Modeling Across the Curriculum

bagimpertinentΠολεοδομικά Έργα

16 Νοε 2013 (πριν από 3 χρόνια και 1 μήνα)

41 εμφανίσεις

Modeling Across the Curriculum

Paul Horwitz, Principal Investigator

Co
-
PI’s: Janice Gobert, Research Director

Bob Tinker &

Uri Wilensky, Northwestern


Other senior personnel:

Barbara Buckley, The Concord Consortium


Chris Dede & John Willett, Harvard University


For more on The Concord Consortium visit
www.concord.org



Funded by the the National Science Foundation and the U.S. Dept. of Education

under a grant awarded to the Concord Consortium (IERI #0115699).

Any opinions, findings, and conclusions expressed are those of the presenters

and do not necessarily reflect the views of the funding agencies.

http://ccl.northwestern.edu

INE/IKIT themes addressed by Modeling
Across the Curriculum (MAC)


Building on intuitive understandings
--
MAC’s
representations leverage from students’ physical intuitions.


Focus on idea improvement
--
MAC focus on progressive
model
-
building.


Comprehending difficult text as a task for collaborative
problem
-
solving
--
Scaffolding difficult learning tasks
(MAC).


Controlling time demands of on
-
line teaching and
knowledge
-
building

Scaffolding knowledge integration
(model
-
building) and transfer
(
MAC).


Project Summary


Context
:

IERI

program

emphasizes

scalability,

“evidence
-
based”

research,

and

emphasis

on

diverse

populations
-

No

Child

Left

Behind

(NCLB)
.



Four

levels

of

studies
-



Level

1
-

focus

on

improving

the

scaffolding

design

through

individual

interviews

of

students

and

teachers
.



Level

2
,

classroom
-
based

studies

to

evaluate

the

impact

of

amount

of

scaffolding
.



Level

3

is

a

longitudinal

study

of

a

3
-
year

implementation

of

materials

in

the

Partner

Schools
.



Level

4
-

we

address

how

this

technology

can

be

scaled

to

include

many

more

schools
.



Project Summary (cont’d)

Doing

this

work

in

three

areas

of

high

school

science
:

Genetics

(BioLogica),


Newtonian

Mechanics

(Dynamics),

and

Gas

Laws

(Connected

Chemistry)


Our

models

in

each

of

these

domains

are

hypermodels
-

models

that

incorporate


core

science

content

that

students

learn

through

exploration

and

scaffolded

inquiry
.


More

about

this

later
.


We

apply

Pedagogica

a

powerful

engine

that

~

drives

all

three

software

tools,


provides

embedded

guidance

and

assessment,

controls

all

aspects

of

the

learners’

interactions

with

the

software

tools

by

changing


the

nature

of

the

scaffolding

and

the

assessments
.



Pedagogica

can

automatically

report

student

progress

through

these

lessons

via

the


Internet,

providing

real
-
time,

fine
-
grained

data

on

student

learning
.

Screen shot from connected Chemistry~

Pressure in a Rigid Box

QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture.
Graphics screen

Monitors

Settings,

Operations

Plots

Research: Level 1
-

Case Studies
with students

Case studies of students with
software tools to assess

conceptual progression of concepts (progressive
model
-
building), development of scaffolding framework
(more later), and HCI issues. Tools:

BioLogica (formerly GenScope, teaches Genetics)

Dynamica (teaches Newtonian /Mechanics)

Connected Chemistry (teaches Gas Laws)


Student Data collection with surveys for case studies:



Science learning survey (mix of items from Schommer, and
items constructed by us).



Students’ Epistemology of Models (Gobert & Discenna, 1997).



Level 1 (cont’d): Teachers

Teacher Data collected with surveys ~
science
teaching style, epistemological understanding, science
“comfort” level, pedagogy with modeling.


Surveys ~



Teachers’ epistemologies of models (adapted from Gobert
& Discenna, 1997)



Teachers’ science teaching survey (adapted from Fishman,
1999) and teachers’ background questionnaire (The CC
Modeling Team).




Research: Level 2
-
Classroom

Years 1
-
2



1) Classroom studies of scaffolding with software tools.


2) Testing out reliability and validity of Science Learning Survey


3) Attempt at developing a quantitative form of the epistemology of
models survey.


Decided to use instead:



VASS
-
views of science survey (cognitive and scientific dimensions;
Halloun and Hestenes, 1998); one form for each biology, physics, and
chemistry.


Students’ epistemologies of models (SUMS, Treagust et al, 2002, adapted
from Grosslight et al, 1991).






Scaffolding Framework for Learning with Models

Effects of epistemology


MBTL and cognitive affordances focus primarily on factors dealing with student’s
cognitive processing but...


