How to analyze the KPS

Findings from the
Kindergarten Parent
Survey (KPS) in
Hamilton
How to analyze the KPS:
Some ideas
DAC Network
Marriott Bloor Yorkville, Toronto
October 8, 2010
The Kindergarten Parent Survey
• A wealth of information on:
•
•
•
•
•
•
Parent’s and child’s experiences with child care
Child’s pre-kindergarten program usage
Parent’s and child’s experiences of kindergarten
Developmental activities parent has shared with child
Family’s mobility/stability and neighbourhood
Family demographics (income, education, etc.)
• Linked to the teacher-provided data from the EDI.
• However, completion is voluntary.
The Kindergarten Parent Survey
How to analyze it in a way that is:
• More than merely descriptive?
• Allows for non-response and possible biases?
Some ideas from Hamilton’s 2008 KPS, with
examples.
Scaling
Several KPS questions involve true/false or yes/no
responses to a series of related statements.
Convert these responses into scales by simply
adding up the “true” or “yes” responses (count
“sometimes true” or “yes, once or twice” as 0.5).
Technically: if questions scored 0-2, then SUM(questions)/2
The EDI is similar – it uses 0, 5, 10 and averages,
rather than 0, 0.5, 1 and adding.
Neighbourhood Rating
This scale summarizes responses to the following nine statements (so
scores range from 0 to 9):
• It is safe to walk alone after dark.
• It is safe for children to play outside during the day.
• There are safe parks, playgrounds and play spaces.
• If there is a problem around here, the neighbours get together and deal
with it.
• There are adults my children can look up to.
• People are willing to help each other.
• When I'm away from home, I know that my neighbours will keep their
eyes open for possible trouble.
• You can count on adults to make sure children are safe and don't get in
trouble.
• There are meeting places where parents get together to talk. (optional)
Neighbourhood Rating by Social Risk
Neighbourhood Rating Components
Neighbourhood Rating
(Rotated Component Scores)
1.000
.800
Parks
Children to play
Safety
Walk af ter dark
.600
.400
Neighbours wat ch
out
Adult role models
.200
Count on adults
Help each other
Neighbours deal
M eeting places
.000
.000
.200
.400
.600
Social Cohesion
.800
1.000
Parent engagement with child
In the PAST 7 DAYS, have you or someone close to your child done
the following activities with your child after school or on weekends?
•
•
•
•
Told or read him/her a story
Taught him/her letters, words, or numbers
Taught him/her songs or music
Worked on arts or crafts with him/her
• Played with your child
Õ nearly 100% “yes”
• Your child played with other children (other than siblings)
• Taken him/her along while doing errands like going to the post
office, the bank or grocery store
• Involved him/her in household chores like cooking, cleaning,
setting the table, or caring for pets
Parent engagement with child
35.0%
30.0%
Vulnerable (Hamilton Baseline)
25.0%
20.0%
15.0%
10.0%
5.0%
.0%
4.5 or less
5.0
5.5
6.0
6.5
Parent engagement with child
7.0
7.5
8.0
Parent engagement with child
45.0%
Removing the two “play” items
40.0%
Vulnerability (Hamilton Baseline)
35.0%
30.0%
25.0%
20.0%
15.0%
10.0%
5.0%
.0%
2.5 or less
3.0
3.5
4.0
4.5
Parent engagement with child
5.0
5.5
6.0
Parent engagement with school
Since September 2007, have you or someone close to your child done any
of the following?
• Attended a parent-teacher conference
• In general, we only expect “yes, once or twice”. Unlike other types of
participation, “yes, many times” is generally a bad sign.
• How should we handle this?
• Attended a general school meeting (e.g. open house, school council
meeting)
• Volunteered in the classroom or school (e.g. class trip, helped with
fundraiser, helped in library)
• Attended a school or class event? (e.g. school concert, play)
Parent engagement with school
45.0%
40.0%
Vulnerability (Hamilton Baseline)
35.0%
30.0%
25.0%
20.0%
15.0%
10.0%
5.0%
.0%
0.0
0.5
1.0
1.5
2.0
2.5
Parent engagement with school
3.0
3.5
4.0
Parent engagement with school
45.0%
Reversing the scoring of parentteacher conference (“yes, many
times” = 1, “yes, once or twice” = 2)
40.0%
Vulnerability (Hamilton Baseline)
35.0%
30.0%
25.0%
20.0%
15.0%
10.0%
5.0%
.0%
0.0
0.5
1.0
1.5
2.0
2.5
Parent engagement with school
3.0
3.5
4.0
Pre-K Program Usage
In the year before your child started Senior Kindergarten, how often did your child
do the following activities outside of school hours?
• Play-based Children's Programs (e.g., drop-ins, Ontario Early Years Centres
(OEYC))
• Literacy / Family Reading Programs (e.g., OEYC, Parent and Family
Literacy Centres, library)
• English as a Second Language Programs
• Organized Team Sports (e.g. hockey, soccer, t-ball)
• Physical Activity & Recreation Programs (e.g. swimming, gymnastics, family
skating)
• Children's Club (e.g. Beavers, Sparks, Boys and Girls Club)
• Music, Dance, or Arts Programs (e.g. music lessons, dance lessons, arts
and crafts)
Responses are frequencies (never, less than once a month, about once a month,
about once a week, more than once a week).
