Follow this and additional works at:

Transcription

1 The University of Toledo The University of Toledo Digital Repository Theses and Dissertations 2013 The effects of two schedules of instruction with constant time delay on the receptive word learning skills of preschool children with developmental delays Margie A. Spino The University of Toledo Follow this and additional works at: Recommended Citation Spino, Margie A., "The effects of two schedules of instruction with constant time delay on the receptive word learning skills of preschool children with developmental delays" (2013). Theses and Dissertations This Dissertation is brought to you for free and open access by The University of Toledo Digital Repository. It has been accepted for inclusion in Theses and Dissertations by an authorized administrator of The University of Toledo Digital Repository. For more information, please see the repository's About page.

2 A Dissertation entitled The Effects of Two Schedules of Instruction with Constant Time Delay on the Receptive Word Learning Skills of Preschool Children with Developmental Delays by Margie A. Spino Submitted to the Graduate Faculty as partial fulfillment of the requirements for the Doctor of Philosophy Degree in in Curriculum and Instruction: Special Education Dr. William F. McInerney, Committee Chair Dr. Laurie A. Dinnebeil, Committee Member Dr. Edward J. Cancio, Committee Member Dr. Mark Wolery, Committee Member Dr. Patricia R. Komuniecki, Dean College of Graduate Studies The University of Toledo May, 2013

3 Copyright 2013, Margie A. Spino This document is copyrighted material. Under copyright law, no parts of this document may be reproduced without the expressed permission of the author.

4 An Abstract of The Effects of Two Schedules of Instruction with Constant Time Delay on the Receptive Word Learning Skills of Preschool Children with Developmental Delays by Margie A. Spino Submitted to the Graduate Faculty as partial fulfillment of the requirements for the Doctor of Philosophy Degree in Curriculum and Instruction: Special Education The University of Toledo May, 2013 Young children with special needs in inclusive settings may be provided special education services by itinerant early childhood special education (ECSE) teachers. These teachers typically travel to the community-based preschool program and work with a child for about 1 hour, 1 day each week. Research suggests that once itinerant ECSE teachers leave the classroom, children with disabilities are rarely provided specialized instruction. Additionally, the Division for Early Childhood (DEC) of the Council for Exceptional Children (CEC), the primary ESCE professional organization in the U.S., recommends that instruction be embedded within everyday routines and that it be distributed across the day and throughout the week. Despite these recommendations, the practice of providing specialized services directly to a child, during one visit per week, continues. This study examined if a once-a-week instructional schedule as used by ECSE teachers providing direct services was sufficient to promote learning in young children with disabilities, especially those with cognitive disabilities. The extent to which instruction should be distributed in order to promote learning in young children with disabilities was addressed by comparing two different intervention intensities (i.e., iii

5 schedules) with the constant time delay strategy. Four preschool children with identified disabilities participated along with their 3 preschool teachers. The ECSE teachers taught children to identify items by pointing to them using the constant time delay strategy with a 1x/week schedule and a 3x/week schedule. An adapted alternating treatments design was used to evaluate the effects of the two schedules. The results indicated that both the 1x/week and 3x/week schedules were effective to teach two of the four children both target words and to teach one child one target word. Examination of the three teachers procedural fidelity data revealed that three s with specific corrective feedback were sufficient to yield PF scores of 100%. While teachers were nearly perfect in their implementation of the CTD procedures, they were less successful in implementing the procedure with the intended dose frequency. iv

6 Acknowledgements When one tugs at a single thing in nature, he finds it attached to the rest of the world. -- John Muir My parents instilled in me a love of learning and a sense of wonder about the world. I am thankful to them for those gifts and for their unwavering love and support. Thank you to my sisters and their families who always encouraged and believed in me and to my in-laws whose warm care has comforted me. In particular I wish to thank my husband and step-daughter, Rod and Kelsey, for always listening, for taking care of me, for never complaining, for being patient. I remember the day that Dr. William McInerney, my committee chair, suggested that I pursue a doctorate degree. I am eternally grateful not only for that initial suggestion but for all the guidance and support he provided. I thank my other committee members, Dr. Laurie Dinnebeil, Dr. Edward Cancio, and Dr. Mark Wolery, for giving so freely of their time. Their advice and feedback kept me on course. In addition, I am thankful for the friendship and support of Dr. Lyn Hale and Dr. Laurence Coleman as well as the contributions of Gwen Weber and Christina Yeager for coding and Dr. Andrea Glesser for assisting with the child assessment. I am also deeply appreciative of the mentorship of Dr. Vladimir Sloutsky during my time at The Ohio State University. I thank the administrators of the Monroe County Intermediate School District and the families of the children for allowing me to work with them. Finally, I can never thank enough the three teachers who gave so generously of their time and efforts, Elyse Bryant, Renee Retli, and Michele Springsteen. I am truly indebted to them for their commitment and attention to detail. v

