Learning and Instruction 33 (2014) 1e11
Contents lists available at ScienceDirect
Learning and Instruction
journal homepage: www.elsevier.com/locate/learninstruc
Do students’ beliefs about writing relate to their writing self-efficacy,
apprehension, and performance?
Joanne Sanders-Reio a, *, Patricia A. Alexander b, Thomas G. Reio, Jr. a, Isadore Newman a
a
b
College of Education, Florida International University, Miami, FL 33199, USA
College of Education, University of Maryland, College Park, MD 20742, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 12 September 2012
Received in revised form
9 February 2014
Accepted 14 February 2014
Available online 18 March 2014
This study tested a model in which beliefs about writing, writing self-efficacy, and writing apprehension
predict writing performance. The Beliefs About Writing Survey, the Writing Self-Efficacy Index, and the
modified Writing Apprehension Test were administered to 738 undergraduates to predict their grade on
a class paper. In a hierarchical regression, beliefs about writing predicted variance in writing scores
beyond that accounted for by writing self-efficacy and apprehension. Audience Orientation, a new belief
associated with expert practice, was the strongest positive predictor of the students’ grade. Transmission,
a belief in relying on material published by authorities, was the leading negative predictor. Writing selfefficacy predicted performance, albeit modestly. The traditional measure of writing apprehension
(anxiety about being critiqued) was not significant, but Apprehension About Grammar, a new construct,
significantly and negatively predicted performance. These results support the possibility that beliefs
about writing could be a leverage point for teaching students to write.
Ó 2014 Elsevier Ltd. All rights reserved.
Keywords:
Writing self-efficacy
Writing apprehension
Beliefs about writing
Expertise
Writing development
1. Introduction
Social cognitive theory established the importance of beliefs in
human learning and performance. The most important of these
beliefs are self-efficacy beliefs, one’s confidence in one’s ability to
perform tasks required to cope with situations and achieve specific
goals. People with high self-efficacy are more likely to take on
challenges, try harder, and persist longer than those with low selfefficacy (Bandura, 1989). People with high self-efficacy tend to be
less apprehensive and to confront anxiety-producing situations to
reduce their threat, while those with low self-efficacy avoid such
situations (Pajares, 1997). Bandura maintains that there are four
sources of self-efficacy, with the most influential being one’s previous successes and mastery experiences (Bandura, 1997).
Thirty years of research with students ranging from fourth
graders to undergraduates supports the linkages between selfefficacy, apprehension, and performance with respect to writing.
Students with high writing self-efficacy write better and are less
apprehensive about writing than those with low writing self-
* Corresponding author. Tel.: þ1 305 348 0124; fax: þ1 305 348 1515.
E-mail addresses: jsanders@fiu.edu, sandersreio@netscape.net (J. Sanders-Reio),
palexander662@gmail.com (P.A. Alexander), reiot@fiu.edu (T.G. Reio), newmani@
fiu.edu (I. Newman).
http://dx.doi.org/10.1016/j.learninstruc.2014.02.001
0959-4752/Ó 2014 Elsevier Ltd. All rights reserved.
efficacy (e.g., McCarthy, Meier, & Rinderer, 1985; Pajares &
Valiante, 1999). Correlations between writing self-efficacy and
writing performance have ranged from .03 (Pajares & Johnson,
1994) to .83 (Schunk & Swartz, 1993), clustering around .35, while
correlations between writing performance and writing apprehension have ranged from .28 (Meier, McCarthy, & Schmeck, 1984)
to .57 (Pajares & Johnson, 1994).
1.1. Beliefs about writing
More recent work has extended the social cognitive view of
writing by exploring whether another type of belief, beliefs about
writing, also relates to writing performance and its established
correlates, writing self-efficacy and apprehension. In contrast to
writing self-efficacy beliefs (i.e., one’s beliefs about one’s own
writing skills), beliefs about writing address what good writing is
and what good writers do. As Graham, Schwartz, and MacArthur
(1993) wrote, “The knowledge, attitudes, and beliefs that students
hold about writing play an important part in determining how the
composing process is carried out and what the eventual shape of
the written product will be” (p. 246). Mateos et al. (2010) similarly
described these beliefs as “filters leading students to represent the
task of.writing to themselves in a particular way,” with the
various models of writing created by these beliefs leading to
“different engagement patterns” (p. 284).
2
J. Sanders-Reio et al. / Learning and Instruction 33 (2014) 1e11
Scholars of both educational psychology and writing and rhetoric have studied beliefs about writing. Palmquist and Young (1992)
conducted one of the first empirical studies of these beliefs, an
examination of the belief that writing is an innate gift that some
have and others lack. Overall, undergraduates who believed that
writing ability is innate were more apprehensive about writing, had
lower estimates of their writing skills and abilities (a belief akin to
self-efficacy), and were less confident in their potential to become
good writers. The authors concluded that the belief in innateness
“appears to make an important, though largely unacknowledged,
contribution to a constellation of expectations, attitudes, and beliefs that influence the ways in which students approach writing”
(p. 159). More specifically, the authors found an interaction between self-appraisals and apprehension, and the belief in innateness. Among students who had low assessments of their own
writing, the belief in the innateness of writing ability strongly
correlated with writing apprehension, while among students with
high appraisals of their own writing, the belief in innateness did not
relate to apprehension. The authors suggested that students with
low assessments of their written work and high writing apprehension might use the belief in innateness to rationalize their poor
performance.
Silva and Nicholls (1993) studied the beliefs underlying six
traditions of teaching writing: those emphasizing (a) personal
involvement, (b) writing for understanding, (c) mechanical correctness, (d) collaboration, (e) cognitive strategies, and (f) models of
good writing. The authors developed two genre-neutral scales, one
based on the characteristics of good writing espoused by each
tradition and the other reflecting the writing strategies that
emerged from each perspective. A principal components analysis
(PCA) of each scale, followed by a second-order PCA of the resulting
components, yielded four emphases: (a) personal meaning and
enjoyment of words, (b) a recursive approach fostering understanding, (c) focus on audience and strategies, and (d) surface
correctness and form.
Lavelle (1993) published a number of studies about students’
approaches to writing, a broad construct encompassing beliefs
about writing, writing self-efficacy, writing goals, and writing
strategies. A factor analysis of college students’ survey responses
yielded five approaches that fell into two categories, deep and
surface. The deep approaches included the elaborationist approach,
characterized by personal and emotional involvement, and the
relative-revisionist approach, with its strong audience awareness
and in-depth revision. The surface approaches were the low selfefficacy approach, with its relative lack of writing strategies; the
spontaneous-impulsive approach, characterized by a one-step
process and lack of personal meaning; and the procedural
approach, with its reliance on strategies. Writers using deep approaches had a stronger sense of audience and revised more, both
globally and locally. Those using surface approaches were less
invested in their writing, used fewer writing strategies, and were
less aware of their audience and writing process.
1.1.1. Transaction and Transmission
White and Bruning (2005) explored whether two established
beliefs about reading, Transaction and Transmission (Schraw &
Bruning, 1996, 1999), influence students’ writing. Writers with
high Transaction beliefs are emotionally and cognitively engaged in
their writing process. They see writing as a means of deepening
their understanding of the concepts they write about and their own
views. By contrast, those with high Transmission beliefs regard
writing as a means for reporting what authorities think. These
writers stick to the information and arguments they find in
established sources. Transaction and Transmission are independent
of one another, so individuals can espouse neither, one, or both of
these beliefs. Students with high Transaction beliefs earned
significantly higher grades for their written work, while those with
high Transmission beliefs received significantly lower scores.
Transaction positively correlated with writing self-efficacy, but did
not relate to writing apprehension. Transmission related to neither
self-efficacy nor apprehension. The authors suggested that these
beliefs influence writing performance via affective (e.g., apprehension), cognitive and behavioral writing processes.
Mateos et al. (2010) extended White and Bruning’s (2005) work
by studying writers’ adherence to Transaction and Transmission
beliefs along with their support of the epistemic beliefs examined
by Schommer-Aikins (2004). As in the White and Bruning (2005)
study, Transaction positively correlated with academic achievement, while Transmission negatively related to achievement.
