Computer Aided Education in India : A survey of the Azim Premji Foundation’s junior school initiatives

1. Computer Aided Education in India:A survey of the Azim Premji Foundation’s junior school initiatives Joyojeet Pal Joyojeet Pal

2. Project Goal and Objective • Preliminary study of APF’ s Computer Aided Learning Centers (CALCs) program to identify factors influencing computer usage • Social issues • Operational issues • Cognitive issues Children in Cuttack running an OPEPA CD with a tracing game Joyojeet Pal

3. MSR and TIER’s Interest in CALC • Research arm that does not have to have product-orientation • Has a section looking at Technology for Emerging Markets • TIER working on several levels of ICT for development research And my own interest… Joyojeet Pal

4. Research Methodology • Short field visits, interview and observation based • Locations selected from three states on basis of: • Language – variations within Karnataka • Condition of local economy • Stage of the program • Feasibility of research Joyojeet Pal

5. Locations • Orissa – 4 locations • Karnataka – 6 locations • Pondicherry – 1 location • Mumbai – 1 location Joyojeet Pal

6. Locations Joyojeet Pal

7. Scope • 9 observations – most Orissa • 130 interviews – range 3 min – 180 min • 18 schools • 15 HTs / HMs • 28 subject teachers • 7 computer teachers • 27 students • 15 parents • 4 VEC/Panchayat • 21 community • 5 government • 8 administrators/agency Joyojeet Pal

8. Research Approach Joyojeet Pal

9. A Computer Center • This is a computer center from Abishekapakkam in Pondicherry – these typically seat between 3-6 children per PC – notice that the teacher has little involvement in the actual class functioning. The aggressive users (often standing) in this case are students from senior classes sneaking in for extra time Joyojeet Pal

10. Resource-strapped • In Baripada, Mayurbhanj district in Orissa, the local school has three computers, to be spread over 500 students – this numerical equation is not uncommon, consequently, as many as 10 students can be using one computer simultaneously – notice the interactive work here Joyojeet Pal

11. Observations: Social Issues • Clear enthusiasm, despite lack of understanding of computing • VEC/SDMC plagued by political issues • Parents do not question school – no PTA involvement • Children seem to be doing better in areas with higher female parent involvement in PTA, though causality not clear here • Child labour a comparatively minor issue, perceived value from schooling more serious issue • Class retention follows a threshold pattern – to lesser extent among DL families • More homogenous or older communities tend to have comparatively successful CALC programs KEY TAKEAWAY: Village community endorsement of a project is not a stable indicator of the sustainability of a project Joyojeet Pal

12. Parent Profile • Except in the most remote areas – only poorest send kids to govt. schools • Majority are DLs • Females more active in schools (often breadwinners) along the more developed areas • Will question private schools, not govt. Parents in Pondi, peeved when midday meal contained copper wire Farmer affords one child in private school, other in govt. school – works in local VEC Joyojeet Pal

13. Dropout profile • Daily wager parents – extreme poverty (Rs. 20 daily) • Generally construction, household help – in rural livestock, cottage • Seasonal absenteeism highly prevalent • Urban rim more prone to child labour, though inland dropout rate higher • Poor school infrastructure • Abuse Joyojeet Pal Child labourer from Raichur

14. Observations: Operational Issues • Timetables highly improvisational • Head Teacher critical – interest in CALC highly varied • In resource strapped areas (infrastructure and teacher time), junior classes tend to be excluded • Implementing agency (APF, Aptech. Etc.) usually seen as the owner of the project – thus operational role critical • Teachers training for CALC very likely to have different expectations from trainers’ motivations • No effective student testing mechanism currently • Financing patterns ad-hoc for CALC services • Equipment maintenance still an issue KEY TAKEAWAY: Most rural schools are not prepared to function without significant handholding Joyojeet Pal

15. Head Teacher profile • Usually not locals • Generally near retirement • Usually lower academic qual., longer experience • Often face social issues in taking up appointments • Play link role with the government machinery • Tend to multi-task classrooms A Maharastrian head teacher in Raichur, Karnataka Joyojeet Pal

16. CC Instructor profile • From local families (teachers often not) • Reasonable to expect Rs. 1500 pm as saleable wage - BUT • Salaries in risk most places • Use of CCs for additional rev. generation not ensured • Empowerment, respect, openness to work away • Play a tenuous role with the rest of the institution A successful CC Instructor from Udupi Dist. Joyojeet Pal

17. Observations: Curricular / Cognitive • Teacher Difficulties (and student difficulties, largely): • English (Pondi, Karnataka) • Geometry (Orissa, Karnataka) • Labs, Geography (Orissa + Encylopedia demands) • Student Benefits • Match concepts fractions / LCM • Geography (Pondicherry CD popular even in Orissa) • CALC teaching generally adaptive – children positioning, time-sharing • Teachers reinforce classroom standings in CCs • High variance in terms of learning / computing ability • Children easily tire of the repetitive content and want to move to games • Picking up Linux OS quite painless for younger children KEY TAKEAWAY: Actual computer time for children is very low, but learning within the limited access time fairly good – optimization an issue. Joyojeet Pal

18. Seating patterns • No conclusive evidence, but enough to merit further investigation • Using the ANOVA test for Statistical Significance we find: • The correlation between the position occupied by the student during the computer class and • the student’s family’s economic position is statistically significant to over 95.1% • and to a student’s performance in class is statistically significant to over 99.8% Joyojeet Pal

19. Position and Family Affluence Class Size: 21 RIGHT SD=0.83 LEFT SD=0.66 CENTER SD=0.48 Joyojeet Pal

20. Position and Classroom Performance Class Size: 21 RIGHT SD=0.86 LEFT SD=0.82 CENTER SD=0.48 Joyojeet Pal

21. Shuffling seating • Children in positions R2 is the smart kid in class, R1 is average, and C and L1 were among the poorer performers – C being the poorest according to the teacher. Before the seating intervention, R2 was the most active, controlling the mouse and running ahead with the game – all the three remaining were inattentive. There was no dialogue. Joyojeet Pal

22. Recommendations QUICK FIX • Shuffling students on an experimental basis • Single set of instructions per active class, streamlined instead of self-paced CURRICULAR • Quick modules that deal with one specific problem • Adapting to number of users at start-up stage • Use of icons as characters in CD content • Games (incremental scope) ORGANIZATIONAL • Headmaster training – preferably on-site • Door-to-door campaigns on computer proficiency • Coordinators and Computer instructor essential in short term (see Pratham model) LONG TERM • Test CALC using current SSA student evaluation metrics • Increase amount of offline teaching – CALC dependable for fraction of time • Support policy level initiatives – Childcare, English earlier Joyojeet Pal

23. ThanksFor Questions: joyojeet@berkeley.edu Joyojeet Pal

24. Thanks • Sukumar Anikar • Kentaro Toyama • Santhosh R. • Lopamudra Jena • Rajashekhar Pandi • Prema • Shankar Joyojeet Pal