Tea 20225
Tea 20225
Tea 20225
574–599 (2008)
Abstract: This study is a meta-interpretative analysis that focuses on research conducted and
published by other researchers. Concepts central to this study include global practical relevance,
curriculum design, and formative situation. We analyzed 35 studies selected from 374 published studies in
the years 2000 and 2001 in three journals referenced in the International Scientific Index. Using a replicable
methodology developed specifically for this research, we found evidence of s clusters of variables that
suggest the existence of transversal traits in the 35 science education research studies. These results form a
reference framework of theoretical and practical knowledge relevant for research and practice pertaining to
teaching and learning science. ß 2008 Wiley Periodicals, Inc. J Res Sci Teach 45: 574–599, 2008
Keywords: science teacher education; curriculum development; science education
Contract grant sponsor: Portuguese Ministry of Science, Technology and Higher Education; Contract grant number:
POCTI/CED/36466/99-00.
Correspondence to: J.B. Lopes; E-mail: blopes@utad.pt
DOI 10.1002/tea.20225
Published online 8 April 2008 in Wiley InterScience (www.interscience.wiley.com).
between researcher and practitioner needs is the problem that we address in this meta-
interpretative study.
The present study seeks to contribute to the construction of a practical and theoretical
reference framework that is enriched in its articulation and usefulness. We sought studies that
linked researcher to researcher and researcher to practitioner. We located and grouped these
studies in clusters according to common characteristics that we further analyzed to identify
transversal traits in educational research. By transversal traits, we mean characteristics that are
common to the research studies, even those with distinctly different theoretical and methodo-
logical foci. Accomplishing our goal required the design of an appropriate methodological
framework.
A total of 374 studies published in 2000 and 2001 were located for this study, which were
taken from three journals referenced in the International Scientific Index (ISI): Science Education
(SE); International Journal of Science Education (IJSE); and Journal of Research in Science
Teaching (JRST). Using criteria related to practical relevance, we selected 35 studies that constitute
the corpus of our meta-interpretative study.
Four concepts were central to this study: global practical relevance; curriculum design;
formative situation; and transversal traits. From an elaboration of the first three concepts, we
obtained 23 variables for analysis. The 35 studies forming the corpus were then analyzed,
compared, and grouped, based on these 23 variables. The groupings, or clusters, were generated
with the STATISTICA software (StatSoft, 1997), and the clusters constituted the data from which we
identified transversal traits.
We explain the various methodological aspects of our analysis, particularly the decisions
related to accuracy, replicability, and stability of the clusters, and the stability of the results formed
from the clusters. From our results we propose a reference framework for teaching and research
and suggest future developments.
Science education research has examined objects of study that capture partial aspects of the
science teaching and learning process. These studies are important, but they fail to capture
the richness of the science classroom. There are arguments in the literature that question whether
the complexity of the classroom is amenable to research on subcomponents of the teaching and
learning process. Martinez-Terrades (1998) noted the need for studies that integrate different lines
of research (in Gil-Perez et al., 1999).
Several researchers concluded that innovations in teaching and learning will only produce
results if attention is paid to the complexity of the processes (Cachapuz et al., 2001; Linn, 1987).
For example, there are studies that focus on assessment of students, but this topic is related to
learning (Black, 1998) and thus to learning contexts (e.g., laboratory work). Likewise, Gil-Perez
et al. (1999) questioned the separation among studies related to concept learning, paper-and-
pencil problem solving, or laboratory work. This separation reflects the traditional compartmen-
talization of the components of instruction (lectures, tutorial, and laboratory classes) at a time
when research needs to be integrated if it is to inform science teaching and learning practice.
1. Interconnections between research lines (a need verified in the literature, see Martinez-
Terrades [1998]), comprehensively related to the complexity of teaching and learning
processes.
2. Evidence of feasibility in formal teaching and learning environments, showing in a
detailed and convincing manner that the proposals made do not imply unreasonable
demands on procedures, resources, time, and curriculum.
3. Explicit assumption of a theoretical framework that guides the research questions and the
finding and discussion of results.
The concept of curriculum design has two components (Roldão, 1999): learning outcomes
and ways to achieve them. To achieve the curricular outcomes, five conditions must be taken into
Journal of Research in Science Teaching
TRANSVERSAL TRAITS 577
account (Lopes, 2004), including: (1) the students’ knowledge and difficulties; (2) the
contextualization of the themes; (3) the conceptual fields and their structure; (4) the competencies
to develop; and (5) the organization of teaching. Recent research studies about curriculum design
are less nomothetic and more descriptive (White, 2002). This shift is necessary to elucidate the
relationships between students’ learning and aspects such as the incorporation of epistemological
reflection in teaching practices; normative curricula analysis; social, technological, and
environmental relevance; and construction and evaluation of didactic materials (Kelly & Lesh,
2000). Studies about curriculum design should lead to better foundations, better teaching, greater
relevance to students, and increased effectiveness (Kelly & Lesh, 2000). Regardless of the
multiplicity of dimensions found in curriculum design research, Martin and Solbes (2001) and
Koliopoulos and Ravanis (2000) identified three essential features, including epistemological
validity (compliance with updated epistemological and subject-matter knowledge), didactic
validity (compliance with updated educational knowledge), and feasibility (possibility of
adaptation and use in real teaching conditions). In addition, sound curriculum design must include
plans for adequate curricular management in the classroom. Features of curriculum management
include formative assessment and assessment of products (Shepard, 2001), incorporation of
students’ knowledge and appropriate contexts (Lopes, 2004), teacher mediation (Dumas-Carré &
Goffard, 1998; Moll, 2002) of the students’ learning, and effective use of conceptual tools by
students.