Additionally, Transaction negatively related to Fixed Ability (intelligence is defined, not malleable), Simple Knowledge (knowledge is
comprised of discrete facts, not complex, conceptual structures),
and Quick Learning (learning occurs immediately or not at all).
Transmission positively related to Simple Knowledge.
1.1.2. Kellogg’s model of writing development
The development of the four-factor beliefs about writing
framework presented here was guided by Kellogg’s (2008) cognitive model of writing development. Kellogg built on Bereiter and
Scardamalia’s (1987) two-stage developmental model of Knowledge Telling and Knowledge Transforming. Knowledge Tellers record what they know about a topic, primarily as their ideas occur to
them. Knowledge Transformers are aware of discrepancies between what they intend to write and what their text actually says.
These writers revise to bridge these gaps, and they refine their
understandings and rethink their ideas as they work. Kellogg added
a third stage, Knowledge Crafting, which describes expert writing.
Knowledge Crafters tailor their writing to an audience they have
richly represented in their minds. They select which information to
include and decide how to present it with this audience in mind.
A major difference between the writers in these three stages is
the number of perspectives and representations they maintain as
they write. Knowledge Tellers have one main perspective, their
own representation of the text, and only a tenuous grasp of what
their paper actually says. Knowledge Transformers consider two
perspectives, their ideal text and their actual manuscript; they
revise to make their paper more like their ideal representation.
Knowledge Crafters juggle three rich and stable representations of
their work: their ideal paper, the text as it actually reads, and the
text as they think their readers will understand it. Writers in this
final stage regulate themselves cognitively, emotionally, and
behaviorally.
Writers move from one stage to the next only after many of their
writing skills have become fluid and their ability to represent their
text, in its ideal and actual forms, well developed and stable.
Because of the heavy demands writing makes on working memory,
particularly central executive function, Kellogg (2008) estimated
that it takes writers about 10 years to master each of the first two
stages. Only experts and those who write extensively reach stage
three, and normally not before adulthood. Even then, they usually
write at this level in only a few genres. Because the oldest students
that Bereiter and Scardamalia (1987) studied were undergraduates,
Kellogg reasoned that most of them were in the first two stages
because they would not have had the time to gain stable executive
control over the skills associated with stage three.
Although Kellogg (2008) clearly delineated these stages, he did
not cast them as discrete. Rather, he described writing skills and
representations as being on a continuum. He allowed that writers
in the first stage may have some conception of their audience and
their actual text, but maintained that such representations are
J. Sanders-Reio et al. / Learning and Instruction 33 (2014) 1e11
sketchy and unstable. Scheuer, de la Cruz, Pozo, Huarte, and Sola’s
(2006) study of children’s conceptions of the writing process supports Kellogg’s views. The Scheuer group examined the conceptions
that kindergarteners, first graders, and 4th graders have about
writing and thinking at four points in the writing process: before
writing, and during writing, revising, and rereading. At each higher
grade, the children had a more complex working conception of
writing. The kindergarteners worked to capture oral language on
paper, while the 4th graders worked to elaborate on and organize
their papers so they made sense, were thematically unified, and
complied with writing conventions. The older children had more
executive control and thought more as they wrote.
1.1.3. The importance of audience
Researchers and experts from numerous disciplines concur with
Kellogg’s emphasis on audience. In the research on beliefs about
writing discussed above, one of the four components generated by
Silva and Nicholls’s (1993) secondary PCA stressed the writers’
rapport with their audience. Similarly, one of the main differences
between writers who used deep approaches and those who used
surface approaches in Lavelle’s (1993) studies was that those who
used deep approaches had a stronger sense of audience. Elsewhere,
Beach and Friedrich (2006), and Miller and Charney (2008) discussed how writers adjust the presentation, content, and tone of
their arguments in response to their audience and whether this
audience is part of the writers’ usual discourse community. Finally,
with regard to instruction, writing teachers have emphasized
rhetoric, the study of how to influence and persuade readers, since
the time of Aristotle (Miller & Charney, 2008).
1.1.4. Writing as a recursive process
Scholars and practitioners also agree that proficient writers use
a recursive process. Silva and Nicholls’s (1993) secondary PCA
yielded a component that emphasized an iterative approach to
writing. Lavelle (1993) found that deep approaches involved indepth revision. Iteration is also foundational to Hayes and
Flower’s (e.g., 1980) process writing model. Similarly, renowned
journalists/writing teachers William Zinsser and Donald Murray
stress the recursive nature of writing. Zinsser (1976), whose On
Writing Well has sold well over a million copies, maintained,
“Rewriting is the essence of writing” (p. 4), while Murray (1991)
declared, “Writing is rewriting” (p. vii).
1.1.5. Examining these beliefs about writing in light of Kellogg’s
model
With respect to Kellogg’s model (2008), White and Bruning’s
(2005) Transmission (i.e., the purpose of writing is to convey information published by authorities) can be seen as a form of stage
one, Knowledge Telling, as writers inform readers about what authorities have written in sources such as textbooks, encyclopedias,
and journals without necessarily engaging with the material or
incorporating their own insights and analyses. Transaction (i.e.,
writers are cognitively and emotionally engaged in their writing
and think through their views as they write) aligns with stage two,
Knowledge Transforming, where writers work to make their actual
papers more in accord with their goals for their papers and their
mental representations of the content. Recursive Process, the belief
that accomplished writers go through multiple versions of their
plans and drafts, also falls under the second stage, where writers
rework their papers to refine their understandings and their presentation of those understandings. Audience Orientation, which
holds that writers should adapt their writing to the needs of their
readers, aligns with Kellogg’s third stage, Knowledge Crafting,
which emphasizes audience.
3
1.2. Writing self-efficacy and writing apprehension
The earliest writing self-efficacy scales emphasized mechanical
writing skills (e.g., Meier et al., 1984). Subsequent measures also
addressed substantive writing skills (e.g., Pajares & Valiante, 1999)
and writing self-regulation (Zimmerman & Bandura, 1994). This
investigation examines all three types of writing self-efficacy. Daly
and Miller (1975) operationalized writing apprehension as avoidance of writing and the expectation of negative evaluations of one’s
written work. In the Pajares group’s (e.g., Pajares & Valiante, 1997)
path analyses, writing self-efficacy directly and positively predicted
writing performance, and indirectly contributed to students’
writing scores by reducing and even nullifying writing apprehension as defined by Daly and Miller (1975). However, Smith, Cheville,
and Hillocks (2006) suggested that there may be another type of
writing apprehension, a fear of making mechanical errors such as
spelling and grammar mistakes.
1.3. Purpose of the study, research questions, and hypotheses
This study builds on the work of researchers of educational
psychology, writing and rhetoric, and expert practice to examine a
model of how beliefs about writing, writing self-efficacy, writing
apprehension, and writing performance relate to one another. In so
doing, it augments White and Bruning’s (2005) work by examining
two new beliefs about writing, Recursive Process and Audience
Orientation, in addition to Transaction and Transmission. This study
also combines and expands existing measures of writing selfefficacy to create a three-component scale that assesses selfefficacy for substantive writing issues, writing mechanics, and
writing self-regulation. Finally, the study operationalizes writing
apprehension more broadly than do existing studies by investigating apprehension about grammar and correctness as well as
avoidance of writing and anxiety about having one’s written work
evaluated.
The following research questions guided the study:
1. What are the relations among beliefs about writing, writing selfefficacy, writing apprehension, and writing performance?
2. What are the unique contributions of beliefs about writing,
writing self-efficacy, and writing apprehension to writing
performance?
