An Exploration of Professional Knowledge Needed for Reading and Responding to Student Writing

Abstract

This study explored the knowledge elementary teachers need for one core practice: reading and responding to students’ writing. Forty-five preservice teachers read and responded to an elementary student’s narrative writing sample. Using teacher noticing as a framework, we first decomposed the practice into five components indicative of differences in teachers’ attention to writing features, reasoning about those features, and suggested responses. We used multiple correspondence analyses to investigate potential underlying relationships among components and developed cases to highlight one underlying relationship that was found. The findings indicate reading and responding draws upon teachers’ pedagogical content knowledge. More specifically, it draws on both knowledge of content and students and knowledge of content and teaching.

Keywords

literacy teacher education preservice teacher education practice-based teacher education teacher knowledge elementary teacher education teacher noticing writing

The widespread adoption of the Common Core State Standards (CCSS) has refocused attention on writing instruction in schools. Previous policies, such as No Child Left Behind, emphasized reading and mathematics achievement and marginalized writing in many classrooms (Harper et al., 2007; McCarthey, 2008). In contrast, the CCSS make writing central to schooling, provide a blueprint for writing instruction across grade levels, and establish ambitious learning goals (Graham & Harris, 2015). The CCSS include Language Standards, which address foundational writing skills, such as spelling, grammar, conventions, and vocabulary. The CCSS also include Writing Standards, which require that children learn to write narrative, informational, and opinion texts for different purposes and audiences; use the writing process—planning, revising, editing—to develop and strengthen a text’s content, organization, and style; use digital tools to produce and publish writing; and use writing to analyze texts and build new knowledge.

Although the writing standards appear straightforward at first glance, perhaps because they are well-organized and specify clear developmental progressions (Shanahan, 2015), learning to write is, in fact, a complex process for elementary school children. Whereas adults have developed automaticity in basic aspects of transcription (the act of putting words on paper), children must exert conscious effort to form letters, spell words, and adhere to capitalization and punctuation rules. This can impede writing production (Hayes & Olinghouse, 2015). Children are also still developing conceptual knowledge about discourse, including the characteristics of good writing generally and genre-specific text features (McCutchen, 2011; Olinghouse & Graham, 2009). While the CCSS do not directly state the discourse knowledge children must learn, students cannot produce narrative, informational, and opinion texts without understanding appropriate genre features (Hayes & Olinghouse, 2015).

Another challenge elementary children face as they learn to write is engaging in the writing process itself. Although children can learn to both plan and revise their writing through explicit strategy instruction (Graham, Harris, & Santangelo, 2015), these processes do not come naturally. When prompted to plan, children often simply begin composing (Bereiter & Scardamalia, 1987), and planning does not necessarily improve the quality of their writing (Olinghouse & Graham, 2009). Children also do not spontaneously revise, and when they do, they tend to make surface-level changes more akin to editing than substantive changes that significantly improve text quality (Bereiter & Scardamalia, 1987). In short, elementary children have a lot to learn to attain the ambitious writing goals established by the CCSS.

Attaining the CCSS’s ambitious writing goals will depend upon skillful teaching, yet teachers report minimal preparation for writing instruction during teacher preparation programs (Cutler & Graham, 2008; Gilbert & Graham, 2010). While this is problematic, it is perhaps unsurprising given the small body of research on what writing teachers must know. Disciplines such as mathematics (e.g., Ball, Thames, & Phelps, 2008) and science (Abell, 2008) have ongoing, substantial lines of research into the knowledge needed for teaching. In contrast, teacher knowledge has not been a major area of inquiry among writing researchers (Reutzel et al., 2011). The paucity of empirical findings offers little guidance for teacher education programs with respect to preparing teachers of writing.

A growing number of teacher educators advocate taking a practice-based approach to understanding the knowledge and skills needed for teaching (Ball & Forzani, 2009; Forzani, 2014; Grossman, Hammerness, & McDonald, 2009; McDonald, Kazemi, & Kavanagh, 2013). A practice-based approach involves first identifying the work teachers do—the core teaching practices that support student learning—then decomposing those practices to specify the “special knowledge, skills, and orientations” (Ball & Forzani, 2009, p. 500) needed for enactment. Decomposition involves “breaking down a complex practice into its constituent parts” (Ball & Forzani, 2009, p. 504) so that novices can more easily learn it. Because decomposition provides a template for how preservice teachers will learn to enact a practice, a practice must be decomposed in a way that is manageable for novices while preserving the aspects that make it effective. Once a core practice has been decomposed, novices can engage in “approximations” (Grossman et al., 2009, p. 283) of it through rehearsals and feedback from teacher educators.

One promising practice that is both possible for novice teachers to begin to master and likely to help children progress toward the CCSS’s writing goals is “reading and responding to student writing.” Reading and responding is a subject-specific enactment of two core practices identified by other scholars: analyzing student work and providing feedback to students (Teaching Works, 2016). It is central to writing instruction because it occurs regularly in classrooms and can positively impact the quality of students’ texts (Graham, Hebert, & Harris, 2015). However, little research has examined reading and responding as a practice or how novice teachers become skilled in it.

In our work with elementary preservice teachers, we have been investigating what reading and responding to student writing entails and how to decompose it so that preservice teachers can learn to skillfully enact it. In this article, we first describe how we have chosen to decompose reading and responding as a practice and how this decomposition has allowed us to explore preservice teachers’ current approximations of practice (Grossman et al., 2009). Then, we present the current study and how it informs our thinking about the knowledge needed for reading and responding effectively.

Reading and Responding Decomposed

Recent scholarship around teacher noticing (Sherin, Jacobs, & Philipp, 2011) offers a useful starting point for decomposing reading and responding as a practice because of its strong conceptual alignment. Much of the noticing literature analyzes how teachers attend to and make sense of students and their thinking in the midst of classroom instruction or discussions. Jacobs, Lamb, and Philipp (2010) have extended this line of research to analyze how teachers read and respond to students’ written mathematical tasks. The teacher noticing literature highlights three interrelated components: attention, reasoning, and determining a response.

The first two components—attention and reasoning—decompose what teachers do while reading student writing. A piece of writing contains many different features. Writing teachers must be able to focus their attention on the most salient features and reason about the instructional significance of those features. Current research notes the variability in writing teachers’ attention and reasoning while reading students’ writing. Some teachers attend almost exclusively to what many teacher educators call surface features (e.g., grammar, usage, and length), those features found in the Language Standards. Other teachers also attend to deep features (e.g., ideas, structure, and style; Ell, Hill, & Grudnoff, 2012; Gibson, 2007; Masuda & Ebersole, 2012), features embedded in the Writing Standards. When teachers reason about these features, they may adopt an evaluative lens, focusing on weaknesses and errors (Masuda & Ebersole, 2012; Matsumara, Patthey-Chavez, Valdes, & Garnier, 2002). Or they might read more generously (Spence, 2010), seeking to interpret students’ purposeful thinking and meaning-making as authors. For example, a teacher may read “this story is about. . . ” at the beginning of a child’s draft, attend to the lowercase “t,” and reason that the child does not know to capitalize the first word of a sentence. In contrast, another teacher may read the same text, attend to how the sentence announces the story topic, and reason that the child is attempting to inform the audience about the subject of the text.

