Teachers’ perceptions of effective school-wide programs and strategies for English language learners

Abstract

Data suggest that teachers lack confidence in teaching English language learners (ELLs). Teachers’ perceived shortcomings have been identified but little is known about why these perceptions exist or how they might be associated with broader measures of efficacy. We surveyed 578 Californian teachers of ELLs to explore the relations among teachers’ perceptions of ELL strategies and school programs, teachers’ ratings of collective efficacy in meeting all students’ needs, and the general school climate. Descriptive data and two multiple regression analyses suggest two primary findings: teachers’ perceived collective efficacy for English language development (ELD) instruction was higher than their individual efficacy (relative to previous studies), and measures of general school-wide collective efficacy when associated with perceived strength of ELD practices and programs. These results suggest that school-wide reforms designed to improve ELL instruction might yield greater collective efficacy. Policies that could enhance teachers’ perceived ELD efficacy are discussed.

Introduction

Improving the education of the 3.5 million English language learners (ELLs) attending U.S. schools is a national educational priority. For many years, ELL achievement has been largely ignored; even the collection of national and state data by language status has been sparse. Despite the dearth of data about their educational successes and shortcomings, Latino students constitute the largest group of ELLs (80–85 %) and thus serve as a proxy for the assessment of ELL achievement (see Téllez 2010 for a review). Of all ELL groups, Latinos have the lowest levels of education and the highest dropout rate. Like other ‘minority’ groups in the United States, Latinos score well below their dominant culture counterparts. Data from National Center on Educational Statistics indicate that Latinos generally score approximately a standard deviation lower on academic achievement tests than do their White counterparts (Rumberger and Gándara 2004).

Overall, fourth- and eighth-grade students’ scores on the National Assessment of Educational Progress assessment increased in 2009 [U.S. Department of Education (Hemphill & Vanneman 2011)]. In contrast, Latinos’ scores continued to be among the lowest and, from 1990 to 2009, the gap in the test scores for Whites and Latinos did not change significantly ((Hemphill & Vanneman 2011). In 2009, these test results indicated that 44 % of White students were at or above proficient, compared to 17 % for “Hispanic” students in eighth-grade mathematics and reading, while 17 % of White and 43 % Hispanic students at that same grade level were below basic proficiency (Aud et al. 2010).

Schools and school systems have employed various professional development strategies designed to improve English language development (ELD) teachers’ instructional capacities and thereby improve educational achievement of ELLs, especially Latino ELLs (Hart and Lee 2003; Téllez and Waxman 2006). Despite these efforts, research indicates that teachers continue to doubt their individual skills and capacities for working with ELLs (Alexander et al. 1999; Gándara et al. 2005). To date, studies of teacher knowledge and skills for ELLs have used teacher self-ratings. We believe the theoretical frame for these studies should be recast in the light of teacher efficacy, which in a topic of considerable attention from researchers who study teacher capacities (Tschannen-Moran and Hoy 2001).

Briefly defined, self-efficacy is a largely cognitive process in which individuals generate beliefs about how their persistence, response to potential failure, and coping strategies affect their performance on a certain task or tasks (Bandura 1997). Building on Bandura’s (1993) concept of self-efficacy, research on teacher efficacy has shown a link between teachers’ beliefs about their own pedagogical skill and their actual performance, although the particular cause-and-effect relations remain equivocal (Washburn 2006). Scholars have recently turned their attention to teacher collective efficacy, which Goddard et al. (2004a) defined as teachers’ judgement of their capacities to organise and execute the courses of action required to have a positive effect on student learning. As Bandura noted, “People do not lead their lives in individual autonomy. Indeed, many of the outcomes they seek are achievable only through interdependent efforts” (Bandura 2000, p. 75).

Promising research on collective efficacy fed our interest in collective efficacy of teachers working in the interest of ELLs. We believe that this new focus is warranted for several reasons. First, teachers’ collective efficacy appears to be more closely related to actual student performance than are measures of individual teacher efficacy. One study revealed that schools with high collective efficacy were able to overcome the normal effects of poverty (Hoy et al. 2002). Second, effective programs in ELD have a positive impact on ELL achievement. School-related variables, such as (a) staff capacity to address ELL needs, (b) school-wide focus on ELD and standards-based instruction, (c) shared priorities and expectations of educating ELLs and (d) systematic, ongoing assessment and data-driven decision making, improve ELL achievement even more than does the particular language program implemented (Parrish et al. 2006).