7 Table of Contents Abstract Acknowledgements Table of Contents List of Tables List of Figures iii v vi x xii I. Introduction 1 A. Itinerant Service Delivery 2 B. Evidence-based Practices 5 C. Treatment Fidelity 6 D. Intervention Intensity 7 E. Summary 9 F. Objectives of the Study 11 II. Literature Review 13 A. The Spacing Effect 13 a. General characteristics of the spacing effect 13 b. Research with adult participants 15 c. Research with child participants 16 d. Research with children with disabilities 20 e. Every-day versus every-other-day instruction 24 f. Summary 25 B. Differential Treatment Intensity 27 a. Definition and components 27 vi

8 b. Recommendations to researchers 30 C. Constant Time Delay 31 a. Effectiveness and fidelity 34 b. Treatment intensity and research reviews 36 c. Treatment intensity and early childhood studies 37 d. Summary 45 III. Methodology 47 A. Participants 47 a. Teacher participants 47 b. Child participants 48 c. Child assessment 51 B. Setting 54 C. Materials 54 D. Target Stimuli 56 E. Response Definitions and Measurement Procedures 57 F. Experimental Design 60 G. Threats to Internal Validity 61 H. Procedures 65 a. Teaching training 65 b. Screening 66 c. Initial probe 67 d. Wait training 68 e. Comparison 71 vii

9 f. Final probe 76 g. Generalization 76 h. Maintenance 77 I. Social Validity 77 J. Interobserver Agreement and Procedural Fidelity 77 K. Data Analysis Procedures 79 L. Determination of a Functional Relationship 84 IV. Results 86 A. Visual Analysis 86 a. Brenda 86 b. Cara 89 c. Fiona 92 d. Jacob 94 e. Median level change 98 f. Percentage non-overlapping data 98 B. Descriptive Statistics 99 a. Effectiveness 99 b. Efficiency 100 c. Generalization 101 d. Maintenance 102 e. Fidelity of procedure and intensity 103 C. Social Validity 107 V. Discussion 112 viii

10 A. General Findings 112 B. Limitations 118 C. Implications for Practice and Future Research 120 References 125 Appendices A. Reinforcer Survey 138 B. Lesson Plans for Teacher Training 139 C. Data Collection Forms for Each Condition 144 D. IOA and PF Data Collection Forms 149 E. Social Validity Survey 156 ix

11 List of Tables Table 1 Definitions and examples of treatment intensity components Table 2 Example of treatment intensity components for a low-frequency and highfrequency intervention Table 3 Example of treatment intensity components for a study holding cumulative intervention intensity constant Table 4 Treatment intensity components of early childhood special education studies using CTD Table 5 Estimates of number of trials each week of ECSE studies using CTD Table 6 Learning criterions for the ECSE studies using CTD Table 7 DAS-II subtests administered and abilities measured Table 8 Descriptive classification labels for composite and subtest scores for the DAS-II Table 9 Child participant composite and subtest scores on the DAS-II Table 10 Target items used with each child Table 11 Research questions and measures for the study Table 12 Sequence of experimental sessions and number of trials Table 13 Procedure for wait training Table 14 Sample schedule for comparison condition Table 15 Procedure for 0-second delay trials in the comparison condition Table 16 Procedure of 3-second delay trials in the comparison condition Table 17 Percentage of IOA for each child participant by condition Table 18 Percentage of PF for planned teacher behaviors for probe sessions x

12 Table 19 Percentage of PF for planned teacher behaviors for 1x/week sessions Table 20 Percentage of PF for planned teacher behaviors for 3x/week sessions Table 21 Median level change between initial probe and comparison conditions by child and schedule Table 22 PND between initial probe and comparison conditions by child and schedule Table 23 Average correct on probes for each child by schedule Table 24 Efficiency data for each child who met criterion Table 25 Average PF for teachers before and after last day of corrective feedback Table 26 Percent of 1x/week periods that correct dose frequency (12 trials/week was provided Table 27 Percent of 3x/week periods that correct dose frequency (1 session/day, 3 days/week, 12 trials/week) was provided Table 28 Average number of trials/week for each child by schedule Table 29 Percentage of four teachers responses on the Social Validity Survey to the overall use of the CTD strategy before and after the study Table 30 Percentage of four teachers responses on the Social Validity Survey to the use of the CTD strategy with a 1x/week and 3x/week schedule Table 31 Actual and intended dose frequency for each schedule Table 32 Treatment intensity components of ECSE studies that used CTD, including the current study xi

13 List of Figures Figure 1 Subtests comprising the Special Nonverbal Composite of the DAS-II Figure 2 Percentage of unprompted correct responses for Brenda during initial probe, comparison, final probe, generalization, and maintenance conditions Figure 3 Percentage of unprompted correct responses for Cara during initial probe, comparison, final probe, generalization, and maintenance conditions Figure 4 Percentage of unprompted correct responses for Fiona during initial probe, comparison, final probe, generalization, and maintenance conditions Figure 5 Percentage of unprompted correct responses for Jacob during initial probe, comparison, final probe, generalization, and maintenance conditions Figure 6 Change in final probe to maintenance probe scores for each child by schedule xii

