Ecobehavioral Assessment in Early Childhood Programs: A Portrait of Preschool Inclusion

Ecobehavioral Assessment in Early Childhood Programs: A Portrait of Preschool Inclusion

William H. Brown

The purpose of this investigation was to describe the nature of preschool children’s experiences in inclusive early childhood programs. The momentary time-sampling data reported in this article represented 3 hours of observational information per child for 112 preschoolers with and without disabilities in 16 community-based, inclusive preschool programs in four states. In general, children with and without disabilities exhibited similar child behaviors and were meaningfully engaged in a variety of adult- and child-initiated activities within similar activity contexts. Two noteworthy between-group differences were that (a) children without disabilities, compared to those with disabilities, participated in more child-child social behaviors and (b) children with disabilities received more adult support and attention than peers without disabilities. These ecobehavioral data begin to “paint a portrait” of preschool inclusion. This “portrait” revealed that children with disabilities were physically included but suggested that if social integration of young children with and without disabilities is a primary goal of inclusion, then additional, focused intervention efforts may be required to establish socially inclusive programs for young children with and without disabilities.

During the last 20 years, early childhood inclusion has been a widely discussed placement strategy for young children with disabilities (e.g., Bailey, McWilliam, Buysse, & Wesley, 1998; Guralnick, 1978; Odom et al., 1996; Peck, Odom, & Bricker, 1993; Wolery et al., 1993). Inclusion of young children with disabilities has had legal (e.g., Individuals with Disabilities Education Act, Public Laws 102-119 and 105-17), scientific (e.g., Buysse & Bailey, 1993; Lamorey & Bricker, 1993; Odom & McEvoy, 1988), and public policy support (e.g., Bredekamp & Copple, 1997; DEC Task Force on Recommended Practices, 1993). Researchers who have conducted comprehensive examinations of the effectiveness of early childhood inclusion have concluded that inclusion has important behavioral and social, but not necessarily intellectual and developmental, benefits for young children with disabilities and their families (Buysse & Bailey, 1993; Lamorey & Bricker, 1993; Odom & McEvoy, 1988; Peck & Cooke, 1983). Many demonstrations of the effectiveness of inclusive early childhood programs have been presented in the professional literature (e.g., Brown, Horn, Heiser, & Odom, 1996; Fox, Dunlap, & Philbrick, 1997; Peck, Killen, & Baumgart, 1989; Rule et al., 1987), and the number of well-specified strategies for supporting inclusive practices has proliferated in both the early childhood and early childhood special education literatures (e.g., Allen & Schwartz, 1996; Brown & Conroy, 1997a; Cook, Tessier, & Klein, 1996; Wolery & Wilbers, 1994). Indeed, given current programmatic evidence and the contemporary zeitgeist for inclusion, New and Mallory (1994) argued that an “ethic of inclusion” (p. 1) has emerged within early childhood education.

Past discussions of and contemporary recommendations for inclusion of young children with disabilities notwithstanding, much remains to be learned about the nature of early childhood inclusion (Kontos, Moore, & Giorgetti, 1998; Odom et al., 1996). Much of the previous research concerning early childhood inclusion has been characterized by Bronfenbrenner’s (1979) indictment of developmental psychology, in which “the science of the strange behavior of children in strange situations with strange adults for the briefest possible periods of time” (p. 19) has predominated. For example, in their review, Buysse and Bailey (1993) identified 22 methodically rigorous investigations of early childhood inclusion. The reviewers noted, however, that most of the studies were conducted in programs that (a) were university-based programs, (b) had low teacher to child ratios, and (c) often enrolled more children with disabilities than children without disabilities. Most often, conditions and services in university-based preschool programs have been in sharp contrast to the conditions and services that typify many community-based preschool programs (e.g., childcare centers, Head Start Programs, not-for-profit and for-profit preschools, and public school preschools). Specifically, most community-based preschool programs have had (a) only a few children with disabilities in each program, (b) higher teacher to child ratios, and (c) fewer specialized and related services (e.g., speech and language therapy, physical therapy; Wolery et al., 1993). These potentially important programmatic differences have limited the ecological validity of much of the extant research on early childhood inclusion, at least with respect to generalizability to inclusion issues in many community-based programs.

Given that annually slightly more than 50% of the preschoolers with disabilities have been served in inclusive preschools (United States Department of Education, 1998) and that the majority of those programs have been community based, explicit descriptions of young children’s experiences in community-based, inclusive preschools are sorely needed to inform our understanding of early childhood inclusion (Kontos et al., 1998; Odom et al., 1996). Although previous research has provided a foundation, conducting a new generation of applied research in typical community contexts may generate important information about specific processes and programmatic outcomes that affect the success or failure of preschool inclusion.

One systematic approach to obtaining specific environmental and behavioral information about young children’s experiences within early childhood programs has been ecobehavioral assessment (Carta, Sainato, & Greenwood, 1988). Ecobehavioral assessment has been defined as “environment-behavior interactions and the ecological contexts of students’ behaviors” (Greenwood, Carta, Kamps, & Arreaga-Mayer, 1990, p. 36). Given that children’s daily interactions within preschool programs may affect their development, examinations of ecobehavioral relationships have broadened our understanding of environmental circumstances that may influence young children’s behavior and subsequent development within early childhood programs (Carta et al., 1988). Ecobehavioral assessment has been employed in several studies in the last 15 years. Investigators have used ecobehavioral assessments to (a) evaluate the effectiveness of early childhood programs (Carta, Greenwood, & Robinson, 1987); (b) investigate children’s transitions between early childhood settings (Carta, Atwater, Schwartz, & Miller, 1990); (c) examine the use of language intervention strategies in preschools (Schwartz, Carta, & Grant, 1996); (d) identify activities that set the occasion for child-child social interaction in early childhood and early childhood special education programs (Odom, Peterson, McConnell, & Ostrosky, 1990); and (e) describe classroom practices and child behavior in early childhood education and early childhood special education settings (Odom, Skellenger, & Ostrosky, 1999).

The ecobehavioral data presented in this article were collected as one facet of a larger ecological systems investigation of preschool inclusion. The overall goal of the ecological systems study was to inform educators’ understanding of preschool inclusion in community-based preschool programs (Odom et al., 1996). The purpose of the ecobehavioral facet of the ecological investigation was to describe with direct observation procedures the nature of children’s experiences in 16 inclusive preschools across the nation. The following specific questions were asked:

1. Do children with and without disabilities participate in different group

arrangements (i.e., number of children within an activity or activity area)

and different peer group compositions (i.e., mixture of children with and

without disabilities) during various activities within inclusive preschool

programs?

2. Do children with and without disabilities participate in different

activities (i.e., activity areas or ongoing activities) within inclusive

preschool programs?

3. Given common preschool activities, are the initiators of those

activities different for children with and without disabilities within

inclusive preschool programs?

4. During common preschool activities, do children with and without

disabilities exhibit different nonsocial and social behaviors within

inclusive preschool programs?

5. During common preschool activities, do teachers exhibit different adult

behaviors within inclusive preschool programs?

Method

Participants and Settings

Participants in the study were 112 children enrolled in 16 inclusive preschools in California, Maryland, Tennessee, and Washington. Purposive sampling procedures were employed to select both the preschool programs within each state and the focal children for observation within each participating program. Preschools were selected to represent a range of (a) program types (e.g., childcare centers, public school classrooms, Head Start programs); (b) inclusive services (e.g., integrated activity model, early childhood special education consultant model, itinerant teaching model, coteaching model with early childhood and early childhood special educators); and (c) geographic diversity (i.e., urban, suburban, and rural sites). Across participating programs, the number of separate preschool classes ranged from a single class to five classes, with the mean number of two classes per participating program. The number of preschoolers per class ranged from 9 to 39, with a mean of 21 children per participating class. The number of adults working in the preschools ranged from 2 to 8 adults, with a mean of 3.5 adults per class. In addition, within each participating preschool program, children were selected for observation based on a range of (a) developmental abilities (i.e., 2 children without disabilities, 3 children with mild to moderate disabilities, 2 children with severe disabilities such as autism, severe mental retardation, and multiple disabilities); (b) socioeconomic status; (c) ethnicity; and (d) language use (e.g., Spanish, Chinese). In some preschools more than 2 of the 5 selected children with disabilities had severe disabilities. Percentages of specific child and family demographic characteristics for the sample of children observed are delineated in Table 1.

TABLE 1. Child and Family Demographic

Characteristics

Child characteristics n %

Age

3-year-olds 30 27

4-year-olds 63 56

5-year-olds 19 17

Gender

Male 65 58

Female 47 42

Ethnicity

Caucasian 61 55

African American 27 24

Latino/Hispanic 9 8

Asian/Pacific Islands 9 8

Native American 1 1

Multiracial 5 5

Disability status

Language impairment 20 18

Developmental delay 13 12

Mental retardation 12 11

Physical impairment 11 10

Autism/PDD 10 9

Emotional, behavioral, or ADHD 6 5

Heating impairment 2 1

Visual impairment 3 2.5

Health impairment 3 2.5

No disability 32 29

Family income

$15,000 or below 38 34

$15,001 to $30,000 17 15

$30,001 to $45,000 14 13

$45,001 to $60,000 18 16

$60,000 or above 15 13

Not reported 10 9

Parental education Mother(a) Father(b)

Less than high school 16 16 8 13

High school graduate 29 29 14 22

Vocational training 3 3 2 3

Some college 15 15 12 19

Associate degree 5 5 2 3

College degree 18 18 18 28

Graduate degree 9 9 8 12

Not reported 4 4 0 0

(a) N= 99. (b) N = 64.

