Printer Friendly

Training teachers to promote pretend play in young children with disabilities.

Children with disabilities engage in less object play than children without disabilities (Blasco, Bailey, & Burchinal, 1993); children with autism display fewer play behaviors and less variety in their play with objects (e.g., Jarrold, Boucher, & Smith, 1996; Ungerer & Sigman, 1981). Object play, however, is important for young children with disabilities for a variety of reasons. Object play sets the occasion for social interactions and communication with peers (McConnell, 2002) and provides a context for implementing instruction (Sandall, Hemmeter, Smith, & McLean, 2005), such as activity-based interventions (Pretti-Frontczak & Bricker, 2004), embedded learning opportunities (Sandall & Schwartz, 2002), and response prompting procedures (Wolery, 2001). Also, play with objects may increase engagement in inclusive settings (Lieber, 1993; Morrison, Sainato, Benchaaban, & Endo, 2002) resulting in higher sociometric ratings by peers without disabilities (Strain, 1985).

With typically developing children, object play can take many forms and frequently involves using objects in nonliteral ways. Nonliteral behaviors involve pretense, and play taxonomies refer to these behaviors as a distinct category. However, there are major inconsistencies in the name and response definition of these behaviors across taxonomies. For example, nonliteral behaviors were referred to as pretend play, symbolic play, functional play, or object play. To be precise and consistent, the term pretense is used here to refer to only and all nonliteral play behaviors.

In a number of studies (e.g., DiCarlo & Reid, 2004; Jarrold et al., 1996) children with disabilities were taught to engage in pretense behaviors using adult prompts. A review of the literature identified 16 intervention studies on teaching pretend play to children with disabilities (Barton & Wolery, 2008). Five major weaknesses were identified. First, pretense behaviors were measured inconsistently across studies. For example, Lifter, Sulzer-Azaroff, Anderson, & Cowdery (1993) measured only functional play with pretense and denoted assigning absent attributes as pretend play. DiCarlo and Reid did not define the target behaviors with replicable precision (i.e., "a single step action that appeared to imitate a real life situation involving objects that corresponded to the toys used in the action", p. 199). They included a variety of examples, but did not include any non-examples. It is difficult to discern which behaviors were not included as pretend play behaviors. Nonliteral play with dolls or with miniature objects was defined as pretend, symbolic, or functional play, depending on the report. For example, walking the doll (Kim, Lombardino, Rothman, & Vinson, 1989), using a banana as a phone (Sherrat, 2002), talking into a miniature phone (DiCarlo & Reid, 2004), bringing an empty cup to mouth (Kasari, Freeman, & Parapella, 2006) were measured as pretend play in the respective studies.

Second, only 3 of 16 studies removed all prompts from the measurement context (Kasari et al., 2006; MacDonald, Clark, Garrigan, & Vangala, 2005; Sherrat, 2002). The use of prompts during measurement limits conclusions about change in pretense behaviors. Behavior cannot be considered acquired, unless it is performed without prompts (i.e., transfer of stimulus control has occurred from the prompts to the toys or objects).

Third, limited measurement of maintenance and generalization of children's pretense behaviors occurred in sessions without adult prompting. For example, 4 of 16 play intervention studies measured generalization. However, only 2 specifically programmed for generalization (Lifter, Ellis, Cannon, & Anderson, 2005; Lifter et al., 1993) and none removed prompts from the generalization sessions. For pretense behaviors to count as generalized pretend play, the behaviors must occur in novel situations in which prompts are not used.

Fourth, in most studies, research staff, rather than classroom teachers, applied the interventions (Barton & Wolery, 2008). Further, only two studies reported the interventionist experience and training. This limits the generality of the pretend play studies.

Fifth, almost no studies measured the fidelity of adults' prompting, including whether the prompts were actually removed. Three of 16 play intervention reports described procedural fidelity (Kasari et al., 2006; Stahmer, 1995; Thorp, Stahmer, & Schreibman, 1995). However, only Kasari et al. met basic standards for procedural fidelity.

A frequently used procedure in this literature was the system of least prompts (Barton & Wolery, 2008). The system of least prompts has been implemented with fidelity in direct instructional arrangements (Doyle, Wolery, Ault, & Gast, 1988) and in play contexts to promote communicative exchanges in young children (Filla, Wolery, & Anthony, 1999). Thus, it was selected for this study. Effective strategies for training teachers to use new interventions include (a) short training sessions with brief procedural manuals (Filla et al., 1999), (b) guided practice and rehearsal (Lifter et al., 2005), (c) video examples (Moore & Fisher, 2007), and (d) verbal and written performance feedback (e.g., Casey & McWilliam, 2008). These components were used to train the teachers to implement the system of least prompts in this study.

This study addresses the methodological weaknesses noted earlier in the literature on teaching children with disabilities to use pretense behaviors. Specifically, children's pretense behaviors were measured and taught using a taxonomy developed from the intervention and assessment literature on pretend play (Barton & Wolery, 2008). Maintenance was measured periodically across the study in probe sessions with the teacher but without prompts. Generalization was measured in two ways: (a) with an adult who was not the child's teacher and who did not use prompts (generalization across adults), and (b) with an untrained set of toys and an adult who was not the child's teacher and did not use prompts (generalization across toys and adults). Classroom teachers were the instructors, rather than research staff, and procedural fidelity data were collected on the teachers' use of the intervention package.

METHOD

PARTICIPANTS

Teacher Participants. Four female teachers from a university-based inclusive preschool participated. Inclusion criteria for the teachers were (a) minimum of 1 year of classroom experience, (b) no previous instruction on promoting pretense behaviors, and (c) study-eligible children enrolled in their classroom. Each teacher was assigned to a child participant in her classroom. Darcy was a 28-year-old lead teacher with a Bachelor's degree in Speech Pathology. She had 6 years of paid experience with children. Amy was a 27-year-old lead teacher with a Master's degree in School Psychology. She had 3 years of paid experience with children. Lucy was a 48-year-old co-teacher in the classroom with Amy. She held a high school diploma with 2 years of college and had 24 years of paid experience with children. Beth was a 24-year-old co-teacher with a Master's degree in Special Education and had 3 years of paid experience with children.

Child Participants. Four preschoolers with disabilities were recruited from the three classrooms of consented teachers. Inclusion criteria for the child participants were (a) a diagnosed disability based on school records; (b) age less than 60 months; (c) minimum mental age of 18 months as measured by the Mullen Scales of Early Learning (MSEL; Mullen, 1995), because children without disabilities begin to engage in pretend play behaviors at about 18 to 24 months (Belsky & Most, 198i; Fein, 1981); (d) consistent school attendance (80% for previous month based on teacher report); (e) ability to participate in a play activity with an adult for 8 min; and (f) less than eight different unprompted play behaviors and no unprompted symbolic behaviors during an adapted version of the Structured Play Assessment (SPA; Ungerer & Sigman, 1981). The last two criteria were measured in a 10-min videotaped play assessment. Table 1 shows this data on the children.

Daniel was a 43-month-old European American male and fraternal twin of Anna. Darcy was his teacher. Daniel had a mental age of 36 months on the MSEL and was the highest functioning of the four children. He regularly used three- to five-word phrases to express wants and needs and to comment and narrate his play. He received speech and language therapy (SLT) from the local school district. He infrequently engaged in social interactions with peers or adults. His teacher described his object play as rigid, repetitive, and restricted to themes involving movie cartoon toys.

Anna was a 43-month-old European American female (twin sister of Daniel) who received a diagnosis of autism at 24 months. Amy was her teacher. Anna had a mental age of 23 months and received special education, SLT, and occupational therapy (OT) from the local school district. She did not initiate conversations, but she was verbally imitative and repeated three- to four-word statements. She did not initiate interactions with peers unless prompted by an adult. Anna rarely played with toys without adult prompting and often was alone in the book or music areas of the classroom turning the pages of books or playing musical instruments.

