Perceived collective teacher efficacy in low performing schools

The challenge of low performing schools continues to rear its ugly head in many countries, including South Africa. The responses to low performance differ from country to country, but none of these have included the enhancement of collective teacher efficacy in their repertoire. Research shows that collective teacher efficacy is positively related to improved academic performance of learners. Schools with a higher sense of collective efficacy outperform schools with a lower sense of collective efficacy. Schools with high efficacy are characterised by strong work ethic, and teachers who persist in the face of difficulty. Moreover, teachers in these schools are more persistent in their efforts, plan more, and view failure as a temporary set-back that does not discourage them. Therefore, if principals and their management teams could find a way to enhance collective teacher efficacy, the challenge of low-performing schools may be overcome. In view of this, quantitative research was conducted with the aim of determining the strength of collective teacher efficacy in low performing schools. Ten randomly selected schools in the Kenneth Kaunda Education District were involved in the research and in each selected school all the teachers were involved (N = 217). Data was collected using a questionnaire (The Collective Teacher Efficacy Scale: short version). The questionnaires were delivered and collected in schools by the researchers. Data was analysed using descriptive statistics, frequencies, percentages and mean scores. The results show that collective teacher efficacy in these schools is medium to high pertaining to group competence, but lower in task analysis.


Introduction and Background
Low-performing schools remain a challenge all over the world.Most governments react by offering more money to uplift low-performing schools in the belief that poverty and lack of resources are the major contributing factors to low performance (McColskey & Monrad, 2004).In the United States of America (USA), low-performing schools are placed under sanctions such as dismissal of the principal, closure of the school, and re-opening of the school as a private school (Woods & Levaçić, 2002).In South Africa, the government prefers to introduce accountability measures such as the Integrated Quality Management System (Education Labour Relations Council, 2003).The Department of Education also provides support to these schools through the institutional support programme (Molale, 1995), training of principals (Masitsa, Van Staden, De Wet, Niemann, Heyns, Brazelle & Niemann, 2004), monitoring the provision of support materials (North West Department of Education, 2004) as well as the visiting of low performing schools by subject advisors (Mogonediwa, 2008).All these measures, however, have had limited success only in turning around low performing schools.
Commentators cite a number of reasons for low performance of schools.In low-performing schools, the principal works in isolation, the School Management Team (SMT) does not meet regularly, there are no subject heads, and where these subject heads are found, they seldom meet for purposes of planning (Mogonediwa, 2008).The challenge seems to lie with the management of the school where teaching is not effectively supervised (Legotlo, Maaga, Sebego, Van der Westhuizen, Mosoge, Niewoudt & Steyn, 2002;North West Department of Education, 2001).This implies that school management tolerates poor teaching, which results in poor academic achievement (Woods & Levaçić, 2002).However, teachers in these low performing schools seem to exhibit a weak teacher efficacy and the school as whole has a weak collective teacher efficacy.
What, then, is teacher efficacy and collective teacher efficacy?Teacher efficacy refers to the individual teacher's belief in his/her capacity to affect student performance (Cheung, 2008;Erawan, 2010;Yeo, Ang, Chong, Huan & Quek, 2008).Rangraje, Van der Merwe, Urbani and Van der Walt (2005:38) come to the conclusion, after analysing the work of Tschannen-Moran, Hoy and Hoy (1998:206), that "teacher efficacy can be conceptualised as teachers' belief that factors under their control ultimately have greater impact on the results of teaching than do factors in the environment or in the student-factors beyond the influence of teachers." Collective teacher efficacy refers to the perceptions of teachers that their efforts as whole will have a positive effect on students (Goddard, Hoy & Hoy, 2004).Elaborating on this definition, Schechter and Tschannen-Moran (2006:481) indicate that collective teacher efficacy means that "teachers in a given school believe that they can make an educational difference to their students over and above the educational impact of their homes and communities."Collective teacher efficacy, therefore, involves the combined perceptions of the staff of a particular school.It is a product of interaction between group members and the emergent property is more than the sum of individual members (Goddard et al., 2004).
Although the concepts teacher efficacy and collective teacher efficacy influence each other reciprocally, the concepts nonetheless differ.While teacher efficacy uses the individual as the unit of analysis, collective teacher efficacy uses the teaching staff or school as its unit of analysis.
Hence, Schechter and Tschannen-Moran (2006) point out that, unlike teacher efficacy, which is the attribute of the individual teacher, collective teacher efficacy is a group attribute that is more than an aggregate of individual teachers' self-efficacy beliefs.Moreover, teacher efficacy beliefs are based on individual perceptions whereas collective teacher efficacy is based on the combined perceptions of the teachers in a school.