14 Chapter One Introduction Before the passage of landmark federal legislation (Education of All Handicapped Children Act, 1975) many young children with disabilities were not eligible for special services. Those children who received services were usually taught in separate classrooms and were segregated from their peers without disabilities (DeMonte, 2010). While parents and professionals were initially happy that these young children were finally receiving educational services, they recognized that the separate settings limited children s abilities to learn and to become accepted as members of the community. Throughout the 1960s and 1970s parent and professional groups fought for equal access to education for children with disabilities. Persistent pressure in the public arena and in the courts ultimately resulted in changes to federal and state policies. While Congress in 1975 mandated that all children with disabilities had the right to a free and appropriate public education (P.L ), it was not until 1986 that Congress amended the earlier act and extended the same protections to preschoolers (3-5 years of age) while laying the foundation for early intervention with young children birth to 3 years of age (P.L ). However, states were not required to make the necessary changes to serve young children with disabilities until the school year (DeMonte, 2010). Since the passage of P.L , an increasing number of young children with special needs receive special education services in inclusive settings with typical peers; both the number of young children receiving services and the percentage of children receiving services in early childhood settings has increased. In 1995 IDEA annual data indicated there were 548,588 young children from 3-5 years of age receiving services 1

15 (U.S. Department of Education, 2010). By 2009 that number had increased by 33% to 731,250 young children (U.S. Department of Education, 2009). With regard to more inclusive settings, in 2006 approximately 44% of the 692,017 young children with disabilities receiving services were educated in regular early childhood programs for more than 80% of their weekly instructional time (U.S. Department of Education, 2006). Within three years almost 50% of the 731,250 preschool-age children with disabilities were in regular early childhood programs for more than 80% of the time (U.S. Department of Education, 2009). As young children with special needs move from separate special education preschool classrooms to more inclusive settings, there is a need for special education teachers to also move out of their classrooms in order to visit these young children and provide services. Itinerant Service Delivery An itinerant model of early childhood service delivery is one that is used to support the inclusion of children with disabilities within the least restrictive environment often within community-based programs (Dinnebeil, Pretti-Frontczak, & McInerney, 2009, p. 436). Itinerant early childhood special education (ECSE) teachers travel to communitybased early childhood classrooms to provide specialized services to eligible children (Dinnebeil & McInerney, 2000; Dinnebeil et al., 2009). In Ohio, for example, a typical pattern of service delivery involves the itinerant early childhood special education teacher visiting inclusive settings once a week for about one hour (Dinnebeil, McInerney, Roth, & Ramaswamy, 2001). The term, itinerant, only denotes the location of the special education service but not the nature of those services (Dinnebeil, McInerney, & Hale, 2

16 2006b; Dinnebeil et al., 2009). So while it is understood that an itinerant ECSE teacher will provide services in a setting other than a self-contained special education classroom, (e.g., Head Start classroom, childcare center, child s home), the term itself does not automatically specify how the services will be provided. With respect to itinerant services, Odom and his colleagues (1999) specified two major types of itinerant services: direct and consultative. With the direct services approach the itinerant ECSE teacher periodically visits the classroom to provide specialized services directly to the child with special needs. Markowitz and colleagues (2006) indicated, however, that once the itinerant ECSE teacher left the classroom, the early childhood teacher rarely provided instruction that addressed the special needs of the child with disabilities. If this once-a-week instruction is the only instruction children receive related to the skills and knowledge targeted in their IEP, then too much forgetting may occur between teaching sessions. The Division for Early Childhood (DEC), recommends that instruction be embedded within everyday routines and that it be distributed across the day and throughout the week (Wolery, 2005). Despite these recommendations, the practice of providing specialized services directly to the child, during one visit per week, continues in some school districts. Another model for providing services is the consultative approach (Odom et al., 1999) in which the itinerant ECSE teacher primarily spends her weekly visit working with and coaching her early childhood teacher partner. Although the itinerant teacher in this model may interact with and work with children, her or his primary responsibility is to help the child s primary teacher or caregiver find ways to provide IEP-based instruction in her absence (Dinnebeil et al., 2006b, p. 153). In the consultative approach, 3

