Effects of an intervention and video interaction counseling with teachers of 78 children from school and centers were investigated. The study involved a quasi-experimental pretest-intervention–posttest control group design. In Condition 1 teachers received the counseling for 12 weeks, followed by withdrawal of this intervention for weeks. In Condition 2 the intervention was in reversed order. The students' competent behavior was assessed by rating their response to social support during engagement in a standardized task, their performance of standardized tasks of cognitive skills, and ratings of their social skills. Results revealed differences in favor of the participants in the two video interaction counseling research conditions on measures of social support and cognitive competence.
In The Netherlands, children with intellectual deficiencies are increasingly included in mainstream schools (a rise from 24% to 45% within 4 years). These placements are often rescinded prior to the transfer of the students to Grade 3 (comparable to Grade 2 in the United States) because social and emotional problems arise in addition to lack of developmental progress (Scheepstra, Pijl, & Stevens, 1997). For integration in the primary school to succeed, professional educators need to have a large body of knowledge about the possible benefits of social interaction for these children. In this study we address the issue of the impact of one important classroom contextual variable, consultative feedback to teachers regarding their videotaped interactions and communications in meeting the needs of their students with intellectual disabilities.
In contemporary views intellectual impairment is defined as a characteristic of the actual way an individual functions within the context of society. For example, Buntinx and Bernhard (1996) interpreted the most recent definition of the American Association of Mental Retardation (now the American Association on Intellectual and Developmental Disabilities—AAIDD) as follows: When individuals have intellectual disabilities, there are substantial impairments in their current functioning that exist concurrently with associated impairments in two or more of the following adaptive domains: Communication, Ability to Live Independently, and Social Competences accompanying some other types of behavior. In this most recent definition of intellectual handicap, the significance of social interaction is acknowledged. The abilities of the individual are described as are the supports that the individual needs from the environment to enable him or her to interact in interpersonal communications and learn the social skills that are required in order to produce competent behavior within that social context (Buntinx & Bernhard, 1996; Merrell & Gimpel, 1998; Polloway, Smith, Chamberlain, Denning, & Smith, 1999; Schalock et al., 1994).
In our study, which includes young children with intellectual impairments within an educational setting, we accept this interpretation. When we use the term competence, we mean all of the cognitive and social skills that can be observed within the classroom, such as the ability to (a) perform a task that matches the child's abilities, (b) seek and maintain contact and communicate with peers, (c) profit from assistance with a task that's presented in the classroom, as well as (d) comprehend and engage in communication with adults in that setting.
The principle concern of those who nurture and rear children with intellectual impairments is the need to address and challenge the capacities of the child. This requires offering these children the opportunities that reflect and stimulate their developmental competencies (Guralnick & Groom, 1987; Nakken & Vriesema, 1993). The developmental model involves finding a balance between the child's current abilities and the demands made by the environment that reflects the understanding that children with an intellectual disability in principle go through the same sequences of social development as do those with typical development, though with slower speed and a lower ceiling (Van Berckelaer-Onnes, 1993).
Two basic ideas have governed research concerning interventions with children who have intellectual impairments. The first is the deficit hypothesis, suggesting that children with intellectual disabilities have something missing or wrong with their general development. From this perspective, interventions that are most effective are those that emanate from the environment and directly target specific behaviors that need to be learned or increased. These types of interventions include prompting and positive empowerment in combination with instruction and feedback to the teacher. Meta-analyses involving these techniques have shown that generalization of training effects more frequently occur when a phase is added to the design when the technique is phased out, if the intervention period is increased, and if the children participate in the intervention at younger ages (Beelmann, Pfingsten, & Losel, 1994; Chandler, Lubeck, & Fowler, 1992; Odom & McConnell, 1996).
