Writing and Writing Motivation of Students Identified as English Language Learners

The purpose of this study was to examine the writing performance and motivational beliefs of students who were identified by their school district as English language learners. The study included 880 students (463 girls; 417 boys) in grades three to eight who wrote an informative/ explanatory essay on information technology and completed a motivational survey assessing their intrinsic, extrinsic, and self-regulation incentives for writing. Ninety-seven percent of students’ scores on the writing measure did not meet grade-level proficiency for writing, girls received higher scores than boys, and writing scores generally increased across the six grade-levels. A Article International Journal of TESOL Studies (2021) Vol. 3 (1) 1-13 https://doi.org/10.46451/ijts.2021.01.01 *Corresponding Author Address: Mary Lou Fulton Teachers College, Arizona State University, P.O. Box 871811, Tempe, AZ, USA Email: steve.graham@asu.edu 2 International Journal of TESOL Studies 3 (1)


Introduction
At least one out of every 10 students in schools in the United States (U.S.) are identified as English language learners (ELL; Soto et al., 2015). These students attend schools where their native language (i.e., Spanish) is not spoken unless it is taught as a foreign language, and they are still in the process of learning English. Students classified as ELL are eligible to participate in programs designed to help them attain proficiency in English so that they can meet the academic standards all students are expected to master. Different labels are used to refer to these children in the U.S. (Garcia, 2009). We used the term ELL in this investigation as it was the term used in the school district where the study took place.
The purpose of this study was to examine the writing performance and writing motivation of students classified as ELL in the United States. National assessments (National Center for Education Statistics, 2011) and individual studies (e.g., Kuball & Peck, 1997;Neugebauer & Howard, 2015) have tested the writing performance of these students, but little research has examined their motivations for writing (Camping et al., in press). We were particularly interested in whether the writing performance and motivations of students classified as ELL were related to their gender and the grade they attended (e.g., grade three, grade four).

Theoretical Framework
The theory underlying our investigation was the Writer(s)-within-Community model (WWC; Graham, 2018aGraham, , 2018b. This model indicates that writers draw on oral language knowledge stored in long-term memory to carry out the processes involved in writing. More specifically, stronger oral language skills provide a better foundation for writing than still developing oral language skills (see Graham, Hebert, et al., 2020). This is because oral language production and writing rely on many of the same underlying knowledge resources (e.g., phonological, morphological, semantic, syntactic, and pragmatic knowledge) and apply similar production processes (e.g., generating and monitoring intentions and production, translating messages/ideas into acceptable language, and bringing this language to life by articulating it or putting it into written form). As a result, it is unlikely that students still acquiring a new language are proficient writers in that language.
The WWC model (Graham, 2018a(Graham, , 2018b also proposes that motivational beliefs are essential to writing. Motivation influences whether writers engage in writing, how much effort they commit, what actions and writing tools they apply, and how writers interact and collaborate with other members of the writing community. Within the WWC model, motivation involves one's intention to write. A writer can be more or less motivated to write a specific paper (i.e., state) or more or less motivated to write generally (i.e., trait). In the current study, we view writing motivation as a trait (habitual and stable intentions to write). It is further assumed that motivation is not a unitary trait because there are different majority of students agreed that intrinsic and extrinsic incentives drive their writing behavior, but only 38% of students indicated that self-regulation incentives had such an effect. Gender was not related to students' motivational scores, but scores for the three motivational incentives declined from lower to higher grades. Recommendations for future research and suggestions for classroom practice were provided.