Third, studies involving theories of human action and cognition have overwhelming concluded that engagement with complex tasks is best understood using a socio-cultural lens. Theories of teacher learning have drawn our attention to the social nature of teachers’ work. For instance, Lave and Wenger’s (1991) overarching theory plays a prominent role in contemporary literature on growth in professional learning, and teacher educators have used their work to describe how teachers best learn new ways of instruction (Lieberman and Wood 2003). Finally, school systems have embraced professional learning communities as the preferred—and sometimes only—method for professional development among teachers. This movement is based loosely on the socio-cultural approach, but has had a much wider influence (Talbert 2009).

Growing evidence suggests that instructional improvements are best realised when the school, rather then the individual teacher, is the focus for research and policy on teaching efficacy to support enhanced student achievement. We argue that this orientation is important in the research on teacher capacity for ELD instruction.

Collective efficacy

Our review of the notion of collective efficacy relates especially to schools as the unit of study. Bandura’s (1993) enduring theory of human agency has influenced a wide range of social sciences. He used the term ‘efficacy’ to describe the belief that people have about their own capabilities. Those with a strong sense of self-efficacy view difficult tasks as challenges that can be overcome if they set goals and stay focused, keep going when obstacles are encountered, and believe that additional knowledge and skills can make success more likely. Bandura’s (1997, 2000, 2001) more recent work suggests that self-efficacy is, in fact, a social context in which groups of people can share a sense of purpose in their work towards common goals. This shared belief in individuals’ collective power “to produce desired results is a key ingredient of collective agency” (Bandura 2001, p. 14). Collective performance is founded on collective efficacy at the group level; it is more than the sum of individual members’ efficacy beliefs.

Goddard et al. (2004b) suggested that Bandura’s theory of collective efficacy can be applied to the group effort and energy found in a school, going so far as to suggest that increasing collective self-efficacy in a school results in higher student achievement.¹ To us, elementary school serves as an ideal laboratory for collective efficacy. First, it is possible for elementary school teachers to conduct their work without much collective agency. The culture in some elementary schools promotes a ‘close the door and teach’ standard; teachers do not work together in any coordinated fashion (Flinders 1988). As a consequence, collective efficacy is of interest to researchers by virtue of its variability alone. Second, teachers in an elementary school are bound to a common mission, even if they choose to work alone. A vertical curriculum, a set of students and families that teachers will share across years and, for better or worse, accountability schemes such as No Child Left Behind that aggregate scores at the school level all point to the need for elementary teachers to work together to improve their students’ academic achievement.

Another factor related to the development of efficacy is the construct of trust, which Hoy and Tschannen-Moran (1999, p. 189) defined as “an individual’s or group’s willingness to be vulnerable to another party based on the confidence that the latter party is benevolent, reliable, competent, honest, and open”. Goddard et al. (2001) concluded that trust has an influence not only on how willing teachers and students are to take risks, but also on performance and effort. Specifically to the school context, Goddard et al. argued that the larger the proportion of poor students in the school, the lower the level of collective trust and, therefore, efficacy among teachers. Having worked in many schools where poverty and teacher efficacy existed side by side, we are somewhat dubious about this claim but, on the off chance the assertion held validity, we wondered if the same were true for the proportion of ELLs. Would a high proportion of ELLs in either the school or classroom reduce (or perhaps increase) collective efficacy in either the strategies or program of a school? We specifically investigated this question in our study.

Collective efficacy is also related to factors such as the capacity of a school to focus on joint tasks, teachers’ recognition of each other’s work, and the ability to achieve goal consensus (Kurz and Knight 2004). In the following section, we explore the research on effective schools for ELLs.

Research on effective schools for English language learners

Interest in the education of ELLs is increasing. Much like researchers of the 1960 s who believed that poverty was a principal cause of school underperformance, we argue that many contemporary policy makers and researchers now view the need to learn English as a daunting but important hurdle to academic performance. Lawsuits that waged war on poverty in previous decades are now fought on behalf of ELLs (Rumberger and Gándara 2004). These critiques generally consider the school as the unit of analysis, thereby focusing on school-wide reforms that provide a high-quality education for ELLs.