Liz was a 50-month-old European American female diagnosed with developmental delays. Lucy was her teacher. Liz had a mental age of 27 months and received SLT and OT from the local school district. She was verbal, but often used gestures, signs, and one-word utterances to request preferred items; she rarely initiated conversations or interactions with adults or peers. The majority of Liz's social interactions were responses to adult behaviors. Liz participated in a feeding program each morning during the study. She displayed a few rigid play behaviors, which emulated her feeding program (using a spoon to scrape the inside of a bowl). She rarely elaborated on this play without adult prompting. Liz rarely played independently with toys in the classroom.

Brian was a 30-month-old African American male with a diagnosis of autism. Beth was his teacher. Brian had a mental age of 18 months and received SLT from a local early intervention provider. He exhibited echolalia and unintelligible speech, which he used to request objects and attention from adults. Brian often used vocalizations and gestures to request adult attention or to be picked up. He rarely initiated interactions with peers, except when attempting to retrieve a preferred object. He often ignored peers' initiations. Brian displayed a few rigid and brief play sequences with the housekeeping toys, which replicated mealtime activities (feeding and cleaning). Brian spent the majority of his time during free play at the sensory table (e.g., shaving cream, Jell-O) or rolling vehicles back and forth. Brian's caregiver abruptly removed him from school before the study ended.

SETTINGS

The initial teacher training session occurred in the conference room at the school. Additional training sessions were in individual teachers' classrooms. The adapted SPA sessions occurred in a school therapy room. All experimental sessions for children were conducted in each child's classroom. During these sessions, other adults and children participated in the typical classroom routines and activities.

MATERIALS

Five toy sets were used (three for instruction, one for generalization, and one for the adapted SPA) and are shown in Table 2. Each set was used with each teacher-child dyad. The toy sets were (a) selected to provide children with opportunities to use pretense behaviors, (b) based on toys available in preschool classrooms, and (c) similar to toys in previous pretend play studies (Lifter et al., 1993; Thorp et al., 1995). Although each toy set contained objects with similar functions (e.g., dolls, spoons), all objects were markedly different in form (e.g., shape, size, color) across toy sets.

RESPONSE DEFINITIONS AND MEASUREMENT

Teacher Behaviors. The intervention package included a system of least prompts, contingent imitation, and a system for reinforcement for all children. However, the specific teacher behaviors varied based on the child's response patterns. Seven teacher behaviors were measured in each experimental session for all teachers. Contingent imitation was the teacher doing the Same behavior (motor or vocal) as the child. A correct model prompt was the teacher doing a pretense behavior with a toy and verbally describing the action at least 12 s but not more than 20 s after a previous prompt or pretense behavior. A model prompt error was the teacher failing to provide a model within 20 s of the last child pretense behavior, modeling a nonpretense behavior, or providing a correct model after 20 s of the last child pretense behavior. A correct physical prompt was the teacher using her hands to guide the child's hands through a pretense behavior within 5 s of a model prompt that did not result in a pretense behavior. A physical prompt error was the teacher failing to deliver a physical prompt within 5 s of a model prompt that did not result in a pretense behavior, using her hands to move a child through a nonpretense behavior, or delivering a physical prompt more than 5 s after a model prompt that did not result in a pretense behavior. A prompt sequence error occurred when the teacher delivered a physical prompt before delivering a model prompt. Missed opportunities occurred when more than 20 s elapsed without a child pretense behavior and the teacher failed to deliver the prompt sequence. Event sampling was used to record each of these behaviors. Data were collected using a digital video recorder, coded using ProCoderDV (Tapp & Walden, 2000), and analyzed using INTMAN software (Tappet al., 2006).

For Amy (Anna's teacher), four additional behaviors were measured. Correct toy presentation was Amy placing two or three toys in Anna's lap within 12 s but not more than 20 s after the last prompt or last pretense behavior. A toy presentation error occurred when Amy failed to place two or three toys in Anna's lap within the time window or only placed one toy in Anna's lap. Correct picture presentation occurred when Amy showed Anna a picture depicting a pretense behavior and verbally described the depicted action at least 12 s but not more than 20 s after the last prompt or last pretense behavior. A picture presentation error occurred when no picture was shown to Anna within the time window or a picture was presented but Amy failed to label the action. For Lucy, two additional behaviors were measured. A correct toy choice was Lucy saying, "Do you want to play with the--or the--?" and holding two toys in front of Liz at least 12 s but not more than 20 s after the last prompt or pretense behavior. A toy choice error occurred when Lucy failed to hold two toys in front of Liz or did not ask Liz to make a choice.

Adapted Structured Play Assessment (SPA). The SPA is used to describe children's optimal play skills (e.g., Mundy, Sigman, Ungerer, & Sherman, 1987; Ungerer & Sigman, 1981). The SPA has six behavior categories and was adapted for this study by dividing functional play into two categories: functional play with and without pretense. Three categories did not indicate pretense: simple manipulation of one object was any child movement of a single object (e.g., mouthing a block); relational manipulation of more than one object was moving two objects in relation to each other in a nonfunctional way (banging a car on top of a block); and functional play was moving an object or moving two objects together in a functional way (dosing doors on a toy house, putting puzzles together). Four categories indicated pretense: functional play with pretense (FPP) was movement representing a nonliteral action of one or more objects (puts spoon to doll's mouth); object substitution (OS) was movement of one object as if it were a different object (putting a block to the doll's mouth as if it were a bottle); imagining absent objects (IAO) was a movement implying an object was present although it was not (putting a fist near a doll's mouth to feed without the bottle); and assigning absent attributes (AAA) was movements or vocalizations assigning roles, emotions, or attributes (picking up a doll and saying, "crying."). Event sampling was used to count the number of play behaviors across each category. The SPA was used as a pre- and posttest, and pretest scores were used as an inclusion criterion.

Pretense Behaviors. An event recording system was used with child pretense behaviors using three steps. First, each pretense behavior was coded as prompted or unprompted to provide evidence for transfer of stimulus control. Unprompted pretense behaviors were those occurring without a prompt in the previous 5 s. Prompted pretense behaviors were those occurring within 5 s of a teacher prompt. Second, each pretense behavior was coded as same or different. Behaviors were coded as different if they had not occurred previously in the session and as the same if they had occurred previously in the session. Thus, the total number of prompted same, unprompted same, prompted different, and unprompted different pretense behaviors was generated for each session to examine the diversity of pretense behaviors. Third, the pretense behaviors were coded by type: functional play with pretense, object substitution, imagining absent objects, and assigning absent attributes (see Barton & Wolery, 2008) using the same definitions as for the adapted SPA.

In addition, the number of sequences and number of behaviors per sequence were measured with event recording. A sequence consisted of at least two related pretense behaviors occurring within 3 s of each other. For each sequence, the number of pretense behaviors was counted. Finally, each vocalization related to a pretense behavior was recorded. Vocalizations were coded as prompted if they were imitative and occurred less than 5 s after a teacher vocalization and unprompted if they occurred more than 5 s after a teacher vocalization, regardless of whether they were imitative or non-imitative.

Interobserver Agreement (IOA). A second observer coded at least 20% of the videotaped sessions for each toy set, condition, and child. Because event recording was used, the formula for calculating IOA was the smaller number of one observer divided by the larger number of the other observer and the quotient was multiplied by 100 (Gast, 2010). The IOA data for child pretense behaviors are shown in Table 3.

DESIGN AND PROCEDURES

This study used a multiple probe design across three toys sets and replicated across four children with disabilities (Gast & Ledford, 2010). The design progressed through the following conditions: (a) initial probe condition measuring each child's behavior with each toy set in sessions with the child's teacher and sessions with a nonteacher adult (generalization across adults) who did not prompt pretense behaviors, (b) initial teacher training session followed by instructional sessions on the first toy set with each child and daily generalization measures with a nonteacher adult who did not prompt pretense behaviors, (c) second probe condition measuring each child's behavior with each toy set in sessions with the child's teacher and sessions with a nonteacher adult (generalization across adults) in which neither used prompts, (d) second teacher training session followed by instructional sessions on the second toy set with each child and generalization measures with a nonteacher adult who did not use prompts, (e) third probe condition identical in procedures to the second probe condition, (f) third teacher training session and instructional sessions on the third toy set and generalization measures as in the second intervention condition, (g) final probe condition identical in procedures to the second and third probe conditions, (h) generalization across toy sets measuring the child's behavior with a new toy set and with the nonteacher adult who did not prompt pretense behaviors, and (i) posttest with the SPA in which pretense behaviors were not prompted. The decision to intervene with each subsequent toy set was based on a clear change in level of the child's number of unprompted pretense behaviors.