17 the early childhood teacher learns to address the child s specialized needs, as identified in the Individualized Education Plan (IEP), and embeds and distributes instruction multiple times a week. Buysse and Wesley (2005) contended that a primary reason for adopting the consultative approach is that itinerant ECSE teachers could assist more students more efficiently than through the direct services delivery model. The consultative approach holds the promise of producing positive change, not only in the client (e.g., the child) but also in the consultee (e.g., the teacher), the program (e.g., the curriculum or learning environment), and potentially, the entire system (Buysse & Wesley, 2005, p. 5). The field of early childhood special education regards models of indirect service delivery (i.e., the consultative approach) as best practice for inclusive education (Horn & Sandall, 2000; Lieber et al., 1997; McWilliam, Wolery, & Odom, 2001). While the consultative approach is the preferred model of instruction recommended by leaders in the field of early childhood special education, it does not appear to be the preferred model for itinerant ECSE teachers. Survey and observational studies found that the overwhelming majority of itinerant ECSE teachers self-reported or were observed providing direct services to children rather than consulting with ECE teachers (Dinnebeil, et al., 2001; Dinnebeil, McInerney, & Hale, 2004; Dinnebeil, McInerney, & Hale, 2006a). Using a direct services approach would be warranted if research found it to be more effective or efficient at facilitating learning. However, to date, there have not been any studies conducted that have directly compared the effect of direct services versus the consultative approach on a child s attainment of priority IEP objectives. 4

18 Evidence-based Practices There is ample evidence to suggest a growing gap between what we know we should do and what we are doing in early childhood intervention (Bruder, 2010, p. 345). Within the field of education there has been a recent emphasis on adopting evidence-based practices. In 2001, the No Child Left Behind (NCLB) Act required schools to utilize scientifically-based research practices for the classroom, while in 2004, the Individuals with Disabilities Education Improvement Act (IDEA) required that special education instruction be based on peer-reviewed research. IDEA identified an insufficient focus on applying replicable research on proven methods of teaching and learning for students with disabilities (IDEA, Section 14(c)(4)) as a major barrier to improving outcomes for children with disabilities. Additionally, the mission for the Division for Early Childhood (DEC) is to promote policies and advance evidence-based practices that support families and enhance the optimal development of young children who have or are at risk for developmental delays and disabilities (from As such DEC has based its recommended practices for the field on two key sources of information. The first source is scientific research on effective practices. The second is the knowledge and experience of individuals who work with young children and partner with their families. Odom and Wolery (2003) echoed this view when they stated that the emphasis on grounding practice in supportive evidence is a critical direction in EI/ECSE (p. 170). Despite these federal mandates and calls for the use of evidence-based practices, teachers are not necessarily utilizing such practices in their classrooms. The dearth of 5

19 research-supported methods being consistently applied to support learning is recognized within special education as a whole (Carnine, 1997; Greenwood, 2001; Kauffman, 1996) and within early childhood special education (Bruder, 2010; McLean, Snyder, Smith, & Sandall, 2002; Odom, 2009) as the research-to-practice gap. Reasons often cited for the gap include inaccessible research findings, contradictory research findings, and complex implementation procedures (Carnine, 1997; Gersten, Morvant, & Brengelman, 1995; Greenwood & Abbott, 2001; Jones, 2009). Researchers found that implementing evidence-based practices in special education classrooms were often challenging and difficult (Abbott, Walton, Tapia, & Greenwood, 1999; Ayres, Meyer, Erevelles, & Park- Lee, 1994; Gersten, et al. 1995). Teachers who are not adequately trained or who find the implementation process too complex may use the strategy with low levels of fidelity or may abandon the strategy altogether (Fitzpatrick & Knowlton, 2009). Treatment Fidelity Treatment integrity (sometimes called treatment fidelity or procedural reliability) refers to the degree to which a treatment is implemented as planned or intended. [It] is concerned with the accuracy and consistency with which independent variables constituting the treatment (as opposed to subject characteristics) are implemented. (Gresham, MacMillan, Beebe-Frankenberger, & Bocian, 2000, p. 198) There are many strategies available to ECSE teachers that are based on research such as those recommended by DEC as child-focused practices (e.g., differential reinforcement, correspondence training, high-probability procedures) (Sandall, Hemmeter, Smith, & McLean, 2005). However, besides choosing an evidence-based 6

20 strategy, implementing the strategy with fidelity is equally important. [I]n order to realize the promise offered by research on intervention and instructional efficacy, teachers must implement practices in their classroom in a way similar to that intended by the purveyors (Odom, 2009, p. 54). Fidelity is facilitated when intervention practices are simple and straightforward and, therefore, easier to learn for early childhood practitioners and parents. In a review of implementation fidelity research, Carroll et al. (2007) found that simple but specific interventions are more likely to be implemented with high fidelity than overly complex or vague ones (p. 5). The degree to which an intervention is implemented accurately and consistently has been found to affect student outcomes (Benner, Beaudoin, Chen, Davis, Ralston, 2010; Holcombe, Wolery, & Snyder, 1994; Mihalic, S., 2004; Stein et al., 2008) with higher levels of fidelity associated with more positive outcomes. Interventions and strategies that are considered evidence-based are not only efficacious but also efficient. Efficacy, or effectiveness, considers whether a student learned, while efficiency considers if one intervention allowed more rapid learning or allowed more content or skills to be learned in the same amount of time. The efficacy of an EBP assumes implementation in the manner and level of intensity used in the efficacy studies (Odom, Collet-Klingenberg, Rogers, & Hattan, 2010, p. 279). As such, the level of intensity or schedule used to implement the strategy may be an important variable that might impact effectiveness and efficiency (Venn, Wolery, & Greco, 1996; Warren, Fey, & Yoder, 2007). Intervention Intensity 7

