Persistence and Fadeout of Preschool Effects
Persistence and Fadeout of Preschool Effects: Evidence from 16 LMICs CIES Annual Conference Mexico City, 28 March 2018
Amber Gove, PhD @AmberGove
1
RTI International is a registered trademark and a trade name of Research Triangle Institute.
www.rti.org
Acknowledgements ▪
Presentation derived from paper (in press) ▪ Early Grade Reading Assessment Datasets, 2008-16 from: Ethiopia Ghana Indonesia Iraq ▪
Jordan Kenya Malawi Mali
Nicaragua Nigeria Philippines Rwanda
Senegal Tanzania Uganda Zambia
Data collected by local teams in partnership with student, teachers, staff and government officials; mostly grade 2 and/or 3 ▪ Co-authored by: –
– – – 2
Eileen Dombrowski Simon King Jennifer Pressley Tara Weatherholt
Context: The world is getting better, but we have some work to do ▪
Good News: – Global under-5 deaths reduced by half since 1990 – Less disease, war, poverty – https://www.gatesfoundation.org/Who-WeAre/Resources-and-Media/Annual-Letters-List
▪
Bad News: – 250 million children NOT reaching: ▪ developmental potential (Lancet 2016) ▪ basic proficiency in reading or math (UNESCO 2014)
SDG 4: Quality Education Indicator 4.1.1 Proportion of children and young people: (a) in grades 2/3; (b) at the end of primary; and (c) at the end of lower secondary… achieving at least a minimum proficiency level in (i) reading and (ii) mathematics, by sex Indicator 4.2.1 Proportion of children under 5 years of age who are developmentally on track in health, learning and psychosocial well-being, by sex Source: http://unstats.un.org/sdgs/metadata/ 5
Mind the Gap
6
(Black, Gove & Merseth, 2017)
Research Questions Three driving (sets of) questions: 1. Does participation in preschool improve later reading outcomes?
7
2.
3.
If so, why? What are the characteristics of preschool systems that seem to have persistent impact on student reading performance?
How can we improve our answers to these questions? What additional data are needed to better inform policy and practice?
Early Grade Reading Assessment (EGRA) ▪
Components aligned with essential reading skills
▪
Administered orally one-on-one with a child
▪
Feasible, actionable and adaptable to multiple alphabetic languages
▪
Complementary questionnaires for teachers, directors, classroom observations –
Child question: Did you attend preschool?
▪
Funded by USAID and World Bank, developed by RTI; open-source
▪
Expert panel: Brown University, Center for Applied Linguistics, AIR, Pratham, Texas A&M, University of Ouagadougou, NICHD/NIH, University of Massachusetts, University of Dundee, University of Texas/Austin, International Reading Association
8
Summary of datasets Country Ethiopia Ghana Indonesia Iraq Jordan Kenya Malawi Mali Nicaragua Nigeria
Level of representation Regional National Project National National Project National National National 4 states: Jigawa, Kaduna, Kano and Katsina Philippines National Rwanda National Senegal National Tanzania National Uganda Regional Zambia National 9
Year 2014 2015 2014 2012 2014 2013 2012 2015 2008 2014
Grades 2 and 3 2 2 2 and 3 2 and 3 1 and 2 2 and 4 2 and 4 2, 3 and 4 2 and 3
No. of students 2,000 7,311 4,812 1,153 2,935 4,222 5,240 2,393 6,649 2,531
2013 2011 2009 2014 2016 2014
3 4 and 6 3 2 1 and 2 2
2,463 840 687 2,266 3,720 4,850
Model Specification
Dependent
Correct Letter Sounds Per Minute (CLSPM) Correct Letter Names Per Minute (CLPM) Oral Reading Fluency (ORF)
Independent
Preschool participation as reported by the child
Controls
Gender Language of child match to language of instruction Socioeconomic status (Quartiles) Location of school (Urban vs Rural)
10
Results
11
Effect of preschool participation on early primary reading skills Country
Indonesia† Jordan Senegal† Tanzania† Ghana Zambia Philippines Uganda Mali† Nigeria Kenya Malawi† Iraq† Rwanda Nicaragua† Ethiopia† 12
CLPM† or CLSPM Attended Intercept Preschool 7.44*** 53.24*** 4.78** 20.99*** 5.33* 25.59*** 2.99*** 1.45* 1.80** 9.21*** 1.99*** 2.52*** 1.77 15.41*** 0.93** -0.08 1.68 10.36*** 1.21 1.13 0.85 18.22*** 0.28 1.01 -0.03 11.08*** -0.15 11.84*** -0.85 36.81*** -2.68 24.94***
Oral Reading Fluency Attended Intercept Preschool 12.83*** 18.60*** 6.23*** 5.38** 5.05 18.25*** 4.67** 8.23*** 1.44** 1.88* 0.64 1.24 8.80*** 46.74*** 0.19 -0.25 1.94 3.25*** 1.13 .30 0.20 7.53*** -0.86 -0.45 -2.40 5.71*** 3.87 10.15*** -0.94 40.03*** 1.66 5.55
Relationship between preschool participation rate and letter sound score effect
13
Relationship between preschool participation rate and letter identification effect
14
Relationship between preschool participation rate and ORF effect
15
Answers to Research Questions 1.