Pivotal research on the relationship between school climate and student learning (e.g. Brookover et al. 1978) showed that the culture that teachers and school leaders create influences academic performance. Extensive work on classroom learning environments by Fraser (1991) demonstrated the association between effective classroom practices and student learning and satisfaction. Thus, it is reasonable to suppose that improving the culture of a school and the quality of classroom instruction will enhance the educational outcomes of ELLs. Indeed, we might anticipate that ELLs, who represent the children of immigrants, could benefit most from improved school culture and classroom practices. Perhaps more than any other subgroup in U.S. schools, ELLs have the greatest potential to grow academically. This very proposition was made by the progressive educator Counts (1922/2010) nearly 100 years ago, when he found that schools were not having disproportionate, positive success with their immigrant students. Unfortunately, Counts’ findings have not been significantly altered (Téllez 2010).

Because ELLs attend primarily segregated schools (i.e. schools serving predominantly low-income, Latino students), we might imagine little variation in the effectiveness of their schools but, in fact, we find wide differences. Such variance might be the result of collective efficacy, which is related to the educational resiliency that a school can promote (Waxman 1992). Studies have shown wide variation in classroom instruction and in classroom learning environments between effective and ineffective schools for ELLs (Waxman et al. 2004). In particular, schools successful in teaching marginalised students such as ELLs are characterized by teachers’ involvement in their jobs, a high level of teacher cohesion, and strong administrative support.

Many of the effective schools recognise ELLs’ unique needs that require more personal attention from teachers and have developed clusters of teachers who work with a particular group of mainstream students (Ancess 2003; Minicucci et al. 1995). López et al. (2001) found that this idea of personal attention often takes the form of home visits, parent education and providing school supplies to families (see also Maden 2001). In a study germane to our work, Guo et al. (2010) examined the relationships among preschool teachers’ self-efficacy, classroom quality and children’s gains in early literacy skills. Teachers’ self-efficacy and classroom quality predicted children’s gains in print awareness but not vocabulary knowledge. A complex set of interactions among the variables suggests that self-efficacy and instructional quality are subtly related. Although Guo et al. assessed individual teacher efficacy, they demonstrated the important role that teacher efficacy can play in literacy development, a chief goal for teachers of ELLs.

One as yet unstudied variable in relation to school effectiveness for ELLs is teacher experience, which has been associated with higher academic performance, although findings are not consistent across subjects and grade levels (Croninger et al. 2007). In any case, we were interested in whether teachers’ ratings of collective efficacy would vary by experience.

Curriculum and instruction in effective schools are built upon collaborative efforts among the teachers and site leaders (Miramontes et al. 1997; Rosebery et al. 1992). Within the school, teachers plan units, lessons and activities that take language teaching into consideration (Ancess 2003; Miramontes et al. 1997; Rosebery et al. 1992). For ELLs, professional development for teachers in successful schools centres on students’ needs and academic learning (Mora 2000).

In the main, the research on effective schools for ELLs demonstrates that high-quality leadership is crucial for success, a focus on standards-based literacy is required for increased student achievement, and successful schools devote time to ELD, particularly with respect to literacy (Williams et al. 2007). This latter point is one that we take up in the Results section, when we consider how collective efficacy might be closely related to effective ELL instruction.

Based on our review of previous research and our own experiences working with schools whose teachers strive to improve the academic achievement of their ELLs, we developed two specific research questions to explore the relationships among ELL teaching and collective efficacy:

• 1. Are teachers’ ratings of the effectiveness of their school’s ELL strategies related to measures of collective efficacy, school climate, teacher experience and the proportion of ELLs in schools and classrooms? If an overall relationship can be found, which factors are most related to ratings of effective ELL strategies?

• 2. Are teachers’ ratings of the effectiveness of their school’s ELL programs related to measures of collective efficacy, school climate, teacher experience and the proportion of ELLs in schools and classrooms? If an overall relationship can be found, which factors are most related to ratings of effective ELL programs?

We chose these two key measures because we wanted to understand how collective efficacy and other variables differ in terms of a measure of effectiveness largely under the control of teachers (strategies) and a measure not typically under their control (program).

Methods

Procedure

Data for this study were obtained from a survey of teachers representing a volunteer sample of elementary schools throughout California. The sample was chosen by inviting all elementary schools with 300 or more students in a large county in California to participate. All members of the Association of California School Administrators Elementary Education and Curriculum, Instruction, and Accountability Councils were also asked to invite elementary schools in their region to participate. By way of compensation, participating schools were given a school licence for the Association of California School Administrators product, The Standard Finder™, a database of the California Content Standards linked to state assessment blueprints.

