Science

Education

Writing to Learn by Learning to

Write During the School Science
Laboratory: Helping Middle and
High School Students Develop
Argumentative Writing Skills as
They Learn Core Ideas

VICTOR SAMPSON,1 PATRICK ENDERLE,2

JONATHON GROOMS,2 SHELBIE WITTE3
1
School of Teacher Education and FSU-Teach, 2 College of Education, and 3 School of
Teacher Education, The Florida State University, Tallahassee, FL 32306, USA

Received 10 October 2012; accepted 16 May 2013

DOI 10.1002/sce.21069
Published online 14 August 2013 in Wiley Online Library ([Link]).

ABSTRACT: This study examined how students’ science-specific argumentative writing

skills and understanding of core ideas changed over the course of a school year as they
participated in a series of science laboratories designed using the Argument-Driven Inquiry
(ADI) instructional model. The ADI model is a student-centered and writing-intensive
approach to laboratory instruction. The study was conducted in two middle school and
two high school courses offered at a university-affiliated laboratory school located in the
southeast USA. The intervention took place over two semesters and consisted of at least
eight laboratory activities in each course. Student learning gains were measured using a
science content assessment and a science-specific argumentative writing assessment that
were administered at the beginning, middle, and end of the school year. Changes in student
performance on the two assessments over time indicate that the students’ science-specific

Correspondence to: Victor Sampson; e-mail: vsampson@[Link]

Contract grant sponsor: Institute of Education Science, U.S. Department of Education.
Contract grant number: R305A100909.
The opinions expressed are those of the authors and do not represent the views of the Institute or the
U.S. Department of Education.


C 2013 Wiley Periodicals, Inc.
644 SAMPSON ET AL.

argumentative writing skills and their understanding of core scientific ideas improved over
the course of intervention. Furthermore, students who participated in a greater number of
ADI activities demonstrated greater and more consistent improvement in their writing. The
implications of the study describe ways to enhance students’ science proficiencies during
school science laboratories and include recommendations for future research.  C 2013

Wiley Periodicals, Inc. Sci Ed 97:643–670, 2013

INTRODUCTION
The formal science curriculum should include opportunities for middle and high school
students to write argumentative texts in science for two important reasons. First, writing
argumentative texts is an important scientific practice (Baker et al., 2008; Carter, 2007;
Pratt & Pratt, 2004; Wallace, Hand, & Prain, 2004; Yager, 2004). Scientists, for example,
must be able to write an argumentative text to communicate the nature and results of their
investigations, and how well they write often influences their future success in science
(Barrass, 2002; Yore, 2004). Middle and high school students should therefore be given a
chance to learn the argumentative writing skills that are used by scientists and individuals
in science-related careers to share and justify new ideas (Breger, 1995; Connolly & Vilardi,
1989; National Governors Association Center for Best Practices and Council of Chief
State School Officers, 2010; National Research Council [NRC], 2012; Neal & Langer,
1992; Yore, Bisanz, & Hand, 2003). Second, the process of writing an argumentative
text can help students develop a better understanding of core ideas of science. Students
often understand the content better after writing an argumentative texts because the mental
activities associated with this type of writing, which include reflection, consolidation,
elaboration, interpretation, synthesis, and justification, requires the writer to think deeply
about the content and why it is valid or acceptable (Connolly & Vilardi, 1989; Glynn &
Muth, 1994; Hand, 2004; Yore et al., 2003).
Despite the potential benefits of integrating argumentative writing into the science
curriculum, current research indicates that science teachers do not devote much instruc-
tional time to it (Fulwiler, 2008; Kiuhara, Graham, & Hweken, 2009; Weiss, Banilower,
McMahon, & Smith, 2001; Yore et al., 2003). Most middle and high school students, as
a result, are not given an opportunity to learn how to write a sound scientific argument or
how to share information about the process or the theoretical underpinnings of an inquiry
in a manner that is consistent with the norms and epistemological commitments of science
(Indrisano & Paratore, 2005; Kelly, Regev, & Prothero, 2008; Saul, 2004). Many students
therefore struggle to communicate their ideas, support a claim with evidence, coordinate
evidence and theory, or provide an adequate challenge to an alternative claim when they
are asked to craft an argumentative text in the context of science (Kelly et al., 2008; Kelly
& Bazerman, 2003; Kelly & Takao, 2002). To complicate matters further, many science
teachers are reluctant to teach students how to write science-specific argumentative texts
because they claim that the amount of instructional time that is required to help students
learn to write will only decrease the amount of time available to address the content they
need to “cover” in an already overcrowded curriculum, or they feel unqualified to teach stu-
dents how to write in science (Galbraith & Torrance, 1999; Holliday, Yore, & Alvermann,
1994).
One way to help address these issues is to develop instructional models that teachers can
use to help students develop science-specific argumentative writing skills and learn about
the core ideas of science at the same time. An example of one such instructional model is
called Argument-Driven Inquiry or ADI (Sampson, Grooms, & Walker, 2011). The ADI
instructional model is not designed to be a curriculum; rather, it is designed to serve as a
Science Education, Vol. 97, No. 5, pp. 643–670 (2013)
 WRITING TO LEARN BY LEARNING TO WRITE IN SCIENCE 645

template or a guide that instructors can use to create school science laboratory experiences
that are more authentic and educative for students. This model is designed to be a more
authentic approach to laboratory instruction than the “cookbook”-style activities that are
often used by teachers (NRC, 2005) because it provides students with an opportunity to
engage in the practices of science. Students who participate in laboratory experiences that
are consistent with the ADI instructional model have an opportunity to design their own
investigations, engage in scientific argumentation as they develop and critique arguments,
write a report about their investigation for a critical and knowledgeable audience, participate
in the peer review process, and revise the report based on the critiques offered by the
reviewers. ADI is also intended to be more educative than the status quo because several
mechanisms that are designed to promote and support student learning have been built into
the model.
In this article, we will first describe the theoretical framework that serves as a foundation
for the design of the ADI instructional model and then provide a review of current research
that examines the relationship between writing and student learning. We will then share
the hypothesis underlying our research, the predictions we made in light of our hypothesis
and the nature of the study, and our specific research questions. Next we will describe the
method we used during our study and our findings. We will then conclude the article with
a discussion of our conclusions and the implications of our research.

THEORETICAL FRAMEWORK
The ADI instructional model is grounded in social constructivist theories of learning
(see Anderson, 2007; Driver, Asoko, Leach, Mortimer, & Scott, 1994; Scott, Asoko, &
Leach, 2007). These theories suggest that learning in science is both social and personal in
nature. The social process of learning in science involves being introduced to how science
is different from other ways of knowing. This process requires input and guidance about
“what counts” from people who are familiar with the ideas, modes of inquiry, standards
of argument, and ways of communicating that define science. Thus, learning in science is
dependent on educative and supportive interactions with people that know about science.
The individual process of learning in science, on the other hand, involves the construction
of knowledge through the appropriation or accommodation of important ideas, modes of in-
quiry, habits of mind, and modes of communication. This requires individuals to make sense
of their experiences and to integrate new views with old views. This theoretical framework
also stresses that learning how to engage in specific scientific practices (such as arguing
from evidence or writing an argumentative text) and how to use specific tools (such as using
specific types of criteria to evaluate the quality of an argument or using important concepts
or models to make sense of natural phenomena) when engaged in these practices is a fun-
damental part of what students need to learn inside the science classroom (Duschl, 2008).
There are several implications that stem from this theoretical framework for current
efforts to help students learn core scientific ideas and write argumentative texts in science.
First, students need to have opportunities to write argumentative texts to learn about them.
Second, students must develop an understanding of the norms and standards that are used to
evaluate the quality argumentative writing in the context of science. It is therefore important
for opportunities to write in science to be educative for students. Instructors can make an
opportunity to write in science more educative for students by showing them what counts
as good science writing (i.e., modeling), providing learners with reminders about the norms
and standards that are used to judge quality in science (i.e., scaffolding), providing students
with feedback about their performance (i.e., coaching), and encouraging students to be
metacognitive as they refine their text over several iterations (Sampson & Walker, 2012).
Science Education, Vol. 97, No. 5, pp. 643–670 (2013)
646 SAMPSON ET AL.

The nature of the writing task that students are expected to do inside the classroom is
also an important issue to consider because the nature of a writing task influences what
students learn from it (Sampson & Walker, 2012). Tasks, for example, need to be as
authentic as possible so students view writing as an important aspect of “doing science”
rather than “doing school” (Jimenez–Aleixandre, Rodriguez, & Duschl, 2000). The writing
that students do should also be linked with opportunities to engage in other scientific
practices, such as designing investigations or analyzing data, so students see the important
role writing plays in science and in the scientific community (Carter, Ferzli, & Wiebe, 2007;
NRC, 2012; Wallace et al., 2005; Yore et al., 2003). Finally, and perhaps most importantly,
the writing tasks that students complete need to focus on “knowledge building” rather than
“knowledge telling” (Bereiter & Scardamalia, 1987). This type of focus is important when
the goal of writing is to help students learn about the core ideas of science (Hand & Prain,
2006; Yore, 2004; Yore et al., 2003).

