Upload
pearsondigital
View
1.435
Download
0
Tags:
Embed Size (px)
DESCRIPTION
This presentation was given by Kathy Maksimov, Curriculum Specialist from the Waterford Institute, at the Pacific District Executive Forum on March 11, 2009. The presentation focused on the ability of well-designed instructional technology to replicate teaching best practices across multiple environments and means of measuring program efficacy.
Citation preview
Technology to TEACH
Kathy MaksimovCurriculum Specialist
Modern Computing Technology
Manufacturers needed consistency
Scientists needed perfect recall and delivery
Businesses needed the ability to scale
Turned to Technology
Consistent, replicable
Perfect recall and delivery
The ability to scale
As educators, we need consistency, perfect recall, the ability to scale, and
…
… the time to focus on just one child.
Our StudentsPre-Literacy Training by First Grade
Marilyn Jager Adams, Beginning to Read, 1990
Middle Class
Low Income
3000 Hours
200 Hours
Our StudentsEarly Predictor
Our Students
- Hart and Risley, Meaningful Differences (1995)
Welfare Parents
13 million words
2:1 negative to positive
Working Class Parents
26 million words
2:1 positive to negative
Professional Parents
45 million words
6:1 encouraging
Vocabulary at age 4
A normal classroo
m
Average Students
Exceptional Students
Troubled Students
The Average School Day
- Eaton H. Conant (1973)
Time at school
Actual Instructio
n
Individualized
Instruction
7 Hours
2 Hours
1 Minute!
“The Work Problem”
w = p*ew is the work produced
by a system
p is the potential of the system to create work
e is efficiency of the system in creating work
Efficiency (e)
• How well workers work.• Maximum = 100%• Examples:
– Lesson manuals– Professional development– Mastery learning– Managed schools, charter schools, etc.– Accountability– Grouping students
Potential (p)
• Workers and tools• Limited only by the
worker or tool• Examples:
– Teacher– Paraprofessionals– Manipulatives– Chalkboards or digital whiteboards– Books– Software …
w = p*eExample: Digging a Foundation
The Work Problem
What if your students need this much work?
An Ideal Solution?• Do I scale?• Can I
individualize?• Am I interactive?
• Perhaps you just need more of me …
• Costs too much …• Can’t find enough experts
…• Too hard to consistently
train existing resources …• Not enough time …
Why can’t we solve the work problem?
The Ideal Solution Described
• Scalable• Affordable • Perfect recall• Consistently replicable• Never tired, impatient, or frustrated• Always the very best performance• Constantly improving
What if your students need this much work?
The Work Problem …
The Work Problem … Solved!
Technology fundamentally changes potential …
Communication example:1. Pony Express2. Telegraph (45.4 million
times faster than a horse)
3. Telephone4. Radio5. Television6. Optical fiber / Internet
Musical Performance
A pioneer father goes to see a concert …
Teaching is a Performance
What if you could capture and always deliver the best teaching?
The Formation of Waterford - Dusty’s Epiphany
Moore’s Law
And so on…
Doubling Checkmate
$184,000,000,000,000,000.00
In the beginning, most people only saw
a penny.
Decision Science
Learner Profiles
• Can we apply decision science in education?
Approximation to Precision
Linear vs. Exponential Growth
Source: Kurzweil 2005, The Singularity
is Near
Leveraging Technology Requires …
•Commitment
•Willingness to change …–How we view the classroom
–How we view the role of the teacher
–What we teach and when
–How we use student data
Leveraging Technology Delivers …
the very best education
individualized
for each student on the curve
“It’s independently run. I turn it on in the morning and it pretty much through the rest of the day, gives them their time on it and evaluates where they need to be the next day.”
- Shannon Skipper, Pre-K Teacher, Gadsen, Alabama
A Teacher’s Perspective …
Teacher Experience
> 450 Hours of Instruction
Ways Programs Deliver Instruction
Menu Linear / Predetermined
Adaptive (Mastery-based)
1
2
3
4
From Chutes and Ladders by Milton Bradley
Waterford Delivers Instruction
Automatically individualized for
each student
From Chutes and Ladders by Milton Bradley
Waterford’s Sequencing
Lesson
Pre-assessmentSongBook
InstructionPractice
Extended PracticeAssessment
Did the student master the learning objective?
