Escolar Documentos
Profissional Documentos
Cultura Documentos
A
'
e
r
a
g
e
S
c
o
r
e
Exhibit reads: The total average score for schools in District 1 was 79 percent
correct.
,emaining Anal$ses
)e analyses presented in tis report provide initial insigt into variation in data7
informed decision7ma"ing s"ills for scools and districts. )e final report +ill contain
findings from te 9ualitative coding of teacer responses to open7ended data scenario
9uestions and provide greater detail about te types of "no+ledge and s"ills tat teacers use
as tey develop 9uestions, read data tables and graps, interpret data, and use data to ma"e
AC
instructional decisions. !lso, for te final report, te teacer data set +ill be e3panded +it
additional teacers intervie+ed in %&&?>&E and +it responses of small groups of teacers
responding collectively rater tan as individuals to te same or similar scenarios.
Summar$
)e data in E3ibit A7C suggest tat teacers generally can locate specific information in
a data table or grap but tat tey e3perience increasing difficulty if tey need to manipulate
te data in some +ay or ma"e comparisons bet+een multiple data points. 5ecessary
manipulations re9uired only basic aritmetic 1subtraction, division2, and teacers +ere
provided +it paper and pencil and +it calculators tey could use if tey +ised. Even +it
tese tools, cognitive researc +ould suggest tat te cognitive load re9uired for "eeping in
mind te issue at and, multiple data points, and intermediate products of data manipulations
ave a role in te observed deterioration of teacers/ accuracy 1!yres %&&82. Upcoming
administrations of data scenarios to small groups of teacers +ill test +eter greater
accuracy in data compreension is acieved by spreading te compreension burden across
multiple individuals.
$n terms of te ability to interpret data, teacers demonstrated limited understanding of
basic statistical concepts suc as variability and reliability. )eacers appeared to ave
particular difficulty 1a2 using individual students/ scores, rater tan teir subgroup mean, as
te basis for predictions and instructional prescriptions and 1b2 "eeping in mind te fact tat
student groups differ from year to year +en loo"ing at multiyear data.
AA
<- Summar$ of the Earl$ %indings
)e concept of using data as a basis for teir decisions is novel to many scool staff
members, and tere are multiple obstacles to institutionali,ation of tis practice. By collecting
data at te scool and district levels, te researc team sougt a Gbottom7upH perspective on
te 9uality of te supports for data7informed decision ma"ing coming from te state and
national levels. Policy7ma"ing in tis area began +it te assumption tat federal
re9uirements and national campaigns 1e.g., te Data Fuality Campaign, grants to states for
enancement of teir data systems2 +ould lead to iger79uality state data systems, +ic in
turn +ould ave a positive influence on te use of data in districts, scools and classrooms.
Although federal and state policies have spurred districts and schools to initiate more
data-informed activity, the study found no evidence of a tightly coupled system. )e vast
maDority of te activity around data systems and data use in districts and scools involves
district, not state data systems, according to te 5E))S survey data. )+o of te case study
districts +ere in states +it systems tat use a uni9ue student identifier tat te district system
also used, facilitating transfer of data across systems. But for te most part, te oped7for
efficiencies to be gained from integrating data systems at te state, district and scool levels
are not apparent from te vantage point of scools and districts. More progress in tis area can
be e3pected as ne+ state systems come on line.
Direct district use of a state7provided soft+are system appears to be relatively
uncommon. )is practice +as found in only one of te nine case study districts. $n tis case,
te state made available a performance improvement mapping system tat te district could
use to obtain test items lin"ed to areas +ere te scool needed to improve. $n anoter case,
te district +as using a state soft+are system for its data +areouse, but te state ad since
moved to a ne+ soft+are system +it additional capabilities, and te district +as left +it an
GorpanH system.
!mong te nine case study districts 1dra+n from nine different states2, only one cited
training on data7informed decision ma"ing made available troug its state education agency.
$n tis case, principal training and some funds for laptop purcases +ere made available by
te state.
$t sould be noted tat a tigtly coupled system is not necessarily te ideal solution in all
cases. )e fact tat no suc system +as found in any of te case study districts, selected on
te basis of perceived e3cellence in data use, suggests tat implementation of suc systems is
difficult. Muc of te value provided by a tigtly integrated system can arguably be provided
also by less tigtly coupled systems.
Case study informants were more likely to describe federal funds supporting data use
activities than state funds doing so. ! number of federal programs ave provided funds tat
ave paid for e3tra staff positions or oter activities to promote a data7using culture. $n one
district, for e3ample, )itle $ funds paid for full7time scool7based instructional advisors +o
elped teacers ac9uire and interpret data: te funds also supported te ig7speed $nternet
access needed to use te data system. !noter district used some of its )itle $ funds for
AB
professional development on te use of D$BE6S
C&
data for instructional decision ma"ing.
Scools in tree districts used #eading @irst grant funding for activities suc as collecting and
trac"ing longitudinal data on primary students/ reading s"ills and for funding reading coaces
+o elped teacers incorporate assessment data into teir practice. !n EE)) grant to a state
consortium for scool use of data +as led by one of te case study districts: te consortium
as provided professional development on data7informed decision ma"ing at no cost to te
district.
University partners are an additional class of facilitators. One district is +or"ing +it a
literacy proDect run by a teacers/ college to learn more about "inds of student data tat can be
collected and o+ tey can be used to guide subse9uent instruction. !noter district ad a
university come in to +or" +it its teacers on o+ to disaggregate and interpret data from a
ponological a+areness assessment.
(rogress in the Use of Data S$stems
)is study encompassed bot district and scool use of student data systems and teir
data use practices to improve instruction. !cross districts and scools in te case study
sample, use of locally generated data to inform instruction and use of electronic data systems
containing student scores on standardi,ed tests appeared to be t+o parallel initiatives rater
tan a single, integrated effort. District and scool use of standardi,ed test scores obtained
from data systems typically focused on accountability concerns and ma"ing sure tat local
curriculum and instruction +ere +ell aligned +it te state/s assessments. 6ocally generated
data often came from instruments connected +it an early literacy program or bencmar"
assessments. )e t+o data activities +ere not necessarily integrated, and scools ma"ing
progress in one of tese areas +ere not necessarily doing muc +it te oter.
Staff members at four of te nine case study districts reported aving identified areas in
need of improvement, corrections needed in curriculum alignment, or areas ripe for ne+
programs on te basis of analysis of data from teir data systems. )+o of te districts
described more mature data use, including using data on an ongoing basis to evaluate +ic
of teir initiatives +ere +or"ing and +ic +ere not. One of tese districts +on a state a+ard
as one of te most improved districts in te state.
)e 5E))S teacer survey and first round of case study data collection ma"e it clear
tat te implementation of data systems and practices for teir use in decision ma"ing at te
scool level is still in its infancy. Bet+een te t+o 5E))S teacer surveys, te proportion of
teacers reporting access to a student data system rose by nearly B& percent. )is large a
cange over a t+o7year period is 9uite remar"able, but te optimism it engenders must be
tempered by concerns over te perceived value of te data to +ic more teacers ave
access. On te latest 5E))S survey, te rougly t+o7tirds of teacers +it access to a
student data system +ere divided in teir perspective on +eter or not scool use of data ad
paid off. A% percent agreed tat teir scool ad been improved troug te use of data, +ile
AB percent neiter agreed nor disagreed +it te statement. Similarly, te case study
C&
)e Dynamic $ndicators of Basic Early 6iteracy 1D$BE6S2 are a set of standardi,ed, individually
administered measures of early literacy development. )ey are designed to be sort 187minute2 fluency
measures used to regularly monitor te development of pre7reading and early reading s"ills.
A=
intervie+s suggest tat e#en in districts )ith a reputation for leadership in using data,
electronic data systems are *arely influencing classroom-le#el decision making. <o+ever,
more progress can be seen in te use of data, coming from oter sources suc as diagnostic
tests included in early reading programs, to sape instruction.
Staff members at all %? case study scools described te use of student data tat +as not
contained on teir data system to guide instruction. @re9uently mentioned +ere data from
early reading assessments given in primary grades 1for +ic state assessment data are
usually unavailable2, performance reports from computer7based instruction, running records,
and e3aminations constructed by scool departments.
&essons &earned for Implementation
)e reflections of te case study informants and te pattern of activity across teir
districts and scools +it different practices suggest a number of guidelines for scools and
districts embar"ing on te implementation of data7informed decision ma"ing.
8. Districts are finding that providing school-level data coaches is an important support for
school-level use of data to inform instruction. $n te maDority of case study districts, some
or all scools ad on7site staff members designated to elp teacers retrieve data from te
data system, interpret data and ma"e instructional decisions based on data. 5ot all districts
started out +it tis intention, but it appears to be an emerging best practice tat bot
encourages more use of data and lessens te li"eliood of misinterpretation of data.
%. A common curriculum and curriculum-aligned benchmark assessments increase the
likelihood that school staff members will make extensive use of a districts data system.