Formative situation is a concept that was introduced by Astolfi, Darot, Vogel, and Toussaint
(1997), and further developed by Lopes (2004). It makes explicit key aspects of a didactic
modeling intended to transform teaching objects into consolidated learning outcomes. A
formative situation must consider teacher mediation as being of paramount importance, determine
the knowledge already available to students, give students real opportunities to perform tasks and
formulate problems, allow students to learn in a progressive and sustained way, and provide
opportunities for students to use newly acquired knowledge.
Formative situation (Figure 1) links the concepts of students’ available knowledge
(Cachapuz, Praia, & Jorge, 2000), situation (e.g. Science–Technology–Society context [Lopes,
2004; see also Gil-Perez et al., 1999]), teacher mediation (Weil-Barais & Dumas-Carré, 1998),
and learning tasks (Howe, 1996). A situation is presented to students for them to explore. Students
then are invited to develop a mental and practical activity, using their available knowledge within a
predetermined time frame. The activities are thus conditioned by the learning tasks, the teacher’s
mediation and involvement of students, and the available material and conceptual resources. The
concept of a formative situation offers a powerful approach for designing and managing curricula
(Lopes, 2004; see also Branco, 2005; Cravino, 2004).
Transversal traits are characteristics that are common to several research studies,
independent of their theoretical or methodological approaches. These characteristics correspond
to emergent knowledge that is not yet necessarily explicit in the individual research reports.
However, when looking at a considerable number of published studies with practical relevance
this knowledge begins to surface. To identify the transversal traits first we chose the relevant
variables (i.e., global practical relevance, curriculum design, and formative situation). Next, we
used a software analysis tool to build clusters or groups of interrelated variables. The occurrence of
a stable cluster identifies the presence of a transversal trait.
components: the need for eliciting transversal traits that link individual reports and the need to
draw from existing research an improved reference framework for practitioners and researchers.
We have thus formulated three research questions to be addressed:
Q1: What methods should be used to elicit transversal traits in a corpus of research studies
conducted in formal educational contexts that have practical relevance?
Q2: Which transversal traits can be inferred from this analysis?
Q3: How can these traits embody a reference framework that informs research and practice?
Research Methods
We describe the methods used in this research, from corpus selection to data analysis. Some
specific procedures were created and improved as the research proceeded.
1. The criterion of relevance was expressed in a list of dimensions of teaching practice. Each
dimension was assigned a representative set of keywords, and each dimension was coded
(1–8) by a digit (see Table 1).
2. The keywords were searched for in each abstract. The occurrence of a dimension (as
indicated by the presence of a keyword) was registered. Thus, for example, a coding
sequence 36247 means that the dimensions of relevance 3, 6, 2, 4, and 7 were found.
3. An abstract was considered as indicating a study with relevance if five or more of digits
(each representing one of the eight dimensions) were present in its coding sequence
(see Table 1).
We compared the results obtained independently by the two investigators regarding the
studies that would or would not be included in the corpus. The first comparison yielded reasonable
agreement, but showed the need for fine-tuning. All six researchers in the team discussed
the agreement issue in order to achieve a broad conceptual consensus at the following levels:
1. Clarifying the elements that must be present in an abstract to maximize the probability
that the correspondent study has relevance.
2. Expanding the set of keywords and dimensions (Table 1 is shown in its final form) and
ensuring the correct allocation within the various dimensions.
3. Rewriting the coding sequence with the digits 3 (task dimension) and 6 (mediation
dimension) at the beginning of the sequence to indicate essential dimensions in the
coding scheme.
4. Considering that a study has relevance if and only if its coding sequence has at least five
digits, starting with digits 3 and 6 (thus, the coding sequence should be of the type
36###).
5. Making sure that the keywords that lead to the allocation of codes 3 and 6 are not out of
context (e.g., the word ‘‘discussion’’ may refer to discussion in the classroom or
discussion of the research results).
With these improvements, the two methods—the full-abstract reading and the keyword
abstract search—finally yielded disagreement of less than 1%. Only one study (out of 103) was
excluded from the corpus using the keyword-search method, but was included in the full-reading
method. If transversal traits do in fact exist, we agreed, they would not fail to show up as a
consequence of such a small margin of error.
We then applied the semiautomated, keyword-search method to the remaining 271 studies
(from the two remaining journals). Starting with an initial set of 374 studies, we built a corpus
comprised of 14 studies from the IJSE, 12 from the JRST, and 9 from SE (see full list in Appendix 1).
1. Level 0 includes three questions aligned with the three general concepts defining
relevance for teaching practice (global practical relevance, curriculum design, and
formative situation).
Table 1
Final list of keywords grouped by dimension of relevance for teaching practice (coding digits between parentheses)
Knowledge,
competences,
Available attitudes and
knowledge Learning Physical values to be Curriculum
(from students) environment Task situation Resources Mediation developed design
understanding
Pupils Group Experimental Real life Physic Questions Student
understanding work Discussions constructing
Misunderstanding Inquiry Learning Everyday Model Dialogue Knowledge
activity Argumentation construction
Peer Modelling Software Assess(ment) Perform(ance)
Autonom(ous) Work Strategies Debate Student learn(ing)
Technology Talk Learn
Textbook
TRANSVERSAL TRAITS 581
2. Level 1 includes questions in which each of the three general concept questions are
unfolded by answering five to eight specific questions.
The specific questions (see Appendix II) further defined the dimensions that we considered
essential for a study to be relevant for teaching practice. These additional distinctions included:
In the dimension ‘‘Global Practical Relevance’’ (P) the study must include more than two
lines of research (P-RL) and features directly related to teaching practice. These features
include: Feasibility (P-F), Methods (P-M), and/or Assessment (P-A) used in teaching;
Skills and Competencies (P-SC) and/or Values and Attitudes (P-VA) that students
develop; and developed Knowledge (P-K).
