ED 160 241
AUTHORTITLE
DOCOBEIT EIS
. ,
Granville, -Atthut,-; -Ove, ohm M. 4-A Process Evaluation df Prdject DevelopmentalContinuity. .Interim Retort VI:,,%-lxetutive Summary.
- RetoMiendations for Continuing. ille-.mPaot Study..INSTITUTION , High/Scope -Educational fesearth-. Foundation,
Apsilanti, Mich.SPONS AGENC1 Offite of Child DeVelopment 1DEEW), Washington D.C.
-Eirly:-Childhood Research Ardivaluation-Btana.A
'PUB DATE Apr. 77CONTRACTNOTE,
HEW-105-75--111424p0;-For related documents, see ED 144 715, PS. ©10163-166', and PS 01t 17C-176; This -series includes allthe public reports generated by this:study.
. .
EDRS PRICE MF-$0.83 D,C-$1,67 Plus Postage. ,DESCRIPTORS Attrition (Research Studies) ; *Deionstration
Programs; 41,Early Childhood Education; *Feasibilit.y.Formative Evaluation; IongitudinalStudieS4_
Matched GroUps; Measurement. InstruMentsvErogramectiveness;,*Program Evaluation; Relietility;
Research Methodology 1 4Researchjroblems;-ValiditYIDENTIFIERS- *Developmental Continuity;-.*Prc ectDevelopmental
-Continuity; Projett Head: Start
ABSTRACTThis brief report sumsar zes the analysis and
conclusions presented in detail in Interim Report VI regarding thefeasibility of conducting a longitudinal stlidy cf ProjectDevelopmental Continuity (PDC). This project is a Bead Startdemonstratien program aimed at 'providing educational anddevelopmental continuity between children's Head Start and primaryschool experiences.- The analyses were carried out cn informationcollected primarily during Fall 1966, and fccased on the reliabilitynd validity of' the measures, the potential comtarability of PDC and
comparison groups, and the adequacy of PDC and ccmtariscr samplesize.. (Author/CM)
****************** *44*eproductions supplied by. EDES are the best tha
from the origiAal document.*********************4**********
*********made
*
U'5 DEPARTMENT OF HEALTHEDUCATION L WELFARENATIONAL INSTITUTE OF_
EDUCATION
HI% IlN51- NT HAS BF_EN PEPE?OutED F xAt i4 y RECEIVED PROMiHF pf P,,nr4 3W ORGANIZA TION ORIGIN,
TIrat. OF WIF vt, OR OPINIONS
ICI NATIOt4A4 IN TrTUTE OFDIE ATF 00 OT KNECA-SSAPILY PEP ft
DFOtJi.TION Po ozioNpoP PO icy
4
f
ConInter Report VI
"vetmenda ions
inning The Impact Study
prAmi-,sioN T17 fiLPROI)1_1( E T144S
l3 IEITIAL FIX; HEEN IFHANTEll Hr
10 THE Filth t\TION 'FAL RESOURCESf,l()N Cr:1\11EW (E111[ 1 AND
(P,a-liS OF THE THU:SYSTEM"'
April 1977
This report a6 irepared for the-Early Childhood ,Research arsi,Evaluation Branch,
Office of Id Deveropment, Department of- Health, Educatiom,:and Welfare,
under C9qtr4t No HEW7105-75-111A (Dr-: Esther Kresh, Project Officer).."
Views or'oon4lusion.sicontained herein should not be interpreted as reflecting
the offici6liposition of the!sponsorihg-agency0
v A PROCESS EVALUATION OF PROJECT D VELOPMENTALCONTINUITY, INTERIM REPORT VI: EXECUTIVE SUMMARY
ReCommek&tionsfor Continuing he act Study
ApriY-1977
Arthur C. Granvilleand
John M. Love
cope Educational Reseauh'Foundation'600 North River Street'
Ypsilanti, Michigan 48197
A
The present volume summari es our assessment of the feasibility ofconducting a longriudintalAludy of Project. Developmental Continuity. The
analysis, and discussions n the measurement of PDC's impact on the socialcomplete interitm Report ttlip a culmination of two years of data colLect?bn,
competence of Head Start Children. We,want'to acknowledge the -Contributions_of the many/persons and groups iho have been part of our evaluation sinceit began Lc, November l 74.