State assessments are typically administered once a year, and scool revie+ and reflection
on te resulting data is typically also a once7a7year event. )eacers in all nine case study
districts loo"ed more fre9uently at interim, bencmar" or end7of7course assessments. $n
some cases, tese data +ere available on a district7provided system, but in many cases tey
+ere accessed on a test developer -eb site, eiter as reports generated by te assessment
soft+are itself or troug a scool7developed database. )o te e3tent tat te district data
system contains more current data tat teacers find relevant to teir instructional
decisions, teacers +ill ave a greater motivation to use te system.
+. !f teacher use of data is the goal, then it is desirable to have curriculum and instruction
staff members involved in the initiative. )e use of student data to inform instruction is
necessarily a systemic effort, lin"ing assessment and accountability, professional
development, information tecnology, and curriculum and instruction functions tat are
typically separate offices +itin medium7si,e and large districts. !ll of tese offices need
to support te initiative. $n t+o of te case study districts, te data use initiative +as led by
te assessment and accountability office, and oter district offices +ere not deeply
involved. $n tese cases, te data system +as not being used to sape instruction.
A. "eacher buy-in for the data system and its use should be sought early and maintained
continuously. !ttention to teacer buy7in +as not uniform across te case study districts.
On one end of te continuum +ere districts tat solicited input from a fe+ principals or
A?
teacers regarding +at tey +anted in te system and ten proceeded to procure and
adopt one. !t te oter end of te continuum +as a district tat instituted a nine7mont
process of convening sta"eolder groups, including teacers, to elp select te ne+ system
and involved te union in te media campaign for its launc. Several districts tat ad not
involved teacers early on +ere dealing +it teacer suspicions tat te data system +ould
be used to identify +ea" teacers and tat data +ould be used against tem.
B. #rofessional development should include training on how to interpret data and how to
translate data into changes in instructional practice. )eacers in seven of te nine case
study districts cited insufficient training on o+ to engage in data7informed decision7
ma"ing practices, o+ to read and interpret te data tat are given to tem, and o+ to
translate data into actionable plans for instruction. )eacer responses in te data scenario
intervie+s suggest tat tese concerns are +ell7founded. )eacers could read data tables
and graps but ad difficulty framing a 9uestion tat could be ans+ered troug a data
system 9uery, and +en tey did ave data in front of tem, tey ad a ard time dra+ing
defensible interpretations or inferences based on te data. On te 5E))S teacer survey, a
large proportion of respondents felt tey could benefit from additional professional
development around te use of data. BE percent tougt tey could benefit from
professional development on o+ to develop diagnostic assessments for teir students,
C8
BB
percent on o+ to adDust te content and approac used in teir class on te basis of data,
B& percent on +at data to collect to monitor scool progress against improvement goals,
and AE percent on o+ to interpret data.
=. District policies should be examined to identify and remove policies and procedures that
undermine teachers use of data to inform instruction. Despite a commitment to te
promotion of teacers/ use of data in ma"ing instructional decisions and individuali,ing
and optimi,ing students/ educational e3periences, districts often ave policies in place tat
run counter to suc efforts. !n e3ample is te imposition of district pacing plans tat do
not provide time for individuali,ation or selective reteacing based on bencmar" or
formative assessment findings. On te 5E))S )eacer Survey, of tose teacers +it
access to a student data system, =& percent said tat no matter +at te system tells tem
about teir students/ learning, tey must "eep up +it state or district pacing plans.
?. $chool leaders need to build teachers mutual trust to a point where teachers are
comfortable working with colleagues to examine data that reflect on their teaching
performance. Muc of te e3amination of data in case study scools occurred in small
groups4during common planning time or grade7level or department meetings. )is
practice +as particularly +ell ingrained in te scools identified as Gig useH by teir
districts. Staff members in seven districts said tat data7informed decision7ma"ing
practices ad moved scool staff members to+ard more open conversations about
instructional practice and increased opportunities to learn from one anoter. One principal
noted, GUsing data eliminates distracters and "eeps everyone focused on +ere tey need
to go. Data also elps to eliminate some of te interpersonal issues tat can arise: it/s not
about +o you are, it/s about +at te data so+s.H Some scools are actually
C8
)eacers are being encouraged to use diagnostic assessments but often e3press frustration +it te fact tat
tey do not ave diagnostic instruments available to tem tat matc teir instructional program.
AE
disaggregating assessment data by teacer to find areas +ere tere is a mar"ed difference
across classrooms so teacers can learn from peers +it te best outcomes in tat area.
Clearly, suc activity re9uires a ig degree of mutual trust, a caracteristic tat scool
reform researc as found to be present in urban scools tat ma"e significant
improvements 1Bry" and Scneider %&&%2. )e principals at t+o of te case study scools
noted tat tey needed to spend several years building up an atmospere of trust and a
Gblame7freeH culture before teir teacers could loo" at data togeter onestly.
)e 5E))S teacer survey data also underscore te importance of peer support in
creating a culture of data use. )eacers +o reported feeling +ell7supported by teir
colleagues in teir efforts to use data +ere more li"ely tan oter teacers to use data in +ays
related to instruction, for e3ample, to identify student s"ill gaps.
Data7informed decision ma"ing is not a simple intervention not only because it involves
so many aspects of education 1e.g., assessment, curriculum, accountability, information
tecnology2 but also because it re9uires fundamental improvements in te degree of mutual
trust and canges in te +ay teacer time is used. )e survey and case study data presented in
tis interim report suggest tat te movement to incorporate student data in local education
decision ma"ing is real 1many districts, scool leaders and teacers are ma"ing good7fait
efforts2, but tere is a significant distance to go before it becomes +ell e3ecuted in practice.
)e picture in %&&=>&? +as one tat so+ed various pieces of te comple3 re9uirements
getting put in place in different locations but no single scool or district demonstrating a
totally integrated, consistent and pervasive continuous7improvement process. Mutual trust
may prove to be te glue needed to old togeter te district and scool practices tat involve
using data to improve instruction and acievement.
A'
,eferences
!yres, P. 6. %&&8. Systematic matematical errors and cognitive load. ontemporary
,ducational -sychology %.1%2. %%?>AE.
Boudett, K. P., and L. Moody. 2005. Organizing for collaborative work. In Data wise, ed. K. P.
Boudett, E. A. City, and R. J. Murnane, 1128. Cambridge, Mass.: Harvard Education Press.
Brunner, C., C. @asca, *. <ein,e, M. <oney, D. 6igt, E. Mandinac, and D. -e3ler. %&&B.
6in"ing data and learning4)e (ro+ 5et+or" study. /ournal of ,ducation for Students
-laced at 0isk, 1&1C2. %A87%=?.
Bry", !. S., and B. Scneider. %&&%. Trust in schools: A core resource for impro#ement2 5e+
Kor". #ussell Sage @oundation.
Camburn, E., B. #o+an, and *. E. )aylor. %&&C. Distributed leadersip in scools. )e case of
elementary scools adopting compreensive scool reform models. ,ducational ,#aluation
and -olicy Analysis %'1A2. CA?>?C.
Coppin, *. %&&%. Data use in practice: ,3amples from the school le#el2 Paper presented at te
annual meeting of te !merican Educational #esearc !ssociation, 5e+ Orleans.
Confrey, *., and I. M. Ma"ar. %&&B. Criti9uing and improving te use of data from ig7sta"es
tests +it te aid of dynamic statistics soft+are. $n Scaling up success: !essons learned from
technology-*ased educational impro#ement, ed. C. Dede, *. P. <onan, and 6. C. Peters, 8'E>
%%=. San @rancisco, Calif.. *ossey7Bass.
Cromey, !. %&&&, 5ovember. Using student assessment data. -at can +e learn from scools;
Policy $ssues 5o. =. Oa" Broo", $ll.. 5ort Central #egional Educational 6aboratory.
Data Fuality Campaign. %&&?. 0esults of %&&( N,A sur#ey of state --1% data collection issues
related to longitudinal analysis. ttp.00+++.data9ualitycampaign.org0surveyJresults0 1clic"
on individual states to retrieve %&&? information: accessed *anuary %, %&&E2.
Datno+, !., L. Par", and P. -olstetter. %&&?. Achie#ing )ith data: 4o) high-performing
school systems use data to impro#e instruction for elementary students2 6os !ngeles, Calif..
University of Soutern California, Center on Educational (overnance.
@eldman, *., and #. )ung. %&&8. Using data based in9uiry and decision7ma"ing to improve
instruction. ,0S Spectrum 151C2. 8&>8'.
<alverson, #., *. (rigg, #. Pricett, and C. )omas. %&&B. The ne) instructional leadership:
reating data-dri#en instructional systems in schools2 Madison, -isc.. -isconsin Center for
Education #esearc, University of -isconsin.
<erman, *., and B. (ribbons. %&&8. !essons learned in using data to support school in6uiry and
continuous impro#ement: 7inal report to the Stuart 7oundation. CSE #eport BCB. 6os
!ngeles, Calif.. UC6! Center for te Study of Evaluation. !vailable at
ttp.00+++.cse.ucla.edu0products0#eports0#EBCB.pdf 1accessed December 8=, %&&E2.