Sequencing within a Lesson
“Successful” SaraContinue to the next lesson
Sequencing within a Lesson
“Needs Help” Sam
Mark this lesson to automatically try again later
Sequencing between Lessons
1 2 3 1 4
1 2 3 2 1
Automatic Review
Automatic Review
Try Again
“Successful” Sara
“Needs Help” Sam
Teacher Reports
• Averages• Student
progress, usage, and skill performance
• Highlighted areas of concern
Meet “Miss Waterford”
• One-on-one instruction tailored for each student …
… Proven methods, endlessly patient, FUN, responsive, private, equitable
The Data
1. Understand types of efficacy studies
2. Review examples with Waterford
Understanding Terms
• Random Assignment: – a technique for assigning subjects to
different treatments (or no treatment).
• Control Group– the group that does not receive the new
treatment being studied.
Study Designs
Quasi Experiments
Pretest Posttest Nonequivalent Group.
– Control and treatment– Group assignment by convenience – Pretest and posttest
Commons Lane Elementary
• Pre and Posttest (Terra Nova)• Participants (K and 1st):
– Commons Lane (treatment) – 20 students per class; approx. 80
– Halls Ferry (control) – 13 students per class; approx. 80
• Non-equivalents– Class size (favors the control)– Pretest scores (Commons Lane kindergarteners had
lower pretest scores)
Commons Lane – Kindergarten Results
Commons Lane =
3 times the gains!
Commons Lane – 1st Grade Results
0
5
10
15
20
25
30
35
Gains
ControlExperiment
Commons Lane = 2 times the gains!
Hecht and Close (Florida)
• Pre and Posttest• Participants: inner city & rural public
schools with low SES (“at risk” students)– 42 Kindergarteners (treatment)– 34 Kindergarteners (control)
• Treatment:– Six months on Reading Level One
Hecht and Close - Definitions
• Effect sizes (ES): tell how different two groups are.– ES = 0.2: small difference – ES = 0.5: medium difference– ES = 0.8: large difference
• Finding: Best Predictors of Future Reading Ability– Segmenting and blending phonemes
Hecht and Close - Results
Segmenting ES = 1.14 Blending ES = 1.13 Word Reading ES = 1.11 Invented Spelling ES = 1.19
Print Concepts Letter Name Letter Sound Letter Writing
“… Computer assisted instruction provides a cost effective way to teach at-risk children.”
Hecht and Close - Results
• Confident in Results:– Exact amount of time
each student used Waterford
– Computer delivers identical experience
– Individualized– Reports show exactly
how students performedStudying Computer-Based
InstructionStudying Classroom Instruction
Hecht and Close
Sited in:Developing Early Literacy: Report of the National Early Literacy Panel
“Found that … Amount of exposure children had to [Waterford] contributed to individual differences in phonemic awareness and spelling.”
Quasi Experiments
• Time Series Designs. – One group of subjects– Pretested and posttested at different intervals. – The purpose might be to determine long-term effect of
treatment and therefore the number of pre- and posttests can vary from one each to many.
Hillcrest Elementary
• Demographic:– Title 1 School (low SES)
• Before Waterford:– Below district average reading scores– 75% of students were in two lowest
reading categories: below basic (more then 50%) and basic
Hillcrest Elementary - Results
• Two years after Waterford: trend reversed.– 75% students in top two categories: Proficient
and Advanced
• Three years after Waterford:– the first class to use all three levels of
Waterford from kindergarten to second-grade reached the third-grade and had the highest reading scores of all 36 schools in the district!
0
5
10
15
20
25
30
35
40
1995 1997 1998
BelowBasicProficientAdvanced
Hillcrest Elementary - ResultsKindergarten
Student Rankings on the Utah State Core Assessment Test
Nu
mb
er
of
Stu
den
ts
1996*
State of Idaho
• Participants: 8 Idaho School Districts– 3,394 students
(treatment)– 2,413 students
(historical control)
• Test scores (IRI) over 4 years
State of Idaho – ResultsMore Use = Higher Gains
3637
3839
44
49
34
36
38
40
42
44
46
48
50
Control 0–1000 1001–1500
1501–2000
2001–2500
2500+
Usage (Minutes)
Av
era
ge
Ga
in
Waterford recommends 15 min per day = 2250 min
State of Idaho - Results
• Lowest third (at-risk) experienced the most gains (>1.0 effect size).
• Finishing the level had a larger effect size than SES, motivation, and tutoring.