The third component of teacher noticing is determining an appropriate response. The noticing literature focuses on the decision-making aspect of responding—a teacher’s intended response—rather than how a teacher carries out a response (Jacobs et al., 2010). We also take this approach because responses are likely to draw on strategies (e.g., conducting mini-lessons, modeling with mentor texts), which in turn might be decomposed and studied as practices. The literature on writing teachers highlights at least three components of response: response topics, response strategies, and response agents. Response topics—the writing features teachers choose to address in their response—vary. Teacher feedback may emphasize surface edits (Matsumara et al., 2002) or may focus on content, process, and style (DeGroff, 1992). Response strategies—the form of the response—also vary. Teachers may suggest very generic response strategies like editing or additional practice (Gibson, 2007). In contrast, they might employ more sophisticated tactics such as teacher modeling, trade book modeling, guided practice, or targeted questioning to help the child consider how to clearly communicate with the audience (Gibson, 2007; Masuda & Ebersole, 2012). Response agents—who are actively participating in the response—vary as well. Teachers might offer teacher-centered responses that promote students’ dependence on the teacher or child-centered responses that promote students’ use of self-regulated writing strategies (Glasswell, Parr, & McNaughton, 2003).

The variability of teachers’ reading and responding reported in the literature is troubling. While there is unlikely to be a single “right” way to read and respond to a given piece of writing, some approaches are more likely than others to support the goals of the CCSS. For example, teachers who attend and respond to a broad range of writing features will help children develop not only language and transcription skills but also conceptual knowledge of writing. Teachers whose reasoning takes children’s goals, thinking, and process into consideration have a broader evidence base for diagnosing children’s learning needs. Teachers who engage children as active participants in learning and equip them with effective writing strategies are likely providing more targeted and meaningful learning opportunities.

Our own research (Ballock & McQuitty, 2014) suggests that preservice teachers nearing the end of their elementary education program exhibit similar variability in attention, reasoning, and response to that found in the literature reviewed above. It is tempting to jump from the identification of this problem to immediate implementation of a solution: to design a pedagogical response within teacher education courses. However, we desired to pursue deeper empirical support for our decision making. Research to date allows us to describe the variability in teachers’ attending, reasoning, and responding as discrete components. What is still unclear is the relationship among these components. Does each component follow an idiosyncratic learning trajectory, or do two or more components develop in concert? Is each component distinctive, requiring a unique set of knowledge and skills that preservice teachers need to learn in isolation from the other components, or is there a common underlying knowledge base upon which all components rely? The research presented in this article begins to address these questions. In this study, we explore what, if any, relationship exists among the various components of reading and responding to student writing and what such a relationship might tell us about the professional knowledge or skill the practice requires and how it might best be decomposed so preservice teachers can learn it.

Research Design

This study included 45 preservice teachers in an elementary education program at one Mid-Atlantic state university. Eighteen participants were just beginning their degree program and were enrolled in either Principles of Learning, Development, and Diversity, an introductory education course, or Writing for Elementary Educators, a course that required them to write a personal narrative, an informational picture book, and an argumentative essay. In contrast, 27 participants were nearing degree completion. They had completed a series of literacy courses focused on children’s literacy development, literacy methods, and literacy assessment and were participating in their year-long Professional Development School internship in several different school systems with exposure to a variety of different literacy curricula. Transcripts of face-to-face interviews served as data for this study. During the interview, each participant read and responded to Reptile Story, a typed narrative authored and illustrated by a second-grade student. We chose this writing sample because it was interesting to read and included numerous features participants could notice and respond to with respect to the expectations of the second grade and even a few of the third-grade standards—for example, characters, plot, dialogue, action, details, and conventions. We asked participants two questions about the sample: (a) What do you notice about the student’s writing? (b) What would be your next steps with this student?

Data Analysis

The statistical analysis conducted for this study drew on a qualitative analysis conducted for a previous study (Ballock & McQuitty, 2014). Therefore, in this section, we first describe our previous qualitative analysis and then describe the analysis conducted for this study. In our initial analysis, we used qualitative content analysis (QCA; Schreier, 2012), a process used to describe the content of qualitative data by systematically applying coding frames that are both concept- and data-driven. Drawing on the teacher noticing literature reviewed above (e.g., Sherin et al., 2011), we developed a coding frame with five main categories (Table 1). We identified subcategories for each main category through consultation of relevant literature followed by selective coding (Schreier) of the data.

Table 1.

Qualitative Content Analysis Coding Frame.

Categories	Subcategories	Sample transcript segment
Topic/ Response Topic	Surface Features	It says “Me and my snake,” it should be “my snake and I”
Topic/ Response Topic	Deep Features	He ended the story with somebody living happily ever after.
Stance	Description	He spoke in first person the whole time.
	Interpretation	He may have recognized something from science and made a connection.
	Positive Evaluation	The content is really good.
	Negative Evaluation	His writing is very choppy and can come off a little bit awkward.
Response Strategy	No Strategy	He needs to work on more of the escalation to the problem.
	Basic Strategy	Focus more on the conventions—mini-lessons on those.
	Specialized Strategy	Do a web . . . to help focus on the main idea.
Response Agent	Teacher	I would show him paragraph formatting.
	Student	Have the kids first go through themselves and edit.
	Both	Make an outline with him . . . Say, “Where do you want to start out with it? What do you want to get across to your reader? And where do you want to end with it?”

The first two main categories—Topic and Stance—focused on participants’ reading of the narrative. Topic addressed participants’ attention: the writing features they noticed while reading. Following precedent set by previous research in literacy teacher education (e.g., Ell et al., 2012; Glasswell et al., 2003), we coded two Topic subcategories: (a) Surface Features, such as grammar, spelling, and punctuation and (b) Deep Features, such as ideas, sequence, details, word choice, and narrative features. Stance referred to participants’ reasoning about the writing features they noticed. Four subcategories, adapted from previous noticing research (van Es & Sherin, 2006), included (a) Description, (b) Interpretation, (c) Positive Evaluation, and (d) Negative Evaluation.