Procedure

Following administrative and parental permission for participation in the study, children were systematically observed in their preschool programs by means of the procedures that follow.

CASPER II Observational System. Research personnel observed children in their inclusive preschool programs with the Code for Active Student Participation and Engagement-Revised (CASPER II) observational system (Brown, Favazza, & Odom, 1995), which is a revision of two previous observational systems, the Code for Active Student Participation and Engagement (CASPER) and the Ecobehavioral System for the Complex Assessment of Preschool Environments (ESCAPE; Carta, Greenwood, & Atwater, 1985). The CASPER II was a direct observational system designed to collect information about preschool environments and behavior of adults and children within those environments. The observational system was a focal child system in which a designated child served as the focus of observations and all decisions about contextual and behavioral coding for a session were made in reference to the observed focal child. The CASPER II consisted of seven ecological variables that provided information about the focal child’s immediate contextual circumstances as well as the focal child’s, peers’, and adults’ behavior within those preschool contexts. Ecobehavioral variables that focused on the focal child’s immediate contextual circumstances included information concerning (a) Group Arrangement, (b) Peer Group Composition, (c) Activity Area or Activity, and (d) Initiator of the Activity. Ecobehavioral variables that cataloged the focal child’s, peers’, and adults’ behavior included information about (a) Child Behavior, (b) Child Social Behavior, and (c) Adult Behavior. The seven ecobehavioral variables with brief definitions and their accompanying environmental and behavioral codes are delineated in Table 2.

TABLE 2. Ecobehavioral Variables with Environmental and Behavioral Codes, Interobserver Agreement, and Kappa Coefficients

Percentage

interobserver

Code agreement

Group arrangements(b)

Overall agreement 95

Solitary (SO) 71

One-to-one with an adult (OO) 77

Small group with 1 or 2 peers (SG) 75

Small group with an adult 71

and 1 or 2 peers (SA)

Large group with 3 or more peers (LG) 66

Large group with an adult 93

and 3 or more peers (LA)

Peer group composition(c)

Overall agreement 96

All children with disabilities (AD) 81

Majority of children 78

with disabilities (MD)

Equal number of children 72

with disabilities (EQ)

Majority of children without disabilities (MN) 88

All children without disabilities (AN) 85

No group (NG) 81

Activity area and activities(d)

Overall agreement 97

Transition (T) 83

Manipulative (M) 89

Large motor (LM) 85

Story time/books (B) 91

Art (A) 86

Pretend play/sociodramatic (P) 89

Large blocks (L) 87

Sensory (S) 86

Dance/music/recitation (D) 86

Snack/meals/food (F) 98

Self-care/self-help (H) 90

Preacademics/3 Rs (R) 86

Computer activities (CP) 90

Circle time/group time (G) 88

Initiator of activity(e)

Overall agreement 98

Adult-initiated activity (AD) 94

Focal child-initiated activity (CH) 92

Typical peer-initiated activity (TP) 14

Peer with disabilities-initiated activity (DP) NA

Child behavior(f)

Overall agreement 93

Books (B) 83

Preacademics/3 Rs (R) 67

Pretend play/sociodramatic play (P) 83

Art (A) 89

Games with rules (GR) 72

Singing/dancing/reciting (D) 76

Self-care/self-help (G) 80

Manipulating (M) 75

Large motor (LM) 72

Cleanup (C) 61

Not engaged (NE) 87

Child social behavior(g)

Overall agreement 94

Social behavior directed to adult (SA) 73

Negative social behavior directed to adult (NA) 88

Social behavior directed to peer (SP) 77

Negative social behavior directed to peer (NP) 69

Peer social behavior directed to focal child (PS) 64

Negative peer social behavior directed 38

to focal child (PN)

No social behavior (NO) 94

Adult behavior(h)

Overall agreement 94

Adult support by teaching or 75

direct assistance (AS)

Adult approval (AA) 62

Adult comment (AC) 55

Group discussion or directions (GD) 80

No adult behavior directed at focal child (NO) 90

Kappa

Code coefficient(a)

Group arrangements(b)

Overall agreement .78

Solitary (SO) .76

One-to-one with an adult (OO) .80

Small group with 1 or 2 peers (SG) .79

Small group with an adult .75

and 1 or 2 peers (SA)

Large group with 3 or more peers (LG) .70

Large group with an adult .79

and 3 or more peers (LA)

Peer group composition(c)

Overall agreement .80

All children with disabilities (AD) .82

Majority of children .78

with disabilities (MD)

Equal number of children .75

with disabilities (EQ)

Majority of children without disabilities (MN) .76

All children without disabilities (AN) .84

No group (NG) .83

Activity area and activities(d)

Overall agreement .92

Transition (T) .91

Manipulative (M) .89

Large motor (LM) .80

Story time/books (B) .93

Art (A) .87

Pretend play/sociodramatic (P) .90

Large blocks (L) .88

Sensory (S) .88

Dance/music/recitation (D) .87

Snack/meals/food (F) .98

Self-care/self-help (H) .92

Preacademics/3 Rs (R) .86

Computer activities (CP) .90

Circle time/group time (G) .90

Initiator of activity(e)

Overall agreement .68

Adult-initiated activity (AD) .70

Focal child-initiated activity (CH) .82

Typical peer-initiated activity (TP) .14

Peer with disabilities-initiated activity (DP)

Child behavior(f)

Overall agreement .85

Books (B) .85

Preacademics/3 Rs (R) .79

Pretend play/sociodramatic play (P) .89

Art (A) .90

Games with rules (GR) .74

Singing/dancing/reciting (D) .78

Self-care/self-help (G) .83

Manipulating (M) .78

Large motor (LM) .74

Cleanup (C) .64

Not engaged (NE) .85

Child social behavior(g)

Overall agreement .79

Social behavior directed to adult (SA) .78

Negative social behavior directed to adult (NA) .89

Social behavior directed to peer (SP) .82

Negative social behavior directed to peer (NP) .70

Peer social behavior directed to focal child (PS) .68

Negative peer social behavior directed .40

to focal child (PN)

No social behavior (NO) .79

Adult behavior(h)

Overall agreement .83

Adult support by teaching or .80

direct assistance (AS)

Adult approval (AA) .65

Adult comment (AC) .59

Group discussion or directions (GD) .81

No adult behavior directed at focal child (NO) .84

(a) point-by-point interobserver agreement scores for the ecobehavioral codes are calculated with both occurrence and nonoccurrence agreements, whereas their accompanying codes are calculated with only occurrence agreements. (b) Group Arrangements are defined by the number of children who are in the same activity area and who are in the vicinity of the focal child (e.g., same table, following one another on tricycles). Group arrangements are defined by both the physical environment (e.g., proximity, presence on equipment) and the focal child’s involvement in a common activity with peers. (c) Peer Group Compositions are defined as the mix of peers with and without disabilities who are found within a group arrangement. To determine the composition of a peer group, the focal child is not included in the group. (d) Activity Area and Activities represent information about the location of the focal child within the physical ecology of an early childhood setting (i.e., activity areas, ongoing activities). In general, activity area codes are determined by where the focal child is within an early childhood setting rather than by what the child is doing (the exception to this rule is the transition and large motor codes). An activity area may be a relatively permanent area (e.g., book area, block area) or a temporary area or activity (e.g., moveable water table, musical instruments). (e) Initiator of Activity is defined by the person (or persons) who selected the activity area where the focal child is located or the activity in which the focal child is involved. Although observers watch the focal child during the 2-second observation interval, they should remember that information about who started the activity may be obtained from adult- or child-initiated behavior before any specific 2-second window of time. (f) Child Behavior is scored when the focal child is attending to, engaging in, or involved in an activity during the 2-second observational interval. If more than one child behavior occurs during the 2-second observation window, code the focal child behavior that is highest in the Child Behavior hierarchy. (g) Child Social Behavior is defined as any socially directed motor/gestural or vocal/verbal behavior that the focal child directs to a person or another child directs to the focal child during a 2-second observational window. (h) Adult Behavior is defined as adult behavior directed toward the focal child during the 2-second observation interval. If more than one adult behavior occurs during the 2-second observation window, code the adult behavior that is highest in the Adult Behavior hierarchy.

Observer Training. Training manuals for observers were sent to data collection coordinators at each of the four sites (i.e., California, Maryland, Tennessee, and Washington). Data collection coordinators from each research site participated in an initial observer training workshop in Nashville, Tennessee. The initial observer training in Nashville and subsequent observer training at each research site were based on a modified version of the direct observation training procedures recommended by Hartmann and Wood (1990). The initial and subsequent observer training procedures included (a) memorization of contextual and behavioral codes for each of the seven ecobehavioral variables as indicated by obtaining 100% accuracy on written tests; and (b) in situ practice employing the CASPER II while observing preschool children in programs not involved in the study. Following initial CASPER II training in Nashville, data collection coordinators at each research site trained additional observers using similar observer training strategies. During observer training at each research site, weekly conference calls were conducted by data collection coordinators to discuss observers’ progress toward learning the coding system and to resolve any problems that arose during observer training. Prior to data collection, observers were trained to a 100% accuracy criterion for observational codes on written tests and to an 85% interobserver agreement criterion for each of the seven ecobehavioral variables. Observer training was conducted three to five times each week for 30 to 60 minutes per session, and training lasted about 8 to 10 weeks at each of the four research sites. After observer training and initiation of data collection at research sites, conference calls were employed periodically to discuss progress toward data collection and to resolve any problems that arose during observations in the field.

