

29

12

Complementary

23

30

26

1

Tendency

21

30

24

1

Neural

14

30

20

1

Susceptible

16

19

17

2

Significantly

12

22

16

7

Eighteen of these 23 terms are more frequently noted by non-science majors than by science majors. As to be expected, large differences occur in typical natural science terms as correlations, physiological, neural and significantly, but also in more general science terms as induced and tendency. Strikingly, correlated and stimuli were noted far more frequently by the science majors than the non-science majors, an effect for which we have no explanation at present.

The listed terms roughly fall into two categories. One category consists of terms that are frequently found in many scientific contexts, such as cognition, empirical and correlation. Although these are rarely used in colloquial language, they can be found in an extensive general dictionary. The other category consists of terms rather specific for the research described in the article, such as psychophysiological, perception-action model and schizotypical. The meaning of these can only be found in specific information sources such as university course books or specialised websites.

Some terms in both categories may seem familiar to students but appear to have a somewhat different meaning in the context of the article, such as baseline and significantly. They pose a different kind of problem since students might not feel tempted to check their (contextual) meaning. Another problem, observed by the teachers, was that students often had difficulty choosing the contextually correct meaning of any term that has multiple meanings such as graded. Eventually, the meaning of all unknown words were found using a combination of dictionary use, students’ discussion and hints by the teacher. It took approximately 30 min discussing the unknown words in class in order to complete students’ glossaries even after a round of dictionary use.

The questionnaire revealed that 43 % of the science majors found looking up their unknown words in a dictionary informative and that 28 % of them found it a positive activity. For the non-science majors, these figures were 57 and 40 %, respectively.

Additional Results from the Questionnaire

At the end of the intervention students were asked their opinion on a number of aspects of the intervention. With regard to the applicability of what they had learned 88 % of the science majors and 66 % of the non-science majors agreed that they would be able to apply what they had learned about the different types of argumentative elements after the intervention. About their ability to independently employ PSL as an information source these figures were 68 and 36 %, respectively; 71 % of the science majors and 55 % of the non-science majors agreed that the intervention had been useful for themselves.

Discussion

Our study suggests that using the presented argumentative analysis framework pre-university students are able to study and analyse PSL to such an extent that they can identify the elements motive, objective, main conclusion and implication and part of the underpinning. After the intervention students focused considerably more on the main conclusions while summarising a PSL paper. The accessibility of the chosen PSL article seemed sufficient for understanding the global message of the authors on the paragraph level. Students’ vocabularies appeared to be insufficient to read and study the article without help.

The science majors were more prolific in correctly finding argumentative elements than the non-science majors. This difference could simply be explained by higher ability of the science students, or by motivational issues since this assignment was met more positively by the science majors than the non-science majors. However, the larger difference between the two groups in frequency of finding supports, especially the secondary supports, indicates that probably also the higher content understanding of the science students plays a role, given the more technical character of these elements. If that is indeed the case, then the students, science or non-science, could possibly reach a more complete analysis, including most of the supports, by providing sufficient attention to the subject content of the articles during lessons.

There are more possible reasons for the low identification frequency of the supports. Supports have no fixed location within the Discussion section of an article, nor a specific vocabulary. Furthermore, since the numbers of supports, counterarguments and refutations vary from one article to another, it is very difficult for an inexperienced reader to decide whether all are identified. Another reason may be that students perceived the assignment less strict than we intended and that they were quite content having found a couple of supports.

Compared with supports, identifying motive, objective, main conclusion or implication is probably less demanding with respect to content understanding. Still, students appeared to have used their content understanding while trying to find these elements, as most of those that failed to correctly identify them still chose text fragments that were similar content to the type of element they were looking for. The motive was an exception, posing considerably more problems than the other three. It was easily confused with the main implication. This is surprising since the motive and the implications are located in the Introduction and in the Discussion, respectively. We assume that the students’ confusion was caused by the article’s particular implication of ‘being able to tell the colour of extinct dinosaurs’ that seemed to be the very reason of conducting this research.

It is likely that students used the main conclusion—once it was found—to try and identify supports nearby, as shown by the relatively high frequency of supports 1, 2, 2.1 and 4. This suggests that students tried to make use of location aspects even for the supports, rather than identifying them by their content and relation to the main conclusion, that define their role in the argumentation. Similarly, Falk and Yarden (2009) showed how students did not fully acknowledge the relations laid out by the authors of an article between the supports and their main conclusion: ‘In most instances, the students apply … to reach their own conclusions from the data, usually overlooking the required coordination with the scientists’ conclusions as exposed in the Discussion section’ (p. 374).