WRITING IN SCIENCE CLASSROOMS: A BRIEF REVIEW OF THE

LITERATURE
There have been numerous calls over the past 30 years to integrate more writing into
the middle and high school science curriculum. As a result, science and literacy educators
have developed numerous approaches that science teachers can use to give students more
opportunities to write in science. Some of these approaches are well aligned with our
theoretical framework and some are not. In this section, we will provide a brief overview
of some of these various approaches. We will also discuss some of the affordances and
constraints of these approaches in terms of helping students developing their science-
specific argumentative writing skills and their understanding of core scientific ideas.
One approach to integrate writing into the teaching and learning of science is to introduce
students to the language conventions of science and the specific formats (e.g., nonnarrative,
multimodal texts) that are used by scientists as part of the formal science curriculum
(Alley, 1996; Halliday & Martin, 1993; Indrisano & Paratore, 2005; Saul, 2004). This
type of approach is often described as learning to write (LtoW) because the main goal
of the teacher is to help students learn a specific but generalizable set of skills that are
needed to write well in science. Students in classrooms where the emphasis is placed on
LtoW in science, as a result, are often taught how to produce texts that are important
in science (e.g., reports, reviews, abstracts) at the beginning of a course and often, but
not always, in decontextualized manner through some sort of framework or instructional
scaffold that makes the characteristics of high-quality writing explicit (see Berber-Jimenez,
Montelongo, Hernandez, Herter, & Hosking, 2008; Friend, 2002; Porter et al., 2010 for
illustrative examples). This type of approach therefore tends to be process focused rather
than content focused.
Current research suggests that this type of approach can help students learn how to
write better in science (Fallahi, Wood, Austad, & Fallahi, 2006; Galbraith & Rijlaarsdam,
1999; Halliday & Martin, 1993; Rijlaarsdam, Couzijn, Janssen, Braaksma, & Kieft, 2006).
In addition, this type of approach can be integrated into a curriculum easily because it
is not tied to specific content, which some argue is an affordance of the approach. This
approach, however, has a number of constraints associated with it. First, it does not give
students an opportunity to learn core scientific ideas or how to use writing as way to make
sense of what they are seeing and doing inside the science classroom (Hand, 2004; Prain,
2006). Second, students often view these types of writing tasks as doing school rather than
doing science (Jimenez–Aleixandre et al., 2000), which can decrease students’ motivation
and have a negative affect on students’ attitudes toward writing (Klein & Kirkpatrick,
Science Education, Vol. 97, No. 5, pp. 643–670 (2013)
 WRITING TO LEARN BY LEARNING TO WRITE IN SCIENCE 647

2010). Third, the gains made by students are often rather small when teachers use this
type of approach (see Gerdeman, Russell, & Worden, 2007; Walvoord, Hoefnagels, Gaffin,
Chemchal, & Long, 2008 as examples). Finally, and perhaps most importantly, the writing
that is associated with LtoW is often not realistic, even though students are expected to
produce texts that are valued in science, because the writing is rarely embedded into the
inquiry process and it is not used by students for the purpose of knowledge building (as it
is in the scientific community).
Another way to promote and support writing in the science classroom is to encourage
students to write to help them develop a better understanding of the content (Connolly
& Vilardi, 1989; Prain, 2006; Saul, 2004; Wallace et al., 2004). This approach is often
described as writing to learn (WtoL) in science because the goal is to help students
learn content by having them complete writing tasks that require them to explain, reflect,
elaborate, and synthesize the laws, theories, principles, and concepts that are introduced to
them in class. In WtoL classrooms, as a result, students are often assigned to write journal
entries in a “learning log” or to write summaries of a text in nonacademic language (Heuser,
2005; Miller & Calfee, 2004; Timmerman, 2002; Towndrow, Ling, & Venthan, 2008; Turner
& Broemmel, 2006) but are not expected to write texts that reflect the nature of the texts
scientists use to share and justify claims. Students are encouraged to complete these types
of writing tasks because writing is viewed as a vehicle that will help students develop a
better understanding of the content (Applebee, 1984; Gunel, Hand, & McDermott, 2009;
Prain, 2006; Sakai & Leggo, 1997). This type of approach therefore tends to be viewed as
content focused rather than process focused because the development of scientific writing
skills is seen as a secondary or even tertiary goal.
Current research suggests that this type of approach, in general, can help students learn
content (Bangert–Drowns, Hurley, & Wilkson, 2004; Rivard, 1994). In addition, teachers
can easily integrate this type of writing into a curriculum because it requires only a small
amount of instructional time during each lesson. It can therefore be used on a regular basis
without sacrificing other learning activities. The flexibility and the ease to which it can
be integrated into a lesson are affordance associated with WtoL. This approach, however,
also has a number of constraints associated with it. First, and foremost, there is often little
focus on the nature of scientific writing in general or the development of science-specific
argumentative writing skills in particular. Students therefore have few opportunities to learn
the purpose, assumptions, procedures, and “hidden” rules of scientific writing or to adopt
the norms and standards that are used to assess the quality of explanations and arguments in
science. The writing tasks that focus on journal writing or summarizing the key points also
do not engage students in the serious writing practices of science. This type of approach, as
a result, will likely not promote the development of science-specific argumentative writing
skills that students need to be able to craft sound and persuasive arguments in science.
Another approach is to engage students in realistic writing tasks as part of the inquiry
process to promote the understanding of content and the development of science writing
skills at the same time. In order for a writing task to be realistic, however, it must do more
than require students to use the formats and language conventions of science (as is the case
with the LtoW approach). The writing task also needs to enable students to use writing
as a way to build or transform knowledge rather than transmit it (as is the focus of the
WtoL approach). Students can be encouraged to make sense of the phenomenon under
investigation during a realistic writing task by requiring them to engage in the serious
writing practices of science. In the context of science, serious writing practices include, but
are not limited to, developing explanations, transforming data into evidence, arguing from
evidence, coordinating theory and evidence, and revising their claim or ideas in light of
critique.
Science Education, Vol. 97, No. 5, pp. 643–670 (2013)
648 SAMPSON ET AL.

Kelly et al. (2002) argue that students must engage in the serious writing practices
of science as part of the inquiry process to transform the writing they do in science
classrooms from “a simple formal approach to science, to active work with scientific
evidence, knowledge, and concepts” (p. 3). They also suggest that realistic tasks that
engage students in the serious writing practices of science allow students to learn

. . . what kinds of evidence is needed to warrant arguments and how that evidence can be
appropriately developed, analyzed, and interpreted given community standards; what kinds
of concepts are appropriately evoked; and what kind of stance authors can appropriately
take as contributors to their fields. (Kelly et al., 2002, p. 3)

This type of approach to writing instruction, as a result, enables students to understand the
language and discourse of science as well as the content under investigation. It therefore
goes beyond LtoW, which views writing as a content-independent skill to be mastered, and
a WtoL approach, which views writing as a technique that can be used to help students
understand or recall content (Bazerman, 1988; Berkenkotter & Huckin, 1995; Carter et al.,
2007). Thus, an approach that is designed to give students an opportunity to complete
realistic writing tasks and engage in the serious writing practices of science in a more
immersive fashion is often described as “writing to learn by learning to write” (Carter et al.,
2007).
An example of a “writing to learn by learning to write” approach is the science writing
heuristic (SWH). The SWH integrates discussion, reading, and opportunities to engage in
the serious writing practices of science into the process of scientific inquiry. A number
of studies suggest that the SWH helps students understand the content at the heart of the
inquiry (Hand, Hohenshell, & Prain, 2007; Poock, Burke, Greenbowe, & Hand, 2007).
There has been far less research; however, that has examined how this approach, and others
like it, influences the development of students’ argumentative writing skills in science (see
Gleason, 1999; Glynn & Muth, 1994 for examples of other approaches). As a result, little is
known about the impact of approaches that emphasize writing to learn by learning to write on
the development of science-specific writing skills or the nature of the modeling, coaching,
scaffolding, and feedback that students need as they learn how to engage in the serious
writing practices of science. Therefore, research that seeks to understand the interaction
between instruction that emphasizes writing to learn by learning to write, as facilitated
through models such as ADI, and students’ science learning and writing skills warrants
exploration, particularly in light of current national reform movements that intertwine these
constructs for mutual enhancement.

HYPOTHESIS, PREDICTIONS, AND THE RESEARCH QUESTIONS

Our hypothesis is that the various activities and tasks that are embedded into the ADI
instructional model provide a mechanism that can improve students’ argumentative writing
skills along with their understanding of core scientific ideas. This mechanism consists of
four interrelated components that are well aligned with the major tenets of our theoretical
framework and reflects current research about writing to learn by learning to write in the
context of science. First, ADI provides a reason for students to learn how to write in science
because students are required to complete a realistic writing task and engage in the serious
writing practices of science as part of the inquiry process. Second, ADI provides students
with an opportunity to read good examples of reports that were written in the same style
and with the same overall goal in mind. This practice is common in science (Yore, 2004);
scientists often read the journals they write for before, during, and after an investigation
Science Education, Vol. 97, No. 5, pp. 643–670 (2013)
 WRITING TO LEARN BY LEARNING TO WRITE IN SCIENCE 649

because the content informs their research and the articles help them improve their own
writing (i.e., use of terms, style, and conventions). Third, the instructional model provides
students with educative information about their current understanding of the content, the
validity of their claim, and the quality of their writing in light of a set standard for “what
counts” as valid or acceptable. Fourth, this instructional model gives students opportunities
to revise their writing in light of critique. This mechanism, as a result, is designed to help
students become more attuned to “what is and what is not acceptable” in science and enable
them to develop a more fluid “grasp of practice” (Ford, 2008) that will allow them to use
their argumentative writing skills in different contexts or in novel situations. This approach
should also help students understand the content they write about because engagement in
the serious writing practices of science requires the writer to think about what they know
and how they know it (Kelly et al., 2002; Yore et al., 2003).
To test this hypothesis, we decided to conduct a study in the classrooms of six sci-
ence teachers. At the time of this study, two of these teachers taught high school chem-
istry, two taught high school biology, one taught middle school physical science, and
one taught middle school life science. These teachers helped us to identify eight fun-
damental concepts that are addressed each year in each course. We then developed a
content assessment for each course that targeted the eight important concepts as well as
a science-specific argumentative writing assessment for each course. Next, we used the
ADI instructional model to develop two lab investigations for each fundamental concept in
collaboration with the classroom teachers who were responsible for teaching the course.1
The classroom teachers then agreed to implement the 16 lab activities as designed over
the course of the school year while the research team tracked the number of lab activi-
ties that were actually implemented, how they were implemented by the teachers, and the
gains students made in their content knowledge and science-specific argumentative writing
skills.
We predicted, given our hypothesis and the nature of the intervention, that the students
enrolled in these courses would make substantial and statistically significant gains on the
assessments of content and science writing over the course of the year if the underlying
mechanism functioned as intended. We also predicted that the students who completed
more of the ADI lab investigations (as designed) would make the greatest gains. Thus, our
objective for this study was to compare these predictions to what actually happened in these
middle and high school courses and to what the students actually learned to assess the merits
of the hypothesis underlying the design of the current iteration of the ADI instructional
model. We were also interested in documenting the challenges that both the teachers and
the students faced when ADI was used in middle and high school science laboratories
because of the exploratory nature of this study. The specific research questions that guided
this study were as follows:

1. How do students’ understanding of important content and their science-specific argu-

mentative writing skills change over time when they participate in a series of school
science lab activities that were designed to emphasize “writing to learn by learning
to write”?
2. Which aspects of science-specific argumentative writing did the middle and high
school students improve on the most and which aspects appear to be the most chal-
lenging for them?