In developing our conceptual model, we were guided first by
researchers (e.g., White & Bruning, 2005) who suggested that students’ beliefs about writing influence their writing process, which
would include their selection of writing strategies. Thus, it seems,
for example, that students who believe that writing is a recursive
process would be more likely to proofread and revise their work
and to attend to instruction on revision, thereby sharpening their
editing skills. Proofreading and revising would in turn improve
writing performance, thereby creating the mastery experiences
that, according to Bandura (1997), enhance self-efficacy. An
exception might occur with maladaptive beliefs, such as the belief
that writing is an innate gift, which are associated with weak
writing performance. Such maladaptive beliefs may operate in a
different way, as rationalizations for poor writing performance, as
Palmquist and Young (1992) suggested. With respect to selfefficacy, we were informed by Lavelle (1993), who found that students with low writing self-efficacy reported that they used few
writing strategies. We thus reasoned that a belief in the helpfulness
of effective writing strategies would precede the development of
writing self-efficacy. Finally, for apprehension, we were informed
by the Pajares group’s path-analytic research (e.g., Pajares &
Johnson, 1994), which indicated that writing self-efficacy often
4
J. Sanders-Reio et al. / Learning and Instruction 33 (2014) 1e11
nullifies the effects of writing apprehension. In sum, we saw beliefs
about writing as affecting writing self-efficacy, and self-efficacy as
affecting apprehension.
For the first research question, based on the extant literature, we
predicted that three beliefs about writing, Transaction, Recursive
Process, and Audience Orientation, would relate significantly and
positively to writing self-efficacy (Hypothesis 1a) and writing performance (Hypothesis 1b), and relate negatively to writing apprehension (Hypothesis 1c). By contrast, we expected the fourth belief,
Transmission, to relate negatively to writing self-efficacy (Hypothesis 1d) and writing performance (Hypothesis 1e) and relate
positively to writing apprehension (Hypothesis 1f). We hypothesized that all three types of writing self-efficacy would positively
relate to writing performance (Hypothesis 1g) and negatively relate
to writing apprehension (Hypothesis 1h). Finally, we predicted that
both types of writing apprehension would negatively relate to
writing performance (Hypothesis 1i).
For the second research question, we expected Transaction,
Recursive Process, and Audience Orientation to significantly and
positively predict writing performance (Hypothesis 2a) but that
Transmission would negatively predict writing performance (Hypothesis 2b). We further hypothesized that all three types of
writing self-efficacy would positively predict writing performance
(Hypothesis 2c) while both writing apprehension subscales would
predict negative variance in writing performance (Hypothesis 2d).
Finally, we predicted that the beliefs about writing would explain
variance in writing performance above and beyond that accounted
for by writing self-efficacy and apprehension (Hypothesis 2e).
2. Method
2.1. Participants
The participants were 738 undergraduates at a large, researchintensive, Hispanic-serving, public university in south Florida.
This study focused on undergraduates because they receive
considerable instruction and practice in writing and because using
undergraduates facilitates comparison with previous research (e.g.,
Zimmerman & Bandura, 1994). The students were enrolled in an
educational psychology class preservice teachers must take. Most
(86%) were women; 68% were Hispanic, 16% were white, 11% were
black, and 2% were Asian. Most were juniors (68%) or seniors (24%).
With respect to their parents’ education, 88% of the participants’
fathers and 91% of their mothers graduated from high school, 32%
of their fathers and 32% of their mothers earned a bachelor’s degree, and 14% of their fathers and 13% of their mothers held an
advanced degree. On the other hand, 7% of the participants’ fathers
and 5% of their mothers had an 8th grade education or less. More
than 37% reported that their first language was Spanish, 31% English, and 3% another language; 28% said they were raised
bilingually.
2.2.2. Modified Writing Apprehension Test
Daly and Miller’s (1975) Writing Apprehension Test (WAT), a 5point Likert scale, was used to facilitate comparison with previous
research. Daly and Miller saw the WAT as assessing polar aspects of
a single factor. Other researchers (e.g., Bline, Lowe, Meixner, Nouri,
& Pearce, 2001) saw this measure as having two subscales, fear and
avoidance of having one’s written work evaluated, and enjoyment
in sharing one’s written work with others. These subscales are often
called “Dislike of Writing” and “Enjoyment of Writing,” and have 13
items each. Reliability estimates have been consistently greater
than .9 (e.g., Bline et al., 2001). Based on the experiences of Smith
et al. (2006), we added three items to the WAT to assess apprehension about making errors with respect to grammar, punctuation, and spelling (I worry that I may make a grammatical error; I’m
afraid that I will make a punctuation error; I’m afraid that I may
miss a misspelled word or typo.) The entire modified scale has 29
items.
2.2.3. Beliefs About Writing Survey
Beliefs about writing were assessed with four subscales (Audience Orientation, Recursive Process, Transaction, and Transmission,
50 items) of the Beliefs About Writing Survey, a 5-point Likert scale
developed for a related study (BWS; Sanders-Reio, 2010). These
subscales align with the Kellogg model (2008), as discussed above.
The BWS is an expansion of White and Bruning’s (2005) Writing
Beliefs Inventory, which had two subscales, Transaction and
Transmission, previously described.
2.2.4. Writing performance
Assessment of student writing in much research and many highstakes tests tends to be based on a single sample, written in 20e
30 min in response to a prompt and graded with a single score
(Hillocks, 2008). Although valuable, such assessments have serious
shortcomings: their often-inflexible format, lack of authenticity,
emphasis on speed over skill, discrimination against students unfamiliar with the topic, and restraints on students’ use of writing
strategies (Coker & Lewis, 2008). Writing researchers (e.g., Murphy
& Yancy, 2008) and the National Council of Teachers of English
(2008) have called for more ecologically valid indicators of
writing performance. Here, writing performance, the dependent
variable, was assessed via the students’ grades on a 5- to 8-page,
structured, take-home assignment, an analysis of a video about
three preschools in light of learning theory.
Participants completed the surveys during 20e40 min of class
time with respect to papers they write at the university. They did so
after they understood the writing assignment, but before they
could begin working on it, as Bandura recommended (Pajares,
1997). Students received extra credit for participating; no one
refused the opportunity. Two professors, each of whom has taught
this course, graded all of the papers. The interrater agreement between the graders, .93, was calculated using correlational analysis,
as directed by Gay (1992).
2.2. Measures
3. Results
2.2.1. Writing Self-Efficacy Index
Writing self-efficacy was assessed with the Writing Self-Efficacy
Index (WSI; Sanders-Reio, 2010), which was based on Zimmerman
and Bandura’s (1994) Writing Self-Regulatory Efficacy Scale of 25
items primarily relating to the self-regulation of writing projects
and processes (e.g., “I can start writing with no difficulty”). Questions addressing substantive and mechanical writing issues were
added, based on the research and practice literature, for a total of 60
items. Participants indicated their confidence in their ability to
perform a task by making a hash mark on a 10-mm line marked 0 to
100 (e.g., Nietfeld & Schraw, 2002).
3.1. Examination of the measures
3.1.1. Writing Self-Efficacy Index and the modified Writing
Apprehension Test
PCAs with varimax rotation were used to examine the structures
of the WSI and WAT. The WSI had three components according to
the Kaiser Criterion and the scree plot: self-efficacy for Substantive,
Self-Regulatory, and Mechanical writing skills. The eigenvalues for
these components were 16.4, 10.9, and 10.2, respectively; they
accounted for 26.9%, 18.5%, and 17.9% of the variance, respectively,
J. Sanders-Reio et al. / Learning and Instruction 33 (2014) 1e11
for a total of 63.3% of the variance explained. Most of the items from
the Zimmerman and Bandura (1994) scale loaded on the SelfRegulatory subscale.
We expected that the three items added to the WAT to assess
apprehension about writing mechanics would strengthen the
Dislike Writing subscale. However, the Kaiser Criterion and the
scree plot indicated that these items formed a separate component,
Apprehension About Grammar. One item from the Enjoy Writing
scale, “I have no fear of my writing being evaluated,” was dropped
because its coefficient was below .30 (12 items remained). Dislike
Writing, Enjoy Writing, and Apprehension about Grammar
accounted for 23.8%, 23.4%, and 9.1% of the variance, respectively,
for a total of 56.3% of the variance explained. The eigenvalues were
6.9, 6.8, and 2.6, respectively.