Problem Statement
As alluded to above, some challenges experienced in schools may be attributed to a lack of effective management.Van Deventer and Kruger (2003) hold the view that teachers can only perform their tasks successfully if a skilled and efficient management team is leading the school.Available research suggests that principals who encourage collaboration among teachers enhance collective teacher efficacy in their schools (Brinson & Steiner, 2007).According to Ross and Gray (2006), principals who adopt the transformational leadership approach are more likely to impact positively on collective teacher efficacy.As an example of the impact of collective teacher efficacy, Brinson and Steiner (2007) show how an elementary school principal changed the school from low performance to high performance within a space of two years by applying strategies that enhance collective teacher efficacy.This means that principals and their management teams play a vital role in achieving a strong sense of collective teacher efficacy.
Although research points to teacher efficacy as a crucial aspect of improving student performance, attempts to turn around low-performing schools have yet to consider enhancing collective teacher efficacy.According to Van der Westhuizen, Mosoge, Swanepoel and Coetsee (2005), many variables influence student achievement, but none is more powerful than the educator in class.Goddard et al. (2004) contend that efficacy of teachers is a powerful construct, which is associated with student achievement.Therefore, if principals and their management teams could find a way to enhance teacher efficacy and collective teacher efficacy, the challenge of low-performing schools may be overcome.A school's sense of collective efficacy can therefore stimulate high levels of academic improvement which can contributes significantly to the level of academic success of the school (Schechter & Tshannen-Moran, 2006).To this end, Schechter and Tshannen-Moran (2006:482) assert: "Collective teacher efficacy influences student achievement because greater efficacy leads to greater effort and persistence that result in better performance." A number of studies have linked collective teacher efficacy with improved student achievement (Cheung, 2008;Klaasen, Tze, Betts & Gordon, 2011;Tschannen-Moran & Barr, 2004;Yeo et al., 2008).Research by Brinson and Steiner (2007) shows that even when race, socio-economic status and gender are taken into consideration, collective teacher efficacy remains a powerful predictor of academic performance.The latter statement based on international studies, brings hope to most South African schools which are burdened with structural and systemic challenges that hinder the academic performance of learners.Enhanced collective teacher efficacy holds promises that these challenges may be overcome.
The purpose of this paper is therefore to investigate the state of collective teacher efficacy in low performing schools.The overall aim is to establish the importance of collective teacher efficacy towards improved learner academic performance in schools.Before we discuss the research methodology, we will consider the underlying theoretical framework of collective teacher efficacy.