1
One lab investigation was designed to introduce the concept and one was designed to give students a
chance to apply the concept in an unfamiliar situation.

Science Education, Vol. 97, No. 5, pp. 643–670 (2013)

650 SAMPSON ET AL.

TABLE 1
Number and Demographics of the Participants
Number of Participants

Sex Race

Course Total Male Female White Black Latino/a Asian Multiracial

Life science 67 31 (46%) 36 (54%) 31 (46%) 18 (27%) 8 (12%) 5 (7%) 5 (7%)
Physical 52 27 (52%) 24 (48%) 25 (48%) 20 (38%) 6 (12%) 0 (0%) 1 (2%)
science
Biology 94 49 (52%) 45 (48%) 49 (52%) 32 (34%) 5 (5%) 4 (4%) 4 (4%)
Chemistry 81 36 (44%) 45 (56%) 47 (59%) 21 (26%) 9 (11%) 1 (1%) 3 (4%)

METHOD
Participants
All of the participants in the study were recruited from a group of students that were
enrolled in the life science course (seventh grade), physical science course (eighth grade),
biology course (9th or 10th grade), or chemistry course (10th or 11th grade) offered at a
K-12 university laboratory school located in the southeast United States at the beginning
of the 2010–2011 academic year. A total of 401 students agreed to take part in this study;
however, owing to high student attrition during the year and student absenteeism on the
days we administered our assessments, a complete data set was collected from 294 (73%)
of these students. An overview of the participants is provided in Table 1.

Procedure
The broader context of this research aimed to refine the ADI instructional model so that
teachers can use it within the context of an existing middle or high school science curriculum
to provide a high-quality laboratory experience for their students. We are using an iterative
outcome-focused approach that is consistent with the major tenets of design-based research
(Brown, 1992; Brown & Campione, 1996; The Design-Based Research Collective, 2003)
to develop and refine the ADI instructional model through several iterative cycles of design,
enactment, analysis, and redesign.

The Stages of Argument-Driven Inquiry

The iteration of the ADI instructional model that was used in this study consists of eight
steps (see Figure 1). In the paragraphs that follow, we will provide a brief overview of each
step. We will, however, devote more space to the steps that are designed to promote and
support the development of argumentative writing skills in science given the focus of this
article.
The first stage is the identification of the task and a guiding question by the instructor.
In the second stage, the students must work in groups to design and implement a method to
collect the data they need to answer the guiding question. The students then analyze their
data and develop a tentative argument (a claim supported by evidence and a justification
of the evidence) during the third stage of the model. Students then share their arguments
and critique the arguments of their peers during an argumentation session that takes place
during the fourth stage of the model.
Science Education, Vol. 97, No. 5, pp. 643–670 (2013)
 WRITING TO LEARN BY LEARNING TO WRITE IN SCIENCE 651

Figure 1. The stages of the ADI instructional model at the time of this study.

The fifth step of ADI is the creation of a written investigation report by each student. This
report gives the students a reason to share the goals and method of their investigation along
with their overall argument. We use a nontraditional report format organized around three
fundamental questions: What were you trying to do and why?, What did you do and why?,
and What is your argument? The intent of this format is to bring the argumentative nature
of science writing to the foreground and to highlight the nonnarrative structure (i.e., text
that is hierarchical or clustered by topic) and the multimodal presentation of information
(e.g., words, figures, tables, equations) that characterize science texts (Bazerman, 1988;
Carter, 2007; Wallace et al., 2004).
The sixth stage of ADI is a double-blind group peer review of these reports. Once
students complete their investigation reports, they submit four typed copies with only an
identification number to the instructor. The instructor then randomly distributes three or
four sets of reports (i.e., the reports written by three or four different students) to each
lab group along with a peer review sheet for each set of reports. The peer review sheet
includes specific criteria to be used to evaluate the quality of an investigation report and
space to provide feedback to the author. The lab groups review each report as a team and
Science Education, Vol. 97, No. 5, pp. 643–670 (2013)
652 SAMPSON ET AL.

then identify the aspects of it that need to be revised based on the criteria provided on the
peer review sheet. Groups are also directed to provide explicit feedback to the author about
what needs to be done to improve the quality of the report and their writing as part of the
review.
The seventh stage of the ADI instructional model is the revision of the report. The
students are required to rewrite their reports based on the reviewers’ comments and sug-
gestions. Once completed, the revised reports (along with the original version of the report
and the peer review sheet) are submitted to the instructor for a final evaluation. This
approach is intended to provide an opportunity for students to improve their mechan-
ics, reasoning, and their understanding of the content without imposing a grade-related
penalty. It also encourages students to adopt new standards for “what counts” as quality
in the context of science, help students to be more metacognitive about what they are
doing when they write, and encourage students to treat writing as a process of iterative
refinement.
The instructional sequence concludes with the teacher leading a reflective and explicit
discussion. The goal of the teacher during this stage of the model is to encourage the
students to reflect on what they did and learned during their investigation. The teacher also
uses this time to discuss relevant aspects of the nature of science and the nature of scientific
inquiry, using the investigation as an illustrative example.

Implementation of ADI by the Six Different Teachers

Although all of the teachers experienced similar professional development regarding the
ADI instructional model in the summer prior to this study, implementation of the model
varied across courses. The chemistry teachers, for example, decided to truncate several
stages of the model related to the writing and revision of the reports to save time. The
chemistry teachers also implemented fewer ADI lab activities than the biology teachers.
As a result, some teachers gave their students’ more opportunities to develop their writing
abilities and engage deeply with content through writing than others. We expected this type
of variation in implementation at the onset of the study given the substantial literature that
indicates that teachers will interpret any curriculum or instructional model in unique ways,
which will result in difference in the way they choose to enact it (Blanchard et al., 2010;
Smith & Southerland, 2007). It is important to note, however, that we did not see this kind
of variation in teacher implementation as a potential limitation or as a flaw in the design of
the study (as is the case in a true experiment; see Murnane & Willett, 2011 for a discussion
of this issue). Rather, we viewed it as an opportunity to link differences in implementation
of ADI to observed differences in students’ learning gains. The variation in implementation
across courses, as a result, provided us with a “natural experiment” (Murnane & Willett,
2011) that allowed us to examine the impact of dosage and adherence on student learning
gains.
To link differences in student outcomes to differences in the teacher implementation of the
ADI model, we needed to document the fidelity of implementation of the ADI instructional
model in each classroom. We therefore video recorded each ADI lab activity to document the
number of ADI lab activities that the teachers chose to implement (dosage) and the number
of times the teachers required the students to write a report in each classroom (adherence).
Table 2 provides information regarding the number of ADI activities implemented by each
teacher and the volume of investigation reports generated in each classroom. As illustrated
in Table 2, dosage and adherence was consistent within each course but there was a great
deal of variation across the four courses.
Science Education, Vol. 97, No. 5, pp. 643–670 (2013)
 WRITING TO LEARN BY LEARNING TO WRITE IN SCIENCE 653

TABLE 2
Number of ADI Labs Implemented by the Classroom Teachers in Each
Semester and the Number of Labs That Included the Generation, Review,
and Revision of the Investigation Report (Stages 5–8 of the Model)
Number of ADI Labs

First Semester Second Semester

Number of With Without With Without

Course Teacher Sections Total Reports reports Total Reports Reports
Life science 1 5 5 4 1 7 5 2
Physical science 2 5 4 4 0 4 2 2
Biology 3 5 7 6 1 7 6 1
4 4 7 6 1 7 6 1
Chemistry 5 6 5 5 0 5 2 3
6 2 5 5 0 5 2 3

Data Sources
Argumentative Writing Assessment. The research team created a science-specific ar-
gumentative writing assessment for each course. This assessment provided a student with
a small amount of background information and a related data table followed by a prompt.
The prompt presented an argument by a scientist (or “expert”) who provides an explanation
involving the data provided but flawed in a noticeable manner. The students were asked to
refute the scientist’s claim using information and data provided in the question and then
provide and support a counterclaim using evidence and a rationale. They were also expected
to write their essay using appropriate style, grammar, punctuation, and spelling. During this
assessment, the students were provided with several pieces of lined paper to help organize
their writing. The students were asked to engage in a prewriting activity to outline their
argument and then generate a rough draft. Students refined any initial drafts or prewriting
exercises to provide a final draft of their argumentative essay addressing the task identified
for the assessment. The students completed this assessment in a single 55-minute class
period at the beginning of the year, at the end of the first semester, and then again at the
end of the second semester of instruction.
To avoid any testing effects (i.e., students explicitly remembering a specific writing
prompt) over the course of the year, three slightly different versions of the writing assess-
ment were created for each course. The variation among these versions involved slight
changes to the data values in the chart, different data categories, and minor changes to the
focus of the “expert” claim. Although these different versions represent unique testing con-
texts, the task for each version of the assessment was not changed. The rubric used to score
the assessment focused on the presence and nature of critical elements of the essay that
spanned versions of the assessment. Thus, comparisons of scores from different versions
of the test were appropriate since those scores were based on the same elements.

Science Content Assessment. To gain deeper insight into the nature of students’ learn-
ing, the research team and teachers developed an open-ended instrument to assess students’
understanding of core scientific ideas for each course. The assessment is composed of
eight multipart items, with each item focusing on a core scientific idea addressed in the
particular course, as determined by the teachers and researchers. Each item includes an
Science Education, Vol. 97, No. 5, pp. 643–670 (2013)
654 SAMPSON ET AL.

opening paragraph that provides a relevant scenario or context, followed by two questions.
One question asks the student to describe the core scientific idea (Know) and the other
asks the student to apply that idea to explain the scenario (Use). The scenarios designed
for these questions typically included a graphical representation of data in various forms
(i.e., graphs, DNA sequence alignments, food webs, spectrum) and supporting contextual
information that students could use in their written response. Students were provided space
after each question to provide their answer and a 55-minute class period to complete the
entire assessment.