In the dimension ‘‘Formative Situation’’ (FS) the study must include: Student Knowledge
mobilized in learning (FS-SK); Learning Environment (FS-LE); students’ Tasks (FS-T),
Physical Situations (FS-PS), and Resources (FS-R) used; modalities of Mediation (FS-
M); Knowledge Developed during learning (FS-KD); and Assessment instruments/
procedures used in mediation (FS-A).
In the dimension ‘‘Curriculum Design’’ (CD) the study must present a perspective of
curriculum design that includes: Epistemological Validity (CD-EV) and Feasibility (CD-
F); justification of the processes and requisites for Learning (CD-L); Explicit Sequence
and articulation of activities (CD-ES); and articulation with Other Knowledge (CD-OK).
The analysis instrument was validated using a two-stage process. First, each researcher
independently applied the instrument to several studies. The results were then compared and the
categories of analysis were fine-tuned. Second, a set of eight studies were selected. Each study
was read by at least two researchers who completed the analysis instrument. The six
investigators compared the results for each study and fine-tuned the categories of analysis.
This second stage of comparison and discussion of the preliminary results obtained allowed
us to fine-tune the interpretation of key questions for each dimension. Answering ‘‘yes’’ or ‘‘no’’ to
each question requires a high level of certainty and confidence. In the process we developed a
better shared understanding of the concepts involved and of the meaning of the terminology used
by each investigator when writing the research abstract. Finally, all the remaining studies were
analyzed by two researchers, working independently at first, coming together to share their
analyses, and then settling occasional emerging disagreements. A consensual agreement was
reached in all cases.
Figure 2. Tree diagram for the 35 cases (all variables) with the identification of the clusters (see Table 2 for
variables code).
We grouped the clusters using Ward’s method, which minimizes the sum of squares of any two
clusters that can be formed at each step of the process of cluster formation. This method is regarded
as very efficient (Statsoft, 1997), even though it tends to create clusters of small size but with
strong links among the correspondent variables. For the calculation of distances between cases we
used percent disagreement [calculated by: distance (x, y) ¼ (number of xi 6¼ yi)/i]. This measure is
particularly useful when the data for the dimensions included in the analysis are categorical in
nature (Statsoft, 1997). From the cluster analysis we obtained dendograms that group variables
(see Figure 2) and can allow us to identify the clusters and see the relationship among them.
Interpreting these clusters allows us to infer and articulate the nature of the transversal traits.
This means that each cluster includes a set of variables that are closely interrelated. For example,
cluster 6 is formed by two variables that share the same value (represented by 0 or 1) in 24 different
studies and are, therefore, somehow associated with each other. This relationship must then
be interpreted.
selection criteria failed to identify. In addition, some studies outside the corpus may have an
indirect relevance for practice due to their theoretical nature or focus. Regardless, we propose
that the low percentage of studies selected for the corpus (10%) deserve the attention of
researchers.
The studies selected in the preliminary analysis (see Appendix 1) were further analyzed in a
more refined way. Performing the cluster analysis with STATISTICA software results in members
of clusters dependent on the linkage distances selected. For example, using a linkage distance of
1.2 we identified only two clusters, and using a distance of 0.1 we identified a number of clusters
similar to the number of variables, which is a meaningless result. We chose for our purposes a
linkage distance that maximized the number of clusters and, in doing so, maximized the meaning
of each cluster. This distance is about 0.5 (as shown in Figure 2), which resulted in the production
of six clusters.
We cannot establish causal relationships from the clusters in the tree diagram, but we can
establish meaningful associations among variables. On inspection of the tree diagram, we see that
there are two groups of clusters, with dimensions 1, 2, 3, on one side and 4, 5, 6 on the other. A large
separation exists between the two blocks. We may build a hierarchy of clusters by analyzing
the matrix Studies Variables (Appendix 3) in terms of the relative density of ‘‘0’’ (meaning
that the correspondent characteristics are absent in the studies) and ‘‘1’’ (meaning that the
correspondent characteristics are present in the studies). This can also be done using the two-way
joining technique of the STATISTICA software. Using this technique, we found that cluster 1 contains
the set of variables for which almost all cases are marked ‘‘1’’ and, on the other extreme, cluster
4 groups that include variables that tagged with ‘‘0’’ in most cases. According to this criterion,
these clusters may be ordered as 1, 2, 3, and 6, 5, 4.
Testing the stability of the results. We asked whether we would we obtain the same
results (clusters) if we increased the number of studies. The point in question is whether the
small number of studies in the corpus (35) limits the findings. To address this question,
we conducted a second cluster analysis with a random sample of only 17 of the 35 studies
(about half of the number of studies in the corpus). Random selection was guaranteed by blind
raffle, picking out coded labels from a closed box. From a comparison of the tree diagrams
obtained with the two samples (17 and 35 studies) we verified the existence of the same structure—
that is, two cluster groups (see Figure 3). We also verified that clusters 1, 2, 4, and 6 appear with
almost the same composition (clusters 1 and 6 are totally coincident). From a comparative analysis
of Figure 3 (17 studies) and Figure 2 (35 studies) we can say that our method produced consistent
results.
Discussion and interpretation of the results. Transversal traits clearly emerged
from the analysis. Aspects that are less stable still subsist, but we interpret them as not
configuring (at least for the moment) transversal traits. Thus, we identified four stable clusters (see
Table 2).
