Page 1
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES
Students’ Voices: The Relationship between Attitudes and Writing Outcomes
for Fourth and Fifth Graders
Andrea Winokur Kotula a
Terrence Tivnan b
Cynthia Mata Aguilar a
a Education Development Center, Inc.
b Harvard University Graduate School of Education
Page 2
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES
Suggested citation:
Kotula, A. W., Tivnan, T., & Aguilar, C. M. (2014). Students’ Voices: The Relationship Between
Attitudes and Writing Outcomes for Fourth and Fifth Graders. Waltham, MA:
Education Development Center, Inc.
Copyright © 2014 by Education Development Center, Inc.
Education Development Center, Inc. is a global nonprofit organization that creates learning
opportunities for people around the world, empowering them to pursue healthier, more
productive lives. For more information, visit edc.org.
Page 3
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES
Abstract
As part of a four-year research project to study a writing curriculum for fourth- and fifth-grade
students, the authors measured students’ writing ability and their attitudes about writing to
determine the relationship between the two variables. A principal component analysis of the 18-
item attitude survey indicated three composites with eigenvalues above 1.0, and these showed
reasonable levels of internal consistency reliability: (1) the Perceived Value of Writing; (2) Self-
Rating as a Writer; and (3) Writing Behaviors. A small but consistent relationship was found
between each component and writing outcomes, with a slightly stronger relationship at the end of
the year, especially for the Self-Rating as a Writer component. This adds to the convergence of
evidence about the relationship between attitudes about writing and writing ability despite the
different ways in which researchers define or measure attitudes. Moreover, this study confirms
this relationship with students who came from primarily high-poverty homes and were on
average poor writers. About half of the students were Hispanic, with 32% either receiving
Limited English Proficiency services at the time of the study or up to two years prior to it.
Because the authors collected writing and attitude data at both the beginning and end of the year,
they were able to determine that student attitudes about writing did not get stronger on average
after a year of writing instruction, even though the correlations between writing ability and
attitudes tended to be a little higher at the end of the year. Girls had significantly more positive
attitudes than boys on all three components in both fall and spring of both grades. They were also
better writers.
Page 4
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES
Students’ Voices: The Relationship between Attitudes and Writing Outcomes for
Fourth and Fifth Graders
1. Introduction
To compete in a global economy, students must possess the skills to think critically and
creatively, solve problems, collaborate, and especially, communicate (Partnership for 21st
Century Skills, 2011). A report from the College Board, the National Writing Project, and Phi
Delta Kappa International states, “Writing has never been more important than in this digital age.
It is almost inconceivable to achieve academic success without good writing skills” (College
Board Advocacy and Policy Center, 2010, p. 2). Moreover, writing is an essential
communication skill in the business sector as well as in the world of education, with many
arguing that poor writing has repercussions in the workplace (cf. Levy & Murnane, 2004;
Murnane & Levy, 1996; National Governors Association Center for Best Practices, 2005).
However, today’s students find writing very challenging. Across all grades, only about
one-quarter of the students in our nation’s schools are proficient in writing. On the most recent
national assessment of student writing in 2011 (National Center for Education Statistics [NCES],
2012), only 27% of students at both grades 8 and 12 performed at or above the proficient level
(“solid academic performance”) in writing. Fourth-grade student writing was most recently
assessed in 2002, when only 28% of these students obtained scores at or above the proficient
level. Moreover, the writing proficiency of students among certain racial and ethnic groups
suffers even more. For example, 34% of White eighth-grade students demonstrated proficiency
in writing compared to 14% of Hispanic and 11% of Black students, and 35% of White twelfth-
grade students demonstrated proficiency as compared to 11% of Hispanic and 9% of Black
Page 5
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 2
students (NCES, 2012). Similarly, 34% of White fourth-grade students demonstrated proficiency
as compared to only 17% of Hispanic and 14% of Black students (Persky, Daane, & Jin, 2003).
1.1. Motivation and student learning
In response to this national student writing proficiency crisis, we engaged in a four-year
research project funded by the US Department of Education, Institute of Education Sciences, to
study a writing curriculum for fourth- and fifth-grade classrooms in six school districts. As part
of that study, we examined the contextual factors (district, school, classroom, or student) that
might influence students’ writing outcomes, and one of the student-level variables we studied
was students’ attitudes about writing. A large literature exists on student attitudes, one
component of motivation.
Research indicates a relationship between motivation and student learning across many
domains (Alexander, Graham, & Harris, 1998; Brophy, 2010; Murphy & Alexander, 2000;
Schunk, Meece, & Pintrich, 2014; Wigfield & Eccles, 1992; Zimmerman, 2011). Moreover,
there is evidence that links motivation with writing ability in particular; that is, students with
higher scores on measures of motivation tend to be better writers (Albin, Benton, & Khramtsova,
1996; Graham, Berninger, & Fan, 2007; Graham, Schwartz, & MacArthur, 1993; Hidi,
Berndorff, & Ainley, 2002; Knudson, 1995; Pajares & Johnson, 1996; Pajares & Valiante, 1999,
2001; Shell, Colvin, & Bruning, 1995; Troia, Harbaugh, Shankland, Wolbers, & Lawrence,
2013). Studies have also demonstrated that student attitudes about their writing ability affect how
often and how well they write (Clark, 2012; Daly & Shamo, 1978; Faigley, Daly, & Witte, 1981;
Nelson, 2007).
Page 6
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 3
1.2. Student attitudes about writing
An attitude is often defined as “a psychological tendency that is expressed by evaluating
a particular entity with some degree of favor or disfavor” (Eagly & Chaiken, 2007, p. 598). The
attitude construct has traditionally been assumed to comprise three parts: cognition, affect, and
behavior. “In this view, the attitude is an unobservable psychological construct which can
manifest itself in relevant beliefs, feelings, and behavioral components” (Fazio & Olson, 2003, p.
139). More recently, attitudes have been thought of as evaluative judgments (Fazio & Olson,
2003; Schwarz, 2007), which are not necessarily permanent traits. However, no firm evidence
exists to determine if people have attitudes per se or if they create them as needed, and this may
depend on the nature of the attitudes, their accessibility in one’s memory, and the context in
which they are measured. Many facets of student attitudes have been studied, including interest,
goal orientation, task value, and perceived causes of failure (Troia et al., 2013). Of these, self-
efficacy —how capable one feels about his or her writing—has been studied the most. In his
review of the literature, Pajares (2003) found a consistent relationship between student self-
efficacy beliefs and writing performance, with effect sizes of .19 to .40 even when previous
writing performance was held constant.
Writing self-concept is another attitude sometimes discussed in the literature, although
less frequently than self-efficacy. An example of a statement that focuses on self-concept might
be “Writing makes me feel inadequate,” while an example of a question that focuses on self-
efficacy might be “How sure are you that you can correctly spell all words in a one-page story or
composition?” (Pajares, 2003, p. 147).
Self-efficacy is associated with other attitude variables, such as how much one values
writing, although in some studies that hold self-efficacy beliefs constant, the perceived value of
Page 7
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 4
writing variable does not make an independent contribution. It may be that one’s self-efficacy
beliefs influence the value placed on writing (Pajares, 2003; Pajares, Miller, & Johnson, 1999).
In our study, we wanted to measure students’ writing ability, determine their self-efficacy as
writers, and assess how much they valued writing.
1.3. Measuring student attitudes about writing
Many researchers have developed measures to assess student attitudes about writing. In
the last two decades alone, such surveys were developed, for example, by Knudson (1991, 1992,
1993a, 1993b, 1995); Pajares et al. (1999); Kear, Coffman, McKenna, and Ambrosio (2000); and
Graham, Berninger, and Fan (2007). The features of these surveys will be explained in more
detail in Section 1.5.
When planning our attitude survey, we regarded attitudes about writing as falling along a
continuum from positive to negative in the same way that other researchers have (Graham et al.,
2007; Kear et al., 2000; Knudson, 1995). We considered attitudes as comprising different
components or beliefs: a sense of how capable one is as a writer (self-efficacy), one’s self-
concept as a writer, and how much one values writing. Our survey and its results add to the
existing research base by including a large number of Hispanic students and students from low-
income families. We also collected data about attitudes and writing outcomes at both the
beginning and end of the year and followed fourth grade students into fifth grade.