Initial Probe Condition. These 8-min sessions occurred before teacher training. The investigator told the teachers to play with their child "as they normally would"; they were free to use verbal, model, or physical prompts. Each toy set was assessed for at least three sessions (nine total sessions).

Teacher Training. The goals were to train the teachers to (a) use contingent imitation during play, (b) use the system of least prompts with the a specific toy set, (c) discriminate nonpretense from pretense behaviors, and (d) identify examples of four types of pretense (functional play with pretense [FPP], object substitution [OS], imagining absent objects [IAO], and assigning absent attributes [AAA]). Contingent imitation was taught to focus the teacher on the child's current behavior and to elicit the child's attention to the adult behavior (Ingersoll & Shreibman, 2006). Immediately after the initial probe condition, the teachers were given a six-page manual 3 days before the initial 45-min training session. The manual included a rationale for the study, a description of the prompts and prompting procedures, a definition and description of contingent imitation, examples of teacher statements to be used, and examples of pretense behaviors. This session occurred with the teachers as a group at the end of the school day and focused intervention procedures with Toy Set 1. In the first 30 min, the investigator and teachers discussed and reviewed the manual and watched the video. The video included two 10-min segments: (1) the investigator modeling contingent imitation and the prompting sequence with a child, and (2) a child without disabilities engaging in pretense and nonpretense behaviors with the first toy set. The teachers then practiced the prompting sequence with nonparticipant children, and the investigator provided feedback and praise for correct use of the procedures.

Training for the second and third toy sets occurred individually with each teacher after her child completed the second and third probe conditions, respectively. These sessions were shorter (20-30 min) and occurred in the teacher's classroom. The investigator reviewed the manual and videos and modeled each of the four types of pretense behaviors with the new toy set. Again, the teacher practiced using the prompting sequence with nonparticipant children, and the investigator provided feedback and praise for correct use of the procedures. In addition, before each child instructional session, the investigator gave the teacher a checklist with at least two examples of each type of pretense behaviors with the current toy set. During the session, the investigator recorded examples of correct use of and errors with the prompting sequence. After each session, the investigator provided verbal and written feedback using a one-page feedback form with space for examples of four pretense behaviors the teachers had prompted and the number of missed opportunities to prompt. The investigator asked teachers to initial that they received the feedback. This feedback form was used for the duration of the study.

Child Instructional Sessions. In the 8-min instructional sessions, the teachers contingently imitated the child, applied the system of least prompts targeting the four types of pretense behaviors (functional play with pretense, object substitution, imagining absent objects, and assigning absent attributes), and praised all pretense behaviors. A system of least prompts was used with three or four levels from least to most intrusive, depending on the child's responses. The first level (independent level) consisted of the presentation of the materials and the verbal statement "Let's play" at the beginning of the session. The teacher waited 12 to 20 s while contingently imitating and observing the child; if the child did not do a pretense behavior, the teacher modeled a pretense behavior and labeled the action. The teachers were asked to prompt all four types of pretense behaviors within each session based on the child's behavior or interest. For example, if the child was banging a spoon on the floor, the teacher imitated banging the spoon, waited 12 to 20 s, moved the spoon into a cup, made a stirring motion, and said, "I'm stirring." Or, if the child was not touching any toys, but looking at the sponges, the teacher picked up a sponge, moved it back and forth on the carpet, and said, "It's a fast car!" If the child did not imitate the model or do a pretense behavior within 5 s, the teacher used full physical, hand-over-hand prompts (the controlling prompt) to assist the child in doing a pretense behavior.

Consistent with a system of least prompts, the prompt sequence was adapted based on the child response patterns. For Anna and Liz, prompt levels were added because of low response rates. These levels occurred after the presentation of the toys and before the physical model in the prompting sequence. For Anna, a picture with a written description of the pretense behavior was presented with the model prompt (introduced during Toy Set 1 and included with Toy Set 2). Anna's teacher and parents suggested adding this visual cue, because visuals had been used successfully in teaching her new skills. With Toy Set 3, specific materials were presented to her with a verbal prompt, "Let's play," 12 to 20 s after the presentation of the toy set. For Liz, a choice between two toys was provided after the presentation of the toys, prior to the physical model (with Toy Set 1 only). Daniel consistently resisted hand-over-hand prompting so the controlling prompt was placing the toy in or near his hands.

If the child engaged in a pretense behavior, the teacher delivered descriptive praise (e.g., "Good feeding the baby, Liz"). However, descriptive praise did not function as a reinforcer for pretend play for Anna. Amy reported that edibles functioned as reinforcement for a variety of other skills for Anna; thus, she presented a small edible to Anna after each prompted and unprompted pretense behavior for Toy Sets 1 and 2 except during probe conditions. With Toy Set 3, Amy thinned the delivery of an edible to every other pretense behavior.

Second, Third, and Final Probe Conditions. These 8-rain sessions occurred with each teacher-child dyad for at least three sessions with each toy set. The procedures were identical to the initial probe condition, except teachers were told to refrain from using the system of least prompts, contingent imitation, or reinforcement (i.e., descriptive praise or an edible) to prompt pretense behaviors. Teachers provided descriptive praise for remaining in the play area.

Unprompted Daily Generalization Probes Across Adults. Generalization of child pretense behaviors was measured in 5-min play sessions with a nonteacher adult prior to the probe and instructional sessions. Two graduate students in special education served in the role of nonteacher adult. The nonteacher adult presented the same toy set used in the instructional sessions, but did not deliver prompts or use contingent imitation. Descriptive praise for sitting (e.g., "I like the way you are sitting") was delivered about every 20 s if the child remained in the specified area. For Anna and Liz, these generalization sessions were conducted intermittently to avoid toy or session satiation. Because of high rates of escape attempts, praise was delivered more often to Brian.

Generalization Probes Across Materials and Adults. One session was conducted per child during the final probe condition. The nonteacher adult used the procedures identical to those of the unprompted daily generalization session with a new toy set (see Table 2).

RESULTS

PROCEDURAL FIDELITY

Three types of procedural fidelity data were assessed. First, each teacher completed a 12-step checklist after each teacher training session. This checklist assessed whether the teacher training sessions were conducted as planned. Data indicated all 12 steps occurred in each training session. Second, the teachers initialed the daily feedback form to verify feedback was delivered. The forms were reviewed, and the teachers initialed 100% of them for each instructional condition. Third, video records of teacher behaviors in probe and instructional sessions were observed directly and coded. A second observer coded at least 20% of the video records for each teacher and condition to measure IOA on the teacher behaviors. IOA was calculated using the smaller divided by the larger method for each teacher behavior. IOA for Darcy's procedural fidelity was 94% (81%-100%) for Toy Set 1 and 100% for Sets 2 and 3; for Amy it was 88% (81%-100%) for Toy Set 1 and 100% for Sets 2 and 3; for Lucy it was 93% (86%-100%) for Toy Set 1, 91% (83%-100%) for Set 2, and 95% (88%-100%) for Set 3; for Beth it was 100% for Toy Set 1 and 92% (90%-100%) for Set 2.

The percentage of correct implementation was calculated for each session by subtracting the number of errors from the total number of prompts and dividing by the total number of prompts. The scores for all sessions were averaged to obtain a mean percentage of correct implementation per toy set. For Darcy, the mean percentage of correct implementation for Toy Set 1 was 88% (range: 69-100) and was 100% for Sets 2 and 3. For Amy, the mean percentage of correct implementation for Toy Set 1 was 84% (range: 67-100), and was 100% for Sets 2 and 3. For Lucy, the mean percentage of correct implementation for Toy Set 1 was 85% (range: 50-100), 94% for Set 2 (range: 83-100), and was 100% for Set 3. For Beth, the mean percentage of correct implementation for Toy Set 1 was 82% (range: 63-100), and 100% for Set 2. Toy Set 3 was not used because Brian withdrew from school unexpectedly. These data indicate the teachers used the intervention package with high fidelity. With each teacher, the lowest percentages of correct implementation occurred with Toy Set 1--indicating the teachers were acquiring the skills needed to use the intervention correctly.

