Year 1 - Faculty of Social Sciences

Report
Does tutoring by foster parents help
foster children in primary school make
gains in reading and math?
First-year results from the RESPs for Kids in Care
randomized field trial.
Robert Flynn, Marie-Pierre Paquet,
Robyn Marquis, & Tim Aubry
School of Psychology
& Centre for Research on Educational and Community Services
University of Ottawa
The RESP for Kids in Care Project is funded in part by the
Government of Canada, Canada Education Savings Program,
Human Resources and Skills Development Canada.
Outline
● Purpose of randomized field trial (RFT)
● Research on low educational achievement of children in
foster care
● Methodology:
►
►
►
►
Participants
Interventions
Design
Outcome measures
● Results in year 1:
► Pre-intervention assessment (October, 2008)
► Post-intervention assessment (June, 2009)
● Conclusion:
► Lessons learned
► Improving the foster-parent tutoring intervention
Purpose of randomized field trial (RFT)


Educational achievement of many
children in out-of-home care lags
behind that of age peers in general
population
Goal of RFT: Evaluate whether tutoring
by foster parents can help foster
children in primary school "catch up" in
reading and math
Funding
Project funded (2007-2010) in
part by Canada Education Savings
Program (CESP), HRSDC
 One of 17 CESP-funded projects to
make RESPs more accessible to
disadvantaged young people

Problem of low educational achievement
of young people in care: US research

Excellent review of academic status of
young people in care (Trout et al., 2008):




3X more likely to be in special education
Up to 80% said by teachers to be at risk
academically & performing below grade level
Most in low/low-average range on measures of
academic achievement
Many require intensive academic assistance
Problem of low educational
achievement: UK research

Jackson (2007):




As in US, widespread educational underperformance
Little research being conducted on basic reasons for
"huge and persistent gap in attainment between
care leavers and others"
Much more attention needed on pivotal role of foster
parents in improving educational performance
Failure of care system in UK to put sufficient
emphasis on education also seen in other Englishspeaking countries, including Canada, US, Australia
Problem of low educational achievement:
Canadian research



Few Canadian studies, but results similar to US and
UK
Flynn & Biro (1998): young people in foster care in
Ontario had much higher rates of suspension and
grade retention than age peers in general population
Flynn et al. (2004): In OnLAC sample of young
people in care:


10-15 years of age: 80% scored in same range as
lowest third of general Canadian population on
parental ratings of reading, spelling, math, and overall
5-9 years of age: 78% scored in same range of
lowest third of Canadian population, on same criteria
Towards a (partial) solution?

Given size of problem, surprisingly few
evaluated interventions exist:




Barth and Ferguson (2004) found only 12
intervention studies
Trout et al. (2008) uncovered only 9 others
Tutoring by foster parents may be a
partial solution
Focus is on home rather than school,
but both are needed
Methodology of RESPs for Kids in Care
Project

Participants (N = 77 foster children):



Year 1 (2008-2009):



Young people in care (grades 2-7, ages 6-13) and their foster
parents or kinship caregivers (tutors)
Randomly assigned to control and intervention groups
Control group (n = 35): RESP only
Intervention group (n = 42): RESP + tutoring by foster parent,
(with Maloney’s direct-instruction educational model)
Year 2 (2009-2010):


Year 1 control group: RESP + foster-parent tutoring
intervention
Year 1 intervention group: RESP + a second year of fosterparent tutoring (with Maloney model)
Maloney’s Direct-Instruction
Educational Model
Michael Maloney,
Quinte Learning Centre
Belleville
Methodology of RESPs for Kids in
Care Project (continued):
2008-2009
school year
2009-2010
school year
(30 weeks)
Control group:
Registered Education
Savings Plan
(30 weeks)
RESP ( with orientation)
RESP (with
orientation)
(Levels 1 & 2 of Teach Your
Children To Read Well & math
software)
(RESP)
Intervention group:
Foster parent tutoring
(with Maloney model)
RESP (with orientation)
tutoring
+ foster parent tutoring
in reading and math
RESP (with orientation)
+ foster parent tutoring
in reading and math
RESP (with orientation)
+ foster parent tutoring
tutoring
in reading and math
(Levels 1 & 2 of Teach Your
Children To Read Well & math
software)
(Levels 3 & 4 of Teach Your
Children To Read Well)
Pre-test
Post-test
Follow-up test
Methodology of RESPs for Kids in
Care Project (continued)

