Abstract
The principal aim of this article is to describe and document a comparison of public and private school effectiveness in the marginalized (often referred to as slum) district of Aguablanca, in the city of Cali, Colombia. The data suggest that teachers and parents are largely satisfied with the current status of education provision in Aguablanca. However in terms of ‘statistically significant’ differences, stakeholder perception of the private sector exceeds that of the public sector for every given indicator. The research suggests that the authorities reconsider their education strategy in light of the positive perceptions associated with the existing system of provision. While it may be necessary to increase system capacity in order to meet the demand, great care should be taken to protect (and even extend) provision for the poor through the existing programme of private school subsidies.
Introduction
It is generally accepted that free public education is required to meet the needs of children in marginalized communities. In Dakar, 2000, governments and agencies committed themselves to ensuring that by 2015 all children will ‘have access to and complete free and compulsory primary education of good quality’ (World Education Forum, 2000: para. 7). However, research has identified a fee paying private (non-state) alternative albeit low cost that has ‘mushroomed’ in developing countries due to the perceived (by parents) inadequacies of state provision (Adelabu and Rose, 2004; Baurer et al., 2002; Rose, 2002; Tooley, 2009).
This growth of private sector schooling has even been witnessed in poor countries, for example, Kenya, Uganda and Malawi where free, state education has been recently introduced, though Rose (2002) and Lewin (2007) question the quality, accessibility and subsequent legitimacy of private institutions in the presence of a state system struggling for scarce resources. The aim of this research therefore is to characterize and compare principal stakeholder (parent and teacher) perceptions of school quality in relation to management type in the context of a developing world environment; more specifically, the Aguablanca marginalized (slum) district of Cali, Colombia’s third city.
Research suggests that Colombia is a highly unstable environment, poverty and social unrest being closely related to gross political and economic inequality (Ortiz, 2002). However, Colombia also represents a particularly pertinent research environment because unlike many developing regions of the world, it actively promotes alternative forms of education provision for the poor beyond the free option offered by the state. Broadly speaking, provision in Colombia is divided into the following categories:
Public schools: those wholly financed through government funds.
Assisted private schools: privately managed institutions in receipt of a state subside (PAC—Programme for Increased Enrolment) that covers (entirely or in part) the cost of a poor child’s education.
Unassisted private schools: privately managed self-funding institutions that receive no financial assistance from the state.
Aguablanca is a 20 year-old slum district on the eastern periphery of Cali. It is composed of four communities (comunas 13, 14, 15 and 16) each subdivided into discrete neighbourhoods and covering a total area of 12 square kilometres. It has a total population in excess of 500,000 people (one-quarter of the city’s total) composed predominantly of urban victims of economic crises and poor migrants, the majority of whom have been forcibly evicted from the surrounding rural provinces. Aguablanca currently accounts for 36 per cent of the city’s poor population, however, the relentless arrival of displaced people creates further pockets of extreme poverty (most notably in comuna 13) provoking confrontation and violence with adjacent, more established neighbourhoods (Hentschel, 2005). This social instability is further exacerbated by gang violence often related to the highly active and resilient drugs trade.
Previous educational research in Aguablanca suggests that the private sector not only accounts for the majority share of institutions and student enrolment but also provides a better service in terms of a range of quantitative input performance indicators, including pupil-teacher ratio, teacher activity and teacher cost with equivalence in material resources (Akerman and AlWindi, 2007).
School Climate
School climate research has its origins in effectiveness studies, devised as a mechanism for improving the learning environment for children from marginalized or poverty stricken backgrounds (Brookover and Lezotte, 1979; Edmonds, 1979). Effectiveness criteria alone however could not account for the range of differences in educational outcomes. Schools are therefore perceived as having definite personalities related to perceptions of the work environment as influenced by the formal and informal organizations, the personalities of the participants and leadership. According to Hoy and Miskel (1996: 141) school climate is: A relatively enduring quality experienced by participants that affects their behaviour and is based on their collective perceptions.
Subsequent research has established a definitive link between school climate and academic achievement (Brown and Medway, 2007; Hoy and Miskel, 1996; Stewart, 2008) most notably, Borger et al. (1985) on the basis of a meta-analysis (205 research reviews) investigating the range of factors that characterize school performance.
Literature review suggests that the assessment of school performance in the developing world is generally defined by input measurement (scale of available resources), test scores and the degree of primary and secondary coverage. The climate scale was therefore selected as a complementary function, providing a broad and efficient means of deriving stakeholder perception of organizational effectiveness. To date, only one climate investigation in the developing world (the Gambia) has been identified, related to secondary school comparison (Sarr and Lewin, 1993). In view of the objectives and the limited description of methods and instruments, this research was however not deemed suitable as a basis for climate assessment in Aguablanca. This study was therefore based on an objective selection of climate tools using criteria specifically related to the delivery of education in the developing world.