21 [H]igh doses of teaching episodes per se may not necessarily be the most effective approach. More learning may result by increasing the dose frequency (i.e., number of sessions per week) while applying lower doses across longer durations for some types of goals, with some types of children, and with some dose forms. (Warren et al., 2007, p. 71) Intervention intensity (or treatment intensity) has usually been defined as the duration of a treatment (e.g., Casto & Mastropieri, 1986; Innocenti & White, 1993), such as the number of hours per week. While Warren et al. (2007) acknowledged that defining intensity in this way allows the concept to be more easily understood by practitioners, parents, and policy makers, this simple definition fails to capture the full complexity of the construct. For example, researchers might report that an intervention was implemented for 3 sessions a week but might not report the number of trials or teaching opportunities within each session. An intervention that utilized 8 trials per session for a total of 24 trials per week might result in vastly different outcomes than an intervention that only used 1 trial per session for a total of 3 trials per week. To provide consistency across research studies and aid in the development of optimally efficacious interventions, Warren et al. (2007) proposed that intervention intensity be more precisely defined as five separate terms: dose, dose form, dose frequency, total intervention duration, and cumulative intervention intensity. Of particular interest to this study are dose, dose frequency, and total intervention duration. Dose is the number of properly administrated teaching episodes during a single intervention session (Warren et al., 2007, p. 71), and dose frequency is the number of times a dose of intervention is provided per day and per week (p. 72). The total intervention duration 8

22 is the period of time that an intervention is provided. For example, a play-based intervention may involve 60 teaching sessions per hour (dose) for 1 session each day for 5 days a week (dose frequency) to be provided for 40 weeks (total intervention duration). Over 30 years ago, Meyen and Lehr (as cited in Faggella-Luby & Deshler, 2008) identified intensity as a necessary factor to consider when designing effective instruction for students with disabilities. They challenged educational researchers to measure and study intensity of instruction. Warren et al. (2007) identified four types of knowledge that could result from the study of different levels of intensity: 1. A treatment that is effective at a given level of intensity may be more or less effective at a different level of intensity. 2. Changing any one of the aspects of intensity may produce different effects on children with different developmental profiles. 3. Some intensity levels may produce side effects such as stress and challenging behaviors. 4. Cumulative intervention intensity plays an important role and when comparing different interventions, it should be controlled. In fact, Warren et al., (2007) recommended that Researchers should commence the systematic study of differential treatment intensities for interventions that have shown promising effects at given intensity levels. Many intervention techniques and approaches have been reported in the literature. We know virtually nothing about their effectiveness at different doses, dose frequencies, and cumulative intensities. (p. 76) Summary 9

23 With increasing numbers of young children with disabilities participating in inclusive settings, ECSE teachers must travel to community-based early childhood classrooms to provide specialized services to these eligible children. Two approaches to providing itinerant services have been identified. The direct services approach is often used by itinerant ECSE teachers even though the consultative approach is recommended by leaders in the field. With the direct services approach, young children in inclusive settings usually receive specialized instruction one day each week. After the itinerant ECSE teacher leaves the setting, the general early childhood education teacher rarely continues to provide specialized instruction to the child with disabilities. This study examined if this once-a-week schedule was sufficient to promote learning in young children with disabilities, especially those children with cognitive disabilities. Chapter 2 examines the issue of once-a-week instruction by reviewing research on distributing instruction (i.e., the spacing effect). Both federal mandates and professional organizations have required the use of evidence-based practices (EBP) within special education to improve outcomes for children with disabilities. But just using an EBP is insufficient; it is also necessary to implement the practice with fidelity. This means that practitioners need to utilize the correct procedures described for the practice as well as the appropriate intervention intensity. Not using the same procedure and intensity that were used when the practice was determined to be effective may result in less than optimal outcomes for children. Chapter 2 examines one EBP (the practice of constant time delay) and the intervention intensities used in previous studies of the effectiveness of this strategy. This study compared two different intervention intensities of the constant time delay strategy. 10

24 Objectives of the Study There are two main objectives of the study. The first was to compare two different intervention intensities of an evidence-based strategy during a receptive word-learning task. The subjects of this study were preschool children who were receiving special education services under an IEP. Two different intervention schedules were used, with each schedule employing different doses and dose frequencies. One of the schedules provided 12 teaching episodes during a single session (dose), with 1 session occurring one day each week (dose frequency). The second schedule provided 4 teaching episodes during a single session (dose), with 1 session occurring each day and 3 sessions occurring each week (dose frequency). In general, the study sought to examine the effectiveness and efficiency of using an evidence-based strategy at two different intervention intensities. The second objective was to determine the types of supports needed by early childhood special education teachers (ECSE) in order to reliably implement the intervention strategy. Before beginning the experimental conditions, the ECSE teachers received the same training in the use of a specific evidence-based teaching strategy (constant time delay). Specifically, the following research questions were addressed: 1. Will young children who were identified with special needs target acquire target behaviors using both instructional schedules (once a week and 3 times a week)? (effectiveness) 2. Will a once-a-week instructional schedule or a 3 times a week instructional schedule result in more rapid acquisition of target behaviors? (efficiency) 11