1.4. Results of previous surveys
Previous surveys have varied by grades and populations of students and the items
included in the survey. The findings from these studies tend to converge, although there were
some differences in how attitudes were defined, the way they were measured, and the
demographics of the students. Knudson studied the relationship between attitudes and writing
Page 8
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 5
across several studies in grades 1–3, 4–8, and 9–12 (1991, 1992, 1993a, 1993b). She found that
girls had more positive attitudes about writing than boys, attitudes became less positive as
students got older, and ethnicity did not appear to have any effect on the results. In 1995,
Knudson undertook a new study with 430 students in grades 1–6 who were administered a
survey and a writing prompt. These students responded to 19 items along a 5-point scale that
ranged from almost always to almost never. The author identified three factors for the grade 4–6
survey: (1) “Positive Attitude Toward Writing,” (2) “Letter/Note Writing,” and (3) “Positive
View of Self as Writer” (p. 92). Some of the 19 items focused on self-efficacy, some on self-
concept, and some on how the writer felt when writing. Knudson found a positive relationship
between students’ attitudes and writing even after controlling for grade.
Pajares et al. (1999) conducted a study with 363 third to fifth graders that compared the
relationship between writing ability and both self-concept and self-efficacy by using separate
surveys. They also included a survey to measure students’ perceived usefulness of writing. In
addition, the authors asked students to compare their writing ability to other girls and boys in
their school.
Pajares et al.’s (1999) skills-based self-efficacy scale comprised nine items rated on a 0–
100 scale (from no chance to completely certain); the self-concept scale included six items rated
along a 6-point scale (definitely false to definitely true); and the perceived usefulness of writing
scale was made up of nine items that were rated along a 6-point scale (extremely unimportant to
extremely important). A multiple regression analysis indicated that self-efficacy predicted
writing performance. Perceived usefulness did not significantly predict writing performance. The
authors found that girls were better writers than boys and had higher self-concept—but not
higher writing self-efficacy.
Page 9
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 6
Kear et al. (2000) developed a 28-item attitude survey to be used across grades 1–12. All
items begin with the stem, “How would you feel,” and most items are specific to writing tasks,
subjects, and genres, so they are aimed at assessing students’ feelings of self-efficacy. Each item
was rated along a 4-point scale using cartoon-like Garfield™ characters to represent being very
happy to very upset.
More recently, Graham et al. (2007) focused their survey on measuring children’s
attitudes about writing for different purposes and in different locations. Participants were 128
first-grade and 113 third-grade children. The questionnaire was made up of only seven items,
each of which began with the stem “How do you feel” and utilized the same Garfield™ pictures
along a 4-point scale from very happy to very unhappy. A factor analysis yielded a single factor
that accounted for 52% of the variance. The authors found that writing attitude predicted writing
achievement and that there was a statistically significant direct path from attitude to
achievement—which was measured before the survey was administered. However, they did not
find a significant difference in writing attitude between the grade-1 and grade-3 students. Girls
had more positive attitudes than boys, but their writing scores were not significantly higher.
1.5. Comparison of previous surveys
While all these surveys measure student attitudes toward writing, there are important
differences in the approaches that were used and in some of the findings from these studies. The
items on Knudson’s (1995) surveys assess self-efficacy, self-concept, and feelings. Pajares et al.
(1999) measured self-efficacy and self-concept with separate surveys. Graham et al.’s (2007)
survey utilized a narrower interpretation of attitudes—how students felt when they were writing
for different purposes and in different places. The authors state that “it was most closely, but not
perfectly, aligned with a mood” (p. 518), and the purposes for writing were general, e.g., “for
Page 10
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 7
fun,” “during free time.” They found that the seven items appeared to load onto one main factor.
Graham et al. tested three models to examine the relationship between writing attitude and
writing ability—a direct path from attitude to writing performance, from writing performance to
attitude, and a bi-directional path between performance and attitude. The direct path from
attitude to performance best described the data. Kear et al. (2000) created a longer survey that
also focused on how students felt when they were writing, but the purposes were more specific
than Graham et al.’s and addressed writing tasks such as “writing a letter to the author of a book
you read,” or writing about something you did in science.” However, the authors did not include
data linking the survey to student writing ability.
1.6. The development of our survey
We looked closely at other researchers’ surveys and decided that we wanted to measure
students’ self-efficacy and the value they placed on writing rather than specific writing tasks or
locations. Although we did not ask students how they felt about writing per se, their attitudes
about it can be inferred from their responses to many of the items in the Value of Writing
component. Because it was important that our survey be easily administered to groups of
students in 10 minutes or less, length was considered. We also wanted to extend the work of
other researchers by comparing results at different data collection points. We believe that
attitudes are potentially dynamic and could change over time; although other researchers have
generally used a more “snapshot” approach and collected writing and attitude data at one time,
we studied students’ attitudes and writing outcomes at the beginning as well as at the end of each
school year. Most of the fifth graders in our study also participated in fourth grade.
Page 11
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 8
2. Method
2.1.Participants
The data collection for the study took place over two years. In Year 1, we studied the
students in fourth-grade classrooms. In Year 2, we studied a second cohort of fourth-grade
classrooms and followed the first cohort of students into their fifth-grade classrooms. We
collected surveys from all student participants at the beginning and end of each school year.
Students came from 367 classrooms (270 fourth-grade classrooms and 97 fifth-grade classrooms)
in 56 schools in six school districts in Massachusetts. In Years 1 and 2 of the implementation
(2010-2012), we collected 8, 234 surveys from students in fourth grade and 2,806 surveys from
students in fifth grade (Year 2 only).
Teacher experience varied from beginning to veteran teachers with over 20 years of
experience within and across schools and districts, and participating teachers taught between one
and five writing classes in their schools. Districts ranged in size from 13,373 students to 4,496
students. All are Title I districts.
2.1.1. Student demographics
There was a high percentage of low-income students in the study (those eligible for free
or reduced lunch), 76%. Thirty-two percent of the students received Limited English Proficiency
(LEP) services during the study or up to two years prior to the study, and 17% were enrolled in
special education during either or both years of the implementation. Twenty-five percent of the
students were White, 52% Hispanic, and 6% Black.
2.1.2. Writing curricula in study schools
Schools employed a variety of writing curricula. The teachers in about one-half of the
participating schools were involved in trying out a new writing curriculum, which includes daily,
Page 12
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 9
scripted writing lessons that highlight targeted skills. The lessons are taught systematically
within a planned hierarchy, scope, and sequence. The curriculum addresses three broad goals:
establishing foundational writing habits, mastering essential skills, and learning the process
approach to writing. By design, this curriculum does not attempt to teach all writing skills; it
deliberately targets specific skills to be covered in depth, within the context of a genre, with
mastery as the goal. The new program provides professional development for teachers and
administrators before the school year begins, and additional workshops and coaching are
provided throughout the year.
Teachers in other schools used a range of writing curricula. These curricula included
Lucy Calkins’ Units of Study for Teaching Writing (e.g., Calkins, 2006) or other variations of the
writers’ workshop; the 6+1 Traits of Writing (Culham, 2003); the Collins Writing Program (e.g.,
Collins, 2007, 2009); First Steps Literacy (e.g., Annandale et al., 2005; Education Department of
Western Australia, 1997); and Picturing Writing (e.g., Olshanksy, 2006, 2007). Other resources
included Ralph Fletcher’s lessons (e.g., Fletcher & Portalupi, 2001, 2007) and Strategies for
Writers (Crawford & Sipe, 2008); professional textbooks and works by noted authors in the field
of writing; and teacher-created prompts and rubrics based on guidelines from the Massachusetts
Department of Elementary and Secondary Education. Professional development and
coaching/mentoring varied.
2.2. Measures
2.2.1. Student writing survey
Students responded to a paper and pencil survey comprising 18 items with stems and four
choices (see Appendix). Each item choice ranged from weakest to strongest. For example, “I
write: not as well as my friends, about the same as my friends, a little better than my friends, a
Page 13
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 10
lot better than my friends.”