Pre-K Program Usage
50.0%
With scores of 0 to 4, to make
“About once a week” (3) = 1, take
SUM(questions)/3. With 7 activities,
this gives a scale from 0 to 9.33.
45.0%
Vulnerability (Hamilton Baseline)
40.0%
35.0%
30.0%
25.0%
20.0%
15.0%
10.0%
5.0%
.0%
0 to 0.33
0.67 to 1.33
1.67 to 2.33
2.67 to 3.33
3.67 to 4.33
Pre-Kindergarten Program Usage
4.67 to 5.33
5.67 or higher
Pre-K Program Usage
45.0%
Alternatively, scale it to time. So
“about once a week” = 1, “about
once a month” = 0.25, etc. With
“more than once a week” = 2, this
gives a scale from 0 to 14.
40.0%
Vulnerability (Hamilton Baseline)
35.0%
30.0%
25.0%
20.0%
15.0%
10.0%
5.0%
.0%
0 to 0.9
1 to 1.9
2 to 2.9
3 to 3.9
4 to 4.9
Pre-Kindergarten Program Usage
5 to 5.9
6 to 6.9
7 or higher
KPS Response Rate 2008
Overall response rate:
53.1% (↑ 18.5%!)
Non-response is not random
Non-response can bias
Non-response, Response Bias and the Kindergarten Parent Survey:
Amongst less vulnerable populations, parents with vulnerable children are much less likely to respond
45.0%
% Vulnerable on any domain (Hamilton Baseline)
40.0%
35.0%
30.0%
25.0%
Responded to KPS
Did not respond to KPS
20.0%
15.0%
10.0%
5.0%
.0%
English (or French) as a Second Language
Not E(F)SL
Vulnerability by Income
Vulnerability by Reported Household Income
45.0%
40.0%
Vulnerable on one or more domains
35.0%
30.0%
25.0%
20.0%
15.0%
10.0%
5.0%
.0%
Less than $17,000
$17,000-30,000
$30,001-50,000
$50,001-75,000
Household Income
$75,001-100,000
More than $100,000
Multiple Imputation
•
Simple imputation replaces missing data with a single number – the
mean of the reported data, an interpolation, or auxiliary data.
•
In EDI analysis, we have seen median household income at the DA
level used as a proxy for individual household income (Oliver, et al.,
2007).
•
In multiple imputation, we use the distribution of family income at the
DA level, and randomly assign levels from that distribution.
•
The key is that we do this multiple times and average across the
results.
Multiple Imputation
•
Using Census or tax file (SAAD) data, calculate the proportion of family
incomes in each of our six categories for each DA (or CT or FSA where
necessary).
•
Assign postal codes to DAs in the EDI with the PCCF.
•
Match the two files.
•
Generate a random number from 0 to 1 for each EDI_id without a KPS
reported income and assign it to the appropriate income category.
•
Repeat at least 5 times.
A low income DA
A middle income DA
A high income DA
0.90
0.90
0.90
0.80
0.80
0.80
0.70
0.70
0.70
0.60
0.60
0.60
0.50
0.50
0.50
0.40
0.40
0.40
0.30
0.30
0.30
0.20
0.20
0.20
0.10
0.10
0.10
0.00
0.00
Under
$17K
$17 to
30K
$30 to
50K
$50 to
75K
$75 to
100K
Over
$100K
0.00
Under
$17K
$17 to
30K
$30 to
50K
$50 to
75K
$75 to
100K
Over
$100K
Under
$17K
$17 to
30K
$30 to
50K
$50 to
75K
$75 to
100K
Over
$100K
Vulnerability by Income
Vulnerability by Household Income
45.0%
40.0%
Vulnerable on one or more domains
35.0%
30.0%
25.0%
20.0%
15.0%
10.0%
5.0%
.0%
Less than $17,000
$17,000-30,000
$30,001-50,000
$50,001-75,000
Household Income
$75,001-100,000
More than $100,000
Vulnerability by Income
Vulnerability by Household Income
45.0%
40.0%
Another way of visualizing it
Vulnerable on one or more domains
35.0%
Imputed Income
30.0%
25.0%
20.0%
Reported Income
15.0%
10.0%
5.0%
.0%
Less than $17,000
$17,000-30,000
$30,001-50,000
$50,001-75,000
Household Income
$75,001-100,000
More than $100,000
How it can make a difference
Vulnerability by JK Attendance by Income
70.0%
Vulnerable on one or more domains
60.0%
50.0%
40.0%
Didn't do JK
30.0%
Did JK
20.0%
10.0%
.0%
Less than $17,000*
$17,000-30,000
$30,001-50,000*
$50,001-75,000
Household Income
$75,001-100,000
More than $100,000
How it can make a difference
Vulnerability by JK Attendance by Income
70.0%
Vulnerable on one or more domains
60.0%
50.0%
40.0%
Didn't do JK
30.0%
Did JK
20.0%
10.0%
.0%
Less than $17,000*
$17,000-30,000
$30,001-50,000*
$50,001-75,000
Household Income
$75,001-100,000
More than $100,000
QUESTIONS?