25 3. Will these young children generalize the target behaviors across 2 different teachers? In the generalization condition, will there be more correct responses from children who learned with the 1x/week or children who learned in the 3x/week schedule? 4. Will these young children maintain the target behaviors 2 weeks after the generalization probe? In the maintenance condition, will there be more correct responses from children who learned with the 1x/week schedule or children who learned in the 3x/week schedule? 5. What types of supports will preschool teachers need in order to implement the intervention strategies reliably? 12

26 Chapter Two Literature Review The Spacing Effect This study examined the effectiveness and efficiency of two different intervention intensities for the CTD strategy. In other words, the study addressed the question of how distributed instruction should be in order to facilitate the learning of young children with disabilities. If instruction were more widely distributed and occurred once a week, was too much forgotten? Additionally, did young children with developmental delays learn when instruction was distributed over three days? The study of distributing instruction and examining its effects on learning was first documented by Ebbinghuas in 1885 (Cepeda et al., 2009). Since that time researchers have consistently found that performance improves if practice sessions are distributed (or spaced) rather than massed. This influence of distributed practice is called the spacing effect, and it is considered one of the best known and most robust phenomena in experimental psychology (Bahrick & Hall, 2005; Dempster & Farris, 1990; Rea & Modigliani, 1985). Researchers continue to explore and refine the spacing effect. Questions remain about this effect, such as what is the optimal amount of spacing for learning certain tasks, what amount of spacing optimizes retention, and do these effects hold for children? What follows is an explanation of general characteristics of the spacing effect along with studies conducted with participants who were adults, children developing typically, and children with disabilities. General characteristics of the spacing effect. A typical spacing effect study involves massed and distributed practice sessions. In each practice session, whether 13

27 massed or distributed, an individual studies the same material. What changes is the schedule of instruction, i.e., how often and when the person practices. An example of massed practice would be if a child repeatedly practiced a skill once a day for 30 minutes. Conversely, with distributed practice, the child might practice the same skill three times a day for ten minutes each time (for a total of 30 minutes). For empirical purposes, what is important is that in each case the child studies the same amount of time (e.g., 30 minutes) or the same amount of material (e.g., 10 words). In these examples, the practice sessions are massed or distributed within one day; however, practices could also occur over multiple days. For example, a child might practice a skill one time a day for four days. Therefore, practice sessions or instruction could be massed within a single day, distributed within a single day, or distributed across days. Within empirical investigations of the spacing effect, two important concepts have been considered: interstudy interval and retention interval. The interstudy interval (ISI) is the amount of time (i.e., interval) separating different study or teaching episodes. For example, if a child practices a skill three times a day with two hours between each session, then the ISI is 2 hours. If a child practices a skill once a day for 30 minutes, then the ISI is zero since there is only one practice session. If a child practices one day a week for 4 weeks, then the ISI is 7 days (1 week). The retention interval (RI) is the amount of time (i.e., interval) between the last instructional/study time and the test. For example, a teacher may practice vocabulary words with a student and then immediately test the student on those words. In this case the retention interval (RI) would be zero. If the teacher waited three hours after the 14

28 review to test the student, the RI would be 3 hours. And if the test were given two weeks after the review (with no other practice by the student), then the RI would be 14 days. Research with adult participants. In two research reviews of the spacing effect (Cepeda, Pashler, Vul, Wisted, & Rohrer, 2006; Donovan & Radosevich, 1999) over 300 experiments involving mostly adults were analyzed. Overall both reviews concluded that spaced practice results in more learning and is, therefore, superior to massed practice. Two other findings are relevant. The first is that too much spacing may not be beneficial. Donovan and Radosevich (1999) found that in some cases, longer ISIs actually negated the beneficial effects of distributed practice. In other words, if too much time elapsed between practice sessions, the individual did not learn as much in the spaced condition as they learned in the massed condition. Cepeda et al. (2006) also noted this effect and discussed it as a relationship between the ISI and the RI. They found that both the ISI and RI contributed to the spacing effect. For a given RI, as the ISI increased, more information was retained; however, if the ISI was increased too much, then retention declined. The second relevant finding is that distributing across days rather than simply within a single day benefits learning. Cepeda et al. (2006) concluded this after examining the effects of ISI alone as well as the joint effects of ISI and RI on level of retention. In the studies they reviewed, there were ISIs that ranged from 1 second to more than 100 days, and RIs that ranged from 1 second to 5 years. While the overall results of their analysis found that distributing practice is more beneficial to learning than massing practice (average benefit of 15%), they also stated that 15