We wanted our survey to focus on self-efficacy, self-concept, and the perceived value of
writing without emphasizing feelings or specific writing tasks or locations. We also wanted it to
be brief enough to be administered in a short time. After examining various surveys, including
those already discussed, we decided that the best fit to our needs was the Motivation to Read
Profile (Gambrell, Palmer, Coding, & Mazzoni, 1996). By changing the format from reading to
writing and rewording a few items, the content and phrasing focused on the kinds of attitudes we
wanted to measure. We also added one item from Graham et al.’s (2007) Writing Attitude
Survey that was a condensed version of their seven items. Our research team wrote three
additional items. Students completed the surveys in class, which took about 10 minutes. Teachers
were permitted to assist with reading as needed, although our pilot work and feedback from the
teachers indicated that this was rarely necessary.
2.2.1.1. Principal component analysis
A principal component analysis of the 18 survey items indicated three composite
variables with eigenvalues above 1.0: (1) the Perceived Value (or usefulness) of Writing; (2)
Self-Rating as a Writer; and (3) Writing Behaviors. Self-Rating as a Writer and Writing
Behaviors both measure a form of self-efficacy, with Self-Rating as a Writer leaning more
toward students’ self-concepts. The correlations among the three components were very
consistent, ranging from .39 to .49 across time of year, Year 1 or 2, and grade. The median
correlation is .44. In other words, the trend is for students with higher (more positive) ratings on
the Perceived Value of Writing component to also have higher ratings on the Self-Rating as a
Writer and Writing Behaviors components.
Page 14
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 11
The Perceived Value of Writing component accounted for about 29% of the variance at
each administration of the survey; the Self-Rating as a Writer component accounted for about
10–12% of the variance; and the Writing Behaviors component accounted for about 7% of the
variance. The three factors together accounted for about 46 to 48% of the variance (see Table 1).
Table 1
Variance Accounted for by Components Principal Components Analysis of Student Writing Survey Items. Proportions of Variance Accounted for by First
Three Components (all Eigenvalues > 1.0).
Year 1 Fall Year 1 Spring Year 2 Fall Year 2 Spring
Component 1
(Value of Writing)
.2907 .2947 .2915 .2904
Component 2 (Self-
Rating as a Writer)
.0987 .1208 .1065 .1208
Component 3
(Writing Behaviors)
.0662 .0687 .0706 .0690
Cumulative
Proportion
.4556 .4843 .4686 .4802
The internal consistency for each of the three composite variables (Cronbach’s alpha) is
shown in Table 2. The reliability of the Perceived Value of Writing component ranges from .79
to .81, and the reliability of the Self-Rating as a Writer component ranges from .79 to .81. The
reliability of the Writing Behaviors component is not as strong, ranging from .55 to .64, and thus
should be interpreted with even greater caution.
Table 2
Internal Consistency of Composite Variables
Internal Consistency Reliability (Cronbach’s Alpha) for Composite Variables at Each Time Point
Value of Writing Self-Rating as a Writer Writing Behaviors
Year 1fall .79 .79 .55
Year 1 spring .81 .81 .64
Year 2 fall .79 .81 .58
Year 2 spring .80 .81 .61
Means and standard deviations for each item are presented in Table 3 on page 12. Two
items were omitted in the final version. Item 8 was originally included with the Writing
Behaviors component, but it did not load strongly onto it. Moreover, removing it greatly
increased the internal consistency of that component. Item 18 was a multi-part item, so it was not
Page 15
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 12
consistent with the other items in the survey. The final survey includes nine items in the
Perceived Value of Writing component, four in the Self-Rating as a Writer component, and three
items in the Writing Behaviors component.
Table 3
Item Information for Principal Component Analysis
Item Information for Principal Component 1: The Value of Writing
Item number Description Mean SD
2 Writing is something I like to do 1.94 .74
4 My best friends think writing is fun 2.46 .89
6 People who write a lot are interesting 3.26 .82
9 Knowing how to write well is important 1.58 .81
11 I think writing is a good way to spend
time
2.23 .97
13 When I grow up I will spend time writing 2.05 .79
14 When I am in a group talking about
writing, I talk about my ideas
2.53 .95
16 I like to share my writing 2.38 .87
17 When people share their writing, I feel
happy
3.43 .70
Item Information for Principal Component 2: Self-Rating as a Writer
1 My friends think I am a good writer 3.08 .80
3 I write better than my friends 2.77 .85
7 I am a poor/very good writer 2.05 .83
15 My teacher thinks I am a poor/very good
writer
3.07 .78
Item Information for Principal Component 3: Writing Behaviors
5 When I am writing, I understand what to
do
3.47 .59
10 When my teacher asks me a question
about what I write, I can never think of an
answer
1.85 .80
12 Writing is easy/hard for me 3.27 .78
Omitted items
8 I worry about what other kids think about
my writing
1.90 .95
18 Check the different places and times for
your writing. You may check more than
one box.
Multi-part item
Page 16
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 13
2.2.2. State test (Massachusetts Comprehensive Assessment System [MCAS])
The MCAS is administered to Massachusetts students in March. The tests include a
variety of response formats, including open-ended questions for students. A writing prompt is
part of the English Language Arts test in grades 4, 7, and 10. In grade 4, the prompt genre has
been narrative. The test is untimed, and students may take as long as they need to complete it.
The Massachusetts Department of Elementary and Secondary Education uses two scoring
rubrics—one for Topic/Idea Development (along a 6-point scale) and one for Standard English
Conventions (along a 4-point scale). Two people rate each writing sample, and the two ratings
are added together for a possible maximum score of 12 or 8 points, respectively.
2.2.3. Released MCAS prompts
2.2.3.1 Fourth and fifth grade pretests
As part of our work in studying the writing curriculum in the schools, we wanted to
obtain assessments of students’ writing at the beginning of each school year. The study’s writing
pretest was an MCAS writing prompt from a previous year. (The prompts used in prior years
become available to the public and are referred to as “released prompts.”) We used the same
fourth grade pretest in both years; a different released prompt was used for the fifth-grade pretest
in Year 3.
In collecting the pretest writing samples, we tried to replicate the conditions used in the
official MCAS assessments, although instead of allowing unlimited time for the writing
assessments we decided (after some pilot work and in consultation with teachers) to put a time
limit of one-hour for completing the responses to the pretests. Teachers followed written
instructions to help students provide identifying information and respond to the prompt.
Students were informed when 20 minutes remained, then 10 minutes. Although teachers were
Page 17
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 14
not allowed to help students write, they were permitted to help read the instructions as needed.
Students also had the option of using dictionaries, as on the state test. Those with Individual
Education Plans were permitted the same accommodations they would have on any test.
2.2.3.2. Fifth grade posttest
The MCAS writing prompt is not administered to fifth graders, so teachers gave another
released prompt as a posttest. They followed the same procedures when administering the fifth-
grade posttest as they did when administering the fourth- and fifth-grade pretests.
2.2.3.3. Holistic scoring
We recruited readers each year to score the released prompts. Most of the readers were
master’s or doctoral students or retired teachers and principals. Others had worked in related
fields and understood fourth- and fifth-grade student work. The training, followed by the scoring,
was undertaken after the end of the school year after the writing samples had all been collected.
We followed a rigorous protocol to ensure that our readers could rate the writing samples
reliably and accurately, using the two MCAS scoring rubrics. Project staff scanned all writing
samples, redacting identifying information. The readers then read and scored these redacted
versions so they could be kept uninformed about any school or classroom information or the time
of year (fall or spring) of the writing samples. Team leaders periodically checked the scores of
each reader against their own scores, generally three times per day. If there was more than a one-
point difference between the scores of the reader and the team leader, the reader received
additional training. After the completion of the training sessions for the scorers, all of the student
writing samples were rated by two people, and the scores were added together as they are on the
MCAS. These rating pairs were chosen randomly for each writing sample, and readers did not
know who their scoring partner was. When a reading pair provided ratings for either rubric with
Page 18
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 15
more than a one-point difference, a member of the research team arbitrated so that the ratings
were no more than one point apart.
2.3. Procedure
Project staff delivered boxes of teacher and principal packets to district liaisons or to
individual schools in August and again in late May. A cover letter for the teachers provided an
overview of the data-collection procedures, with the required dates for returning documents to
school offices. Student writing surveys and writing pretests (or fifth grade posttests) were
included in the packets.
Teachers could administer the surveys and tests anytime within a two-week period in
September or a one-week period in June.