B8
<offman, *. %&&?. Num*ers and types of pu*lic elementary and secondary education agencies
from the ommon ore of Data: School year %&&'8&. 9N,S %&&(-+'+:. U.S. Department
of Education. -asington, D.C.. 5ational Center for Education Statistics.
ttp.00nces.ed.gov0pubsearc0pubsinfo.asp;pubidM%&&?CBC 1accessed December 8=, %&&E2.
Mandinac, E. B., M. <oney, D. 6igt, *. <ein,e, and 6. #ivas, 6. %&&B. )ecnology7based tools
tat facilitate data7driven decision ma"ing. $n To)ard sustaina*le and scala*le educational
inno#ations informed *y the learning sciences, ed. C. I. 6ooi, D. *onassen, and M. $deda,
%=?>?A2. !msterdam. $OS Press.
Mars, *. !., *. @. Pane, and 6. S. <amilton. %&&=. Making sense of data-dri#en decision making
in education2 Santa Monica, Calif.. #!5D.
Palaic, #. M., D. (. (ood, and !. van der Ploeg. %&&A. State education data systems tat
increase learning and improve accountability. Policy $ssues 5o. 8=. 5aperville, $ll.. 6earning
Point !ssociates.
Scmo"er, M. *. 8''=. 0esults: The key to continuous school impro#ement2 !le3andria, La..
!ssociation for Supervision and Curriculum Development.
)orn, C. !. %&&%. Data use in the classroom: The challenges of implementing data-*ased
decision-making at the school le#el2 Paper presented at te annual meeting of te !merican
Educational #esearc !ssociation, 5e+ Orleans.
U.S. Department of Education, Office of Educational )ecnology. %&&A. To)ard a ne) golden
age in American education: 4o) the Internet, the la) and today$s students are
re#olutioni;ing e3pectations2 -asington, D.C.. !utor.
U.S. Department of Education, Office of Planning, Evaluation and Policy Development. %&&Ea.
National educational technology trends study: !ocal-le#el data summary. -asington, D.C..
!utor.
444. %&&Eb. Teachers$ use of student data systems to impro#e instruction %&&' to %&&(.
-asington, D.C.. !utor.
-ayman, *. C. %&&B. $nvolving teacers in data7driven decision ma"ing. Using computer data
systems to support teacer in9uiry and reflection. /ournal of ,ducation for Students -laced
At 0isk 1&1C2. %'B>C&E.
-ellman, B., and 6. 6ipton. %&&A. Data-dri#en dialogue: A facilitators guide to colla*orati#e
in6uiry. Serman, Conn.. Mira Lia 66C.
B%
Appendi A
Case Stud$ and Sur#e$ Data Sources
BC
Ehi!it A-'
Definitions of Terms
Term Definition
.ata systems >ectronic information systems to assist in the organization and
management of data+ 3hey consist of hard1are and soft1are that
pro'ide many different functions to users such as storing current and
historica data, rapidy organizing and anayzing data !e+g+, e6amining
reationships 1ithin data, specifying data subgroups#, and de'eoping
presentation formats or interfaces+
7nteroperabiity 3he abiity of different data systems or soft1are packages to
communicate 1ith one another+
Soft1are appications Soft1are appications consist of a 1ide range of speciaized products
to faciitate access to data, data anaysis and interpretation, and
presentations of data !e+g+, formatted reports, graphing functions#+
Schoos 7nteroperabiity
%rame1ork !S7%#
.ata e6change standards for N"12 education soft1are de'eoped to
promote data sharing bet1een different appications 1ithout further
soft1are inter'ention+ S7% 1as aunched in 1((- as a nonprofit
initiati'e of the soft1are and information technoogy industry to ensure
that N"12 instructiona and administrati'e soft1are appications 1ork
together more effecti'ey+
.ataAinformed decision
making
A process that integrates the anaysis of educationa data, typicay
stored in educationa data systems, to support decisions intended to
impro'e teaching and earning at the schoo and cassroom e'es+
3he practice entais reguar data coection and ongoing
impementation of impro'ements+
/ongitudina student
data
.ata on indi'idua students coected o'er time that ao1s users to
compie an academic history for each student+ 3his type of data
enabes more robust anayses of student performance to hep
differentiate instruction and impro'e student achie'ement+
Kuery too Soft1are that ao1s for customized and adAhoc data re?uests such
as Bdri do1nC capabiity to efficienty e6amine a subset of data at a
grade, cassroom or student e'e+
.ata ?uery A re?uest for specific records from a data system+
3ransaction capture <eaAtime accounting of daiy schoo functions such as attendance
and schoo unch counts+
>ectronic grade books 2nine toos to hep teachers manage cassroom acti'ities !e+g+,
generating seating charts, recording grades and test scores,
managing attendance, tracking skis and standards, panning essons,
and generating report cards#+
>ectronic portfoios 2nine storage of student 1ork sampes+
>ectronic
communication toos
3oos to faciitate communication of information such as use of eAmai,
de'eopment of 0eb sites, and eectronic discussion groups or
Bmessage boardsC 1here users from mutipe ocations can discuss
issues+
%ormati'e assessment Assessment conducted during instruction to pro'ide feedback that
can be used to adIust ongoing teaching and earning to impro'e
student outcomes+ %ormati'e assessment can be contrasted 1ith
summati'e assessment, 1hich takes pace after a period of
instruction, to Iudge ho1 much earning has occurred+
BB
Ehi!it A-+
Data A#aila!le %rom the State Data S$stem for Each Case Stud$ District in +//2
District ' + 7 ; < 0 2 > ?
Data Elements in State
Information S$stem
CA @A MD TB ,I MA =C C9 M9
Lni?ue state1ide student identifier
that connects student data across key
databases across yearsO
Pes Pes ;o Pes Pes Pes ;o Pes ;o
Abiity to match indi'idua students*
test records from year to year to
measure academic gro1thO
Pes Pes ;o Pes Pes Pes Pes Pes Pes
State can measure yearAtoAyear
gro1th for indi'idua students in any
subIect !state tests in the same
subIect in consecuti'e years and is
abe to connect data across years#
;o Pes ;o Pes Pes Pes Pes Pes Pes
StudentAe'e enroment,
demographic, and program
participation informationO
Pes Pes Pes Pes Pes Pes Pes Pes ;o
StudentAe'e transcript information,
incuding information on courses
competed and grades earnedO
;o ;o ;o Pes ;o ;o Pes ;o ;o
A teacher identifier system 1ith the
abiity to match teachers to studentsO
;o ;o ;o ;o Pes ;o ;o ;o ;o
7dentifies 1hich schoos produce the
strongest academic gro1th for their
students
Pes Pes ;o Pes Pes Pes ;o Pes ;o
Nno1s 1hat achie'ement e'es in
midde schoo indicate that a student is
on track to succeed in rigorous
courses in high schoo
;o ;o ;o Pes ;o ;o ;o ;o ;o
A state data audit system assessing
data ?uaity, 'aidity and reiabiityO
Pes Pes Pes Pes ;o Pes Pes Pes Pes
;umber of the 10 .K) essentia
eements presentO
7 7 $ ( 6 - 7 6 4
Exhibit reads: In District 1, in 2007, the state data system has a unique statewide student identifier that connects student data across key databases across years.
Note: An asterisk identifies six of the 10 essential elements most relevant to the data issues discussed in the case studies. There are four additional elements not
listed.
Source: Data Quality Campaign (2007).
B=
A
-
3
Ehi!it A-7
Case Stud$ District Data S$stems
3urn/eaf Achie'ement &anagement System ser'es as the data
1arehouse and assessment systemH eS7S is the student information
systemH ;0>A onine database for anaysis and reporting toos and
instructiona materias+
eSchoar is the data 1arehouseH S7&S is student information systemH
StarQDase for student recordsH 3est0iz managed by data&etrics
Soft1areH 3P. eectronic registrarH P7&S performance impro'ement
mapping system from the S>AH >dine for reporting teacher data to
students and parents+
Schoo;et Account !report generation#H Assess !assessment bank#
and Aign !data 1arehouse# moduesH Pentamation is the student
information systemH Specia >ducation &anagerH discipine and
attendance systems+
>ducation .ecision Support /ibrary is the data 1arehouse !>.S/ is
/>A de'eoped#H Denchmark Assessment and <eporting 3oo !DA<3
1as Iointy de'eoped by />A and Princeton <e'ie1#H Schoos
Administrati'e Student 7nformation !SAS7# from Pearson, >ducationa
)urricuum Assessment <esource 3oo !>)A<3 is being de'eoped by
/>A to repace DA<3#+
S7&S is the state data system and ser'es as the />A data 1arehouse
!1i be repaced by ;) 07S>#H )A)ASS for specia education dataH
.7D>/S onine database for anaysis and reporting toos !data
upoaded from handheds#+
.ata 0arehouse 1as de'eoped by the />AH 7nstructiona
&anagement System !7&S designed by A.&7;+)2&#H Professiona
.e'eopment 2nine !P.2 1as buit by />A and a pri'ate company#+
<egistration !Reg 1as de'eoped by the />A# is a student information
system and data 1arehouseH Socrates )<& !the ibrary data system#H
Phonoogica A1areness /iteracy Screening !PA/S# 0eb site for
anaysis and reporting toos+
7nfinite )ampus houses student data and has a suite of reporting toos
and parent portaH >d5ate assessment system is being pioted+
)ognos !ac?uired by 7&D# is the />A*s data 1arehouseH Range is a
0ebAbased transactiona system to enter data !attendance, unch,
district test scores#H Student Assessment .ata System !SA.S#
pro'ides historica test data and standard reportsH >ncore is the
specia education system !/>AAde'eoped systems#+
B?