EFFECTS ON PRETENSE BEHAVIORS

Acquisition and Maintenance of Pretense Behaviors. The number of unprompted and prompted pretense behaviors for Daniel, Anna, Liz, and Brian are shown in Figures 1 to 4, respectively. Total unprompted pretend play included any of the four types of pretense behavior (FPP, OS, IAO, & AAA). Unprompted pretend play increased with the introduction of the intervention; a clear functional relation was established between the intervention package and unprompted pretend play across toy sets and children.

For Daniel, Figure 1 shows that his unprompted pretense behaviors were low (less than three behaviors) in each session of each toy set during the initial probe condition. With the intervention for Toy Set 1, his prompted and unprompted pretense behaviors increased with an accelerating trend for unprompted pretense behaviors. In the second probe condition, his unprompted pretense behaviors for Toy Set 1 remained high (mean of 18.7 for the three sessions) and his unprompted pretense behaviors on other sets remained low. This pattern was replicated across the two subsequent toy sets. Thus, for each toy set, the unprompted pretense behaviors were low in probe conditions prior to intervention, increased during intervention, and maintained at high levels in probe conditions after intervention.

For Anna, Figure 2 shows that unprompted pretense behaviors were zero in each session of each toy set during the initial probe condition. She had a few prompted pretense behaviors in this condition for each toy set. With the intervention for Toy Set 1, both her prompted and unprompted pretense behaviors increased slightly. With the introduction of the visual prompt (instructional session 14) and the edible reinforcer (instructional session 18), prompted pretense behaviors increased and an accelerating trend is noted for unprompted pretense behaviors toward the end of the condition. In the second probe condition, her unprompted pretense behaviors for Toy Set 1 displayed an immediate and large increase in level (i.e., mean of 9.3 for the last three instructional sessions and 17.7 for the three probe sessions); unprompted pretense behaviors for Toy Sets 2 and 3 remained at zero. The introduction of the intervention (including the visual prompt and edible reinforcer) for Toy Set 2 resulted in an abrupt increase in prompted pretense behaviors and a gradual increase in unprompted pretense behaviors, with transfer of stimulus control occurring during the last three sessions. The visual prompt was discontinued after the tenth intervention session. During the third probe condition, unprompted pretense behaviors for Toy Set 2 remained high at a level slightly lower than the last three intervention sessions; Toy Set 1 maintained at levels similar to the second probe condition, and Toy Set 3 remained at zero. With the introduction of intervention for Toy Set 3 (including the presentation prompt and edible reinforcement delivered for every other prompted or unprompted pretense behavior), an abrupt increase in prompted pretense and an accelerating trend in the number of unprompted pretense behaviors occurred. In the final probe condition, an accelerating trend in the number of unprompted pretense behaviors occurred with the last data point slightly higher than the last data point during the instructional condition for Toy Set 3. Unprompted pretense behaviors for Toy Sets 1 and 2 maintained at high levels. Thus, for each toy set, the unprompted pretense behaviors were low in probe conditions prior to intervention, increased during intervention, and maintained at high levels in subsequent probe conditions (without prompts or edible reinforcement).

[FIGURE 1 OMITTED]

[FIGURE 2 OMITTED]

For Liz, Figure 3 shows that her unprompted pretense behaviors were low (less than six behaviors) in each session of each toy set during the initial probe condition. She had several prompted pretense behaviors for Toy Sets 1 and 3 (between 10 and 20), and few prompted pretense behaviors for Toy Set 2 (less than five). With the intervention for Toy Set 1, her unprompted pretense behaviors abruptly increased and then had a decelerating trend. With the introduction of the choice, unprompted pretense slightly increased and stabilized around 13 behaviors per session. In the second probe condition, her unprompted pretense behaviors for Toy Set 1 remained high and displayed an accelerating trend (mean of 27). No increase occurred in unprompted pretense behaviors for Toy Sets 2, and a slight increase occurred for Toy Set 3. These remained at zero behaviors for Toy Set 2 and below six behaviors for Toy Set 3. The introduction of intervention for Toy Set 2 (without the choice prompt) resulted in an accelerating trend in the number of unprompted pretense behaviors. In the third probe condition, unprompted pretense behaviors for Toy Set 2 remained high. Toy Set 1 started lower than during the third probe condition and showed an accelerating trend similar to the second probe condition. A slight increase occurred for Toy Set 3, but the number of unprompted pretense behaviors was less than eight. With the introduction of intervention for Toy Set 3, an accelerating trend in the number of unprompted pretense behaviors occurred and then stabilized at a level higher than during the third probe condition. In the final probe condition, the numbers of unprompted pretense behaviors were high across toy sets. Thus, for each toy set, the unprompted pretense behaviors were low in probe conditions prior to intervention, increased during intervention, and maintained during subsequent probe conditions.

[FIGURE 3 OMITTED]

For Brian, Figure 4 shows that his unprompted pretense behaviors were low for all toy sets in the initial probe condition. He had several prompted pretense behaviors during this condition for Toy Set 1. With the introduction of the intervention with Toy Set 1, unprompted pretense behaviors immediately increased and prompted pretense behaviors had a decelerating trend. In the second probe condition, he had more unprompted pretense with Toy Set 1 than in the first probe condition, but slightly lower than during intervention. Toy Sets 2 and 3 remained low during the second probe condition. With the intervention for Toy Set 2, the unprompted pretense behaviors increased to a higher level than during previous probe sessions. He was absent for several days and returned for one final instructional session before he withdrew from school. His unprompted pretense behaviors for Toy Set 2 were above previous probe conditions and lower than last session of intervention.

Types of Pretense Behaviors. Four types of pretense behaviors were measured (FPP, OS, IAO, and AAA). For each child, the percentage of each type of unprompted pretense behaviors was similar across intervention and generalization sessions within toy sets; the percentages for each type are shown in Table 4. Daniel had higher percentages of object substitution behaviors than the other three types with Toy Set 1, more assigning absent attributes behaviors with Toy Set 2, and similar percentages of object substitution and assigning absent attributes behaviors with Toy Set 3. He displayed few unprompted functional play with pretense behaviors across the toy sets. Anna displayed more unprompted functional play with pretense than object substitution with Toy Sets 1 and 3. With Toy Set 2, Anna displayed more unprompted object substitution than functional play with pretense. She displayed no imagining absent objects or assigning absent attributes. Liz and Brian displayed more functional play with pretense than object substitution, imagining absent objects, and assigning absent attributes across the three toy sets.

[FIGURE 4 OMITTED]

Diversity of Pretense Behaviors. The intervention was designed to teach children a diverse. repertoire of pretense behaviors from four categories rather than specific responses. Within the session, the teachers were free to vary the type of pretense they prompted. Thus, the diversity (defined as the number of different pretense behaviors per session) of pretense behaviors was a critical measure. The mean and ranges of the number of different pretense behaviors are shown in Table 5 for probe conditions with the teacher and with the nonteacher adult (generalization sessions). Teachers did not prompt pretense behaviors during the probe conditions; thus, probe condition data were analyzed because the diversity measure could be influenced by the teachers' prompts during intervention sessions. Three conclusions are noted: (a) the number of different pretense behaviors increased for all children on all toy sets after intervention was applied, (b) the children consistently had more different pretense behaviors with the teacher than with the nonteacher adult, and (c) the effects maintained across probe conditions. In summary, all children learned to engage in a variety of pretense behaviors in each condition and their performance maintained in sessions with teacher prompts removed.

Sequences of Pretense Behaviors. None of the children demonstrated sequences of pretense during the first probe condition. Overall, Daniel, Anna, and Liz displayed less than 10 sequences across subsequent conditions and toy sets. Brian displayed no sequences with Toy Sets 1 or 2. Thus, the intervention did not appear to promote sequences of pretense behaviors.