Method
Based on a range of educational research studies initiated by Hanushek (1986) and Lockheed and Levin (1993), a set of general factors were considered essential for the creation of an effective school in the developing world. This list of ‘effectiveness’ factors was used as selection criteria for the research in order to ensure that the chosen set of teacher and parent instruments were readily suited to climate evaluation in the developing world. A complete description of the research design is provided below.
A thorough review of climate research has revealed a range of quantitative instruments considered suitable for learning environments different to those in the slums of Cali. However, a high degree of alignment to the Hanushek effectiveness factors resulted in the following selection. Teacher: The Organisational Health Inventory (OHI) as defined by Hoy et al. (1991) is based on the following seven climate factors: institutional integrity; consideration; principle influence; initiating structure; resource allocation; morale; and academic emphasis. Parent: The Houston Independent School District survey (HISD) as described by Stevens and Sanchez (1999) is based on the following four climate factors: instructional environment; instructional process; communications; and administration.
This predominantly quantitative assessment is based on a questionnaire format. Both the teacher and parent instruments contain closed questions as a mechanism for gathering perceptions of the school climate prior to conversion into manageable data profiles. The questionnaire approach was selected primarily for reasons of suitability and convenience; a limited and controlled range of answers simplifying the data gathering and processing tasks. This approach was further supplemented by an ‘additional comments’ section on the questionnaires and unstructured interviews (with teachers and parents) as a means of providing a qualitative dimension to the research. Note that the climate assessment doesn’t include an evaluation of student perceptions due to time and resource constraints.
The OHI was originally piloted on a diverse set of 75 schools (Hoy and Feldman, 1987). A factor analysis indicated that the OHI has high construct clarity and fit, that is, instrument items are closely related to their factors. An internal consistency check (applying Cronbach’s alpha coefficient) optimized the number of test items while further testing of the refined OHI indicated that the alpha coefficient for each of the associated factors was of the order of 90 per cent, suggesting a high degree of internal reliability in the given context.
The Houston Independent School District survey (HISD) developed a climate tool to identify student/parent/community perception of district schools (Stevens and Sanchez, 1999). While no formal reliability and validity checks were conducted on the HISD instrument, the items used were firmly based on theoretical constructs derived from thorough review of associated literature and grounded research (Stevens and Sanchez, 1999).
Although these instruments are used to assess the school environment (from the perspectives of teachers and parents) and have similar questions, the OHI and HISD questionnaires do not come as a validated set. Hence, teacher and parent opinions could not be compared directly.
The most up-to-date list (March 2009) of private and public institutions in the Aguablanca district was obtained from the local education authority in central Cali. Unlike other developing nations, there is no sizeable informal/unregistered segment in urban Colombia, therefore the list is believed to represent a ‘complete’ school population (Akerman and AlWindi, 2007). In view of the constraints associated with a short 2-month data-gathering period, the climate assessment was focused on a single community (comuna 13).
While there was an intention to compare ‘like for like’ in terms of institution size (based on student numbers), a review of the data soon revealed that this would not be possible, all state schools being several orders of magnitude (at least three times) larger than any private equivalent. The review also identified six different grades ranging from pre-school through to adult education. In order to constrain the number of test variables to a manageable quantity, the sample population was limited to schools that taught at both primary and secondary grades only (25 schools out of a community total of 58). A complete population list of target comuna 13 schools was compiled and divided by type (public, private) and size. A final total of 17 institutions (30 per cent of the community total) were then drawn ‘randomly from a hat’ with the management type breakdown illustrated in Table 1.
Research sample
A glance at the sampled institutions and student numbers indicates that the private sector in Aguablanca is heavily subsidized by the state (56 per cent of students within the private ‘assisted’ sector sample are covered by the PAC programme). While the majority of school principals expressed satisfaction with the efficiency of the programme, according to qualitative evidence a number of institutions are receiving ‘less financial support than anticipated’ and/or ‘late payments’ with knock-on effects particularly in relation to teachers’ salaries. The majority of schools in the sample satisfied the predefined response requirement with 30 per cent of total teacher population and 1 per cent of total parent population (based on an assumption of one student per parent) completing the questionnaires, as illustrated in Table 2. It is important to the note that Aguablanca is a particularly dangerous district with an extremely high rate of homicides. Data collection was therefore limited to daylight hours and the international researchers were always accompanied by a local guide.