=ETTS Sur#e$ Data
)e primary survey data included in tis report come from te U.S. Department of
Education/s 5ational Educational )ecnology )rends Study 15E))S2. 8,&C' district
tecnology directors surveyed during spring %&&? and %,B&' teacers surveyed in spring %&&?.
)e teacers +ere clustered in scools sampled from districts participating in te 5E))S
district survey.
C%
)e 5E))S district sample of 8,&C' districts +as nationally representative
+it respect to poverty status, student enrollments, and location 1urban or rural status2. )e =&
largest urban scool districts across te country +ere selected +it certainty 1i.e., included in
te sample from te outset2. Districts composed entirely of special education scools and
vocational7tecnical scools, as +ell as independent carter scools tat are teir o+n
districts, +ere e3cluded from te district sampling frame because of teir dissimilarity to
GtypicalH districts. )e teacer sample +as created by dra+ing a probability sample of '?B
scools from respondents to te district survey, stratified by scool type 1elementary or
secondary2, and poverty level 1ig or lo+2. Scools +ere randomly sampled in proportion to
te number of teacers and in inverse proportion to district si,e to produce a sample of
scools +ose selection probabilities +ere rougly independent of te si,e of teir district/s
enrollment.
<iger7poverty scools +ere oversampled 1%CC of te '?B scools2 to obtain more
precise data about teir tecnology use. @or scools, Giger povertyH +as defined as above a
specified cutoff in terms of te percentage of students +o +ere eligible for te free or
reduced7price lunc program. )e dividing line bet+een iger7poverty and lo+er7poverty
scools +as selected to ensure tat for eac scool type 1elementary, middle or ig scool2,
tere +ould be te same number of teacers in te iger7poverty and te lo+er7poverty
groups, as reported in te 5ational Center for Education Statistics Common Core of Data
1CCD2. Elementary scools +it %'.? percent of teir students eligible for free or reduced7
price lunces +ere classified as iger poverty. @or middle and ig scools, te poverty
tresolds +ere %A.C percent and 8B.' percent, respectively.
5E))S researcers obtained teacer rosters for te '?B scools +itin te districts
selected for te 5E))S district survey. )o be eligible for te teacer sample, a teacer ad to
be teacing at te same scool in te scool year before survey administration 1i.e., teacers
ne+ to te scool +ere e3cluded2. )eacers +o did not teac core academic subDects also
+ere omitted from te sample. )argets of four teacers from eac of te scools +ere
randomly selected for te teacer sample. )e final teacer sample in %&&? consisted of 8,??'
teacers from E=B scools.
Response rates were 94 percent for the district survey in 2007 and 85 percent for the
teacher survey. Sampling weights were applied to the teacher data to obtain nationally
representative estimates. The district and teacher surveys were initially administered during
the 200405 school year to the same sample of districts and schools that provide the basis of
comparison between 2007 and 2005 for the teacher survey data. In fall 2005, 6,017 teachers
were surveyed with a response rate of 82 percent; the larger sample was designed to provide
C%
Districts +ere sampled from among te 8%,AEC districts tat received federal Enancing Education )roug
)ecnology 1EE))2 funds in %&&C and from an additional %,%C' districts tat ad not received EE)) funds.
BE
robust, school-level estimates of technology use. For additional information on the
comparisons between 2007 and 2005 teacher survey responses, see Teachers Use of Student
Data Systems to Improve Instruction: 2005 to 2007 (U.S. Department of Education 2008b).
B'
Appendi 6
Scenario Ehi!its, Items, ,u!rics, Item Score Distri!ution and Mean Scores
=8
Scenario A
;o1 7*m going to describe a hypothetica situation and a computerAbased student data system to you and
7*d ike to see 1hat kind of information you*d ike to get from the system+
Suppose it*s Manuary 2007 and you*re one of the fourthAgrade teachers in a schoo that 1as surprised by
fourth graders* reati'ey o1 performance on the state reading test ast year !spring 2006#+ Pour principa
has encouraged you to use student data to gain insights into ho1 you can get higher 5rade 4
achie'ement this year+
3he data system a'aiabe to you Sshow Figure A screen mockupT contains data on both current !2006"
07# fourth graders and ast year*s !2005"06# fourth graders !show Student Groups table, pointing to
student groups as you name them# as 1e as other student groups+ %or each student, the data system
has scores on ast spring*s state reading test and scores on a district test gi'en in the fa !point to
relevant cells in Student Variables table#+ 7t aso has other information about students that can be used to
create subgroups 1ithin a grade if you 1ant to see ho1 different subgroups compare+ %or e6ampe,
ethnicity, gender, and 1hether the student is eigibe for free or reduced price unch !%<P/#+
=C
%igure A8
Student 5roups
200506 Third graders 200506 Fourth graders 200506 Fifth graders
200607 Third graders 200607 Fourth graders 200607 Fifth graders
Student Uariabes
Ethnicity 200506 Grade 3 state spring reading achievement score
Gender 200506 Grade 4 state spring reading achievement score
Free or reduced-price lunch applicant 200506 Grade 5 state spring reading achievement score
Year entered district 200607 Grade 3 district fall reading test
Grade 3 teacher 200607 Grade 4 district fall reading test
Grade 4 teacher 200607 Grade 5 district fall reading test
Assessment System 2utput
=A
A-'- So no1 in Manuary 2007, 1hat specific data 1oud you 1ant to get from this system to hep you
decide ho1 to impro'e your fourth graders* performance4 !Follow-up probes+# 0hat data 1oud
you 1ant to see first4 3e me 1hat group of students and 1hat test score or student
characteristic you 1oud ike to see for those students+
,u!ricC
'C 3eacher picks a ogica group !either 2005"06 fourth graders or 2006"07 fourth graders# A=D
seects a ogica measure for that group !either Spring 2006 5rade 4 state test scores or %a
2006 5rade 5 district fa test scores for 2005"06 fourth graders 2< either Spring 2006 5rade $
state scores or %a 2006 5rade 4 district scores for 2006"07 fourth graders+
/-<C 3eacher identifies group and scorer can infer measure based on teacher response+H 9,
3eacher picks a ogica group !either 2005"06 fourth graders or 2006"07 fourth graders#H 9,
3eacher seects a ogica measure for that group !either Spring 2006 5rade 4 state test scores or
%a 2006 5rade 5 district fa test scores for 2005"06 fourth graders 2< either Spring 2006
5rade $ state scores or %a 2006 5rade 4 district scores for 2006"07 fourth graders+
/C 7ncorrectH 3eacher does not kno1+H Scorer cannot concude 1hether ans1er is correct or
incorrect+
n Teachers
Score %re.uenc$ 3(roportion4
M / /-< '
-5 26 !+$1# 2$ !+27# $6 !+42# +56
A-+- 0oud you ike to make any other ?ueries of the data system4 Are there other kinds of data you
1oud ike to ha'e to hep impro'e this year*s fourth graders achie'ement that you don*t see
represented in this system4
=o ,u!ricC Coded onl$, not scored
=B
Scenario 6
3his !show Figure B# is the kind of data tabe that some student data systems produce+ 7*m going to ask
you to find some information from the tabe and then 7*d ike for you to te me on a scae of 1 to 10, 1ith
10 being e6tremey difficut, ho1 hard or easy it 1as to find the information in the tabe+ <eady4
%igure D
2005"06 Score /e'esE>ngish /anguage Arts !>/A#
English Language Arts
Hamilton Elementary
Grade Gender Ethnicity
Number of
Students
Tested
Percent
of Tested
Students
Mean
Scale
Score
Number Students at
Each Proficiency Level
Belo
Basic Basic Proficient Advanced
!