A Midwest School (Indiana)
• Pretest and Posttest• Participants:
– 46 first grade students (year 2001) - Treatment
– 47 first grade students (year 2000) - Control
• Historical control is nice because it reduces the variance from teachers.
A Midwest School - Results
• Evaluated students by how they performed on the pretest:– High scores– Moderate scores– Low scores
• All treatment students outperformed their control counterparts … but the low treatment outperformed the moderate control on the posttest!
A Midwest School - Results
660
640
620
600
580
560
540
520
500Grade 1 Grade 2
Control – High
Control – Moderate
Control – Low Exp – High
Exp – Moderate
Exp – Low
True Experiments
• “The Gold Standard”• Random treatment and
control• Testing to measure change
in both groups• Only research method that
can adequately measure the cause and effect relationship
True Experiments
• Post Equivalent Groups.– Treatment and control– Randomized assignment to groups– Posttest administered to measure
differenceR = Randomized participants
N = Not-randomized participants
O = Test
X = Treatment
True Experiments
• Pretest Posttest Equivalent Groups– Treatment and control– Randomized assignment to group– Pretest to measure difference before the
study takes place– Posttest to measure effect of treatment
R = Randomized participants
N = Not-randomized participants
O = Test
X = Treatment
True Experimentsin Education
One review showed that not even 1 percent of dissertations in education or of the studies archived in ERIC Abstracts involved randomized experiments. http://www.hoover.org/publications/ednext/3384446.html
True Experimentsin Education
Challenges for True Experiments in Education
• Random sample– Parents– School staff / Well-
meaning teachers
• Fidelity of implementation– Teacher abilities– Classroom set up – Scheduling
Tucson – Math and Science
• Participants:– 5 Title 1 schools in Tucson Unified School
District• Free and reduced lunch rate 88.5%-97.5%• 22 classrooms• 338 students total
– Treatment and Control– Random assignment of classrooms– Pretest and Posttest
• SAT10 Math and the environment (science) tests
Tucson Results by classroom
School Gain Diff Free & Reduced Lunch
ELL
A WEMS Control
8.84 .99
7.85 88.5% 16.0%
B WEMS Control
13.43 2.99
10.44 90.6% 39.2%
C WEMS Control
14.23 2.86
11.37 98.3% 22.5%
D WEMS Control
13.02 6.92
6.10 97.5% 49.0%
E WEMS Control
6.21 1.26
4.95 92.6% 29.1%
Tucson Results define terms
• NCE (Normal Curve Equivalent)– Where a student falls on a normal curve
• Indicates a student’s rank compared to other students on the same test
– Range from 1-99 with mean of 50– In a normally distributed population, if all
students make exactly one year of progress, NCE gain would be zero even though raw score increased.
Tucson Resultsby subject
Math
Science
Math - Tucson Results
35
37
39
4143
45
47
49
51
Pretest Posttest
Me
an
NC
E
Waterford
Control
Science - Tucson Results
35
37
39
41
43
45
47
49
Pretest Posttest
Me
an
NC
E
Waterford
Control
Tucson Resultsby genderTucson Results - by Gender
11.71
10.00
5.84
0.45
0.00
2.00
4.00
6.00
8.00
10.00
12.00
14.00
Boys Girls
Mea
n N
CE
Gai
nWaterford
Control
Tucson Resultsby ELL status
Tucson Results - ELL Status
0
2
4
6
8
10
12
14
16
18
20
Waterford ELL Waterford Non ELL Control ELL Control Non-ELL
Mea
n N
CE
Gai
n
Tucson Resultsby ELL status
Tucson Results - ELL status
30
35
40
45
50
55
Pretest Posttest
Mea
n N
CE
Waterford ELL
Waterford Non ELL
Control ELL
Control Non-ELL
Waterford ELL students had the lowest pretest scores and the highest posttest scores!
Qualitative Research• Uses “naturalistic” methods
– interviewing– observation– focus groups
• No statistical or quantitative procedures
• Goals – behavior in natural setting– perspective of the research
participant– meanings people give to their
experience
Madisonville Consolidated Independent School District
• Teachers report higher interest in reading– They report that children now argue over who is allowed to go
to the reading centers, when previously there was little interest shown in reading activities.
• Teachers report improved home/school connection; parents support program 100% (survey)
• Increased student academic self-esteem• Waterford supports and supplements existing
curriculum• Waterford is user friendly• Anecdotes of improved phonemic awareness and
reading readiness skills
Questions?