Scoring the Assessments

Science-Specific Argumentative Writing Assessment. A base rubric was developed to
guide the scoring of the argumentative essays the students wrote, to facilitate comparisons
across courses. This rubric, with a maximum score of 28 points, focused on three main
aspects of the essay. The first aspect was the complexity of the overall argument (6 points
possible). The complexity of the argument referred to the basic structure of the argument
at the heart of the essay (i.e., it is a one-sided or two-sided argument; it includes multiple
reasons to support or challenge a claim). The second aspect was the content of the argument
(10 points possible). This aspect of the rubric targeted the quality and relevance of the
reasons provided, the author’s interpretation of the information provided in the prompt,
and if the language used by the student was consistent with the norms and epistemological
commitments of science. The last aspect of the rubric targeted the mechanics of the essay
(12 points possible). This aspect focused on correct punctuation, grammar, and the overall
technical quality of the writing. These three aspects of the rubric, when used together, were
designed to represent the unique nature of argumentative writing in science. Once the base
rubric was developed and refined, we modified the content aspect of the rubric to create a
specific rubric that was aligned with the writing prompt for each course.
Four teams, consisting of two researchers each, were created to score the sets of writing
assessments (pre-, mid-, and postversions) for each course. The sets of assessments were
randomly assembled and blinded to students’ identity and time of administration. Each
team scored at least 10% of the set together to calculate interrater reliability. The intraclass
correlation coefficient (ICC), a measure of reliability similar to Cohen’s κ and interpreted
using the same scale, was then determined for each team. The ICC (two-way random effects,
absolute agreement) was 0.71 for the biology team, 0.62 for the chemistry team, 0.74 for
the physical science team, and 0.76 for the life science team. These scores demonstrated
substantial agreement between the raters (Landis & Koch, 1977). The ICC for chemistry,
however, was lower than we desired, so the chemistry scoring team scored each assessment
individually and then generated a consensus score for each student (and the consensus
scores were used in all analyses). The remainder of the assessment set for biology, physical
science, and life science courses was divided among the raters and scored individually.

Science Content Assessment. In a similar fashion, the research team developed a rubric
for scoring students’ free responses on the science content assessment. The rubric for this
assessment was developed from answers provided for the questions by a collection of expert
scientists who were also involved in reviewing and validating the assessments. A separate
rubric was developed for the content assessment for each course. Using the extensive
expert response, the rubric was crafted by categorizing elements of the response into either
basic or sophisticated elements to a response. A panel of scientists and science educators
then reviewed these categorizations. The content assessment rubrics were standardized by
Science Education, Vol. 97, No. 5, pp. 643–670 (2013)
 WRITING TO LEARN BY LEARNING TO WRITE IN SCIENCE 655

developing a three-point scale based on the number of basic and sophisticated elements
found in a student’s response. This scaling allowed for more uniform comparisons to be
made among students within the same course and across courses. A students’ score was
developed from the rubric based on correct description of several content elements identified
in the expert’s answer to the question.
Teams of researchers were paired together for scoring the science content assessments
and given all assessments (pre, mid, and post) for a particular course. The sets of assessments
were randomly assembled and blinded to student identity each time. Each team scored at
least 10% of the set together to calculate interrater reliability. ICC was then determined
for each team. The ICC (two-way random effects, absolute agreement) was 0.89 for the
biology team, 0.90 for the chemistry team, 0.97 for the physical science team, and 0.98 for
the life science team. These scores demonstrated substantial agreement between the raters
(Landis & Koch, 1977). The remainder of the assessment sets for all courses were divided
among the raters and scored individually.

RESULTS AND DISCUSSION

We have divided the presentation of the results into two subsections. The first subsection
documents the observed changes in students’ content assessment scores, and the second
subsection provides the observed changes in the science-specific argumentative writing
scores. Both of the subsections will include a brief overview of how we chose to analyze
the scores and the results of each analysis. We will also include a brief interpretation of the
results as part of each subsection.

Content Scores
A series of paired-samples t-tests were conducted to determine whether the students’
scores on the content assessments changed in each course at the end of the first and second
semesters of instruction. To control for Type I error across the 16 tests, alpha was set at
.003 (.05/16). Table 3 provides the means and standard deviation for the content assessment
scores by semester as well as the results of each t-test. The results of this analysis indicate

TABLE 3
Overall Scores on the Content Knowledge Assessment by Semester for Each
Course
Score

Pre Post

Course N Semester M SD M SD t df p d
Life science 67 First 1.97 1.27 4.21 2.13 8.56 66 <.001 1.05
Second 2.51 2.36 8.07 4.17
10.99 66 <.001 1.34
Physical science 52 First 1.17 0.99 1.29 1.18 0.636 51 .53 –
Second 2.52 2.14 7.19 3.03
12.13 51 <.001 1.68
Biology 94 First 2.57 2.65 8.26 3.29
16.69 93 <.001 1.72
Second 1.82 2.00 9.70 4.19
19.08 93 <.001 1.97
Chemistry 81 First 1.61 1.35 4.37 2.21
10.51 80 <.001 1.16
Second 0.71 1.18 4.65 3.56
11.48 80 <.001 1.27
√
Note. The effect size for the paired samples’ t-test is d, where d = t/ N. d values of 0.2 =
small, 0.5 = moderate, and 0.8 = large (Green & Salkind, 2005).

Science Education, Vol. 97, No. 5, pp. 643–670 (2013)

656 SAMPSON ET AL.

TABLE 4
Overall Scores on the Content Knowledge Assessment Separated by Aspect
for Each Semester
Score

Pre Post

Course Semester Aspect M SD M SD t df p d

Life science First Know 0.76 0.74 1.90 1.14 6.95 66 <.001 0.85
Use 1.21 0.95 2.31 1.29 6.16 66 <.001 0.75
Second Know 0.94 1.10 3.93 2.10 10.87 66 <.001 1.33
Use 1.57 1.50 4.15 2.55 8.39 66 <.001 1.02
Physical science First Know 0.21 0.49 0.42 0.83 1.67 51 .10 –
Use 0.97 0.84 0.87 0.74 − 0.71 51 .48 –
Second Know 1.48 1.42 3.63 1.33 10.45 51 <.001 1.45
Use 1.04 1.12 3.56 2.34 7.85 51 <.001 1.08
Biology First Know 1.11 1.41 4.68 2.32 15.30 93 <.001 1.58
Use 1.47 1.59 3.57 1.56 11.03 93 <.001 1.14
Second Know 0.39 4.27 1.43 5.44 15.84 93 <.001 1.63
Use 1.42 1.56 5.44 2.33 16.72 93 <.001 1.73
Chemistry First Know 0.98 0.96 2.70 1.48 9.27 80 <.001 1.02
Use 0.63 0.69 1.67 1.09 8.67 80 <.001 0.96
Second Know 0.43 0.88 2.40 2.29 8.81 80 <.001 0.97
Use 0.28 0.55 2.24 1.82 10.31 80 <.001 1.14

that the students scored significantly better on the content assessments each semester
with one exception; the students enrolled in physical science did not make any significant
improvement during the first semester of instruction, t(51) = 0.636, p = .53. The effect
sizes associated with each significant comparison were also rather large (minimum = 1.05;
maximum = 1.97) although the raw scores at the end of each semester were not.
Given the significant overall gains on the content assessments in each course, we de-
cided to conduct a follow-up analysis to determine whether the students improved in their
knowledge of scientific explanations and their ability to use these explanations during each
semester of instruction. We therefore conducted a paired-samples t-test for each aspect of
the content assessment and set alpha at .0015 (.05/32) to control for Type I error across the
32 comparisons. Table 4 provides the means and standard deviations for each aspect of the
content assessment for each semester of instruction. Table 4 also includes the results of each
t-test. This analysis indicates that the students’ knowledge of scientific explanations and
their ability to use scientific explanations to explain natural phenomena improved in every
course each semester with the exception of the students enrolled in physical science. The
students in physical science did not make any significant improvement in their knowledge
of, t(51) = 1.67, p = .10, or in their ability to use scientific explanations, t(51) = –0.71,
p = .48, during the first semester of instruction. The effect sizes were once again large
(minimum = 0.75; maximum = 1.73) for each course when the students made significant
improvements.

Science-Specific Argumentative Writing Scores

A one-way within-subjects analysis of variance (ANOVA) was conducted for each course
to determine whether there were any significant differences in the students’ performance
Science Education, Vol. 97, No. 5, pp. 643–670 (2013)
 WRITING TO LEARN BY LEARNING TO WRITE IN SCIENCE 657

TABLE 5
Overall Scores on the Argumentative Writing Assessment by Semester for
Each Course
Pre Mid Post
Wilks’ Multivariate
Course N M SD M SD M SD  F df p η2

Life science 67 11.21 4.89 13.36 5.22 15.66 4.91 0.61 20.40 2, 65 <.001 .38
Physical 52 13.48 4.62 17.23 4.34 17.60 4.48 0.60 16.62 2, 50 <.001 .40
science
Biology 94 18.82 3.63 20.73 3.77 22.53 3.44 0.54 38.84 2, 92 <.001 .46
Chemistry 81 15.32 4.42 18.00 3.64 17.91 3.77 0.62 24.59 2, 80 <.001 .38

Note. The effect size for the multivariate test associated with Wilks’ lambda () is the multivariate
eta squared (η2 ). This statistic ranges from 0 to 1, and η2 values of 0.01 = small, 0.06 = moderate,
and 0.14 = large (Green & Salkind, 2005).

on the writing assessment over time. To control for Type I error across the four tests, alpha
level was set at .0125 (.05/4). The means and standard deviations in writing scores by
course at the beginning of the year (pre), end of the first semester (mid), and end of the
second semester (post) are provided in Table 5 along with the results of each multivariate
test (which do not require the assumption of sphericity). The results of all four ANOVAs
indicated a significant time effect. The effect sizes were all large and ranged in value from
0.38 to 0.46.
Given the significant effect of time in each course, we conducted a series of paired-
samples t-tests to examine changes in the writing scores by semester for each course. Alpha
was set at .004 (.05/12) for this analysis to help control for Type I error across the 12
pairwise comparisons. Table 6 provides the mean differences in scores for each comparison
and the results of each t-test by course. Table 6 also provides the effect size for each of
the differences observed, in the form of the Cohen’s d statistic. Overall, the results of this
analysis indicate that the students who were enrolled in the life science course and the

TABLE 6
Mean Differences between Pre-, Mid-, and Postintervention Scores on the
Argumentative Writing Assessment
Course N Comparison Mean Difference t df p d
Life science 67 Pre–Mid 2.15 3.2 66 <.001 0.39
Mid–Post 2.30 3.5 <.001 0.42
Pre–Post 4.45 6.43 <.001 0.78
Physical science 52 Pre–Mid 3.75 4.83 51 <.001 0.67
Mid–Post 0.37 0.53 .60 –
Pre–Post 4.11 5.48 <.001 0.76
Biology 94 Pre–Mid 1.92 4.82 93 <.001 0.49
Mid–Post 1.80 4.85 <.001 0.49
Pre–Post 3.71 8.85 <.001 0.91
Chemistry 81 Pre–Mid 2.68 5.95 80 <.001 0.66
Mid–Post − 0.09 − 0.19 .85 –
Pre–Post 2.59 6.21 <.001 0.69
√
Note. The effect size for the paired samples’ t-test is d, where d = t/ N. d values of 0.2 =
small, 0.5 = moderate, and 0.8 = large (Green & Salkind, 2005).