Social Cognitive Theory as Theoretical Framework
Collective teacher efficacy is a derivative of Bandura's Social Cognitive Theory (Bandura, 1977(Bandura, , 1997)).Key to this theory is the existence of human agency which defines the way people exercise some level of control over their lives.The exercise of control is related to the person's sense of efficacy in that a person believes in his/her capabilities to influence a course of action to produce a given goal (Goddard et al., 2004).Teachers in a school, for example, exercise some control over the functioning of the school and thus believe that they will influence the outcomes of the school through their actions.
In dealing with collective teacher efficacy, the concept of human agency is replaced with the concept organisational agency, which refers to the combined control that members exert on the organisation.Organisational agency is based on the premise that individuals do not act in a vacuum or as "social isolates" but are influenced by the actions of the social group (Sørlie & Torsheim, 2011:176).This is reflected in the decisions that groups make in the light of their collective capability to reach a given goal.Thus, considering their capabilities, teachers in a school may decide to pursue high standards of teaching, attain excellence in sports and/or improve the academic performance of their learners.
Social cognitive theory posits four sources of efficacy-shaping information: social persuasion, vicarious experience, mastery experience and affective state (Goddard et al., 2004), of which social persuasion and vicarious experience can be beneficial to principals wishing to enhance collective teacher efficacy.Social persuasion seems likely to shape collective teacher efficacy, because the school is an organisation in which members interact on a daily basis (Tschannen-Moran & Barr, 2004).Through collective teacher efficacy, norms are developed in the school and teacher behaviour and actions are evaluated within the context of these norms.A robust collective teacher efficacy influences the way in which teachers manage their classrooms, what expectations they have about student achievement, and how they teach.In schools with a high collective teacher efficacy teachers believe that all students are teachable regardless of their socio-economic backgrounds (Schechter & Tschannen-Moran, 2006).Teachers in a school with a high collective efficacy sense believe that they, not the environment, have the greatest influence on student achievement.Vicarious experience means that a sense of efficacy is gained from learning from other people (Sørlie & Torsheim, 2011).Teachers listen to and share stories of successes and failures with their colleagues whenever they meet in conference or workshops.Principals and their teachers may visit other schools to see how things are done at these schools and either copy or model what these other schools are doing.One of the strategies for sharing experiences is through school clusters.Research shows that clustering of schools has benefits for the individual and the school (Delport, A & Makaye, 2009;Giordano, 2008).Among the advantages of clustering schools, A Delport and Makaye ( 2009) identify exchange of expertise, forging of relationships between previously isolated teachers, collaborative problem-solving and improved staff development.Mastery experience is recognised as the most influential source of efficacy, and may include prior school performance (Zakeri, Rahmany & Labone, 2016:160).In essence, mastery experience refers to the belief that one can perform due to having mastered a previous task or venture . Ramos, Costa e Silva, Pontes, Fernandez and Nina (2014:180) posit that the affective state provides a source of collective teacher efficacy, which according to Bandura, states that people by judging their capabilities, partially place trust in their emotional state.An example is that of high levels of stress weakening group functioning, which lowers a sense of self confidence in the capabilities of other members.
The literature reveals that very little research has been conducted on the construct collective teacher efficacy.Henson (2002) and Tschannen-Moran et al. (1998) bemoan the dearth of collective teacher efficacy studies and the limited nature of research in this direction.Echoing this idea, Klaasen et al. (2011) states that more research has been conducted on teacher efficacy than on collective teacher efficacy.The limited research that has been conducted in developed countries, shows that collective teacher efficacy has been researched from different points of view, linking it in particular to student achievement (Goddard et al., 2004;Parker, Hannah & Topping, 2006) and the mediating factors to collective teacher efficacy (Bruce, Esmonde, Ross, Dookie & Beatty, 2010;Goddard & Skrla, 2006;Ross & Gray, 2006;Ross, Hogaboam-Gray & Gray, 2004).
Studies on collective teacher efficacy are more numerous in developed countries, but scant in developing countries.In fact, most of the studies such as those by Rangraje et al. (2005) in South Africa, Onderi and Croll (2009) in Kenya, and Yeo et al. (2008) in Singapore, concentrate on teacher efficacy rather than on collective teacher efficacy.Therefore the present study will present a view of collective teacher efficacy from a developing country, thereby adding to the sparse literature on this subject, and allowing for comparison with studies conducted internationally.