Keywords
Extrinsic motivation, intrinsic motivation, self-regulatory motivation, writing, English language learners kinds and amounts of motivations . We further assumed that motivational beliefs operate independently while influencing each other (Cerasoli et al., 2014), and that motivations are best viewed as domain specific (Alexander, 2003).
In the WWC model (Graham, 2018a(Graham, , 2018b, motivation is promoted or restricted by a variety of beliefs writers hold in long-term memory. This includes beliefs about why one engages in writing , the importance of outperforming others and avoiding performing more poorly than others Elliott, 1999), the value and utility of writing (Eccles, 2005), attitudes and interest towards writing (Ekholm, Zumbrunn, De-Busk-Lane, 2017;Hidi & Boscolo, 2006), confidence in one's writing capabilities (Pajares, Miller, & Johnson, 1999), why one is or is not successful when writing (Weiner, 1985), and one's identities as a writer.
In the current study, we were interested in students' beliefs about why they write. We focused on three motivational incentives for writing Zimmerman & Risemberg, 1997): (a) intrinsic incentives, (b) extrinsic incentives, and (c) self-regulatory incentives. Intrinsic incentives involved writing because it is inherently satisfying or enjoyable. Our intrinsic motivation measure included items assessing curiosity (writing because of an interest in the topic) and involvement (writing because it provides a positive experience). Extrinsic incentives involved writing for external reasons. Our extrinsic measure included items assessing the influence of grades (writing to do better in school), competition (writing to do better than others), and social recognition (writing to receive praise). Self-regulatory incentives involved writing to regulate emotions and psychological traits. Our self-regulation motivation measure included items assessing emotional regulation (writing to cope with negative or unwanted emotions) and relief from boredom (writing to fill time or overcome boredom). Intrinsic incentives are moderately and positively related to literacy outcomes, whereas extrinsic incentives are not related or negatively related to literacy performance (Schiefele et al., 2012). The self-regulatory incentives assessed in our study were positively related to literacy outcomes in several prior investigations (Camping et al., in press;Schiefele & Schaffner, 2016).
Our focus on intrinsic, extrinsic, and self-regulatory incentives does not mean that the other motivational beliefs identified in the WWC model (Graham, 2018a(Graham, , 2018b are unimportant. We purposefully focused on these incentives because the other beliefs about writing identified in the WWC model (e.g., value/utility, attitudes) typically serve as antecedents, consequences, or both for these three incentives (see Schiefele & Schaffner, 2016;Schiefele et al., 2012). For example, the power of intrinsic motivators depends on antecedents such as competency beliefs ("I know how to write about this topic."); value, utility, and attitude beliefs ("Writing is important to me and I like to write."), and identity beliefs ("I am a really good writer."). Likewise, these same beliefs can be impacted by intrinsic motivators. For instance, intrinsic incentives are likely to increase how often one writes. The success of these new writing efforts may modify a writer's beliefs about writing competence, value and utility of writing, interest and attitudes toward writing, reasons for success, and identities as a writer.

Research Questions and Predictions
We addressed the following two research questions in this study: 1. Does the writing performance of students classified as ELL differ by gender and grade? 2. Do intrinsic, extrinsic, and self-regulatory motivational incentives of students classified as ELL differ by gender and grade? The writing performance measure administered to students in this study was the school district's annual writing assessment. In addition to providing an overall score for writing quality, the test included benchmarks allowing us to determine if students evidenced proficiency, partial proficiency, or minimal proficiency in writing at their grade level. We anticipated that few of the participating grade three to eight students classified as ELL in this study would evidence grade-level writing proficiency for two reasons. One, as noted earlier, it is unlikely that students still acquiring a new language will demonstrate proficiency in writing in that language (Graham, 2018a(Graham, , 2018b. Two, on national assessments in the U.S. (National Center for Education Statistics, 2011) and in individual studies (e.g., Kuball & Peck, 1997;Neugebauer & Howard, 2015;O'Conner, Abedi, & Tung, 2012), students classified as ELL evidenced below grade-level English writing.
We also anticipated that girls would be better writers than boys. While the mechanisms underlying gender differences are not fully understood (Pyne, 2020), girls generally score higher on writing measures than boys in studies conducted in the United States (National Center for Education Statistics, 2011;Reilley et al., 2019). We further expected that students in later grades would evidence higher writing scores than students in lower grades, because they had more opportunities to strengthen their writing competence over time.
Previous research has shown that students classified as ELL view the writing they do at school in a neutral or negative light, especially if such writing was not shared with others (McCarthey & Garcia, 2005;Rueda & Moll, 1994), but they are more positive about the writing they do outside of school (Sturtevant & Kim, 2009). The motivation measure used in the current study did not reference writing at school except when asking about extrinsic incentives (specifically items asking about grades and competition). Because of these findings and the nature of the motivation measure applied in the current study, we made no predictions about whether the participating students would agree or disagree that intrinsic, extrinsic, and self-regulatory incentives drive their writing behavior.
We also did not make predictions about the relationships between grade-level and each of the three motivational incentives. In previous studies (see Graham, 2006), data on whether writers become more motivated over time was mixed, with some studies finding that motivation increases with age and others finding that it declines. We did, however, anticipate that girls would be more motivated writers than boys. Given the consistent findings that girls are better writers than boys (National Center for Education Statistics, 2011;Reilley et al., 2019), it is reasonable to expect that girls are more motivated writers as well.
An unusual feature of the present investigation was that the writing test and the motivational assessment were administered to all students in grades three to eightas part of the participating school district's regular practices. Grade one and two students were not administered the motivational assessment because it was assumed that children this young could not accurately gauge their motivational incentives to write. Students beyond grade eight were not included because high school students did not attend this school district. This unique feature of the study allowed us to include almost all third to eighth grade students classified as ELL in the school district. We were unable to include, however, students who were absent the day the measures were administered. We purposefully decided not to include students classified as ELL with a disability in the analysis. Our goal in this investigation was to focus on typically developing youngsters.