Science Education, Vol. 97, No. 5, pp. 643–670 (2013)

658 SAMPSON ET AL.

Figure 2. Mean writing scores for students in each course at the beginning of the year (pre), end of the first
semester (mid), and the end of the second semester of instruction.

biology course made significant improvements in their writing skills from pre- to midyear
and then again from mid- to postyear. The scores of the students who were enrolled in
the physical science course and the chemistry course, however, did not improve during
the second semester of the intervention although these students in these courses made
significant improvements during the first semester.
The change in writing scores over time for each course is illustrated in Figure 2. For
the students in the biology course and life science course, the mean score on the writing
assessment as a whole shows gradual improvement over time (pre–mid, mid–post). These
results suggest that the students in these courses made continuous improvements in their
science-specific argumentative writing skills over the duration of the course. These students
also wrote the most investigation reports during the intervention (9 in life science and 12
in biology). The students enrolled in the physical science course and the chemistry course,
in contrast, made significant improvements only during the first semester of instruction.
The lack of improvement in writing scores during the second semester also corresponds
with decreased opportunities to write in these two courses. The students enrolled in the
physical science and chemistry course, as noted earlier, were required to write at least four
investigation reports during the first semester of the intervention but were required to write
only two investigation reports during the second semester. The students who were enrolled
in the physical science course and the chemistry course, as a result, wrote fewer reports
overall (six in physical science and seven in chemistry) and most of the writing they did
Science Education, Vol. 97, No. 5, pp. 643–670 (2013)
 WRITING TO LEARN BY LEARNING TO WRITE IN SCIENCE 659

occurred during the first semester (four of the six reports in physical science and five of the
seven reports in chemistry).
Given the significant overall gains on the writing assessments in each course, we decided
to conduct a follow-up analysis to determine whether the students improved on all three
aspects of argumentative writing (argument complexity, content of the argument, and writ-
ing mechanics) or just some of the aspects and to determine whether there were differences
across the courses. We therefore conducted a paired-samples t-test for each aspect of scien-
tific writing by time in each course. To control for Type I error across the 36 comparisons,
we set alpha at .0013 (.05/36). Table 7 provides the means and standard deviations for each
aspect of the writing assessment for each semester of instruction. Table 7 also includes the
results of each t-test and the effect size of the comparisons that were significant.
This analysis indicates that the students enrolled in all four courses wrote an essay with
a more complex argument at the end of the school year when compared to the essay they
wrote at the beginning of the school year. Growth in this aspect of argumentative writing,
however, was not uniform over time across the courses. The students enrolled in the life
science course and the chemistry course made significant gains on this aspect of argu-
mentative writing during the first semester of instruction but not the second (see Table 7).
The students enrolled in the biology course made significant improvements during both
semesters, and the students enrolled in physical science course made a significant improve-
ment only when their end of the year scores were compared to their scores at the beginning
of the year (see Table 7).
This analysis also indicates that the content of the arguments (i.e., quality of the reasons
provided, interpretation of the available data, and the use of language that was consistent
with the norms of science) written by the students at the end of the school year was
significantly better than it was at the beginning of the school year in all four courses.
However, growth in this aspect of scientific writing was, once again, not uniform over time
in each course. The students enrolled in the life science course made significant gains on
this aspect of argumentative writing only during the second semester of instruction (see
Table 7). The students enrolled in the physical science course and the chemistry course, in
contrast, made significant gains only during the first semester of instruction. This result,
however, was expected, given the limited amount of writing these students did during the
second semester of the intervention and the lack of improvement observed in their overall
scores during the second semester (see Table 6 and Figure 2). The students enrolled in
the biology course made significant improvements during both semesters, which was also
expected given the amount of writing they did during each semester of the intervention.
The mechanics of the essays written by the students (i.e., grammar, punctuation, word
choice) had the least amount of growth when compared to the other two aspects of scientific
writing. The students enrolled in the life science course and the biology course made
significant gains on this aspect of argumentative writing only when the scores from the end
of the year were compared to scores at the beginning of the year (see Table 7). The students
enrolled in the physical science course made significant gains during the first semester of
instruction but not the second, and the students enrolled in the chemistry course did not
make any significant gains at all (see Table 7). The students enrolled in the biology course
and the chemistry course, however, were already scoring relatively high on this aspect of
scientific writing at the beginning of the year (Mbiology = 10.76; Mchemistry = 8.96; 12 points
possible) and, as a result, there was not much room for improvement. The students in the
life science course, in contrast, made small improvements each semester, which resulted
in a significant effect when the two semesters were combined. Finally, the students in the
physical science course did very little writing during the second semester of instruction, so
it should not be surprising that improvement was limited to the first semester.
Science Education, Vol. 97, No. 5, pp. 643–670 (2013)
660 SAMPSON ET AL.

TABLE 7
Mean Differences between Pre-, Mid-, and Postintervention Argumentative
Writing Assessments by Aspect
Aspect of
Argumentative Mean
Course Writing Comparison Difference t df p d
Life science Argument complexity Pre–Mid 0.94 4.38 66 <.001 0.53
Mid–Post 0.42 1.95 66 .06 –
Pre–Post 1.36 5.77 66 <.001 0.70
Argument content Pre–Mid 0.63 1.93 66 .06 –
Mid–Post 0.96 2.96 66 <.001 0.36
Pre–Post 1.61 4.99 66 <.001 0.61
Mechanics Pre–Mid 0.61 1.59 66 .11 –
Mid–Post 0.81 2.54 66 .01 –
Pre–Post 1.42 4.13 66 <.001 0.50
Physical Argument complexity Pre–Mid 0.71 2.50 51 .02 –
science Mid–Post 0.27 0.96 51 .33 –
Pre–Post 0.98 3.58 51 <.001 0.49
Argument content Pre–Mid 1.44 3.87 51 <.001 0.54
Mid–Post 0.17 0.49 51 .62 –
Pre–Post 1.62 4.55 51 <.001 0.63
Mechanics Pre–Mid 1.62 4.71 51 <.001 0.65
Mid–Post − 0.09 − 0.29 51 .77 –
Pre–Post 1.52 4.02 51 <.001 0.56
Biology Argument complexity Pre–Mid .42 3.04 93 <.001 0.31
Mid–Post 0.57 4.02 93 < .001 0.41
Pre–Post 0.99 6.39 93 <.001 0.67
Argument content Pre–Mid 1.42 5.68 93 <.001 0.59
Mid–Post 1.03 4.53 93 <.001 0.47
Pre–Post 2.45 9.76 93 <.001 1.00
Mechanics Pre–Mid 0.28 2.12 93 .04 –
Mid–Post 0.14 1.05 93 .30 –
Pre–Post 0.42 2.67 93 <.001 0.28
Chemistry Argument complexity Pre–Mid 1.19 6.37 80 <.001 0.70
Mid–Post 0.28 1.25 80 .21 –
Pre–Post 1.39 7.99 80 <.001 0.88
Argument content Pre–Mid 1.19 5.01 80 <.001 0.55
Mid–Post − 0.15 − 0.61 80 .54 –
Pre–Post 1.05 4.25 80 < .001 0.47
Mechanics Pre–Mid 0.31 1.37 80 .18 –
Mid–Post − 0.21 − 0.96 80 .34 –
Pre–Post 0.10 0.47 80 .64 –
Note. The p value for this analysis was set at .0013 for this analysis √ to control for Type I
error. The effect size for the paired samples’ t-test is d, where d = t/ N. d values of 0.2 =
small, 0.5 = moderate, and 0.8 = large (Green & Salkind, 2005).

LIMITATIONS AND CONCLUSIONS

All research designs involve trade-offs, and it is therefore important to discuss some of
the decisions that we made when we designed this study. One such decision was to work
with a small group of teachers. We felt that this decision was important because it enabled
Science Education, Vol. 97, No. 5, pp. 643–670 (2013)
 WRITING TO LEARN BY LEARNING TO WRITE IN SCIENCE 661

us to develop the new lab activities in a more collaborative manner. This decision also
made it possible for us to video-record them teaching during each lab activity, which in
turn, enabled us to document how each of these teachers implemented the ADI instructional
model. This choice, however, resulted in a small sample size on the classroom level. We
also decided to look for differential learning gains that could be linked to differences in
teacher implementation across courses instead of making comparisons to a “business-as-
usual” control group. We decided that this approach would be informative for two reasons.
First, we felt that it would be uninformative to compare students who wrote a great deal in
science over the course of a year to students who were not required to write in science at all
because there is already a substantial amount of literature that indicates that students do not
learn how to write in science as part of their science coursework unless writing is a specific
focus of the curriculum (Duschl, Schweingruber, & Shouse, 2007; Norris & Philips, 2003;
Yore et al., 2003). Second, we felt that conducting a quasi-experiment with an insufficient
sample size at the classroom level would do little to eliminate alternative explanations for
the outcomes of interest and, therefore, all findings that would result from such a study
would be meaningless.
Given these research design considerations, our conclusions need to be viewed in light
of three important limitations. First, to test the current iteration of the model and the
mechanism embedded in it, we needed to implement it in actual classrooms. We therefore
worked with a small group of teachers to change the nature of four courses, all of which
were offered at the same school, so it was possible to explore what students learn when
lab instruction is aligned with the ADI instructional model. Our findings, therefore, might
be context dependent or atypical. Second, we need to acknowledge that the social norms
inside each classroom were different and changed within each classroom over time. These
factors, therefore, might have also contributed to changes we observed. Finally, we can only
speculate about how much change we would have observed in these courses in response
to no intervention at all because we, as noted earlier, did not include a comparison group
in this study to help control for the influence of time or the impact of other courses at
each grade level (although the chemistry and physical science courses serve as informative
comparisons because the students in these courses did so little writing during the second
semester of the intervention). With these limitations in mind, we will now provide a tentative
answer for each research question based on the results outlined earlier.