Observational Procedures. Observers used a momentary time-sampling procedure for collecting contextual and behavioral information. Specifically, observers watched a designated child for a 2-second period of time and then, during the next 28 seconds, recorded one of the mutually exclusive or hierarchical codes for each of the seven ecobehavioral variables. The 2-second observational intervals and the recording of a code for each ecobehavioral variable were repeated every 28 seconds during each 30-minute session. Data for the CASPER II were collected on handheld computers using a customized software data collection program. Similar momentary time-sampling procedures have been employed effectively in a number of observational studies in preschool settings (e.g., Carta et al., 1987; Odom et al., 1990).

Prior to observations of a focal child, data collectors observed and assessed the focal child’s preschool program and noted relevant environmental circumstances (e.g., where activity centers were located, regular routines, materials used only during designated activities). If questions arose about the nature of specific activities, activity areas, routines, or materials, observers asked the adults in the classroom to clarify the purpose of those activity areas, activities, routines, and materials. In addition, observers initially calibrated their codes for each preschool setting by observing a child who was not a focal child in the study. These preobservation calibration sessions allowed data collectors and children to acclimate to one another and, when indicated, observers to discuss any idio-syncrasies within the focal children’s preschools with other adults or among themselves prior to the actual observations. Teachers were asked to designate which time periods were best for observing children’s active participation in typical preschool activities, and those suggested times were used for observations. Each focal child was observed for six 30-minute observation sessions (i.e., a total of 3 hours of direct observation per child). Observers spread their observations across at least 3 days, with no more than two observational sessions conducted in any given day for a focal child.

Interobserver Agreement Procedure

To assess interobserver agreement, two observers simultaneously but independently collected data on two of the six observational samples for each focal child (i.e., a total of 32% of the entire sample). Interobserver agreement scores were calculated using both the percentage of point-by-point interobserver agreement (i.e., agreements divided by the total number of agreements and disagreements multiplied by 100) and Cohen’s Kappa coefficient. Listed in Table 2 are the point-by-point interobserver agreement scores and Kappa coefficients for each ecobehavioral variable and its accompanying observational categories.

Results

Because the contextual and behavioral codes from the CASPER II were either mutually exclusive or hierarchical in nature, the number of intervals for each ecobehavioral code was quantified as a percentage of intervals observed for each child. Because multiple forms of behavior might be performed during the 2-second observational windows, we established behavioral hierarchies and coded the highest form of behavior in the hierarchy for the three behavioral dimensions. For data analyses, at the individual child level, the percentage of intervals observed for each observational category was aggregated and divided by six observational sessions. This initial procedure yielded a mean for each child observed. Following the computation of individual child means for each observational category, group means were determined by aggregating individual child means and dividing by the number of children within each respective group of children with and without disabilities. This second procedure yielded a mean percentage of observed intervals for each CASPER II code for each group of children (i.e., a mean of the child incorporates data for children with and without disabilities).

Table 3 delineates the mean percentage of observed intervals for each contextual and behavioral code in the CASPER II observation system. In addition, Table 3 presents the F and t values, p values, and [eta.sup.2] effect size estimates that we obtained when systematically comparing children with and without disabilities. When determining statistical significance for the resultant data, we adjusted for possible Type 1 errors related to making multiple contrasts with a Bonferroni procedure, which divided an acceptable alpha level (.05) by the number of comparisons made within each ecobehavioral variable (Kirk, 1995). The reported [eta.sup.2] effect size estimates are a variance-accounted-for statistic that is related to [R.sup.2] (Thompson, 1999). Although the reporting of effect size indices has been explicitly encouraged (American Psychological Association, 1994), standard conventions for interpreting the noteworthiness of various effect size estimates have not been agreed upon and problems have remained in interpreting the practical usefulness of various effect size indices (Rosenthal, 1994). We decided, given that an [eta.sup.2] of. 10 accounts for about 32% of the variability associated with a factor, that it was reasonable to use. 10 as a mathematical cutoff for interpreting the practical significance of any obtained differences between and within groups of children with and without disabilities.

TABLE 3. Mean Percentage of Observed Intervals and Accompanying Statistics

Child

with without

Ecobehavioral variable (Code) disabilities (%) disabilities (%)

Group arrangement codes

Solitary (SO) 10 6

One-to-one w/adult (OO) 8 2

Small group (SG) 7 13

Small group w/adult (SA) 10 7

Large group (LG) 6 9

Large group w/adult (LA) 58 62

Peer group composition codes

All children w/D (AD) 8 4

Majority of children 10 9

w/D (MD)

Equal number of 4 5

children w/D (EQ)

Majority of children 33 43

w/o D (MN)

All children w/o D (AN) 26 31

No group (NG) 17 8

Activity area and

activity codes

Transition (T) 11 9

Manipulative (M) 9 8

Large motor (LM) 12 11

Story time/books (B) 8 6

Art (A) 6 7

Pretend play/ 5 6

sociodramatic (P)

Large blocks (L) 4 3

Sensory (S) 3 3

Dance/music/ 7 6

recitation (D)

Snack/meals/food (F) 11 11

Self-care/self-help (H) 4 3

Preacademics/3 Rs (R) 5 7

Computer activities (CP) 1 2

Circle time/group 15 16

time (G)

Initiator of activity codes

Adult-initiated 57 54

activity (AD)

Focal child-initiated 42 44

activity (CH)

Peer w/o D-initiated 0.29 0.13

activity (TP)

Peer w/D-initiated 0 0

activity (DP)

Adult-initiated vs.

child-initiated for children

with disabilities(c) 57 vs. 42

Adult-initiated vs.

child-initiated for 54 vs. 44

children without

disabilities

Child behavior codes

Books (B) 4 5

Preacademics/3 Rs (R) 3 4

Pretend play/sociodramatic 2 5

play(P)

Art(A) 3 5

Games with rules(GR) 0.33 1

Singing/dancing/ 3 4

reciting(D)

Self-care/self-help(G) 10 10

Manipulating(M) 19 15

Large motor(LM) 8 7

Cleanup(C) 2 2

Not engaged(NE) 45 41

Child engagement 54 58

composite score

Child social behavior codes

Social behavior directed 11.4 6.3

to adult (SA)

Negative social behavior 0.60 0.05

directed to adult (NA)

Social behavior directed 5.30 13.7

to peer (SP)

Negative social behavior 0.45 0.31

directed to peer (NP)

Peer social behavior 2.5 4.4

directed to focal

child (PS)

Negative peer social

behavior directed 0.17 0.10

to focal child (PN)

No social behavior (NO) 78 74

Adult behavior codes

Adult support (teaching/ 15.2 5.2

direct assistance) (AS)

Adult approval (AA) 2.2 0.62

Adult comment (AC) 2.4 1.4

Group discussion/ 20.9 26.1

directions (GD)

No adult behavior 58 65

directed at focal

child (NO)

F or t values

Ecobehavioral variable (Code) and p-values

Group arrangement codes

Solitary (SO) F(1,110) = 6.62, p = .011

One-to-one w/adult (OO) F(1,110) = 16.28, p < .001(*)

Small group (SG) F(1,110) = 13.62, p < .001(*)

Small group w/adult (SA) F(1,110) = 5.60, p = .020

Large group (LG) F(1,110) = 6.41, p = .013

Large group w/adult (LA) F(1,110) = 1.32, p = .254

Bonferroni adjustment p = .008(b)

Peer group composition codes

All children w/D (AD) F(1,110) = 5.06, p = .026

Majority of children F(1,110) = 0.03, p = .856

w/D (MD)

Equal number of F(1,110) = 0.05, p = .818

children w/D (EQ)

Majority of children F(1,110) = 3.03, p = .085

w/o D (MN)

All children w/o D (AN) F(1,110) = 0.62, p = .431

No group (NG) F(1,110) = 16.54, p < .001(*)

Bonferroni adjustment p = .008

Activity area and

activity codes

Transition (T) F(1,110) = 1.55, p = .216

Manipulative (M) F(1,110) = 0.96, p = .329

Large motor (LM) F(1,110) = 0.11, p = .745

Story time/books (B) F(1,110) = 0.84, p = .362

Art (A) F(1,110) = 1.81, p = .182

Pretend play/ F(1,110) = 0.50, p = .480

sociodramatic (P)

Large blocks (L) F(1,110) = 0.05, p = .862

Sensory (S) F(1,110) = 0.09, p = .767

Dance/music/ F(1,110) = 0.47, p = .495

recitation (D)