To/Our governm project-officer, C her Kresh, we owe a specialdebt o 'gratitude. Her support, guidanc and advice on key issues andher c ntinuin,g encouragement and cooperation have formed a mainstay forour valuatiOn effort.
/To nation OCD program staff, including Ray Collins, Jenny KleinJ aniIaiDenni Laura Dittmann, Soledad'Arenas, Judy Ramire2 and Georglanna'cGuire; */ tend our thanks for clarifying program goals and expandingur knowled base of PDC activities.
To th.. OCD regional representatives, grantee staff, and to the PDC,,1
coordina rs, parents, teachers, administrators, community representatives,and hil en at every PDC site across the nation, 'whO have given their 4
tim talents to shape the PDC program, we Can'only express our gratitude.Withou the commitment and contribution of each of these indivjduals therewould,Vof ,course, be no PDC program to evaluate. Wi,thout exception, theircOoperiation with the demands of the evaluation has been very gratifying.
We are also grateful to Charles Billings, Ricardo-Cornejo, Edward/Gottst, Robert Egbert, J. Ward Keesling and Eugene Litwak, who as membersof .our National Advisory Panel, have read alljnterim reports and providedvaluable criticism, suggestions and recommendations.
/Finally, Jae are indebted to our many colleagues at High/Scope whohave provided valuable input and support throughout the course of the study.In,particular, we want to acknowledge the contributions to the presentreport.of Judy McNeil, Mel Shelly, Mary Morris, Judy Meece, Sally Wackerand Lynn Spencer,
Arthur C. Granville
-John M. Love
Table of Contents'
INTRODUCTION: . . , .
Overview Project Developmental Continuity (PDC)
Purposes. of the POC'Evaluation'
Purpose of this Report.
II. METHODS
Measures
Data Collect on
Data Ana1ysis
1.11-. FINDINGS
1
2
5
7
,, Are the ftasuring Instruments Appropriate to th6 Task?. . 7
Reliability 7
Validity 7
Sensitivity to change 8
Relevance to social competence.
Suitability for'usb in the higher grades= 9
,Are--the PDC and Comparison Groups Really CoMparable? 9)
Will targe Enough PDC and Comparison Samples Remain toPermit a Longitudinal Study'? .tp
Ease of-administration
V. CONCLUSIONS 13
SuMmary 14
I.Table 1 Composition OF the Samples, Fall 1976 . . .k.. 15
Table / ,Reliab4lityof the Child Measures: Cronbach's Alpha(Internal Consistency) for Fall 1976 Meadtart'Children 16
Table 3:. Comparability of PDC and Comparison Groups at EachSite and for Samples 'Aggregated Across - Sites, Fall'1976. 17
Table .Projected Retention of Cohort '2 Children-for Each Yearof the Prospective Longitudinal Study. = = = ; --
Tabu_ 5 `Summary of Findings on Characteristics of M0aYuresIncluded in the Fall 1976 Battery. . . 19
INTRODUCTION
Overview of Project Developmental Continuity (PDC)
The Office of Child Development originated Project ,DevelopmentalContinuity (PDC) in 1974 as a Head Start demonstration program "aimedat promoting greater continuity of education and comprehensive childdevelopment servrce9 for childvn as they make the transition from pre-school to school ". The single n1ost importan't effect of this u ertaking,it is hoped, will be to-enhanEe the social competence of the c ildrenservedthat is, to increase their everyday effectiveness in (leafingwith their envirormtint (at school, at home, in the community il4
society). PDC also aims to bring about broader and more in ensivp in-volvement of parents and teachers in the governance of scho 1 affairs,and to promote poAitive change in the institutional process even beyondthe people who may occupy the institution at a given time. ,
_
As part of the own] Head Start improvement and iinnov tion effort,PDC.emphasizes the involvemelt of dadminisirators classroom s aff, andparents in formulating educational gois/and developing a co prehensive
'curriculum. The object is to ensureTTTat children receive continuousindividualized attention as they progress from Head Start through theearly primary grades. If the rogram is successful, existing disc tinu-ities between Head Start and ementary school experiences will,
mgreduced
by PDCmechaniisms at encou e communication and mutual decision g
4`.,
among preschool anN elementary.sc fool teachers, administrators, andparents. ---V
School organizations at 15 sit around the country received OCDfunding during 1974-75-to design and'plan implementation of seven_pre-.scribed componnts:
Administtration; administrative coordination between and withinHead Start and elementary school;
-Education: coordination of curriculum approaches and educationalgoals;
Trainin-: preservicetraining in program-r
-d in9ervice eaCher, staed areas;
1
and parent
Developmental support services:, comprehensive services medical,nutritidna and social to children and families;
oil Parent involvement: parent participation in policy-making., homeschool activities, and classroom visits or volunteering.;
Services. for the handicapped= services for handicapped childrenand children with learningidisabilities;
Bilin ualibicultural and multicultural_ education: programs forbilingualibitultural or multicultural children.'