How Do Students’ Understanding of Important Content and Their

Science-Specific Argumentative Writing Skills Change Over Time
When They Participate in a Series of School Science Lab Activities
That Were Designed to Emphasize Writing to Learn by Learning to
Write?
Students who experienced science lab instruction, structured using the ADI model,
improved in their understanding of important science content and their ability to write in
a scientific manner over the course of a school year. In all four courses described in this
study, students made significant learning gains as measured by the content assessment,
which involved students writing out descriptions of core scientific ideas, such as natural
selection or density, and also crafting explanations for different scenarios using those
ideas (see Tables 3 and 4). Similarly, students who participated in this study significantly
improved their science-specific argumentative writing skills as measured by an assessment
that required students to craft a counterargument using the same data and information
employed in supporting an erroneous claim from another scientist (see Table 5). The
magnitude of these improvements demonstrates that the change observed in these students’
Science Education, Vol. 97, No. 5, pp. 643–670 (2013)
662 SAMPSON ET AL.

Figure 3. Observed intervention effect sizes on each outcome for each course by semester (Cohen’s d: small =
0.2, moderate = 0.5, large = 0.8).

scores reflects a meaningful shift in what the students know and can do. Figure 3 provides
a graphical representation of the effect sizes related to gains made in each course on both
assessments, further distinguished by semester.
The effect sizes determined for the writing and content results demonstrate that in
some of the courses studied, significantly large improvements were seen. As effect size
serves as a metric describing change in terms of standard deviations, the large effect sizes
observed in all four courses with regard to students’ content understanding provide evidence
that students experienced substantial and meaningful gains in their understanding of core
scientific ideas and their ability to apply them. Looking across the four courses, the largest
amount of progress occurred in the courses where students also engaged in the largest
volume of report writing for ADI-related activities. The largest effect comes from the
biology courses studied, where students engaged in writing 12 reports over the course of
the year, followed by both life science, where students produced nine reports, and chemistry,
with students writing seven reports. Chemistry students demonstrated significant gains in
content understanding in both the first and second semesters, although they wrote only two
reports in the second semester. Similarly, physical science students produced significant
gains in content understanding during the second semester in which those students wrote
only two reports for ADI activities.
The improvement in content understanding observed in the chemistry and physical
science students suggests that perhaps the volume of report writing occurring in ADI-
modeled instruction does not directly influence students’ understanding of core scientific
ideas. However, although these students did not engage in creating reports in a higher
Science Education, Vol. 97, No. 5, pp. 643–670 (2013)
 WRITING TO LEARN BY LEARNING TO WRITE IN SCIENCE 663

volume, they did still fully engage with the other aspects of the ADI instructional model.
Although not as extensive in the amount of writing as crafting an investigation report, several
other stages of the model do necessitate students’ engaging in creating and evaluating
coherent written products. Specifically, Stage 3 of the ADI model requires the development
of fully explicated scientific arguments that involve making sense of the content topics being
addressed in relation to data and evidence obtained through the students’ investigations. And
in a similar function to latter peer review activities in Stage 6, the argumentation session in
Stage 4 also provides an opportunity for students to explain their group’s argument, written
out for display to others for their evaluation. Thus, chemistry and physical science students
still participated in events where they had an opportunity to make sense of and use core
scientific ideas through the creation of these written arguments.
The learning gains observed in the second semester for chemistry and physical science
can also be attributed to another non-ADI–related contextual pressure. Both of these courses
coincide with the grade level where state-level standardized tests are administered for the
evaluation of the school. The teachers in these courses were therefore responsible for making
sure their students were prepared for a comprehensive test that covered not only the science
content for their courses, but for all of the content expected for either middle school or high
school. As a result of this pressure, these teachers altered the nature and course of their
instruction to allow for test preparation. The decrease in the amount of writing occurring
in the chemistry courses was a direct result of this effort to dedicate more time to test
preparation and broad topic coverage. The physical science teacher, for example, took the
first half of the second semester to conduct an intensive, 6-week review session to prepare his
students, relegating the majority of his ADI activities that semester toward the end, which in
turn limited the amount of time that this teacher dedicated to writing investigation reports.
These additional pressures highlight that there are other contextual issues, in addition to
instructional approaches, that influence the amount of content understanding that students
achieve. However, the life science students, the youngest and least experienced writers
who created nine reports and experienced 12 ADI activities, made similar magnitudes of
gains in their content understanding as the chemistry students, the oldest and presumably
most experienced writers who wrote only seven reports and experienced 10 ADI activities.
It is also important to note that the biology students made the largest of meaningful
improvements in content understanding (a whole point larger in effect size), and these
students were given the most opportunities to experience ADI instruction (14 activities)
and wrote the largest number of investigation reports (12) for any course.
The results observed in the increase in content understanding in the four science courses
indicate a potential dosage effect in regard to the number of full ADI activities that stu-
dents experience. This dosage effect is further supported by the magnitudes and timing of
learning gains observed in students’ achievement on the argumentative writing assessment.
Perhaps the most striking comparison drawn from Figure 3 involves the amount of writing
improvement that occurred during the second semester of instruction. The life science and
biology courses, where investigation reports were continuously written during the year,
demonstrated a consistent improvement in students’ science-specific argumentative writ-
ing skills across both semesters. However, in the physical science and chemistry courses,
both of which decreased their rate of report writing by 50% or greater during the second
semester, students demonstrated improvement in argumentative writing only during the
first semester of the year, with scores and learning gains flat lining during the second
semester.
The effect sizes for the significant changes in students’ argumentative writing scores offer
the opportunity to compare the progression of improvement in writing among the courses.
During the first semester, improvements in students’ science-specific argumentative writing
Science Education, Vol. 97, No. 5, pp. 643–670 (2013)
664 SAMPSON ET AL.

skills ranged from a small-medium effect (life science: d = 0.39) to a medium-large effect
(physical science: d = 0.67). Thus, the results observed in students’ scores demonstrate
that increased engagement in writing activities embedded in the ADI model can help
students improve their science-specific argumentative writing skills, but the magnitude of
improvements is limited. Yet the dosage trend emerging from the content assessment results
continues to be supported by the writing assessment results when considering the second
semester.
The chemistry and physical science teachers limited the amount of ADI activities they
completed during the second semester. They also decreased the number of reports they
required the students to write when they implemented an ADI lab activity. The argumen-
tative writing assessment results described in Figure 3 demonstrate that during that time
students in these courses did not experience any meaningful learning gains in regard to
their science-specific argumentative writing skills. In contrast, students in the biology and
life science courses demonstrated continued improvement in their science-specific argu-
mentative writing skills, making progress over the second semester when they experienced
more than double the opportunities to craft investigation reports and engage in peer review
activities than their peers in chemistry and physical science.
Our findings, in conclusion, offer some support for the validity of our original hypothe-
sis. The observed improvements in the students’ assessment scores suggest that the various
tasks and activities that are embedded in the ADI instructional model can help improve
students’ understanding of science content and their science-specific argumentative writing
skills. In addition, the trends of improvement appear to be consistent across all four courses,
particularly in light of the “dosage effect” that emerges in scores when we consider the
differentiated volume of investigation reports that were written by the students in each
course. It is important to stress, however, that the nature of this study establishes a corre-
lational relationship only between ADI instruction and the observed gains in assessment
scores.

Which Aspects of Science-Specific Argumentative Writing Did the

Middle and High School Students Improve on the Most and Which
Aspects Appear to Be the Most Challenging for Them?
We also examined changes in the students’ scores on the argumentative writing assess-
ment by each aspect. These aspects included argument complexity, which concerned the
components students included in their arguments; argument content, which focuses on
students’ appropriate application of science knowledge in their writing in ways that align
with practices of the scientific community, and mechanics, which includes considerations
of grammar, punctuation, and tone. Figure 4 provides a graph that illustrates differences in
the effect size of the intervention by aspect for all four of the courses.
The effect sizes provided in Figure 4 demonstrate a less congruent pattern of learning
when compared to the content results. However, the lessened rate and magnitude of im-
provement can also be attributed to the novelty of the style of writing. Much of the writing
students were engaged in, with other course work, involved very descriptive and narrative
styles of writing. The nature of the writing called for by the investigation reports presented
the students with the challenge of writing in a more argumentative and persuasive style. The
students also needed to engage in the serious writing practices of science, such as trans-
forming data into evidence and arguing from evidence, as they learned how to write a report
to share the process and product of their investigations with their classmates. The novelty
of this type of realistic writing task and the complex nature of the serious writing practices
that the students needed to engage in to write in this manner could be a moderating factor
Science Education, Vol. 97, No. 5, pp. 643–670 (2013)
 WRITING TO LEARN BY LEARNING TO WRITE IN SCIENCE 665

Figure 4. Intervention effect sizes on each aspect of science writing by course for each semester (Cohen’s d:
small = 0.2, moderate = 0.5, large = 0.8).

influencing the nature of learning gains made by students in terms of their science-specific
argumentative writing skills.
During the first semester of the school year, when the volume of report writing was
highest for all four courses, the students made significant improvements with moderate
effect sizes on at least one of the three subscales, with the students in the physical science
course improving on all three subscales. The biology and chemistry students did not produce
any significant gains on the mechanics subscale, yet that was more likely due to the high
scores they attained for this subscale on the first administration of the assessment. The
life science students produced significant gains on the argument complexity subscale only
during the first semester. The overall life science results, however, offer an interesting look
at the development of science-specific argumentative writing skills in younger learners.
The students in the life science course represent the youngest in the sample, and these
students presumably had the least amount of formal writing instruction when compared to
the other students in this study. This student group demonstrated moderate but significant
improvement in the first semester in their ability to compose an argument that included
a claim and supporting reasons, also identifying those in the counterargument embedded
in the writing assessment. However, during the second semester, the life science students
produced significant gains only on the argument content subscale. The small to moderate
effect size calculated for this result suggests a potential learning progression for these
younger students in regard to their proficiency in crafting coherent arguments using valid
and relevant science content. That is, these students had to learn how to craft the argument
by recognizing and combining the structural elements (claims, counterclaims, evidence,
etc.) before they were able to appropriately employ science content knowledge within
those argument structures.
Science Education, Vol. 97, No. 5, pp. 643–670 (2013)
666 SAMPSON ET AL.