The second model is the developmental hypothesis. Proponents of this hypothesis believe that children with intellectual impairments do not develop differently but, rather, more slowly than do those with a typical pattern of development. Supporters of this view have observed that interventions such as social skills training fail to eliminate the child's deficits or impairments. The child instead becomes more dependent upon the environment. According to Van Gemert (1998), a so-called “handicapism” arises whereby the nurturer is less likely to make an effort to teach the child typical skills, which has the consequence of increasing the developmental deficit of the child. This hypothesis explains the frequently passive attitude of children with intellectual impairments (Rosenthal & Klein, 1999). Studies in which the developmental hypothesis was tested have included interventions in which the investigation consisted of attempts to speed up the development of the child by improving the quality of the environment; that is, the environment, rather than specific behaviors of the child, was the primary source of manipulation. For example, effects of video self-modeling and self-assessment training revealed that when inappropriate teaching behavior was reduced, students made more choices and increased their active involvement in tasks (Reamer, 1995). Another example is the increased value placed on integrated care in schools (Perrig-Chiello, 1999; Sims, Hutchins, & Taylor, 1997).
Both the deficit and development hypotheses ignore the possible influences of the child on his or her own development. According to Waber et al. (2003), the child's complement of skills, as a biological risk, can interfere with his or her ability to adapt successfully to the demands and expectations of the educational context. This, then, is the third or adaptation hypothesis. This hypothesis refers to the systemic model that emphasizes the dynamic developmental interaction between the child and the process of schooling. From this point of view, developments are the result of dynamic, ongoing, reciprocal processes that occur between the child's genetic propensities and environmental influences that either enhance or obstruct the child's abilities, regardless of any limitations of resources or presence of special needs. Improving the environment would then be expected to increase opportunities for the child to experience successful interactions with the teacher and, thereby, enable the child to use and further develop abilities.
In this study we tested the adaptation hypothesis. We expected that teachers of children with intellectual impairments would profit from suggestions offered to help them improve their communications with their students and that the students, in turn, would increase their social and cognitive competence in response to their teachers' improved communications.
Participants were 78 children from two types of care settings. Ethnic backgrounds included Dutch, Turkish, Moroccan, Surinam, and “others,” which represent the typical population in The Netherlands. The first type of setting was four special schools for children with intellectual impairments; 43 children, whose average age was 75.8 months (SD = 14.74), from these settings were all in the first grade. The criteria for placement of students in these school were that they were at least 4 years of age and showed evidence of cognitive (IQ < 60) or social (below average on a standardized measure of social behavior) impairment, which may be combined with documentation of disorders of language and/or motor skill or of pervasive developmental disorder as described in the Diagnostic and Statistical Manual of Mental Disorders, 4th ed., revised—DSM-IV-TR (American Psychiatric Association, 2001). The mean group size for these students was 8 to 12.
The second type of setting was special care centers; 35 children from preschool groups whose average age was 59.1 months (SD = 11.14). The criterion for placement of these children was primarily evidence of multiple handicaps, often including severe cognitive impairment in combination with motor skill impairments that required physical care in an institution. The mean group size for these students was 6 to 8.
The special schools were recruited through the special education training course of the Fontys Teacher Education School, Tilburg. The 4 participating teachers had completed the Teacher Training College for Primary School Teachers as well as a special education training course. We recruited participants from the special care centers in the cities in which the schools were located through direct contact with the directors of the institutions. The 4 caretakers from the special care centers were educated and trained at the level of intermediate professional and vocational education.
Because it was not possible to assign participants randomly to the groups, we used a quasi-experimental design. In each institution all of the students of the participating teacher were included. The design consisted of a pretest, intervention, and posttest as well as a nontreatment control group. Video interaction counseling was the independent variable (see description below). The independent variables were type of care (special school vs. special care), order of intervention (video interaction counseling followed by no video interaction counseling vs. no video interaction counseling followed by video interaction counseling), socioeconomic status (SES), and pedagogic quality of the institutions. Socioeconomic status and pedagogic quality of the institutions were measured in order to establish post-hoc whether the intervention effect could have been influenced by the sociocultural circumstances in which the children were being raised and where they received their education. The video interaction counseling intervention for these institutions was conducted as follows: Condition 1—video interaction counseling for 12 weeks following the first measurement (Time 1). The second measurement (Time 2) was administered, followed by a period of 12 weeks without video interaction counseling. This condition was concluded with administration of the third measurement (Time 3). In Condition 2, the order was reversed. Table 1 illustrates the design of the study.