Snack/meals/food (F) F(1,110) = 0.13, p = .719

Self-care/self-help (H) F(1,110) = 3.80, p = .054

Preacademics/3 Rs (R) F(1,110) = 2.02, p = .158

Computer activities (CP) F(1,110) = 1.69, p = .196

Circle time/group F(1,110) = 0.11, p = .740

time (G) Bonferroni adjustment p = .004

Initiator of activity codes

Adult-initiated F(1,110) = 0.22, p = .637

activity (AD)

Focal child-initiated F(1,110) = 0.20, p = .655

activity (CH)

Peer w/o D-initiated F(1,110) = 0.63, p = .428

activity (TP)

Peer w/D-initiated F(1,110) = 0.40, p = .530

activity (DP)

Adult-initiated vs.

child-initiated for children

with disabilities(c) t (79) = 2.74, p = .008(*)

Adult-initiated vs.

child-initiated for t (31) = 1.02, p = .314

children without Bonferroni adjustment p = .008

disabilities

Child behavior codes

Books (B) F(1,110) = 0.03, p = .851

Preacademics/3 Rs (R) F(1,110) = 2.23, p = .138

Pretend play/sociodramatic F(1,110) = 6.06, p = .015

play(P)

Art(A) F(1,110) = 4.72, p = .032

Games with rules(GR) F(1,110) = 5.28, p = .023

Singing/dancing/ F(1,110) = 2.02, p = .158

reciting(D)

Self-care/self-help(G) F(1,110) = 0.01, p = .931

Manipulating(M) F(1,110) = 3.22, p = .0.76

Large motor(LM) F(1,110) = 0.41, p = .525

Cleanup(C) F(1,110) = 0.02, p = .886

Not engaged(NE) F(1,110) = 1.76, p = .186

Child engagement F(1,110) = 1.94, p = .167

composite score Bonferroni adjustment p = .004

Child social behavior codes

Social behavior directed F(1,110) = 8.95, p = .003(*)

to adult (SA)

Negative social behavior F(1,110) = 6.03, p = .016

directed to adult (NA)

Social behavior directed F(1,110) = 56.84, p < .001(*)

to peer (SP)

Negative social behavior F(1,110) = 0.88, p = .351

directed to peer (NP)

Peer social behavior F(1,110) = 16.29, p < .001(*)

directed to focal

child (PS)

Negative peer social

behavior directed F(1,110) = 2.07, p = .153

to focal child (PN)

No social behavior (NO) F(1,110) = 3.43, p = .067

Bonferroni adjustment p = .007

Adult behavior codes

Adult support (teaching/ F(1,110) = 26.63, p < .001(*)

direct assistance) (AS)

Adult approval (AA) F(1,110) = 6.08, p = .015

Adult comment (AC) F(1,110) = 3.51, p = .064

Group discussion/ F(1,110) = 3.12, p = .080

directions (GD)

No adult behavior F(1,110) = 6.42, p = .013

directed at focal Bonferroni adjustment p = .01

child (NO)

Effect sizes(a)

Ecobehavioral variable (Code) [Eta.sup.2]

Group arrangement codes

Solitary (SO) .06

One-to-one w/adult (OO) .13

Small group (SG) .11

Small group w/adult (SA) .05

Large group (LG) .06

Large group w/adult (LA) .01

Peer group composition codes

All children w/D (AD) .04

Majority of children .00

w/D (MD)

Equal number of .00

children w/D (EQ)

Majority of children .03

w/o D (MN)

All children w/o D (AN) .01

No group (NG) .13

Activity area and

activity codes

Transition (T) .01

Manipulative (M) .01

Large motor (LM) .00

Story time/books (B) .01

Art (A) .02

Pretend play/ .00

sociodramatic (P)

Large blocks (L) .00

Sensory (S) .00

Dance/music/ .00

recitation (D)

Snack/meals/food (F) .00

Self-care/self-help (H) .03

Preacademics/3 Rs (R) .02

Computer activities (CP) .02

Circle time/group .00

time (G)

Initiator of activity codes

Adult-initiated .00

activity (AD)

Focal child-initiated .00

activity (CH)

Peer w/o D-initiated .01

activity (TP)

Peer w/D-initiated .00

activity (DP)

Adult-initiated vs.

child-initiated for children

with disabilities(c) .09

Adult-initiated vs.

child-initiated for .03

children without

disabilities

Child behavior codes

Books (B) .00

Preacademics/3 Rs (R) .02

Pretend play/sociodramatic .05

play(P)

Art(A) .04

Games with rules(GR) .05

Singing/dancing/ .02

reciting(D)

Self-care/self-help(G) .00

Manipulating(M) .03

Large motor(LM) .00

Cleanup(C) .00

Not engaged(NE) .02

Child engagement .02

composite score

Child social behavior codes

Social behavior directed .08

to adult (SA)

Negative social behavior .05

directed to adult (NA)

Social behavior directed .34

to peer (SP)

Negative social behavior .01

directed to peer (NP)

Peer social behavior .13

directed to focal

child (PS)

Negative peer social

behavior directed .02

to focal child (PN)

No social behavior (NO) .03

Adult behavior codes

Adult support (teaching/ .19

direct assistance) (AS)

Adult approval (AA) .05

Adult comment (AC) .03

Group discussion/ .03

directions (GD)

No adult behavior .06

directed at focal

child (NO)

(a) percentages within ecobehavioral variables may not add up to 100% because of rounding. (b) A Bonferroni adjustment for Type 1 errors was employed with an alpha level of .05 divided by the number of comparisons made for each ecobehavioral variable. (c) This analysis represents a within group for adult- versus child-initiated activities.

(**) p < .05. D = disability.

Group Arrangement Codes

As presented in Table 3, both children with and without disabilities spent a majority of time in Large Group with Adult arrangements with three or more peers (i.e., 58% vs. 62% of intervals). Children with disabilities were more likely than children without disabilities to be involved in One-to-One with an Adult arrangements. Children without disabilities, however, were more likely than children with disabilities to be involved in Small Group arrangements with one or two peers. The effect sizes for One-to-One with Adult arrangements and Small Group arrangements accounted for .13 and .11 of the variance, respectively, for those two group arrangement codes.

Peer Group Composition Codes

As shown in Table 3, children with and without disabilities spent approximately 17% and 8% of their time, respectively, in activities without an immediate peer group, and children with disabilities were more likely to be involved in solitary activities. The effect size for the No Group Peer Group Composition code accounted for. 13 of the variance for that social grouping. Children with and without disabilities spent 59% and 74% of their time, respectively, in two peer group compositions, Majority of Children without Disabilities and All Children without Disabilities.

To assess the extent of peer group integration of children with and without disabilities during various preschool activities, we developed a Peer Group Integration Index for children with disabilities by aggregating the percentage of intervals observed from the Majority of Children with Disabilities, Equal Number of Children without Disabilities, Majority of Children without Disabilities, and All Children without Disabilities codes. Similarly, for children without disabilities an index was produced by totaling the percentage of intervals observed from the Equal Number of Children with Disabilities, Majority of Children without Disabilities, Majority of Children with Disabilities, and All Children with Disabilities codes. The aggregation of these codes showed how often children with disabilities were involved in social groups with one or more peers without disabilities as members of their immediate peer group. Likewise, the combination of the codes for children without disabilities indicated how frequently they were involved in groups with at least one peer with disabilities as a member of their immediate social group. For children with disabilities, the Peer Group Integration Index indicated that they were in integrated peer groups during 73% of the observed intervals. Children without disabilities participated in integrated peer groups during 61% of the observed intervals. Although children with and without disabilities spent the majority of their time involved in integrated social groups, as indicated by the Peer Group Integration Index, children with disabilities were more likely to be involved in integrated peer groups than children without disabilities, F (1,110) = 12.08, p [is less than] .001. The effect size for the comparison of children with and without disabilities with the Peer Group Integration Index accounted for .10 of the variance for integrated social groupings. In spite of frequent peer group integration, children with disabilities were in segregated peer groups on 31% of the observed intervals, whereas children with disabilities participated in segregated peer groups on 8% of the intervals. These data are shown in Figure 1.

[Figure 1 ILLUSTRATION OMITTED]

Activity Area and Activity Codes

As presented in Table 3, children with and without disabilities spent their time in a variety of different preschool activity contexts and activities. Of the 14 different activity contexts and activities children with and without disabilities were observed in, none of the activities yielded statistically significant differences between groups. The most frequently observed activities for both groups were circle time, large motor, snack/food, and transition activities. The least frequently observed activities for both groups were computer, sensory, self-care, and large block activities.

Initiator of Activity Codes

As shown in Table 3, children with and without disabilities spent a majority of their time in activities that were initiated by either adults (i.e., 57% vs. 54% of intervals) or the focal child (i.e., 42% vs. 44% of intervals) and very little time in activities initiated by peers with and without disabilities (i.e., less than 1% for both groups in both categories of peer initiation). None of the four Initiator of Activities codes yielded statistically significant between-group differences. Because we were interested in both between- and within-group comparisons of adult-initiated and child-initiated activities, we conducted two-tailed t tests for independent and paired samples and compared adult-initiated and focal child-initiated codes. Children with disabilities were more likely to be involved in adult-initiated activities than in child-initiated activities (i.e., 57% vs. 42% of intervals). For children with disabilities, the effect size for the obtained difference in adult and child-initiated activities accounted for .09 of the variance for those types of activities. Statistically significant within-group differences in adult-initiated and child-initiated activities were not evident for children without disabilities (i.e., 54% vs. 44% of intervals).