Also important are students’ epistemological understanding of the nature of
models and the nature of science, both of which have been found to affect their
success in building models of phenomena (Gobert & Discenna, 1997) and their
knowledge integration (Songer & Linn, 1991).


Specifically, learners who have a sophisticated view of the nature of models
generally outperform students who have less sophisticated views (Gobert &
Discenna, 1997) and can use their epistemological understanding to drive deeper
content understanding (Gobert, in preparation).


With the survey data and data from log files we hope to be able to detect
differences in students’ use of MAC activities depending on their epistemologies of
models, e.g., those with more sophisticated epistemologies may use different
knowledge acquisition strategies and model
-
building strategies (log files can
provide an index of this). Example haphazard versus systematic experimentation in
BioLogica.




Model
-
Based Learning in situ

Intrinsic Learner
Factors


Epistemology of models

Attitudes & Self
-
efficacy




Intrinsic Teacher Factors

Epistemology of models

Teaching experience

Background


Classroom Factors

Implementation of MAC activity use (logged)

Teacher practices (reported via Classroom
Communique)


Research: Level 3
-

Longitudinal

D.V.’s
-

Cumulative gains on students’ content knowledge,
modeling skills, epistemological knowledge, and
attitudes towards science.





Research: Level 3
-

Longitudinal



In

September

2003
,

we

began

a

longitudinal

study

of

three
-
year

implementations

of

project

materials
.

The

longitudinal

study

is

designed

to

answer

five

research

questions
:


Content

learning
.

Do

students

who

are

exposed

to

greater

numbers

of

activities

in

the

three

areas

using

our

modeling

tools

achieve

a

deeper

understanding

of

content?


Epistemological

understanding
.

Do

students

who

are

exposed

to

greater

numbers

of

activities

in

the

achieve

a

deeper

understanding

of

models?



Modeling

skills

and

transfer
.

Do

students

who

are

exposed

to

a

greater

number

of

activities

able

to

use

their

understanding

of

models

to

provide

reasoned

explanations

of

new

phenomena?



Attitudes
.

Do

students

who

are

exposed

to

a

greater

number

of

activities

have

increased

motivation

for

science?


Learning

sequence

effects
.

Are

there

differences

in

any

of

the

measures

depending

on

the

sequence

of

courses?



School

effects
.

How

do

the

varying

levels

of

assistance

to

the

schools

influence

learning

outcomes?




Level 3
-

Longitudinal (cont’d)

Research with log files.



Pedagogica generates logs for every student interaction, including all assessments for all
students over three years.


Data be used as input include:
which activities were used, for what length of time, the
pattern of use (consecutive or intermittent days), and pre and post
-
test dates.



We can also generate a profile for the class in terms of their understanding at pivotal
points in the curriculum. These data will be used to derive teacher reports, important for
formative and summative evaluations.






Research: Level 4
-

Scalability


What kinds of technology infrastructure and data logging capacities are
necessary to provide high level, conceptually
-
based feedback to
teachers about their students?



What kinds of additional support (professional development, on
-
line
support, etc) is necessary for teachers to succeed?



How can we scale up from 3 partner schools to many schools across
the U.S. where we deliver software and collect data from schools with
modest support?


Sample: Levels of Partnerships

Fine
-
Grained Data


Time
-
series data collected online as students work
through activities


Log files stored locally (thin client on schools’ servers
send data to Concord


Data includes pre
-

and post
-
tests


Capacity to treatments (levels of scaffolding) fully
randomized within classrooms~part of original
design.

School Details


First 2 years of project (3 partner schools)
~1000 students, 22 teachers


large urban, suburban, small urban, very mixed SES


Growth curve:


Adding 10 schools this year


Any number of schools can join in subsequent
years; is this viable?


Technology Enhanced Formative Assessment for
Teachers’ and Students’ Use

Observations (level 1 only)

Explicit assessment items
-
post

Implicit assessment items
-
logs


Manipulable models


Time & tries to success


What steps they take


What info or help they seek


Graphic Adapted from
Knowing What Students Know
(2001)

Cognition

Interpretation

Effective

Assessment

Other Research Issues
-

Hypermodels

In MAC we are also leveraging the affordances of technology to:




further develop our hypermodel technology.



characterize model
-
based learning with technology.



provide individualized, technology
-
scaffolded learning that can be faded as
student needs less scaffolding.


assess conceptual understanding with technology and provide formative
feedback to teachers.



In summary, how are we leveraging
technology…

Technology allows for dynamic simulations ~ not


possible with static representations

Broadly accessible using our client and Pedagogica

Pedagogica~ can make changes to all activities in all

schools, etc when it is initiated,

~ creates teacher reports for formative assessment

~ sends back all data for researchers.

With the scaffolding and the on
-
line support, should

be “infinitely” scalable