3.1.2. Beliefs About Writing Survey
As the BWS had not been validated, we sought to support the
construct validity of this scale via a sequential exploratory factor
analysis (EFA)-confirmatory factor analysis (CFA) (Worthington &
Whittaker, 2006). Using the SPSS 20.0 random data-splitting protocol, we divided the dataset into two subsets of equal size
(ns ¼ 369). We then employed an EFA on the first dataset to
investigate the factor structure, reduce the number of items as
appropriate, and distinguish items that align with a simple structure (Nimon, Zigarmi, Houson, Witt, & Diehl, 2011). Finally, we
conducted a CFA with the second subsample, using the remaining
items specified by the EFA.
3.1.2.1. EFA (subsample 1). Because of the hypothesized underlying
theoretical structure and our expectation that the factors would
correlate, we used principal-axis factoring and promax rotation for
the EFA (Nimon et al., 2011). According to the scree plot and the
Kaiser criterion, the EFA revealed four interpretable factors
explaining 43.2% of the variance in the 50 submitted items. The top
four writing belief components aligned with Kellogg’s model
(2008), supporting their use in this study. Table 1 lists the items in
these four subscales.
Audience Orientation, which advocates focusing on one’s
readers and their interests. (22.3% of the variance; eigenvalue,
6.9)
Recursive Process, which reflects an iterative approach to
writing (8.1% of the variance; eigenvalue, 2.5)
Transaction (White & Bruning, 2005), which reflects affective
and cognitive engagement (7.1% of the variance; eigenvalue, 3.7)
Transmission (White & Bruning, 2005), which advocates sticking closely to the arguments, information, and quotes provided
by authorities (5.7% of the variance; eigenvalue, 1.7)
Note that the items forming Transaction and Transmission are
not identical to those used by White and Bruning (2005). Two items
that White and Bruning included in Transaction loaded on Recursive Process. In addition, each subscale contains a new item:
“Writing helps new ideas emerge” in the case of Transaction, and
“When writing, it’s best to use proven formats and templates, and
then fill in the important information” for Transmission. Finally,
several items were dropped from each scale during the exploratory
factor analysis (see below).
Factor intercorrelations ranged from .10 to .54. Analysis of the
pattern and structure coefficients that cross-loaded or were below
.40 suggested that 12 of the items could be deleted without
affecting content coverage (Hair, Anderson, Tatham, & Black, 1998).
Audience Orientation had 14 items; Recursive Process, 5 items;
Transaction, 7 items; and Transmission, 5 items, in the final
subscales.
5
Table 1
Beliefs About Writing Survey: subscales and items.
Transmission: 5 items, a ¼ .65
Good writers include a lot of quotes from authorities in their writing.
Writing should focus on the information in books and articles.
The key to successful writing is accurately reporting what authorities think.
The most important reason to write is to report what authorities think about a
subject.
When writing, it’s best to use proven formats and templates, and then fill in
the important information.
Transaction: 7 items, a ¼ .78
Writing is a process involving a lot of emotion.
Writing helps me understand better what I’m thinking about.
Writing helps me see the complexity of ideas.
My thoughts and ideas become more clear to me as I write and rewrite.
Writing is often an emotional experience.
Writers need to immerse themselves in their writing.
Writing helps new ideas emerge.
Recursive Process: 5 items, a ¼ .72
Writing requires going back over it to improve what has been written.
Good writing involves editing many times.
Writing is a process of reviewing, revisioning, and rethinking.
Revision is a multi-stage process.
The key to good writing is revising.
Audience Orientation: 14 items, a ¼ .85
Good writers make complicated information clear.
Good writers are sensitive to their readers.
Good writers support their points effectively.
Good writers adapt their message to their readers.
The key to good writing is conveying information clearly.
Good writers keep their audience in mind.
Good writers thoroughly explain their opinions and findings.
Good writers are oriented toward their readers.
Good writers are logical and convincing.
Good writers are reader-friendly.
Good writing sounds natural, not stiff.
Good writers don’t let their choice of words overshadow their message.
It’s important to select the words that suit your purpose, audience, and
occasion.
Good writers anticipate and answer their audience’s questions.
3.1.2.2. CFA (subsample 2). We arranged the 31 items identified by
the EFA in four empirically identified factors to conduct the CFA
with intercorrelated factors on the second subsample. Overall,
the goodness-of-fit indices revealed that the CFA model had an
acceptable fit to the data. The c2 analysis results (c2
[428] ¼ 986.64, p < .001) suggested that the data did not fit the
model adequately; yet, c2 tests have been shown to be especially
sensitive to larger sample sizes, for which statistical significance
is often the result (Kline, 1998). However, other methods
demonstrated evidence of factorial validity. The c2/df ratio yielded a value of 2.31, considered acceptable fit (less than 3.0 is best;
Kline, 1998). The root mean square error of approximation
(RMSEA ¼ .059; CI 90 ¼ .054, .061) was less than .06, also suggesting acceptable fit (less than .06 is recommended; Hu &
Bentler, 1999). The comparative fit index (CFI ¼ .91) and
Adjusted Good-of-Fit-Index (AGFI ¼ .90) both indicated acceptable fit (Kline, 1998) as well, although a criterion value closer to
.95 is desirable (Worthington & Whittaker, 2006).
3.1.2.3. Reliability (overall sample). The Cronbach’s alphas of the
Beliefs About Writing Survey subscales ranged from .65 to .85 (see
Table 2). All but one of these subscales (Transmission) exceeded the
.7 minimum that researchers prefer (e.g., Hair et al., 1998; Nunnally,
1967). The Cronbach’s alphas of the Writing Self-Efficacy Index
subscales ranged from .94 to .98, and those of the modified WAT
ranged from .87 to .92. Table 2 lists the means and standard deviations of each subscale.
6
J. Sanders-Reio et al. / Learning and Instruction 33 (2014) 1e11
Table 2
Reliabilities and sample items for the subscales of the Beliefs About Writing Survey, the Writing Self-Efficacy Index, and the modified Writing Apprehension Test.
Subscale
Cronbach’s a
Sample item
5
.65
The most important reason to write is to report what
authorities think about a subject.
No. items
Beliefs About Writing Survey
Knowledge Telling
Transmission
Knowledge Transforming
Transaction
Recursive Process
Knowledge Crafting
Audience Orientation
7
5
.78
.72
Writing helps me understand what I’m thinking about.
Writing is a process of reviewing, revisioning, and rethinking.
14
.85
Good writers anticipate and answer their audience’s questions.
Writing Self-Efficacy Index
Substantive
Self-Regulatory
Mechanical
25
19
11
.98
.94
.95
I can logically make the points I want to convey.
I can start writing with no difficulty.
I can correctly punctuate the papers I write.
Modified Writing Apprehension Test
Enjoy Writing
Dislike Writing
Apprehension About Grammar
12
13
3
.92
.92
.87
I like seeing my thoughts on paper.
I expect to do poorly in composition classes even before I enter them.
I’m afraid that I may make a punctuation error.
3.1.3. Grading
All participants submitted their papers to Turnitin.com to check
for plagiarism. Two professors, one the actual instructor and
another who has taught this course, assigned each paper a letter
grade ranging from A to F, including pluses and minuses. Note that
the College of Education required the students to earn at least a C on
this paper to pass the course. Those who fell short could rewrite.
Papers were evaluated with a rubric assessing course content,
substantive writing skills (i.e., development, clarity, and organization), and mechanical writing skills (i.e., usage and grammar). Students had to demonstrate basic competence in all three of these
areas to pass. For this study, grades were coded on a scale of 1e12
(from 1 for an F and 2 for a D- to 12 for an A). The mean grade of the
two professors was 8.1 (B-), with 30.2% receiving an A or A-, and
28.9% earning less than the C required. The actual course grades
were almost identical (mean grade of 8.2 [B-]). As part of the battery
of measures, students predicted the grade that they would receive
for this assignment. The correlation between the grade students
predicted and the grade they received was .13 (p < .001) (Table 3).
Table 3
Means and standard deviations of the independent and dependent variables.