Cluster 1 is formed by seven variables of the ‘‘Global Practical Relevance’’ dimension (Result
1—Table 2). These seven variables share the value 1 for the majority of the studies. This means that
the characteristics associated with the seven variables are present in most studies and are somehow
associated with each other. From this relationship we infer a transversal trait, meaning a strong
connection between the research reported in the studies and a concern with its relevance for
professional practice. This association of variables points to the importance of the articulation
Journal of Research in Science Teaching
584 LOPES ET AL.
Figure 3. Tree diagram for the 17 cases (all variables) with the identification of the clusters that reappear.
between research and practice. In a more specific way this transversal trait corresponds to two
essential features of research (see cluster 1, Result 1—Table 2), namely presenting proposals to
practitioners that are feasible and clearly show how to improve students’ knowledge, skills, and
values, and crossing two or more research lines, which is of potential interest to the practitioner as
well as the researcher.
Cluster 2 is formed by six variables from the ‘‘Formative Situation’’ dimension (Result 3—
Table 2). In this cluster, 22 studies share the same value for all cluster variables (value 1 in 15 studies;
value 0 in 7 studies). This means that the characteristics associated with the six variables are
somehow associated with each other by either their presence or absence in the studies. From this
relationship we infer a transversal trait linked to the teaching characteristics and learning
conditions, which can be interpreted in several ways. We emphasize one interpretation that is
especially useful to the practitioner, namely research studies that: present evidence about
knowledge acquired by students; describe the learning tasks students completed, the resources
used, and the mediation carried out by teacher; and explain the nature of the learning environment
(see cluster 2, Result 3—Table 2). Thus, the structural aspects of a formative situation, when
present, are all articulated. From this finding we may infer that the research concerned with
relevance for practice should not focus on partial aspects of teaching and learning but instead
should focus on structural aspects and their articulation. We emphasize that finding this articulation
(among the tasks proposed, the resources used, the mediation carried out by the teacher, and the
learning environment) in various independent studies confirms the importance of the concept of
formative situation for practitioners and researchers.
Journal of Research in Science Teaching
TRANSVERSAL TRAITS 585
Table 2
Summary of main results
Stable Clusters of Variables Summary results of cluster analysis
Cluster 1
P - Global Practical relevance of the paper. Result 1 - Cluster 1, as a stable cluster (see Fig. 2
P-RL - Relevant for teaching Practice - the paper and Fig. 3), configures a transversal trait (the
involves more than two Research Lines. characteristics associated to the 7 variables are
P-F - Relevant for teaching Practice – Feasibility present in most papers).
of proposals.
P-M - Relevant for teaching Practice – Methods Result 1.1 - P-M and P-F are very closely connected
used in teaching are described. (see Fig. 4), so they configure a specific
P-SC - Relevant for teaching Practice - Skills transversal trait of result 1.
and Competences that students develop are
described.
P-VA - Relevant for teaching Practice - Values Result 2 - Cluster 1 is connected to cluster 2 (see
and Attitudes that students develop are Fig. 2 and Fig. 3), so together may configure a more
described. general transversal trait.
P-K - Relevant for teaching Practice - Knowledge
that students develop is described.
Cluster 2
FS – The paper develops and/or studies a Result 3 - Cluster 2, as a stable cluster (see Fig. 2
Formative Situation in a global sense. and Fig. 3), configures a transversal trait (the
FS-LE – Aspect of Formative Situation - The characteristics associated to the 6 variables are
paper elucidates/describes the Learning somehow associated to each other, by either their
Environment. presence or absence in the papers).
FS-T - Aspect of Formative Situation - The paper Result 3.1 - FS-T and FS-M are very closely
elucidates/describes the students’ Tasks used. connected (see Fig. 4 and Fig. 5), so configure a
FS-R - Aspect of Formative Situation - The paper specific transversal trait of result 3.
elucidates/describes the Resources used.
FS-M - Aspect of Formative Situation - The paper Result 3.2 - FS-LE and FS-R are very closely
elucidates/describes the modalities of connected (see Fig. 4 and Fig. 5), so configure a
Mediation done by the teacher. specific transversal trait of the result 3.
FS-KD - Aspect of Formative Situation - The
paper elucidates/describes the Knowledge
Developed during learning.
Cluster 4
CD - The paper presents a perspective of Result 4 - Cluster 4, as a stable cluster (see Fig. 2
Curriculum Design. and Fig. 3), configures a transversal trait (the
CD-EV - Aspect of Curriculum Design – The characteristics associated to the 4 variables are
paper discusses Epistemological validity of the absent in most papers).
curriculum proposed.
CD-F - Aspect of Curriculum Design – The Result 4.1 - CD-L and CD-F are very closely
paper discusses the Feasibility of connected (see Fig. 5), so they configure a
its proposal. specific transversal trait of the result 4.
CD-L - Aspect of Curriculum Design – The
paper justifies the necessary processes and
requisites for learning.
Cluster 6
P-A - Relevant for teaching Practice - Result 5 - Cluster 6, as a stable cluster (see Fig. 2
Assessment used in teaching is described. and Fig. 3), configures a transversal trait (the
FS-A - Aspect of Formative Situation - The paper characteristics associated to the 2 variables are
elucidates/describes how Assessment absent in most papers).
instruments/practices were used in mediation.
Cluster 4 is formed by four variables from the ‘‘Curriculum Design’’ dimension (Result 4—
Table 2). The four variables in this cluster share the value 0 for the majority of the studies. This
means that the characteristics associated with the four variables are absent in most studies and are
somehow associated with each other. This stable association forms a transversal trait related to
curriculum design and its relevance for practice. In general, the few studies that present a proposal
for curriculum design make the effort to show its relevance for professional practice. This
transversal trait consists of an articulation among the three curriculum dimensions (see variables
of cluster 4, Result 4—Table 2) of feasibility, features necessary for learning, and epistemological
validity. A curriculum design relevant for practitioners therefore should not include only one or
two components but all three. In fact, this transversal trait appears explicitly postulated in research
studies about curriculum design in science education (e.g., Koliopoulos & Ravanis, 2000; Martin
& Solbes, 2001).
