Title: Determinants of performance in last grade ICFES test in Medellns schools
1(No Transcript)
21 What is bad performance in basic and secondary
school in Colombia due to?
- - The agency problems in the regulation-control-fi
nancing system, bad school management, and lack
of teachers incentives associated with their
performance (Wiesner, 1997 Misión Alesina, 2001
and Tobón et. al., 2003). - - The need to improve the institutions,
parents-directives-teachers interactions, and the
public school management, which have not been
subjected to measurement (Gaviria and Barrientos,
2001 Nuñez, 2002 Restrepo and Alviar, 2003).
31 What is bad performance in basic and secondary
school in Colombia due to?
- - The fact that when teacher characteristic
indexes, incentives (monetary and non monetary
ones) and school environment area measured, the
school characteristics and students are more
important than the teachers career and the
incentives (Tobón et.al., 2003). - - The latest work indicating that schools can not
compensate the differences in students and
families characteristics (Correa, 2004).
42 Relevant studies indicate that
- - To emphasizes in incentives, autonomy and
decision power of teachers is important, but so
are the home characteristics (Mizala and
Romaguera MR-,2002 Vegas, 2003 MR and
Reinaga, 2004) - - Teaching and pedagogy theories have not been
considered (Scheerens, 1999). - - School and teachers cultural-violence-socioecon
omic context must be considered (Fuller and
Clark, 1994 Heneveld and Craig, 1996, Grogger,
1997 Garner and Raudenbush, 1991) - - Basic and secondary school performance change
due to individual and contextual factors
(Cervini, 2002 y 2005).
52 Relevant studies indicate that
- - Institutions must be understood, as efficiency
in public expenses in education. Achievement
tests are an institutional instrument which
motivates the effort of all the actors toward a
goal (Bishop, 2002). - - The better the institutional environment, the
better the students achievements but it is
better to give group incentives to teachers, for
stimulating a trust and cooperative environment
instead of a rivalry one (Mcmeekin, 2001). - In general, education and socioeconomic
characteristics of the family are more
determinant. School and teacher characteristics
as well as the role of expenses in education are
more controversial. - Researches that include inputs are numerous, but
those including classroom and school managing
processes are more consistent (Piñeros, 2002).
63 Methodology
- Information revelation mechanisms (preferences)
to know the actions and interactions among
teaching process actors. - Hierarchical models. Nesting appears in the
educational process. Variables are grouped
hierarchically which indicate different levels
and the interest settles in searching the
variability due to each one of them. - So the institutional contest can be naturally
introduced.
73. Methodology
- Qualitative variables main components (Catpca).
For transforming qualitative variables into
quantitative ones. - Factor analysis. Permits adding variables to
indexes, due to their correlations, and
diminishes the transformed data variables and
co-variables matrix rank.
84 Used information
- A survey applied to 840 last grade teachers and
directives of 105 schools in Medellín 45
privates and 60 public schools 6 male, 15
female, and 85 male and female ones (mainly
officials) covering 8,345 students (49.5 F y el
50.5 M). - The survey allows us to know managing and
teaching human capital, their interactions,
school institutionally, as well as the
relationships with the local administration. It
also indirectly allows us to know the
nutritional, affective and cognitive conditions
of the students. - ICFES form 2003 on socioeconomic student
conditions. - DANEs form on school resources and
infrastructure.
94 Used information
- Hierarchical level 2. indexes using surveys
applied to teachers and principals were built.
Input variables are individual and standardized - Hierarchical level 1. Variables are individual,
mostly numeric or dichotomy, except those
associated with parent human capital.
105 Indexes construction
11(No Transcript)
12Anova model (Empty model)
13Intra-School Coefficient Correlation (ICC)
- School affects 38 in the differences presented
in the total variation of the school achievement.
14First Level Variable Estimation
15First level variable explained variability
achievement proportion
16First Level significant variables estimation plus
second level variables
17Variable proportion explained by the average
achievement adjusted in schools, compared to the
ANOVA
Unfortunately, the included first level variables
do not contribute much to explain
achievements. Regrettably, the Saber tests for
5th and 9th grade were only recently applied to
all the schools of Medellín. Moreover, the
registration files at the Secretary of Education
contain very little information on students.
18A greater of better infrastructure does not
necessarily help the achievement
19Variability in the marginal contribution on each
explanatory variable
20Variability in the marginal contribution on each
explanatory variable
- When the variability in the marginal contribution
on each explanatory variable over the achievement
depend on each school, the only significant
variable was the economic stratification.
21Definitive hierarchical model
22Adjusted proportion of the variability achievement
The included first level variables do not
contribute much to explain achievements
23Conclusions
- - In Colombia we are subject to scarce
information on processes and conditions of the
students. - - The constructed second level variables are
highly explanatory in opposition to the second
level ones. - - The parent human capital and the initial
student conditions are much more important than
the teacher human capital. - - To have and use resources has a greater
importance, though the information on
infrastructure is bad. - - A fruitful analysis via, in the teaching
economy, is to advance in the designing of
surveys for revealing the institutionally and
governability in the schools.