Setting
This investigation took place in an urban school district in the Southwest United States. The school district enrolled over 11,000 students at the time the study took place, and 51% of students were Hispanic, 21% White, 12% Black, 7% American Indian or Alaska Native, 2% Asian, and 1% Pacific Island. Three out of every four students in the school district were eligible for free or reduced-fee lunch.
The school district was diverse with over 74 different languages spoken by students. For students whose native language was not English and proficiency in this language was still developing, the school district employed a sheltered English language immersion model to provide English instruction to these students. The school system labeled these youngsters as English language learners (ELL). The goal of the sheltered immersion program was for students to acquire English as rapidly as possible so they could participate fully in regular classroom instruction which was provided in English. The program provided four hours of instruction daily in speaking, listening, reading, and writing in English. The program was delivered by teachers who held an endorsement such as a structured English immersion, bilingual education, or second language teacher.
Students were enrolled in the sheltered English language immersion program when their parents indicated to the school system that a language other than English was spoken in the home. Students' proficiency in English was then evaluated using the Arizona English Language Learner Assessment. This test assessed English proficiency in speaking, listening, writing, and reading. Coefficient alpha for the test was high (0.91; Arizona Department of Education, 2016). If a student scored below the proficient level in English on this test, the parents were notified. Parents had to provide consent for their child to participate in the sheltered English language immersion program. Students' English proficiency was tested yearly, and they exited the program once they scored at the proficient level.

Participants
A total of 880 students in grades three to eight classified by the school district as ELL participated in this study. The students were drawn from 22 different schools, and they constituted virtually all youngsters in the school district who had been classified as ELL. It did not include, however, students receiving services for special education or students who were absent the day the district writing test and the writing motivation measure was administered.

Measures
District writing test. For the district writing test, all students were asked to write an informative/ explanatory essay on informational technology. Students were told they had as long as they needed to complete their essay, except that it had to be written before the school day ended. All essays were scored on three dimensions: purpose/focus and organization (4 possible points), evidence and elaboration (4 possible points), and conventions (2 possible points). The three dimensions were summed to provide a total score for the test, with a score of 8 to 10 indicating writing proficiency at grade-level; 5 to 7 representing partial writing proficiency at grade-level; and 0 to 4 signaling minimal proficiency at gradelevel. Scoring reliability (based on kappa) for the three dimensions were at acceptable levels, ranging from 0.77 to 0.85 for the three dimensions. Motivation measure. The Writing Motivation Survey included 28 items that measured three primary motivational constructs (Graham, Harbaugh, et al., 2020). This included intrinsic motivation, which included items assessing curiosity ("I write because I can learn about things that interest me."; "I write because I like to think about particular topics."; "I write because I can write about topics interesting to me."; "I write because I can write about topics important to me.") and involvement writing incentives ("I write because I like to create a character that I can identify with."; "I write because it allows me to imagine everything so well."; "I write because I can create and experience adventures in my mind.").
It also included extrinsic motivation, which included items assessing grades ("I write in order to get better grades at school."; "I write because it helps me get better in school."; "I write because it helps me perform well in school."; "I write because it is important to how well I do at school."), competition ("I write because it is important for me to know more than other students."; "I write because it is important to me to write better than other students."; "I write because it is important to me to be among the best students."; "I write because it helps me perform better in school than my classmates."), and social writing incentives ("I write because I know that my friends write a lot."; "I write because one gets praise for writing well."; "I write because I like it when other people think I am a good writer.").
It further included self-regulation motivation, which included items assessing boredom ("I write in order to avoid being bored."; "I write because it helps me pass the time."; "I write is there is nothing better to do.") and emotional regulation incentives ("I write because it cheers me up when I'm in a bad mood."; "I write because it helps me calm down."; "I write because it makes me feel better."; "I write so that I can think about something that bothers me.").
Students responded to each item on the survey by choosing one of four options: Very True, Mostly True, Sometimes True, or Not True At All. These were scored from one to four for each item, and then reverse coded for analysis so that higher scores indicated higher motivation. Coefficient alphas for intrinsic, extrinsic, and self-regulation motivation for students in this study were 0.81, 0.83, and 0.81, respectively.

Procedures
In November, the school district administered the district writing test and motivation measure to all students on a single test day. The Writing Motivation Survey was administered first, followed by the district writing test. Both tests were administered as part of normal school procedures.