Cluster 6 is formed by two variables related to Assessment feature (Result 5—Table 2). These
two variables share the value 0 for the majority of studies. This means that most research studies in
this cluster fail to address the role of the feature Assessment. This stable association of variables
reveals a paradoxical transversal trait, namely the independence of assessment and the learning
process. The important and influential role of assessment is well supported in the literature (e.g.,
Black, 1998; Shepard, 2001). Nevertheless, cluster 6 makes evident that research with explicit
concerns for practical relevance neither emphasizes the importance of assessment nor makes
explicit how summative assessment is related to goals and methods. We consider as an indicator of
the internal consistency of our study the fact that, as shown in Figure 2, cluster 6 lacks a link to
cluster 2, formed by variables from the ‘‘Formative Situation’’ dimension. As shown in Figure 2, if
we choose a linkage distance of greater than 1.2, a connection appears between cluster 1 and
cluster 2. This linkage configures the general transversal trait (Result 2—Table 2) that if a
formative situation (or some of its characteristics) is present, then it is almost sure that the research
is relevant for practice.
The groups of variables type FS (cluster 2) and P (cluster 1) also are connected (see Table 2
and Figure 2). A cluster analysis of this group of 15 variables results in the tree diagram in Figure 4.
Taking a linkage distance approximately equal to the value that we have been using (0.5), we
recognize clusters 1, 2, and 6 (see Table 2), which confirms the internal consistency of the method.
A new cluster appears, formed by two variables of the ‘‘Formative Situation’’ dimension (see
Result 6.1—Table 3), which were not part of any of the previous clusters. This new cluster can
be interpreted as an indicator of an important practical aspect, namely students’ knowledge is
more easily enhanced if real physical (biological, chemical, or other) contexts are used. This
interpretation supports the STS (science–technology–society) studies, which emphasize the role
of contextualized situations in learning.
An examination of the a smaller linkage distance in Figure 4 (see dotted circles) reveals three
subclusters, which are identified in Table 2 as Result 1.1, Result 3.1, and Result 3.2. Result 3.1
suggests ways to improve the teacher’s role as mediator of students’ learning. Mediation is more
effective if instruction begins with students performing rather than beginning with a conversation
between the teacher and students that lacks any common experiential referent for teachers and
students. Results 1.1 and 3.2 present connections among variables that may seem obvious. The
feasibility of a didactic research study is closely related to the methods used, and the resources are
related to the learning environment. The emergence of these obvious results offers additional
support for the internal consistency of the results because the analytic procedures produced both
anticipated results as well as new findings.
We then evaluated the tree diagram obtained with the Formative Situation (FS) and
Curriculum Design (CD) variables (Figure 5).
Journal of Research in Science Teaching
TRANSVERSAL TRAITS 587
Figure 4. Tree diagram for the 35 cases considering only the variables of type P and FS.
Figure 5. Tree diagram for the 35 cases considering only the variables type FS and CD.
Three clusters appear that were identified and interpreted previously, namely clusters 2, 3, and
4 (see Figure 2). At a finer level, Results 3.1 and 3.2 also reappear. The appearance of these clusters
further confirms the internal consistency of the study. A new result (Result 6.2—Table 3) emerges
that complements Result 6.1 (Table 3). Result 6.2 identifies the conditions necessary to enhance
students’ knowledge, namely the use of physical situations and/or the incorporation of knowledge
of other subjects in the school curriculum.
There are clusters that maintain their stability in two additional analyses, specifically an
analysis of all the variables for 17 of 35 cases and an analysis of restricted groups of variables for
all 35 cases. These clusters are interpreted as transversal traits—see Results 1–5 in Table 2.
Cluster 5 reveals instability. From a fine and partial analysis, two clusters are identified that point
to possible transversal traits, specifically Results 6.1 and 6.2 in Table 3. These are compatible but
not coincident with Result 6.3 (see Figure 2 and Table 3). Further analyses may detect complementary
associations. There are stable associations among variables that occur inside clusters (Results 1.1, 3.1,
3.2, and 4.1 in Table 2) that configure more specific transversal traits, which are compatible with the
most general ones identified.
Table 3
Summary of results to explore in future researches
Other Possible Clusters of Variables Summary results of cluster analysis
FS-SK - Aspect of Formative Situation - The paper Result 6.1 - It is a possible transversal trait (to
elucidates/describes Student Knowledge explore in other researches) (see Fig. 4).
mobilized in learning.
FS-PS - Aspect of Formative Situation - The paper
elucidates/describes the Physical Situations used.
FS-SK - Aspect of Formative Situation - The paper Result 6.2 - It is a possible transversal trait (to
elucidates/describes Student Knowledge explore in other researches) (see Fig. 5).
mobilized in learning.
CD-OK - Aspect of Curriculum Design – The paper
discusses the articulation with Other Knowledge
mobilized within classroom.
CD-FS - Aspect of Curriculum Design – The paper Result 6.3 - It is a possible transversal trait (to
discusses and makes explicit the sequence and explore in other researches) (see cluster 5
articulation of teaching activities or Formative in Fig. 2).
Situations.
CD-OK - Aspect of Curriculum Design – The
paper discusses the articulation with Other
Knowledge mobilized within classroom.
FS-SK - Aspect of Formative Situation - The paper
elucidates/describes Student Knowledge
mobilized in learning.