3. Data Analysis
We began with a general description of the students’ writing scores and their responses to
the three writing attitude components at the beginning and at the end of the school year for each
cohort of students. Our next step was to examine the relationships between students’ attitudes
and their scores on the writing-prompt assessments. For this step we studied the correlation
coefficients, looking at the overall results for each cohort and then separately at the coefficients
for several subgroups of students (girls and boys, for example). We then focused in more detail
on the students’ attitude scores through a series of multi-level regression models in which we
could incorporate information about the potential clustering of students within classrooms and
schools. We used a three-level model, investigating the differences in students’ attitudes across
school districts, grade levels, and a set of student-level characteristics: gender, special education
status, LEP status at the time of the study, eligibility for free/reduced lunch, and ethnicity. We
conducted these regression analyses separately for the students’ attitudes at the beginning of the
Page 19
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 16
school year and at the end of the school year. We then repeated these analyses by looking for
changes in students’ attitudes over time. This series of steps helped us to investigate our main
questions: 1. What are the attitudes about writing among fourth- and fifth-grade students? 2.
How are their attitudes related to their scores on assessments of their writing performance? 3.
What differences are there in the writing attitudes of students with different characteristics? 4.
How do students’ attitudes about writing change over the course of the school year?
4. Results
4.1 Mean scores for all students: Writing assessments and attitudes about writing
4.1.1. Writing test scores
Descriptive statistics (means and standard deviations) for the student writing scores and
the three attitude components in the fall and spring are presented in Table 4 on page 17. In both
Year 1 and Year 2, the fourth-grade Topic/Idea Development writing scores averaged just under
5 points in the fall, and they rose to an average of about 6.5 points in the spring. According to the
MCAS scoring guide, scores of 5 or 6 indicate limited to rudimentary topic development, with
basic supporting details and simplistic language, so although the average scores did increase, the
levels of writing achievement were still relatively low. The scores on the use of Standard
English Conventions show a similar pattern of overall increase, from a mean of about 4.5 on the
fall pretest to a mean of about 6.25 in the spring. These scores indicate that students were able in
the spring to produce essays with relatively few errors in grammar usage and mechanics, and
students were doing relatively better on the Standard English Conventions scale compared to
their weaker performances on the Topic/Idea Development scale.
Page 20
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 17
Table 4
Means and Standard Deviations for Writing Scores and Attitude Composite Scores
Year 1 Grade 4 Year 2 Grade 4 Year 2 Grade 5
n Mean SD n Mean SD n Mean SD
Writing scores
Pretest Topic/Idea Development
(highest possible score = 12)
2176 4.96 1.51 2145 4.90 1.62 1396 5.24 1.88
Pretest Standard English
Conventions (highest possible
score = 8)
2176
4.65
1.32
2145
4.53
1.33
1396
4.86
1.57
Posttest Topic/Idea
Development
2209
6.50
1.58
2164
6.75
1.51
1479
5.78
1.74
Posttest Standard English
Conventions
2210
6.21
1.39
2164
6.30
1.26
1479
5.43
1.25
Attitude composite scores
Value of writing (fall) 2260 2.91 0.51 2163 2.90 0.52 1460 2.80 0.48
Self-rating as a writer (fall)
2260
2.83
0.64
2164
2.81
0.66
1460
2.75
0.60
Writing behaviors (fall)
2260
2.30
0.53
2163
2.30
0.54
1460
3.27
0.51
Value of writing (spring)
1918
2.83
0.52
1861
2.83
0.51
1346
2.71
0.50
Self-rating as a writer (spring)
1918
2.81
0.62
1861
2.76
0.62
1345
2.72
0.61
Writing behaviors (spring)
1918
2.35
0.53
1861
3.30
0.53
1343
3.28
0.54
Page 21
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 18
For the grade-5 students, the gains in the writing scores were even more modest than
those of the grade-4 students. On both the Topic/Idea Development and Standard English
Conventions scores there were small increases from the fall to the spring. The average score on
Topic/Idea Development moved from 5.2 to 5.8, which is still considered at a “rudimentary”
level of topic development. The Standard English Conventions scores moved from 4.9 to 5.4, so
on average they were still producing errors in grammar and usage that would interfere with the
effective communication of their ideas.
4.1.2. Students’ attitudes about writing
The mean scores on the three writing attitude composites are presented in Table 4 on
page 17. These scores were scaled using a range of 1 (low scores or poor attitude) to 4 (high
scores or positive attitude). We have no “norms” or other comparative information for these
scores, so it is difficult to interpret the means in the table, but it does appear that students were
above the mid-point (2.5 is the midpoint) on their ratings of the Value of Writing (means of
about 2.85) and on their ratings of themselves as writers (means of about 2.8). It is interesting to
note that students’ ratings on these two dimensions always dropped slightly from fall to spring;
there are slightly lower mean scores for both grade-4 and grade-5 students. When examining the
ratings of their Writing Behaviors, however, we see a slightly different pattern. Here, the overall
ratings were slightly more positive, with the mean for students during the second year of the
project exceeding 3.0 on the 4-point scale. So the students tended to rate their behaviors
(understanding what to do when writing, being able to respond to teacher questions) as more
positive than their ratings of their own skills or success as writers. These ratings of their
behaviors showed modest increases over the course of the school year, while their ratings of their
own skills showed small decreases. From this look at the mean scores, it appears overall that
Page 22
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 19
students demonstrated small improvements in their writing skills as judged by independent
reviews of their responses to writing prompts. On their own assessments of their writing
attitudes, however, the patterns were a bit different. Students reported very small increases
(positive changes) in their writing behaviors but little or no differences (actually some very slight
decreases) in their self-ratings of themselves as writers. We examine these relationships at the
individual-student level in our next set of analyses.
4.2. Relationship between attitudes and writing: Beginning and end of year
We examined the relationships between each of the three components of students’ writing
attitudes and their writing ability both at the beginning and end of the year (see Table 5 on page
20). For fourth graders, small but consistent correlations were found between each of the attitude
components and the assessments of student writing (the median coefficient is about .2). The
correlations were slightly stronger when student attitudes were measured at the end of the year,
particularly with the Self-Rating as a Writer component (coefficients of about .25). To a lesser
extent, this pattern is similar for the fifth-grade students.
We carried out analyses of these patterns for each grade level and also by comparing the
coefficients for boys to those of girls, and the patterns were generally very consistent, with no
statistically significant differences between the grade levels or the genders. We also found no
important differences in the correlation coefficients across classrooms using different types of
writing curricula.
Page 23
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 20
Table 5
Correlations of Student Writing Survey Results with Writing Achievement Scores Year 1 Grade 4 Beginning-of-School-Year Attitudes
(n = 2146) End-of-School-Year Attitudes
(n = 1831)
Writing Test Value of Writing
Self-Rating as a Writer
Writing Behaviors
Value of Writing
Self-Rating as a Writer
Writing Behaviors
Pretest Topic Development
.13 .18 .25 .10 .27 .26
Pretest Standard Conventions
.10 .17 .23 .08 .29 .28
MCAS Prompt Topic Development
.17 .20 .22 .13 .26 .24
MCAS Prompt Standard Conventions
.15 .20 .22 .13 .28 .24
Year 2
Grade 4
Beginning-of-School-Year Attitudes (n = 2115)
End-of-School-Year Attitudes (n = 1844)
Pretest Topic Development
.15 .17 .22 .16 .25 .25
Pretest Standard Conventions
.11 .17 .20 .12 .25 .21
MCAS Prompt Topic Development
.11 .15 .20 .13 .24 .24
MCAS Prompt Standard Conventions
.11 .19 .22 .13 .26 .27
Year 2
Grade 5
Beginning-of-School-Year Attitudes (n = 1366)
End-of-School-Year Attitudes (n = 1291)
Pretest Topic Development
.07 .17 .21 .09 .21 .20
Pretest Standard Conventions
.01 .13 .18 .04 .18 .15
End-of-Year Prompt Topic Development
.07 .17 .19 .12 .22 .22
End-of-Year Prompt Standard Conventions
.01 .18 .17 .08 .20 .21
Note. Actual sample sizes for some pairs of variables differed slightly due to occasional missing data. All fourth
grade coefficients are statistically significant at p < .001. Grade 5 coefficients greater than .04 are statistically
significant at p < .05; coefficients greater than .09 are statistically significant at p < .001.