The remaining three coding categories—Response Topic, Response Strategy, and Response Agent—focused on the participants’ response to the narrative. Response Topic referred to the focus of the next steps participants suggested. As with Topic, we applied two subcategories: Surface Features and Deep Features. Response Strategy addressed the teaching strategies participants suggested for their next steps. To identify subcategories, we began with Cutler and Graham’s (2008) list of instructional strategies for primary-grade writing. Through selective coding of the data, we narrowed this list down to two subcategories: (a) Basic Strategies, which any layperson might suggest, such as editing, teaching foundational writing skills, and mini-lessons with no clearly specified learning activities and (b) Specialized Strategies, which would require more specialized knowledge, such as graphic organizers, reading to support writing, writing prompts, goal-oriented writing conferences, and clearly specified learning activities. Finally, Response Agent referred to who would carry out the suggested next steps. For this category, we identified three subcategories, adapted from noticing research (van Es & Sherin, 2006): Teacher, Student, or Both.

Focusing on one category at a time, the first two authors independently segmented each transcript into coding units using thematic criterion (Schreier, 2012). Then, we coded each transcript, applying only one code to each segment. Overall interrater reliability was 90%. We resolved coding conflicts through consensus.

Coded transcripts served as the basis for describing participants’ skill in each component of reading and responding, the goal of the previous study. We first focused on Topic. Because every participant noticed a variety of Surface Features, we focused on describing participants’ attention to Deep Features. We identified whether each participant attended to a Narrow Range (0-2 Deep Features), a Medium Range (3-4 Deep Features), or a Broad Range (5 or more Deep Features). We then determined each participant’s primary stance toward the narrative: (a) a Negative Stance (50% or more of word count coded as Negative Evaluation), (b) a Positive Stance (50% or more of word count coded as Positive Evaluation), or (c) a Broad Stance (all subcategories accounted for less than 50% of the word count). Next, we characterized each participant’s response. We identified whether each participant suggested No Topic, only Surface Features, a Medium Range (Surface Features and one Deep Feature), or a Broad Range (Surface Features and two Deep Features). We determined whether each participant identified No Strategy, only Basic Strategies, or at least one Specialized Strategy. We determined whether participants were Teacher-Focused (teacher as sole agent of all next steps) or Student-Focused (student as collaborator or primary agent in at least one next step).

This analysis approach was effective for our initial research, allowing us to describe the characteristics of each individual participants’ reading and responding and to compare participants within each coding category. However, it was insufficient for the purposes of this study: exploring the interrelationship of the five coding categories to look for underlying structures that might shed light on reading and responding as a practice. Several factors guided our choice of method for this analysis. First, because we entered the study with the goal of discovery instead of confirmation, we required a procedure that could reveal multivariate structure rather than confirm a priori hypotheses, as one might with structural equation modeling. Second, because we were interested in exploring both the clustering of participants and categories, we needed a procedure that could be both person-centered and variable-centered in approach. Neither Factor Analysis nor Cluster Analysis would allow us to uncover person and variable clusters simultaneously. Third, because our data are categorical instead of continuous, we needed a method that was flexible enough to accommodate the granularity of our data. Factor Analysis and Cluster Analysis assume data are at least continuous in some sense and generally require larger samples sizes than what we have. In contrast, multiple correspondence analysis (MCA) is useful for discovering clusters of persons/variables and accommodates categorical data. Therefore, we elected to use MCA for this study. We used each coding category from the qualitative analysis (e.g., Stance), and the values assigned to transcripts for each category (e.g., Negative, Positive, Broad) as categorical variables for the MCA.

MCA is a variant of correspondence analysis (CA), an analysis technique useful for discovering whether particular categories of the variables tend to co-occur in a way that reveals common meaning. It is sometimes referred to as a principal components analysis for nominally scaled variables (Clausen, 1998). MCA is an extension of CA from the case of two variables in a contingency table to the case in which there are multiple categorical variables in higher order contingency tables.

Any underlying structure between two variables discovered through CA is based on the association between the various categories of one variable and the categories of a second variable. The observed association may reflect a shared feature, or features. Extending CA to MCA, categories of multiple variables may collectively be associated, because of shared features across multiple variables. Whereas shared variable associations are called factors in Factor Analysis, in CA/MCA, these shared features are called dimensions.

MCA proves advantageous to this study because it can represent each categorical variable and participant simultaneously in a multivariate space which spans the number of dimensions that are extracted. However, careful interpretation is necessary to assess the meaningfulness of each identified dimension, because it is possible that one or more dimensions identified through MCA may simply be the result of chance associations among the data. Clausen (1998) suggests that quantitative indices, such as Contribution and Squared Correlation, are the first tools to aid interpretation. Categories with large Contribution values strongly influence the placement of dimension axes, while categories with large values of Squared Correlation are those categories particularly well explained by a dimension. Visual displays, such as scatterplots, also guide interpretation. CA/MCA results are often displayed visually as a scatterplot with dimensions as axes and variable categories and/or participants as points on the scatterplot. This results in a spatial representation of (a) the interrelationships among variables and categories and (b) how participants relate to those interrelationships. For example, variable categories which tend to co-occur in the same interview transcript would, through MCA, be located in close proximity to each other on the scatterplot, whereas categories unlikely to co-occur would be located further away from each other. Similarly, associations of individual participants with categories or dimensions are observable by virtue of their location in the scatterplot. The alignment of variable categories with dimensions reveals structure because it is assumed that correlations among variables and categories reflect common threads of meaning. Supplementary variables, though not direct contributors to the MCA dimension coordinate solutions, can also be plotted and used as confirmatory evidence of dimension meanings. We used FactoMineR (Husson, Josse, Le, & Mazet, 2015), a software package for R (R Core Team, 2016), to conduct the MCA.

As we sought to explore underlying structures or associations among QCA coding categories (Topic, Stance, Response Topic, Response Strategy, and Response Agent), we treated each coding category as an active categorical variable in the MCA. We carefully assessed the meaningfulness of each identified dimension through triangulation of MCA outputs, QCA results, and careful re-reading of select transcripts. Finally, we added two supplementary variables to the MCA to aid interpretation: (a) Transcript Length (four categories; quartiles of total transcript word count) and (b) Program Status (Early, meaning enrolled in initial coursework, or Late, meaning enrolled in the final internship year).

Findings

The purpose of this study was to explore what preservice teachers need to learn to effectively read and respond to student writing. To do so, we explored possible interrelationships among categories of attending, reasoning, and responding. MCA allowed us to investigate the likelihood that certain characteristics of reading and responding would coexist in the same participant. For example, would participants who notice a Broad Range of Topics be likely to also suggest a Broad Range of Response Topics? Or would participants who take a Positive Evaluative Stance toward the narrative be likely to recommend similar Response Strategies? Are there particular ways of attending, reasoning, and suggesting responses that cohere?