During Year II,'197576, 14 sites (one-had withdraWn voluntarily),comprising a total of 42 Head Start centers and elementary schools,began their "start-up" year, pilot testihg their adaptations of- the PDCprogram. In 1976-77 NC js supposed to exist inwature form af_the 13sites that were refunded for Year III. if a longitudinal study of PDCis commissioned, the program's effects will be examined throughout theperiod beginning with the present year and continuing until the endof the 1980-81 school" year. During this period chi] en in the currenttesting samples (Cohort 2) will progress from Head through'grade 8.
Purposes of efie_PDC Evaluation
The major purpose of the'POC evaluation is to aid.the Office ofChild Develdpment in its efforts to design effective programs for earlychildhood education. To accomplish this, the evaluation will ultimatelyhave to assess PDC's impact in four areas: children's social competence,parent participation and attitudes, teacher attitudes and work stylesand the organizational climates oft the schools.
In addition to describing the consequences -of PDC, the evaluationwill describe and-analyze the processes that,led to those consequences.1It is- important to emphasize here that the aims of the total. evaluationare to produce conclusions about what happened (impact) Snd how and whyit happened'(proc4ss). This information will facilitate future decisionsabout whether the program should be replicated, and if so, how replicationcan best be aFomplished in the light of past experience.
Purpose
I
This report summarize. the analySes and conclusions of interim ReportVI, The analyses were carried out on information collected primarily
Volume 2 of Interim Report IV (August 1976) describes the plans andprocedure: of the Implementation Stu'cly.
2
duringthe fall of 1976.0n. children-who entered PDC er comparison Head.Start programs, and were directed toward answering three criticeqUeStiOns related to assessing impect.on children:
1. Are the Measuring instruments appropriate to thetask?'
Are the PDC and comparison groups really comparable? ,
Will large enough PDC and comparison sampleeremain to permita longitudinal study?
Preliminery `information related to these three questions was gathered irr
1975=76., 0nthe basis of answers.gained that year, some instruments were-eliminated from the battery, and:others were-modified.; recommendations,:were.made for Head Start enrollment procedures in order to balance RDand comparison groups ,in.certein ifflportent-respectS; and it was decithat the samples would, probably. remain large enough over a five-yearterm to permit group comparisons, but that attrition should be fe-assessedin-1976777.1
in'-this report, the fall 1976 firldIng are integrated with previousfindings and their implications for a longitudiAal study, of POC's impactare considered. The discussion concludes with recommendations for thefuture oif the evaluation. 1,0
ese cbnclusions are documented and discussed in Inieri Reportolume1 Pilot Year Impact Study: Instrument Characters tics and
Ataritign Trends_, August 1976.,
L
METHODS
Measures
The measures on-which this 'report focuses are tests of'cOgnilik4 andlanguage development, psychomotor development, and social-emotional-behavior (see list in Tab-+e ) 1 Additional data on the program's impacton children are obtained through teacher and tester ratings of children'sbehavior and a classroom observation system.