The research task for this study was the Jiggling Tower game, which was provided by the researchers. This game (generally available in stores) consists of a plastic tower with removable round base; ledges colored blue, red, green, or yellow; and plastic dolls in these same colors. The throw of a colored dice determines which doll can be placed on the tower. The goal is to place the dolls so that they balance and do not topple the tower.
The researchers took the tower apart while the teacher observed how the ledges and the round base were put together. The teacher was free to decide to use the game rules or to deconstruct the tower in case she considered the balancing part as too difficult for the children. The task was played for 10 min and recorded on video. This activity took place in a separate room of the institution, where it was administered to small groups of 3 children chosen randomly from the classroom. Once chosen, these groups were maintained for each of the total of three measurements with the same task. Depending upon the size of the class, it was possible that the teacher repeated this task up to three or four times to assure participation of all the students of her class. This task was not used other than for pretest and posttests.
Video interaction counseling
The decision to use video interaction counseling reflects the assumption that a child's intentions and willingness to learn are enhanced by being given opportunities to think and reflect on their actions rather than being subjected to passive guidance (Janssens & Kemper, 1996). The video interaction counseling intervention consisted of provision of feedback to the teachers according to a standardized format by a counselor. He or she was a professional trainer who was certified by an agency that offers training on how to consult with teachers by video interaction counseling. The consulting therapists were familiar with conducting video interaction counseling with teachers from regular primary schools. They had no experience with teachers who instruct children with intellectual impairments. The therapists were aware of reciprocal processes during social interactions but mainly those that were concerned with teacher– child interactions. Therefore, to test the validity of the intervention, we included data collection with respect to the video interaction counseling sessions.
The content of the feedback was based on videotapes of the teacher engaged in a classroom activity with the children (Van den Heykant, 1997). A typical video interaction counseling session required approximately one hour and proceeded as follows: The counselor invited the teacher to watch a selection of the video relating to the activity. After several minutes of watching and offering opportunity for comment, the counselor provided an explanation of examples of optimal communication, which was defined as verbalizations or actions that were adapted to the needs expressed by the children during the interaction. An example selected from an actual session was when a teacher offered a die to one of the children in the group, waited for eye contact, and then said: “Look ___ (child's name); this is the die that we need for the game.” Another example was when a teacher waited until all the children watched her and then suggested that the game could start. A third example was when a teacher waited for a reaction from the child after encouraging him or her to take a turn. Stills of the video recording were used to illustrate instances of optimal interaction.
Next, a few examples of suboptimal interactions were reviewed, and the counselor discussed with the teacher how the interaction could be improved. An example selected from an actual session involved a teacher who continued to talk to a student who failed to respond to her initial overtures with eye contact. The counselor then reviewed the following points to increase opportunities for more optimal interactions: follow the student's initiative, try to match the student with tone of voice and pacing of language utterances, respond to the student's initiatives with confirmations that they were noticed and understood, provide names for actions and feelings, promote turn-taking, and provide opportunities for students to share leadership of the group activities. The counselor and the teacher then discussed possibilities for using these suggestions, and the teacher was invited to choose two or three of these to implement during the next 2 weeks.
The researchers attended the first counseling session of each participating teacher to ensure that video interaction counseling was implemented as intended. The researchers also discussed all the video interaction counseling sessions with the counselors following the intervention as further assurance of correct implementation. The first video interaction counseling material that was used was the recording of the first small group task. After the first video interaction counseling meeting, 2 weeks of practice followed, and then a new video recording was made. Video interaction counseling was provided in this way prior to the first week and during the 3rd, 5th, and 9th weeks of the intervention period.