The pre- and post-intervention summaries of Warren (1983) showed a notably increased presence of the main conclusions, indicating that students had become aware of the predominance of this type of element, that ‘bears the message’ of the article. This was the case for both main conclusion 1—that was central during the analysis of the paper—and for main conclusion 2—that received much less attention. Our students also showed a growing tendency to present main conclusions in their summaries without mentioning supports. If students do include supports of a main conclusion in their summaries, they probably find that the main conclusion should be accompanied by some of its associated argumentative elements. But many of our students may also have found supports as such important enough to include them, as the results in Table 4 seem to indicate: even in the post-test still 49 % of them presented one or more supports without their associated main conclusion. We can explain this by the similarities between the first main conclusion and some of its supports, in both of which Warren (1983) refers to the bacteria he had found in human stomachs. Many students were probably still unaware of the difference in epistemic status between seemingly interchangeable phrases as: ‘The stomach must not be viewed as a sterile organ with no permanent flora’ and ‘… I have observed small curved and S-shaped bacilli in 135 gastric biopsy specimens’ (Warren 1983, p. 1273) which do not convey exactly the same message, being a main conclusion and a support, respectively.

The almost unchanged frequency of the ‘Candida’ fragment in the summaries, together with the decline of the neutral ‘antrum’ fragment, indicates that during the post-test students did not close their eyes to relevant text fragments other than the set of previously identified elements, but probably had produced their own independent understanding of the text.

The higher average number of argumentative elements in the post-test summaries was not surprising after the argumentative analysis during the intervention. In any case it was consistent with a higher awareness in students of the argumentative structure of the paper since the assignment did not specify that the summary would have to contain argumentative elements. The shorter summaries in the post-test seem to indicate an increased confidence in participants’ ability to decide which text fragments were important.

The accessibility of the Introduction section of Campbell et al. (2009) was sufficient for most students to successfully link paragraphs with their summaries in the matching assignment, indicating that they got the ‘general message’ of each paragraph. This should be sufficient for studying the rest of the article. The results from the interviews also suggest that superficial reading incidentally may have hampered understanding the Introduction section. It is likely that not all students were aware of the level of accuracy required for reading PSL, even for experienced readers. In this light, the high number of students that found reading the text easy indicates that they underestimated this task. If the required high level of reading accuracy necessary for PSL is indeed unusual for our students, then this could also explain their difficulties identifying supports in Vinther et al. (2008).

The vocabulary of the students, as was shown by the results of their lists of unknown words in Campbell et al. (2009) was such that they could benefit from a glossary. Since the list of unknown words was relatively short, providing them with a pre-composed glossary would have been entirely feasible. Of course the length of such a list depends on the choice of article. Articles with a higher number of unknown words would require longer lists and more preparatory work for teachers, making the choice for suitable articles smaller. On the other hand, the decision for providing a glossary anyway would enhance the number of suitable articles. A teacher could indicate which terms in the list are important outside the scope of its own article and thus worth memorising. Providing a glossary would not necessarily undermine the authenticity of our approach since even graduate students would have to look up, or ask for, the meaning of new terms. A glossary merely saves time.

Our students did not always seem to be aware or appreciative of the opportunity studying PSL offers to learn new and important scientific vocabulary, although many of the terms in Table 5 help shape a useful vocabulary especially for future science students. In this respect we fully concur with Wellington and Osborne (2001, p. 2) in their premise ‘Learning the language of science is a major part (if not the major part) of science education. Every science lesson is a language lesson’.

The questionnaire (part B, after the intervention) showed students’ optimism for themselves or for their classmates to be able to use PSL as a personal information source. Their opinion was based on studying only three selected PSL articles, but their optimism may be not unjustified since for browsing ‘the latest scientific news’ from primary sources it is often sufficient to find the objective, the main conclusion, and the implications. The identification of these three types of elements appeared to be fairly easy for our students, keeping in mind that the articles we used had been translated into their native language. Since in our approach of PSL through argumentative analysis only parts of an article are studied extensively, retrieving information from PSL is not necessarily an exclusive component of multidimensional scientific literacy in terms of Bybee’s framework (Bybee 2010), but rather a basic technique that can be used for either a quick scan of an article or as the start of an extensive evaluation.