School principals were offered one of two methods for administering the survey. Either the researcher(s) would administer and collect all surveys during a staff meeting held at the school or the surveys could be distributed at a staff meeting at which the researcher was not present; if the latter option was chosen, principals were instructed to advise their teachers to return the completed survey to a teacher at the school who agreed to be responsible for mailing the surveys (in a stamped self-addressed envelope) to the researchers. Based on principals’ choices, surveys were administered by the researcher(s) at seven schools; at all of the remaining schools in the county, the principals distributed the surveys and had them collected and returned by a teacher.² Surveys were completed using an optical character recognition form and scored electronically. All surveys, as well as the database of survey results, were visually scanned for response inconsistencies.

Participants

Principals of 50 elementary schools³ agreed to participate, yielding a full dataset of 950 teachers. The present study limited this dataset to those teachers whose classrooms had at least 10 % ELLs,⁴ resulting in a subsample of 578 teachers. Teachers in this study were asked to share their years of experience at their school and in the profession (these data are reported in our findings), as well as their gender (88 % are women).⁵

Participants in the study also completed the Collective Efficacy Scale (Goddard 2002) and a measure of school climate (Marzano et al. 2005). Finally, they rated their perceptions of the effectiveness of the ELD instructional strategies of their school, as well as the success of the programs for ELLs.

Measures

This study employed the 12-item Collective Efficacy Scale developed by Goddard (2002; see “Appendix 1”). This scale, which evolved from a 21-item survey developed by Goddard et al. (2000a), represents an attempt create a more efficient measure with a better balance of variables inside and outside the control of the group (Goddard 2002). These two halves are task analysis (TA), which measures the constraints and opportunities embedded in teaching students (Goddard 2002), and group competence (GC), which assesses judgements about the capabilities that a school staff bring to bear upon teaching tasks. For example, the statement “These students come to school ready to learn” is an example of a TA item, whereas “Teachers in this school are able to get through to difficult students” is an example of a GC item (Goddard 2002). Cronbach’s alpha reliability was calculated for each scale (TA = 0.84; GC = 0.75).

We also administered a measure of school climate (Marzano et al. 2005) using three of the survey subscales that assess school focus, culture and affirmation.⁶ The following items comprise each subscale, and Cronbach’s alpha was computed for each subscale.Items for Focus construct (Cronbach’s alpha = 0.89):

• 1. Concrete goals for achievement have been established for students at this school.

• 2. In my school, we have designed goals for our curriculum.

• 3. We have specific goals for instructional practices at my school.

• 4. We have established specific goals for the assessment practices in my school.

Items for culture construct (Cronbach’s alpha = 0.83):

• 1. Teachers in my school regularly share ideas.

• 2. There is a strong spirit in my school.

• 3. In my school, we have a common language that is used by administrators and teachers.

• 4. In my school, we have a shared understanding of our purpose.

• 5. In my school, we share a vision of what we could be like.

Items for Affirmation construct (Cronbach’s alpha = 0.77):

• 1. The accomplishments of individual teachers are recognized and celebrated.

• 2. The accomplishments of the students and the school in general are recognized and celebrated.

• 3. At this school, we systematically acknowledge our failures and celebrate our accomplishments.

The following items represented the teachers’ views about the instruction for ELs: ⁷, ⁸, ⁹

• 1. In my school, teachers use effective strategies for ELLs.

• 2. Our school program for ELLs is working well.

All items used for a 4-point scale (i.e. 1 = Strongly Disagree; 4 = Strongly Agree). Some items were reverse-coded and required recoding.

We also developed an interview protocol (see “Appendix 2”) and gathered data from approximately 20 teachers.¹⁰ We conducted interviews to generate qualitative data to augment our quantitative findings but did not consider them to be specific tests of our research questions.

Results

Teachers rated the ELD strategies at their school as being highly effective (3.16/4). They also rated the ELD programs positively (2.73/4). Ratings on collective efficacy were divided between two constructs assessed on the 12-item survey. GC measured factors internal to the efficacy of the school, while TA measured factors beyond the control of the school. Data indicate the teachers generally rated their efficacy as positive. Of the subscales derived from the climate survey, the Affirmation subscale was rated less well than the other two subscales. Details of teachers’ responses are presented in Table 1.