Female
African American 18 26% 439 5 7 5 1
Asian/Pac Islander 1 1% 610 0 0 0 1
Latino 17 24% 428 5 6 5 1
!ite 34 49% 449 4 13 11 6
"otal Female 70 100% 444 14 26 21 9
#ale
African American 18 23% 436 6 6 5 1
Asian/Pac Islander 2 3% 452 0 1 0 1
Latino 31 40% 430 8 7 14 2
!ite 27 35% 448 6 11 7 3
"otal #ale 78 100% 438 20 25 26 7
"
Female
African American 18 24% 441 3 8 5 2
Asian/Pac Islander 2 3% 462 1 0 1 0
Latino 36 47% 436 8 12 12 4
!ite 20 26% 472 2 7 8 3
"otal Female 76 100% 447 14 27 26 9
#ale
African American 16 23% 442 2 8 5 1
Asian/Pac Islander 0 0% $A 0 0 0 0
Latino 29 42% 438 5 12 10 2
!ite 24 35% 456 3 13 5 3
"otal #ale 69 100% 445 10 33 20 6
#
Female
African American 19 26% 463 2 6 8 3
Asian/Pac Islander 1 1% 317 1 0 0 0
Latino 34 47% 452 4 14 10 6
!ite 19 26% 507 1 6 7 5
"otal Female
73 100% 467 8 26 25 14
#ale
African American 17 23% 449 2 6 6 3
Asian/Pac Islander 3 4% 560 0 0 1 2
Latino 34 46% 448 7 13 11 3
!ite 20 27% 468 3 6 8 3
"otal #ale 74 100% 467 12 25 26 11
==
6-'- 0hat 1as the mean !or a'erage# scae score for the /atino fifthAgrade girs 1ho took the test4
Go1 easy 1as it to find this information in the tabe4
,u!ricC
'C 452
/C 7ncorrectH 3eacher does not kno1+H Scorer cannot concude 1hether ans1er is correct or
incorrect+
n Teachers Score %re.uenc$ 3(roportion4 M
/ /-< '
4( 5 !+10# ;@A 44 !+(0# +(0
6-+- 0hich student group had the highest a'erage or mean >/A scae score in 5rade 44 Go1 easy
1as it to find this information in the tabe4
,u!ricC
'C B0hite femaesC or B1hite girsC
/C 7ncorrectH 3eacher does not kno1+H Scorer cannot concude 1hether ans1er is correct or
incorrect+
n Teachers Score %re.uenc$ 3(roportion4 M
/ /-< '
4( 6 !+12# ;@A 4$ !+--# +--
6-7- Go1 many African American thirdAgrade boys took the test4 Go1 easy 1as it to find this
information in the tabe4
,u!ricC
'C 1-
/C 7ncorrectH 3eacher does not kno1+H Scorer cannot concude 1hether ans1er is correct or
incorrect+
n Teachers Score %re.uenc$ 3(roportion4 M
/ /-< '
4- 0 ;@A 4- !1+00# 1+00
6-;- 2'era, based on the 5rade $ data in this tabe, 1oud you say that there 1as a difference
bet1een boys and girs in >/A test performance4
=o ,u!ricC Coded onl$, not scored
=?
Scenario C
3his !show Figure C# is the kind of data chart that some student data systems produce+ 7*m going to ask you to find some
information from the graph and then 7*d ike for you to te me on a scae of 1 to 10, 1ith 10 being e6tremey difficut, ho1
hard or easy it 1as to find the information in the chart+ <eady4
%igure )
>/A Dar 5raph
8,ADE 7 8,ADE ; 8,ADE <
n 9 ;umber of test takers
6
8
S
r
i
C-' 3II4- /ooking at this graph, can you find the a'erage !or mean# >ngish /anguage Arts scae score for
the /atino fifthAgrade girs 1ho took the test4 Go1 easy 1as it to get this information from the
chart4
,u!ricC
'C Any ans1er from 44( to 455
/C 7ncorrectH 3eacher does not kno1H Scorer cannot concude 1hether ans1er is correct or
incorrect
n Teachers Score %re.uenc$ 3(roportion4
/ /-< '
(7 25 !+26# ;@A 72 !+74#
C-+- )an you find the mean score for Asian@Pacific 7sander fourthAgrade boys 1ho took the test4
Go1 easy 1as it to get this information from the chart4
,u!ricC
'C B;oC or Bnone took the testCH B3here*s no data+C
/C 7ncorrect or B3here*s no information+CH 3eacher does not kno1+H Scorer cannot concude
1hether ans1er is correct or incorrect+
n Teachers Score %re.uenc$ 3(roportion4 M
/ /-< '
(7 $0 !+$1# ;@A 67 !+6(# +6(
C-7- 0hich student group had the highest a'erage >/A score in 5rade 44 Kuestion8 Go1 easy 1as
it to get this information from the chart4
,u!ricC
'C B0hite femaesC or B1hite girsC
/-<C B0hiteCH Borange bar for 1hiteC
/C 7ncorrectH 3eacher does not kno1+H Scorer cannot concude 1hether ans1er is correct or
incorrect+
n Teachers Score %re.uenc$ 3(roportion4 M
/ /-< '
(5 4 !+04# 10 !+11# -1 !+-5# +(1
C-;- Go1 many African American thirdAgrade boys took the test4 Go1 easy 1as it to find this
information in the tabe4
,u!ricC
'C 1-
/C 7ncorrectH 3eacher does not kno1H Scorer cannot concude 1hether ans1er is correct or
incorrect+
n Teachers Score %re.uenc$ 3(roportion4 M
/ /-< '
(5 10 !+11# 0 -5 !+-(# +(0
='
Scenario D
3his !show Figure # is the kind of data tabe that some student data systems produce+ 7*m going to ask
some ?uestions about ho1 you 1oud interpret the data in this tabe+
%igure .
2005"06 Score /e'esE>ngish /anguage Arts !>/A#
English Language Arts
Hamilton Elementary
Grade Gender Ethnicity
Number
of
Students
Tested
Percent
of
Tested
Students
Mean
Scale
Score
Number Students at
Each Proficiency Level
Belo
Basic Basic Proficient Advanced
!
Female
African American 18 26% 439 5 7 5 1
Asian/Pac Islander 1 1% 610 0 0 0 1
Latino 17 24% 428 5 6 5 1
!ite 34 49% 449 4 13 11 6
"otal Female 70 100% 444 14 26 21 9
#ale
African American 18 23% 436 6 6 5 1
Asian/Pac Islander 2 3% 452 0 1 0 1
Latino 31 40% 430 8 7 14 2
!ite 27 35% 448 6 11 7 3
"otal #ale 78 100% 438 20 25 26 7
"
Female
African American 18 24% 441 3 8 5 2
Asian/Pac Islander 2 3% 462 1 0 1 0
Latino 36 47% 436 8 12 12 4
!ite 20 26% 472 2 7 8 3
"otal Female 76 100% 447 14 27 26 9
#ale
African American 16 23% 442 2 8 5 1
Asian/Pac Islander 0 0% $A 0 0 0 0
Latino 29 42% 438 5 12 10 2
!ite 24 35% 456 3 13 5 3
"otal #ale 69 100% 445 10 33 20 6
?&
D-'- Suppose you*re 1orking 1ith thirdAgrade teachers at this schoo and they*re interested in
e6amining ho1 their students performed in terms of the anguage skis measured on this test+
0hat does the data in this tabe indicate about 1hether boys and girs performed differenty in the
third grade4 SGet open responseT
=o ,u!ricC Coded onl$, not scored
D-+- 2N+ ;o1 7*m going to read a series of statements that peope might make about different aspects
of the 5rade $ data in this tabe+ 7*d ike you to te me for each statement 1hether you agree or
disagree and the reasons 1hy+ <emember to think aoud as you*re deciding on your ans1er+
a+ A maIority of third graders at this schoo ha'e not achie'ed proficiency in >/A as measured by this test+
,u!ricC
'C Agree S57: are ess than proficientH -5@14- 9 !14J26J20J25#@!70J7-#T
/-<C Agree !no e6panation#
/C .isagreeH 3eacher does not kno1+H Scorer cannot concude 1hether ans1er is correct or
incorrect+
n Teachers Score %re.uenc$ 3(roportion4 M
/ /-< '
(7 16 !+16# 2 !+02# 7( !+-1# +-2
b+ 7n 5rade $, boys 1ere more ikey than girs to score Deo1 Dasic on this assessment+
,u!ricC
'C Agree 25+6: boys V 20: girs !20@7- V 14@70#
/-<C Agree !20 boys V 14 girs#H Agree !no e6panation#
/C .isagreeH 3eacher does not kno1+H Scorer cannot concude 1hether ans1er is correct or
incorrect+
n Teachers Score %re.uenc$ 3(roportion4 M
/ /-< '
(6 1- !+1(# 47 !+4(# $1 !+$2# +5-
c+ 2f those students 1ho scored Deo1 Dasic in 5rade $, most 1ere /atino+
,u!ricC
'C .isagree !!ess than hal" of a of the students 1ho scored Deo1 Dasic in 5rade $ are /atino+#
/-<C .isagree 1ith no e6panation
/C AgreeH 3eacher does not kno1+H Scorer cannot concude 1hether ans1er is correct or
incorrect+
n Teachers Score %re.uenc$ 3(roportion4 M
/ /-< '
(6 71 !+74# 1 !+01# 24 !+25# +26
?8
D-7- ;o1 et*s assume that you*re a fourthAgrade teacher and these 5rade 4 data are for midAyear performance on
a benchmark test+ 7f there is a particuar group of fourth graders you think 1i be most ikey to ha'e troube
scoring basic or abo'e on the state test at the end of the year, coud you point out their data in this tabe4 Sif
appropriate probe 1ith one of the foo1ingT 0hich group 1oud you be concerned about and 1hat data
trigger that concern4 2< 0hy don*t you think the subgroup data in the tabe point to a priority need for any
particuar subgroup4
,u!ricC
'C 3hose scoring Deo1 Dasic
/C 7ncorrectH 3eacher does not kno1+H Scorer cannot concude 1hether ans1er is correct or
incorrect+
n Teachers Score %re.uenc$ 3(roportion4 M