Child Behavior Codes

As presented in Table 3, children with and without disabilities exhibited a number of different child behaviors during a variety of preschool activities. Children with and without disabilities engaged in very similar behaviors in their inclusive preschool programs, with no statistically significant differences occurring between groups. The most often observed child behaviors were manipulating, self-care/self-help, and large motor behaviors, whereas the least frequently observed child behaviors were games with rules and cleanup behaviors.

Another strategy for assessing young children’s behavior has been to collect a global measure of their active and meaningful participation in multiple preschool activities (i.e., child engagement during common program activities; e.g., McWilliam & Bailey, 1995; McWilliam, Trivette, & Dunst, 1985). For comparison of children with and without disabilities, we aggregated the percentage of observed intervals for 10 of the Child Behavior codes to create a composite score for Child Engagement (i.e., all nonsocial behavior codes except Not Engaged). This procedure indicated that children with and without disabilities were meaningfully engaged in preschool activities 54% and 58% of the observed intervals, respectively. The between-group difference in Child Engagement was not statistically significant.

Child Social Behavior Codes

As shown in Table 3, children with and without disabilities had several statistically significant differences in their social behavior within inclusive preschool programs. Specifically, children with disabilities were more likely to socially interact with adults than were children without disabilities. Nevertheless, the effect size for social behavior directed to adults accounted for only .08 of the variance in that behavioral code. In addition, children without disabilities directed social behavior to peers much more often than did children with disabilities, and they were more likely to receive social initiations from peers. The effect sizes for these two social behavior codes accounted for .34 and .13 of the variance, respectively. Negative social behavior directed to peers and negative social behavior directed to the focal child were observed in less than 1% of the observed intervals for all children. When the positive social behavior directed to peers and positive peer social behavior directed to the focal child codes were aggregated, children without disabilities were interacting with other children on 18.1% of the intervals observed, whereas children with disabilities were interacting with peers on 7.8% of the intervals observed.

Adult Behavior Codes

As presented in Table 3, adult behavior toward children with and without disabilities differed on only one dimension. Children with disabilities were three times more likely than children without disabilities to receive the highest level of teacher assistance, Adult Support (i.e., 15.2% vs. 5.2% of intervals). The effect size for the Adult Support code accounted for. 14 of the variance for that behavioral code. At much lower proportions, children with disabilities also received more Adult Approval (i.e., 2.2% vs. 0.62% intervals) and Adult Comments (i.e., 2.4% vs. 1.4% intervals) than children without disabilities. When all forms of individual adult attention to the focal child (i.e., Adult Support, Adult Approval, and Adult Comments) were aggregated, children with disabilities received nearly three times as much adult attention as children without disabilities (i.e., 19.8% vs. 7.2% intervals). Group Discussion/Directions was the most frequent adult behavior for children with and without disabilities (i.e., 20.9% vs. 26.1% intervals), although the between-group difference was not statistically significant.

Discussion

We assessed the contextual and behavioral circumstances of 112 young children with and without disabilities in 16 inclusive preschool programs in four states. This initial descriptive analysis has allowed us to begin to “paint a portrait” of early childhood inclusion in the 16 participating preschool programs. Our general conclusions are (a) activities and activity contexts for children with and without disabilities were very similar; (b) children with disabilities were meaningfully engaged in a variety of preschool activities as often as their peers without disabilities; (c) children with and without disabilities participated in a mixture of adult- and child-initiated activities; (d) children with and without disabilities engaged in similar types of nonsocial behavior; (e) children without disabilities participated in child-child socially directed behaviors much more frequently than peers with disabilities; and (f) children with disabilities received much more adult support and attention than peers without disabilities.

On the one hand, the overall ecobehavioral analysis has some “good news” for proponents of preschool inclusion. On the other hand, some results clearly remind us of the known developmental needs of many young children with disabilities. Moreover, the contextual and behavioral differences we found illuminate programmatic areas that ought to be addressed if we want to establish optimal inclusive preschool settings for promoting and supporting the development of young children with disabilities (Bricker, 1995). For the purpose of discussion, the ecological data we collected can be subdivided into contextual variables and behavioral variables. Both contextual and behavioral variables have potential for informing our understanding of children’s preschool experiences in inclusive settings.

Contextual Variables

With respect to group arrangements, we found differences between children with and without disabilities on two of six group arrangements. Children with disabilities were more likely to be involved with adults in one-to-one activities than were children without disabilities. At least part of this finding may be attributed to the fact that periodically children with disabilities were observed during therapy sessions that were integrated into their preschool schedule. Moreover, adult support was more frequent for children with disabilities than their peers without disabilities. Children without disabilities were involved in small groups without adults more often than were peers with disabilities. The different proportion of intervals children without disabilities spent in small group activities with one or two peers without adults present might have been an indication of their peer-related social competence relative to peers with disabilities (Guralnick, 1992; Odom & Brown, 1993; Odom, McConnell, & McEvoy, 1992). These two differences in group arrangements differed from the results of another observational study of inclusive preschools in which Kontos et al. (1998) found that children without disabilities were more likely to be alone with teachers in inclusive preschools (i.e., slightly over one third of the time). Kontos and her colleagues acknowledged that their teacher involvement finding was “puzzling” (p. 45), and they speculated that the adults might have been directing children with disabilities to activities without direct adult supervision and that children without disabilities may be more effective in evoking and maintaining adults’ attention than children with disabilities.

The measures of peer group composition clearly indicated that children with and without disabilities were involved in developmentally integrated social groups in the majority of the intervals observed. Specifically, children with disabilities were in social groups with at least one or more children without disabilities in their immediate peer group during 73% of the intervals observed. Similarly, children without disabilities participated in social groups with at least one or more children with disabilities in their peer group during 61% of the intervals observed. In addition, children without disabilities were more likely to be in social groups with only peers without disabilities (i.e., about one third of the intervals), whereas children with disabilities were more likely to be alone during preschool activities. For children without disabilities, their social groupings with peers without disabilities might have been related to their relatively sophisticated peer-related social competence. Likewise, for children with disabilities, their solitary behavior might have been associated with their relatively less sophisticated peer-related social competence.

Our observations also showed that children with and without disabilities spent similar amounts of time within a variety of common preschool activities. With respect to who initiated most preschool activities, adults initiated activities for both children with and without disabilities a majority of the time. Focal child-initiated activities were also frequent for both groups, however. For children with and without disabilities, between-group comparisons did not show systematic differences in adult-initiated (i.e., 57% vs. 54% of intervals) or child-initiated activities (i.e., 42% vs. 44% of intervals). Within-group analyses, however, indicated that children with disabilities were more likely to be involved in teacher-initiated than child-initiated activities (i.e., 57% vs. 42% of intervals). In an investigation of early childhood and early childhood special education preschools, Odom et al. (1999) determined that teacher-directed activities occurred more frequently in early childhood special education settings than in the early childhood programs. However, teachers were also the most frequent initiator of activities in early childhood preschools. In an observational study of child engagement in inclusive preschools, McCormick, Noonan, and Heck (1998) found that child-directed activities were more common for children with and without disabilities (i.e., between 50% and 60% of the time). Similarly to those of Odom et al. (1999), our results indicated more prevalent adult-directed activities. Nevertheless, the findings from these two studies were not radically different from those of McCormick et al. Previous philosophical concerns about adult-initiated activities notwithstanding (e.g., Bredekamp, 1987), it appears that activities in many preschools have continued to be initiated by teachers. Although early childhood professionals have not explicitly recommended nor established criteria for appropriate proportions of child-initiated and adult-initiated activities within preschool settings, the pattern of initiator of activities in our observations and others’ (e.g., McCormick et al.; Odom et al.) appeared to be reasonable given contemporary philosophical shifts toward–or at least explicit statements of–a “balance between children’s self-initiated learning and adult guidance and support” (Bredekamp & Copple, 1997, p. 17).

We believe that, overall, the observational data for contextual variables indicated that children with disabilities were included physically within common preschool activities and contexts. As Odom and McEvoy (1988) argued, however, placement of children has represented only one dimension of preschool inclusion, and another critical facet, social integration, has been essential for effectively including young children with disabilities (see Bricker, 1995; Brown & Conroy, 1997b; Odom et al., 1996 for similar arguments). Whereas physical inclusion was reflected by the contextual variables collected during the CASPER II observations, children’s social integration was best examined with its behavioral variables.

Behavioral Variables

Our observations indicated that children with and without disabilities exhibited similar nonsocial behaviors during a variety of preschool activities. This general finding for children’s behavior differed from the results of two recent observational studies. For example, Kontos et al. (1998) found that children with disabilities engaged more frequently in art, books, and manipulative behaviors than peers without disabilities. In addition, children without disabilities participated more often in computer/science and dramatic activities than peers with disabilities. In their comparison of early childhood and early childhood special education settings, Odom et al. (1999) reported that children in early childhood special education programs were more likely to participate in preacademic activities, whereas children in early childhood settings were more likely to be involved in play activities. However, in contrast to our comparisons within inclusive preschools, Odom et al. compared one group of children with disabilities in their early childhood special education settings and another group of children without disabilities, who were “matched” on gender and age, in their early childhood preschools.