Table 1 Correlation matrix (zero order), means, and standard deviation for variables included in both models

Multiple regression analyses

We used two ordinary least squares regression models to study the effects of individual variables on both ratings of EL strategies and EL programs. Model A used the study variables to predict teachers’ ratings of the use of effective strategies in ELD in their school. The overall model was significant (F = 26.20, p > 0.001; adjusted R ² = 0.275). Significant variables in the model included the school climate measures of focus (β = 0.266, p < 0.001), culture (β = 0.196, p < 0.05) and GC (β = 0.193, p < 0.001). The TA scale was also significant (β = −0.113, p < 0.05) but was negatively related to effective practice, suggesting that teachers who believe that their school promotes good instruction also believe they can work to overcome student and family characteristics beyond their control. Results for the contributions of individual variables to the overall regression model are presented in Table 2.

Table 2 Individual variable contributions to overall regression model, Model A

Model B used the study variables to predict teachers’ ratings of the use of effective strategies in ELD in their school. The overall model was significant (F = 18.22, p > 0.001; adjusted R ² = 0.206). Significant variables in the model included the school climate measures of focus (β = 0.205, p < 0.05), culture (β = 0.154, p < 0.05), and GC (β = 0.111, p < 0.05). Unlike in Model A, the TA scale in Model B was not significant (β = −0.091, p > 0.05). For details, see Table 3.

Table 3 Individual variable contributions to overall regression model, Model B

We conducted interviews with randomly-selected study participants. Teachers were chosen to be interviewed by stratifying the sample by the percentage of ELLs in the teacher’s class, and then selecting three teachers at random from each strata or quartile. These data add a corollary analysis to the research, which we explore in the Discussion section.

Discussion

First, we found that teachers in this sample were quite confident in the ELD strategies used at their school. This finding contradicts previous studies, which showed that individual teachers rate their capacity to provide high-quality instruction for their ELLs as low. The teachers in this sample are less sure about their confidence in the ELD program of their school, but were positive nevertheless.

The regression models demonstrate that teacher efficacy for ELLs is related to group efficacy and other school climate measures, while controlling for the level of ethnic diversity of the school, the percentage of ELLs in teachers’ classrooms or the school, or teacher experience at a specific school. The climate survey subscale measuring school focus accounted for a significant proportion of the variance in the group efficacy for ELL strategies. This finding suggests that ELL strategies were considered as part of the overall goals for instructional practices at schools. The climate variable measuring culture was also associated with ratings of effective practice. The Affirmation subscale, which assesses whether a school routinely recognises individual achievements, however, was not found to be related to effective practice for ELLs. These data suggest that teachers who work in schools with a positive climate developed a strong belief in the ELD instruction program at their school. With respect to collective efficacy, the GC scale was associated with perceived strength of both strategies and program. This finding suggests that schools with strong collective efficacy tended to also have strong perceived ELD practices. The TA scale was also associated with perceived strength of ELD instruction but the negative beta weight indicates an inverse relationship. This finding is not surprising because the TA portion of the Collective Efficacy Scale measures factors outside the control of the school staff (e.g. “Students at this school come to school ready to learn”).

Equally important are those variables found not to be significant in both models. First, collective efficacy in schools has been shown to decrease with the number of poor students (Goddard et al. 2001). Although our study did not measure individual student-level variables, we might consider the percentage of ELLs in a school or a teacher’s classroom as a representation of poverty; national data show that more than 80 % of all ELLs participate in the federal free and reduced-price lunch program (Téllez 2010), but our data indicated no significant variation explained by the percentage of ELLs in either the school or a teacher’s classroom. Similarly, we might have also predicted that schools and classrooms filled with ELLs would perceive with less confidence their strategies and programs for ELD, but this was not the case. The concentration of ELLs in either the school or individual teacher’s classroom did not influence perceptions of the ELD program overall. Teacher experience, typically an important variable in teacher efficacy, was not associated with perceptions of effective ELD strategies or programs.

Perceptions of collective efficacy are important because recent research suggests that program coherence and clarity of goals at the school level are the most important factors in a successful language education program. Program coherence and clarity of goals are more important than whether the program is based on a bilingual model or one based on English instruction alone (Parrish et al. 2006). This finding suggests that a teacher’s individual assessment of skill or preparedness for working with ELLs is less important than the collective assessment of capacity at the school for language instruction, although we recognise that they are clearly associated.