/ /-< '
(6 -6 !+(0# 5 !+05# 5 !+05# +0-
=oteC
7tem responses did not correate 1ith those to other .ata 7nterpretation items+ Scores not incuded in
anayses+
D-;- ;o1 et*s go back to the 5rade $ data+ <emember that these are for ast year*s third graders+ 7f there ha'e
been no maIor changes in the schoo*s student body, teachers, or curricuum 1oud you e6pect that8
a+ 2n the basis of ast year*s test scores, girs can be e6pected to score higher than boys 1hen this test is
gi'en to this year*s third graders+
,u!ricC
'C .isagree S5rade $ girs and boys scored about the same, e+g+, !21J(#@70 9 !26J7#@7-T
/-<C .isagree 1ith no e6panation
/C AgreeH 3eacher does not kno1+H Scorer cannot concude 1hether ans1er is correct or
incorrect+
n Teachers Score %re.uenc$ 3(roportion4 M
/ /-< '
($ 44 !+47# 0 4( !+5$# +5$
b+ 3hese data suggest that ne6t year*s thirdAgrade Asian@P7 girs 1i score better than other third graders on
this test+
,u!ricC
'C .isagree !)annot predict ne6t year*s scores based on the one 5rade $ Asian@P7 student
represented in the tabe#
/-<C .isagree 1ith no e6panation
/C AgreeH 3eacher does not kno1+H Scorer cannot concude 1hether ans1er is correct or
incorrect+
n Teachers Score %re.uenc$ 3(roportion4 M
/ /-< '
($ $( !+42# 2 !+02# 52 !+56# +57
?%
Scenario E
3his !show Figure ## is a different kind of data dispayEa bar graph of 5rade 4 mathematics achie'ement separated into t1o components
!computation and probem so'ing# as 1e as their tota, for a schoo and its district for each of three years+ Again, 7*m going to ask you to find
some information on the dispay and then te me ho1 easy or hard that 1as to do on a scae from 1 to 10 1ith 10 being Be6tremey difficut+C
%igure >+ 3rend .ata Dar 5raph
7
3
E-'- 0hat 1as 2ak Schoo*s a'erage 3ota &ath Score in 200$"044
,u!ricC
'C Any specific ans1er from $$$ to $$-
/-<C Uague, but correct ans1er range pro'ided !e+g+, Bbet1een $$0 and $40C#
/C 7ncorrectH 3eacher does not kno1+H Scorer cannot concude 1hether ans1er is correct or
incorrect+
n Teachers Score %re.uenc$ 3(roportion4 M
/ /-< '
(1 4 !+04# $ !+0$# -4 !+(2# +(4
E-+- 0hat 1as the difference in the district*s tota math score in 2005"06 compared 1ith 200$"044
,u!ricC
'C Any ans1er from 1 to $ pointsH Ba fe1 pointsC or Bup a itte bitC or Babout the sameC or BsimiarC
/-<C Be6acty the sameCH Bthe sameC
/C 7ncorrectH 3eacher does not kno1+H Scorer cannot concude 1hether ans1er is correct or
incorrect+
n Teachers Score %re.uenc$ 3(roportion4 M
/ /-< '
(1 4 !+04# 10 !+11# 77 !+-5# +(0
;o1 7*m going to ask you some ?uestions about ho1 you 1oud interpret the data in this chart+
E-7- /ooking at the chart as a 1hoe, 1hat 1oud this data te you about fourth graders* mathematics
achie'ement at this schoo4 SGet open response$=o ,u!ricC Coded onl$, not scoredT 2N+ ;o1
7*m going to read a series of statements that peope might make about the data in this graph and
7*d ike you to te me for each one 1hether you agree or disagree and the reasons 1hy+ <emember
to think aoud as you*re deciding on your ans1er+
a+ 2ak Schoo does better than the district as a 1hoe in 5rade 4 mathematics+
,u!ricC
'C .isagree !2ak Schoo tota math scores are o1er than the district*s math scores+#
/-<C .isagree 1ith no e6panation
/C AgreeH 3eacher does not kno1+H Scorer cannot concude 1hether ans1er is correct or
incorrect+
n Teachers Score %re.uenc$ 3(roportion4 M
/ /-< '
(1 2 !+02# - !+0(# -1 !+-(# +($
?A
E-7 3continued4
b+ 2ak Schoo has made some impro'ement in 5rade 4 mathematics o'er this time period+
,u!ricC
'C Agree !2ak Schoo tota math scores are sighty higher in 2005"06 than in 200$"04+#
/-<C Agree 1ith no e6panation
/C .isagreeH 3eacher does not kno1+H Scorer cannot concude 1hether ans1er is correct or
incorrect+
n Teachers Score %re.uenc$ 3(roportion4 M
/ /-< '
(1 $1 !+$4# $ !+0$# 57 !+6$# +64
c+ <eati'e to the district as a 1hoe, 2ak Schoo fourth graders ha'e been getting better in their
probem so'ing skis+
,u!ricC
'C Agree !based on clearly identi"iable comparison of probemAso'ing scores for 2ak Schoo and
the district#
/-<C Agree !Dased ony on 2ak Schoo dataH no cear, identifiabe comparison 1ith district#H Agree
1ith no e6panation
/C .isagreeH 3eacher does not kno1+H Scorer cannot concude 1hether ans1er is correct or
incorrect+
n Teachers Score %re.uenc$ 3(roportion4 M
/ /-< '
(2 $- !+41# 10 !+11# 44 !+4-# +5$
d+ <eati'e to the district as a 1hoe, 2ak Schoo fourth graders ha'e been getting better in their
computation skis+
,u!ricC
'C .isagree !Dased on clearly identi"iable comparison of computation scores for 2ak Schoo and
the district#
/-<C .isagree !Dased ony on 2ak Schoo dataH no cear, identifiabe comparison 1ith district#H
.isagree 1ith no e6panation
/C AgreeH 3eacher does not kno1+H Scorer cannot concude 1hether ans1er is correct or
incorrect+
n Teachers Score %re.uenc$ 3(roportion4 M
/ /-< '
-- $2 !+$6# - !+0(# 4- !+55# +5(
?B
E-7 3continued4
e+ 2ak Schoo*s progress in narro1ing the 5rade 4 math achie'ement gap 1ith the rest of the district
has been in probem so'ing rather than computation+
,u!ricC
'C Agree !5ap bet1een 2ak Schoo and district scores decreases more for probemAso'ing than
for computation+#
/-<C Agree 1ith no e6panation
/C .isagreeH 3eacher does not kno1+H Scorer cannot concude 1hether ans1er is correct or
incorrect+
n Teachers Score %re.uenc$ 3(roportion4 M
/ /-< '
-( 26 !+2(# 6 !+07# 57 !+64# +67
E-;- Suppose 2ak Schoo had started using a ne1 mathematics program at the beginning of the 2004A05
schoo year 1hie the rest of the district continued 1ith the od program+ /ooking at these data, 1hat
are your thoughts about the ne1 curricuum4 SGet open response$=o ,u!ricC Coded onl$, not
scoredT 0hich of these statements 1oud you agree 1ith and 1hy4
a+ 3he math program appeared to impro'e achie'ement the first year it 1as used but the benefit
disappeared the ne6t year+
,u!ricC
'C .isagree !2ak Schoo is sti performing better in 2005"06 than it 1as in 200$"04+#D Agree 1ith
reser'ations !3eacher takes issue 1ith 1ord BdisappearC+ %or e6ampe, B7 agree, it did impro'e
achie'ement the first year and scores are o1er in the second year, but the benefit did not totay
disappear+C#
/-<C .isagree 1ith no e6panation
/C AgreeH 3eacher does not kno1+ Scorer cannot concude 1hether ans1er is correct or incorrect+
n Teachers Score %re.uenc$ 3(roportion4 M
/ /-< '
-- 5$ !+60# 1 !+01# $4 !+$(# +$(
?=
E-; 3continued4
b+ 3he ne1 math program appeared to hep students 1ith their computation skis+
,u!ricC
'C .isagree !Dased on comparison of computation scores for 2ak Schoo and the districtH 5ap
stays reati'ey constant each year+#
/-<C .isagree !ook at 2ak Schoo ony#H .isagree 1ith no e6panationH Agree 1ith reser'ations
!i+e+, Bgap is pretty consistent, but the program heped a itte bitC#
/C Agree and focus ony on 2 years of data !0$"04 and 04"05#H Agree and say that B05"06 is
better than 0$"04C but do not say Bsighty betterCH 2ther agreeH 3eacher does not kno1+H Scorer
cannot concude 1hether ans1er is correct or incorrect+
n Teachers Score %re.uenc$ 3(roportion4 M
/ /-< '
(0 6( !+77# 15 !+17# 6 !+06# +15
c+ Pou can*t be sure 1hether the program is ha'ing an effect because there may be differences
bet1een the different casses of fourth graders+
,u!ricC
'C Agree !e+g+, ackno1edges that there are differences in the popuations of students each year#
/-<C Agree 1ith no e6panation
/C .isagree !e+g+, Program 1i ha'e an effect regardess of the student popuation in'o'ed+# H
3eacher does not kno1+H Scorer cannot concude 1hether ans1er is correct or incorrect+
n Teachers Score %re.uenc$ 3(roportion4 M
/ /-< '
(0 22 !+24# 20 !+22# 4- !+5$# +64
=oteC
7tem responses did not correate 1ith those to other .ata 7nterpretation items+ Scores not incuded in
anayses+
??