1. The keywords used to select the studies to form the corpus were accurate and effective.
2. The analysis instrument was adequate and functionally operative (Appendix 2) and
allowed us to characterize the 35 studies in the corpus, despite the differences in framework
and perspective that existed among them.
3. The cluster analysis method had heuristic power and allowed us to identify groups of
variables and to use these groups to make visible features common to the different studies
as well as their importance for practice.
4. The analysis withstood tests of internal validation, in particular the reanalysis of 17 of
35 studies and of all and only a subset of the variables. These tests corroborated the stability
and consistency of the clusters, thereby strengthening the claim that interpretations and
inferences were grounded in solid analytic methods.
5. It was possible to identify and characterize transversal traits among studies with different
theoretical orientations and research lines.
indicators point to the use physical situations (chemical, biological, STS, or other) and the
activation of other types of knowledge (CD-OK) in the design of curriculum.
Structuring principle concerning the relationship between assessment and learning. Our
study shows that most studies separate assessment from the learning process, even though this
separation should not occur in practice. Research literature (e.g., Black, 1998; Shepard, 2001) has
recommended the use of assessment to improve students’ learning. Assessment should be another
structural principle for developing students’ knowledge. A valuable direction for future studies
with relevance for practice is to articulate assessment with tasks, resources, and teacher mediation.
It is crucial to articulate the nature of the learning task, the conceptual field, students’ prior
knowledge, the competencies and knowledge to be developed, and guides to managing
the learning tasks.
It is crucial to match the resources available in the classroom with the demands of the
learning tasks, but it is equally important to explore new resources and new learning tasks.
It is imperative that teachers mediate students’ activities and learning. In addition,
teachers should provide different forms of mediation for different students, depending on
their knowledge and the teaching context.
The findings of our study offer implications for initial and in-service teacher training. They
provide a framework for deciding which dimensions of professional development need to be given
high priority.
Reference framework for research. Research in science education intended to influence
practice tends to be holistic, both in theoretical and methodological orientation. Usually, these
research studies are framed from a multidimensional theoretical perspective and describe in detail
the operant processes and contexts. Typically, research studies characterize states rather than
variables and identify tendencies rather than relationships. They often raise epistemological concerns
that must be considered, as well as challenge the researcher to develop reliable evaluation tools (not
necessarily instruments designed to control variables). Our results also show that learning tasks
and the teachers’ mediation of these tasks play a key role in science education research designed to
influence teacher practice. As a general theoretical framework, we propose that research should be
structured along multidisciplinary and multidimensional lines. These lines should include the
learning tasks provided for students, teachers’ mediation of these tasks, and development of
students’ knowledge and competencies.
Journal of Research in Science Teaching
TRANSVERSAL TRAITS 591
Notes
1
Our study is part of a field frequently called ‘‘Didactics of Science’’ and its linguistic
correspondents in German, Danish, Spanish, French, Dutch, Italian, Norwegian, Portuguese, and
Swedish (Sjøberg, 1996). In English, ‘‘didactic/didactics’’ frequently has a more restricted and even,
sometimes, pejorative meaning; in this study, that is not the intended meaning and we more often use
the designation ‘‘science education’’ instead, to facilitate a better understanding in English.
2
Evans (2002) has defined pragmatization of research as ‘‘(. . .) a planned process involving
analysis, presentation and dissemination that is directed at transforming research findings into
viable, specific ideas and recommendations for policy and practice’’ (p. 202).
The authors thank the anonymous reviewers for their helpful suggestions and the
kind and competent support of Professor Frank Crawley in the language revision of the
manuscript.
Crismond D (2001). Learning and using science ideas when doing investigate-and-redesign
tasks: A study of naive, novice, and expert designers doing constrained and scaffolded design
work. Journal of Research in Science Teaching, 38(7), 791–820.
Warren B; Ballenger C; Ogonowski M; Rosebery AS; Hudicourt-Barnes J (2001). Rethinking
diversity in learning science: The logic of everyday sense-making. Journal of Research in Science
Teaching, 38(5), 529–552.
Moje EB; Collazo T; Carrillo R; Marx RW (2001). ’’Maestro, what is ’quality’?‘‘: Language,
literacy, and discourse in project-based science. Journal of Research in Science Teaching, 38(4),
469–498.
Windschitl M (2001). The diffusion and appropriation of ideas in the science classroom:
Developing a taxonomy of events occurring between groups of learners. Journal of Research in
Science Teaching, 38(1), 17–42.
Hart C; Mulhall P; Berry A; Loughran J; Gunstone R (2000). What is the purpose of this
experiment? Or can student s learn something from doing experiments? Journal of Research in
Science Teaching, 37(7), 655–675.
Costa J; Caldeira H; Gallastegui JR; Otero J (2000). An analysis of question asking on
scientific texts explaining natural phenomena. Journal of Research in Science Teaching, 37(6),
602–614.
Mistler-Jackson M; Songer NB (2000). Student motivation and Internet technology:
Are students empowered to learn science? Journal of Research in Science Teaching, 37(5), 459–
479.
Chin C; Brown DE (2000). Learning in science: A comparison of deep and surface
approaches. Journal of Research in Science Teaching, 37(2), 109–138.
Chinn PWU; Hilgers TL (2000). From corrector to collaborator: The range of instructor roles
in writing-based natural and applied science classes. Journal of Research in Science Teaching,
37(1), 3–25.
Does the study show/elucidate the environment in which the learning takes place, namely the types of
activities, work organization and the degree of freedom granted to students?
III.2.3 Tasks
Does the study describe the tasks performed by students and elucidate the role these activities play in the
current learning?
III.2.4 Physical situations
Does the study describe the physical situations to which the tasks performed are reported, or other teaching
activities, and does it elucidate the role they play in the current learning?