Overall, our findings show that for this particular pre-university group it was indeed possible to gain access to the PSL article of Vinther et al. (2008) through argumentation analysis. This article was certainly representative for PSL with respect to its highly technical Methods section, formal style, and the specific, in this case cytological and paleontological, vocabulary. Still, generalising these findings for using PSL on the basis of only three different articles is not possible in terms of probable success with any other school group or with any choice of articles. In this respect we mainly gave a proof of possibility. The framework presented here positively enabled students to independently study and analyse PSL. The description of the types of elements and the application of the framework raised no questions. The framework appeared to be easy to use and apply.

The role of the teacher should not be underestimated. The difficulties that students encounter are such that the teacher should have considerable (academic) experience in reading PSL in order to be able to guide his or her students through this matter. We do not think that this argues against the use of primary literature. With respect to APL Falk and Yarden (2009, p. 380) remark: ‘future developers … should not regard the complexity and the inherent difficulty associated with this genre as negative factors … The complexity can be considered, to a certain extent, an elicitor of authenticity at the level of the context and of the applied epistemic practices, provided that the difficulties are claimed and can be answered by teachers and peers in a suitable manner’.

Now that it appears to be possible to introduce pre-university students to PSL through argumentation analysis, one should ask whether it is useful. In the February 2009 special issue of Research in Science Education on the reading of scientific texts, in particular APL, several authors show how they implemented the use of APL in classrooms pursuing different goals that cannot be reached using normal textbooks. As Goldman and Bisanz (2002, p. 39) stated: ‘Typically, neither textbooks nor training manuals convey the sociohistorical context of the presented information and they provide limited information about the process by which knowledge claims are made and evaluated’.

In the same issue of Research in Science Education, Osborne gives a response to the presented research providing a critical view on the use of APL in classrooms (Osborne 2009). We will consider two of his important statements that challenge PSL just as much as APL. Osborne writes (Osborne 2009, p. 399) ‘It could be argued, rather, that the texts that most students will commonly encounter in their future lives are media reports of science and that this type of literature should be the focus of science education’. This holds for most students, but for pre-university students we may expect that someday they will have to extensively study primary literature in at least their own discipline. Even for those who will eventually not pursue an academic career their experiences with PSL will help them shape a more realistic image of some important ingredients of scientific work.

Osborne (2009, p. 398) also states: ‘Their [successful textbooks’] argument is not reliant on primary data sources (from which they are far removed) but on a multi-modal account of how the world is’. With Osborne we acknowledge that this valuable account should perhaps be accompanied by offering ‘students some insight into the context of justification’, while maintaining ‘a clear and transparent explanation of the contemporary understanding’ (Osborne 2009, p. 398). This is exactly what we tried to do, fitting within the scope of Science in Society. Students studied the underpinnings scientists provide for their claims revealing their tentative character. In focusing on the argumentative structure, much more than on the content, we clearly demonstrated the inherent uncertainty of basically all scientific endeavour, leaving the explication of other basic concepts of science to textbooks of other science subjects.

The findings of our study may have some pedagogical implications but also raise some new questions. As for students’ personal vocabularies we found that a number of terms frequently occurring in almost every field of science were yet unknown by a considerable part of our students: empirical, correlations/correlated, induced, complementary, significantly. We pose the question that should pre-university education pay more attention to the important scientific literacy aspect of acquiring a basic scientific vocabulary? If so, the use of PSL offers a good possibility to teach these very general and common scientific terms from authentic materials.

An important challenge for improving the presented curriculum unit is to help students identify the supports of the main conclusion of a scientific publication. For these it is not sufficient to search for identifier terms or to look in specific locations of the article. In contrast to the element types of motive, objective, main conclusion and implication, supports probably call for a stronger focus on the content, requiring in students a more detailed understanding of the technical aspects of the research presented in the article. Further study may reveal whether just prolonged training is sufficient for enabling students to successfully recognise and analyse the underpinning of the main conclusion in PSL or that an additional heuristic method is needed.

During this report, we have paid little attention to the relations between the argumentative elements that were to be identified. These relations had been explained to the students and implicitly used by them during their analysis. We will report later on students’ understanding of the relations between the argumentative elements.

Reading PSL offers students a new view on science: vocabulary and other writing conventions, epistemology, the intrinsic uncertainty of scientific claims—these are all new to most pre-university students. We may assume that reading and analysing PSL somehow changes students’ ideas about science. In the future, we will report on participants’ individual views on science before and after the intervention.

Notes

Acknowledgement

The research for this article was funded by Platform Bèta Techniek through project RUG1 of the Dudoc programme.

