IMPROVING ACHIEVEMENT AND CLOSING GAPS BETWEEN GROUPS:

1

• IMPROVING ACHIEVEMENT AND CLOSING GAPS BETWEEN
  GROUPS
• How Can Researchers Help?

IES June, 2006

2
As a group, educators are a pretty conservative
lot. Its hard to get them to change what they
do.
3
Not clear that more and better research will
cause many of them to do things differently, but
it certainly can help.
4
Would help, however, if it focused on the
questions most on the minds of educators today.

• Here are the ones I get most often.

5
1. What differentiates the teachers who are most
effective in growing student learning from those
who are less effective?
6
Students in Dallas Gain More in Math with
Effective Teachers One Year Growth From 3rd-4th
Grade
Source Heather Jordan, Robert Mendro, and Dash
Weerasinghe, The Effects of Teachers on
Longitudinal Student Achievement, 1997.
7
Cumulative Teacher Effects On Students Math
Scores in Dallas (Grades 3-5)
Beginning Grade 3 Percentile Rank 57
Beginning Grade 3 Percentile Rank 55
Source Heather Jordan, Robert Mendro, and Dash
Weerasinghe, The Effects of Teachers on
Longitudinal Student Achievement, 1997.
8
Some of the things we most need to knowin order
to act

• Are teachers who are high value-added on state
  assessment also same when second measure is
  introduced?
• How stable are value-added measures?
• What are the characteristics and practices of
  teachers with high value added?
• What can we do to produce/attract/select more of
  them?

9
2. What is it about unusually effective
schoolsespecially those serving low-income
students and students of color-- that makes them
so effective?
10
Source Education Trust analysis of data from
National School-Level State Assessment Score
Database (www.schooldata.org).
11
Source Education Trust analysis of data from
National School-Level State Assessment Score
Database (www.schooldata.org).
12
Source Education Trust analysis of data from
National School-Level State Assessment Score
Database (www.schooldata.org). Data are from 2002.
13
Not all these schools are consistently high
performers in multiple years and subjects. But
some certainly are.
14
Centennial Place Elementary SchoolAtlanta,
Georgia

• 528 students in grades K-5
• 92 African American
• 62 Low-Income

Source School Information Partnership,
http//www.schoolmatters.com
15
Centennial PlaceHigh Achievement for All
StudentsGrade 5 Math, 2005
Source School Information Partnership,
http//www.schoolmatters.com
16
Centennial PlaceHigh Achievement for All
StudentsGrade 5 Reading, 2005
Source School Information Partnership,
http//www.schoolmatters.com
17
Frankford ElementaryFrankford, Delaware

• 29 African-American
• 34 Latino
• 34 White
• 76 Low-Income

Source Delaware Department of Education Online
School Profiles, http//issm.doe.state.de.us/prof
iles/EntitySearch.ASPX
18
Frankford ElementaryClosing Gaps, Grade 5 Math
Source Delaware Department of Education, DSTP
Online Reports, http//dstp.doe.k12.de.us/DSTPmar
t/default.asp
19
University Park High SchoolWorcester, MA

• Grades 7-12
• 70 poverty
• 50 ELL
• Most students enter at least two grade levels
  behind.

20
University Park Results 2004

• 100 of 10th graders passed MA high school exit
  exam on first attempt.
• 87 passed at advanced or proficient level.
• Fifth most successful school in the state,
  surpassing many schools serving wealthy students.
  

21
3. What explains the big differences in how the
same groups of children perform across
different districts and states?
22
Low-Income African American Students do Better
in Some Districts (Urban NAEP Grade 4 Reading,
2005)
Proficient Scale Score 238
Houston and Austin are not included due to high
exclusion rates for students with disabilities
and English Language Learners
Source National Center for Education
Statistics, NAEP Data Explorer,
http//nces.ed.gov/nationsreportcard/nde
23
Low-Income African American Students do Better
in Some Districts (Urban NAEP Grade 8 Math, 2005)
Proficient Scale Score 281
Source National Center for Education
Statistics, NAEP Data Explorer,
http//nces.ed.gov/nationsreportcard/nde
24
Latino Students Do Better In Some Districts
(Urban NAEP Grade 4 Reading, 2005)
Proficient Scale Score 238
Houston and Austin are not included due to high
exclusion rates for students with disabilities
and English Language Learners. Latino scores
are not available for Atlanta
Source National Center for Education
Statistics, NAEP Data Explorer,
http//nces.ed.gov/nationsreportcard/nde
25
NAEP 2005 Grade 8 Math, African American Scale
Scores
Proficient Scale Score 299
MA
RI
Source National Center for Education
Statistics, NAEP Data Explorer,
http//nces.ed.gov/nationsreportcard/nde
26
NAEP 2005 Grade 8 Math, Latino Scale Scores
Proficient Scale Score 299
TX
CA
Source National Center for Education
Statistics, NAEP Data Explorer,
http//nces.ed.gov/nationsreportcard/nde
27
4. None of these educators, schools, districts
or states are yet achieving what we need them to
achieve, especially for the kids who enter
behind. What do we knowfrom learning science,
psychology, brain research and the likethat
could help them go further, faster?
28
Average High School Percent of Instructional
Time in Reading Intensive Courses
Below Grade Level Students On Grade Level Students Advanced Students
24 29 35
29
Surprise Gaps Grow.
30
We know that this isnt smart. But what,
exactly, is?
31
What many educators think they know from
research

• If theyre not good at reading or math, let them
  express themselves with another intelligence
• If you dont get them by the time they are 3 (or
  3 months) those synapses will never grow or
  connect
• If the students are black or brown, virtually
  every lesson must be culturally relevant (and
  if that means taking the content or rigor out,
  well then so be it).

32
You can do better!

• Really important to both get research out there
  AND to guard against deeply dumbed down
  versions of your research.

33
Getting Results USED What DOESNT Work

• Writing it up in journals
• Expecting Ed School faculty to integrate your
  findings into their instruction.

34
Getting Results Used What CAN work

• First, get Russ Whitehurst to give you lots of
  money to ensure that the findings are solid
• Then, get the federal government (or, if you
  prefer) the Gates Foundation to put lots of money
  behind training people to do it and,
• Make really, really sure that your trainers dont
  dumb it down.

35
The Education Trust

• Download this Presentation
• www.edtrust.org
• Washington, DC 202-293-1217
• Oakland, CA 510-465-6444