Page 24
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 21
The interesting finding from these analyses is that there appear to be small-but-consistent
relationships between students’ attitudes toward writing and their scores on writing assessments.
Most coefficients were in the .20 range, so there is not evidence that students’ attitudes are
strongly related to their writing abilities, but there does appear to be a small and statistically
significant relationship between student attitudes and writing performance, which is consistent
with previous studies of the relationship between student attitudes and writing performance. In
particular, students’ responses to questions about the value of writing (e.g., “knowing how to
write well is important,” “writing is something I like to do”) showed the lowest correlation with
actual writing scores. Many students stated that writing is valuable or important, even though
their own writing skills were relatively low. The coefficients for students’ ratings of their own
writing (e.g., “my friends think I am a good writer,” “my teacher thinks I am a good writer”)
were also not strong, but they were consistently stronger than the coefficients for the writing-
behavior attitudes.
Comparing the scores from the beginning of school year to the end, students’ scores on
the writing assessments did increase modestly, but we did not see corresponding changes in
students’ overall attitudes about themselves as writers. In the next step, we carried out some
further analyses of the attitude scores to look for differences in writing attitudes among
subgroups of students.
4.3. Differences in writing attitudes among student subgroups
We carried out a series of regression analyses to investigate potential differences in
students’ attitudes about writing. These were multi-level models to take into account the
clustering of students within classrooms and classrooms within schools. We included some
school-level characteristics to account for the different school districts and also an indicator for
Page 25
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 22
whether the school was taking part in the new writing curriculum we were studying or was part
of the group of schools continuing to use their usual curricula (hereafter referred to as “usual
curricula.” Most of our attention in these analyses, however, was on the student-level
characteristics. (The differences between schools and between classrooms turned out to be very
modest, so most of the variation in attitudes about writing appeared at the student level.) We
looked for differences in students’ attitudes at the beginning of the school year and then again at
the end of the school year, as we were interested in how students may vary in the fall and how
things might look a bit different in the spring. We also focused directly on individual changes in
attitudes over time (described in Section 4.4).
Table 6 on pages 23 and 24 presents results for the grade-4 and grade-5 cohorts. One
important finding for grade 4 is the relatively small variance estimates for the school and
classroom-within-school levels in these analyses. These are the conditional estimates, and so
they indicate that most of the variance is accounted for by student-level factors. Students differed
in their attitudes about writing, but there were relatively small differences in attitudes among
schools and even among classrooms; within most classrooms there is considerable variation
among students in their attitudes about writing. The coefficients presented in the table represent
differences in the mean attitude scores across the corresponding subgroups of students. (The
scores are on a 4-point scale.) The results for the grade-4 cohorts were relatively consistent.
Girls consistently demonstrated more positive attitudes than boys, with statistically significant
differences at both the beginning and end of the school year. (The difference averaged about .20
scale points, which is not huge, but in effect-size terms these correspond to approximately .30 to
.40 standard deviations.) The differences between the attitudes of the girls and the boys are the
most consistent and the largest in magnitude of the differences for the fourth-grade students.
Page 26
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 23
Table 6
Regression Results (Mean Scores of Subgroups)
Year 1 Grade 4 Beginning-of-School-Year Attitudes
(n = 1962)
End-of-School-Year Attitudes
(n = 1658)
Variable Value of
Writing
Self-Rating
as a Writer
Writing
Behaviors
Value of
Writing
Self-Rating
as a Writer
Writing
Behaviors
Intercept 2.80 2.78 3.34 2.81 2.88 3.47
Girls .25*** .29*** .15*** .18*** .22*** .06*
White -.03 -.07 .04 -.02 .00 .01
Hispanic -.07~ -.09~ -.03 -.09* -.05 -.09*
F/R Lunch .00 -.03 -.03 -.02 -.10* .07~
LEP .05 .10* -.09* .07 -.20*** -.15***
SPED -.07* -.17*** -.16*** -.11** -.19*** -.19***
New Curriculum -.07* -.06 .03 -.04 -.09* -.03
District B .10 .02 .04 -.03 .05 -.03
District C .00 -.01 -.13 .11~ -.03 -.10~
District D 0 0 0 0 0 0
District E -.06 -.02 -.03 -.10 .03 -.01
New Curriculum x District ns ns ns ns ns ns
Mean Changes for
Usual Curricula
New Curriculum
2.95
2.88
2.87
2.81
3.32
3.29
2.85
2.81
2.77
2.85
3.36
3.33
Variance Estimates for
Schools (n = 40)
Classrooms (n = 123)
Residual
.00008
.01131**
.2373***
.00261
.01321*
.3676***
0
.00651*
.2583***
0
.0144***
.2341***
0
.0045
.3506***
0
.00395
.2614
Year 2 Grade 4 Beginning-of-School-Year Attitudes
(n = 2137)
End-of-School-Year Attitudes
(n = 1854)
Intercept 2.77 2.76 3.37 2.75 2.85 3.37
Girls .24*** .29*** .13*** .20*** .21*** .11***
White -.07~ -.13* -.05 -.02 -.10~ -.04
Hispanic -.07~ -.09~ -.07~ -.08* -.13** -.12**
F/R Lunch .00 -.05 -.04 .02 -.07~ -.05
LEP .14*** .04 -.05 .03 -.12** -.09*
SPED -.04 -.04 -.19*** -.01 -.06 -.17***
New Curriculum .01 -.05 -.03 .02 .02 .01
District A 0 0 0 0 0 0
District C .11 .03 .02 .08 .01 .07
District D -.01 -.03 -.00 -.06 -.06 .04
District E .03 .04 .03 -.03 -.00 .07
District F .07 .05 -.13* -.02 -.00 -.04
Page 27
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 24
Note. ~ p < .10. * p < .05 ** p < .01 *** p < .001
New Curriculum x District ns ns ns ns ns ns
Mean Changes for
Usual Curricula
New Curriculum
2.89
2.89
2.85
2.80
3.32
3.29
2.80
2.82
2.76
2.77
3.29
3.30
Variance Estimates for
Schools (n = 40)
Classrooms (n = 132)
Residual
.0024
.0094
.2336***
0
.0029
.4049***
0
.0009
.2710
0
.0202***
.2213***
.0007
.0117*
.3456***
0
.0031
.2656***
Year 2 Grade 5 Beginning-of-School-Year Attitudes
(n = 1361)
End-of-School-Year Attitudes
(n = 1254)
Value of
Writing
Self-Rating
as a Writer
Writing
Behaviors
Value of
Writing
Self-Rating
as a Writer
Writing
Behaviors
Intercept 2.80 2.77 3.29 2.63 2.74 3.32
Girls .19*** .26*** .08** .16*** .19*** .06*
White -.03 .00 .14** .05 .06 .09
Hispanic -.10* -.04 -.02 -.06 -.04 -.04
F/R Lunch -.10* -.16*** -.06 -.07~ -.17** -.09*
LEP .10** -.04 -.12** .12** -.06 -.10*
SPED -.10** -.15*** -.19*** -.07* -.14** -.19***
New Curriculum .03 .06 -.04 .02 .01 .00
District C .15** .08 .01 .19** .14~ .08
District D 0 0 0 0 0 0
District E -.02 .08 -.01 .04 .07 .03
New Curriculum x District ns ns ns ns ns ns
Mean Changes for
Usual Curricula
New Curriculum
2.76
2.79
2.69
2.76
3.28
3.24
2.69
2.70
2.71
2.72
3.26
3.26
Variance Estimates for
Schools (n = 28)
Classrooms (n = 91)
Residual
.0011
.0010
.2164***
.0042~
.00
.3305***
.0041~
.00
.2441***
.0017
.0038
.2285***
.0063~
.00
.3419***
.0021
.00
.2728***
Page 28
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 25
There were smaller differences across the ethnic groups, with Hispanic students showing
slightly less-positive attitudes than other students. More noticeable were the differences for
students receiving special education services and, to a slightly lesser degree, students receiving
LEP services, although the special education differences were more prominent in the Year-1
cohort than in Year 2. These differences were slightly more prominent in the attitudes of the
students at the end of the school year.
We do not see strong or consistent differences between the classrooms that were adopting
the new writing curriculum and those who used their usual curricula. Nor were there major
differences across the school districts, and there were no significant interactions of districts and
writing curriculum, so we did not see any curriculum differences in students’ attitudes either
overall or in particular school districts.
