Conceptual review and meta-analysis of school effectiveness

Reviews of school effectiveness research in developing countries

The most recent review of school effectiveness studies in developing countries is the one by Fuller and Clarke (1994), see table 10.

Table 10: School input and process variables that showed significant positive associations with achievement in at least 50% of the studies in developing countries, analyzed by Fuller and Clarke, 1994^*)

	Number of significant effects divided by the number of analyses
School/teacher factor	Primary Schools	Secondary Schools
School spending Expenditure per pupil Total school expenditure	3/6 2/5	3/5 -
Specific school inputs Average class size School size Availability of textbooks Supplementary readers Exercise books Teaching guides Desks Instructional media Quality of facilities School library Science laboratories Child nutrition and feeding	9/26 7/8 19/26 1/1 3/3 0/1 4/7 3/3 6/8 16/18 5/12 7/8	2/22 1/5 7/13 2/2 - - 0/1 - 1/1 3/4 1/1 1/1
Teacher attributes Total years of schooling Earlier measured achievement Tertiary or teacher college In-service teacher training Teacher subject knowledge Teacher gender (female) Teacher experience Teacher salary level Teacher social class	9/18 1/1 21/37 8/13 4/4 1/2 13/23 4/11 7/10	5/8 1/1 8/14 3/4 - 2/4 1/12 2/11 -
Classroom pedagogy and organization Instructional time Frequent monitoring of pupil performance Class preparation time Frequency homework Teacher efficacy Cooperative learning task student	15/17 3/4 5/8 9/11 1/1 -	12/16 0/1 ½ 2/2 0/1 3/3
School Management School cluster membership Principal’s staff assessment Principal’s training level School inspection visits Tracking or pupil segregation	2/2 3/4 3/4 2/3 1/1	- 0/1 1/2 0/1 -

^*) Source: Fuller & Clarke, 1994.
The review considered about 100 studies and drew upon earlier reviews by Fuller (1987), Lockheed & Hanushek, 1988, Lockheed & Verspoor and an analysis of 43 studies in the period 1988-1992 conducted by the authors themselves.
Only studies that controlled achievement for students’ family background were included; and only significant associations at the 5% level were reported.
What table 10 indicates is, first of all, that there were more studies about primary schools than about secondary schools. Also, financial, material and human resource input variables were investigated more frequently than school and classroom process variables, with the exception of instructional time.
This predominance of relatively easily assessable input characteristics is also evident from table 11 where the number of times a particular variable was included in a total of 43 studies is indicated.

Table 11: The number of times out of a total of 43 studies conducted between 1988 and 1992 (primary and secondary schools taken together) a particular type of school input or process variable was investigated. Source: Fuller & Clarke, 1994

Enrolments/staff School size Class size	6 25
Teacher variables Teacher training Teacher salaries Teacher experience Teacher preparation Teacher efficacy Teacher gender Inservice training	24 3 9 1 1 5 7
Instruction Instructional time Homework Specific pedagogy Testing of pupils	13 3 12 5
School organization Public/Private Tracking Headmaster supervision	4 1 3
Equipment and facilities Library facilities General facilities and equipment	3 15

On the basis of their review of significant positive effects Fuller and Clarke (ibid) conclude that rather consistent school effects can be found in three major areas: availability of textbooks and supplementary reading material, teacher qualities (e.g. teachers’ own knowledge of the subject and their verbal proficiencies) and instructional time and work demands placed on students.

Policy relevant factors that showed inconsistent or lack of effects appeared to be class size and teacher salaries.

The findings summarized in tables 10 and 11 once more underline the predominance of production function type of effectiveness studies in developing countries. Riddell (1997), in a more methodologically oriented review, observes that a “third wave” of school effectiveness research in developing countries is “in danger of being lost without ever having been explored”. By this third wave she refers to, what I have described as “integrated school effectiveness studies”, comprising resource inputs, organizational factors and instructional characteristics, in which multi-level modeling is a vital methodological requirement.

An innovative set of suggestions that Fuller & Clarke develop in their interpretation of the research evidence, is to pay more attention to cultural contingencies when studying school effectiveness in developing countries. Such contingencies might help in explaining why school and classroom level variables “work” in one country but not in the next. They distinguish four broad categories of cultural conditions:

1. 
   the local level of family demand for schooling;
   
2. 
   the school organization’s capacity to respond to family demand “while offering forms of knowledge that are foreign to the community’s indigenous knowledge” (Fuller & Clarke, 1994, p. 136);
   
3. 
   the teacher’s capacity and preference for mobilizing instructional tools;
   
4. 
   the degree of consonance between the teacher’s pedagogical behavior and local norms regarding adult authority, didactic instruction and social participation within the school (ibid, p. 136).
   

These ideas, as well as the appeal to overcome other weaknesses of school effectiveness studies (lack of cost benefit analyses, shortage of longitudinally designed studies) have demanding implications for the design of studies. According to Riddell (1997) Fuller and Clarke fail to present clear research alternatives.