Vocalizations. For all children in all probe conditions of each toy set, no vocalizations occurred prior to instruction on the respective toy sets. For all children during each instructional condition, some vocalizations occurred; however, for all children except Daniel on Toy Sets 1 and 3, some instructional sessions had no vocalizations. Vocalizations occurred for all children in all post-instruction probe conditions, except for Liz on Toy Set 2 (third and fourth probe conditions). All children had fewer vocalizations during generalization sessions than during instructional sessions, and all children had some generalization sessions with no vocalizations.

Generalization Across Adults. Unprompted pretense was measured with a nonteacher adult for Toy Sets 1, 2, and 3. For Daniel and Anna (see Figures 1 and 2) in probe conditions before intervention, the frequencies of unprompted pretense behaviors were low in generalization assessment sessions; in probe conditions after intervention for all toy sets, the frequencies occurred at higher levels than before intervention and maintained across subsequent probe conditions. For Liz, Figure 3 shows that in probe conditions before intervention, the frequencies of unprompted pretense behaviors were low in generalization sessions for Toy Sets 1 and 2; in probe conditions after intervention for Toy Set 2, the frequencies occurred at higher levels than before intervention. In the second and third probe conditions, the frequencies of unprompted pretense behaviors were low in generalization assessment sessions for Toy Set 3; in the probe conditions after intervention for Toy Set 3, the frequencies occurred at higher levels than before intervention. For Brian, Figure 4 shows that during instruction with Toy Sets 1 and 2, the frequencies of unprompted pretense behaviors in the generalization sessions displayed an accelerating trend at levels higher than before intervention.

For Daniel, Anna, and Liz, the frequencies of unprompted pretense behaviors in daily generalization sessions during instructional conditions tended to be variable and at lower levels than in instructional sessions. With Anna, the frequencies were particularly low. With Brian, there tended to be an accelerating trend of pretense behaviors in generalization sessions, but his withdrawal from school precluded sufficient replications.

Generalization Across Toy Sets. For Daniel, Anna, and Liz, generalization to a new toy set with the nonteacher adult and without prompts for pretense behavior was assessed in an 8-min session during the final probe condition. Daniel engaged in 32 pretense behaviors, with 17 assigning absent attributes and 11 object substitution behaviors. Anna only demonstrated three pretense behaviors--two functional play with pretense and one object substitution behavior. Liz engaged in 31 pretense behaviors, and 26 functional play with pretense behaviors. Daniel had 24 vocalizations, Anna had none, and Liz had one.

Posttest of the SPA. As shown in Table 1, on the SPA pretest, Daniel had two different unprompted functional play with pretense behaviors and no unprompted symbolic behaviors; Anna, Liz, and Brian had no unprompted behaviors in either category. On the posttest, Daniel's number of unprompted functional play with pretense behaviors were essentially the same, but he had 18 symbolic behaviors. Anna increased substantially on functional play with pretense behaviors, and she had two symbolic behaviors. Liz increased substantially in both categories. Brian did not have the posttest because he had withdrawn from school.

DISCUSSION

This study examined the relation between the teachers' use of the system of least prompts and contingent imitation and pretense behaviors of children with disabilities. The effects of the intervention were evaluated on children's acquisition, maintenance, and generalization of the pretense behaviors. A number of measures were used, including the number of prompted and unprompted pretense behaviors during intervention sessions with their teachers; the types, the diversity, and sequences of pretense behaviors; and the use of vocalizations related to pretense behaviors. Six findings are discussed.

First, as shown in Figures 1 to 4, the intervention was functionally related to an increase in the number of unprompted pretense behaviors across children in the intervention sessions, which maintained across probe conditions when no prompts were used and at lower levels in the generalization-across-adult sessions, again without prompts. The intervention produced consistently higher levels of pretend play across children and toy sets in the intervention sessions and in probe conditions after intervention. The levels of pretense behaviors in the final intervention sessions for all participants were at least three times, and often more than four or five times, the levels prior to instruction. Daniel had means of 0, 3, and 3.25 unprompted pretense behaviors prior to instruction and means of 16.7, 17.3, and 14 for the final three instructional sessions across Toys Sets 1, 2, and 3, respectively. Anna displayed no unprompted pretense behaviors prior to instruction and means of 9.3, 15.7, and 14.3 for the final three instructional sessions across Toys Sets 1, 2, and 3, respectively. Liz had means of 1.6, 0, and 4.3 unprompted pretense behaviors prior to instruction and means of 13.7, 19.3, and 18.7 for the final three instructional sessions across Toys Sets 1, 2, and 3, respectively. Brian had means of 1 and 2 unprompted pretense behaviors prior to instruction and means of 19.7 and 12.7 for the final three instructional sessions across Toys Sets 1 and 2, respectively. Thus, substantial transfer of stimulus control occurred from the adult prompts to the materials.

This adds to the literature indicating systematic instructional practices can be embedded into play activities of young children (Filla et al., 1999) and to the growing literature indicating play, including pretend play, can be taught systematically (Lifter et al., 2005). Also, this study expands earlier research on using systematic prompting to teach pretend play (DiCarlo & Reid, 2004; Lifter et al., 1993, 2005). For example, DiCarlo and Reid demonstrated children with disabilities acquired increased rates of unprompted pretense with the introduction of a least-to-most prompting hierarchy. In their study, three participants acquired rates of 0.6, 1.0, and . 10 unprompted pretense behaviors per min during instructional sessions. They did not report measuring pretend play in contexts without prompts. In the current study, Daniel averaged 1.7, 2.3, and 1.8 unprompted pretense behaviors per min during instructional sessions across the three toy sets respectively. Anna averaged 0.5, 0.5, and 0.8, unprompted pretense behaviors per min across the three toy sets respectively. For Anna, these averages take into account the delayed treatment effects with several initial instructional sessions at zero. Liz averaged 1.9, 1.8, and 2.0, unprompted pretense behaviors per min across the three toy sets respectively. Brian averaged 2.7 unprompted pretense behaviors per min with Toy Set 1, and 1.4 with Toy Set 2, before he was abruptly removed from the school. However, it should be noted DiCarlo and Reid defined pretend play with some ambiguity, making interpretations and comparisons difficult.

In addition, the intervention produced consistently higher levels of pretend play in probe sessions without prompting across all children. For example, the levels of pretense behaviors in the probe conditions following intervention for all participants were at least two times the levels in the probe condition prior to intervention, and in most cases close to three times the level. For Toy Set 1 with Daniel, Toy Sets 1, 2, and 3 with Anna, and Toy Set 2 with Liz, unprompted pretend play went from not occurring prior to intervention to occurring in every session of every probe condition after intervention. These findings are important and significant for advancing the literature. Unlike other studies, the maintenance and generalization of pretend play was measured in repeated probe conditions across toy sets. Only three studies have measured pretend play in context without prompts (Kasari et al., 2006; MacDonald et al., 2005; Sherrat, 2002). Severe methodological flaws limit the interpretation of Sherrat's study.

No other studies measured generalization of pretend play concurrently across adults. The generalization across adults of pretend play was measured during probe and instructional conditions. The children showed increased, yet lower, levels in the generalization-across-adult sessions; this has important implications for both practice and research. The nonteacher adult in this study did not prompt or respond to the child's pretense behaviors-suggesting that systematic teaching of pretense behaviors might generalize to settings without a responsive adult. This may have practical value for families. For example, children's ability to engage independently in pretend play may free up time and reduce demands on caregivers.

Second, the participants demonstrated all four types of pretense behaviors (see Table 4) in both intervention and generalization-across-adult sessions, except Anna who did not show assigning absent attributes. Again, in the instructional sessions teachers prompted all four types of pretense, but were free to vary the amount of prompting of each type. The results demonstrate patterns in the types of pretense each child emitted across the toy sets, which is consistent with the pretend play literature. Anna, Liz, and Brian displayed more functional play with pretense than the three substitution behaviors. Conversely, Daniel had higher percentages of object substitution, assigning absent attributes, and imagining absent objects. It is interesting to note that he was the highest functioning (based on mental age and behavioral repertoire) of the four participants. This finding builds on previous research; for example, Lifter et al. (1993) and Lifter et al. (2005) targeted developmentally appropriate play for young children with autism and pervasive development disorders (PDD), respectively. For most children, the targets were functional play with pretense behaviors. Kasari et al. (2006) found that children with autism who participated in a pretend play intervention showed a significant increase in differential unprompted symbolic behaviors (similar to object substitution, assigning absent attributes, and imagining absent objects). However, the overall average play level for the children with autism after treatment was functional play with pretense. Future studies should examine whether a sequence of instruction is more efficient; for example, focusing on functional play with pretense and object substitution and then shifting to assigning absent attributes and imagining absent objects. Perhaps this sequence would produce a more evenly distributed use of the four categories of pretense.