The writing scores for the biology students in the research sample provide the most
compelling evidence for the importance of continued engagement in activities that require
the use of argumentative writing skills in the context of science. These students generated
12 reports over the course of a year in their classes related to ADI activities. These students
were also the only subgroup to evidence significant learning on the two argument-related
subscales on the writing assessment during both semesters. Again, lack of improvement
on the mechanics subscale is attributed to early demonstration of facility with those funda-
mental elements. The correlational relationship existing between volume of reports written
and enhanced science-specific argumentative writing skills supports the notion of a “dosage
effect” in regard to the impact of ADI instruction on science learning. This effect garners
further support from the absence of success seen in the writing assessment scores in the
chemistry and physical science courses during the second semester, when students produced
only two reports in relation to ADI activities.

IMPLICATIONS
Middle and high school students, as noted in the introduction of this article, have few
opportunities to write reports that offer an explanation for a natural phenomenon, to read the
reports of others, or to offer a formal review of other students’ work as part of the inquiry
process. Therefore, it should not be surprising to science educators that many of these
students struggle to write in a manner that is consistent with the norms and epistemological
commitments of science. One way to address this issue is to encourage students to write
more during school science laboratories. Science teachers, however, must do more than just
encourage students to write more; they must also consider the purpose and the nature of the
writing that they expect students to do and what activities that students must engage in as
they write. Writing short answer responses to analysis questions at the end of a lab activity,
writing reflective journals, or summaries of research articles alone does not seem to result
in better scientific writing skills, although these types of writing tasks can help students
develop a better understanding of important content (Baker et al., 2008; Gunel et al., 2009;
Prain, 2006)
To improve secondary students’ science-specific argumentative writing skills, and un-
derstanding of the content at the same time, the writing students do during school science
laboratories needs to be more authentic and educative. To be more authentic, the writing
tasks need to be realistic, embedded into the inquiry process, and engage students in the
serious writing practices of science. These serious writing practices include, but are not
limited to, transforming data into evidence, arguing from evidence, coordinating theory
and evidence, and refining a text in light of a critique. All of these serious writing practices
are common in science but tend to be neglected in middle and high school science class-
rooms. For these tasks to be more educative, instructors need to show students examples of
what counts as science-specific argumentative writing (i.e., modeling), provide prompts or
reminders to learners as they write (i.e., scaffolding), and encourage students to be more
metacognitive as they engage in the iterative refinement of their text (i.e., coaching). This
modeling, scaffolding, and coaching, however, must focus on the transformation of ideas or
claims and the development of a sound argument as well as the improvement of grammar,
spelling, and punctuation to help students learn the content and how to write well in the
context of science at the same time (Yore et al., 2003).
The results of this study support this claim. Our findings suggest that the ADI instructional
model, which gives students a chance to engage in the serious writing practices of science
and receive feedback about their performance, can help learners learn content and how
to write better in science at the same time. We conjecture that this model helps students
Science Education, Vol. 97, No. 5, pp. 643–670 (2013)
 WRITING TO LEARN BY LEARNING TO WRITE IN SCIENCE 667

develop their science-specific argumentative writing skills because the activities and tasks
embedded into the model helps them to become more attuned to “what is and what is
not acceptable” in science. This, in turn, appears to enable students to develop better
argumentative writing skills over time and a more fluid “grasp of practice” (Ford, 2008)
that allows them to use their skills in different contexts or in novel situations. The ADI
instructional model also seems to promote and support the learning of content at the same
time because the realistic writing tasks that students complete, and the serious writing
practices that they must engage in as they complete these tasks, require students to think
deeply about the content (Hand, 2004; Kelly et al., 2002; Yore et al., 2003)
The stages of the ADI instructional model also seem to help move the writing that occurs
during a school science laboratory from “doing school” to “doing science” (Jimenez–
Aleixandre et al., 2000) and give students more opportunities to write for original means
(Bazerman, 1988). The realistic writing task embedded into the model, in other words,
provides an opportunity for students to write for an audience consisting of knowledgeable
peers, craft scientific arguments, provide feedback to their peers, and revise their ideas and
writing in light of feedback. These types of opportunities are important because middle and
high school students are rarely encouraged to make sense of a topic under investigation by
writing about it in a manner that reflects the nature of the writing that takes place within the
scientific community. Without engaging students in these types of serious scientific writing
practices, we argue, writing a lab report is little more than an assessment of what students
did during a lab activity rather than an opportunity for them to write to learn by learning
to write. Overall, the results of this study, although clearly not definitive, suggest that
when teachers require students to complete realistic writing tasks and engage in the serious
writing practices of science throughout a course, students can develop their science-specific
argumentative writing skills and learn important content at the same time.
Our results also suggest that science educators and science teachers will need to be mind-
ful about issues related to dosage when attempting to help students develop science-specific
argumentative writing skills when using this type of approach to laboratory instruction. At
the beginning of this article, we described how important it is to give students multiple
opportunities to complete realistic writing tasks and engage in the serious writing practices
of science because science-specific argumentative writing skills are complex and take time
to develop. Our findings clearly support this notion; students will need to write more than
twice a semester to make significant gains in their science-specific argumentative writing
skills. A writing to learn by learning to write approach, as a result, needs to be a major
focus of the curriculum and sustained throughout a course in order for it to have in impact
on student learning.
We also identified several issues that will be fruitful lines of inquiry in the future. The first
issue concerns the relationship between how a teacher implements a model of instruction,
such as ADI, and what students learn. In this study, for example, some teachers did not
require their students to write a report as part of each laboratory activity. This decision
to omit this requirement resulted in the elimination of several educative elements of the
ADI model (such as peer review and report revision), which in turn, seemed to influence
the learning gains made by the students. Therefore, more research is needed to identify
the elements of ADI that can be altered or adapted by teachers and which elements are
“nonnegotiable” and must be enacted as intended. Second, it will be useful to determine
how many ADI lab activities are needed to promote meaningful learning gains and if this
number varies by age or grade level. Finally, researchers will need to examine how various
student characteristics such as culture, home language, and past experiences influence what
learners do and learn during an ADI, to ensure that this type of instruction is equitable. All
three of these issues, we argue, are important, and we must develop a better understanding
Science Education, Vol. 97, No. 5, pp. 643–670 (2013)
668 SAMPSON ET AL.

of each one to increase the promise and potential of ADI and other instructional approaches
that are designed to promote and support writing to learn by learning to write in a discipline.

REFERENCES
Alley, M. (1996). The craft of scientific writing. New York: Springer Science+Business Media.
Anderson, C. (2007). Perspectives on science learning. In S. K. Abell & N. Lederman (Eds.), Handbook of
research in science education (pp. 3 – 30). Mahwah, NJ: Erlbaum.
Applebee, A. (1984). Contexts for learning: Studies of secondary instruction. Norwood, NJ: Ablex.
Baker, W. P., Barstack, R., Clark, D., Hull, E., Goodman, B., Kook, J., et al. (2008). Writing-to-learn in the
inquiry-science classroom: Effective strategies from middle school science and writing teachers. The Clearing
House, 81(3), 105 – 108.
Bangert-Drowns, R., Hurley, M., & Wilkson, B. (2004). The effects of school-based writing-to-learn interventions
on academic achievement: A meta-analysis. Review of Educational Research, 74, 29 – 58.
Barrass, R. (2002). Scientists must write: A guide to better writing for scientists, engineers and students. Portland,
OR: Taylor & Francis.
Bazerman, C. (1988). Shaping written knowledge: The genre and activity of the experimental article in science.
Madison: University of Wisconsin Press.
Berber-Jimenez, L., Montelongo, J., Hernandez, A., Herter, R., & Hosking, D. (2008). Helping students write
better conclusions. The Science Teacher, 75(3), 56 – 61.
Bereiter, C., & Scardamalia, M. (1987). The psychology of written composition. Hillsdale, NJ: Erlbaum.
Berkenkotter, C., & Huckin, T. (1995). Genre knowledge in disciplinary communication: Cognition/culture/power.
Mahwah, NJ: Erlbaum.
Blanchard, M., Southerland, S., Osborne, J., Sampson, V., Annetta, L., & Granger, E. (2010). Is inquiry possible in
light of accountability? A quantitative comparison of the relative effectiveness of guided inquiry and verification
laboratory instruction. Science Education, 94(4), 577 – 616.
Breger, D. (1995). The inquiry paper. The Science Scope, 19(2), 27 – 32.
Brown, A. L. (1992). Design experiments: Theoretical and methodological challenges in creating complex inter-
ventions in classroom settings. Journal of the Learning Sciences, 2, 141 – 178.
Brown, A. L., & Campione, J. C. (1996). Psychological theory and the design of innovative learning environments:
On procedures, principles, and systems. In L. Schauble & R. Glaser (Eds.), Innovations in learning: New
environments for education. Mahwah, NJ: Erlbaum.
Carter, M. (2007). Ways of knowing, doing, and writing in the disciplines. College Composition and Communi-
cation, 58(3), 385 – 418.
Carter, M., Ferzli, M., & Wiebe, E. (2007). Writing to learn by learning to write in the disciplines. Journal of
Business and Technical Communication, 21, 278 – 302.
Connolly, P., & Vilardi, T. (Eds.). (1989). Writing to learn in mathematics and science. New York: Teachers
College Press.
Driver, R., Asoko, H., Leach, J., Mortimer, E., & Scott, P. (1994). Constructing scientific knowledge in the
classroom. Educational Researcher, 23, 5 – 12.
Duschl, R. (2008). Science education in three-part harmony: Balancing conceptual, epistemic, and social learning
goals. Review of Research in Education, 32, 268 – 291.
Duschl, R., Schweingruber, H., & Shouse, A. (Eds.). (2007). Taking science to school: Learning and teaching
science in grades K-8. Washington, DC: National Academies Press.
Fallahi, C., Wood, R., Austad, C., & Fallahi, H. (2006). A program for improving undergraduate psychology
students’ basic writing skills. Teaching of Psychology, 33(3), 171 – 175.
Ford, M. (2008). Disciplinary authority and accountability in scientific practice and learning. Science Education,
92(3), 404 – 423.
Friend, R. (2002). Summing it up: Teaching summary writing to enhance science learning. The Science Teacher,
69(4), 40 – 43.
Fulwiler, B. (2008). Writing in science: How to scaffold instruction to support learning. Portsmouth, NH: Heine-
mann.
Galbraith, D., & Rijlaarsdam, G. (1999). Effective strategies for the teaching and learning of writing. Learning
and Instruction, 9, 93 – 108.
Galbraith, D., & Torrance, M. (Eds.). (1999). Knowing what to write: Conceptual processes in text production.
Amsterdam, The Netherlands: Amsterdam University Press.
Gerdeman, R., Russell, A., & Worden, K. (2007). Web-based student writing and reviewing in a large biology
lecture course. Journal of College Science Teaching, 36(5), 46 – 52.