MCA Dimensions

The five QCA categories (i.e., Topic, Stance, Response Topic, Response Strategy, Response Agent) were submitted as active variables to MCA. Each variable consisted of two to five categories, collectively creating a 45 (person) × 16 (categories of variables) table for analysis. Extraction of two dimensions accounted for 37% of the variance across participants, 22% for Dimension 1 and 15% for Dimension 2. While not large, the proportion of variance explained suggests that intercorrelations among categories are potentially interpretable. MCA output identified a third dimension; however, it did not explain sufficient variance to warrant interpretation. Table 2 displays quantitative indices of the relationship between variable categories and dimensions, whereas Figure 1 visually represents the MCA output as a set of scatterplots. In each scatterplot, categories are plotted to show how strongly associated they are with each dimension. Confidence ellipses reflect differences between centroids of categories with respect to dimension association. Nonoverlapping ellipses imply categories which are spatially significantly different from each other on the two dimensions. Finally, points indicate where individual participants fall with respect to the dimensions.

Table 2.

Multiple Correspondence Analysis Coefficients for Two Extracted Dimensions.

	Dimension 1				Dimension 2
	Coordinate	Contribution	Squared correlation	V test	Coordinate	Contribution	Squared correlation	V test
Topic
Narrow range	−0.54	0.04	0.13	−2.42	−0.78	0.11	0.27	−3.47
Medium range	−0.41	0.02	0.09	−1.94	0.83	0.14	0.34	3.88
Broad range	0.86	0.11	0.41	4.25	−0.10	0.00	0.01	−0.47
Stance
Negative	0.16	0.01	0.06	1.62	−0.15	0.01	0.05	−1.43
Positive	−1.08	0.07	0.22	−3.07	−0.49	0.02	0.04	−1.39
Interpretive	−0.41	0.00	0.01	−0.73	0.80	0.03	0.05	1.42
Broad	0.93	0.03	0.08	1.92	1.38	0.10	0.18	2.85
Response Topic
No topic	−0.75	0.01	0.03	−1.08	0.72	0.01	0.02	1.03
Surface features	−0.63	0.08	0.38	−4.07	−0.27	0.02	0.07	−1.72
Medium range	0.12	0.00	0.00	0.44	0.99	0.15	0.32	3.75
Broad range	1.40	0.18	0.56	4.98	−0.65	0.06	0.12	−2.32
Response Strategy
No strategy	−0.99	0.09	0.28	−3.52	−0.87	0.10	0.22	−3.08
Basic strategies	−0.25	0.01	0.07	−1.69	0.72	0.16	0.54	4.85
Specialized strategies	1.31	0.19	0.62	5.22	−0.65	0.07	0.15	−2.59
Agent
Teacher focused	−0.53	0.07	0.39	−4.12	−0.14	0.01	0.03	−1.05
One or more student focused	0.73	0.09	0.39	4.12	0.19	0.01	0.03	1.05

Figure 1.

Multiple Correspondence Analysis Scatterplots.

Because our interpretation of Dimension 2 is tentative and needs further investigation, we focus this report of findings on Dimension 1. We describe how we used a combination of MCA outputs and three cases selected based on their MCA dimension coordinates as the basis for interpreting Dimension 1. Then, we articulate our conclusions as to the meaning of Dimension 1.

Dimension 1

The spatial organization of categories—indicated by the sign and magnitude of the Coordinate column (Table 2) and visual inspection of categories plotted across the horizontal axis of each scatterplot (Figure 1)—provides initial evidence for interpreting Dimension 1. Five categories are clearly associated with the positive pole—Broad Topic, Broad Stance, Broad Response Topic, Specialized Strategies, and Student-Focused Response Agent—while another six categories are clearly associated with the negative pole—Narrow and Medium Topic, Positive Evaluative Stance, No Response Topic, Surface Features Only, No Response Strategy, and Teacher-Focused Response Agent. The positioning of the most sophisticated categories and least sophisticated categories of each variable at opposite poles leads us to infer that Dimension 1 represents a continuum of sophistication in reading and responding.

Supplemental variables add support to this interpretation. This is important as supplementary variables were not used to compute dimension coordinates but were plotted after dimension solutions had already been established. First, participants plotted near the positive pole were more likely to be further along in their degree program (Program Status, Late Program Status; Figure 1) and, therefore, had completed more literacy coursework and field placements. We would expect these participants to have more professional knowledge and experience to draw on when reading and responding and therefore to exhibit greater sophistication. Participants plotted near the positive pole also had more to say about the narrative sample, as indicated by longer transcript length (Transcript Length, Figure 1). While verbosity should not be conflated with sophistication, review of the data shows that participants speaking fewer than 300 words attended and responded to a narrow or medium range of features and only one suggested a specialized student-focused strategy, while participants speaking over 1,000 words all attended to a broad range of features and suggested specialized student-focused strategies.

Contribution and Squared Correlation coefficients (Table 2) also aid interpretation. Although Contribution and Squared Correlation coefficients tend to vary together, they reflect different points of view. Since Contribution values sum to 1.00 down columns, they reflect the relative importance of categories to a dimension. Two categories are the strongest determinants of Dimension 1: Specialized Response Strategies (0.19) and Broad Response Topic (0.18). In contrast, Squared Correlations indicate the extent to which dimensions explain categories. Three categories are particularly well explained by Dimension 1: Specialized Response Strategies (0.62), Broad Response Topic (0.56), and Broad Topic (0.41). V test values greater than |2.00| indicate coordinates significantly different from zero (Le, Josse, & Husson, 2008). These data suggest that three variables are particularly key to interpreting Dimension 1: Topic, Response Topic, and Response Strategies.

As a final aid to Dimension 1 interpretation, we developed three case descriptions. To select the three cases, we examined the location of individual participants along the Dimension 1 axis. Paying particular attention to the three variables noted above (i.e., Topic, Response Topic, and Response Strategies), we noted substantive qualitative differences among participants at two points along the Dimension 1 axis. This allowed us to group participants into three profiles (Table 3). Despite the considerable diversity within each profile, it is clear that Profile 2 participants focus on more deep features than Profile 1 participants, and Profile 3 participants are characterized by an even broader focus on deep features. Profile 3 participants are unique in suggesting specialized response strategies. Here, we present case descriptions to illustrate each profile. Then, we present and discuss our conclusions as to what Dimension 1 represents.

Table 3.

Reading and Responding Profiles.