Third, the intervention taught children to exhibit different pretense behaviors during each session, which maintained, at a lower level, in generalization across adult sessions (see Table 5). Further, each child emitted pretend play from all four categories (except Anna, see previous discussion). This diversity maintained across toy sets and probe conditions for all children. This finding suggests the participants were acquiring and generalizing a general class of responding that could be characterized as pretend play. This finding argues for engaging in systematic instruction of play, including pretend play, for children who do not display such behaviors. These results are remarkable and significant for advancing the literature on play in young children. Of the three studies measuring pretend play in contexts without prompts, only one measured the diversity of play behaviors (Kasari et al., 2006). Given the decreased levels of diversity of play behaviors in young children with disabilities, particularly children with autism, when compared to children with typical development, the diversity of play behaviors appears to be an important, if not essential goal for play interventions.

Fourth, the number of vocalizations increased across all participants, although not specifically prompted or reinforced. When the teachers applied the system of least prompts, they modeled and physically prompted actions on toys and described the actions being prompted. This suggests that the combined prompt (i.e., action with the verbal description) may have corollary effects of teaching both the action and vocalizations. The finding also raises the question of whether specific vocabulary (e.g., names of toys or actions) could be taught through the verbal descriptions of the actions. Frey (2006) modeled expansions of children's actions on toys and described the actions; this produced an increase in the number and diversity of actions. She did not measure children's vocalizations. A future study should evaluate whether prompting the action without the verbal description also increases children's vocalizations. Also, the effects of describing and not describing the actions with the prompt should be evaluated.

Fifth, the intervention did not change the number of pretense sequences. The prompts were provided for single pretense behaviors rather than for a sequence of behaviors (e.g., feeding a doll with a block, and then putting the doll to bed and covering it with an imaginary blanket). Future studies should evaluate whether adding themes to the play or systematically prompting sequences would increase the number of pretense sequences.

Sixth, children generalized the pretense behaviors to novel toy sets with a nonteacher adult who did not prompt their behaviors. This occurred both with the generalization toy set and on the posttest of the adapted SPA (see Table 1). Although earlier studies measured generalization across toys (e.g., Lifter et al., 1993; Stahmer, 1995; Thorp et al., 1995), none did so in play sessions without adult prompting. This study adds to the literature by suggesting generalization will occur in contexts where prompts were not used. However, this finding should be interpreted with caution as the generalization toy set was not measured during the initial probe condition. Future studies should evaluate whether this generalization occurs across other times of the day when an adult is not directly with the child (e.g., free play periods).

TEACHERS' USE OF THE INTERVENTION PROCEDURES

A secondary purpose of this study was to evaluate the teachers' acquisition of the intervention procedures (i.e., contingent imitation and system of least prompts). Thus, the study had two levels of independent variables: the teachers' use of the intervention with children, and the investigator's training and feedback to the teachers. Although each teacher received three training sessions staggered across time, the second and third sessions focused on using the previously trained intervention With new materials (toy sets). The current design is weak in terms of drawing conclusions about the existence of a functional relation between the training and the teachers' use of the intervention. However, the measurement of teachers' use of the intervention shows that they used the procedures with fidelity. They made fewer procedural errors as the study progressed across toy sets. Further, the training used in this study is feasible and extends the research on professional development (e.g., didactic training with feedback) to training teachers to use the system of least prompts with pretend play.

It is important to note that although the teachers had at least 3 years of paid experience working with children, they had a variety of formal education backgrounds as is typical in early childhood programs. It is unclear whether the training package would result in less educated or experienced individuals using the intervention with fidelity. Future studies should examine strategies for training less experienced or educated teachers to use systematic procedures for teaching pretend play.

The intervention was complex and the behavioral categories targeted involved subtle discriminations. As a result, the training was designed to be powerful based on the professional development literature (Filla et al., 1999; Moore & Fisher, 2007); it involved a number of components, including a short manual, video examples, focused discussions, practice with nonparticipant children with feedback, a checklist of example behaviors before each intervention session, and feedback after each intervention session. Although each component is a recommended training practice, we cannot identify from this study which (if any) components were responsible for teachers' use of the intervention. Future component analysis studies can identify the active elements, but it is our subjective opinion that the feedback seemed critical.

In addition, future studies should focus on the frequency with which prompts should be used. In this study, teachers were taught to implement a prompt in a window of 12 to 20 s from the last pretense behavior or last prompt. It is unclear whether a larger window (e.g., 30-60 s) would also be effective and easier to implement. This intervention included contingent imitation to focus the teacher's prompting on materials being used by the child with the assumptions that child attention would be maximized and the teacher could teach an additional behavior with the same materials. It is unclear whether these assumptions were met; thus, evaluating the intervention without contingent imitation should be addressed in future studies.

LIMITATIONS

This study has a number of limitations. The sessions were all conducted with one teacher and one child in a designated area of the classroom. The nonteacher adult who conducted most of the generalization-across-people sessions was the same individual across the study; thus, she was not a novel adult by the end of the study. Further, the generalization sessions occurred in the same classroom as the intervention sessions. Generalization across settings was not assessed except for the posttest on the adapted SPA. Likewise, generalization across toys only was measured during the final probe condition to avoid toy satiation. However, this limits comparisons to baseline sessions, prior to intervention. Finally, during the intervention sessions, the number of prompts did not necessarily decrease. Teachers continued to prompt the children frequently. This indicated children had opportunities to engage in more pretense behaviors than they did; however, it is unclear how frequently typically developing children with similar toys would use pretense behaviors. Future studies should measure the social validity of these outcomes by comparing to the frequency and types of pretense behaviors exhibited by peers.

IMPLICATIONS FOR PRACTICE

This study has a number of implications for practice. First, when teachers use the system of least prompts systematically, children may engage in more pretense behaviors of different types, maintain the increased frequency, and generalize it to situations with adults who do not prompt their pretense behaviors. In this study, children were taught to engage in the same behaviors with the same toys as children with typical development. These findings have both functional and practical value for children. For example, this increases the possibility for positive social interactions and communication with peers, provides a context for learning and practicing other skills, and might reduce time spent engaged in maladaptive behaviors. As such, this intervention is recommended for use.

Second, this study provides clear, operationalized categories for pretend play, which were functional for teaching and measuring pretend play. This has important implications for practice and research. Across the pretend play literature, the definition of pretend play was inconsistently and vaguely operationalized, which limits the replication and interpretation of pretend play studies. The taxonomy used in this study can be used in classrooms to measure and teach play and applied in future research.

Third, teachers can be taught to use a complex intervention to promote children's use of pretense behaviors in preschool classrooms. Regardless of their level of education or experience, all four teachers in this study implemented a system of least prompts with high fidelity across three toy sets. Further, they were responsive to the child's behaviors and used appropriate prompts based on the child's behaviors. This is an important finding because most pretend play studies used trained interventionists rather than teachers, and only two of these studies reported the interventionists' experience or training.

In summary, the results of this study provide promising data regarding the systematic prompting of pretense behaviors by preschool teachers in "real world" inclusive classrooms. Teachers implemented the intervention package with high fidelity, and the children engaged in more unprompted pretense behaviors. Specific, intentional interventions, such as the one described here, are necessary to ensure children with special needs engage in pretense behaviors alongside their peers in inclusive, preschool classrooms.

The authors wish to acknowledge the assistance provided by Kristin McCole and Brian Reichow in the conduct of this study. The contributions of Ann Kaiser, Patricia Snyder, and Mary Louise Hemmeter on earlier drafts of this study were appreciated. The study was greatly facilitated by Dr. Ruth Wolery, Director, and staff of the Susan Gray School, Peabody College, Vanderbilt University.