Subscale
Beliefs About Writing Survey
Knowledge Telling
Transmission
Knowledge Transforming
Transaction
Recursive Process
Knowledge Crafting
Audience Orientation
Item M
Subscale M
SD
2.4
11.8
3.1
4.0
4.1
27.8
20.7
3.8
2.8
4.1
57.0
6.8
73.6
59.8
67.9
1840.1
1136.4
747.4
407.3
309.7
213.1
Modified Writing Apprehension Test
Enjoy Writing
Dislike Writing
Apprehension About Grammar
3.4
2.3
2.8
41.0
29.8
8.4
9.6
9.9
3.5
Writing Performance
Grade
8.1a
Writing Self-Efficacy Index
Substantive
Self-Regulatory
Mechanical
3.0
(A ¼ 12, A ¼ 11. F ¼ 1. 8.3 is equivalent to a B.).
a
The writing grade is a letter grade from A to F (with pluses and minuses).
3.2. Examination of the correlations between the variables
Table 4 lists the correlations between the variables, which
indicate that students who received higher grades for their papers
had higher writing self-efficacy and lower writing apprehension, as
they did in previous studies (e.g., Pajares & Valiante, 1997). Correlations between writing self-efficacy and writing performance were
within the range reported in previous research, but somewhat
lower than the norm (e.g., Pajares & Valiante, 1999). Notably, the
new writing apprehension subscale, Apprehension About
Grammar, had a stronger negative correlation with writing performance than did Dislike Writing, the traditional measure of
writing apprehension (Daly & Miller, 1975).
Only the two new beliefs about writing, Audience Orientation
and Recursive Process, positively correlated with both writing
performance and the three writing self-efficacy subscales, and also
negatively related to both measures of writing apprehension.
Audience Orientation had the strongest positive relations with
writing performance and self-efficacy. Transaction, the belief in
engaging with the writing process, was the strongest positive
correlate of all three writing self-efficacy subscales as well as Enjoy
Writing. However, it did not relate to writing performance. Transmission, which advocates basing one’s papers on the arguments
and information published by authorities, negatively related to
both writing performance and self-efficacy, and positively related
to writing apprehension. Transmission was the only belief about
writing that positively correlated with Apprehension About
Grammar. All three writing self-efficacy subscales positively related
to writing performance and negatively related to writing apprehension, including Apprehension About Grammar. Both Dislike
Writing and Apprehension About Grammar negatively related to
writing performance. These results differ from White and Bruning’s
(2005), where Transaction related positively to writing performance and self-efficacy, but not to apprehension, and where
Transmission negatively related to writing performance but did not
relate to self-efficacy or apprehension.
3.3. Relations among the research variables
To test the relations among the research variables Hypothesis 1,
we first ran a series of simultaneous regressions (see Table 5). In
each case, the hypothesis (Hypotheses 1a-i) was at least partially
supported. The effect sizes (R2) associated with most of the
regression results ranged from moderate (.06e.24) to large (.25;
J. Sanders-Reio et al. / Learning and Instruction 33 (2014) 1e11
7
Table 4
Correlations among beliefs about writing, writing self-efficacy, writing apprehension, and writing performance.
Grade
Grade
Transmission
Transaction
Recursive
Audience
SE Substantive
SE Self-Regulatory
SE Mechanical
Enjoy Writing
Dislike Writing
Apprehension
About Grammar
1
.20***
.01
.12**
.18***
.18***
.15***
.23***
.11**
.17***
.26***
Beliefs About Writing
Knowledge
Telling
Knowledge
Transforming
Transmission
Transaction
1
.03*
.26***
.03
.15***
12**
.13**
.11**
.20***
.15***
Self-efficacy
Apprehension
Knowledge
Crafting
Recursive
Audience
Substantive
Self-Regulatory
Mechanical
Enjoy
Writing
Dislike
Writing
1
.76***
.20***
1
Apprehension
About Grammar
1
.12**
.40**
.32***
.30***
.24***
.45***
.32***
.00
1
.09**
.11**
.04
.04
.07*
.10**
.00
1
.30***
.24***
.25***
.24***
.19***
.01
1
.83***
.72***
.57***
.58***
.22***
1
.63***
.60***
.59***
.20***
1
.42***
.48***
.43***
.40***
1
Note. N ¼ 738. *p < .05; **p < .01; ***p < .001.
Cohen, 1988). (When referring to the scales as they relate to the
regressions, we are referring to the scales’ scores.)
We first tested our hypothesis that Audience Orientation,
Recursive Process, and Transaction would positively predict all
three writing self-efficacy subscales, while Transmission would be a
negative predictor (Hypotheses 1a, d). The equations indicated that
the beliefs about writing predicted self-efficacy for the following:
Substantive writing skills: R ¼ .40, F(4, 733) ¼ 34.83, R2 ¼ .16,
p < .001.
Writing Self-Regulation: R ¼ .35, F(4, 733) ¼ 24.94, R2 ¼ .12,
p < .001.
Writing Mechanics: R ¼ .32, F(4, 733) ¼ 21.17, R2 ¼ .10, p < .001.
In each of these regressions, Audience Orientation and Transaction were significant positive predictors, while Transmission was
a significant negative predictor, as hypothesized. However, Recursive Process was not a significant predictor in any of these
equations.
We ran additional simultaneous regressions to test our hypothesis that Audience Orientation, Recursive Process, and Transaction would positively predict Enjoy Writing and negatively
predict Dislike Writing and Apprehension About Grammar (Hypothesis 1c), while Transmission would positively predict Dislike
Writing and Apprehension About Grammar and negatively predict
Enjoy Writing (Hypothesis 1f). The beliefs about writing predicted
the following:
Enjoy Writing: R ¼ .46, F(4, 733) ¼ 49.47, R2 ¼ .21, p < .001.
Dislike Writing: R ¼ .38, F(4, 733) ¼ 30.41, R2 ¼ .14, p < .001.
Apprehension About Grammar: R ¼ .16, F(4, 733) ¼ 4.49,
R2 ¼ .02, p ¼ .001.
As expected, Audience Orientation and Transaction positively
predicted Enjoy Writing, while Transmission was a negative predictor. Additionally, as hypothesized, Audience Orientation and
Transaction were negative predictors of Dislike Writing, while
Transmission was a positive predictor. Transmission was the only
significant predictor of Apprehension About Grammar. Recursive
Process did not attain statistical significance in any of these
regression analyses.
In the regression equation testing whether Audience Orientation, Recursive Process, and Transaction positively predict writing
performance (Hypothesis 1b), and Transmission negatively predicts
writing performance (Hypothesis 1e), beliefs about writing did
predict the students’ writing grades (R ¼ .29, F(4, 733) ¼ 16.83,
R2 ¼ .08, p < .001). Audience Orientation was a positive predictor,
and Transaction and Transmission were negative predictors. Again,
Recursive Process was not a significant predictor.
A simultaneous regression confirmed Hypothesis 1g, which
predicted that all three writing self-efficacy subscales would positively predict writing performance (R ¼ .24, F(3, 734) ¼ 14.31,
R2 ¼ .06, p < .001). As expected, Self-Efficacy for Mechanical writing
skills positively predicted the writing grades, but, contrary to our
hypotheses, the remaining writing self-efficacy scales did not.
Hypothesis 1h predicted that the writing self-efficacy subscales
would each negatively predict Dislike Writing and Apprehension
About Grammar, and positively predict Enjoy Writing. Writing SelfEfficacy predicted the following:
Enjoy Writing: R ¼ .61, F(3, 734) ¼ 145.37, R2 ¼ .37, p < .001.
Dislike Writing: R ¼ .61, F(3, 734) ¼ 148.36, R2 ¼ .38, p < .001.
Apprehension About Grammar: R ¼ .45, F(3, 734) ¼ 63.35,
R2 ¼ .21, p < .001.
Self-Efficacy for Substantive writing skills and Writing SelfRegulation positively predicted Enjoy Writing and negatively predicted Dislike Writing, as hypothesized, but did not predict
Apprehension About Grammar. As expected, Self-Efficacy for Mechanical writing skills was a negative predictor of Dislike Writing, a
powerful predictor of Apprehension About Grammar, and a predictor of Enjoy Writing.
Hypothesis 1i predicted that Dislike Writing and Apprehension About Grammar would negatively predict the students’
writing grades while Enjoy Writing would be a positive predictor.