Questionnaire responses
The research procedure was conducted in line with a set of ethical guidelines (BERA, 2004) that promote the efficacy of the study while protecting the needs of the participants. A set of communiqués were prepared in advance of arrival, addressing the following aspects: voluntary participation, guarantee of privacy, purpose of the research, methods employed, the right to withdraw and publication information.
Results and Analysis
The principal aim of this section is to present, compare and assess parent and teacher perceptions of school effectiveness (through the climate measure) for the private (assisted and unassisted) and public school management types in Aguablanca.
In essence, this investigation is a study of service satisfaction measured in terms of effectiveness through school climate factors. Data analysis is based on a Likert scale containing the following response categories: ‘1’ (rarely—least favourable response); ‘2’ (sometimes); ‘3’ (often) to a maximum score of ‘4’ (almost always—most favourable response). The data are supported by a ‘significance’ value (ρ) and an ‘effect size’ (d), quantifying the magnitude of the mean difference between the two management types.
Teacher Factors
The public and private school data derived from the teacher questionnaires are presented in the form of box plots as illustrated in Figure 1.
Teacher climate factors
Institutional Integrity (integ)
This factor is a measure of the schools ability to manage the interface between itself and the community, meeting the needs of the parents while maintaining the educational consistency and integrity of its programmes.
The data suggest that the teachers in both the private and state sectors are largely ‘satisfied’ with the degree of institutional integrity. However, the mean factor score is higher for the private sector schools than the public equivalent by a significant value of 0.38 (ρ < 0.001, d = 0.788).
Qualitative evidence generally reflected a preference for the private sector in terms of institutional integrity. A number of teachers (and parents) indicated that security in the public sector was ‘inadequate’ and that ‘the size of public institutions was an issue’, as they tended to attract more negative attention in the forms of violence and drug problems related to local gang interference. While public institutions are very aware of these problems and have responded with conspicuous levels of ‘front gate’ security, the perception is that they are more vulnerable than the smaller, lower profile private equivalents. Furthermore, several school principals have indicated that a relatively high turnover rate of students from displaced families also undermines the integrity of public institutions. Interviews with principals in the private sector indicated that late or non-payment of PAC subsides are the most significant threat to institutional integrity. Teachers in private schools, conversely, were mostly concerned with the ‘lack of support from parents’ some of whom appeared to ‘delegate their responsibilities to the school’. This seems to reflect how private managers place fees and costs highly on their agenda, while teachers, working at the ‘front end’ put the parental relationship at the top of their concerns.
Consideration (consid)
This factor is a measure of principal behavior that is supportive and collegial. It represents a genuine concern on the part of the school principal for the welfare of the teacher.
The data suggest that teachers in both sectors are largely ‘satisfied’ with the degree of principal consideration and there is no statistically significant difference between the different management types. Moreover qualitative evidence suggests that broadly speaking, both teachers and parents perceive the principal as ‘accessible and courteous’.
Principal Influence (prin)
This factor is a measure of the school principal’s ability to influence the actions of superiors. The ability to persuade and gain consideration from ‘superiors’ and not to be impeded by the management hierarchy are important aspects of school administration.
The data suggest that teachers in both sectors are largely ‘satisfied’ with the degree of principal influence with a narrow and consistent perception particularly in the public sector. The climate factor score is higher for the private sector by a mean, statistically significant value of 0.36 (ρ < 0.001, d = 0.5).
Qualitative evidence from one particular public school (associated with all the outlining and extreme values) did suggest that ‘the principal is more focused on presenting a positive public image of the institution rather than addressing the deep seated problems of violence and drugs which continually undermine teacher and student performance’.
Initiating Structure (init_struc)
This factor is a measure of school principal behavior that is both task and achievement orientated. Work expectations, standards of performance and procedures are clearly articulated by the principal.
The data suggest that teachers in the public and private sectors are largely ‘satisfied’ with the schools’ initiating structure. The climate factor score is higher for the private sector by a mean, statistically significantly value of 0.33 (ρ < 0.05, d = 0.55).
It is hypothesized that the size of public institutions makes them more difficult to manage than the smaller private equivalents, particularly in a context of higher pupil to teacher ratios and surrounding social instability. This view was not however substantiated by definitive qualitative evidence from the teachers.
Resources Allocation (res_alloc)
This factor is a measure of material resource availability. Adequate classroom supplies and instructional materials are allocated to teachers and extra materials are readily supplied if requested.