E-; 3continued4
d+ Scores mo'e around from year to year, but the ne1 math program appears promising and shoud
be monitored for more years+
,u!ricC
'C Agree !e+g+, some indication that program benefits studentsH need more than 2 years of data to
determine if a program is 1orking#
/-<C Agree 1ith no e6panation
/C .isagreeH 3eacher does not kno1+H Scorer cannot concude 1hether ans1er is correct or
incorrect+
n Teachers Score %re.uenc$ 3(roportion4 M
/ /-< '
(1 10 !+11# 15 !+16# 66 !+7$# +-0
E-<+ Are there other issues or possibe e6panations that shoud be taken into account as 1e4
=o ,u!ricC Coded onl$, not scored
?E
SCE=A,I9 %
Suppose you*re in a meeting to discuss 2005"06 reading data from the state assessment for your
schoo*s third grade and they hand out this data dispay+ !Show Figure F%#
%igure %
Gistogram
Grade !
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
%elo& %asic %asic Proficient Ad'anced
Le'el
P
e
r
c
e
n
t
a
(
e
?'
%-'#'- Dased on this chart, 1hat percentage of the schoo*s third graders ha'e achie'ed proficiency4
,u!ricC
'C Any ans1er from 44 to 46
/-<C >'idence that correct math is being used, but no fina ans1er !e+g+, B7 need to add $-: and
6:+C#
/C 7ncorrectH 3eacher does not kno1+H Scorer cannot concude 1hether ans1er is correct or
incorrect+
n Teachers Score %re.uenc$ 3(roportion4 M
/ /-< '
47 2- !+60# 1 !+02# 1- !+$-# +$(
%-'#+- Dased on this chart, 1hat percentage of the schoo*s third graders 1ere ess than Proficient in
reading4
,u!ricC
'C Any ans1er from 64 to 66
/-<C >'idence that correct math is being used, but no fina ans1er !e+g+, B7 need to add 2$: and
42:+C#
/C 7ncorrectH 3eacher does not kno1+H Scorer cannot concude 1hether ans1er is correct or
incorrect+
n Teachers Score %re.uenc$ 3(roportion4 M
/ /-< '
4( 12 !+24# 6 !+12# $1 !+6$# +6(
%-+- 2ne of your coeagues, after ooking at these data says, B3here*s something 1rong 1ith this chart+C
0oud you agree4 0hy or 1hy not4
,u!ricC
'C 7f agrees A;. says that Bnumbers don*t add to 100C or Bnumbers add to more than 100C
/C 7ncorrect !)an*t find anything 1rongH 7dentifies something ese as B1rongC#H 3eacher does not
kno1+H Scorer cannot concude 1hether ans1er is correct or incorrect+
n Teachers Score %re.uenc$ 3(roportion4 M
/ /-< '
(6 42 !+44# ;@A 54 !+56# +56
E&
Scenario 8
Suppose you*re teaching in a district that re?uires students to attain eighthAgrade proficiency in
mathematics in order to enro in Agebra 7 in high schoo+ /ooking at students* performance the
preceding year, you found the resuts in this tabe !show Figure G# for the /atino and African
American students 1ho make up your schoo*s entire student body+
W A score of 65: on the district math test is considered proficient, and
W A schoo is considered Bo1 performingC if ess than 50: of students in any student
subgroup reach proficiency+
%igure 5
2005"06 Achie'ement in 5rade - &athematics
5roup ;umber of Students &ean &ath Score Percentage
Proficient
;umber Proficient
/atino 22- 67+5 61 1$(
African American $1 66 4- 15
E8
8-'o- 0hat do these data te you about ho1 1e students are doing at your schoo4
,u!ricC
'C Any correct obser'ation or inference from the tabe, incuding BAbout 60: of /atino students
are proficient+C BA smaer proportion of African American than of /atino students are proficient+C
B3he a'erage score in the schoo is Iust a itte abo'e the proficiency criterion+C
/C 1"2 correct genera statements, but ceary is not reading the tabe correcty !e+g+, B/atinos are
doing better than African Americans, but both groups are not proficient+C#H 2ther 7ncorrect !.o not
reference anything in the tabe, 'ery 'ague#H 3eacher does not kno1+H Scorer cannot concude
1hether ans1er is correct or incorrect+
n Teachers Score %re.uenc$ 3(roportion4 M
/ /-< '
-2 24 !+2(# ;@A 5- !+71# +71
8-'- ;o1 7*m going to read you some statements again and 7*d ike you to te me 1hether you agree or
disagree 1ith each and 1hy+
a+ 3he maIority of our eighth graders are proficient in eighthAgrade math+
,u!ricC
'C Agree S5(: 9 154@25( 9 !1$(J15#@!22-J$1#T
/-<C Agree 1ith no e6panationH Agree 1ith incorrect reasoningH .isagree 1ith different definition
of maIority !e+g+, says 5(: is not a maIorityH may state correct numbers in'o'ed in cacuation,
but sti disagree#
/C 2ther .isagreeH 3eacher does not kno1+H Scorer cannot concude 1hether ans1er is correct or
incorrect+
n Teachers Score %re.uenc$ 3(roportion4 M
/ /-< '
-5 16 !+1(# 41 !+4-# 2- !+$$# +57
b+ 2ur schoo is not getting enough African American students to proficiency, but /atino students are
meeting the re?uired performance standard+
,u!ricC
'C Agree !61: for /atinos is greater than the 50: re?uirementH 4-: for African Americans is ess
than the 50: re?uirement#
/-<C Agree 1ith no e6panation
/C .isagreeH 3eacher does not kno1+H Scorer cannot concude 1hether ans1er is correct or
incorrect+
n Teachers Score %re.uenc$ 3(roportion4 M
/ /-< '
-6 $0 !+$5# - !+0(# 4- !+56# +60
E%
8-' 3continued4
c+ 2ur schoo is cassified as o1 performing based on these mathematics scores+
,u!ricC
'C Agree !African American subgroup is ess than 50: proficient+#
/-<C Agree 1ith no e6panation
/C .isagree !e+g+, B5(: of a students are proficient+C#H Agree based on 65: criteria !e+g+, BDoth
groups are beo1 65: proficient#H 2ther agree 1ith incorrect reasoningH 3eacher does not kno1+H
Scorer cannot concude 1hether ans1er is correct or incorrect+
n Teachers Score %re.uenc$ 3(roportion4 M
/ /-< '
-6 25 !+2(# 7 !+0-# 54 !+6$# +67
8-+- 0hat actions shoud your schoo consider to a'oid being abeed Bo1 performingC in the coming
year4 Sa"ter open responseT 0hich of these statements do you agree 1ith based on these data4
!>6pain your ans1er for each+# SFor each statement, give teacher an inde& card with the
statement on it to keep handy as he'she looks at the data%T
a+ 3his year a eighth graders shoud get more intensi'e instruction in mathematics+
b+ 3his year a African American students shoud get more intensi'e instruction in mathematics+
,u!ric
'C 5i'es same ans1er for 5+2a and 5+2b 1ith appropriate e6panation
/-<C Same ans1er for 5+2a and 5+2b, but no e6panation for one of the itemsH 5i'es the same
ans1er for 5+2a A;. 5+2b 1ith no e6panation for either item
/C 5i'es the same ans1er for a and b 1ith an incorrect rationae !B7 disagree because the tabe
doesn*t sho1 a students in the schoo+CH B7 agree because a of the students are not up to the
proficient e'e+C#H 5i'es different ans1ers for 5+2a and 5+2bH 3eacher does not kno1+H Scorer
cannot concude 1hether ans1er is correct or incorrect+
EC
n Teachers Score %re.uenc$ 3(roportion4 M
/ /-< '
7( 25 !+$2# 20 !+25# $4 !+4$# +56
8-+ 3continued4
c+ 3eachers shoud obtain a detaied breakdo1n of ast year*s test resuts by item or content
standard+
,u!ric
'C Agree !1ith reasonabe e6panation about why ha'ing detai is important#
/-<C Agree 1ith no e6panation
/C .isagreeH 3eacher does not kno1+H Scorer cannot concude 1hether ans1er is correct or
incorrect+
n Teachers Score %re.uenc$ 3(roportion4 M
/ /-< '
77 0 20 !+26# 57 !+74# +-7
d+ 2ur eighthAgrade mathematics program isn*t necessariy Bbroken,C there*s a good chance that 50:
or more of African American students 1i meet the proficiency re?uirement this year+