The three subscales related to writing apprehension did predict
writing performance (R ¼ .27, F(3, 734) ¼ 18.66, R2 ¼ .07,
p < .001). Apprehension About Grammar negatively predicted
writing performance, as expected, while Enjoy Writing and
Dislike Writing did not contribute significantly to the regression
equation.
3.4. Predicting writing performance
In accord with hierarchical regression protocol (Hair et al.,
1998), we used the research literature to guide the order of entry of the independent variables in the hierarchical regression
analysis (beliefs about writing, writing self-efficacy, and writing
apprehension) for predicting the criterion variable, writing performance (see Table 6). As discussed above, the order in which we
8
J. Sanders-Reio et al. / Learning and Instruction 33 (2014) 1e11
Table 5
Summary of simultaneous regression analyses predicting writing self-efficacy,
writing apprehension, and writing performance.
Variable
B
Beliefs About Writing Predicting
Writing Self-Efficacy (Substantive)
Audience
12.51
Recursive
3.59
Transaction
24.47
Transmission
.18.54
Writing Self-Efficacy (Self-Regulatory)
Audience
7.08
Recursive
3.17
Transaction
18.81
Transmission
.12.40
Writing Self-Efficacy (Mechanical)
Audience
5.83
Recursive
2.32
Transaction
9.43
Transmission
.9.38
Writing Apprehension (Enjoy)
Audience
.11
Recursive
.02
Transaction
1.03
Transmission
.31
Writing Apprehension (Dislike)
Audience
.12
Recursive
.02
Transaction
.72
Transmission
.62
Writing Apprehension (Grammar)
Audience
.01
Recursive
.05
Transaction
.01
Transmission
.18
Writing Performance
Audience
.09
Recursive
.06
Transaction
.07
Transmission
.19
SE B
b
Table 6
Summary hierarchical regression analysis predicting writing performance from
beliefs about writing, writing self-efficacy, and writing apprehension.
Writing Grade
Variable
2.23
5.05
3.96
4.60
.21***
.03
.23***
.14***
1.73
3.93
3.08
3.58
.16***
.03
.23***
.12**
1.24
2.73
2.12
2.49
.19***
.03
.17***
.14***
.05
.12
.09
.11
.08*
.01
.41***
.11**
.06
.12
.10
.11
.08*
.01
.28***
.20***
.02
.05
.04
.04
.02
.04
.01
.16***
.02
.04
.03
.04
.22***
.06
.09***
.20***
.00
.00
.00
.23***
.41***
.01
.00
.00
.00
.23***
.34***
.10*
Writing Self-Efficacy Predicting
Writing Apprehension (Enjoy)
Substantive
.01
Self-Regulatory
.01
Mechanical
.00
Writing Apprehension (Dislike)
Substantive
.01
Self-Regulatory
.01
Mechanical
.01
Writing Apprehension (Grammar)
Substantive
.00
Self-Regulatory
.00
Mechanical
.01
Predicting Writing Performance
Substantive
.00
Self-Regulatory
.00
Mechanical
.01
.00
.00
.00
.18**
.01
.57***
.00
.00
.01
.07
.05
.23***
Writing Apprehension Predicting
Writing Performance
Enjoy
.01
Dislike
.02
Grammar
.19
.02
.02
.03
.02
.07
.23***
Note. N ¼ 738. *p < .05; **p < .01; ***p < .001.
entered the variables was guided first by researchers (e.g., White
& Bruning, 2005) who suggested that students’ beliefs about
writing influence their writing process, which includes their selection of writing strategies. Using sound writing strategies in
turn creates mastery experiences that, according to Bandura
(1997), enhance self-efficacy. Thus, we entered the beliefs about
writing in the first block before the writing self-efficacy beliefs in
the second block. With respect to apprehension, the Pajares
group’s path-analytic research (e.g., Pajares & Johnson, 1994)
Step 1: Beliefs About Writing
Knowledge Telling
Transmission
Knowledge Transforming
Transaction
Recursive Process
Knowledge Crafting
Audience Orientation
Block
Step 2: Writing Self-Efficacy
Substantive
Self-Regulatory
Mechanical
Block
Step 3: Writing Apprehension
Enjoy Writing
Dislike Writing
Apprehension About Grammar
Block
Total R2
b
DR2
.15***
.11***
.07*
.19***
.08***
.04
.02
.09*
.03***
.02
.01
.19***
.03***
.15***
Note. N ¼ 738. *p < .05; **p < .01; ***p < .001.
indicated that writing self-efficacy can nullify writing apprehension. Therefore, we entered the writing apprehension scores
in the third block after the writing self-efficacy beliefs.
Together, the three blocks of variables explained 15.0% of the
variance in writing performance [R ¼ .39, F(10, 737) ¼ 12.30,
p < .001]. As a group, the beliefs about writing in the first block
explained 8.4% of the variance in writing grades (R ¼ .29, F(4,
733) ¼ 16.83, p < .001). Each of the four beliefs independently
and significantly predicted writing performance. Audience
Orientation, the most powerful predictor, and Recursive Process
positively predicted performance, while Transmission and
Transaction were negative predictors (Hypotheses 2a, b). The
three writing self-efficacy scales, entered in the second block,
explained an additional 3.3% of the variance (R ¼ .34, FD (3,
730) ¼ 9.23, p < .001, Hypothesis 2c); Self-Efficacy for Mechanical
writing skills was the only significant predictor (positive). Finally,
the three writing apprehension variables entered in the third
block increased the variance explained by a final 3.3% (R ¼ .39, FD
(3, 727) ¼ 7.69, p < .001, Hypothesis 2d). As in previous studies
(e.g., Pajares & Valiante, 1997), the original WAT (the Dislike
Writing and Enjoy Writing subscales here) did not independently
predict performance. However, Apprehension About Grammar,
the new subscale, negatively predicted incremental variance in
performance above and beyond the effects of both beliefs about
writing and writing self-efficacy (b ¼ .19, p < .001).
Inasmuch as beliefs about writing independently explained
variance in writing performance, Hypotheses 2a and 2b were
supported. Likewise, because both writing self-efficacy and
apprehension uniquely predicted writing performance beyond
beliefs about writing, Hypotheses 2c and 2d were supported as
well. Overall, beliefs about writing explained the most variance in
writing performance, with Audience Orientation being the most
powerful predictor.
3.4.1. Predictive validity
To demonstrate further that beliefs about writing, a form of
domain-related beliefs, can predict writing performance beyond
the more researched writing self-efficacy and writing apprehension
J. Sanders-Reio et al. / Learning and Instruction 33 (2014) 1e11
variables included in this study (Hypothesis 2e), we conducted
another hierarchical regression analysis. In this case we entered the
two more extensively studied sets of variables before the beliefs
about writing, and entered self-efficacy first, followed by writing
apprehension, and then beliefs about writing. The findings provided evidence of the predictive validity of the model presented
here as beliefs about writing explained an additional 5.8% of the
variance in writing performance after controlling for writing selfefficacy (5.5%) and writing apprehension (3.2%). Notably, beliefs
about writing explained the most variance in the regression
equation even when entered in the third and final step of the
equation. Finally, we ran the entire regression model (primary)
using only the original items of Zimmerman and Bandura’s Writing
Self-Regulatory Efficacy Scale (1994); this measure explained less
variance (1.3%) than the more comprehensive Writing Self-Efficacy
Index (3.3%).
4. Discussion
This study investigated whether beliefs about writing, including
two new beliefs, Audience Orientation and Recursive Process, relate
to writing performance and two of its established correlates,
writing self-efficacy and apprehension. We hypothesized that the
beliefs about writing would predict variance in writing grades over
and above the effects of writing self-efficacy and writing apprehension. We further hypothesized that a new subscale assessing
apprehension about making grammatical and other mechanical
writing errors would strengthen the traditional measure of writing
apprehension (Daly & Miller, 1975), which assesses anxiety about
showing others one’s written work and having it critiqued.
4.1. Beliefs about writing
The participants’ beliefs about writing did relate to their writing
self-efficacy (Hypotheses 1a, 1d), apprehension (Hypotheses 1c, 1f),
and performance (Hypotheses 1b, 1e), and they did predict unique
variance in the students’ grades for their written work (Hypothesis
2a, 2b). We have used these results to develop the following profiles
of the four beliefs about writing we examined.