Science Education, Vol. 97, No. 5, pp. 643–670 (2013)

 WRITING TO LEARN BY LEARNING TO WRITE IN SCIENCE 669

Gleason, M. (1999). The role of evidence in argumentative writing. Reading and Writing Quarterly, 15(1), 81 – 106.
Glynn, S., & Muth, K. (1994). Reading and writing to learn science: Achieving scientific literacy. Journal of
Research in Science Teaching, 31(9), 1057 – 1073.
Green, S. B., & Salkind, N. J. (2005). Using SPSS for Windows and Macintosh: Analyzing and understanding
data (4th ed.). Upper Saddle River, NJ: Pearson-Prentice Hall.
Gunel, M., Hand, B., & McDermott, M. (2009). Writing for different audiences: Effects on high-school students’
conceptual understanding of biology. Learning and Instruction, 19, 354 – 367.
Halliday, M., & Martin, J. (1993). Writing science: Literacy and discursive power. London: Taylor & Francis.
Hand, B. (2004). Cognitive, constructivist mechanisms for learning science through writing. In C. Wallace,
B. Hand, & V. Prain (Eds.), Writing and learning in the science classroom (pp. 21 – 32). Dordrecht, The
Netherlands: Kluwer.
Hand, B., Hohenshell, L., & Prain, V. (2007). Examining the effect of multiple writing tasks on Year 10 biology
students’ understandings of cell and molecular biology concepts. Instructional Science, 35, 343 – 373.
Hand, B., & Prain, V. (2006). Moving from border crossing to convergence of perspectives in language and science
literacy research and practice. International Journal of Science Education, 28(2 – 3), 101 – 107.
Heuser, D. (2005). Learning logs: Writing to learn, reading to assess. Science and Children, 43(3).
Holliday, W., Yore, L., & Alvermann, D. (1994). The reading-science learning-writing connection: Breakthroughs,
barriers and promises. Journal of Research in Science Teaching, 31(9), 877 – 893.
Indrisano, R., & Paratore, J. (Eds.). (2005). Learning to write and writing to learn: Theory and research in practice.
Newark, DE: International Reading Association.
Jimenez-Aleixandre, M., Rodriguez, M., & Duschl, R. A. (2000). “Doing the lesson” or “doing science”: Argument
in high school genetics. Science Education, 84(6), 757 – 792.
Kelly, G., Regev, J., & Prothero, W. (2008). Analysis of lines of reasoning in written argumentation. In S. Erduran
& M. Jimenez-Aleixandre (Eds.), Argumentation in science education: Perspectives from classroom-based
research (pp. 137 – 157). Dordrecht, The Netherlands: Springer.
Kelly, G. J., & Bazerman, C. (2003). How students argue scientific claims: A rhetorical-semantic analysis. Applied
Linguistics, 24(1), 28 – 55.
Kelly, G. J., Bazermann, C., Skukauskaite, A., & Prothero, W. (2002). Rhetorical features of student science
writing in introductory university oceanography. Paper presented at the International conference of Ontological,
Epistemological, Linguistics and Pedagogical Considerations of Language and Science Literacy: Empowering
Research and Informing Instruction.
Kelly, G. J., & Takao, A. (2002). Epistemic levels in argument: an analysis of university oceanography students’
use of evidence in writing. Science Education, 86(3), 314 – 342.
Kiuhara, S., Graham, S., & Hweken, L. (2009). Teaching writing to high school students: A national survey.
Journal of Educational Psychology, 101(1), 136 – 160.
Klein, P., & Kirkpatrick, L. (2010). A framework for content area writing: Mediators and moderators. Journal of
Writing Research, 2(1), 1 – 46.
Landis, J., & Koch, G. (1977). The measurement of observer agreement for categorical data. Biometrics, 33,
159 – 174.
Miller, R., & Calfee, R. (2004). Making thinking visible. Science and Children, 42(3), 20 – 25.
Murnane, R., & Willett, J. (2011). Methods matter: Improving causal inferences in educational and social science.
New York: Oxford University Press.
National Governors Association Center for Best Practices and Council of Chief State School Officers. (2010).
Common core standards. Washington, DC: National Governors Association Center for Best Practices and
Council of Chief State School Officers.
National Research Council. (2005). America’s lab report: Investigations in high school science. Washington, DC:
National Academy Press.
National Research Council. (2012). A framework for K-12 science education: Practices, crosscutting concepts,
and core ideas. Washington, DC: National Academies Press.
Neal, J., & Langer, M. (1992). A framework of teaching options for content area instruction: mediated instruction
of text. Journal of Reading, 36(3), 226 – 231.
Norris, S., & Philips, L. (2003). How literacy in its fundamental sense is central to scientific literacy. Science
Education, 87(2), 224 – 240.
Poock, J., Burke, K., Greenbowe, T., & Hand, B. (2007). Using the science writing heuristic in the general chemistry
laboratory to improve students academic performance. Journal of Chemical Education, 84, 1371 – 1378.
Porter, R., Guarienti, K., Brydon, B., Robb, J., Royston, A., Painter, H., et al. (2010). Writing better lab reports:
A teacher research project to improve the quality of students’ writing. The Science Teacher, 77(1), 43 – 48.
Prain, V. (2006). learning from writing in secondary science: Some theoretical and practical implications. Inter-
national Journal of Science Education, 28, 179 – 201.

Science Education, Vol. 97, No. 5, pp. 643–670 (2013)

670 SAMPSON ET AL.

Pratt, H., & Pratt, N. (2004). Integrating science and literacy instruction with the common goal of learning science
content. In W. Saul (Ed.), Crossing borders in literacy and science instruction: Perspectives on theory and
practice (pp. 395 – 405). Newark, DE: International Reading Association.
Rijlaarsdam, G., Couzijn, M., Janssen, T., Braaksma, M., & Kieft, M. (2006). Writing experimental manuals
in science education: The impact of writing genre and audience. International Journal of Science Education,
28(2 – 3), 203 – 234.
Rivard, L. (1994). A review of writing to learn in science—Implications for practice and research. Journal of
Research in Science Teaching, 31, 969 – 983.
Sakai, A., & Leggo, C. (1997). Knowing from different angles: Language arts and science connections. Voices
from the Middle, 4, 226 – 230.
Sampson, V., Grooms, J., & Walker, J. (2011). Argument-driven inquiry as a way to help students learn how to
participate in scientific argumentation and craft written arguments: An exploratory study. Science Education,
95(2), 217 – 257.
Sampson, V., & Walker, J. (2012). Argument-driven inquiry as a way to help undergraduate students write to learn
by learning to write in chemistry. International Journal of Science Education, 34(10), 1443 – 1485.
Saul, E. W. (Ed.). (2004). Crossing borders in literacy and science instruction: Perspectives on theory and practice.
Arlington, VA: NSTA Press.
Scott, P. H., Asoko, H., & Leach, J. (2007). Students conceptions and conceptual learning in science. In S. K.
Abell & N. Lederman (Eds.), Handbook of research in science education (pp. 31 – 56). Mahwah, NJ: Erlbaum.
Smith, L., & Southerland, S. (2007). Reforming practice or modifying reforms? Journal of Research in Science
Teaching, 44(3), 396 – 423.
The Design-Based Research Collective. (2003). Design-based research: An emerging paradigm for educational
inquiry. Educational Researcher, 32(1), 5 – 8.
Timmerman, B. (2002). Keeping science current. The Science Scope, 25(6), 12 – 15.
Towndrow, P., Ling, T., & Venthan, A. (2008). Promoting inquiry through science reflective journal writing.
Eurasia Journal of Mathematics, Science & Technology Education, 4(3), 279 – 283.
Turner, T., & Broemmel, A. (2006). 14 writing strategies. The Science Scope, December, 27 – 31.
Wallace, C., Hand, B., & Prain, V. (Eds.). (2004). Writing and learning in the science classroom. Boston: Kluwer.
Wallace, C., Hand, B., & Yang, E.-M. (2005). The science writing heuristic: Using writing as a tool for learning
in the laboratory. In W. Saul (Ed.), Crossing borders in literacy and science instruction. Arlington, VA: NSTA
Press.
Walvoord, M., Hoefnagels, M., Gaffin, D., Chemchal, M., & Long, D. (2008). An analysis of calibrated peer
review (CPR) in a science lecture classroom. Journal of College Science Teaching, 37(4), 66 – 73.
Weiss, I. R., Banilower, E. R., McMahon, K. C., & Smith, P. S. (2001). The 2000 national survey of science and
mathematics education. Chapel Hill, NC: Horizon Research.
Yager, R. E. (2004). Science is not written, but it can be written about. In W. Saul (Ed.), Crossing borders in
literacy and science instruction: Perspectives on Theory and Practice (pp. 95 – 107). Newark, DE: International
Reading Association.
Yore, L. (2004). Why do future scientists need to study the language arts? In W. Saul (Ed.), Crossing borders in
literacy and science instruction: Perspectives on theory and practice (pp. 71 – 94). Newark, DE: International
Reading Association.
Yore, L., Bisanz, G., & Hand, B. (2003). Examining the literacy component of science literacy: 25 years of
language arts and science research. International Journal of Science Education, 25(6), 689 – 725.

Science Education, Vol. 97, No. 5, pp. 643–670 (2013)