	Profile 1 (n = 18)	Profile 2 (n = 14)	Profile 3 (n = 13)
Topic	Narrow or Medium Range	Medium or Broad Range	Broad Range
Stance	Positive or Negative Evaluation	Negative Evaluation	Negative Evaluation
Response Topic	None or Surface Features	Surface Features or Medium Range	Broad Range
Response Strategy	None or Basic	Basic	Specialized
Response Agent	Teacher-Focused	Teacher-Focused or Student-Focused	Student-Focused

Profile 1 case

Typical of participants in Profile 1, Megan attended to a narrow range of writing features:

Some of the sentences needed periods, and he didn’t add plurals to the ends of some nouns. He needs to work on the present tense and the past tense. And endings. I saw something that needed an “ed” at the end. He would say “me and my snake” pretty often. But . . . he understands how to use “I” sometimes . . . He knows how to use dialogue, I think. He knows to put the quotation marks after a punctuation mark. And then he knows to capitalize the letter. He knows how to use his commas when he’s going into a dialogue.

Megan’s comments exclusively addressed surface features such as punctuation, plurality, tense, capitalization, and sentence structure. She seemed to interpret the research task as a directive to assess the writing’s correctness, evaluating strengths and errors almost equally. Although Megan had recently studied features of effective narratives to support her own narrative writing in the Writing for Elementary Teachers course, she did not note any of these deep features in the student’s narrative.

Given Megan’s focus on surface features and evaluative stance, it is not surprising that her next steps emphasized instruction to correct mistakes:

When he said, “Can I borrow this snake for a week,” and then he says “no”—when he says “no” he needs to say who is saying that. I would help him with making complete sentences, not run-on sentences. I would help him with run-on sentences, because I think he has some of those. And then I would help him with using the past and present tense.

Megan identified only three response topics—identifying the speaker in dialogue, sentence structure, and verb tense—and each of these involved conforming the writing to conventional rules. She indicated she would “help” the student but provided no instructional strategies and gave no indication of the role the student might play in making these corrections. Thus, her instructional ideas were confined to surface features and unclear strategies.

Profile 2 case

While Lisa primarily addressed surface features, such as tense, punctuation, and articles, she also highlighted a few deep features, including the sequence of ideas in the story:

“My friend eat all the donuts.” “Everyone go home.” And “I open all my presents, my gifts.” He definitely needs the tenses. And being able to know when to use commas . . . Then he used “a” a couple of times. I gave my snake a snake food. It sounds weird. All the grammar rules are wrong . . . I also thought it was a random story. It’s a pretty decent sequence because it talks about making a party and then what the snake can do. And then he’s bringing the snake for show and tell. Then he went home. I think he has a good gist of how to write a story . . . I just think that the grammar and the tenses and all of that is the main issue.

Lisa also commented on the illustration drawn underneath the text of the story:

I guess that’s the snake. Maybe that’s snake food, but I have no idea the background to it. I would have asked the student what that was. Maybe when he turned it in, or when I gave it back to him, “Could you describe what’s going on in the picture?”

Lisa’s stance toward the writing was primarily negatively evaluative, as demonstrated in her description of grammar and usage errors. However, she took a more interpretive stance toward the picture. She did not evaluate it even though she did not fully understand it. Instead, she wanted to know more about the child’s intended meaning and suggested asking him what he intended the picture to communicate. This shows a beginning acknowledgment of the child author as a communicator with an intentional thought process.

Lisa’s remaining instructional suggestions addressed surface features. Noteworthy, however, is Lisa’s identification of grammar as “the main issue.” This language suggests a weighing and prioritizing of features that might lead to improvement. In the previous case, Megan prioritized certain surface features over others in her suggested response, but gave no indication she noticed any deep features. In contrast, Lisa noticed other features, but indicated she felt the conventions were the most important moving forward. Lisa indicated she would “go over, have a mini-lesson . . . where he could try to make it better.” Then she added, “I think pulling him aside and going over the tenses and when to use plurals or just figuring that all out.” Like Megan, Lisa focused on correcting the writing to “make it better.” She offered basic unelaborated teaching strategies, simply indicating she would use a “mini-lesson” and “go over” tenses and plurals without indicating any particular role for the student.

Profile 3 case

Attention to a wide range of deep features and identification of specialized instructional strategies are unique characteristics of participants in Profile 3. Keri’s interview illustrates this sophistication. In addition to noticing surface features (capitalization, punctuation, paragraph indentations, articles, and plurality), Keri attended to quite a range of deep features. For example, she began by noticing a stylistic feature, stating, “I notice that his writing is very choppy and can come off a little bit awkward.” She continued by highlighting the narrative’s structure:

He has five paragraphs, and it’s clear that he has an introduction, and a conclusion, and a body, so maybe there was an organizer used . . . “Me and my snake live happily ever after.” So he has the endings. I think it’s a good conclusion, actually.

Most notable was Keri’s extensive consideration of the author’s ideas:

He seems very persistent on someone stealing his snake, which is cute. I’m not sure what the exact prompt was, but if they were doing a story based on what they were learning in science, I think that this is probably more fantasy than realistic because snakes do not jump. I think it’s good that he used the name “Python.” It means he may have recognized something from science and made a connection . . . I wonder what the snake food is. Just out of curiosity, because of science . . . He talks about making a doughnut, a cupcake, a cherry pie, and a cake for his birthday, but then he doesn’t ever mention the pie again. And it’s interesting that he made donuts, because that’s hard . . . When I’m reading it, I would like to know more. He says, “I get lots of presents.” Like, what? So, I think that he’s lacking some details that could maybe engage the reader more.

Keri’s response to the ideas in the writing sample suggests she read the narrative through two lenses. First, as a reader, she enjoyed the story and sought to make meaning as she read. She seemed to enjoy the student’s “persistence” in indicating the snake was stolen. She was interested in the inclusion of making donuts, an activity she considered “hard.” Furthermore, she reflected on her experience as a reader and her desire for more details. She treated the narrative as an authentic text, intended to engage an audience. Second, she read through the eyes of a teacher assessing the student’s written piece. She noted how the listed party foods were not integral to the story, questioned whether the ideas fit the assigned genre, and considered interdisciplinary connections reflected in the narrative.

In addition to noticing a wide range of features, Keri was one of just four participants who took a broad stance toward the writing. She did make evaluative comments: “His writing is very choppy,” “He should have a comma and a question mark after right,” and “It’s a good conclusion.” However, she also employed an interpretive stance, attempting to understand the logic in what the student did. For example, she wondered about the snake’s name: “‘I give him a name called Python.’ I wonder if he’s confusing Python as a name or if it is the type of snake . . . ” She went on to hypothesize that the use of “Python” may have come from something the child learned in science. She inferred “maybe there was an organizer used,” acknowledging a process that may have led to the narrative’s clear organization. Keri seemed to view the child as an author with agency. She moved beyond evaluation of text features to consider the student’s intentions and process.