III.2.5 Resources
Does the study describe the resources (theoretical or material) made available to the students, in which way
they were used and does it indicate their role in the current learning?
III.2.6 Mediation
Does the study describe and elucidate the interaction teacher-students (and student-student), the negotiation
of meanings, the help given, the questions asked to deepen knowledge, the chirurgical synthesis and does it
indicate their role in the current learning situation?
III.2.7 Knowledge developed
Does the study show/elucidate the knowledge and/or skills and competences and/or values and attitudes
that were developed with the proposed teaching activities?
III.2.8 Assessment
The study describes and elucidates assessment instruments and/or practices, indicating their role in the
forms of mediation used in the current learning situation?
P FS CD P-RL P-K P-VA P-SC P-A P-M P-F FS-SK FS-LE FS-T FS-PS FS-R FS-M FS-DK FS-A CD-EV CD-F CD-L CD-FS CD-OK
JRST-Seiler et 1 1 0 0 1 1 1 0 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0
IJSE-Tunnicliffe 1 1 0 0 1 1 1 0 1 1 0 1 1 1 1 0 1 0 0 0 0 0 0
SE-Hayes et 1 1 0 1 0 1 1 1 1 1 0 1 1 1 0 1 1 1 0 0 0 0 0
JRST-Hart et 1 1 1 1 0 0 1 1 1 1 0 1 1 0 0 1 1 1 1 0 0 1 1
IJSE-Bennett et 0 0 0 0 0 0 0 1 1 1 0 0 0 0 0 0 0 1 0 0 0 0 0
JRST-Chin et 0 1 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 0 0 0 0 0 0
SE-Dori et 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
JRST-Warren et 1 1 0 1 0 0 1 0 1 1 1 1 1 1 0 1 1 0 1 0 0 1 1
IJSE-Bell et 1 1 1 1 1 0 1 0 1 0 1 1 1 0 1 1 1 1 0 0 1 1 1
IJSE-Davis et 1 1 0 1 1 0 1 0 1 1 0 1 1 0 1 1 1 1 0 0 0 1 1
SE-Polman et 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 1 1 1 0 1
JRST-Chinn 1 1 0 1 1 1 1 1 1 1 0 1 1 0 0 1 1 1 0 0 0 1 1
SE-Blank 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 0 1 1 0 1
JRST-Cavallo et 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1 1 0 0 1
JRST-Mistler et 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1 1 1 1
JRST-Bouillion et 1 1 1 1 1 1 1 0 1 1 1 1 0 1 1 0 1 0 1 1 1 1 1
SE-Koch 1 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1 1 1 1 0 1 0
IJSE-Fouzder et 1 1 1 1 1 1 1 0 1 1 1 1 1 0 1 1 0 0 0 1 1 0 1
IJSE-Simoneaux 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 1
IJSE-Rodrigues 1 1 1 1 1 0 1 0 1 1 0 1 1 0 1 1 1 0 1 1 1 0 1
IJSE-Buncick et 1 1 1 1 1 1 1 0 1 1 0 1 1 1 1 1 1 0 1 1 1 1 0
IJSE-Stodart et 1 0 0 1 1 1 1 1 1 1 1 0 1 0 0 0 0 0 0 0 0 0 0
IJSE-Yip 1 0 1 1 1 1 1 0 1 1 0 0 0 0 0 0 0 0 1 1 1 1 1
IJSE-Tsai 1 0 1 1 1 1 1 1 1 1 1 0 0 0 0 0 1 0 1 1 1 0 1
IJSE-Buckley 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 0 1 1 1 1 1
TRANSVERSAL TRAITS 597
References
Alsop, S., Benzce, L., & Pedretti, E. (2005). Analysing exemplary science teaching:
Theoretical lenses and a spectrum of possibilities for practice. London: Open University Press.
Adúriz-Bravo, A., & Izquierdo, M. (2002). Acerca de la didáctica de las ciencias
como disciplina autónoma [About science education as an autonomous discipline]. Revista
Electrónica de Enseñanza de las Ciencias, 1 (3). Retrieved June 14, 2006, from http:/ /
www.saum.uvigo.es/reec.
Adúriz-Bravo, A., Duschl, R., & Izquierdo, M. (2003). Science Curriculum development as a
technology based on didactical knowledge/El desarrollo curricular en ciencias como una
tecnologı́a baseada en el conociniento didáctico. Journal of Science Education/Revista de
Educación em Ciencias, 4, 64–69.
Astolfi, J.-P., Darot, E., Vogel, Y., & Toussaint, J. (1997). Pratiques de Formation en
Didactique des Sciences [Teacher Education in Science Education]. Paris: De Boeck & Larcier.
Bennett, J. (2003). Teaching and learning science. A guide to recent research and its
application. London: Continuum.
Black, P. (1998). Assessment by teachers and the improvement of students’ learning. In B.J.
Fraser & K.G. Tobin (Eds.). International handbook of science education (pp. 811–822).
Dordrecht: Kluwer.
Botton, C., & Brown, C. (1998). The reliability of some VOSTS items when used with preservice
secondary science teachers in England. Journal of Research in Science Teaching, 35, 53–71.
Branco, J. (2005). Propriedades e Aplicaçes da Luz: Concepção de um currı́culo didáctica e
epistemologicamente fundamentado e sua avaliação [Properties and applications of light: Design
and evaluation of a curriculum didactically and epistemologically grounded]. Unpublished
master’s thesis, Universidade de Trás-os-Montes e Alto Douro, Vila Real, Portugal.
Cachapuz, A., Praia, J., & Jorge, M. (2000). Perspectivas de Ensino das Ciências
[Perspectives in science teaching]. Porto: Centro de Estudos de Educação em Ciência.