Does participation in preschool improve later reading outcomes? Sometimes yes. • 7 countries: letter skills • 5 countries: ORF
16
2.
If so, why? What are the characteristics of preschool systems that seem to have persistent impact on student reading performance? Can’t tell for sure from these data but, some ideas…
3.
How can we improve our answers to these questions? What additional data are needed to better inform policy and practice? Longitudinal data System data Household data
Summary ▪
Systems with higher quality and higher levels of access seem to show preschool effects (Indonesia, Jordan, Philippines) ▪ Contrary to results from HIC, effects persist in through the middle of primary ▪ Need to better understand… – – –
17
How and why lower versus higher level skills are sticking (or not) What the distribution of skills looks like and how that may affect the results How to collect and act on these data in a timely fashion
More Information Amber Gove Director, Research International Education RTI International [email protected] @AmberGove shared.rti.org earlygradereadingbarometer.org globalreadingnetwork.net
18
Amber Gove, PhD @AmberGove
1
RTI International is a registered trademark and a trade name of Research Triangle Institute.
www.rti.org
Acknowledgements ▪
Presentation derived from paper (in press) ▪ Early Grade Reading Assessment Datasets, 2008-16 from: Ethiopia Ghana Indonesia Iraq ▪
Jordan Kenya Malawi Mali
Nicaragua Nigeria Philippines Rwanda
Senegal Tanzania Uganda Zambia
Data collected by local teams in partnership with student, teachers, staff and government officials; mostly grade 2 and/or 3 ▪ Co-authored by: –
– – – 2
Eileen Dombrowski Simon King Jennifer Pressley Tara Weatherholt
Context: The world is getting better, but we have some work to do ▪
Good News: – Global under-5 deaths reduced by half since 1990 – Less disease, war, poverty – https://www.gatesfoundation.org/Who-WeAre/Resources-and-Media/Annual-Letters-List
▪
Bad News: – 250 million children NOT reaching: ▪ developmental potential (Lancet 2016) ▪ basic proficiency in reading or math (UNESCO 2014)
SDG 4: Quality Education Indicator 4.1.1 Proportion of children and young people: (a) in grades 2/3; (b) at the end of primary; and (c) at the end of lower secondary… achieving at least a minimum proficiency level in (i) reading and (ii) mathematics, by sex Indicator 4.2.1 Proportion of children under 5 years of age who are developmentally on track in health, learning and psychosocial well-being, by sex Source: http://unstats.un.org/sdgs/metadata/ 5
Mind the Gap
6
(Black, Gove & Merseth, 2017)
Research Questions Three driving (sets of) questions: 1. Does participation in preschool improve later reading outcomes?
7
2.
3.
If so, why? What are the characteristics of preschool systems that seem to have persistent impact on student reading performance?
How can we improve our answers to these questions? What additional data are needed to better inform policy and practice?