Audience Orientation, the most adaptive belief about writing
studied here, is associated with expert writing practice (Kellogg,
2008). This belief reflects a concern for the needs and interests of
one’s readers. Items addressing classic characteristics of good
writing, such as development and clarity, loaded on this belief. This
makes sense because clarity and development imply an audience
for whom the text is clear, understandable, and for whom concepts
and information are explained with appropriate detail. The adaptiveness of this belief confirms Ong’s (1975) view that being able to
interpret one’s text from a reader’s point of view “is one of the
things that separates the beginning graduate student or even the
brilliant undergraduate from the mature scholar” (p. 19).
Recursive Process, a belief related to stage two of Kellogg’s
(2008) model, sees each aspect of the writing process as a time to
rethink, revise, and revision. We suspect that this belief would be
even more adaptive in the context of longer assignments and
writing that is held to particularly high standards, such as dissertations and articles written for publication.
Transaction, which maintains that writing involves cognitive
and emotional engagement, positively predicted writing performance in White and Bruning’s (2005) study, but it was a negative
predictor here. However, Transaction was a strong positive correlate of Enjoy Writing and a strong and negative correlate of both
types of writing apprehension. The enjoyment of writing that
Transaction seems to engender may keep writers working productively during writing instruction and while crafting papers that
9
entail many iterations. Thus, we expect that this belief, too, may be
more adaptive in the context of more complex assignments written
for higher standards.
Transmission, which endorses the practice of relying on authorities and their published arguments and quotes, was maladaptive as those who embraced it received lower grades on their
papers and were less self-efficacious and more apprehensive about
writing, particularly with respect to grammar and writing mechanics. This belief could easily foster a mechanical and/or safe, selfprotective, and detached approach to writing that entails stringing
other writers’ quotes together, plugging new text into established
formats, or simply using new words to convey established lines of
argument laid out by authorities in encyclopedias and textbooks.
We acknowledge that the correlations between the beliefs about
writing and writing performance are modest. However, we contend
that these relations are meaningful. Even strong adherence to a
belief about writing does not imply the skill or the will to act on
that belief. As Kellogg (2008) states, appreciating that one has an
audience does not mean that one is able to see one’s own text from
the perspective of that audience or that one has the skills, motivation, or executive control to adapt one’s message to that
audience.
4.2. Writing self-efficacy and writing apprehension
As expected, participants with high writing self-efficacy had low
writing apprehension and enjoyed writing more, and those who
had low writing self-efficacy enjoyed writing less and were more
apprehensive about writing (Hypothesis 1h). Participants who
were more apprehensive received lower grades on their papers,
while those with high writing self-efficacy received higher grades
on their papers, but the magnitude of the association was modest
(Hypothesis 2c). In several of the Pajares group’s studies, writing
self-efficacy nullified the influence of writing apprehension (e.g.,
Pajares & Valiante, 1997). That occurred here as well, but only with
respect to the traditional measure of apprehension that Pajares
used. Apprehension About Grammar, a new subscale unavailable to
Pajares, accounted for significant, unique, negative variance in the
regression equation.
The disparity between the magnitude of these self-efficacy results and those in the literature may stem from the differences
between the studies. For example, the participants in this study
were undergraduates enrolled in an upper-level course, while most
previous research involved younger students. This study also used
the grade students received on a take-home assignment to measure
writing performance. Grading standards may have been more
demanding than they would be for short assignments written on
demand in 15e20 min. In addition, the wide disparity between the
grade the participants predicted they would earn and the grade
they actually received (r ¼ .13) indicates that their self-efficacy
judgments may not have been well calibrated. Many reported
that they had commonly received high grades (usually A’s) on
written assignments at other schools, but their writing performance here was not in line with these reports.
4.3. Limitations
Although we investigated a number of beliefs about writing,
others are possible. Beliefs about writing are culturally constructed
and disseminated, and may relate to the pedagogy of teaching
writing (Silva & Nicholls, 1993), previous writing experiences
(White & Bruning, 2005), genre, and context. Thus, students’ beliefs
about writing may change as they work with new genres or media,
or as teachers develop new methods of writing instruction. In
addition, the instruments developed for this study may need
10
J. Sanders-Reio et al. / Learning and Instruction 33 (2014) 1e11
further refinement (e.g., replication with more diverse populations
in terms of academic discipline and writing expertise). Further, the
study used a correlational design, which does not allow the
exploration of causal relations. The participants were primarily
Hispanic females, limiting the generalizability of the results to
other groups. Finally, writing performance, the dependent variable,
was operationalized as the grade participants received on only one
paper, which does not reflect the variance in students’ writing
performance (Hayes, Hatch, & Silk, 2000).
4.4. Future directions
Further work with these variables is needed with other writing
assignments and contexts, and with participants who are more
varied and balanced with respect to gender, race, ethnicity, and
writing expertise. Research is needed to identify the mechanisms by
which beliefs about writing may affect writing performance. It may
be the case, as this study indicates, that there is an affective link, as
certain beliefs about writing seem to foster writing apprehension or,
by contrast, increase the extent to which students like to write. It is
also possible, as many of the researchers cited here theorized, that
there is also a cognitive link mediated by the writer’s choice of
strategies or a student’s openness to instruction in specific strategies. For example, students might not attend to instruction in
revision if they believe that good writers do not revise but write it
right on their first attempt. If beliefs about writing do relate to
openness to instruction, it will be valuable to determine whether
these beliefs are amenable to change and whether certain changes
in a student’s beliefs foreshadow improvements in attitudes about
writing and writing performance itself.
These results also support the possibility that beliefs about
writing could be a worthwhile new leverage point for teaching
students to write. It may be useful to modify writing instruction to
emphasize the mindsets and approaches associated with adaptive
beliefs and minimize those related to maladaptive and ineffective
beliefs. For example, assignments can be structured to encourage
students to have a stronger sense of audience. Additionally,
teachers can assign fewer papers and more revision so that onedraft writing becomes the exception and revision cycles the
norm. Strategies, such as taking notes from outside texts, selecting
and incorporating quotations, and varying and increasing the sophistication of one’s vocabulary, can be presented so that they
remain flexible and do not deteriorate into mechanical cutting and
pasting. Finally, the strong negative relation of Apprehension About
Grammar to writing performance indicates that we may need to be
less indignant about mechanical errors and develop approaches to
teaching grammar and correctness that are less likely to produce
counterproductive levels of anxiety.
4.5. Conclusions
These results support Bandura’s (1997) views about the importance of beliefs in that a relatively new type of belief, beliefs about a
domain, accounted for significant, unique variance in performance.
The fact that these beliefs predict performance in conjunction with
self-efficacy beliefs supports the possibility that constellations of
beliefs (e.g., domain-specific beliefs coupled with epistemic beliefs
[Mateos et al., 2010] or self-efficacy beliefs) may affect performance
in tandem.
Acknowledgements
We would like to acknowledge Lucia Mason, as well as the
three anonymous reviewers who so thoroughly and productively
reviewed this manuscript.
References
Bandura, A. (1989). Human agency in social cognitive theory. American Psychologist,
44, 1175e1184. http://dx.doi.org/10.1037/0003-066X.44.9.1175.
Bandura, A. (1997). Self-efficacy: The exercise of control. New York: W. H. Freeman.
Beach, R., & Friedrich, T. (2006). Response to writing. In C. A. MacArthur, S. Graham,
& J. Fitzgerald (Eds.), Handbook of writing research (pp. 222e234). New York:
Guilford.
Bereiter, C., & Scardamalia, M. (1987). The psychology of written composition. Hillsdale, NJ: Erlbaum.
Bline, D., Lowe, D. R., Meixner, W. F., Nouri, H., & Pearce, K. (2001). A research note
on the dimensionality of Daly and Miller’s Writing Apprehension Scale. Written
Communication, 18(1), 61e79. http://dx.doi.org/10.1177/0741088301018001003.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.).
Hillsdale, NJ: Erlbaum.