Data Collection
observatiop-s and ratings were completed in fall 1976 by 36testers/oPserver hired from-each of the PDC communities and trained byHigh/Scope staff.- Standard procedures for checking accuracy of testingprocdures and quality ape data were followed.2
Data collection began in mid7September and was completed within nineor ten weeks at each site Across the 12 sites, l',219 children were testedand rated, and observations were completed in 80 classrooms. The compositionsof the samples in terms of dekolraphiC characteristics are presented bysite in Table 1.
Data Analysis
The data analyses procdeded through a al4uence Of six steps whichfocused first on' the characteristics of the in truments and then on thecharacteristics of the samples. These steps p oyided information on thefolloWing:
,
lb °reliability of the instruments',
validity of Che instruments,
cross -time and cross-sample congruence of reliability and validity'., findings (examination of fall 1976 data in relation tO idata from
fall 19-75 and spr_ g 1976),
.-1The rationale for selecting these measures dbcumented in Interim Repirt
II: Recommendations 'for measuring pro_gram im ct, June 1975.
2These procedures are described fully in Interim Report VI and in the FieldProcedures Manual prepared for the High/Scope field staff.
a. factor structure of the battery.,
comparesi1itY of 4,PDC and Comparison samples,
a adequacy of present sample sizes in vixew of projected a i ion
FINDINGS
The three questions addressed in this report are particularlycritical_as the PDC evaluation begins what could become a five-yearlonOtudinal study of the(benefits that result from children's partici-pation in Developmental Continuity programs from Head' tart throughthird grade. In this section key findings relating to the threequestions are summarized following a brief discuss.ion of the importanceof each question.
Are the Measuring Instruments Appropriate to the Task?
Since an important goal of PDC is to enhance the social competenceof children, it is essential that the instruments used yield measuresthat, collectively, represent social competence in an, accurate and meaning-ful way. Six criteria have been used for judging the adequacy of theinstruments: reliability (internal consistency), validity (congruencewith, expectations),= sensitivity to change over time, apparent relevanceto social competence, suitability for use in higher grades, and ease ofadministration.
Reliability. Table 2 summarizes the reliability findings for thetests. The internal consistency reliability coefficient was .65 orgreater for all measures in both the English- and Spanish-dominant samples'.Most measures. have remained constantin their reliability indices acrossthe three timepoints at which they have been administered during thisevaluation. Charii:s in scoring have increased the reliability ofitwomeasures, but the reliability-of another measure has declined slightly.
Validity. The validation procedures involved determining the expectedrelatiOnship ofeach measure with each of.the others, then comparing these ,
expectations with the relationships; that actually appeared in the data.Under this convergent-discriMinate method of assessing validity, theassumption is made:. that if.an instrument is actually measuring the con-struct it is intended to measure, the results will= correlate,highly withother measures of the same general construct, will correlate moderatelywithmeasures ofsimilareonstructs, and will not correlate at all withmeasures of -independent constructs. All the instruments examined areacceptably valid for Head Start children, as eyidenced,by the stabilityof their validity indices across two cohorts and threeltimepoints,
etsILLLit.charige. Since t \Impact Study depends upon thePDC battery of measures to detect ch nges that can-be attributed to pro-,gram differences, three types of analyses based on the pilot samples'fall 1,75 and spring 1976 data were carried out=
The correlation of ea&h measure in the fall and the spring
withchild age at the time of'testing was calculated todetatmine'the age-relatedness of the measures; the correlationstended to beldw,,p,ositiveT, awd significant, with coefficientsgenerally between x15 and 0. )
Tb.e diffprence between the fall mean score and the spring meanscore on eats measure was analyzed to a'scerlain if the scoresincreased significant) `from fall to spring; all-measqre's exceptthe Bilingual Syntax Measure-English showed a significant fall-to-spring increase.
4 ^
A regression procedure was used to determine whether the observedspring mean on a measure-Was equal to or greater than the ex-`petted, or predicted, spring men; more than'half of the childrenobtained an actual spring score equal to or greater than their.
`'expected spring score indicating that the tests are sensitive tochange due to educational experience as 'well as to experiencethat is simply a function of increased age.