The data suggest that resources represent a significant climate issue for teachers. While there is a degree of ambivalence and a large variance of opinion in the private sector, the public sector characteristic suggests a united teacher perception verging on ‘dissatisfaction’. The climate factor score is higher for the private sector by a mean, statistically significantly value of 0.73 (ρ < 0.001, d = 0.9).
While a fundamental shortage of resources is a familiar and significant factor undermining school effectiveness throughout the developing world (Harber and Davies, 1997), qualitative evidence from state school teachers in Aquablanca suggests a further unequal division of the ‘limited’ resources available; ‘allocation favoring certain teachers and subjects over others’. Finally, the box plot suggests a small but collective response by certain individuals in a state school.
Morale
This factor is a measure of the collective sense of friendliness and enthusiasm in the school and trust among faculty members. Teachers like each other, like their jobs and assist each other. They are proud of their school and feel a sense of accomplishment in their jobs.
The data suggest that teachers in both sectors perceive high levels of institutional morale while the climate factor result indicates no statistically significant difference between management types. When interviewed, no teacher from either sector referred directly to problems associated with morale. It would therefore appear that despite the presence of multiple contextual issues that could undermine teacher perception (including ‘late payment’ within the private sector), morale has been maintained at an impressively high level throughout the system.
Academic Emphasis (acad_emp)
This factor is a measure of the school’s drive toward academic excellence, that is, high but achievable academic objectives are set for the students, the learning environment is orderly and serious, teachers believe in their students’ ability to achieve and students work hard and respect those who do well academically.
The data suggest that teachers in the private sector are ‘satisfied’ with the academic emphasis while the perception in the public sector tends towards ‘ambivalence’. The climate factor score is higher for the private sector by a mean, statistically significant value of 0.56 (ρ < 0.001, d = 1.1).
It is hypothesized that larger pupil to teacher ratios in public schools complicate the teacher’s efforts to maintain order and discipline in the class. Qualitative evidence also suggests that teachers in the public sector would appreciate ‘additional support that offers alternative pedagogical techniques for students with emotional problems’.
Teacher Climate (index)
The overall climate index represents a mean total of the entire set of teacher responses from all the schools divided in terms of management type.
The data suggest that teachers are largely ‘satisfied’ with overall school climate in both the public and the private sectors. The climate index score is higher for the private sector than the public equivalent by a mean, statistically significant value of 0.37 (ρ < 0.001, d = 0.9).
Teacher climate factor analysis would suggest that the principal issues in the public sector are institutional integrity, resources and academic emphasis. The principal issue in the private sector is resource allocation.
Parental Factors
The data derived from the parent questionnaires are presented in the form box plots as illustrated in Figure 2.
Parent climate factors
Instructional Environment (instr_env)
This factor is a measure of the conditions at the school that affect parents or their children. School issues considered include security, facilities and teacher support.
The data profile suggests that parents are largely ‘satisfied’ with the Instructional Environment in both sectors. The climate factor value is higher for the private sector than the public equivalent by a mean, statistically significantly value of 0.28 (ρ < 0.001, d = 0.535).
Qualitative evidence suggests that a selection of parents both inside and outside of the public sector have particular concerns with respect to the following environmental issues: (1) ‘the lack of security’; (2) ‘the lack of resources’; and (3) ‘the lack of student discipline and respect’.
Communication (comms)
This factor is a measure of the quality of interaction between the school, the parents and the students.
The data profile suggests that parents are ‘satisfied’ with communications in both private and public sectors schools. Furthermore, the climate factor score suggests that there is no statistically significant difference between private and public factor scores for ‘communication’.
Qualitative evidence suggests both private and public sector schools need to provide parents with ‘more information’ regarding child related problems (learning and discipline) and school activities.
Instructional Process (process)
This factor is a measure of parental perception in relation to the teacher planning and classroom activities associated with student learning in the classroom.
The data profile suggests that parents are largely ‘satisfied’ with the Instructional Process in both the private and public sectors. The climate factor for private schools is higher by a mean, statistically significantly value of 0.41 (ρ < 0.001, d = 0.643).
Qualitative evidence suggests concern regarding ‘high pupil to teacher ratios’ and ‘less efficient use of classroom time’ in the public sector, while parents in both sectors would like to see an ‘improvement in general academic standards, more psychological support for children and a curriculum expanded to include foreign languages and more sport’.
Administration (admin)
This factor is a measure of the schools ability to involve parents in the decision-making process.
The data profile suggests that parents are largely ‘satisfied’ with the school administration with a climate factor result that shows no statistically significant difference between public and private sectors. No additional qualitative evidence was noted in relation to the administration factor.