We also conducted a smaller corollary investigation of teachers’ views of their work with ELLs by conducting a random sample of teacher interviews. Using the protocol presented in “Appendix 1”, interviews focused primarily on the teachers’ views of the intersection of professional development and student achievement. Even though we did not specifically invite responses regarding ELD instruction, at least 30 % of the teachers interviewed referenced the professional development for ELLs at their school. Using a standard coding scheme that we connected to our quantitative analysis (see Driscoll et al. 2007), we found that many schools were asking teachers to engage in professional learning community (PLC) time, during which ELD was a primary focus. The following are representative samples of themes found in the full corpus of interview data. One teacher reported that

Every Wednesday, we have our PLC time and we discuss different strategies—our focus is language arts, reading because we are a Decile 1 school and that’s one of our biggest areas. Plus, our population being pretty much English learners, so we do meet on a regular basis.

Another teacher reported that efforts of the PLC committees were not necessarily useful in promoting more productive ELD instruction:

We have consultants that [sic] do ELD training—we had a lot of ELD training because of our population. Then you have the PLC committee—and, you know, just whatever they want to throw at you. So that is very distracting.

Still other teachers revealed a group commitment to the task of better serving ELLs:

They [ELLs] have some added challenges because there is so much text, reading of text, involved in math—so we’re dealing not only with the concept but we’re dealing with them being able to read the text. So that’s very tricky and challenging but, overall, that’s what we want [emphasis added].

Another teacher remarked:

I’d want you to see how we are approaching ELD in our district or how we have, in the past couple of years, been learning to approach it. Looking at how we use writing to help these students be proficient on the CELDT and CST and then it would be nice for you to see not only the classrooms, but working as colleagues and how well we work together. And you could see different venues and different grade levels, each of those components.

Several teachers commented that the ELD program itself drew teachers together:

We get together to plan our ELD, because we have a new ELD program and it takes a lot more work than the typical. We don’t have a manual that you can just flip open and just follow. So now we’re just used to meeting together, and we look forward to it.

These qualitative data support our other findings in that discussion around ELD seemed to be related to a school-wide strategy.

Limitations

All self-reported data must be considered with some circumspection. People tend to answer survey items in ways that square with their own beliefs, but anonymous surveys like ours mitigate this impulse. Second, because our data are clearly correlational, we cannot be sure, for instance, whether collective efficacy causes confidence in the ELD approaches and programs of a school, or vice versa. We contend (and our regression analysis assumes) that collective efficacy and other school climate variables cause strong methods and programs for ELLs, but it is possible that the opposite is true. Third, we were not able to obtain data on achievement linked to ELLs at either the school or classroom level. We contend, however, that previous research (Goddard et al. 2000b) has sufficiently established this link to allow us to assume that those schools with high collective efficacy were also those with achievement scores relative to matched schools. In an era of accountability, test scores are no doubt important, but we would argue that confidence in one’s professional activities, as collective efficacy measures, should be a goal in and of itself.

Implications and conclusions

Perhaps the widest instructional goal in the elementary school is the acquisition of literacy. Thus, the ELD program of a school is necessarily bound to this larger instructional enterprise and thus ‘seeps; into every aspect of school function. As one of our interviewees reported, even in mathematics instruction, there is “so much text, reading of text involved in math[ematics]” that the ELD programs matter. We were therefore not surprised to find efficacy for ELD so closely tied to general measures of school focus, culture and collective efficacy. Our essential finding—that collective efficacy, school culture and focus, and perceptions of school-wide effective instruction and programs for ELLs are all associated—is not surprising, but our study is the first to document these relationships. The more important question turns on how elementary schools can enhance the collective self-efficacy of their teachers and thus the quality of their ELL program. Before addressing this topic, we explore the previous research in light of our findings for consistencies and contradictions.

First, we were somewhat surprised to find that teachers generally rated the strategies for ELD at their school highly. Ratings for the program model at their school were less positive, but nearly all teachers agreed that the program at their school was working well. These data run counter to earlier findings that teachers are not confident ELD instructors, but the previous research involved individual teachers in rating effectiveness. We wondered if this finding might be true for other content or instructional areas. For instance, do elementary teachers report low self-efficacy in their individual mathematics instruction while holding a high regard for the mathematics teaching at their school?

Although we cannot answer that specific question, general research on the relationship between self-efficacy and group efficacy suggests that group efficacy grows when members acquire, store, manipulate and exchange information about each other and their task, context, process and prior performance (Gibson 2001). These same collective processes seem to occur during the development of self-efficacy. With more data to draw upon, collective efficacy might always be greater than self-efficacy, regardless of the task.