Aelevance to social com tence. Since the PDC battery was constitutediith the intent of measuring the traits thal comprrSe\social cot4eteniC
an analysis was performed for the August-1976 Impact Study report thatexamined the relationship of spring 976 test scores to ad hoc criteria=of social competence. The-criteria were established by factor analyzing
.ratings completed by each.child's teacher and tester, and then creatingfactor scores for each child that represented his or her status on eachof the "social competence" factors. The assessments provided by the'teachers and testers are based Soon observations of each child's behaviorin a variety Of forma.) ,and informal situations, and thuslogically comedose to representing measures cif the child's "everyday effectiveness",
social competence:.
The object of the analysis (a linear regression vroeedure) was todetermine the magnitude of the relationship existing between the tests
.incloded in the PDC battery and the "social competence",eriteria. The
more relevant the tests are to social competence, the stronger therelationship expected. All tests except Arm Coordination were found tobe substantially lissoa.iated with the collective "social competence"criteria. Thus, these tests, originalle selected for their theoreticalrelevance to social competence, seem to provide measu res that are
`eempirically relevant o'social competence as well.
8
Suitabili use in the higher radeS. During the 1975-76 testingperiods, app-roximately 25 children per grade (kindergarten through grade 3)were tested at the Georgia site as part of the cross-seetional designthere. In addition, 30 third graders were tested in Ma-ryland.
Conclusions about the suitability of the child measures for use ateach of these grades were based on four factors: response distributionson the items of each measure, mean scores or#each measure, reliability(internal consistency), and validity. Based on these factors, all of themeasures appear to be useful through grade 3, either in their presentforms or'with modifications.
Ease of administration. One of the-factors taken into considerationwhen tests were being reviewedrfot the PDC Impact Study was their generalsuitability for adOlistration by a paraprofssional. In general,monitoring of test,ets during training and data collection indicates thatthe tests have pdt been difficult to administer. Tester performanceimproves with practice and administration(difficulties are more apparentwith new testers than with-experienced:ones.
Are the PDC and Oom-arison Grou Beall arable?
The effects of PDC upon children will be determined primarily bycomparing the performance of chfldren%in PDC testing samples with theperformance of children who are similar, but who do not participate inPDC (i:e., a comparison group). The assumption implicit in this comparisonis that the children' in the two groupS would remain parallel were it notfor the intervention of PDC,'and thus the way children in the comparisongroup perform in the future stands for the way PDC children would haveperformed without the presumed advantage of PD,C. Whether this assumptionitself stands or falls depends upon the initial equi_valerice of the twogroups. Unless they are very similar to begin with, or can legitimaWybe eqUalized-by statistical means, no sensible comparison can be made.
For each situ and for each variable appearing in Table 3, the assump-tion of PDC-comparison group equality was jested statistically (usingthe chi-square teOnique,for categorical variables and7 t tests for metric
variables). All available data entered into each analysis, meaning that''even It data were missing for a particular child on one or mor'e'veriables,data obtained for that child on other variables did enter in(b theyrespective analyses. A difference was declared to exist between PDC andcomparison ar,nps if analysis indicated the .ehance probability of the.observed difference to he less than one in 100 (p.01). These analyses
show that:
1
At-the individual site level the groups appear siMilar;are differences on background variables in only one site'.
performance measures, of the 13 comparisons made (the Spanishand English sampleS in California and Texas were testedseparately in these analyses), ter; showed either no groupdifferences or differences only on the POCL Or Hei-ght and Weight;
th' only two,sites had group differences on more than one chL1-d
measure-
At the aggregate level/CFsiMilarities of the groups are more
prominent than their 4ifferences. In the -5-oiglish-dominant sample,
there are no significant group differences on the backgroundvariables and only one difference in test performance. In the
Span -ish- dominant samPle, the groups differed on only one back-ground variable, a there was no difference on/any of the per-formance measure$.
ill.Lar Enou PDC and Comstudy?
arisen Sam les- Re ain to- Permit a Longitudinal
In addition to the requirement of comparability,t is important thatthe PDC and comparison testing samples remain large enough as time passesto permit continuing analoe's of their relative performance. Atritionwill inevitably occur, an,dfithe smaller the groups become, th>ore difficultit will be te6eparate PD('s effects from the effects of the many otherfactors that contribute to the performance of ,the children.