Parent Climate (index)
The overall climate index represents a mean total of the entire set of parental responses from all schools divided in terms of management type.
The data suggest that parents in both the public and private sectors are largely ‘satisfied’ with school climate. The climate index for private schools is higher than the public equivalent by a mean, statistically significantly value of 0.28 (ρ < 0.01, d = 0.534). Qualitative evidence would suggest that the principal issues in the public sector are related to the institutional environment and academic standards. The principle issue in the private sector is related to academic standards.
School Size-Climate Index Correlation
The aim of this part of the analysis is to assess how school climate perception (climate index) changes in relation to school size. Size is considered a significant factor due to the conspicuous division between public and private institutions; public schools are several orders of magnitude larger than the private equivalent.
Teacher Perception
School climate changes as perceived by the teachers are illustrated in Figure 3. The mean values suggest a decreasing degree of teacher satisfaction in relation to climate as school size increases, bordering on ‘ambivalence’ in the public sector. The most obvious divergence occurs between ‘large’ and ‘very large’ schools or the boundary between private and public institutions. At this point the difference between mean scores is 0.31 and becomes statistically significant (ρ < 0.002, d = 0.7), as is the overall result, that is, climate is related to institutional size and type according to teachers.
Size—climate index (teacher perception)
Parent Perception
School climate changes as perceived by the parents are illustrated in Figure 4. The mean values suggest that parents are largely ‘satisfied’ with service as measured by school climate factors despite the increases in school size. The private sector (small, medium, large schools) however scores marginally better than the public sector (very large schools) in terms of mean values and a ‘cut-off’ also occurs when comparing large private schools with very large public, the difference in mean score is 0.24 and is statistically significant (ρ < 0.009, d = 0.454), as is the overall result that is, climate is also related to size and type according to the parents (and teachers).
Size—climate index (parent perception)
Conclusion and Recommendations
Colombia is a resource rich nation with the world’s 27th largest economy, putting it on par with Sweden and Turkey in absolute terms. However as a consequence of its historical legacy, the nation is riven with gross political and economic divisions that manifest as social instability, poverty and extreme levels of violence (Ortiz, 2002). Education could have a significant role to play in changing this dynamic, increases in human capital producing social benefits in terms of employment, health, mobility and political participation (World Education Report, 2008). However, research evidence suggests that divisions in Colombian society are no more apparent than in the provision of education where the poor are largely consigned to low quality state schools (Heyneman and Todoric-Bebic, 2000; Rivas and Montenegro, 2005).
As a means of expanding access to education beyond the traditional state sector, consecutive conservative governments have applied policy reforms that support the introduction of private investment, including the use of state subsides (PAC) as a vehicle for equality of access. Despite the stated reservations of private education from certain quarters regarding quality and accessibility, research suggests that in the context of the Aguablanca slum district the private sector is delivering a service that is, at the very least comparable with the public sector. In the first instance, Akerman and Alwindi (2007) performed an input assessment producing a positive set of responses for the private sector (when compared to the public sector) for the majority of performance metrics. This second complementary investigation in Aguablanca provided a measurement of organizational effectiveness in the form of school climate assessment. While the results (mean values) point toward a general level of satisfaction in relation to overall provision, an evaluation of teacher and parent climate perceptions would suggest a preference for the low-cost private sector, that is, where a statistically significant difference in quantitative climate scores has been identified, the outcome consistently favours the private schools. A further interpretation of effect size suggests that teacher perception (d > 0.9) of management (centred on the issues of integrity, resources allocation and academic emphasis) is more clearly divided than parental opinion (d = 0.53).
The evidence would therefore suggest that the low-cost private sector has a positive and constructive role to play in the provision of education in the marginalized district of Aguablanca. It is therefore recommended that the government support and consolidate private education facilities for the poor through the continued promotion and possible expansion of the PAC initiative.
In view of the number of children (in the region of 100,000) not attending school in Cali (Hentschel, 2005) there would appear to be a genuine need for additional capacity in Aguablanca, hence local government plans to build a further three very large public schools in the district by 2011. As climate research suggests that teachers and parents are largely satisfied with the way the education system is currently developing, it is recommended that transition planning be designed not to disturb the delicate equilibrium of existing education provision based on a healthy mixture of public and private institutions, that is, avoid a ‘crowding out’ of the private sector.
Finally, the climate instruments have proven to be a convenient and efficient means of assessing school management quality and were well received by all participants. It is therefore recommended that climate assessment be used (in conjunction with input measurement) more widely, as a vehicle of characterizing and understanding the delivery of education throughout the developing world.