In addition, Keri’s suggested instructional strategies stand out for their detailed elaboration and clear student roles. For example,

I would do a mini-lesson on colorful language. I had to do this with my students, when I taught a personal narrative in second grade. I’d hand out paint strips for different words that they have in their narrative. So, when he uses the “big upper boy,” instead of using the word “big,” he would write “big” at the bottom in the lightest color of the paint strip, and you would look up and write down different descriptive words that have the same meaning, so he has a variety to pick from that make his writing more colorful. But he could also use it with different times. He could brainstorm with other things.

Keri highlighted a strategy very specific to enhancing the vocabulary used in the narrative, and she communicated details about how she would orchestrate the mini-lesson. This strategy implies the teacher’s role is to equip rather than correct. Keri would create the opportunity for the child to develop a tool, but the decisions about words used in the narrative would belong to the student author. This strategy would likely improve this narrative, but Keri also suggested the student author could use the words he generated in this mini-lesson to brainstorm at other times. Thus, she seems cognizant of not only helping the child improve this story but also equipping the child with strategies to use over time.

Even when Keri suggested the need for editing the narrative, her approach was very different from Megan’s and Lisa’s:

What else could we work on? Forming our sentences . . . maybe even having the child read the story after he wrote it down, out loud a few times, and seeing if by taking away or adding certain articles, like “a” or “the” or “and” it makes it flow better, or it makes more sense . . . doing it in pairs, so that they can both give each other feedback.

This strategy would actively and meaningfully involve the student, either experimenting independently with adding and removing different articles or getting feedback from a partner. Again, this frames the student as the active agent in writing.

Dimension 1 interpretation

The differences across these three cases led us to conclude that Dimension 1 reveals something about participants’ underlying knowledge for reading and responding to student writing. First, these cases seem to highlight differences in participants’ “knowledge of content and students” (KCS; Ball et al., 2008). KCS is a type of knowledge integrating knowledge of students with knowledge of content. It includes knowledge about children’s typical patterns of understanding, including common conceptions, misconceptions, and error patterns at various ages or points of development. Megan demonstrated minimal KCS. Despite successfully writing her own narrative incorporating dialogue, action, and chronological sequence just weeks before her interview, Megan attended only to surface features. She did not seem to recognize these deep features in the child’s writing. Perhaps she did not expect children could demonstrate genre knowledge while still developing their language and transcription skills. Or perhaps she did not recognize how genre features manifested themselves in the child’s narrative. In fact, Megan did not acknowledge the student author at all except to suggest working with the child to correct errors.

In contrast, Keri seemed to have a deeper understanding of what children can accomplish as writers. She recognized a variety of genre-specific features, treating the student’s writing as an authentic narrative. More sophisticated KCS was also particularly evident in Keri’s stance toward the writing sample. Viewing the student as an author, communicator, meaning-maker, and thinker, she used her KCS to consider possibilities about the child’s thought processes. She pointed out possible connections to science and questioned whether the student understood Python as a name of a snake or a type of snake. Thus, KCS appears to undergird effective reading and responding by helping teachers to (a) recognize novice forms of various writing features, (b) recognize that students are engaged in intentional thought processes, and (c) interpret and diagnose students’ misunderstandings.

Second, these cases seem to highlight differences in participants’ “knowledge of content and teaching” (KCT; Ball et al., 2008). KCT integrates knowledge of teaching with knowledge of content. Teachers use this knowledge to make instructional decisions about how to make content accessible to students—to support students in developing understanding and proficiency. Participants drew on KCT as they weighed the importance of the various strengths and weaknesses they noticed in the writing sample and determined response topics and strategies. Whereas Megan and Lisa demonstrated minimal KCT by suggesting vague basic strategies (e.g., do a mini-lesson, pull the child aside) aimed at correcting errors in the current draft, Keri demonstrated greater KCT by identifying specific strategies to address both surface and deep features (e.g., peer editing, detailed word choice mini-lesson). She also took a long-term view of instruction, suggesting the development of a tool that would simultaneously improve the current draft as well as support the student in future writing. Thus, KCT undergirds effective reading and responding because it helps teachers determine when and how to teach students the knowledge, skills, and strategies that can support their ongoing growth as writers.

KCS and KCT are two smaller components of a broader category of teacher knowledge: pedagogical content knowledge (PCK). Although researchers have described PCK in different ways (Ball et al., 2008), Shulman (1986) first defined it as the subject matter knowledge needed for teaching, an amalgam of subject matter knowledge and pedagogical knowledge, the content knowledge “that embodies the aspects of content most germane to its teachability” (p. 9). It is the specialized content knowledge that distinguishes a good writing teacher from a good writer, including an understanding of what makes certain topics or skills easy or difficult to learn, the ability to diagnose and respond to misunderstandings when they arise, and a repertoire of representations of content that can support student learning. Because KCS and KCT explain key differences among cases, we conclude that the best interpretation for Dimension 1 is that it represents PCK and plots participants along a PCK continuum. In other words, the closer a participant is plotted toward the positive pole of Dimension 1, the more developed his or her PCK.

Discussion

The purpose of this research was to investigate the relationship among the components of reading and responding to student writing to better understand the underpinnings of the practice and how teacher educators might decompose it so that preservice teachers can learn to skillfully enact it. MCA results suggest that all five components are interrelated, but that three may be particularly interrelated and central to supporting children in meeting the CCSS’s learning goals for writing: Topic, Response Topic, and Response Strategy. Identifying the critical components of a core practice is important because, given the time constraints of teacher education programs, preservice teachers may not have enough time to learn all of a practice’s nuances. However, by learning the most crucial and powerful components, they can enact the practice in a way that still supports student learning.

The three cases demonstrate the importance of the Topic, Response Topic, and Response Strategy components of reading and responding. For example, Megan attended to only the surface features of the child’s writing, and her only suggested response was to edit mechanical errors. Thus, the student had no opportunity to develop his narrative using effective techniques, details, or event sequences (Standard 3) or to consider the text’s clarity, coherence, development, organization, or style (Standard 4). In addition, Megan’s response that she would simply “help” the student edit makes it unclear whether he would have developed an understanding of language skills that he could apply to his future writing. Her response was troubling in light of evidence that children need explicit instruction about revision and editing strategies (Graham, Harris, & Santangelo, 2015) so they learn strategies that will guide their writing of future texts.