29 Distributing learning across different days (instead of grouping learning episodes within a single day) greatly improves the amount of material retained for sizable periods of time; the literature clearly suggests that distributing practice in this way is likely to markedly improve students retention of course material. (Cepeda et al., 2006, p. 371) Cepeda et al. (2006) noted educational implications for their findings. The goal of education is to have students remember information for extended periods of time such as months and even years. Their review indicated that long-term retention can be maximized by separating learning sessions by at least one day rather than massing all learning into one session. Additionally, more information was retained after one month if learning sessions were spaced across weeks or months instead of across one day. Despite these promising findings, they cautioned that it was too early to extend the primarily adult data to situations with children. About 88% of the data in their analysis were based on adults (18 years of age or older), while preschoolers made up 29 assessments of distributed practice or only about 3% of the data. Even though most of the studies with children did show a benefit for distributing practice, the retention intervals were only about one day. Until empirical data examining the distributed practice effect in children are collected, using retention intervals of months or years and ISIs of days or months (no usable data meeting these criteria currently exist, to our knowledge), we cannot say for certain that children s long-term memory will benefit from distributed practice. (371) Research with child participants. Most of the studies examining the spacing effect utilized adult participants and experimental tasks conducted in laboratory settings. 16

30 However, there have been studies conducted with young children that employed more typical learning tasks. The first set of studies examined distributing practice within a day, and the next set of studies analyzed the effects of distributing practice across multiple days. Distributed within a day. Rea and Modigliani (1987) examined the spacing effect with 96 children from four age groups (preschool, kindergarten, first grade, and third grade). Children were asked to recall nouns and pictures after learning occurred in massed and distributed practice conditions. In the massed condition items were presented successively with no distracter items presented in between items (ISI=0). In the distributed condition, items were presented with one (ISI=1) or three (ISI=3) intervening items. A test to see how many items could be recalled was given immediately after the last item was presented (RI=seconds). Rea and Modigliani found that while older children recalled more than younger children, all age groups experienced a benefit of spacing. Further, the number of items recalled increased as the spacing increased. As the ISI increased from 0, to 1, to 3 items, the number of items recalled increased from 0.79, to 1.08, to A study that used a more typical task and setting was conducted by Seabrook, Brown and Solity (2005). Their study involved teaching reading through phonics instruction to 34 young children with an average age of 5 years. Instruction occurred in regular classrooms over a two-week period with massed practice occurring in 6-minute sessions each day and distributed practice occurring in three two-minute sessions each day. A pre-test was administered before the instruction began, and a final test was given one day after the instruction was concluded (RI=1 day). Results indicated that distributed 17

31 practice resulted in greater learning than massed practice. The average improvement score for massed practice was 1.3, while for distributed practice it was 8.3. After two weeks, the children in the distributed condition had test scores that were six times greater than their counterparts in the massed condition. Distributed across days. Childers and Tomasello (2002) taught 36 two-and-onehalf year olds novel words over the course of one month during sessions that lasted five to ten minutes. There were multiple conditions with ISIs that ranged from 0 days (massed) to 3 days (distributed). Three tests were given. One test was given immediately after training was completed, the second occurred one day after training, and the third was given one week after training. The results indicated an effect for ISI but not for RI; if the child learned the novel word they remembered it the next day and the next week. Other results indicated that children learned the words best when practice was distributed over four days regardless of the number of intervening days. The worst learning occurred in the two massed conditions when all presentations were done in a single day. Ambridge, Theakston, Lieven, and Tomasello (2006) required 48 young children, aged 3.5 to almost 6 years old, to construct complex grammatical sentences. For example, after observing puppets perform some actions, the child would state, It was the cup that the frog took. This task differed from the others examined so far in that this was a higher level mental task. The child was not simply recalling or recognizing information but had to use the observed information to construct a unique sentence. Three learning conditions were used: (1) a massed condition where 10 training trials were presented one after the other on a single day; (2) a distributed-pairs condition where 10 training trials were presented with 2 trials per day for 5 consecutive days, and (3) a distributed condition 18

32 where 10 training trials were presented with 1 trial per day for 10 days. A final test was given immediately after the last trial (RI=seconds). Even with this complex task, children performed better in the distributed and distributed-pairs conditions than in the massed. Rovee-Collier, Evancio, and Early (1995) described a set of two experiments. In both experiments infants of approximately 90 days old were taught in two 15-minute sessions to kick to activate on overhead crib mobile. In the first experiment there were four conditions that depended on when the second training session was conducted either 1, 2, 3, or 4 days after the initial training session. A long-term retention test was conducted about 7 days after the initial training session (RI=3, 4, 5 or 6 days). They found significant retention if the second session occurred within 1, 2, or 3 days of the first session. If the second session occurred 4 days after the first, the effects of the two sessions did not cumulate, and retention was no better than that of infants who received only a single training session (p. 73). In the second experiment by Rovee-Collier et al. (1995), the same task was used, but there were only three conditions or groups. The first group received its training 2 days after the initial training, while the second received the second training 4 days after the initial training. Both of these groups received a reactivation treatment 14 days after the initial training (and one day before the long-term retention test). During the reactivation treatment, the experimenter (not the infant) activated the mobile and attempted to recreate the infant s performance from his second session. The third group was a control group. This group received a second training session 2 days after the initial training but did not receive the reactivation treatment. The results indicated a clear advantage for the group whose second session occurred 2 days after the first and who received the reactivation 19