Cachapuz, A., Praia, J., Gil-Perez, D., Carrascosa, J., & Martinez-Terrades, F. (2001). A
emergência da didáctica das ciências como campo especı́fico de conhecimento [The emergence of
science education as a specific field of knowledge]. Revista Portuguesa de Educação.
Universidade do Minho, 14, 155–195.
Costa, N. (2003). A Investigação Educacional e o seu Impacto(e) nas Práticas Educativas: o
caso da investigação em Didáctica das Ciências [Education research and its impact in educational
practices: The case of Science Education Research]. Unpublished manuscript, Universidade de
Aveiro, Aveiro, Portugal.
Costa, N., Marques, L., & Kempa, R. (2000). Science teachers’ awareness of findings from
education research. Research in Science & Technological Education, 18, 37–44.
Costa, N., Praia, J., & Marques, L. (1999). Educação em Ciências: binómio Investigação-
Formação. Uma experiência de investigadores portugueses [Science Education: Research–
teacher education binomial. An experience of Portuguese researchers]. In C. Linhares & I.
Fazenda & V. Trindade (Eds.), Os lugares dos Sujeitos na Pesquisa Educacional (pp. 305–323).
Campo Grande, Brazil: Editora UFMS.
Cravino, J.P. (2004). Ensino da Fı́sica Geral nas Universidades Públicas Portuguesas e sua
Relação com o Insucesso Escolar—Caracterização do Problema e Desenho, Implementação e
Avaliação de uma Intervenção Didáctica [Teaching introductory physics in Portuguese public
universities and academic failure—characterization of the problem and design, implementation
and evaluation of a didactical intervention]. Unpublished doctoral dissertation, Universidade de
Trás-os-Montes e Alto Douro, Vila Real, Portugal.
Journal of Research in Science Teaching
598 LOPES ET AL.
Dumas-Carré, A., & Goffard, M. (1998). Objectivation des pratiques de tutelle d’un
enseignant au cour des séances de résolution de problèmes en Physique [Objectivation of
mediation practices during physics problem solving sessions]. In A. Dumas-Carré & A. Weil-
Barais (Eds.), Tutelle et Médiation dans l’éducation scientifique. Bern: Peter Lang.
Evans, L. (2002). Reflective practice in educational research: Developing advanced skills.
London: Continuum.
Gilbert, J. (2002). Science education and research. In S. Arons & R. Boohan (Eds.), Teaching
science in secondary schools (pp. 217–222). London: Routledge Falmer.
Gil-Perez, D., Furió-Mas, C., Valdés, P., Salinas, J., Martinez-Torregrosa, J., Guisáosla, J.,
González, E., Dumas-Carré, A., Goffard, M., & Pessoa de Carvalho, A. (1999). Tiene sentido
seguir distinguiendo entre aprendizaje de conceptos, resolución de problemas de lápiz y papel y
realización de prácticas de laboratorio? [Does it make sense to go on making a distinction among
conceptual learning, pencil and paper problem solving and laboratory practical work?].
Enseñanza de las Ciencias, 17, 311–320.
Gwimbi, E.M., & Monk, M. (2003). Study of classroom practice and classroom contexts
amongst senior high school biology teachers in Harare, Zimbabwe. Science Education, 87, 207–
223.
Hammersley, M. (2002). Educational research, policymaking and practice. London: Sage.
Howe, A.C. (1996). Development of science concepts within a Vygotskian framework.
Science Education, 80, 35–51.
Ireson, G. (1999). A multivariate analysis of undergraduate physics students’ conceptions of
quantum phenomena. European Journal of Physics, 20, 193–199.
Kelly, A., & Lesh, R. (2000). Handbook of research design in mathematics and science
education. Mahwah, NJ: Lawrence Erlbaum.
Kempa, R. (1992). Research in chemical education: Its role and potential. In chemistry for
science teachers (pp. 45–67). Milton Keynes, UK: Open University Press.
Kempa, R. (2002). Research and research utilisation in chemical education. Chemistry
Education: Research and Practice in Europe, 3, 327–343.
Kjærnsli, M., & Lie, S. (2004). PISA and scientific literacy: Similarities and differences
between the Nordic countries. Scandinavian Journal of Educational Research, 48, 271–286.
Koliopoulos, D., & Ravanis, K. (2000). Élaboration et évaluation du contenu conceptuel d’un
curriculum constructiviste concernant l’approche énergétique des phénomènes mécaniques
[Elaboration and evaluation of the conceptual content of a constructivist curriculum about an
energy approach of mechanic phenomena]. Didaskalia, 16, 33–56.
Linn, M. (1987). Establishing a research for science education: Challenges, trends and
recommendations. Journal of Research in Science Teaching, 24, 191–216.
Lopes, J.B. (2004). Aprender e Ensinar Fı́sica [Learning and teaching physics]. Lisboa:
Fundação Calouste Gulbenkian.
Lopes, J.B., & Costa, N. (2007). The evaluation of modelling competences: Difficulties
and potentials for the learning of the sciences. International Journal of Science Education, 29,
811–851.
Martin, J., & Solbes, J. (2001). Diseño y Evaluatión de una propuesta para la Enseñanza del
concepto de campo en Fisica [Design and evaluation of a proposal for teaching the concept of field
in physics]. Enseñanza de las Ciencias, 19, 393–403.
Martinez-Terrades, F. (1998). La didactica de las ciencias como campo especı́fico de
conocimientos. Génesis, estado actual y perspectivas [Science Education as specific field of
knowledge. Genesis, state of the art and perspectives]. Unpublished doctoral dissertation,
Universitat de València, Valencia, Spain.
Journal of Research in Science Teaching
TRANSVERSAL TRAITS 599