In contrast to Megan, Keri attended to the both the surface and deep features of the child’s writing and suggested specific, specialized instructional strategies. In addition to mechanics, she considered the writer’s ideas, the narrative’s organization, and whether the details were appropriate for the genre (Standards 3 and 4). This attention to genre features is important because children’s understandings of discourse knowledge contribute to their success as writers (Hayes & Olinghouse, 2015; McCutchen, 2011; Olinghouse & Graham, 2009). Keri also considered the process that might have led the child to produce a text with these features (Standard 5). Furthermore, she suggested instructional responses that would allow the child to learn strategies that he could apply not just to the current text but also to his future writing as well.

Our findings provide evidence that differences in PCK contribute to differences in participants’ reading and responding. While PCK has been examined extensively in science and mathematics, this study is among the first to empirically identify how PCK pertains to writing instruction. Unlike many studies, we did not set out to describe or measure PCK. Therefore, the emergence of PCK from our analysis is particularly important, adding credibility to PCK as a relevant construct for the teaching of writing.

More specifically, this study suggests that at least two subsets of PCK undergird skillful reading and responding: KCS and KCT. Teachers draw on KCS when they read students’ writing. Well-developed KCS supports teachers in attending to a broad range of features and reasoning about those features to identify children’s learning needs. In contrast, teachers draw on KCT when responding to students’ writing. Well-developed KCT supports teachers in selecting targeted instructional strategies to address children’s learning needs. Our data suggest that KCS and KCT, while related, may not develop simultaneously. For example, teachers can attend and respond to deep features in the writing sample without identifying targeted strategies likely to improve those features. The reverse does not seem true. Participants who attended and responded only to surface features did not identify any specialized response strategies. Thus, it appears that the application of KCT depends, at least in part, upon the development of KCS. Future research should further explore the relationship between KCS and KCT within the context of reading and responding.

The findings of this study have several implications for teacher education. First, it is clear that content knowledge, such as that represented in the first three Writing CCSS, is insufficient for effective practice. Preservice teachers can know the features of narratives and use them proficiently in their own writing without recognizing them in students’ writing. While content courses, such as Freshman Composition or the Writing for Elementary Educators course taken by participants in this study, provide an important foundation by teaching genre features and writing strategies generally, teacher education programs must extend this knowledge by helping teacher candidates become familiar with how these features typically appear in children’s writing (KCS).

Analyzing children’s writing is one way to bridge preservice teachers’ content knowledge about writing to the KCS needed for reading and responding because it provides opportunity to examine children’s novice, emerging attempts at writing. Preservice teachers can analyze children’s writing with respect to research and theory on children’s writing development and with respect to exemplars provided in Appendix C of the CCSS (http://www.corestandards.org/assets/Appendix_C.pdf). Studies have shown that preservice teachers can learn to analyze children’s writing in increasingly sophisticated ways over time (Davenport, 2006; Gibson, 2007; Moore & Seeger, 2009). However, more research is needed to further clarify how teachers develop skill in analyzing children’s writing.

Similarly, instructional strategies cannot only be taught in isolation if preservice teachers are to learn to read and respond effectively. This prompts a rethinking of traditional methods courses that teach pedagogical skills without necessarily connecting to children’s writing development. Our research makes clear that reading and responding to student writing is one integrated practice: instructional responses are contingent upon a prioritization of the features teachers recognize in a child’s text. We assert that preservice teachers need opportunities to rehearse reading and responding as an integrated practice so they can receive feedback on their attempts to link responses with identified student needs and further develop their PCK in context. Rehearsals of this integrated practice may take on a variety of forms in a methods course, such as using protocols to guide structured collaborative discussions of writing samples in class, collaboratively planning lessons based on a class set of student writing samples, planning and rehearsing writing conferences in response to writing samples, or completing assignments requiring both analysis of and response to a piece of writing. Rehearsals should also occur within field placements. These might include structured conversations about student texts with cooperating teachers or grade-level teams, cofacilitating student writing conferences, videotaping and analyzing the enactment of reading and responding during student writing conferences, and reflecting on the results of lessons planned in response to student texts.

A final recommendation is that rehearsals of reading and responding place particular emphasis on the deep features of children’s writing. Research with both preservice (Ballock & McQuitty, 2014) and inservice (Matsumara et al., 2002) teachers demonstrates the tendency of some to attend and respond to primarily surface features. Yet, the CCSS require children to produce texts with substantive ideas, clear organization, and genre-specific features—attributes that extend far beyond conventions. Thus, meeting the expectations of the CCSS requires teachers to both recognize deep features (or lack thereof) in children’s writing and to respond with pedagogies that teach deep features. Furthermore, because the CCSS address narrative, informational, and persuasive/argumentative writing, preservice teachers must learn to read and respond in genre-specific ways. This study focused on the knowledge needed to read and respond to a narrative, but ultimately, teacher education courses must address KCS and KCT for reading and responding to other genres as well.

We recommend that future research examine how teachers learn to read and respond to different genres at different grade levels. Given the importance of KCS to the practice, we wonder how much difference exists between, for example, reading and responding to kindergartener’s personal narratives and reading and responding to third graders’ personal narratives. Similarly, how much difference exists between reading and responding to fifth graders’ personal narratives and fifth graders’ opinion texts? Is there a point at which a teacher is skillful enough in reading and responding that she can easily learn to apply the practice to new genres and grade levels, or must she specifically learn how each genre manifests in the writing of students at each level? Answers to these questions could provide useful additional guidance for literacy teacher educators.

In summary, this study highlights reading and responding to student writing as a core practice that when enacted skillfully is likely to support children in meeting the CCSS’s challenging demands. This study provides insight into the nature of reading and responding, specifically in terms of its reliance on the development of teachers’ PCK. As teacher educators work to develop pedagogies of teacher education to teach reading and responding as a practice, it will be important not only to decompose reading and responding into component parts that can be rehearsed in isolation but also to support novice teachers in enacting reading and responding as an integrated whole, connecting their knowledge of teaching and content with their knowledge of teaching and students to make targeted instructional decisions.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Author Biographies

Ellen Ballock is an associate professor in the Department of Early and Elementary Education at Gordon College in Wenham, MA. Her research focuses on preservice and inservice teacher professional development with a specific focus on learning through collaboratively examining artifacts of practice.

Vicki McQuitty is an assistant professor in the Department of Elementary Education at Towson University in Towson, MD. Her research and teaching center on preparing preservice and inservice teachers to teach writing.

Scot McNary is associate professor in the Department of Educational Technology and Literacy at Towson University in Towson, MD. His research and teaching are focused on quantitative methods and social interactions in learning environments.