Small group tasks/games
Playing games is part of the regular program of both the special schools and centers. Small group games can either have rules, such as Memory and Bingo, or be creative activities, such as painting, singing, or sports. None of these games was comparable to the research task. Prior to the intervention, we asked the teachers to provide lists of games they already played in their classes. From these we composed one master list that was provided to all of the participating teachers the week before the intervention. Both research conditions included games to be played four times per week in small groups by any 3 students and their teacher. The teachers were asked to make a list of the names of the games they selected as well as the names of the students involved, to ensure that all of the students participated with equal frequency during the first and second intervention periods.
During each intervention period, videos of four small group games were made; the first, after completion of the research task. For this, the teacher was asked to choose a game from the master list. This game was presented to one of the groups who had already completed the research task. The regular game was then played for a maximum of 10 min and recorded on video. In the video interaction counseling condition, the second video was made during the 3rd week, the third video during the 5th week, and the fourth video during the 9th week. For the intervention condition without video interaction counseling (control), three small group games were videotaped, the first after completion of the research task, the second during the 5th week, and the third during the 9th week. By recording videotapes of all conditions, we assured that the small group games were being used.
We started this study at the beginning of the school year. Two institutions from each type of care were randomly chosen to be the first to receive video interaction counseling. The other two institutions were used for the control condition. After the first intervention, the intervention conditions were reversed.
The SES of the participants was defined by determination of the levels of education and occupations of their parents. The SES scores were ranked and summed using the software program International Socio-Economic Index of Occupational Status (Bakker, Sieben, Nieuwbeerta, & Ganzeboom, 1997), which provides weights for the education and occupation of both mothers and fathers, with a maximum score of 14 per child.
The pedagogical quality of each institution was determined from scores on the Early Childhood Environment Rating Scale. This instrument, which was translated into Dutch by Reiling, Verhoeven, and Tavecchio (1995), consists of a total of 37 items from seven categories relating to school activities and teacher–student interactions, specifically, personal care routines (5 items), furnishings and displays for the children (7 items), language–reasoning experiences (5 items), fine and gross motor activities (6 items), creative activities (8 items), social development (6 items), and adult needs (4 items). The observer rates each item of a scale ranging from 1 to 7, with totals varying from 42 to 294. To check the reliability of the observations made for this study, two different observers visited two classrooms twice. The Pearson correlations for these observations were .87 and .84, showing a satisfactory level of agreement. Table 2 contains the data for the contextual variables discussed above by types of care groups.
We performed tests with the data from Table 2 for both conditions as independent groups to establish whether the conditions were sufficiently comparable to be combined and treated as one research group. The results were significant for age, T(71.56) = 5.58, p = .000, and pedagogical quality, T(76) = −9.63, p = .000. Because of the statistical significance of these differences, we were not able to combine the data, which are, therefore, described separately.
Small Group Measures of Competence
Data for documenting competence based on observations of students engaged in small group tasks were collected from videotapes made of the teachers supervising their students in these activities. The three variables assessed using this approach were social awareness, social support, and response to social support.
Social awareness (
Van der Aalsvoort & Gossé, 1999)
This variable was represented by a score that reflected occurrences of social awareness per half min of videotaped observation. The child could score a maximum of four points for each session based on the following four categories: (a) The student is oriented towards interaction with the other students and/or the teacher and visibly joins in the tasks (4 points). (b) The student is attentive in response to stimulation by the teacher (3 points). (c) The student is not attentive despite stimulation by the teacher (1 point). (d) The student is not attentive and the teacher offers no stimulation to solicit attention (0 points). The total scores for each small group session were used for the statistical tests. The internal consistency of the instrument was determined to be sufficient at alpha .79.
Social support (
Erickson, Sroufe, & Egeland, 1985)
This variable was derived from the subscales of Supportive Presence, Respect for Autonomy, Structure, Hostility, and Quality of the Instruction of Social Support. A score could be assigned for each subscale that ranged between 1 and 7, with a total that ranged between 7 and 35 per child. The scores for each small group were summed, and this total score was used for the statistical tests. The internal consistency of this instrument was determined to be sufficient at alpha .89.
Response to Social Support (
Erickson et al., 1985)
The subscales Avoidance, Reliance on Help, Perseverance, and Compliance were used from Response to Social Support. Each subscale resulted in a score between 1 and 7, with a total score of 4 to 28 per child per measurement. The scores from the students in the small groups were summed. The resulting score was included in the statistical tests. Internal consistency of the instrument was sufficient, alpha = .89.