On another of our behavioral measures, the overall level of child engagement was not statistically different for children with and without disabilities. If children without disabilities can be used as a standard for determining a “normative” level of child engagement within inclusive preschools and if engagement is viewed as a primary indicator of children’s access to important learning opportunities in preschool settings (McWilliam et al., 1985), then these data suggest that children with and without disabilities have available similar learning opportunities. Other researchers have reported that young children with and without disabilities in inclusive preschools exhibited similar levels of child engagement (e.g., Kontos et al., 1998; McCormick et al., 1998; Odom, Favazza, Brown, & Horn, in press). In a more precise analysis of child engagement, however, McWilliam and Bailey (1995) found that young children with disabilities were engaged for less time and at less sophisticated levels of engagement than children without disabilities. However, even with similar levels of child engagement for children with and without disabilities, children with disabilities may not be making progress on their Individualized Education Programs (IEPs) and individualized interventions may be warranted. An important future step for researchers will be to carefully examine the relationships between levels of child engagement and children’s developmental progress within inclusive preschools (Carta et al., 1988).

At first glance, the levels of child engagement for children with and without disabilities that we obtained (i.e., 54% and 58% of observed intervals) might be a concern for those who expect young children to be engaged most of the time. The levels of child engagement found with our observations may be attributable to several factors. First, we observed across multiple preschool activities and within many preschool settings that young children spend a considerable amount of time in transitions (e.g., about 10% of observed intervals in our sample). Second, the definition of child engagement has varied across studies and a common definition has not been forthcoming. For example, the definition we used in this observational study did not include children’s passively attending or listening to adults or peers, except when adults read books or were singing, dancing, and reciting with children. Also, given that children spent the majority of their time in large groups with adults in our study (i.e., 58% and 62% of intervals), passive attending and listening might have been frequent, and not recording those forms of behavior might have contributed to lower proportions of child engagement than reported in studies in which researchers coded attentional engagement (e.g., McCormick et al., 1998). Using a conceptually broader definition of child engagement than the CASPER II behavioral codes, which included children’s attentional engagement, would have probably produced higher proportions of engagement.

As mentioned, from our initial analysis of the contextual data, it appeared that on the dimension of physical inclusion, children with disabilities were relatively well integrated into common preschool activities and at least had access to learning opportunities that were similar to those of their peers without disabilities. On the important measure of social integration, however, our findings indicated that children with disabilities exhibited fewer social initiations to and received fewer social bids from peers than did children without disabilities. Indeed, two of the most noteworthy findings of the study were that children without disabilities directed much more social behavior to peers when compared to children with disabilities (i.e., 13.7% vs. 5.3% of intervals, [eta.sup.2] = .34), and peers directed more social behavior to children without disabilities than to children with disabilities (i.e., 4.4% vs. 2.5% of intervals, [eta.sup.2] =. 13). These two results within community-based inclusive preschools replicated a well-documented finding in the social interaction literature for young children with and without disabilities in analogue and university-based settings (e.g., Guralnick & Groom, 1988; Guralnick, Conner, Hammond, Gottman, & Kinnish, 1995; Jenkins, Odom, & Speltz, 1989).

With respect to adult behavior within inclusive preschools, our findings showed that children with disabilities had much more frequent adult support (i.e., 15.2% vs. 5.2% of intervals, [eta.sup.2] =. 19) and attention than children without disabilities. The presence of adults, as well as frequent adult support and other forms of attention for children with disabilities, was reasonable and was not unexpected. Nor was it necessarily a negative outcome. Indeed, increased adult support and attention may be one important indication that the individual needs of children with disabilities were being addressed within the context of their inclusive preschools. Recently, other investigators have noted the importance of adult support for promoting learning opportunities for young children with disabilities in inclusive settings (e.g., McCormick et al., 1998; Sontag, 1997), and it may be that the ultimate success or failure of inclusive preschools will be tied to the feasibility of individualized programs and additional teacher support for many children with disabilities for at least some part of their daily schedule. An important finding of the present study was that individualized adult support did not represent the most frequent adult behavior with children with and without disabilities. Both children with and without disabilities were in large groups with adults for the majority of the time (i.e., 58% and 62% of intervals), and the most frequent adult behavior was group discussion and directions (i.e., 20.9% and 26.1% of intervals).

Whereas inclusive settings may be a supportive context for effective intervention strategies (e.g., peer-mediated interventions, observational learning), they are not necessarily sufficient to address the behavioral and developmental needs of all children with disabilities. Given the developmental requirements of many young children with disabilities, we, along with others, have argued that explicit and well-targeted intervention strategies might be most beneficial for improving children’s behavior and development in critical areas such as peer-related social competence (Brown & Conroy, 1997b; Guralnick, 1992; McEvoy, Odom, & McConnell, 1992; Odom & Brown, 1993) and language (McLean & Cripe, 1997; Notari & Cole, 1993; Schwartz et al., 1996). For example, given physical inclusion with peers without disabilities in common preschool activities and the limited peer interactions exhibited by many young children with disabilities within those integrated contexts (e.g., Guralnick et al., 1995; Guralnick & Groom, 1988), improved social inclusion may require activity-based intervention strategies that promote children’s social interaction (Bricker, Pretti-Frontczak, & McComas, 1998), particularly those strategies that focus on peer interactions such as friendship activities (e.g., Brown, Ragland, & Fox, 1988; McEvoy et al., 1988), buddy skills-training programs (e.g., English, Goldstein, Shafer, & Kaczmarek, 1996; Goldstein, English, Shafer, & Kaczmarek, 1997), incidental teaching (e.g., Brown, McEvoy, & Bishop, 1991; McGee, Almeida, Sulzer-Azaroff, & Feldman, 1992), and group coaching interventions (e.g., Mize, 1995; Odom & McConnell, 1993). Indeed, from our perspective, the dearth of peer-initiated activities for both peers with and without disabilities (i.e., less than 1% of intervals) clearly indicates that peer-initiated interventions may be an extremely important and apparently “untapped” strategy for promoting young children’s social integration within inclusive preschools (Brown & Odom, 1994; McEvoy et al., 1992).

Limitations of Existing Research and Future Research

Our ecobehavioral investigation has at least three limitations. First, although the 112 children observed were purposely selected to be a relatively diverse sample of children and families in 16 programs from across the nation, we do not know how representative the children were of the general population of preschool children who are served in inclusive preschools. Second, the CASPER II observational system has yielded a reliable albeit relatively global assessment of contextual and behavioral dimensions of preschool programs. Given the momentary time-sampling procedures used and the present ecobehavioral assessment codes, we do not know the precise quantitative and qualitative nature of some of the observational categories, particularly the behavioral codes. For example, although we measured adult support (e.g., teaching interactions, verbal and physical guidance), we do not know if those adult-child interactions actually represented recommended or effective teaching practices (Odom & McLean, 1996). Similarly, for peer interactions, we do not know the precise nature of the child-child social interactions and whether or not they represented high-quality, sustained peer interactions or relatively brief social exchanges (Brown, Odom, & Holcombe, 1996). Finally, because the present ecological investigation was a descriptive study and not a process-product investigation, we were not able to link the contextual and behavioral variables to well-specified behavioral or developmental outcomes for children (Carta et al., 1988; Kontos et al., 1998).

Our perusal of the observational codes employed in contemporary investigations of inclusive preschools clearly indicated that operational definitions differ markedly and that those differences have presented challenges for interpreting results across studies. For example, McCormick et al. (1998) included attentional engagement in their observations, whereas we only recorded children’s attention to teachers when adults read, recited, sang, and danced with children. As another example, Kontos et al. (1998) defined teacher involvement using the Howes Involvement Scale (Howes & Stewart, 1987). This approach defined teacher behavior very differently from the adult behavior codes of the CASPER II and ESCAPE observational systems. Definitional differences were most notable on the observational codes for recording the social groups of children. For example, unlike our differentiation of group arrangements into small-group contexts with one or two peers and large-group contexts with more than two peers as well as identification of the developmental status of peers within groups and the presence or absence of adults, Kontos et al. did not differentiate between the relative size of the children’s immediate peer group nor the developmental status of the peers within those groups. Even when investigators coded small and large social groups, definitions of those groupings were radically different. For example, McCormick et al. defined small groups as those with fewer than six children (versus our definition of one or two peers) and large groups as those with more than six peers (versus our definition of three or more children).

Similarly to our observational procedures, other investigators have varied their observations across preschool activities throughout the day (e.g., McCormick et al., 1998; Sontag, 1997). Given their interests and research questions, however, Kontos et al. (1998) restricted their observations to freeplay circumstances. Even when investigators have procedurally varied their observations across preschool activities, this generic strategy has not necessarily precluded collection of markedly different observational samples from dissimilar circumstances and contexts.