References

  1. Almeida, C. A., & Liotta, L. J. (2005). Organic chemistry of the cell: an interdisciplinary approach to learning with a focus on reading, analyzing, and critiquing primary literature. Journal of Chemical Education, 82, 1794–1799.CrossRefGoogle Scholar
  2. Baggott la Velle, L., & Erduran, S. (2007). Argument and developments in the science curriculum. School Science Review, 88(324), 31–39.Google Scholar
  3. Bazerman, C. (1988). Shaping written knowledge: the genre and activity of the experimental research article in science. Madison: Wisconsin University Press.Google Scholar
  4. Boersma, K. Th., van Graft, M., Harteveld, A., de Hullu, E., de Knecht-van Eekelen, A., Mazereeuw, M., et al. (2007). Leerlijn biologie van 4 tot 18 jaar. Uitwerking van de concept-contextbenadering tot doelstellingen voor het biologieonderwijs. Utrecht: Nibi.Google Scholar
  5. Brill, G., & Yarden, A. (2003). Learning biology through research papers: a stimulus for question-asking by high-school students. Cell Biology Education, 2, 266–274.CrossRefGoogle Scholar
  6. Bybee, R. W. (2010). The teaching of science: 21st century perspectives. Arlington: NSTA Press.Google Scholar
  7. Campbell, M. W., Devyn Carter, J., Proctor, D., Eisenberg, M. L., & de Waal, F. B. M. (2009). Computer animation stimulate contagious yawning in chimpanzees. Proceedings of the Royal Society London, B, 276, 4255–4259.CrossRefGoogle Scholar
  8. Chinn, C. A., & Malhotra, B. A. (2002). Epistemologically authentic inquiry in schools: a theoretical framework for evaluating inquiry tasks. Science Education, 86, 175–218.CrossRefGoogle Scholar
  9. Collins, A., Brown, J. S., & Holum, A. (1991). Cognitive apprenticeship: making thinking visible. American Educator, 15(3), 6–46.Google Scholar
  10. Collins, A., Joseph, D., & Bielaczyc, K. (2004). Design research: theoretical and methodological issues. The Journal of the Learning Sciences, 13(1), 15–42.CrossRefGoogle Scholar
  11. Epstein, H. T. (1972). An experiment in education. Nature, 235, 203–205.CrossRefGoogle Scholar
  12. Falk, H., & Yarden, A. (2009). Here the scientists explain what I said. Coordination practices elicited during the enactment of the Results and Discussion sections of adapted primary literature. Research in Science Education, 39(3), 349–383.CrossRefGoogle Scholar
  13. Ford, M. J., & Wargo, B. M. (2007). Routines, roles, and responsibilities for aligning scientific and classroom practices. Science Education, 91, 133–157.CrossRefGoogle Scholar
  14. Gillen, C. M. (2006). Criticism and interpretation: teaching the persuasive aspects of research articles. CBE Life Sciences Education, 5(1), 34–38.CrossRefGoogle Scholar
  15. Goldman, S. R., & Bisanz, G. L. (2002). Toward a functional analysis of scientific genres: implications for understanding and learning processes. In J. Otero, J. A. León, & A. C. Graesser (Eds.), The psychology of science text comprehension (pp. 19–50). Mahwah: Erlbaum.Google Scholar
  16. Hoskins, S. G., Lopatto, D., & Stevens, L. M. (2011). The C.R.E.A.T.E. approach to primary literature shifts undergraduates’ self-assessed ability to read and analyze journal articles, attitudes about science, and epistemological beliefs. CBE Life Sciences Education, 10, 368–378.CrossRefGoogle Scholar
  17. Janick-Buckner, D. (1997). Getting undergraduates to critically read and discuss primary literature. Journal of College Science Teaching, 27(1), 29–32.Google Scholar
  18. Kelly, G. J., Regev, J., & Prothero, W. (2008). Analysis of lines of reasoning in written argumentation. In S. Erduran & M. P. Jimenez-Aleixandre (Eds.), Argumentation in science education (pp. 137–157). Berlin: Springer.Google Scholar
  19. Kuhn, D. (2010). Teaching and learning science as argument. Science Education, 94, 810–824.CrossRefGoogle Scholar
  20. Kuldell, N. (2003). Read like a scientist to write like a scientist: using authentic literature in the classroom. Journal of College Science Teaching, 33(2), 32–35.Google Scholar
  21. Muench, S. B. (2000). Choosing primary literature in biology to achieve specific educational goals: some guidelines for identifying the strengths and weaknesses of prospective research articles. Journal of College Science Teaching, 29(4), 255–260.Google Scholar