Interassessor reliability was determined for each instrument as well. There were three observers, each responsible for one instrument. Of the 69 video recordings of the administration of the research task, we chose 8 at random (25%) for each instrument. Two observers scored each tape independently. Reliability was calculated with Pearson correlation coefficients. For the scales Social Support and Response to Social Support, the mean coefficient was .81, SD = .13, with a range of .51 to .96. For Social Awareness, we used Cohen's Kappa, with an average of .81 and a range from .67 to .87. The interassessor reliability of these measures was, therefore, shown to be adequate.
Individual Measures of Competence
We used two individual measures of competence: the PsychoEducational Profile-Revised (Van Berckelaer-Onnes, van der Ploeg, & Dijkxhoorn, 1995a) and the Vineland School Observation Scales (Van Berckelaer-Onnes, van der Ploeg, & Dijkxhoorn, 1995b). The first instrument includes the following components: Perception, Fine Motor Skills, Gross Motor Skills, Hand–Eye Coordination, Performance, and Verbal Ability. The second author assessed all of the students, assigning one of the following evaluations for each item: successful, with help, and failed. Only the items with a successful score were included in the data analysis. We evaluated the internal consistency of this test because it has not yet been standardized for the Dutch population (Krayer & van der Plas, 1997). From the data-reduction analysis, it appeared that one factor could be inferred from the seven developmental domains, meaning that the raw scores for the whole test could be summed and transformed into percentages. These scores ranged from 1 to 7 for each child for each measure. The total score was used for statistical analysis. The internal consistency of the instrument was found to be adequate at alpha .94.
The Vineland School Observation Scales (the Dutch translation of the Vineland Adaptive Behavior Scales; Van Berckelaer-Onnes et al., 1995b). Teachers were asked to complete this instrument because they were the individuals who were the best informed about the daily functioning of the children. The Vineland contains questions about the child's functioning in four developmental domains: Communication, Daily Living Skills, Socialization, and Motor Skills. Because the instrument has not yet been standardized for the Dutch population, we conducted a data-reduction analysis and one factor remained. Therefore, for this test, we summed the raw scores for each domain and then transformed them into percentages. The scores ranged from 1 to 4 for each child for each domain. The total scores were used in the statistical analysis. Internal consistency was found to be adequate at alpha .90.
The parents of the participating students were asked to give their consent for their child's inclusion in the study prior to the intervention. Information about the education and occupation of both parents was obtained at this time. They were also asked for their approval for video recordings for video interaction counseling and were assured by formal letter that the sessions would be supervised and the tapes protected by the researcher. Following an introductory conversation with the head of the institution and, as much as possible, with the teachers involved, the pedagogic quality of the institution was determined with the Early Childhood Environment Rating Scale. For that purpose, one visit was made to each institution by the researchers to collect the individual assessments for each child. The administration of the PsychoEducational Profile-Revised required approximately 60 minutes per child. During the week that the children were tested, the teachers completed the Vineland School Observation Scales. The intervention period then began. For Condition 1, the video interaction counseling was based on the first videotape of a small group game. Participants in all research conditions played games four times per week. When we noticed that not all of the students were able to participate with equal frequency each week, we reviewed the lists of activities carried out at the end of each week to be sure that every student was included at least once a week. The many events that occurred during a period of 12 weeks, however, often resulted in the teachers not adhering to the schedule. Also, in all of the conditions, many students were taken out for individual therapy sessions, which made it difficult to plan the small group activities. Moreover, special events such as the appearance of Santa Claus, birthday celebrations, and school festivities made it difficult to organize the sessions four times a week. In some institutions, the desired number of group activities was achieved after 12 weeks, whereas in others it took 15 weeks. Therefore, we decided to start the second round of measures after completion of 48 small group games per institution, whenever this occurred. Once completed, the Condition 2 intervention took place. One of the special schools withdrew after Condition 1 because the teacher went on maternity leave and her substitute did not wish to participate in the study. Therefore, the data from this institution were not available for the second intervention.