Analysis
To examine if there were gender and grade differences in the writing and motivation of students 7 Steve Graham et al. who were designated as ELL, we conducted four 2 (gender) X 6 (grade) Analysis of Variance tests (ANOVA). One ANOVA was conducted for the district writing test (research question 1), and the other three ANOVAs focused on one of the three motivational constructs (research question 2). We applied a Bonferroni adjustment and set the significance level at p < .0125 (.05/4). Table 2 presents means and standard deviations by gender and grade for district writing scores. The average scores for this test ranged from 3.78 in grade three to 5.98 in grade seven. Fifty-one percent of students' scores indicated minimal mastery of writing at grade-level, with another 45.7% of scores representing partial mastery of writing at grade-level.

District writing test
The main effects for gender (F (1, 868) = 19.33, p < .001) and grade (F(5, 868) = 37.45, p < .001) were statistically significant. Girls had higher scores (M =4.64, SD = 1.50) on the district writing test than boys (M =4.19, SD = 1.41). Bonferroni post hoc analysis (all p's < .05) revealed that grade eight students scored higher than grade three to seven students. Grade seven students scored higher than grades three and four students. Grade five students scored higher than grade six as well as grades three and four students, whereas grade four students scored higher than grade three students. The interaction effect for the 2 X 6 ANOVA was not statistically significant. Table 2 presents means and standard deviations by gender and grade for intrinsic motivation. Sixtyone percent of the ELL students felt that it was true or mostly true that intrinsic incentives (curiosity and involvement motivation incentives) drive why they write.

Intrinsic incentives
The main effect for grade (F(5, 868) = 31.25, p < .001) was statistically significant. Bonferroni post hoc analysis (all p's < .02) revealed that grade eight students scored higher than grade three to seven students. Grade three students scored higher than all grades except grade five students. Grade four students scored higher than grade seven and eight students, whereas grade five students scored higher than grades six to eight students. Grade six students scored higher than grade eight students. The main effect for gender and the interaction between gender and grade for the 2 X 6 ANOVA were not statistically significant.

Extrinsic incentives
The means and standard deviations by gender and grade for extrinsic motivation are provided in Table 2. Fifty-three percent of the ELL students felt that it was true or mostly true that intrinsic incentives (grades, competition, and social incentives) drive why they write.
The main effect for grade (F(5, 868) = 27.78, p < .001) was statistically significant. Bonferroni post hoc analysis (all p's < .04) revealed that grade three students scored higher than grade four to eight students. Grade four students scored higher than grade six to eight students, whereas grade five students outscored grades seven and eight students. Grade six students had higher scores than grade eight students. The main effects for gender and the interaction between gender and grade for the 2 X 6 ANOVA were not statistically significant.

Self-regulation incentives
Lastly, Table 2 includes means and standard deviations by gender and grade for self-regulation motivation.

NS
Note: District Writing score range was 0 to 10. Scores for Intrinsic incentives ranged from 7 to 28 (1 to 4 for 7 items). Scores for Extrinsic incentives ranged from 11 to 44 (1-4 for 11 items). Scores for Self-regulation incentives ranged from 7 to 28 (1 to 4 for 7 items). Higher scores represent greater agreement about the importance of the incentive. *p < .05, **p < .01, ***p < .001, NS = not significant. There were no statistically significant interactions between grade and gender. Steve Graham et al. Thirty-eight percent of the ELL students felt that it was true or mostly true that self-regulation incentives (boredom and emotional regulation) drive why they write.
The main effect for grade (F(5, 868) = 31.25, p < .001) was statistically significant. Bonferroni post hoc analysis (all p's < .05) revealed that grade three students scored higher than grade four to eight students. Grade four students scored higher than grade six to eight students, whereas grade five students outscored grades seven and eight students. The main effects for gender and the interaction between gender and grade for the 2 X 6 ANOVA were not statistically significant.

Discussion
The current investigation examined the writing and writing motivation of students in grades three to eight who had been classified as ELL by the participating school system and were eligible for English language services. We tested whether the writing performance and writing motivation of these students differed by gender and grade.