The first column of Table 4 shows, for each site and for allSitescollectively, the number of "children that were available for fall 1976testing at PDC and comparion Head Start centers, These children consti-
/ tute the full sample of Cohort 2, the cohore,whose progress
followed through grade 3, On the average, these grqups are about,9smaller ,than the sites had estimated they would be.' Moreover, the meanretention rate - determined this fall for Cohort 1 children (Cohort 2's
-pilot-year predecessors, now in kindergarten) is,lower:.tgan was antici-:-
pated.,s
tThe number of PDC and comparison children who will actually be available
for testing in't,he future i likely to be lower thctin these projections,
for a number of reasons. First, the figures represent children in the
full sample. Consideration of handicap and language factors would r'equ,'rethe elimination of- some children ft arm the analytic sbmple; which is the
source of the data used for statistical analysis. --About 11", of the full-
sample were excluded from trhe, afalyti,sample for reasons of handicap or
language.
1
Taking these factors into cons-t4erAtion, the number of PDC-land com-parfson children from Cohort 2 who are likely to remain 1n ovc Analytidsample of English-dominant children through grade 3 'an be estimated'at about 375 (205 PDC children, 170 comparison children). The numberof children likely to remain in the Spanish- doming sample is about 40(20 PDC, 20 comparison children).
IV
CONCLUSIONS
The findings presented In'Interim Report Vi and reviewed here supportthe ollowing answers to the three questions posed for this phase of thee;ialuation.
Are t e Ieas ring InstruMen,., Appropriate to Ta8k? \\ ,
YES, it can be said with few reservations that all of the instrumentsincluded in the battery satisfy all the criteria that have been used injudging them,(see Table 5). In addition, althOugh the factor structureof the battery does not correspond exactly to the a priori categorizationof the tests (cognitive -'language, social-emotional, psychomotor),' thefactors that emerge are similar to those expected for both the English-dominant and Spanish-dominant groups, and indicate that. the battery doesprovide coverage ofr.thtsefareaT. in addition, the observation systemappears acceptable as a means of assessing the .classroom environment.
Are the PVC and Coat cz2 Z.son Groupe HealPy Comparable?
YES, at the site 1 vel, Cohort 2 PDC and comparison groups seembe more similar than were Cohort '1 groups. The comparability of the
aggregated English- and Spenis-h-dominant 'DC and comparison groups seemsquite
mt
satisfactory,for gnalytik., purposes. In future analyses of testscO're gain, thd-variables oh which; grOups differ initially can beadjusted wi.thout difficulty t(a,make allowances for initial status.
Will Large Enough etc lee Remain for a LongEtudin Study?
YES, by aqpregating PDC and comparison groups across sites, a suf-
ficient sample can be constituted to allow analyses to-continue through
1980-81, when Co'ho'rt 2 will be in grade 3. This, is certainly true for
the English- dominant sample, at least; it is less certainly true for the
spenish-dominpnf Sample. However, eyen'for the latter, analyses could
proceed for a few yearslong enough to allow preliminary conclusions
to be drawn about the effects of PDC. It is obvious from attrition
projections, however, that if the evaluation depended upon4site-levelanalyses of PDC's effects on children, the sample sides available atmost sites would be inadequate by the time Cohort 2 reaches first grade.
Summary
The preliminary phase of the PDe evaluation has achieved its primaryobjective of determining the feasibility of a longitudinal study: suitablemeasures have been selected, adapted, triad out, modified andanalyzed;VcoMparion gro -ups have been located at each site that ara41-Matched tothe PDC groUps; and PDC sites have been successful in recruiting enoughchildren for the PDC and'comparison groups to permit, at least in theaggregates,,a study of PDCJs effectsjon the children's progress throughthird grade.