Early Grade Reading Assessment (EGRA) ▪
Components aligned with essential reading skills
▪
Administered orally one-on-one with a child
▪
Feasible, actionable and adaptable to multiple alphabetic languages
▪
Complementary questionnaires for teachers, directors, classroom observations –
Child question: Did you attend preschool?
▪
Funded by USAID and World Bank, developed by RTI; open-source
▪
Expert panel: Brown University, Center for Applied Linguistics, AIR, Pratham, Texas A&M, University of Ouagadougou, NICHD/NIH, University of Massachusetts, University of Dundee, University of Texas/Austin, International Reading Association
8
Summary of datasets Country Ethiopia Ghana Indonesia Iraq Jordan Kenya Malawi Mali Nicaragua Nigeria
Level of representation Regional National Project National National Project National National National 4 states: Jigawa, Kaduna, Kano and Katsina Philippines National Rwanda National Senegal National Tanzania National Uganda Regional Zambia National 9
Year 2014 2015 2014 2012 2014 2013 2012 2015 2008 2014
Grades 2 and 3 2 2 2 and 3 2 and 3 1 and 2 2 and 4 2 and 4 2, 3 and 4 2 and 3
No. of students 2,000 7,311 4,812 1,153 2,935 4,222 5,240 2,393 6,649 2,531
2013 2011 2009 2014 2016 2014
3 4 and 6 3 2 1 and 2 2
2,463 840 687 2,266 3,720 4,850
Model Specification
Dependent
Correct Letter Sounds Per Minute (CLSPM) Correct Letter Names Per Minute (CLPM) Oral Reading Fluency (ORF)
Independent
Preschool participation as reported by the child
Controls
Gender Language of child match to language of instruction Socioeconomic status (Quartiles) Location of school (Urban vs Rural)
10
Results
11
Effect of preschool participation on early primary reading skills Country
Indonesia† Jordan Senegal† Tanzania† Ghana Zambia Philippines Uganda Mali† Nigeria Kenya Malawi† Iraq† Rwanda Nicaragua† Ethiopia† 12
CLPM† or CLSPM Attended Intercept Preschool 7.44*** 53.24*** 4.78** 20.99*** 5.33* 25.59*** 2.99*** 1.45* 1.80** 9.21*** 1.99*** 2.52*** 1.77 15.41*** 0.93** -0.08 1.68 10.36*** 1.21 1.13 0.85 18.22*** 0.28 1.01 -0.03 11.08*** -0.15 11.84*** -0.85 36.81*** -2.68 24.94***
Oral Reading Fluency Attended Intercept Preschool 12.83*** 18.60*** 6.23*** 5.38** 5.05 18.25*** 4.67** 8.23*** 1.44** 1.88* 0.64 1.24 8.80*** 46.74*** 0.19 -0.25 1.94 3.25*** 1.13 .30 0.20 7.53*** -0.86 -0.45 -2.40 5.71*** 3.87 10.15*** -0.94 40.03*** 1.66 5.55
Relationship between preschool participation rate and letter sound score effect
13
Relationship between preschool participation rate and letter identification effect
14
Relationship between preschool participation rate and ORF effect
15
Answers to Research Questions 1.
Does participation in preschool improve later reading outcomes? Sometimes yes. • 7 countries: letter skills • 5 countries: ORF
16
2.
If so, why? What are the characteristics of preschool systems that seem to have persistent impact on student reading performance? Can’t tell for sure from these data but, some ideas…
3.
How can we improve our answers to these questions? What additional data are needed to better inform policy and practice? Longitudinal data System data Household data
Summary ▪
Systems with higher quality and higher levels of access seem to show preschool effects (Indonesia, Jordan, Philippines) ▪ Contrary to results from HIC, effects persist in through the middle of primary ▪ Need to better understand… – – –
17
How and why lower versus higher level skills are sticking (or not) What the distribution of skills looks like and how that may affect the results How to collect and act on these data in a timely fashion
More Information Amber Gove Director, Research International Education RTI International [email protected] @AmberGove shared.rti.org earlygradereadingbarometer.org globalreadingnetwork.net
18