Writing of students classified as ELL
As expected, very few of the students in this study met grade-level expectations on the standardized writing assessment administered by the school district. Three percent of the students classified as ELL scored at the proficient-level. The scores of 46% of the students indicated partial mastery of gradelevel writing expectations, with the remaining 51% of students evidencing only minimal mastery of these objectives. These findings are consistent with outcomes from earlier experiments where students classified as ELL had lower writing scores than their native English-speaking classmates (e.g., Kuball & Peck, 1997;Neugebauer & Howard, 2015;O'Conner, Abedi, & Tung, 2012).
The relatively low English writing performance of students classified as ELL supports the theoretical proposition that writers draw on their knowledge of oral language to write (Graham, 2018a(Graham, , 2018bGraham, Hebert, et al., 2020). If this proposition is valid, students with stronger oral language skills should produce qualitatively better writing than students with less developed oral language skills. This was the case in the current study, as students still acquiring competence in a new language were not proficient when writing in that language.
Additional research is needed to replicate and extend this finding with students across a broader array of grades and writing measures. It is important to document the oral language skills of students classified as ELL as well as establish the association of these language skills with their respective writing skills. Such research will provide a more nuanced picture of the relationship between oral language and writing.
As predicted, girls classified as ELL had higher scores on the writing assessment than boys with this same classification. This finding is consistent with previous research showing that girls in the U.S. are better writers than boys (National Center for Educational Statistics, 2011;Reilley et al., 2019). Additional research is needed to determine the factors that contribute to gender differences in writing for native language speakers and for students who are still acquiring the language used at school. It is particularly important to examine if such differences occur across different languages and cultures.
Also consistent with our expectations, students classified as ELL generally had higher writing scores in later than earlier grades. With the exception of writing scores in grades five and six, students evidenced increasingly stronger writing performance across the six grades. Writing scores in grade five exceeded writing scores in grade six and were similar to writing scores in grade seven, whereas grade six writing scores did not exceed writing scores in grades three and four. It is not completely clear why the writing scores of students in grades five and six did not follow our predicted progression, although it is possible that this was a consequence of the cross-sectional nature of the investigation. Future research is needed to examine the writing development of students classified as ELL longitudinally.

Writing motivations of students classified as ELL
An especially noteworthy outcome was that 61% of the students classified as ELL in this study believed that intrinsic incentives drive their writing behavior. Previous research has shown that intrinsic beliefs are positively related to students' literacy outcomes (Schiefele et al., 2012). In contrast, only 38% of these same students believed that self-regulatory incentives played an important role in their reasons for writing.
It is possible that a majority of students placed less emphasis on self-regulatory incentives because we asked them only about using writing to combat boredom or control emotions. If students answered the self-regulation incentives items by focusing specifically on the school context, where the survey was administered, they may have minimized the importance of combating boredom and controlling emotions because it is likely they are provided few opportunities to do so during the school day. Additional research is needed to replicate our findings regarding intrinsic and self-regulation writing incentives for students identified as ELL, explore other self-regulation incentives for writing (e.g., environmental structuring), and to examine motivational incentives in-school and out-of-school contexts.
We further found that slightly more than one-half of the students in this study (53%) believed that extrinsic incentives drive their writing behavior. Because prior research found that extrinsic motivators are negatively or not related to students' literacy outcomes (Schiefele et al., 2012), future research needs to determine why some students identified as ELL emphasize extrinsic motivators. It is also important to directly examine if the effects of extrinsic incentives on the writing of students identified as ELL are positive, negative, or neutral.
Contrary to our predictions, we did not find gender writing motivational differences between boys and girls. It is unclear why gender effects were not obtained, as girls are generally better writers than boys (National Center for Educational Statistics, 2011;Reilley et al., 2019), and this was the case in this study. Presumably this would increase the likelihood that girls would be more motivated writers. Possible differences in the writing motivational incentives of boys and girls requires careful examination in future studies.
Unfortunately, there was a declining trend by grade in scores for intrinsic, extrinsic, and selfregulatory incentives. As a result, students classified as ELL in this investigation placed increasingly less emphasis from grades three to eight on these motivators to write. Research is needed to replicate these findings, and if replicated, to explore why these motivators become less prominent reasons for writing. We also need to explore the consequences of these declines in motivation.

Conclusion
The current study demonstrated that upper-elementary grade and middle school students identified as ELL in the U.S. are not likely to be proficient writers when composing in English. Even so, a majority of the students in this study emphasized the role of intrinsic motivators for writing. It is important that schools emphasize instruction that improves these students' writing, while taking advantage of their interest to write because it is an inherently satisfying and enjoyable activity. Moreover, a sizable minority of students in our study emphasized that they write to avoid boredom and to regulate their emotions. For these students, teachers should leverage these beliefs by allowing them to use writing for these purposes at school. We think it is also important for teachers to be sensitive to gender differences in the writing of students identified as ELL, and to combat the decline in students' motivations for writing.
Finally, slightly more than one-half of participating students' emphasized the role of extrinsic motivators in their writing (competition, grades, and social comparison incentives). While it is possible that such incentives can have a positive effect, there is evidence that they can negatively impact students' literacy performance (Schiefele et al., 2012). We encourage teachers to pay close attention to the impact