It is worth considering why many ELD teachers believe themselves to be under-prepared for the crucial work of teaching language (Alexander et al. 1999). ELLs present both a linguistic and cultural challenge for teachers, the vast majority of whom share neither an ethnic background nor native language with their ELLs (Téllez 2004). As anyone who enters an unfamiliar culture discovers, self-confidence is quickly challenged. Our data suggest that perhaps, when teachers consider their collective capacity, a more confident voice is found.

The power of a language teaching program at a school cannot be underestimated. Strong evidence for this claim can be found in a statewide study conducted in the aftermath of Proposition 227 in the form of a comprehensive study, which revealed that bilingual instructional approaches were not statistically different from structured English immersion approaches in terms of improving ELL performance (Parrish et al. 2006). Although the study revealed no clear evidence to support of the superiority of one ELL instructional approach over another, the researchers found that the quality of the program was crucial, maintaining that four features of the language teaching program produced the highest academic gains: (a) staff capacity to address ELL needs; (b) a school-wide focus on ELD and standards-based instruction; (c) shared priorities and expectations in regard to educating ELLs; and (d) systematic, ongoing assessment and data-driven decision making.

The upshot of our work is that ELLs and their specific instructional needs must become a focus of teacher learning communities. The teacher inquiry intentions held by several of the conceptualisations of teacher learning become easily conflated and misinterpreted as accountability tools to support test data analysis. This practice and misinterpretation has worried many, but counterarguments often become caught up in ideological arguments, which can be easily distorted, causing a cycle of well-intentioned ideas but flawed actions.

We suggest that one way in which to interrupt this cycle is to reappraise the role of the ELLs and their families in the teachers’ learning community. This approach takes not only a conceptual step, but also moves teacher learning and school practice forward. We believe that we must reimagine ELLs in the centre of our learning communities, not at the margins, especially for learning community conceptualisations that value teacher inquiry as a primary mode of learning. To turn our earlier theorising upside down, we argue that teacher learning and thus collective efficacy should focus on meeting the needs of students at the margins, especially students such as ELLs who are counting on the school to equip them with the tools that they will need for full participation in the economic, social and political life in the United States. Aiming first at an abstraction known as collective efficacy makes little sense when schools can instead meet the needs of ELLs and thereby gain collective efficacy as a result.

Appendices

Appendix 1: Collective Efficacy Scale survey items

• Teachers in this school are able to get through to difficult students (GC).

• Teachers here are confident they will be able to motivate their students (GC).

• Teachers in this school really believe every child can learn (GC).

• If a child doesn’t want to learn, teachers here give up (GC).

• Teachers here don’t have the skills needed to produce meaningful student learning (GC).

• Students at this school come to school ready to learn (TA).

• Home life provides so many advantages, the students here are bound to learn (TA).

• Students here just aren’t motivated to learn (TA).

• The opportunities in this community help ensure that students will learn (TA).

• Learning is more difficult at this school because students are worried about their safety (TA).

• Drug and alcohol abuse in the community make learning difficult for students here (TA).

• Teachers in this school do not have the skills to deal with student disciplinary problems (GC).

Adapted from Goddard (2002), A theoretical and empirical analysis of the measurement of collective efficacy: The development of a short form, Educational and Psychological Measurement, 62, Copyright 2002 by Sage Publications.

Appendix 2: interview protocol

• 1. If you were to take me on a tour of your school, what are the things you’d most want me to see and notice?

• 2. Tell me why you picked those items to share.

• 3. In this study, I’m particularly wanting to learn about the ways that teachers work together. In what ways do you work collaboratively with other teachers? With all teachers? How is this different from other schools in which you’ve worked?

• 4. What has your principal done that influences the ways teachers are able to work together? What approaches to working more closely together have been tried but just didn’t seem to work?

• 5. In general, what goals do you have for your students? What helps you to stay focused on those goals?

• 6. What kinds of discussions do you have at school about current theories and promising educational practices? Who starts those discussions?

• 7. If you wanted the staff to start thinking about a new idea how would you get that to happen?

• 8. What values or outlooks do you think are shared by all the teachers in this school?

• 9. How are those values communicated to teachers who are new to this school?

• 10. Tell me about a time that you were able to accomplish more than what you thought you could. What inspired you to do that? Who helped you to do that?

• 11. In what ways are the accomplishments of students, faculty, and the school as a whole acknowledged? Who makes sure that happens?

• 12. What else would you like to tell me about your school?