References

Abell

S. K.

(2008). Twenty years later: Does pedagogical content knowledge remain a useful idea? International Journal of Science Education, 30, 1405-1416. doi:10.1080/09500690802187041

Ball

D. L.

Forzani

F. M.

(2009). The work of teaching and the challenge for teacher education. Journal of Teacher Education, 60, 497-511. doi:10.1177/0022487109348479

Ball

D. L.

Thames

M. H.

Phelps

(2008). Content knowledge for teaching: What makes it special? Journal of Teacher Education, 59, 389-407. doi:10.1177/0022487108324554

Ballock

McQuitty

(2014, April). Preservice teachers’ noticing in elementary student writing. Presented at the Annual Meeting of the American Educational Research Association, Philadelphia, PA.

Bereiter

Scardamalia

(1987). The psychology of written composition. Hillsdale, NJ: Lawrence Erlbaum.

Clausen

(1998). Quantitative applications in the social sciences, 121: Applied correspondence analysis: An introduction. Thousand Oaks, CA: SAGE.

Cutler

Graham

(2008). Primary grade writing instruction: A national survey. Journal of Educational Psychology, 100, 907-919. doi:10.1037/a0012656

Davenport

N. A.

(2006). Connecting preservice teachers with students using email to build skills for teaching writing. Journal of Reading Education, 31(2), 13-19.

DeGroff

(1992). Process-writing teachers’ responses to fourth-grade writers’ first drafts. Elementary School Journal, 93, 131-144.

10.

Ell

Hill

Grudnoff

(2012). Finding out more about teacher candidates’ prior knowledge: Implications for teacher educators. Asia-Pacific Journal of Teacher Education, 40(1), 55-65. doi:10.1080/1359866X.2011.643760

11.

Forzani

F. M.

(2014). Understanding “core practices” and “practice-based” teacher education: Learning from the past. Journal of Teacher Education, 65, 357-368. doi:10.1177/0022487114533800

12.

Gibson

S. A.

(2007). Preservice teachers’ knowledge of instructional scaffolding for writing instruction. Mid-Western Educational Researcher, 20(2), 9-15.

13.

Gilbert

Graham

(2010). Teaching writing to elementary students in grades 4-6: A national survey. Elementary School Journal, 110, 494-518. doi:10.1086/651193

14.

Glasswell

Parr

J. M.

McNaughton

(2003). Four ways to work against yourself when conferencing with struggling writers. Language Arts, 80, 291-298.

15.

Graham

Harris

K. R.

(2015). Common Core State Standards and writing: Introduction to the special issue. Elementary School Journal, 115, 457-463.

16.

Graham

Harris

K. R.

Santangelo

(2015). Research-based practices and the Common Core: Meta-analysis and meta-synthesis. Elementary School Journal, 115, 498-522.

17.

Graham

Hebert

Harris

K. R.

(2015). Formative assessment and writing: A meta-analysis. Elementary School Journal, 115, 523-547.

18.

Grossman

P. L.

Hammerness

McDonald

(2009). Redefining teaching, re-imagining teacher education. Teachers and Teaching: Theory and Practice, 15(2), 273-289.

19.

Harper

Platt

Naranjo

Boynton

(2007). Marching in unison: Florida ESL teachers and No Child Left Behind. TESOL Quarterly, 41, 642-651.

20.

Hayes

J. R.

Olinghouse

N. G.

(2015). Can cognitive writing models inform the design of the Common Core State Standards? Elementary School Journal, 115, 480-497.

21.

Husson

Josse

Mazet

(2015). FactoMineR: Multivariate exploratory data analysis and data mining (R package version 1.29) [Software]. Retrieved from http://CRAN.R-project.org/package=FactoMineR

22.

Jacobs

V. R.

Lamb

L. L. C.

Philipp

R. A.

(2010). Professional noticing of children’s mathematical thinking. Journal of Research in Mathematics Education, 41, 169-202.

23.

Josse

Husson

(2008). FactoMineR: An R package for multivariate analysis. Journal of Statistical Software, 25(1), 1-18. Retrieved from http://factominer.free.fr/docs/article_FactoMineR.pdf

24.

Masuda

Ebersole

(2012). Beginning teachers’ study groups: A place to continue critical conversations about teaching writing. The New Educator, 8, 160-176. doi:10.1080/1547688X.2012.670570

25.

Matsumara

L. C.

Patthey-Chavez

G. G.

Valdes

Garnier

(2002). Teacher feedback, writing assignment quality, and third-grade students’ revision in lower and higher achieving urban schools. Elementary School Journal, 103, 3-25.

26.

McCarthey

S. J.

(2008). The impact of No Child Left Behind on teachers’ writing instruction. Written Communication, 25, 462-505. doi:10.1177/0741088308322554

27.

McCutchen

(2011). From novice to expert: Implications of language skills and writing-relevant knowledge for memory during the development of writing skills. Journal of Writing Research, 3, 51-68. doi:10.17239/jowr-2011.03.01.3

28.

McDonald

Kazemi

Kavanagh

S. S.

(2013). Core practices and pedagogies of teacher education: A call for a common language and collective activity. Journal of Teacher Education, 64, 378-386. doi:10.1177/0022487113493807

29.

Moore

R. A.

Seeger

(2009). Dear sincerely: Exploring literate identities with young children and preservice teachers through letter writing. Literacy Research and Instruction, 48, 185-205. doi:10.1080/19388070802226246

30.

Olinghouse

N. G.

Graham

(2009). The relationship between the discourse knowledge and the writing performance of elementary-grade children. Journal of Educational Psychology, 101, 37-50.

31.

R Core Team. (2016). R: A language and environment for statistical computing [Computing software]. Vienna, Austria: R Foundation for Statistical Computing. Available from http://www.R-project.org/

32.

Reutzel

D. R.

Dole

J. A.

Read

Fawson

Herman

Jones

C. D.

. . . Fargo

(2011). Conceptually and methodologically vexing issues in teacher knowledge assessment. Reading and Writing Quarterly, 27, 183-211. doi:10.1080/10573569.2011.560098

33.

Schreier

(2012). Qualitative content analysis in practice. Thousand Oaks, CA: SAGE.

34.

Shanahan

(2015). Common Core State Standards: A new role for writing. Elementary School Journal, 115, 464-479.

35.

Sherin

M. G.

Jacobs

V. R.

Philipp

R. A.

(Eds.). (2011). Mathematics teacher noticing. New York, NY: Routledge.

36.

Shulman

L. S.

(1986). Those who understand: Knowledge growth in teaching. Educational Researcher, 15, 4-14.

37.

Spence

L. K.

(2010). Generous reading: Seeing students through their writing. The Reading Teacher, 63, 634-642. doi:10.1598/RT.63.8.2

38.

Teaching Works. (2016). High leverage practices. Retrieved from http://www.teachingworks.org/work-of-teaching/high-leverage-practices

39.

van Es

E. A.

Sherin

M. G.

(2006). How different video club designs support teachers in “learning to notice.” Journal of Computing in Teacher Education, 22, 125-135.