Main outcome measures:
Wide Range Achievement Test (WRAT4):
Word reading, reading comprehension,
spelling, & math
 Conners’ short form (CADS-P):
Attention and hyperactivity
 Child Behavior Checklist (CBCL):
Internalizing and externalizing behaviours

Pre-intervention equivalence of
intervention & control groups
Only 1/35 pre-intervention comparisons
between control and intervention groups
was statistically significant
 This was within chance levels
 Thus, randomization "worked", in
creating equivalent pre-intervention
groups

100
95
Control
(n=35)
Intervention
(n=42)
90
85
80
Se
nt
e
at
io
n
lin
g
cu
l
Ca
l
h
M
at
nc
e
C
Sp
el
on
om
pr
eh
R
ea
en
si
di
n
g
75
W
or
d
MEAN STANDARD SCORE
Pre-intervention results (Sept.- Oct., 2008) on
Wide Range Achievement Test (WRAT4)
WRAT4
Year 1 post-intervention results
(June, 2009):



September-October, 2008: we assessed &
randomly assigned:
 42 children to tutoring intervention
 35 children to wait-list control
June, 2009: we re-assessed and compared
results of:
 30 children who had actually received the
tutoring intervention, and
 34 children who remained in control group
No differential attrition: 0/35 differences in
June, 2009, between the groups on the preintervention measures were statistically
significant
Word Reading results at end of year 1.
(Effect size at post-test: Cohen’s d = .19, p = .19 (1-tailed), ns.
Post-test scores were adjusted for pre-test scores and age.)
Mean Standard Score
Tutoring (n = 30)
Control (N = 34)
105
104
103
102
101
100
99
98
97
96
95
Sept.-Oct., 2008
June, 2009
Assessment Occasion
Sentence Comprehension results at end of year 1.
(Effect size at post-test: Cohen’s d = .39, p = .035 (1-tailed).
Post-test scores were adjusted for pre-test scores and age.)
Mean Standard Score
Tutoring (n = 30)
Control (n = 34)
105
104
103
102
101
100
99
98
97
96
95
Sept.-Oct., 2008
June, 2009
Assessment Occasion
Spelling results at end of year 1.
(Effect size at post-test: Cohen’s d = -.04, p = .882 (2-tailed), ns.
Post-test scores were adjusted for pre-test scores and age.)
Mean Standard Score
Tutoring (n = 30)
Control (n = 34)
105
104
103
102
101
100
99
98
97
96
95
Sept.-Oct., 2008
June, 2009
Assessment Occasion
Math Calculation results at end of year 1.
(Effect size at post-test: Cohen’s d = .46, p = .009 (1-tailed).
Post-test scores were adjusted for pre-test scores and age.)
Mean Standard Score
Tutoring (n = 30)
Control (n = 34)
95
94
93
92
91
90
89
88
87
86
85
84
83
October, 2008
June, 2009
Assessment Occasion
Conclusion:
Lessons learned during Year 1

Tutoring by their foster parents enabled foster
children to make statistically significant and
practically important gains in two areas:



Reading: Sentence Comprehension
Math: Math Calculation
In light of different criteria for magnitude of
effect sizes:

Cohen (1992): d of .20 = small, .50 = medium, .8 =
large



Math Calculation: d of .46 is close to medium
Sentence Comprehension: d = .39 is closest to medium
Ferguson (2009): d of .41 = practically significant


Math Calculation: d of .46 = practically significant
Sentence Comprehension: d of .39 = practically significant
Conclusion (continued):
Improving the foster-parent tutoring intervention

Improving implementation of model: e.g.,




Increase intensity of training of foster parents
Increase low year-1 rate of participation by foster
parents in monthly coaching teleseminars
Increase foster parents’ use of child reward system
Evaluate group tutoring as alternative mode
of delivering Michael Maloney’s educational
model:



Tutor groups of 4-5 children in care
Train CAS staff or others (e.g., university students)
as tutors
Train tutors intensively (e.g., during summer)
For further information, please contact:
Lisa Peeke, Coordinator,
RESPs for Kids in Care Project:
[email protected]
(613)562-5800 ext. 8860

similar documents