  22. Mulnix, A. B. (2003). Investigations of protein structure and function using the scientific literature: an assignment for an undergraduate cell physiology course. Cell Biology Education, 2(4), 248–255.CrossRefGoogle Scholar
  23. National Research Council. (1996). National science education standards. Washington: National Academy Press.Google Scholar
  24. Norris, S. P. (1992). Practical reasoning in the production of scientific knowledge. In R. Duschl & R. Hamilton (Eds.), Philosophy of science, cognitive psychology, and educational theory and practice (pp. 195–225). Albany: State University of New York Press.Google Scholar
  25. Norris, S. P., & Phillips, L. M. (2003). How literacy in its fundamental sense is central to scientific literacy. Science Education, 87, 224–240.CrossRefGoogle Scholar
  26. Norris, S.P., Stelnicki, N., Vries, G. de (2011). Teaching mathematical biology in high school using adapted primary literature. Research in Science Education. doi:10.1007/s11165-011-9215-8.
  27. Osborne, J. (2009). The potential of adapted primary literature (APL) for learning: a response. Research in Science Education, 39(3), 397–403.CrossRefGoogle Scholar
  28. Osborne, J., Erduran, S., & Simon, S. (2004). Enhancing the quality of argumentation in school science. Journal of Research in Science Teaching, 41(10), 994–1020.CrossRefGoogle Scholar
  29. Peck, W. H. (2004). Teaching metastability in petrology using a guided reading from the primary literature. Journal of Geoscience Education, 52, 284–288.Google Scholar
  30. Phillips, L. M., & Norris, S. P. (2009). Bridging the gap between the language of science and the language of school science through the use of adapted primary literature. Research in Science Education, 39(3), 313–319.CrossRefGoogle Scholar
  31. Sampson, V., & Clark, D. B. (2008). Assessment of the ways students generate arguments in science education: current perspectives and recommendations for future directions. Science Education, 92, 447–472.CrossRefGoogle Scholar
  32. Suppe, F. (1998). The structure of a scientific paper. Philosophy of Science, 65, 381–405.CrossRefGoogle Scholar
  33. Swales, J. M. (1990). Genre analysis: English in academic and research settings. Cambridge: Cambridge University Press.Google Scholar
  34. Tamir, P. (1985). Content analysis focusing on inquiry. Journal of Curriculum Studies, 17(1), 87–94.CrossRefGoogle Scholar
  35. Tenopir, C., & King, D. W. (2001). The use and value of scientific journals: past, present and future. Serials, 14(2), 113–120.Google Scholar
  36. Thompson, D. K. (1993). Arguing for experimental facts in science: a study of research article results sections in biochemistry. Written Communication, 10, 106–130.CrossRefGoogle Scholar
  37. Toulmin, S. (1958). The Uses of Argument. London: Cambridge University Press.Google Scholar
  38. Van Gelder, T. (2005). Teaching critical thinking: some lessons from cognitive science. College Teaching, 53(1), 41–46.CrossRefGoogle Scholar
  39. Venville, G., & Dawson, V. (2010). The impact of an argumentation intervention on grade 10 students’ conceptual understanding of genetics. Journal of Research in Science Teaching, 47(8), 952–977.Google Scholar
  40. Vinther, J., Briggs, D. E. G., Prum, R. O., & Saranathan, V. (2008). The colour of fossil feathers. Biology Letters, 4, 522–525.CrossRefGoogle Scholar
  41. Warren, J. R. (1983). Unidentified curved bacilli on gastric epithelium in active chronic gastritis. Lancet, i, 1273–1275.Google Scholar
  42. Wellington, J. J., & Osborne, J. (Eds.). (2001). Language and literacy in science education. Buckingham: Open University Press.Google Scholar
  43. Yarden, A. (2009). Reading scientific texts: adapting primary literature for promoting scientific literacy. Research in Science Education, 39, 307–311.CrossRefGoogle Scholar
  44. Yarden, A., Brill, G., & Falk, H. (2001). Primary literature as a basis for a high-school biology curriculum. Journal of Biological Education, 35(4), 190–195.CrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media Dordrecht 2012

Authors and Affiliations

  • Marcel Koeneman
    • 1
  • Martin Goedhart
    • 1
  • Miriam Ossevoort
    • 1
  1. 1.Faculty of Mathematics and Natural SciencesUniversity of GroningenGroningenThe Netherlands

Personalised recommendations