Finally, different observational strategies might have made studies difficult to interpret. To date, researchers have typically employed time sampling versus real time continuous observational strategies, and their sampling techniques have varied greatly. For example, we used a momentary time-sampling procedure with 2 seconds to observe and 28 seconds to simultaneously record the seven ecobehavioral variables during six 30-minute observational periods across 3 days. Sontag (1997) also employed a momentary time-sampling technique with 15 seconds to watch and simultaneously code only three of a larger set of observational codes. Within each minute of observation, Sontag was able to code one 15-second sample of her variables of interest. The number of Sontag’s observation sessions ranged from three to seven sessions, which were about 40 minutes in duration. Kontos et al. (1998) used a scan-sampling technique that emphasized collecting a number of behavior samples per participant versus coding within well-specified and timed observation sessions. Using this procedure they collected 2-second observational samples for focal children with disabilities and 2 comparison children without disabilities. Specifically, they observed the children’s behavior, activity settings, social contexts, and their teachers’ behavior for 2 seconds and then recorded the observed categories for 15 seconds. Observations were continued until a sufficient number of behavioral episodes were collected. McCormick et al. (1998) employed a 10-second period to observe followed by 10 seconds to record procedure during two 5-minute observational periods. Even between the two observational systems that employed relatively similar momentary time-sampling techniques (i.e., CASPER and ESCAPE), there existed differences in how (e.g., all seven variables coded simultaneously vs. several variables coded consecutively across time) and when (e.g., 15-second observe and record versus 2-second watch and 28-second record) researchers collected information. Differences in investigators’ sampling strategies may have accounted for some of the differential results in the present database for preschool inclusion. In summary, the differences in operational definitions, observational circumstances, and observational strategies previously noted have made, and may continue to make, the emerging database on the nature of children’s inclusive preschool experience difficult to interpret.

The extant ecobehavioral assessment information and future investigations of inclusive preschools should allow us to acquire a better understanding of the contemporary nature of inclusion for young children with and without disabilities. In future analyses of our existing ecobehavioral database, we plan to perform sequential analyses to determine the precise relationships among specific contextual and behavioral variables (e.g., Odom et al., 1990; Sontag, 1997; see Bakeman & Quera, 1995, for a methodology). For example, which contextual variables (e.g., group arrangement, peer group composition) promote children’s opportunities to interact with one another and their social behavior with peers? “Fine-grained” analyses of the present ecobehavioral database may provide important information concerning which specific ecobehavioral variables (or combinations of variables) might set the occasion for children’s behavioral and developmental progress. Once important environment-behavior relationships are clearly identified within inclusive preschools, investigators might manipulate those robust variables to determine their practical significance for children’s behavioral and developmental progress (Brown, Odom, & Holcombe, 1996; Carta et al., 1988). These types of process-product investigations should inform researchers’ and practitioners’ understanding of preschool inclusion. Moreover, a new generation of ecobehavioral information and future investigations may provide us with the much-needed information to enhance the quality of young children’s early childhood services within inclusive preschools (Bricker, 1995; Bricker, Peck, & Odom, 1993; Kontos et al., 1998; Odom et al., 1996).

AUTHORS’ NOTES

(1.) This research and preparation of the manuscript were supported by Grant No. HO24K60001-99 (Early Childhood Research Institute on Inclusion) from the Office of Special Education Programs, U.S. Department of Education. The content and opinions expressed herein do not necessarily reflect the position or policy of the U.S. Department of Education, and no official endorsement should be inferred.

(2.) We thank Robin McWilliam, an anonymous reviewer, and Lynn Fuchs for their helpful comments and recommendations on previous drafts of the manuscript.

REFERENCES

Allen, K. E., & Schwartz, I. S. (1996). The exceptional child: Inclusion in early childhood education (3rd ed.). Albany, NY: Delmar.

American Psychological Association. (1994). Publication manual of the American Psychological Association (4th ed.). Washington, DC: Author.

Bailey, D. M., McWilliam, R. A., Buysse, V., & Wesley, P. W. (1998). Inclusion in the context of competing values in early childhood education. Early Childhood Research Quarterly, 13, 27-47.

Bakeman, R., & Quera, V. (1995). Analyzing interaction: Sequential analysis with SDIS and GSEQ. New York: Cambridge University Press.

Bredekamp, S. (1987). Developmentally appropriate practice in early childhood programs serving children from birth through age 8. Washington, DC: National Association for the Education of Young Children.

Bredekamp, S., & Copple, C. (Eds.). (1997). Developmentally appropriate practice in early childhood programs serving children from birth through age 8. Washington, DC: National Association for the Education of Young Children.

Bricker, D. (1995). The challenge of inclusion. Journal of Early Intervention, 19, 179-194.

Bricker, D. D., Peck, C. A., & Odom, S. L. (1993). Integration: Campaign for the new century. In C. A. Peck, S. L. Odom, & D. Bricker (Eds.), Integrating young children with disabilities into community programs: Ecological perspectives on research and implementation (pp. 271-276). Baltimore: Brookes.

Bricker, D., Pretti-Frontczak, K., & McComas, N. (1998). An activity-based approach to early intervention (2nd ed.). Baltimore: Brookes.

Bronfenbrenner, U. (1979). The ecology of human development: Experiments by nature and design. Cambridge, MA: Harvard University Press.

Brown, W. H., & Conroy, M. (Eds.). (1997a). Inclusion of preschool children with developmental delays in early childhood programs. Little Rock, AR: Southern Early Childhood Association.

Brown, W. H., & Conroy, M. (1997b). Promoting and supporting peer interactions in inclusive preschools: Effective strategies for early childhood educators. In W. H. Brown & M. Conroy (Eds.), Inclusion of preschool children with developmental delays in early childhood programs (pp. 79-108). Little Rock, AR: Southern Early Childhood Association.

Brown, W. H., Favazza, P. C., & Odom, S. L. (1995). Code for Active Student Participation and Engagement-Revised (CASPER II): A training manual for observers (Unpublished observer training manual). Nashville, TN: Vanderbilt University, Early Childhood Research Institute on Inclusion.

Brown, W. H., Horn, E. M., Heiser, J. G., & Odom, S. L. (1996). Project BLEND: An inclusive model for early intervention services. Journal of Early Intervention, 20, 364-375.

Brown, W. H., McEvoy, M. A., & Bishop, J. N. (1991). Incidental teaching of social behavior: A naturalistic approach to promoting young children’s peer interactions. Teaching Exceptional Children, 24, 35-58.

Brown, W. H., & Odom, S. L. (1994). Strategies and tactics for promoting generalization and maintenance of young children’s social behavior. Research in Developmental Disabilities, 15, 99-118.

Brown, W. H., Odom, S. L., & Holcombe, A. (1996). Observational assessment of young children’s social behavior with peers. Early Childhood Research Quarterly, 11, 19-40.

Brown, W. H., Ragland, E. U., & Fox, J. J. (1988). Effects of group socialization procedures on the social interactions of preschool children. Research in Developmental Disabilities, 9, 359-376.

Buysse, V., & Bailey, D. B. (1993). Behavioral and developmental outcomes in young children with disabilities in integrated and segregated settings: A review of comparative studies. The Journal of Special Education, 26, 434-461.

Carta, J. J., Atwater, J. B., Schwartz, I. S., & Miller, P. A. (1990). Applications of ecobehavioral analysis to study transitions across early childhood. Education and Treatment of Children, 13, 298-315.

Carta, J. J., Greenwood, C. R., & Atwater, J. B. (1985). ESCAPE: Ecobehavioral system for complex assessments of preschool environments. Kansas City: University of Kansas, Juniper Gardens Children’s Project, Bureau of Child Research.

Carta, J. J., Greenwood, C. R., & Robinson, S. L. (1987). Application of an ecobehavioral approach to the evaluation of early intervention programs. In R. Prinz (Ed.), Advances in behavioral assessment of children and families (Vol. 3, pp. 123-155). Greenwich, CT: JAI Press.

Carta, J. J., Sainato, D. M., & Greenwood, C. R. (1988). Advances in the ecological assessment of classroom instruction for young children with handicaps. In S. L. Odom & M. B. Karnes (Eds.), Early intervention for infants and children with handicaps (pp. 217-239). Baltimore: Brookes.

Cook, R. E., Tessier, A., & Klein, M. D. (1996). Adapting early childhood curricula for children in inclusive settings (4th ed.). Englewood Cliffs, NJ: Merrill.

DEC Task Force on Recommended Practices. (1993). DEC recommended practices: Indicators of quality in programs for infants and young children with special needs and their families. Reston, VA: Council for Exceptional Children.

English, K., Goldstein, H., Shafer, K., & Kaczmarek, L. (1996). Promoting interactions among preschoolers with and without disabilities: Effects of a buddy system skills-training program. Exceptional Children, 63, 229-243.

Fox, L., Dunlap, G., & Philbrick, L. A. (1997). Providing individual supports to young children with autism and their families. Journal of Early Intervention, 21, 1-14.

Goldstein, H., English, K., Shafer, K., & Kaczmarek, L. (1997). Interaction among preschoolers with and without disabilities: Effects of across-the-day peer intervention. Journal of Speech and Hearing Research, 40, 33-48.

Greenwood, C. R., Carta, J. J., Kamps, D., & Arreaga-Mayer, C. (1990). Ecobehavioral analysis of classroom instruction. In S. Schroeder (Ed.), Ecobehavioral analysis and developmental disabilities: The twenty-first century (pp. 33-63). New York: Springer Verlag.

Guralnick, M. J. (1978). Early intervention and the integration of handicapped and nonhandicapped children. Baltimore: University Park Press.

