Changes in educational achievement and educational inequality in Lesotho: a relative distribution analysis

Abstract

The fourth goal of the SDGs calls for the increase in access to quality education and the redress of educational inequalities. Monitoring progress towards this goal requires paying attention to changes in educational quality and inequality, not just to changes in average quality. Between 2000 and 2007, Lesotho implemented a multifaceted Free Primary Education programme that included fee eliminations, school-building, and teacher-recruitment components to increase school access and minimise the adverse effects on education quality. During this period, enrolment and average educational achievement increased. However, we do not know whether the increase in average performance was driven by all or just a few gifted students and how it affected educational inequality. This paper fills this knowledge gap by using grade 6 standardised test scores and employing the relative distribution method to analyse changes in educational achievement and educational inequality between 2000 and 2007 in Lesotho. Results show that, although educational achievement of all students increased during this period, much of the increase in overall educational achievement was attributable to improved performance of low- and high-achieving students. This increase in performance at the lower and upper tails of the performance distribution led to an increase in educational achievement and educational inequality, especially in reading proficiency. Further, changes in students’ compositional changes explain the increase in educational achievement.

Keywords:

• Educational achievement
• educational inequality
• relative distribution
• Lesotho

JEL CLASSIFICATION CODES:

• I21 J24

Disclosure statement

No potential conflict of interest was reported by the author(s).

Acknowledgement

I appreciate constructive and detailed comments by the editor (Willem Boshoff) and two anonymous reviewers. I am grateful to Cally Ardington and Patrizio Piraino, and participants at the UNU-MERIT 2nd International Conference on Sustainable Development, and seminars at the National University of Lesotho, and the University of Cape Town for helpful comments and suggestions. All remaining errors are my own.

Notes

1 Educational quality and educational achievement are used interchangeably throughout this paper.

2 Southern and Eastern African Consortium for Monitoring Educational Quality (SACMEQ).

3 Based on calculations from the SACMEQ data described in the data section.

4 These are the two measures of average educational quality developed by the authors that combine grade completion and academic performance in reading and mathematics, respectively.

5 Members are Botswana, Kenya, Lesotho, Malawi, Mauritius, Mozambique, Namibia, Seychelles, South Africa, Swaziland, Tanzania (Mainland), Tanzania (Zanzibar), Uganda, Zambia, and Zimbabwe.

6 For more on the sampling strategy, see (Ross et al., 2005).

7 See Ferreira and Gignoux (2014) and Spaull (2011) on how IRT is used to construct standardised test scores.

8 The covariates used to construct the SES index are parental education, household assets (e.g., radio, TV, car, bicycle, telephone, etc.), whether the household has electricity and piped water, source of lighting, type of material used in the floor, wall, and roof of the house, family structure, and whether the child has regular meals.

9 The “resampled” schools (42 in all) are a panel of schools that have been included in both surveys. Figure A2 shows the test scores’ distributions in 2000 and 2007 by gender of the student. It shows that girls perform better than (same as) boys in English reading (mathematics).

10 This presentation draws heavily from Handcock and Morris (1998, 1999).

11 Ben Jann’s $\mathit{reldist}$

Stata program is used for this analysis (Jann, 2021).

12 Evans and Mendez Acosta (2021) document more evidence about the positive effects of school resources on schooling outcomes.

13 Note that Figures 4(a) and 6(b) are exactly the same as Figures 2(a) and 3(a), respectively.