Research Methodology
This research adopts the positivistic approach which aims to predict human conduct and to evaluate the social world objectively (Maree & Pietersen, 2010).Data were collected using a survey questionnaire.Survey research offers a quantitative or numeric depiction of opinions, attitudes, or trends of a population by studying a sample thereof (Creswell, 2009).This approach was followed because research on this topic is still scant in South Africa, and this will allow for an overview of the state of collective teacher efficacy which is deemed to be suitable as a first step towards opening the topic for further research.A quantitative research method was used because it is a structured process, which allow for the gathering of large samples of quantitative data as for possible generalizability (Morrell & Carroll, 2010).In addition, the researchers were able to identify trends and issues pertaining to collective teacher efficacy in the schools under investigation (Brinson & Steiner, 2007).Moreover, administering the questionnaire was simplified as the respondents simply followed the directives on the questionnaire itself without needing assistance from the researchers.
The survey instrument used was the short version of the validated Collective Teacher Scale developed by Goddard (2002).Goddard (2002:108) found that the use of a 12-item scale "is equally as effective as using the original 21-item scale."He also found the 12-item scale to be more tight-fisted using 43% less items than the original.The high correlation (r = .983)between the short form and the original scale suggests that they are strongly related.The validity tests conducted by Goddard (2002) for mathematics achievement found that "the short form of the Collective Efficacy Scale was a significant predictor of between-school differences in student mathematics achievement" (Goddard, 2002:107).This instrument in the first place, tests the judgments of teachers about groupcompetence (GC) including teaching methods, skills, training and experience.Secondly, it represents judgments about task analysis (TA), which are perceptions of constraints and opportunities in the task at hand.This includes teachers' beliefs about the support students get at home and in the community.The instrument contains an equal number of items for GC and TA, and in each category there are three negatively-worded items to eliminate socially-acceptable responses.The participants responded by ticking the number that best represented their opinion on a four-point Likert scale, where 1 = totally disagree and 4 = totally agree.
Of the schools in the Province, 16 were listed by the Department of Education as low-performing.At an average of 35 teachers per school, the population of teachers for the 16 schools was N = 560.For purposes of collecting data, it was deemed convenient to select 10 schools randomly due to logistic challenges involving schools in difficult to access rural areas.All the teachers (N = 350) in the 10 schools were sampled for the questionnaire survey (Leedy & Ormrod, 2005:221).
Of the 350 questionnaires distributed to schools, 217 teachers responded, which was only 62% response rate -this despite researchers' personal distribution and retrieval of questionnaires following day.However, the return rate was deemed acceptable as asserted by CSL Delport (2002:172) and was considered representative of the 16 schools' population of teachers.
Data was analysed using descriptive statistics involving the frequencies, mean scores and standard deviation.Morrell and Carroll (2010:138) describe descriptive statistics as "summarizing [sic] the characteristics of a group."A t-test was applied between rural and urban schools and between male and female responses, but no significant results emerged from this comparison.Data were analysed according to the four categories indicated by the research instrument: general competence positive (GC+), general competence negative (GC-), task analysis positive (TA+) and task analysis negative (TA-).In the tables below, responses in the anchor of "agree" and "totally agree" will consistently show a low collective teacher efficacy (CTE) for negative statements.

Participants
The data produced a good gender distribution, with 49.3% males and 50.7% females who participated in the research.Majority (52.56%) of the participants were in the age bracket 40 to 49 years followed by those that are 30 to 39 years (24.18%),50 to 59 years (16.28%) with only a few that were 20 to 29 years (4.18%) and 60 years and above (2.8%).Majority (52.33%) possessed a B-degree, followed by 30.57% who had an Honours or BEd degree with 2.59% who had a Masters quali-fication.There was a substantial number of participants (14.51%) who only possessed a Grade 12 certificate.The majority (81.73%) of the participants were on Relative Education Qualification Value (REQV) 14 to 16.The majority (39.53%) of the participants had 11 to 20 years teaching experience, followed by those with 0 to 10 years (32.56%),21 to 30 years (20.47%) and those with 31 years and more (7.44%).As far as the location was concerned, 69.48% were from township schools, 22.07% rural and 8.45% urban schools.

Responses to Items on General Competency (Positive) (GC +)
According to Table 1, responses to the general competency (positive) statements show that the majority of the respondents agreed with the given statements, showing a relatively strong CTE.Responses agree and totally agree accounted for 77.1% of the responses.This shows that the majority of respondents agreed and totally agreed with the statements, thus showing a relatively strong CTE.This is supported by the mean scores which were all above the 2.50 cut-off point.Responses to the statements that 'teachers are confident that they will be able to motivate their learners' (B2) and that 'teachers in the school really believe that every child can learn' (B3) showed a strong CTE with mean scores that border on "agree" (2.96 and 2.99, respectively).Attention, however, is drawn to item B1, which shows that 32.2% of the respondents disagreed and totally disagreed with the statement.This means that a sizeable number of respondents were of the opinion that teachers were unable to get through to difficult learners.The mean score for this item was also the lowest in this category (2.72), which shows a weaker CTE than for other items in this category.This weaker CTE could partly explain the low performance at the schools, which could be attributable to contextual factors in play.