Table 1'
Composition of the Salriples, Fall 1976
AH E
anW
de
ETHNICITY SEX
mHX
H
0u
o >4i
4=4
q:1
Cr)
dp 1--14
r-E
PDCCALIFORNIA-English
Camp
37
25
0
0
3
4
87
630 11
33
0
0
19
38
81
62
PDCCALIFORNIA - Spanish
7:om-
-7
15
0
0
0
0
100
100 0
0 0
0
43,4049
56
57
60
51
44
,PDCDO
Com p 32
56
57
7
6
7
5
75
69
q
0
18
25
PDCCONNECTICUT
com11 46
84
39
9
2
0
13
7
0
0
59
51
41
49
GEORGIA PDC 46 7 65
i7TT
87
Q 0 35 0 50 50
P DCFLORIDA
'am
4739
0
100
0
0
3
00
6447
3653
PDCIOWA
Com50
54
50
24
0 46
70
r
2
50
52
50
48
PDCMARYLAND'
Com44
58
30
22
46
36 0
39
29
5
10
4852'
52
48
PDCMICHIGAN
Comp6664
26
2U
12
9
0
5
75
'4
0
3
31
75
0
0-0
22
65
25
0
0
0
5845
42
55
PDCTEXAS-English
Comp
46
50
54
50
PDCTEXAS-Spanih
Comp38
37
0
H
0
0
100
97
65
41
'35
60
PDCUTAH
Comp58
61
10
10
3
5
15
18 8
79
0
56
49
44
51
PDCWASHINCTON.
58 15
13 430 14 53
42_
4758
PDCWEST VIRGINIA
Comp46
37 14
0
0
0
0
91
87 0
55i
38.1
44
62
pc'TOTALS HCGROUP
Comp
644
975=9 33
36 25
!2
2
37
36 1 47 53
TOTALS, ALL GROUPS COMBINED ,1211 9 35 25 P 36 50 50
i5
DOMINANTLANGUAGE
U]
cla
U)
y-0
86 11 3
96 4 0
37
24
43 57' 0
0 100 0
100100 0
61 36
93 5
0
0
4
2
7
15
51
30
37
55
100 0 0 43
100 090 10
4539
9810004 04664 17 18
100 0 - 0
100 0 0
0
2
0
4951
31
45
10099 0
97 2
0 0
.58.
5826
19
3834
2
97 '0 4
99 0
10o 0 0
100 0 0
88 11
86 12 2
87 11 2
61
55,
49
66ti
42
29575
520
1095
1 Table 2
Reliability of the Child Mea.SuresaCronbaph's Alpha (Internal Consistency
for Fall 1976,Head Start Children
Cronbachls Alpha
MeasuresEnglish-Dominant
Children;..anish-Dominan_
Children
n r nA..____COGNITIVE-LANGUAGE
, Btlifigual Syntax Measure-English
_HIL
. 997 ...*84 39 .93Bilingual Synax:Measure-Spanishb -16 .95 ,8(2 ,86
:Block Design (WPPSI1 999 .77 94 .82Vprbal.;Fluency (MSCA) 975 92 .87
Verbal Memory-1 (MSCA) 997 . 94 .89Verbal Memory-2 (MSCA) 989 .82 93 .84
Draw-A-Child (MSCA) 978 .84 92 .78
PSYCHOMOTOR
Arm Coordination (MSCA) 976 6'F) 89_.
SOCIAL-EMOTIONALI
POCL-Total High/Scope) 1001- .ds 94 .97
POCL -1 (Hi /Scope) 1001 . 95 94 *96POCL-2 High/Scope) 1001 .0V.--- 94 -.96
aTwo instruments are utt included: the scoring of the Preschool Inter-personal Problem Solving Test does not lend jtself to computing alpha,and the reliability of the classroom observation system was deterwineddifferently.
-Texas and California only (Bilingual/Bicultural Demonstration Sitesr
Table
Comparability of PDC and Comparison Croup's at Each Siteand for Samples Aggregated Across Sites, Fall 1976
(P .01)1
Key: SITE:FA,0IX00
tr,
H140Z
CO
&!rz:
C.9
roi%00
cil
HZ4-),rlicn
TZI
&I (t1r4,-1LI t71
14 1.4
r4 0.0 000-1-)
03
0 ..-1z 73fgl
* -= Statistically significantgroup difference (p(.01)
= no significant differencebetween groups
data insufficient f'oranalysis
= test not-appropri _
1
rW
4
<H
ft,H.-1
<U
C
<H-0[LiH.-1
<U
00
000u
ElnuHE-4
0H.-,
CJ
%
H-
W
"c40i=1 ki.,
3
, 0H ,.