33 treatment. The group that had their second session 4 days after the first performed at baseline on the long-term retention test, just like the control group. In other words, the group whose second session occurred 2 days after the first (ISI=2 days) had the test 13 days later (RI=13 days), while the other group had an ISI of 4 days and a RI of 11 days. Even though there were 2 fewer days between the second study session and the test, the group with the 4-day ISI performed poorer than the group with the longer RI. Overall, Rovee-Collier s (1995) work reveals the critical issue in session-spacing effects is whether or not a second session or study trial will be integrated with the memory of the preceding session or study trial. If it is, then individuals will exhibit a retention advantage; if it is not, then individuals will exhibit retention no better than had they received only a single training session. If sessions are too widely spaced to be integrated, then they are represented in memory as completely independent events that are neither associated nor otherwise linked. (p. 150) Research with children with disabilities. Researchers within the field of special education have also examined the question of massed and distributed practice. Participants have ranged from preschoolers with developmental delays (e.g., Chiara, Schuster, Bell, & Wolery, 1995; Wolery, Doyle, Gast, Ault, & Simpson, 1993) to adolescents with moderate to severe disabilities (e.g., Collins, Evans, Creech-Galloway, Karl, & Miller, 2007). Five studies were reviewed. Three of these studies contrasted massed and distributed practices (Chiara et al., 1995; Jameson, McDonnell, Johnson, Riesen, & Polychronis, 2007; Wolery et al., 1993), one compared two different 20

34 distributed practice schedules within one day (Polychronis, McDonnell, Johnson, Riesen, & Jameson, 2004), and one examined the effects of every day versus every-other-day instruction (Venn et al., 1996). Of particular interest in these studies are the longer RIs. Maintenance tests were conducted from 1 week to 16 weeks after the last day of instruction. Massed versus distributed practice within a day. Two of these studies utilized preschoolers as participants (Chiara et al., 1995; Wolery et al., 1993), while one study involved middle school students with developmental disabilities (Jameson et al., 2007). The tasks in each of the studies employed discrete skills (rather than chained behaviors), and the teaching strategies were a variation of time delay. Wolery et al. (1993) presented massed trials consecutively at the beginning of the morning, while distributed trials were presented during transitions between activities throughout the morning. Chiara et al. (1995) presented the massed trials during small group instruction and presented the distributed trials individually. Jameson et al. (2007) presented the massed trials during children s special education classes and distributed trials during their general education classes. In the massed conditions of each study, trials were always presented consecutively with the same number of trials presented as during the matched distributed condition. In the distributed conditions, 16 trials were spaced throughout the morning with a minimum intertrial interval of 3 minutes (Wolery et al., 1993); 10 trials were spaced throughout a 6.5 hour day with intertrial intervals ranging from minutes (Chiara et al., 1995), or; 3 trials were spaced throughout a single general education class (Jameson et al., 2007). Wolery et al. (1993) reported that instructional trials were 21

35 conducted 4 days/week, while the other two studies simply stated that instruction occurred every day. Results of all three studies indicated that children learned to criterion whether they were in massed or distributed conditions. In other words, both massed and distributed practice sessions were effective for learning. With regard to efficiency, the results were mixed. Wolery et al. (1993) found no substantial differences in efficiency between progressive time delay massed trials and transition-based teaching distributed trials. Chiara et al. (1995) found slightly more efficient learning with the individually distributed conditions (fewer trials to criterion and lower error percentages) compared to the small group massed conditions. However, if just the data from the 2 children with documented deficits in cognition are considered, then the differences between the conditions are more pronounced. Across all children the number of trials to criterion averaged 56 for the massed condition and 45 for the distributed; however, for the 2 children with documented cognitive deficits, the average number of trials to criterion was 110 for the massed and 41 for the distributed. With Jameson et al. (2007) there were differences in efficiency by student. For two of the students, the massed instruction within the special education class resulted in fewer trials to criterion, but for the third student, the embedded condition within the general education class resulted in more efficient learning. For the fourth student, there were no differences between the two conditions. Generalization and maintenance were tested by Wolery et al. (1993) at 1 week for one child and at 1 and 3 weeks after the final probe for another child. Chiara et al. (1995) tested generalization and maintenance at 1, 3, and 5 weeks after the final probe. Both 22