Manuscript received March 2009; manuscript accepted November 2009.

REFERENCES

Barton, E. E., & Wolery, M. (2008). Teaching pretend play to children with disabilities: A review of the literature. Topics in Early Childhood Special Education, 28, 109-125.

Belsky J., & Most, R. (1981). From exploration to play: A cross sectional study of infant free play behavior. Developmental Psychology, 17, 630-639.

Blasco, P. M., Bailey, D. B., & Burchinal, M. A. (1993). Dimensions of mastery in same-age and mixed-age integrated classrooms. Early Childhood Research Quarterly, 8, 193-206.

Casey, A. M., & McWilliam, R. A. (2008). Graphical feedback to increase teachers' use of incidental teaching. Journal of Early Intervention, 30, 251-268.

DiCarlo, C. F., & Reid, D. H. (2004). Increasing pretend toy play of toddlers with disabilities in an inclusive setting. Journal of Applied Behavior Analysis, 37, 197-207.

Doyle, P. M., Wolery, M., Ault, M. J., & Gast, D. (1988). System of least prompts: A review of procedural parameters. Journal of the Association for Persons with Severe Handicaps, 13, 28-40.

Fein, G. G. (1981). Pretend play in childhood: An integrative review. Child Development, 52, 1095-1118.

Filla, A., Wolery, M., & Anthony, L. (1999). Promoting children's conversations during play with adult prompts. Journal of Early Intervention, 22, 93-108.

Frey, J. (2006). Use of play expansions to increase the diversity of object play in young children with disabilities (Unpublished master's thesis). Peabody College, Vanderbilt University, Nashville, TN.

Gast, D. L. (2010). Single subject research methodology in behavioral sciences. New York, NY: Routledge.

Gast, D. L., & Ledford, J. R. (2010). Multiple baseline and multiple probe designs. In D. L. Gast (Ed.), Single subject research methodology in behavioral sciences (pp. 276-328). New York, NY: Routledge.

Ingersoll, B., & Shreibman, L. (2006). Teaching reciprocal imitation skills to young children with autism using a naturalistic behavioral approach: Effects on language, pretend play, and joint attention. Journal of Autism and Developmental Disorders, 36, 487-505.

Jarrold, C., Boucher, J., & Smith, P. (1996). Generativity deficits in pretend play in autism. Journal of Developmental Psychology, 14, 275-300.

Kasari, C., Freeman, S., & Parapella, T. (2006). Joint attention and symbolic play in young children with autism: A randomized controlled intervention study. Journal of Child Psychology and Psychiatry, 47, 611-620.

Kim, Y. T., Lombardino, L. J., Rothman, H., & Vinson, B. (1989). Effects of symbolic play intervention with children who have mental retardation. Mental Retardation, 27, 159-165.

Lieber, J. (1993). A comparison of social pretend play in young children with and without disabilities. Early Education and Development, 4, 148-161.

Lifter, K., Ellis, J., Cannon, B., & Anderson, S. R. (2005). Developmental specificity in targeting and teaching play activities to children with pervasive developmental disorders. Journal of Early Intervention, 27, 247-267.

Lifter, K., Sulzer-Azaroff, B., Anderson, S., & Cowdery, G. E. (1993). Teaching play activities to preschool children with disabilities: The importance of developmental considerations. Journal of Early Intervention, 17, 139-159.

MacDonald, R., Clark, M., Garrigan, E., & Vangala, M. (2005). Using video modeling to teach pretend play to children with autism. Behavioral Interventions, 20, 225-238.

McConnell, S. R. (2002). Interventions to facilitate social interaction for young children with autism: Review of available research and recommendations for educational intervention and future research. Journal of Autism and Developmental Disorders, 32, 351-372.

Moore, J. W., & Fisher, W. W. (2007). The effects of videotape modeling on staff acquisition of functional analysis methodology. Journal of Applied Behavior Analysis, 40, 197-202.

Morrison, R. S., Sainato, D. M., Benchaaban, D., & Endo, S. (2002). Increasing play skills of children with autism using activity schedules and correspondence training. Journal of Early Intervention, 25, 58-72.

Mundy, P., Sigman, M., Ungerer, J., & Sherman, T. (1987). Nonverbal communication and play correlates of language development in autistic children. Journal of Autism and Developmental Disorders, 17, 349-364.

Mullen, E. M. (1995). Mullen Scales of Early Learning (AGS ed.). Circle Pines, MN: American Guidance Service.

Pretti-Frontczak, K., & Bricker, D. (2004). An activity-based approach to early intervention (3rd ed.). Baltimore, MD: Paul H. Brookes.

Sandall, S., Hemmeter, M. L., Smith, B. J., & McLean, M. E. (2005). DEC recommended practices: A comprehensive guide for practical application in early intervention early childhood special education. Longmont, CO: Sopris West.

Sandall, S., & Schwartz, I. S. (2002). Building blocks for teaching preschoolers with special needs. Baltimore, MD: Paul H. Brookes.

Sherrat, D. (2002). Developing pretend play in children with autism. Autism, 6(2), 169-179.

Stahmer, A. (1995). Teaching symbolic play skills to children with autism using pivotal response training. Journal of Autism and Developmental Disorders, 25, 123-141.

Strain, P. S. (1985). Social and non-social determinants of acceptability in handicapped preschool children. Topics in Early Childhood Special Education, 4, 47-58. Tapp, J., Ticha, R., Kryzer, E., Gustafson, M., Gunnar, M. R., & Symons, E (2006). Comparing observational software with paper and pencil for time-sampled data: A field test of Interval Manager (INTMAN). Behavior Research Methods, 38, 165-169.

Tapp, J., & Walden, T. (2000). PROCODER: A system for collection and analysis of observational data from videotape. In T. Thompson, D. Felce, & F. J. Symons (Eds.), Behavioral observation: Technology and applications in developmental disabilities (pp. 61-70). Baltimore, MD: Paul H. Brookes.

Thorp, D., Stahmer, A., & Schreibman, L. (1995). Effects of socio-dramatic play training on children with autism. Journal of Autism and Developmental Disorders, 25, 265-282.

Ungerer, J., & Sigman, M. (1981). Symbolic play and language comprehension in autistic, mentally retarded, and normal children. Developmental Psychology, 20, 293-302.

Wolery, M. (2001). Embedding constant time delay in classroom activities. Young Exceptional Children Monograph, 3, 81-90.

ERIN E. BARTON

MARK WOLERY

Vanderbilt University

Erin E. Barton is now Assistant Professor, Department of Special Education and Clinical Sciences, University of Oregon, Eugene.

Correspondence concerning this article should be addressed to Erin E. Barton, 5261 University of Oregon, Department of Special Education and Clinical Sciences, Eugene, OR, 97403-5253 (e-mail: ebarton@uoregon.edu).

ERIN E. BARTON (Oregon CEC), Doctoral Student, Department of Special Education, Peabody College, Vanderbilt University, Nashville Tennessee. MARK WOLERY (Tennessee CEC), Professor, Department of Special Education, Vanderbilt University, Nashville, Tennessee.
TABLE 1
Demographic and Assessment Data for Child Participants

                  Chronological   Mental Age
Child    Gender   Age in Months   in Months (a)   Disability

Daniel     M           43             36          Language delay
Anna       F           43             23          Autism
Liz        F           50             27          Developmental delay
Brian      M           30             18          Autism

         Pretest. Structured Play    Posttest Structured Play
         Assessment (b)              Assessment (b)

                       # Different                  # Different
         # Different   Unprompted    # Different     Unprompted
         Unprompted     Symbolic      Unprompted      Symbolic
Child        FPP        Behaviors        FPP         Behaviors

Daniel        2             0         3             18
Anna          0             0        15              2
Liz           0             0        16             12
Brian         0             0        Not assessed   Not assessed

Note. FPP = Functional play with pretense.

(a) Mullen Scales of Early Learning. (b) Structured Play Assessment
was adapted for this study based on Ungerer & Sigman (1981).