H=U1-i
-
-,-4-4
W_
>WE.
-i
xrtel
0
4-4,
(J)5'
-
H,,-..
0I:4
:..
'''El
(i)
-N for analytic sample
PDc 37 7 51 37 43 4 49 31 58 26 61 49 42 A 4_ 4
_comp. 24 15 30 55 -- 39 51 45 8 19 34 55 66 2 471 49;71
BACKGROUND C1lARACTERISTICS1
EthnicitySexAgePxior Preschool ExperienceNumber of Siblings
-- Mother's Educations
V?
v
,--,
0
C3
.16
0
101
4'
*
TEST4 PERFORMANCE
Cognitive-Language Measures1
.1(
?'
./.
,I
I1
'
-/
*
/ '
-
*'/
.
-
*
-
V
.V
V.
-(/
*
.
/
/''//
,/
-
,*
')/
/VV
.
/
VI
/r
/
/'
/
i
BSM7EngliShBSM- nigh .
Block Design (WPPSI)Verbal Fluency', ,
Verbal Memory-1Verbal Memory-3'DraW-_-Child _Z
; A,
)F',-_-Lychomot_)r Measure
Arm Coordination
Social- Emotional Me suresPTPS-SolutionnPOCL-TotalFOOL -1
pocs,-2
Health- Nutrition Mee4sp s
HeightWeight
,
Table 4
Projected Retention of Cohort 2 Child f; h for EachYear of.the Prospective Longitudinal Study
California
Colorado
Connecticut
Florida
Georgia
Iowa
qr0and
Michigan
Teas
Utah
Washington
West Virginia
PDC
Comp
PDC
Comp
PDC
Comp
PDC
-Comp
PDC
PDC
Comp
PDC
Comp
PDC
Comp
PDC
Comp
PDCComp
PDC
Comp
PDCComp
AGGREGATEPDCComp
1976-77 .1977-78 1978 -79 1979-80 980-8Head Start' 1 2
N
'100 44 31. 59 26
100 40. 42, 17 35 14
100 5'5 49 27 41 23
100 32 42 13 '35 ii
100 50 50 28 42 24
100 57 42 4 35 20
100 26 12
100.47
39 65 -25 55 21
100 46 77 35 65 30
100 50 51 2 6 43 22
100 54 38 21 32' -17
100 :44 73 32 61 29
100 58 77 45 65 38
100 66 86 57 72 . 48100 64 69 44 58. 37
100 64 84 54 71
100 57 76: 43 64 36
100 68 62 42 52 35
100 61* 36 U2 30 IV
100 58 62 .36 52 c
1D0 70 53 59, 45
100 1 25 46 21
.100 37 61 2.3
100 644. 63 408 54 345100 58 334 49 280
NOTE:
3
54 24 .44 1932. 13 26 lq
38 21 31 17
32 10 26 8
39 22, 32 18
32 18 - 2-6 15
24
50
5
20
20
41
49 23
39 20 32 16
29 16 24 13
56 25 46
59 34 49 28
6 44 . 54 36
34 3 25-
6 42 53 44
59 34 48
48 33 39
28 1'7 23 14
48 28 39 23
54 41 44 33
42 29 35 16
56 21 46 i7
49 316 40 258
45 258 36 209
fi
n" represents proportion of original group remaining, "N" representssize of grbup remaining. In the 1976-77 column, N . original samplesize and T, = 100', necessarily. The qgures in successive columns'areprojections based on the actual 1976-77 figures.
Table 5
Summary of Findings on Characteristics'of Measures Included L ithe Fall 1976 B ttery
BSM-English
BSI anish
Blp k 'Design (M.S11\
Verb af 1Fluency
' r i
'Verbal Memory-I'
Verbal Memory-3
Arm'Coordination
Dralk-A-Child
IPS
POC14.
Relevance
IntIrnal Sensitiviti. to Social DevelOprnental Easm.of
Consis:tency Validity to Chalgc Comp etcridea
Span Atnin i 5t rationa
/ / (4
/ /
Acceptable
ProvisiOnally acceptable
Not examined
aDetermined in earlier analys-_ of spring 1A76 date.