Coker, D., & Lewis, W. E. (2008). Beyond writing next: a discussion of
writing research and instructional uncertainty. Harvard Educational Review,
78(1), 231e251.
Daly, J. A., & Miller, M. D. (1975). The empirical development of an instrument
to measure writing apprehension. Research in the Teaching of English, 9, 242e
249.
Gay, L. R. (1992). Educational research: Competencies for analysis and application (4th
ed.). New York: Merrill.
Graham, S., Schwartz, S. S., & MacArthur, C. A. (1993). Knowledge of writing and the
composing process, attitude toward writing, and self-efficacy for students with
and without learning disabilities. Journal of Learning Disabilities, 26, 237e249.
http://dx.doi.org/10.1177/002221949302600404.
Hair, J. F., Jr., Anderson, R. E., Tatham, R. L., & Black, W. C. (1998). Multivariate data
analysis (5th ed.). Upper Saddle River, NJ: Prentice Hall.
Hayes, J. R., & Flower, L. S. (1980). Identifying the organization of writing processes.
In L. W. Gregg, & E. R. Steinbert (Eds.), Cognitive processes in writing (pp. 3e30).
Hillsdale, NJ: Lawrence Erlbaum Associates.
Hayes, J. R., Hatch, J., & Silk, C. M. (2000). Does holistic assessment predict writing
performance? Estimating the consistency of student performance on holistically scored writing assessments. Written Communication, 17(1), 3e26. http://
dx.doi.org/10.1177/0741088300017001001.
Hillocks, G., Jr. (2008). Writing in secondary schools. In C. Bazerman (Ed.), Handbook
of research on writing (pp. 311e329). New York: Erlbaum.
Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes to covariance structure
analysis: conventional criteria versus new alternatives. Structural Equation
Modeling, 6, 1e55.
Kellogg, R. T. (2008). Training writing skills: a cognitive developmental perspective.
Journal of Writing Research, 1(1), 1e26.
Kline, R. B. (1998). Principles and practice of structural equation modeling. New York:
Guilford Press.
Lavelle, E. (1993). Development and validation of an inventory to assess processes
in college composition. British Journal of Educational Psychology, 63, 489e499.
http://dx.doi.org/10.1111/j.2044-8279.1993.tb01073.x.
Mateos, M., Cuevas, I., Martin, E., Martin, A., Echeita, G., & Luna, M. (2010). Reading
to write and argumentation: the role of epistemological, reading and writing
beliefs. Journal of Research in Reading, 34, 281e297. http://dx.doi.org/10.1111/
j.1467-9817.2010.01437.x.
McCarthy, P., Meier, S., & Rinderer, R. (1985). Self-efficacy and writing: a
different view of self-evaluation. College Composition and Communication,
36, 465e471.
Meier, S., McCarthy, P. R., & Schmeck, R. R. (1984). Validity of self-efficacy as a
predictor of writing performance. Cognitive Therapy and Research, 8(2), 107e
120. http://dx.doi.org/10.1007/BF01173038.
Miller, C. R., & Charney, D. (2008). Persuasion, audience, and argument. In
C. Bazerman (Ed.), Handbook of research on writing (pp. 583e598). New York:
Erlbaum.
Murphy, S., & Yancy, K. B. (2008). Construct and consequence: validity in writing
assessment. In C. Bazerman (Ed.), Handbook of research on writing (pp. 365e
385). New York: Erlbaum.
Murray, D. M. (1991). The craft of revision. Fort Worth, TX: Harcourt Brace
Jovanovich.
National Council of Teachers of English, James R. Squire Office of Policy Research.
(2008). Writing now: A policy research brief produced by the National Council of
Teachers of English. Retrieved January 15, 2009, from http://www.ncte.org/
library/NCTEFiles/Resources/PolicyResearch/WrtgResearchBrief.pdf.
Nietfeld, J. L., & Schraw, G. (2002). The effect of knowledge and strategy training on
monitoring accuracy. The Journal of Educational Research, 95(3), 131e142. http://
dx.doi.org/10.1080/00220670209596583.
Nimon, K., Zigarmi, D., Houson, D., Witt, D., & Diehl, J. (2011). The work cognition
inventory: initial evidence of construct validity. Human Resource Development
Quarterly, 22, 7e33.
Nunnally, J. C. (1967). Psychometric theory. New York: McGraw Hill.
Ong, W. J. (1975). The writer’s audience is always a fiction. PMLA, 91, 9e21.
Pajares, F. (1997). Current directions in self-efficacy research. In M. Maehr, &
P. R. Pintrich (Eds.), Advances in motivation and achievement (pp. 1e49).
Greenwich, CT: JAI Press.
Pajares, F., & Johnson, M. J. (1994). Confidence and competence in writing: the role
of self-efficacy, outcome expectancy, and apprehension. Research in the Teaching
of English, 28(3), 313e331.
J. Sanders-Reio et al. / Learning and Instruction 33 (2014) 1e11
Pajares, F., & Valiante, G. (1997). Influence of self-efficacy on elementary students’
writing. The Journal of Educational Research, 90(6), 353e360. http://dx.doi.org/
10.1080/00220671.1997.10544593.
Pajares, F., & Valiante, G. (1999). Grade level and gender differences in the writing
self-beliefs of middle school students. Contemporary Educational Psychology, 24,
390e405. http://dx.doi.org/10.1006/ceps.1998.0995.
Palmquist, M., & Young, R. (1992). The notion of giftedness and student expectations
about writing. Written Communication, 9, 137e168. http://dx.doi.org/10.1177/
0741088392009001004.
Sanders-Reio, J. (2010). Investigation of the relations between domain-specific beliefs
about writing, writing self-efficacy, writing apprehension, and writing performance
in undergraduates. Unpublished doctoral dissertation. College Park: University
of Maryland.
Scheuer, N., de la Cruz, M., Pozo, J. I., Huarte, M. F., & Sola, G. (2006). The mind is not
a black box: children’s ideas about the writing process. Learning and Instruction,
16, 72e85. http://dx.doi.org/10.1016/j.learninstruc.2005.12.007.
Schommer-Aikins, M. (2004). Explaining the epistemological belief system:
introducing the embedded systemic model and coordinated research
approach. Educational Psychologist, 39, 19e29. http://dx.doi.org/10.1207/
s15326985ep3901_3.
Schraw, G., & Bruning, R. (1996). Readers’ implicit models of reading. Reading
Research Quarterly, 31, 290e305. http://dx.doi.org/10.1598/RRQ.31.3.4.
11
Schraw, G., & Bruning, R. (1999). How implicit models of reading affect motivation
to read and reading engagement. Scientific Studies of Reading, 3(3), 281e302.
Schunk, D. H., & Swartz, C. W. (1993). Goals and progress feedback: effects on selfefficacy and writing achievement. Contemporary Educational Psychology, 18,
337e354. http://dx.doi.org/10.1006/ceps.1993.1024.
Silva, T., & Nicholls, J. G. (1993). College students as writing theorists: goals and
beliefs about the causes of success. Contemporary Educational Psychology, 18,
281e293. http://dx.doi.org/10.1006/ceps.1993.1021.
Smith, M. W., Cheville, J., & Hillocks, G., Jr. (2006). “I guess I’d better watch my
English”: grammars and the teaching of the English language arts. In
C. A. MacArthur, S. Graham, & J. Fitzgerald (Eds.), Handbook of writing research
(pp. 263e274). New York: Guilford.
White, M. J., & Bruning, R. (2005). Implicit writing beliefs and their relation to
writing quality. Contemporary Educational Psychology, 30, 166e189. http://
dx.doi.org/10.1016/j.cedpsych.2004.07.002.
Worthington, R. I., & Whittaker, T. A. (2006). Scale development research: a content
analysis and recommendations for best practices. The Counseling Psychologist,
34, 806e838.
Zimmerman, B. J., & Bandura, A. (1994). Impact of self-regulatory influences on
writing course attainment. American Educational Research Journal, 31, 845e862.
http://dx.doi.org/10.3102/00028312031004845.
Zinsser, W. (1976). On writing well. New York: HarperCollins.