Responses to Items on General Competency (Negative) (GC-)
In statements in Table 2 with a negative loading paint a different picture.The total responses for disagree and totally disagree accounted for 40.6% of responses, showing a low CTE.A total of 80.8% agreed and totally agreed that if a child does not want to learn, teachers give up (B4).An even higher percentage (85.6%)indicated that they do not have the skills needed to produce meaningful learner learning (B5).This is in line with the low mean score of 1.76 for these statements, indicating a low CTE.The response to the statement that 'teachers do not have the skills to deal with learner disciplinary problems' (B12) confirms the research findings of Maphosa and Shumba (2010:395) that, with the banning of corporal punishment, teachers are struggling to maintain discipline in schools.A total of 71.9% indicated that teachers in their schools do not have the skills to deal with learner discipline problems.The percentage of respondents who disagreed and totally disagreed is, however, higher than the others in this category, indicating that almost 30% of the respondents still felt confident about their competency to deal with learner disciplinary problems.Responses to Items on Task Analysis (Positive) (TA+) Table 3 presents a different view of the CTE in comparison with the general competency (GC).Of the total respondents, 71.5% disagreed and totally disagreed with the statements whereas only 28.5% agreed and totally agreed.The mean scores also showed low figures and range between 2.12 and 2.26, which is far below the expected 2.50.This is indicative of a weak CTE.According to the responses it seems the challenge lies with the community from which the learners come.For example, the statement that the 'home-life presents so many advantages that learners are bound to learn' (B7) had the lowest mean score of 1.91.Moreover, the statement that 'the opportunities in this country help ensure that our learners will learn' (B9) had a mean score of 2.12 with a total of 68.5% of the respondents disagreeing and totally disagreeing.It seems CTE becomes stronger where learners are concerned, as indicated by the statement that 'our learners come to school ready to learn' (B6); with a total of 33.8% in the agree and totally agree columns, and a mean score of 2.26 − in fact the highest mean score in this category.
Responses to Items on Task Analysis (Negative) (TA-) Table 4 shows that responses between TA+ and those to TA-are in agreement.The total figure for disagree and totally disagree was 53.0%, whereas the figure for agree and totally agree was slightly lower at 47.0 percent.In this regard one would say the closeness of results here shows a medium CTE.However, the responses to items in this category stand in stark contrast to one another.For example, the statement that 'learning in this school is more difficult because learners are worried about their safety' (B10) showed a weak CTE with responses for agree and totally agree, totalling 74.1 percent.
In contrast, the statement that 'drugs and alcohol abuse in the community make learning difficult for learners' (B11) showed a strong CTE with responses in the disagree and totally disagree anchors notching 80.9 percent.

Discussion
The dichotomy of responses to a high CTE for general teacher competence and a low CTE for task analysis clearly emerges from the above results.Apparently, teachers believe in their competence to produce the desired results but are short-circuited by events and factors outside their control.The general expectation about CTE is that it should be strong even in the face of potentially limiting factors such as the home background of the learners.This finding is consistent with findings on teacher efficacy showing no difference whether it refers to individual beliefs or collective beliefs.Teachers' beliefs in their general competence to produce the desired results, is somewhat surprising because of their differing qualifications.In fact, the highest mean scores are found in this category.Mosoge (2012) points out that some schools are staffed with unqualified and underqualified teachers whereas Spaull (2013) argues that teachers cannot teach what they do not know.What gives respondents their strong CTE in this category seems far-fetched.It is possible that the measures instituted by the North-West Education Department to improve performance in low performing schools, enhances the CTE of teachers.Mosoge (2012) concludes that lack of sufficient qualifications did not deter teachers from seeing themselves as competent.One would also conclude that the response is obvious, for no teacher would present himself or herself as incompetent.Thus it seems as though the low performance of the school must be sought elsewhere and not in their competence.The responses to task analysis items produce intriguing results.While the responses for task analysis generally show a medium CTE, there are cases where it is high and where it is low.For example, the statement that 'learning in this school is more difficult because learners are worried about their safety' (B10) shows a weak CTE in that teachers believe that learners are worried about their safety.Thus, it is not surprising that the CTE would be weak for this item considering, the reports of violence in and around schools and perceptions of moral decay in schools and society.
It seems CTE becomes stronger where learners are concerned as indicated by the statement that 'our learners come to school ready to learn' (B6), with a total of 33.8% in the agree and totally agree columns, and a mean score of 2.26 − in fact the highest mean score in this category.The mean

Table 1
Responses to General Competency questions (positive statements)

Table 3
Responses to Task Analysis questions (positive statements)

Table 4
Responses to Task Analysis questions (negative statements)