TABLE 2
Toys by Type of Toy Set

                         Instructional Toy Sets

Toy Set 1                Toy Set 2                Toy Set 3

2 baby dolls             doll house               2 adult figures
2 bottles                adult male figure        2 sponges
2 sippy cups             adult female figure      4 plates
2 sponges                baby figure              4 bowls
2 plates                 cat figure               2 teaspoons
2 sheets of paper        dog figure               2 forks
2 bowls                  2 bear figures           2 serving spoons
2 pieces of ribbon       doll house fence         2 spatulas
2 square blocks          2 doll house chairs      2 ladles
2 wooden rods            doll house bed
2 small lotion bottles   baby stroller
                         rubber bands
                         4 toothbrushes
                         2 sheets of card stock
                         2 paper coin rollers
                         2 container lids

Generalization Toy Set   SPA Toy Set

2 baby dolls             2 baby dolls
2 cloths                 doll house
2 bowls                  2 plates
2 cups                   2 cups
2 sponges                2 sippy cups
2 large spoons           2 stacking rings
2 trays                  1 bottle
2 adult figures          3 wooden rods
4 animal figures         2 sheets of paper
2 square blocks          2 cars
                         2 small telephones
                         2 brushes
                         2 sponges
                         2 square blocks
                         2 rectangle blocks

SPA = Structured Play Assessment.

TABLE 3
Percentage of Interobserver Agreement for Teacher and Nonteacher
Adult Generalization Sessions by Child Behavior

                  Condition (Sessions With
                  Teacher)

Participant       Probe          Instructional
Behavior          Mean (range)   Mean (range)

Daniel
  Prompted        100            95 (93-100)
  Unprompted      97 (95-98)     95 (93-100)
  Type            93 (85-100)    96 (95-100)
  Same/Differ     96 (88-100)    97 (90-100)
  Sequences       100            100
  Vocalizations   95 (78-100)    96 (83-100)
Anna
  Prompted        96 (93-100)    94 (80-100)
  Unprompted      98 (93-100)    95 (88-100)
  Type            96 (90-100)    97 (93-100)
  Same/Differ     91 (87-100)    97 (90-100)
  Sequences       100            100
  Vocalizations   100            100
Liz
  Prompted        87 (78-92)     88 (80-92)
  Unprompted      99 (97-100)    90 (82-95)
  Type            100            89 (82-100)
  Same/Differ     100            88 (77-95)
  Sequences       100            100
  Vocalizations   100            100
Brian
  Prompted        89 (78-100)    97 (88-100)
  Unprompted      98 (90-100)    95 (85-100)
  Type            97 (88-100)    93 (88-100)
  Same/Differ     100            100
  Sequences       100            100
  Vocalizations   100            100

                  Condition (Generalization
                  Sessions With Nonteacher Adult)

Participant       Probe          Instructional
Behavior          Mean (range)   Mean (range)

Daniel
  Prompted        100            100
  Unprompted      100            100
  Type            100            99 (91-100)
  Same/Differ     100            91 (80-95)
  Sequences       100            100
  Vocalizations   100            100
Anna
  Prompted        100            100
  Unprompted      100            96 (80-100)
  Type            100            100
  Same/Differ     93 (89-99)     100
  Sequences       100            100
  Vocalizations   100            90 (82-98)
Liz
  Prompted        100            100
  Unprompted      95 (94-100)    91 (85-100)
  Type            93 (90-100)    91 (80-100)
  Same/Differ     92 (89-100)    90 (80-100)
  Sequences       100            100
  Vocalizations   100            100
Brian
  Prompted        100            100
  Unprompted      100            94 (80-100)
  Type            100            91 (84-100)
  Same/Differ     100            91 (82-100)
  Sequences       100            100
  Vocalizations   100            100

TABLE 4
Percentage of Each Type of Unprompted Pretense Behaviors by Toy
Set During Intervention Conditions

                         Toy Set 1            Toy Set 2
                        Type Pretense       Type Pretense
Participant
Condition          FPP   OS   IAO   AAA   FPP   OS   IAO   AAA

Daniel
  Intervention       1   73    20     5     4   34     3    59
  Generalization     2   79    18     1     2   26     0    72
Anna
  Intervention      82   18     0     0    47   53     0     0
  Generalization    89   11     0     0     0    0     0     0
Liz
  Intervention      72   16     0    12     5   55     8    32
  Generalization    95    2     1     2     0   36     0    64
Brian
  Intervention      72   18     5     5    40   24     1    35
  Generalization    70   25     5     0    53   27     0    20

                          Toy Set 3
                        Type Pretense
Participant
Condition          FPP   OS   IAO   AAA

Daniel
  Intervention       8   39    23    30
  Generalization    42   42    16     0
Anna
  Intervention      78   22     0     0
  Generalization    80   10     0     0
Liz
  Intervention      74   17     2     7
  Generalization    90   10     0     0
Brian
  Intervention      --   --    --    --
  Generalization    --   --    --    --

Note. FFP = functional play with pretense; OS = object substitution;
IAO = imagining absent attributes; AAA = assigning absent attributes.

TABLE 5
Number of Different Unprompted Pretense Behaviors by Probe Condition
for Teacher and Generalization Sessions Across Toy Sets

                      Initial  Probe            Second Probe

Participants     Instruct.         Gen.           Instruct.
Toy Set        Mean (range)    Mean (range)     Mean (range)

Daniel
  Toy Set 1    0                .66 (0-1)      7.3 (6-9)
  Toy Set 2     .67 (0-1)      0               2.0
  Toy Set 3     .33 (0-1)      0                .67 (0-1)
Anna
  Toy Set 1    0               0               6.67 (5-14)

  Toy Set 2    0               0               0
  Toy Set 3    0               0               0
Liz
  Toy Set 1    1 (0-2)         4 (2-6)         7 (5-10)
  Toy Set 2    0               1.7 (1-3)       0
  Toy Set 3    1.3 (1-2)       2.7 (1-4)       1 (0-3)
Brian
  Toy Set 1     .33 (0-1)      3 (2-4)         7.67 (4-11)
  Toy Set 2    0               1.7 (0-3)       1
  Toy Set 3    0               1.7 (0-4)       0

                Second Probe              Third Probe

Participants        Gen.           Instruct.           Gen.
Toy Set         Mean (range)     Mean (range)      Mean (range)

Daniel
  Toy Set 1    3.33 (2-5)       4.3 (4-5)         3.67 (2-6)
  Toy Set 2    1.0              8.3 (6-11)        4.0
  Toy Set 3    1.5 (1-2)        2.6 (1-4)          .33 (0-1)
Anna
  Toy Set 1    4                7.67 (6-9)        --

  Toy Set 2    0 MR             6                 3.33
  Toy Set 3    0                0                 0
Liz
  Toy Set 1    4 (3-5)          5.67 (4-8)        5 (4-5)
  Toy Set 2    0                8.67 (6-11)       6
  Toy Set 3    0                1.67 (0-5)        0
Brian
  Toy Set 1    4.33 (3-7)
  Toy Set 2     .67 (0-1)
  Toy Set 3     .33 (0-1)

                       Final Probe

Participants     Instruct.          Gen.
Toy Set         Mean (range)    Mean (range)

Daniel
  Toy Set 1     8.0 (7-9)       1.5 (1-2)
  Toy Set 2    10.3 (3-18)      4.3 (2-7)
  Toy Set 3     5.0 (1-9)       1 (0-2)
Anna
  Toy Set 1     7 (6-7)         --

  Toy Set 2     6               6
  Toy Set 3    11 (5-12)        5
Liz
  Toy Set 1    10 (6-13)        --
  Toy Set 2     3 (2-4)         --
  Toy Set 3     5.67 (5-7)      7
Brian
  Toy Set 1
  Toy Set 2
  Toy Set 3

Note. Dark lines on table show where intervention occurred.
COPYRIGHT 2010 Council for Exceptional Children
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2010 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Barton, Erin E.; Wolery, Mark
Publication:Exceptional Children
Geographic Code:1USA
Date:Sep 22, 2010
Words:11574
Previous Article:Incorporating student input in developing alternate assessments based on modified academic achievement standards.
Next Article:Planning instruction and self-regulation training: effects on writers with autism spectrum disorders.
Topics:

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters