Determinants of Dropout in Primary Education in Rural Cambodia:
A Multilevel Analysis
1
October 2015
HENG Boret, TANIGUCHI Kyoko and HIRAKAWA YukikoGraduate School for International Development and Cooperation
(IDEC)Hiroshima University, [email protected]
I. Research Background
UNESCO (Education for All- Global Monitoring Report 2014)
“Despite improvements in getting children intoschool, dropout before the last grade remains a serious problem in many low and middle income countries”.
2
I. Research Background (Cont.)
3Source :MoEYS (2009,2010,2011,2012,2013,2014)
94.4 94.8 95.2 96.4 97 95.6
120.2 125.4 116 123.3 123.4 116.1
8.7 8.7 8.3 3.7 10.5
59.33 61.73 61.2
33.1 37.2 36.2
020406080
100120140
NERGERDropout rateSurvival rate grade 1-6Survival rate grade 1-9
II. Literature review (Research method)
4
1. Interview asking: “Why did you dropout?” =“Why do you think…”
DropoutDo you have a TV? Do you have motorbikes?
Family economic status measured by possession of goods
See whether a student have dropped out or not during the research period.
Prospective data
Retrospective data
My family was poor, so I had to work.
2. Survival analysis using prospective data (data before dropping out).
- The answer is an “opinion”. - The answer can include self-
justification, social consideration, etc.
- No opinion. The result is more objective. - Effect size can be identified.- Process of dropout (risk factors, trigger
events and symptoms) can be identified.
II. Literature review (Research method)
• Event history method or survival analysis is considered as the most appropriate for studying a dynamic pattern of dropout (Willett and Singer, 1991).
• There are a series of research on causes of dropout using survival analysis implemented in developed countries.
Eg. Rumberger, R. W. 1983. Dropping out of High School: The Influence of Race, Sex, Family Background. American Educational Research Journal 20 (2), 199-220.
• In developing countries, no ready made data set: it is difficult to obtain reliable data.
5
No Fata a) Area: Kampong Cham Provinceb) Sample: 5 primary schools and
5 lower secondary schoolsG1- 4 276 students (77 dropouts)
3 cohorts: G4-7 282 (124)G7-9 392 (163)
c) Periods: 2008-2011d) Method of analysis: Cox regression (time variant model)
6
II. Literature reviewSurvival analysis in Cambodia
II. Literature reviewSurvival analysis in Cambodia
Outline of findings: significant factors
7
Cohort G1-4 G4-7 G7-9
Divorced parents *** *Age of first entry ** *Relation among students **Absence * * *Ethnicity ***Relative achievement *** **Repetition * **Homework completion *Self-esteem * *Preschool experience * ***
No poverty, no child labor?
II. Literature review
8
Researcher Type of research Type of data Findings Sample
Valesco(2001)
Interview and Focus Group Discussion, Qualitative
Retrospective data
Poverty and necessity to work
6 provinces242 stakeholders
Keng (2003,2004)
Causal Comparative Analysis (2003), Regression (2004)
Retrospective data
Size of farm land (2004) 1 school 31 female students
World Bank(2005)
Correlational Study of Districts Using Household Survey
Retrospective data
Poverty, late school entry, school facilities
All provinces
No (2012)No & Hirakawa (2012)
Survival Analysis, Longitudinal Study(3 years)
Prospective data
G4 and G7- Relative achievement, repetition, and self-esteemD4- Ethnicity and gender
1 Province (Kampong Cham)5 schools for G1 and G45 schools for G7
No et al (2012)
Survival Analysis, Longitudinal Study (1 year)
Prospective data
Both G1 and G5- Relative achievementG1- Parents education
1 Provemce (Prey Veng)5 schools for G1and G5
Ang (2012) Survival Analysis, Longitudinal Study (1 year)
Prospective data
G4: Relative achievement, repetition
1 Province (Battambang)5 schools for G4
II. Literature review
• No’s studies could identify student and family factors, but school effect and school factors could not be identified.
• No’s research found that achievement in class (relative achievement) is important,
but do students dropout
- because they felt disappointed by low acquisition of subject matters or
- because they were discouraged by low rank in class?
9
III. Research objective
• Research questions• Do schools make a difference in the dropout? If yes, how large is it ?
• What school, individual and family factors significantly increase the odds of the dropout?
