Presentation on the KiuFunza Initiative by Youdi Schipper of Twaweza. This was presented at the Commission for Science and Technology (COSTECH) in Dar es Salaam, Tanzania, on June 19, 2014, to an audience of researchers.
4. Introduction: Realities of Primary
School Education
2010 2011
Kiswahili 32.7% 28.3%
English 7.7% 11.5%
Mathematics 30.3% 30.4%
Uwezo Results: Percentage of Standard III
pupils who were able to complete the
Standard II exercises by year and subject
6. • Teacher Data
•The average teacher salary is TZS575,000
(€254)/month, pre-tax, inclusive of allowances.
•47% of teachers report that, if they could start
over, they would not choose teaching as a
career.
•1/3 of teachers report having missed class at
least once in the past week.
Evidence from our baseline 2013
7. • Large expansion in primary school enrollment
• High student to teacher ratios
• Teachers often not in class
• Capitation Grant payments fall short of policy
• Poor learning outcomes
• What should be done?
Descriptive data show
9. •A portmanteau of two Kiswahili
words
•Kiu which means thirst
•Kujifunza which means to learn
•Together KiuFunza means thirst to
learn
KiuFunza
10. • Capitation Grant, current policy (input): transfer
payments for inputs directly to school accounts
• TZS 10,000 (€4.4) per student
• Normal policy rules on spending, accountability apply
• Cash on Delivery, policy innovation (teacher
incentive)
• Grades 1, 2 and 3
• Test subjects: Kiswahili, English, Mathematics
• TZS 5,000 (€2.2) per test passed is paid to responsible
teacher; TZS 1,000 (€0.44) per test passed is paid to Head
Teacher
• Combination arm
Intervention Design
11. • 11 Districts, randomly sampled
• 35 schools in each district
• 4 intervention groups in each district
• COD: 7 schools
• CG: 7 schools
• COD & CG: 7 Schools
• Control: 14 schools
RCT Design
COD
Yes No
CG
Yes
7 7
No
7
14
(control)
Number of Schools per
treatment group per district
16. • Appropriate and legitimate
• Goldilocks: Not too hard and not too easy, but Just
Right.
• Based on the Uwezo test, but with more grade-
specific variation.
• According to Tanzanian syllabus
• Test made by university specialists on curricular and
test development
KiuFunza Tests
22. • Test set randomly assigned to student
• We created a Random Numbers Table
• Test set to be administered to student was
determined using the Random Numbers Table
• First letter of student’s last name and
• Second digit of the day, selected test set
• The tests were also conducted one-on-one and
with independent implementers not the
teachers
Test Integrity
23.
24.
25. • Choice between:
• Bank account (teacher liabilities)
• Mobile money account (innovative, but not without
its problems)
• Irregularities in payment information delayed
payment
• Two teachers deceased before payment was
arranged
• Next of kin issues
• Third party payments not allowed
Payment of Incentives
26. • Blamed any dismal results on student anxiety
and fear
• Students are not used to the one-on-one testing style
• Blamed any dismal results on many exogenous
factors
• Rain, testing after end-of-year tests were conducted
• Want teachers themselves to test the students
• Overall positive feedback and feel the tests were
transparent and fair
Teacher Sentiments
27. • KiuFunza is an intervention aimed at improving
learning outcomes in early-grade education in
Tanzania designed as an RCT
• Aim is to demonstrate how current government
policy can be implemented effectively and
• Identify whether paying teachers for the
performance of their students has impact on
learning outcomes
• Measuring and rewarding performance has
practical constraints that can be non-trivial
Take Aways
Expansion in primary school enrollment
2001: abolished primary school fees
2002: Primary Education Development Program (PEDP) established with main goal of achieve universal primary enrollment (UPE)
from 59% (2000) to 97% (2010); 94% (2011)
High student to teacher ratios
from 46:1 to 59:1, with some as high as 74:1
Schools are adopting double-shift schedules to deal with high # of students
Causing decrease in daily classroom hours from 6 to 3.5
Poor Learning Outcomes
Uwezo results
Shortcomings of current CG funding policy
funds reaching schools are well below policy commitment (10,000Tsh) per primary pupil per year and are declining, from TZS. 6,654 (USD 6) in 2004 and TZS. 6,665 in 2009 to less than TZS. 3,000 in 2011 (Education Sector Performance Report 2004, 2009, 2011; see also Reinnika and Svensson, 2004 for Uganda)
http://hakielimu.org/files/publications/document102edu_poverty_reduction_tz_en.pdf
Expansion in primary school enrollment
2001: abolished primary school fees
2002: Primary Education Development Program (PEDP) established with main goal of achieve universal primary enrollment (UPE)
from 59% (2000) to 97% (2010); 94% (2011)
High student to teacher ratios
from 46:1 to 59:1, with some as high as 74:1
Schools are adopting double-shift schedules to deal with high # of students
Causing decrease in daily classroom hours from 6 to 3.5
Poor Learning Outcomes
Uwezo results
Shortcomings of current CG funding policy
funds reaching schools are well below policy commitment (10,000Tsh) per primary pupil per year and are declining, from TZS. 6,654 (USD 6) in 2004 and TZS. 6,665 in 2009 to less than TZS. 3,000 in 2011 (Education Sector Performance Report 2004, 2009, 2011; see also Reinnika and Svensson, 2004 for Uganda)
http://hakielimu.org/files/publications/document102edu_poverty_reduction_tz_en.pdf
We made a big effort during our baseline to inform all stakeholders, at the school, community and government levels, of each intervention. We used many different information sharing techniques, from cartoon booklets sent home with students to formal letters to all levels of government. This was done for a number of reasons:
To ensure government buy in and understanding of the intervention in order to ease scale up efforts if intervention is found to be effective.
Increase accuracy of evaluation by ensuring all stakeholders completely understand the details of the intervention.
CG
Make payments (Tsh 10,000/enrolled pupil) directly into school bank account
2 equal installments of Tsh 5,000/enrolled pupil each
For quality improvement only (in line with current government policy re: use)
Accountability: comply with existing government protocol
Sharing information with communities
COD
Focused on early grade learning (Standard 1, 2 & 3)
Offer of incentive publicised at the beginning of school year/first school visit.
TZS 5,000/child if s/he performs Uwezo like (English, Swahili & Numeracy) tests up to standard
Maximum incentive per child: TZS 15,000
Plus TZS 1,000/child/subject to head teacher
Incentive paid after students assessed at end of yr.
Combo
A result of extensive consultation: incentive alone may not work where inputs are significantly inadequate/unavailable.
COD is implemented on top of the Capitation Grant program
School receive capitation grant for all pupils enrolled at the beginning of the year
COD focused on early grade (STD 1, 2 & 3) also implemented at school.
Monitor results for STD IV and VII: what changed?
Analysis:
Allows for a comparison between the impact of paying for inputs (current policy) and paying an incentive for outputs (innovation), while using combination intervention allows us to measure the interaction effect of both the current policy and the innovation.
This is a mostly representative sample across the country
Random assignment gives high internal validity
Random assignment in a representative sample gives external validity across Tanzania
While the sample is 10 districts, there is implementation in an 11th district outside of the evaluation.