,u!ricC
'C Agree !e+g+, ackno1edges that sampe size is sma and 1oud be possibe for 1"2 students to
do better thus bumping the percent proficient o'er 50:H states that 4-: is cose to 50: and
1oud be possibe to get to 50: this year#
/-<C Agree 1ith no e6panation !no detai, missing e6panation regarding why African Americans
coud meet 50: e'e#
/C .isagreeH 3eacher does not kno1+H Scorer cannot concude 1hether ans1er is correct or
incorrect+
n Teachers Score %re.uenc$ 3(roportion4 M
/ /-< '
7( 27 !+$4# 10 !+1$# 42 !+5$# +5(
EA
Scenario *
;o1 7*m going to ask you to think about some data from a hypothetica cassroom+
Suppose you*re teaching mathematics and at the end of a unit on measurement, you ga'e a 100Apoint
test on measurement concepts and skis and your students obtained the scores sho1n in this cass ist+
SShow Figure (+T Pou kno1 that students from this schoo ha'e had troube 1ith measurement items on
the state test in pre'ious years, and you*re 1ondering 1hether you need to do more teaching in this area
or can mo'e on to the ne6t topic+ Pou takes these scores into the teachers* ounge and ask coeagues to
take a ook+ 0hen they ask about the test you e6pain that you designed it so that if a student gets a 70:
or better on it, you are reay ?uite confident that he or she understands the concepts+ 0hen a student*s
score is o1er than that, you fee there is something they sti don*t understand+
2ne of your coeagues pus out his cacuator and sho1s that the mean for these scores is 70+5+ B3he
mean score is greater than 70+ Pou*'e done your Iob+ &o'e onX 3here*s ots more math to co'er+C
%igure G
)assroom .ata Set
Student
Number Score
1 95
2 65
3 85
4 55
5 90
6 70
7 100
8 80
9 65
10 80
11 65
12 90
13 65
14 55
15 45
16 15
17 70
18 65
19 75
20 85
21 65
EB
*-'- .o you agree 1ith this coeague4 0hy or 1hy not4
=o ,u!ricC Coded onl$, not scored
*-+- 0hat do you think 1oud happen if you ga'e the same cass of students another test on
measurement the ne6t day4
=o ,u!ricC Coded onl$, not scored
*-7- 7f you 1ere the teacher, 1hat 1oud you do4
=o ,u!ricC Coded onl$, not scored
E=
Appendi C
Scenario De#elopment and ,elia!ilit$ Information
E?
Data Scenario De#elopment
Because of the projects interest in getting at multiple, distinct facets of teachers
use of data (question posing, data location, data comprehension, data interpretation and
data use), an essentially criterion-referenced approach was taken to the development of
the data scenario assessment items. The data scenarios were developed for this study with
the advice and review of experts in data-informed decision making and assessment. The
experts judged that the items represented the intended constructs. In addition, the
scenarios were pilot-tested to make sure that they were comprehensible and tapped the
intended skills and concepts. The mean scores for the individual items are displayed in
Exhibit C-1.
E'
Ehi!it C-'
%re.uenc$ Distri!utions, Mean Scores and ,elia!ilities for Scored Items
Item ID
n
3Teachers4
Score %re.uenc$
M
3SE4
Alpha
/ /-< '
%orm
'
%orm
+ %orm 7
Kuestion Posing !1 item#
A+1 -5 26 2$ $6
+56
!+046#
+46( ;@A +5-(
.ata /ocation !( items#
%+1'1 47 2- 1 1-
+$(
!+071#
+520 ;@A ;@A
%+1'2 4( 12 6 $1
+6(
!+062#
;@A +461 ;@A
)+2 (7 $0 ;@A 67
+6(
!+047#
;@A +450 +566
)+1 (7 25 ;@A 72
+74
!+045#
;@A +4($ +604
)+$ (5 4 10 -1
+(1
!+025#
;@A +4-1 +602
D+2 4( 6 ;@A 4$
+--
!+047#
+47- ;@A +$74
D+1 4( 5 ;@A 44
+-(
!+0$2#
+466 ;@A +175
)+4 (5 10 0 -5
+(0
!+044#
;@A +451 +56-
>+1 (1 4 $ -4
+(4
!+02$#
+471 ;@A +612
.ata )omprehension !16 items#
>+4b (0 6( 15 6
+15
!+0$1#
+475 ;@A +5-(
.+2b (6 1- 47 $1
+57
!+0$6#
;@A +414 +5(5
5+1a -5 16 41 2-
+57
!+0$-#
;@A +45- +610
>+4a -- 5$ 1 $4
+$(
!+052#
+502 ;@A +564
>+$c (2 $- 10 44
+5$
!+04(#
+$7- ;@A +526
>+$d -- $2 - 4-
+5(
!+050#
+46( ;@A +570
%+2 (6 42 ;@A 54
+56
!+051#
+47( +414 ;@A
5+1b -6 $0 - 4-
+60
!+050#
;@A +4$4 +5(5
>+$b (1 $1 $ 57
+64
!+050#
+46( ;@A +564
'&
Ehi!it C-' 3continued4
%re.uenc$ Distri!utions, Mean Scores, and ,elia!ilities for Scored Items
Item ID
n
3Teachers4
Score %re.uenc$
M
3SE4
Alpha
/ /-< '
%orm
'
%orm
+ %orm 7
.ata )omprehension !continued#
5+1c -6 25 7 54
+67
!+04(#
;@A +455 +562
>+$e -( 26 6 57
+67
!+04-#
+4(6 ;@A +5(6
5+1o -2 24 5- -2
+71
!+051#
;@A +455 +605
>+4d (1 10 15 66
+-1
!+0$6#
+4(4 ;@A +5-6
.+2a (7 16 2 7(
+-2
!+0$-#
;@A +456 +5-$
>+2 (1 4 10 77
+(0
!+026#
+4(6 ;@A +5((
>+$a (1 2 - -1
+($
!+021#
+474 ;@A +5(1
.ata 7nterpretation !5 items#
.+2c (6 71 1 24
+26
!+044#
;@A +4$4 +567
5+2ab 7( 25 20 $4
+56
!+04-#
;@A +471 +5-7
.+4a ($ 44 0 4(
+5$
!+052#
;@A +444 +60$
5+2d 7( 27 10 42
+5(
!+052#
;@A +4-0 +5($
.+4b ($ $( 2 52
+57
!+051#
;@A +4-0 +55-
.ata Lse !1 item#
5+2c 77 0 20 57
+-7
!+025#
;@A +471 +57$
5ote. )ere +ere tree forms of te data scenario intervie+ and eac item appeared on t+o of te
tree forms.
'8
,elia!ilit$ and Scores for Data Scenario Euestions
For purposes of analysis, those questions on the data scenarios with right and wrong
answers were considered to be assessment items. In a case such as the present one, in
which the intent is to measure performance relative to a standard rather than performance
relative to peers, professional test publishers would estimate reliability through
techniques associated with item response theory. These procedures require several
hundred examinees for each item, however, and thus were out of reach for a research
project of the present scope.
Classical test theory can be applied with fewer respondents and examines the
relationship of each item to performance on the rest of the items. The complication in the
current instance is the matrix sampling of items done to manage the burden imposed on
individual teachers. There were three forms of the data scenario interview, and each data
scenario appeared on two of the three forms. To provide some indication of reliability,
analysts computed alpha coefficients, examining the relationship between an individual
item and the rest of the items in the data scenario interview, for each item and each of the
two forms of the interview that included it. These coefficients and shown in Exhibit C-1.
However, the purpose of the data scenario interviews was not to produce individual
teacher-level scores. The goal was to provide an estimation of data literacy at the school
level. Different teachers took different assessment forms containing different subsets of
the items, and different teachers total scores are not directly comparable.
Scales were analyzed at the school rather than the individual teacher level. That is,
the scores for all the teachers at a school who took item 1 were averaged and that mean
score was assigned to the school; likewise for item 2, etc. Through this process, the data
set was structured as 27 school-level records with a score for every item.
Reliability was estimated by computing an alpha coefficient for the total score.
Despite the small school sample and the intentional inclusion of distinct abilities in the
assessment, the reliability for the total score was fairly high, alpha = .74.
For all statistical comparisons reported in the body of the report, the school-level
total score was the variable used in the analysis.
'%