Results for the Special Schools
Conditions 1 and 2 were compared with respect to gender, age, nationality, and pedagogical quality. There were no significant differences between the two groups; therefore, any remaining effects from this study are attributed to the intervention. Table 3 contains the means and SDs for the five measures of competence.
A two-way within-subjects analysis of variance was conducted to evaluate the effect of video interaction counseling and intervention on competence. The dependent variable was competence. The within-subjects factors were intervention (video interaction counseling or no video interaction counseling) and three levels of time (Time 1, Time 2, and Time 3). We used a MANOVA to test the time main effect and the Intervention × Time interaction, using the multivariate criterion of Wilks' Lambda (Λ). The factor Time was significant for Social Support, Λ = .18, F(2, 23) = 54.22, p = .000, multivariate η2 = .83, Cognitive Competence, Λ = .11, F(2, 22) = 86.56, p = .000, multivariate η2 = .89, and Social Competence, Λ = .43, F(2, 21) = 14.47, p = .000, multivariate η2 = .58. There was also an interaction between the factor Time and Intervention on Social Support, Λ = .12, F(2, 23) = 81.67, p = .000, multivariate η2 = .88, and Cognitive Competence, Λ = .54, F(2, 22) = 9.30, p = .001, multivariate η2 = .49. The results did not support the hypothesis that the students showed either more social awareness or response to social support than did their peers after video interaction counseling.
Three paired t tests were conducted to follow up the significant interaction. We controlled for family-wise error rate across these tests, using Holm's sequential Bonferroni approach. Differences in mean scores between the two conditions with respect to social support were significant between Times 1 and 2, t(25) = −3.41, p = .002, and between Times 1 and 3, t(27) = −10.70, p = .000. These results support the hypothesis that teachers in Condition 1 offered more support than those in Condition 2 after Time 2, but the results do not support the same hypothesis for Condition 2.
Differences in mean scores between the two conditions with respect to cognitive competence were significant between Times 1 and 2, t(33) = −13.78, p = .000, between Times 2 and 3, t(27) = −.6.29, p = .000, and between Times 1 and 3, t(24) = −10.88, p = .000. Differences in mean scores between both conditions with respect to social competence were significant between Times 1 and 2, t(32) = −4.38, p = .000, between Times 2 and 3, t(27) = −3.22, p = .003, and between Times 1 and 3, t(23) = −4.84, p = .000. These results support the hypothesis that the students showed more cognitive and social competence after video interaction counseling.
Results for the Special Care Centers
We first tested whether the background information for both groups was comparable. Table 4 includes the means and SDs for these instruments. There were significant differences between the group with regard to SES, t(25) = −.827, p = .009, and pedagogical quality, t(32, 169) = −2.84, p = .008. This means that any intervention effects can also be attributed to differences between the groups that existed at the start of the intervention. With respect to SES, it is possible to conduct an analysis with this variable as a covariate; however, pedagogical quality can not be included as a covariate because the resulting scores would be attributable to the institutions rather than to the individual participants.
We tested the effect of video interaction counseling using MANOVA, with the factors Time (three measurement points) and intervention (Conditions 1 and 2) as indicators of competence. We tested the factors Time and Intervention condition using the multivariate criterion Wilks' Lambda (Λ), with SES as a covariate. Table 5 includes the means and SDs for both conditions.
The factor Time was significant for social support, Λ = .56, F(2, 18) = 6.99, p = .006, multivariate η2 = .44, and cognitive competence, Λ = .50, F(2, 20) = 10.03, p = .001, multivariate η2 = .50. The results did not support the hypothesis that the students would show more social awareness, response to social support, and social competence than would their peers after video interaction counseling. Three linked t tests were performed to establish the significant interaction. We controlled for errors by using Holm's sequential Bonferroni approach. Differences in mean scores between the two conditions with respect to social support were significant between Times 1 and 2, t(25) = −3.41, p =. 002, and between Times 1 and 3, t(26) = −10.38, p = .000. Differences in mean scores between both conditions with regard to cognitive competence were significant between Times 1 and 2, t(33) = −13.78, p =. 000, between Times 1 and 3, t(24) = −10.88, p = .000, and between Times 2 and 3, t(27) = −6.29, p = .000. These results support the hypothesis that the teachers offered more social support and that the students showed increased cognitive competence after video interaction counseling.