• Which achievements affect dropout decision? a) Relative achievement measured by class tests and converted to z score orb) Achievement measured by a researcher made achievement test
10
IV. Method of research (sampling)
11
Research Area
Capital City
• Research Area• Preah Net Preah District- Banteay Meanchey
Province• Sangker District – Battambang Province• Thbaung Khmum District – Kampong Cham
Province
• Randomly sampled29 public primary schools30 public lower secondary schoolsStudents in G1, G4 and G7 at the initial data collection (3 cohorts)
IV. Method of research (instrument/implementation)Initial data collection February- April, 2014• Guided interviews for grade 1 and 4 students ⇒ student/family factors
Questionnaire for G7 students• Questionnaires for principals and teachers ⇒ school factors• Observation checklist ⇒ school factors• School record ⇒ relative achievement (student factor)• Researcher made tests (Khmer test and math test) ⇒ achievement (student
factor)• Representative students’ notebook ⇒ day of teacher absence (school factor) The instruments were based on No (2012), except the researcher made test. Follow up data collection February- April, 2015• Name list and interview to classmates 12
IV. Method of research (analysis)
• Multilevel logistic regression (HLM version 7) • As the dependent variable is dichotomous, Bernoulli
model is used for identifying factors of dropout.
13
V. Findings (G4 cohort)
Null model
14See Appendix B
School level, 11%
Student level, 89%
About 11% was explained by school variance.
About 89% was explained by student level.
V. Findings (G4 cohort)• Model 3: Full model (Individual + School)See Appendix
• In school level (Level 2)- Mean socioeconomic status (B = -1.1, OR = 0.333, *p < .05)More students dropout from schools situated in poor community.
- Mean achievement measured by researcher made tests (B = 0.09, OR = 1.094, ***p < .001)
More students dropout from good achievement schools.
Teacher absenteeism was not included in the model. 15See Appendix B
V. Findings (G4)• Model 3: Full model (Individual + School)See Appendix
• In individual level (Level 1)- Educational Aspiration (B = -0.424, OR = 0.655, **p < .01)
Students with high aspiration have lower probability of dropout. - Distance from school to home(B = 0.303, OR = 1.35, *p < .05)Students who live near schools have lower probability of dropout.
- Absolute achievement measured by the researcher made tests was not found significant.
Relative achievement was not included in the model. 16See Appendix B
VI. Further studyFor this study: • Explores other school-level factors that might influence
school dropout in G4.• Explores other individual-level factors that might influence
school dropout in G4, especially relative achievement. • Do analysis of other cohorts (G1 and G7). Derivative topic: The researcher made tests revealed low student achievement. ⇒ Go to Taniguchi, K.’ presentation.
17
ReferencesAng, B. (2012). Duplication study on determinants of dropouts at basic education level at rural schools in Cambodia. Unpublished master thesis, Hiroshima University (HU).Heng, B., Ang, B., & Yukiko, H. (2010). Student and Class Factors Influencing School Dropout in Cambodia (ICER Proceedings 2013). Keng, C. (2005). Decision-making on Education : The role of children1. Education, Southeast Asian Development, International, 5(1), 20–32.MoEYS. (2009). Education Statisitcs & Indicators 2008/2009. Department of Planning.MoEYS. (2010). Education Statisitcs & Indicators 2009/2010. Department of Planning.MoEYS. (2011). Education Statisitcs & Indicators 2010/2011. Department of Planning.MoEYS. (2012). Education Statisitcs & Indicators 2011/2012. Department of Planning.MoEYS. (2013). Education Statisitcs & Indicators 2012/2013. Department of Planning.MoEYS. (2014). Education Statisitcs & Indicators 2013/2014 (Vol. 2014). Department of Planning.No, F. Sam, C., Hirakawa, Y. 2012. Revisiting primary school dropout in rural Cambodia. Asia Pacific education Review 13. pp. 573-581No, F., Hirakawa, Y. 2012. Identifying causes of dropout through longitudinal quantitative analysis in rural Cambodian basic schools. Journal of International Development and Cooperation 19 (1), 25-39.No, F. (2012, September 25). School Dropout at Basic Education Level in Rural Cambodia : A Longitudinal Study. Doctoral dissertation, HU. retrieved from http://ir.lib.hiroshima-u.ac.jp/ja/list/doctoral_thesis/%E5%8D%9A%E5%A3%AB%28%E5%AD%A6%E8%A1%93%29/item/34054Rumberger, R. W. 1983. Dropping out of High School: The Influence of Race, Sex, Family Background. American Educational Research Journal 20 (2), 199-220.Sam, C. (2010). Factors influencing student dropout at rural primary schools in Cambodia. Hiroshima University. Unpublished master thesis, HU.UNESCO. (2012). EFA Global Monitoring Report 2012. Paris.Velasco, E. (2001). Why are Girls Not in School ? Perceptions , Realities and Contradictions in Changing Cambodia, (September 2001).Willett, J. B., & Singer, J. D. (1991). From Whether to When: New Methods for Studying Student Dropout and Teacher Attrition. Review of Educational Research, 61(4), 407–450. doi:10.3102/00346543061004407World Bank. (2005). Quality Basic Education For All. Washington DC 18