Guralnick, M. J. (1992). A hierarchical model for understanding children’s peer-related social competence. In S. L. Odom, S. R. McConnell, & M. A. McEvoy (Eds.), Social competence of young children with disabilities: Issues and strategies for intervention (pp. 37–64). Baltimore: Brookes.

Guralnick, M. J., Conner, R. T., Hammond, M., Gottman, J. M., & Kinnish, K. (1995). Immediate effects of mainstreamed settings on the social interactions and social integration of preschool children. American Journal on Mental Retardation, 100, 359-377.

Guralnick, M. J., & Groom, J. M. (1988). Peer interactions in mainstreamed and specialized classrooms: A comparative analysis. Exceptional Children, 54, 415-425.

Hartmann, D. P., & Wood, D. D. (1990). Observational methods. In A. S. Bellack, M. Hersen, & A. E. Kazdin (Eds.), International handbook of behavior modification and therapy (pp. 107-138). New York: Plenum.

Howes, C., & Stewart, P. (1987). Children’s play with adults, toys, and peers: An examination of family and child-care influences. Developmental Psychology, 23, 423-430.

Individuals with Disabilities Education Act Amendments of 1991, 20 U.S.C. [sections] 1400-1485.

Individuals with Disabilities Education Act Amendments of 1997, Pub. L. No. 105-17.

Jenkins, J. R., Odom, S. L., & Speltz, M. L. (1989). Effects of social integration on preschool children with handicaps. Exceptional Children, 55, 420-428.

Kirk, R. E. (1995). Experimental design: Procedures for the behavioral sciences (3rd ed.) Pacific Grove, CA: Brooks/Cole.

Kontos, S., Moore, D., & Giorgetti, K. (1998). The ecology of inclusion. Topics in Early Childhood Special Education, 18, 38-48.

Lamorey, S., & Bricker, D. D. (1993). Integrated programs: Effects on young children and their parents. In C. A. Peck, S. L. Odom, & D. Bricker (Eds.), Integrating young children with disabilities into community-based programs: Ecological perspectives on research and implementation (pp. 249-269). Baltimore: Brookes.

McCormick, L., Noonan, M. J., & Heck, R. (1998). Variables affecting engagement in inclusive preschool classrooms. Journal of Early Intervention, 21, 160-176.

McEvoy, M. A., Nordquist, V. M., Twardosz, S., Heckaman, K., Wehby, J. H., & Denny, R. K. (1988). Promoting autistic children’s peer interaction in an integrated early childhood setting using affection activities. Journal of Applied Behavior Analysis, 21, 193-200.

McEvoy, M. A., Odom, S. L., & McConnell, S. R. (1992). Peer social competence intervention for young children with disabilities. In S. L. Odom, S. R. McConnell, & M. A. McEvoy (Eds.), Social competence of young children with disabilities: Issues and strategies for intervention (pp. 113-134). Baltimore: Brookes.

McGee, G. G., Almeida, C., Sulzer-Azaroff, B., & Feldman, R. S. (1992). Promoting reciprocal interactions via peer incidental teaching. Journal of Applied Behavior Analysis, 25, 117-126.

McLean, L. K., & Cripe, J. W. (1997). The effectiveness of early intervention for children with communication disorders. In M. J. Guralnick (Ed.), The effectiveness of early intervention (pp. 349-428). Baltimore: Brookes.

McWilliam, R. A., & Bailey, D. B. (1995). Effects of classroom social structure and disability on engagement. Topics in Early Childhood Special Education, 15, 123-147.

McWilliam, R. A., Trivette, C. M., & Dunst, C. J. (1985). Behavior engagement as a measure of efficacy of early intervention. Analysis and Intervention in Developmental Disabilities, 5, 59-71.

Mize, J. (1995). Coaching preschool children in social skills: A cognitive-social learning curriculum. In G. Cartledge & J. F. Milburn (Eds.), Teaching social skills to children and youth: Innovative approaches (3rd ed.; pp. 237-261). Boston: Allyn & Bacon.

New, R. S., & Mallory, B. L. (1994). Introduction: The ethic of inclusion. In B. L. Mallory & R. S. New (Eds.), Diversity & developmentally appropriate practices: Challenges for early childhood education (pp. 1-13). New York: Teachers College Press.

Notari, A., & Cole, K. (1993). Language intervention: Research implications for service delivery. In C. A. Peck, S. L. Odom, & D. Bricker (Eds.), Integrating young children with disabilities into community programs: Ecological perspectives on research and implementation (pp. 17-37). Baltimore: Brookes.

Odom, S. L., & Brown, W. H. (1993). Social interaction skills interventions for young children with disabilities in integrated settings. In C. A. Peck, S. L. Odom, & D. Bricker (Eds.), Integrating young children with disabilities into community programs: Ecological perspectives on research and implementation (pp. 39-64). Baltimore: Brookes.

Odom, S. L., Favazza, P., Brown, W. H., & Horn, E. M. (in press). Ecobehavioral assessment of early childhood programs for children with disabilities. In T. Thompson, D. Felce, & F. Symons (Eds.), Behavioral observation: Innovations in technology and applications in developmental disabilities. Baltimore: Brookes.

Odom, S. L., & McConnell, S. R. (1993). Play time/social time: Organizing your classroom to build interaction skills. Tucson, AZ: Communication Skill Builders.

Odom, S. L., McConnell, S. R., & McEvoy, M. A. (1992). Peer-related social competence and its significance for young children with disabilities. In S. L. Odom, S. R. McConnell, & M. A. McEvoy (Eds.), Social competence of young children with disabilities: Issues and strategies for intervention (pp. 3-36). Baltimore: Brookes.

Odom, S. L., & McEvoy, M. A. (1988). Integration of young children with handicaps and normally developing children. In S. L. Odom & M. B. Karnes (Eds.), Early intervention for infants and children with handicaps: An empirical base (pp. 241-267). Baltimore: Brookes.

Odom, S. L., & McLean, M. E. (Eds.). (1996). Early intervention/early childhood special education: Recommended practice. Austin, TX: PRO-ED.

Odom, S. L., Peck, C. A., Hanson, M., Beckman, P., Kaiser, A., Lieber, J., Brown, W. H., Horn, E., & Schwartz, I. S. (1996). Inclusion at the preschool level: An ecological analysis. Social Policy Report: Society for Research in Child Development, 10(2, 3), 18-30.

Odom, S. L., Peterson, C., McConnell, S. R., & Ostrosky, M. (1990). Ecobehavioral analysis of early education/specialized classroom settings and peer social interaction. Education and Treatment of Children, 13, 316-330.

Odom, S. L., Skellenger, A., & Ostrosky, M. (1999). Procedural variations in early childhood education and early childhood special education classes. Manuscript submitted for publication.

Peck, C. A., & Cooke, T. P. (1983). Benefits of mainstreaming at the early childhood level: How much can we expect? Analysis and Intervention in Developmental Disabilities, 3, 1-22.

Peck, C. A., Killen, C. C., & Baumgart, D. (1989). Increasing implementation of special education instruction in mainstream preschools: Direct and generalized effects of nondirective consultation. Journal of Applied Behavior Analysis, 22, 197-210.

Peck, C. A., Odom, S. L., & Bricker, D. (Eds.). (1993). Integrating young children with disabilities into community programs: Ecological perspectives on research and implementation. Baltimore: Brookes.

Rosenthal, R. (1994). Parametric measures of effect size. In H. Cooper & L. V. Hedges (Eds.), The handbook of research synthesis (pp. 231-244). New York: Russell Sage Foundation.

Rule, S., Stowitschek, J. J., Innocenti, M., Striefel, S., Killoran, J., Swezey, K., & Boswell, C. (1987). The social integration program: An analysis of the effects of mainstreaming handicapped children into day care centers. Education and Treatment of Children, 10, 175-192.

Schwartz, I. S., Carta, J. J., & Grant, S. (1996). Examining the use of recommended intervention practices in early childhood special education. Topics in Early Childhood Special Education, 16, 251-272.

Sontag, J. C. (1997). Contextual factors influencing the sociability of preschool children with disabilities in integrated and segregated classrooms. Exceptional Children, 63, 389-404.

Thompson, B. (1999). Improving research clarity and usefulness with effect size indices as supplements to statistical significance tests. Exceptional Children, 65, 329-337.

United States Department of Education. (1998). To assure the free and appropriate public education of all children with disabilities: Twentieth Annual Report to Congress on the implementation of the Individuals with Disabilities Education Act. Washington, DC: Author.

Wolery, M., Holcombe, A., Venn, M. L., Brookfield, J., Huffman, K., Schroeder, C., Martin, C. G., & Fleming, L. A. (1993). Mainstreaming in early childhood programs: Current status and relevant issues. Young Children, 49, 78-84.

Wolery, M., & Wilbers, J. S. (Eds.). (1994). Including children with special needs in early childhood programs. Washington DC: National Association for the Education of Young Children.

William H. Brown, University of South Carolina

Samuel L. Odom, Indiana University

Shouming Li, University of Maryland

Craig Zercher, San Francisco State University

Address: William H. Brown, Department of Educational Psychology, College of Education, University of South Carolina, Columbia, SC 29208.3

COPYRIGHT 1999 Pro-Ed

COPYRIGHT 2000 Gale Group