The grade-5 results presented in Table 6 are generally consistent with the grade-4 results.
The overall mean attitude scores for the grade-5 students are actually slightly lower than the
grade-4 students’, and we see only relatively small differences among schools and classrooms.
Most of the variability in students’ attitudes is present within classrooms. Consistent with the
grade-4 results are the differences between the girls and boys, with girls continuing to show
slightly more-positive attitudes toward writing.
There are relatively few and relatively small differences based on ethnicity, but larger
differences based on special education, LEP and eligibility for free/reduced lunch. One
interesting pattern here was that the students receiving LEP services had slightly higher (more
positive) attitudes about their Value of Writing (differences of about .10 in both fall and spring),
yet they had slightly lower scores on the other attitude scales, averaging about .10 lower than
other students on their ratings of their own writing behaviors. The coefficients for the special
Page 29
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 26
education students were consistently negative, reflecting the slightly less positive attitudes of
students receiving special education services in grade 5. As we also saw in the grade-4 results,
there were no consistent differences across classrooms using different writing curricula, and few
important differences among the school districts, although one district did show slightly higher
attitudes on the value of writing.
Perhaps the main results from these analyses are not a major surprise. We found that
students’ attitudes toward writing vary, and we see relatively small differences among schools
and classrooms. Differences among classrooms based on their use of a particular writing
curriculum were hardly noticeable. The notable trend was for girls to indicate more positive
attitudes than boys. Among other subgroups, students receiving special education or LEP
services often demonstrated somewhat different attitudes, and these may well reflect their
perceptions of differences in their skills in writing in school.
4.4. Changes across the year
The earlier tables presented the student-attitude analyses separately for the beginning and
end of the year. Table 7 on page 28 shows results of analyses in which we examined changes in
students’ attitudes from the fall to the spring. Overall there were very modest changes over time,
and in fact there were small trends toward less-positive attitudes. There was significant
variability in students’ attitudes, however, with most of the variance at the student level and
relatively small variability across schools or classrooms. Among the grade-4 cohorts, girls had
more-positive attitudes than boys, but girls actually changed a bit less than the boys. These
differences in the changes in attitudes were small, however, and were not always statistically
significant, although the coefficients were consistently negative. A statistically significant
difference for the grade-4 students was for the students receiving LEP services to show larger
Page 30
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 27
decreases in their attitudes, most noticeably on the Self-Rating as a Writer component. The only
other statistically significant coefficient in the fourth-grade analyses was for a more positive
change for Year 1 classrooms on the Self-Rating as a Writer component that were using the new
writing curriculum. Unfortunately this was not found among the second grade-4 cohort, nor was
the difference consistent across the different writing attitude variables.
There were no important differences in the changes in students’ attitudes for any of the
other characteristics. Ethnicity differences were nearly zero. Even students receiving special
education services, whose attitudes were often less positive than those of other students, did not
demonstrate changes that differed significantly from those of other students.
The analyses of changes in attitudes among the grade-5 students were generally similar.
There were very small overall differences, with attitudes moving slightly in a less-positive
direction. The girls’ attitudes changed slightly less than boys’. The only other significant
coefficient occurred for students receiving special education services, who showed a small
positive change in the attitudes about the value of writing, although they still remained less
positive overall than other students.
The analyses of the changes in students’ attitudes over the school year indicated that there
were only small overall changes. Many students did show some changes, but these differences
were not large and were not associated with either student-level characteristics or the different
schools or school districts. The attitudes may be relatively slow to change for these young
students in grades 4 and 5, who are somewhat new in tackling the challenges of writing
assignments and writing instruction in school.
Page 31
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 28
Table 7
Mean Score Changes in Student Attitudes Beginning to End of Year
Year 1 Grade Year 2 Grade 4 Year 2 Grade 5
Variable Value of
Writing
Self-Rating
as a Writer
Writing
Behaviors
Value of
Writing
Self-Rating
as a Writer
Writing
Behaviors
Value of
Writing
Self-Rating
as a Writer Writing Behaviors
Intercept -.00 .07 .13 -.02 -.10 .08 -.15 -.07 .01
Girls -.06* -.05 -.08** -.03 -.06* .02 -.03 -.07* .02
White .02 .09 -.01 .02 .00 -.04 .04 .06 -.01
Hispanic -.03 .06 -.04 -.02 -.04 -.03 -.00 .02 .01
F/R Lunch -.01 -.05 -.03 -.01 .03 -.04 .02 -.02 -.03
LEP -.01 -.28*** -.08 -.06 -.12* -.05 .03 .03 -.01
SPED -.04 -.01 -.05 .01 .05 .04 .09* -.01 .02
New Curriculum .04 .15** .02 .01 .05 .04 -.01 -.04 .06
District A 0 0 0
District B -.16* -.02 .02
District C .10 -.00 .04 -.07 -.06 -.00 .04 .07 .07
District D 0 0 0 -.05 -.05 .03 0 0 0
District E -.03 .04 .04 -.12 -.05 .00 .06 .02 .06
District F -.09 -.04 .12~
New Curriculum x
District ns ns ns
ns ns ns
ns ns ns
Mean Changes for
Usual Curricula
New Curriculum
-.13
-.08
-.11
.04
.00
.03
-.08
-.07
-.08
-.03
-.02
.02
-.07
-.08
-.02
-.06
-.04
.02
Variance estimates for
Schools (n = 40)
Classrooms (n = 123)
Residual
.00
.0166**
.2626***
.00
.0255***
.4055***
.00
.0075*
.3106***
.00
.0080*
.2676***
.00
.0095*
.3891***
.00
.0004
.3205***
.0072
.0022
.1873***
.0114*
.0041
.3057***
.0029
.0023
.2737***
Note. ~ p < .10. * p < .05 ** p < .01 *** p < .001
Page 32
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 29
5. Discussion
Findings from this study support the existence of a small relationship between student
attitudes about writing and writing performance that other studies have found (Graham et al.,
2007; Knudson, 1995; Pajares et al., 1999). This adds to the convergence of evidence about this
relationship despite the different ways in which researchers define or measure attitudes. Other
studies have used 4-, 5-, 6- and 100-point scales, as well as verbal descriptors or Garfield™
pictures; and items that measure self-efficacy or self-concept, focus on student emotions, stress
specific skills, purposes for writing, or to a lesser extent items that measure perceived usefulness.
In all cases attitudes predict writing performance. Therefore, we can study different aspects of
motivation or use different instruments, or both, and still arrive at similar conclusions.
5.1. Students from low-income families
Of interest too is that we were able to confirm the relationship between attitudes and
writing performance with students who came from primarily high-poverty homes. In the most
recent study of student attitudes, Graham et al. (2007) used mothers’ and fathers’ educational
levels to measure socioeconomic status (SES). Ninety-three percent of the mothers and 83% of
the fathers had at least some higher education, indicating a population with higher SES. We used
eligibility for free and reduced lunch as a proxy for SES. In our study, some 76% of the students
were eligible for free or reduced lunch, indicating a population with lower SES. Therefore, our
results offer some additional information about the question posed by Graham et al. about
whether poverty may mediate the relationship between attitudes and writing performance. Our
results indicate that there is indeed a relationship between attitudes and writing ability for
students from families with low incomes.
Page 33
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 30
5.2. Poor writers
The students in the Graham et al. (2007) study were also better writers than those in our
study. Whereas the scores of the Graham et al. students were solidly in the average range on a
standardized test of written expression, the students in our study had low mean scores on the
MCAS and the released prompts, indicating that these students were weaker writers overall. This
addresses another question raised by Graham et al. about whether the relationship between
attitudes and writing outcomes would hold true for poor writers or those with disabilities. Our
results indicate that the relationship does hold true.
5.3. Different student populations
The students in our study also differed from that of those in the Graham et al. (2007)
study demographically. Sixty-five percent of the students in that study were White compared to
25% in our study and 1% were Hispanic compared to 52% in our study. Graham et al. did not
address the percentage of students receiving LEP services, if any; 32% of the students in our
study either received these services at the time of the study or up to two years prior to it. We
were thus able to find similar results as previous researchers with a different population of
students.
5.4. Attitudes from beginning to end of year
Student attitudes about their writing ability did not get stronger on average after a year of
writing instruction. This issue was not addressed in other studies because they measured both
attitudes and writing ability at the same time. Our study had the advantage of being able to assess
these variables at the beginning and end of the year. Interestingly, although student attitudes
were not stronger at the end of the school year, their beliefs were more closely aligned to their
actual writing ability; that is, the correlations between attitudes and writing ability tended to be
Page 34
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 31
higher at the end of the year. Although the differences in the magnitudes of the coefficients were
not large, there was a consistent trend for end-of-year coefficients to be a bit stronger than
beginning-of-year coefficients. Perhaps this is a reflection of the students having experienced a
year of instruction from their writing teachers, and the feedback from and interactions with the
teachers (along with other classroom experiences) provided them with a slightly firmer basis for
making judgments about their own writing attitudes. This finding is worth pursuing in other
studies of individual students and how their attitudes and writing skills and abilities evolve and
develop over time.
Also of interest, although students’ attitudes about the Value of Writing component
appeared to show reasonable levels of internal consistency (Cronbach’s alpha coefficients of
.80), that attitude variable demonstrated the lowest correlation with student writing. Students’
self ratings, and their reports of their own writing behaviors, showed consistently stronger
coefficients. It appears that although students may value writing, and they recognize that it is an
important skill that is valued in school, their performance did not correlate highly with it. Even
students who were weaker writers often reported that they recognized the value of writing. This
finding too should be followed up by longitudinal research that tracks students’ attitudes and
writing development over longer periods of time.
5.5 Girls vs. boys
Our results indicate that girls have more positive attitudes than boys, a finding that is
consistent with previous research. Girls had more positive attitudes than boys on all three attitude
components regardless of when the survey was administered. They began the year with more
positive attitudes and maintained them or became only slightly less positive, whereas boys
started and ended each year with less positive attitudes. Girls in both grades were also better
Page 35
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 32
writers than boys, which supports Pajares et al.’s (1999) and Knudson’s (1995) findings but
differs from the Graham et al. (2007) results.
5.6. Changes in attitudes with grade
Past research has been equivocal about attitudes declining as children get older (Pajares,
2003). Knudson (1991, 1992, 1993) confirmed the decline with surveys in grades 1–3, 4–8, and
9–12 as did Pajares and Valiante (1999) with students in grades 6 and 7, but Graham et al. (2007)
did not find this to be true with their first and third graders. Nor did we find it to be true for our
fourth and fifth graders. A direct path from attitude to performance best described Graham et
al.’s data, but the third-grade students did not have more positive attitudes than the first graders,
as the authors expected. Our study also differed from others in that many of our students were
followed over two years.
5.7. Limitations and implications for future research
One limitation of this study is that it focused only on one aspect of motivation—
attitudes—and that our definition of attitudes was different from the way other researchers define
it. While we were pleased to see the convergence of evidence of the relationship between
attitudes and writing outcomes regardless of the definitions offered thus far, it is also possible
that there is another aspect of attitudes, or in fact of motivation, that might not have that
relationship. It is also possible that future research with the same three components we studied
might have different results with older or younger students.
Another limitation is that all of the districts that participated in this study were eligible
for Title 1 funds and comprised primarily students from low-income families. This
complemented the student population in the Graham et al. (2007) study, but it should be noted
that we studied grades 4 and 5, whereas Graham et al. studied grades 1 and 3. Future research
Page 36
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 33
studies might over time include a wider range of grades as well as students from families across
the income spectrum over a series of studies.
It should also be noted that student attitudes about writing are potentially dynamic.
However, they have generally been measured at only one time point. Although we attempted to
address this issue by collecting data at two points each year, additional research at multiple data
points, spanning more grades, could add important information to the field.
6.0. Conclusions and implications
This article focused on student attitudes about writing, and in particular about three
components of those attitudes: students’ Perceived Value of Writing, Self-Rating as a Writer,
and Writing Behaviors. We found a small but consistent relationship between all three
components and writing, and this relationship was a little stronger at the end of the year,
especially for the Self-Rating as a Writer component. Girls had significantly more positive
attitudes than boys on all three components in fall and spring of both grades. However, their
attitudes changed slightly less than boys from the beginning to end of the year on the Perceived
Value of Writing and Writing Behaviors components.
We were able to confirm the relationship that other researchers have found between
attitudes and writing outcomes although the students in our study came primarily from low-
income families, had a high percentage of Hispanic students, and were on average poor writers.
Moreover, we found that the relationship was stronger by the end of the year.
Although various researchers have defined attitudes differently and studied different
components, there appears to be a convergence of the findings. It is also possible that other
aspects could be studied that would produce different results. Moreover, attitudes are only one
facet of motivation, and other facets could also lead to different results.
Page 37
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 34
We suggest that future research studies follow students over time to document the
dynamics involved in the progression of their writing ability as well as their attitudes about
writing and the relationship between the two variables. We believe that our examination over
time was a positive feature of this study, and we hope it will encourage other researchers to study
a variety of populations over multiple data collection points using a range of methodologies. In
that way, a body of research can accumulate to include a wider range of grades with families
across the income spectrum.
Finally, we wonder if it is appropriate to expect writing attitudes to change over time.
Might this not depend on student, teacher, and classroom contexts? Additional research could
detect not only the change in attitudes, if any, but the reasons behind the positive or negative
changes. This would require qualitative as well as quantitative data collection to present a more
complete picture. We would encourage follow-up studies on the role that teachers and schools
might play in influencing student attitudes. While the causal role of attitudes on writing
performance has not yet been clearly determined, surely positive attitudes would lead to a more
harmonious learning environment.
NOTE:
The research reported here was supported by the Institute of Education Sciences, US Department of
Education, through Grant Number R305A090479 to Education Development Center, Inc. The
opinions expressed are those of the authors and do not represent views of the Institute or the US
Department of Education.
Page 38
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 35
References
Albin, M. L., Benton, S. L., & Khramtsova, I. (1996). Individual differences in interest and
narrative writing. Contemporary Educational Psychology, 21, 305–324.
Alexander, P. A., Graham, S., & Harris, K. R. (1998). A perspective on strategy research:
Progress and prospects. Educational Psychology Review, 10(2), 129–154.
Annandale, K., Bindon, R., Broz, J., Handley, K., Johnston, A., Lockett, L. . . . Rourke, R.
(2005). First steps literacy: Writing map of development and CD-ROM pack (2nd
ed.). Perth,
Australia: Steps Professional Development.
Brophy, J. (2010). Motivating students to learn (3rd
ed.). New York, NY: Routledge.
Calkins, L. (2006). Units of study for teaching writing, grades 3–5. Portsmouth, NH:
Heinemann.
Clark, I. L. (2012). Concepts in composition: Theory and practice in the teaching of writing (2nd
ed.). New York, NY: Routledge.
College Board Advocacy and Policy Center. (2010). Teachers are the center of education:
Writing, learning and leading in the digital age. New York, NY: College Board.
Collins, J. J. (2007). Improving student writing performance through writing and thinking across
the curriculum. West Newbury, MA: Collins Education Associates.
Collins, J. (2009). Collins writing program: Using best-practice research and time saving
techniques to improve student performance. West Newbury, MA: Collins Education
Associates.
Crawford, L. W., & Sipe, R. B. (2008). Strategies for writers. Columbus, OH: Zaner-Bloser.
Culham, R. (2003). 6+1 traits of writing: The complete guide, grades 3 and up. New York, NY:
Scholastic.
Page 39
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 36
Daly, J. A., & Shamo, W. (1978). Academic decisions as a function of writing apprehension.
Research in the Teaching of English, 12(2), 119–126.
Eagly, A. H., & Chaiken, S. (2007). The advantages of an inclusive definition of attitude. Social
Cognition, 25(5), 582–602.
Education Department of Western Australia. (1997). Writing developmental continuum (First
Steps). Portsmouth, NH: Heinemann.
Faigley, L., Daly, J., & Witte, S. (1981). The role of writing apprehension in writing
performance and competence. Journal of Educational Research, 75, 16–21.
Fazio, R. H., & Olson, M .A. (2003). Attitudes: Foundations, functions, and consequences. In M.
A. Hogg & J. Cooper (Eds.), The SAGE handbook of social psychology. London, England:
Sage.
Fletcher, F., & Portalupi, J. (2001). Writing workshop: The essential guide. Portsmouth, NH:
Heinemann.
Fletcher, R., & Portalupi, J. (2007). Craft lessons: Teaching writing K–8 (2nd
ed.). Portland, ME:
Stenhouse.
Gambrell, L. B., Palmer, B. M., Codling, R. M., & Mazzoni, S. A. (1996, April).
Assessing motivation to read. The Reading Teacher, 49(7), 518–533.
Graham, S., Berninger, V., & Fan, W. (2007). The structural relationship between writing
attitude and writing achievement in first and third grade students. Contemporary
Educational Psychology, 32, 516–536.
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(4), 237–249.
Page 40
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 37
Hidi, S., Berndorff, D., & Ainley, M. (2002). Children’s argument writing, interest, and self-
efficacy: An intervention study. Learning and Instruction, 12(4), 429–446.
Kear, D. J., Coffman, G. A., McKenna, M. C., & Ambrosio, A. L. (2000). Measuring attitude
toward writing: A new tool for teachers. The Reading Teacher, 54(1), 10–23.
Knudson, R. E. (1991). Development and use of a writing attitude survey in grades 4 to 8.
Psychological Reports, 68, 807–816.
Knudson, R. E. (1992). Development and application of a writing attitude survey for grades 1 to
3. Psychological Reports, 70, 711–720.
Knudson, R. E. (1993a). Effects of ethnicity in attitudes toward writing. Psychological Reports,
72, 39–45.
Knudson, R. E. (1993b). Development of a writing attitude survey for grades 9 to 12: Effects of
gender, grade, and ethnicity. Psychological Reports, 73, 587–594.
Knudson, R. E. (1995). Writing experiences, attitudes, and achievement of first to sixth graders.
Journal of Educational Research, 89(2), 90–97.
Levy, F., & Murnane, R. J. (2004). The new division of labor: How computers are creating the
next job market. Princeton, NJ: Princeton University Press.
Murnane, R., & Levy, F. (1996). Teaching the new basic skills: Principles for educating children
to thrive in a changing economy. New York, NY: Free Press.
Murphy, P. K., & Alexander, P. A. (2000). A motivated explanation of motivation terminology.
Contemporary Educational Psychology, 25, 3–53.
National Center for Education Statistics. (2012). The nation’s report card: Writing 2011 (NCES
2012-470). Washington, DC: US Department of Education.
http://nces.ed.gov/nationsreportcard/pubs/main2011/2012470.asp.
Page 41
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 38
National Governors Association Center for Best Practices. (2005). Reading to achieve: A
governor’s guide to adolescent literacy. http://www.nga.org/cms/home/nga-center-for-best-
practices/center-publications/page-edu-publications/col2-content/main-content-list/reading-
to-achieve-a-governors-g.html.
Nelson, N. (2007). Why write? A consideration of rhetorical purpose. In G. Rijlaarsdam (Series
Ed.) and P. Boscolo & S. Hidi (Volume Eds.), Studies in Writing, Volume 19, Writing and
Motivation (pp. 17–30). Oxford, England: Elsevier.
Olshansky, B. (2006). Teaching the art of writing. Educational Leadership, 63. Retrieved from
http://www.ascd.org/publications/educational-leadership/ summer06/vol63/num09/Teaching-
the-Art-of-Writing.aspx
Olshanksky, B. (2007). Impact of an art-and-literature-based approach on writing and reading
skills of at-risk learners: Sixteen years of evidence. Arts & Learning Research Journal,
23(1), 1–30.
Pajares, F. (2003). Self-efficacy beliefs, motivation, and achievement in writing: A review of the
literature. Reading and Writing Quarterly, 19, 139–158.
Pajares, F., & Johnson, M. J. (1996). Self-efficacy beliefs and the writing performance of
entering high school students. Psychology in the Schools, 33, 163–175.
Pajares, F., Miller, M. D., & Johnson, M. J. (1999). Gender differences in writing self-beliefs of
elementary school students. Journal of Educational Psychology, 91(1), 50–61.
Pajares, F., & Valiante, G. (1999). Grade level and gender differences in the writing self-beliefs
of middle school students. Contemporary Educational Psychology, 24(4), 390–405.
Page 42
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 39
Pajares, F., & Valiante, G. (2001). Gender differences in writing motivation and achievement of
middle school students: A function of gender orientation? Contemporary Educational
Psychology, 26, 366–381.
Partnership for 21st Century Skills. (2011). Framework for 21
st century learning. Washington,
DC: Author. http://www.p21.org/overview/skills-framework.
Persky, H. R., Daane, M. C., & Jin, Y. (2003). The nation’s report card: Writing 2002, (NCES
2003–529). Washington, DC: Government Printing Office.
Schunk, D. H., Meece, J. L., & Pinrich, P. R. (2014). Motivation in education (4th
ed.). Boston,
MA: Pearson.
Schwarz, N. (2007). Attitude construction: Evaluation in context. Social Cognition, 25(5), 638–
656.
Shell, D. F., Colvin, C., & Bruning, R. H. (1995). Self-efficacy, attribution, and outcome
expectancy mechanisms in reading and writing achievement: Grade-level and achievement-
level differences. Journal of Educational Psychology, 87(3), 386–398.
Troia, G. A., Harbaugh, A. G., Shankland, R. K., Wolbers, K. A., & Lawrence, A. M. (2013).
Relationships between writing motivation, writing activity, and writing performance: Effects
of grade, sex, and ability. Reading and Writing, 26(1), 17–44.
Wigfield, A., & Eccles, J. S. (1992). The development of achievement task values: A theoretical
analysis. Developmental Review, 12(3), 265–310.
Zimmerman, B. J. (2011). Motivational sources and outcomes of self-regulated learning and
performance. In B. J. Zimmerman & D. H. Schunk (Eds.), Handbook of self-regulation of
learning and performance (pp. 49–64). New York, NY: Routledge.
Page 43
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 40
Appendix
Student Writing Survey *
Name ______________________________ Date _____________________________
District ____________________________ School ___________________________
Teacher ____________________________ Grade _________ Male Female
Please answer all questions by checking off the box.
1. My friends think I am:
a very good writer
a good writer
an OK writer
a poor writer
2. Writing is something I like to do:
Never
Not very often
Sometimes
Often
3. I write:
not as well as my friends
about the same as my friends
a little better than my friends
a lot better than my friends
Page 44
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 41
4. My best friends think writing is:
really fun
fun
OK to do
no fun at all
5. When I am writing, I understand what to do:
almost always
sometimes
hardly ever
never
6. People who write a lot are:
very interesting
interesting
not very interesting
boring
7. I am:
a poor writer
an OK writer
a good writer
a very good writer
Page 45
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 42
8. I worry about what other kids think about my writing:
every day
almost every day
once in a while
never
9. Knowing how to write well is:
not very important
sort of important
important
very important
10. When my teacher asks me a question about what I write, I:
can never think of an answer
have trouble thinking of an answer
sometimes think of an answer
always think of an answer
11. I think writing is:
a boring way to spend time
an OK way to spend time
an interesting way to spend time
a great way to spend time
Page 46
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 43
12. Writing is:
very easy for me
kind of easy for me
kind of hard for me
very hard for me
13. When I grow up I will spend:
none of my time writing
very little of my time writing
some of my time writing
a lot of my time writing
14. When I am in a group talking about writing, I:
almost never talk about my ideas
sometimes talk about my ideas
almost always talk about my ideas
always talk about my ideas
15. My teacher thinks I am:
a very good writer
a good writer
an OK writer
a poor writer
Page 47
RELATIONSHIP BETWEEN ATTITUDES AND WRITING OUTCOMES 44
16. I like to share my writing:
every day
almost every day
once in a while
never
17. When people share their writing, I feel:
very happy
sort of happy
sort of unhappy
unhappy
18. Check the different places and times for your writing. You may check more than one box:
before school
after school
free time at school
writing time in school
school vacations
at home
* Adapted from:
Gambrell, L. B., Palmer, B. M., Codling, R. M., & Mazzoni, S. A. (1996). Assessing motivation to
read. The Reading Teacher, 49(7), 518-533 (Items #1-14).
Graham, S., Berninger, V., & Fan, W. (2007). The structural relationship between writing attitude and writing
achievement in first and third grade students. Contemporary Educational Psychology, 32, 516-536 (Item #18).
Items #15–17 were added by the research team.