In this study we investigated the effects of videotaped feedback to teachers of 78 Dutch children with intellectual impairments on the cognitive and social functioning of the children. The study was designed to test the adaptation hypothesis that emphasizes the influence of dynamic interaction between the child and the educational context on children's development. We expected that the teachers would profit from the suggestions offered to help them improve their communication with their students. We also anticipated that an optimized teaching environment would result in an increase in the social and cognitive skills of the students.
The results supported the hypothesis for the teachers in the special schools in Condition 1 (video interaction counseling followed by no video interaction counseling), who offered their students more support than the teachers in Condition 2 (no video interaction counseling followed by video interaction counseling) after Time 2. The results did not support the same hypothesis for Condition 2 not the hypothesis that the students in the special schools and special centers would show either more social awareness or response to social support following video interaction counseling. This suggests that increased teacher support during small group tasks did not affect student behavior during the research task. It is possible that the lack of effect may have resulted from the fact that the teachers were already familiar with engaging their students in small group activities and did not need further support. Although the teachers confirmed before the study that they were familiar with this approach, the frequency of the activity each week seemed to introduce a novel variable that may have confounded the video interaction counseling intervention. The children and the teachers in Condition 2 may have practiced longer to achieve optimal interaction, which would have reduced the potential effect of video interaction counseling. Although the increase in the scores between measurements may reflect maturation because the children were followed over the period of a school year, we cannot conclude this from the findings. The scores obtained were raw scores; there are no norm tables for the instruments. There may also have been a ceiling effect for the measure used to assess cognitive competence (PsychoEducational Profile-Revised): The maximum score can only be obtained if the child can read and write, which is highly improbable for these students.
Significance of the Results for the Care of Children With Intellectual Impairment
In The Netherlands, children with intellectual deficiencies are increasingly included in mainstream schools (a rise from 24% to 45% within 4 years). For integration in the primary school to succeed, professional educators need to have a large body of knowledge about the possible benefits of social interaction for these children. Our findings show that it may be easy for teachers in schools for students with intellectual impairments to elicit competent behavior merely by offering small group tasks four times a week because even without feedback to the teachers, the social support of the students increased. The video images of the activities in small groups showed that when children were given sufficient space and time for the processing of hints, they were able to benefit from the experience. This would seem to be an obvious conclusion because it is known that having an intellectual impairment means that the processing of information is by definition slower. However, it can be difficult to get this processing to occur in practice. Students with intellectual impairments seem to derive particular benefit from the responsive comments of their teachers to their verbal and nonverbal initiatives (Guralnick, Connor, Hammond, Gottman, & Kinnish, 1996; Rotholz & Ford, 2003), especially in the case of children from low SES backgrounds (Silsby, 2004). Moreover, the experiences of teachers in special institutions may not be comparable to those working in mainstream schools. It may, therefore, be necessary to prepare both students and their teachers for reintegration of the students (Hunt, Soto, Maier, & Doering, 2003; Mathes, Fuchs, Roberts, & Fuchs, 1998; Schepis, Reid, Ownbey, & Parsons, 2001).
Alternative Explanations for the Results
Our choice of statistical techniques makes it impossible to account for the possible influence of SES of the parents of the participating children on the results of this study. Two hypotheses related to parts of these data were previously tested and reported elsewhere. The first hypothesis was that the choice of the research task would lead to an artifact in which the behavior of the teacher and the children was not representative of authentic social interaction. To address this, we collected data from the video recordings that were made during the video interaction counseling sessions. The observations from these recordings were analyzed for both pre- and posttest performance. This led to a major loss of data because for every child the data for at least two sessions had to be available, and the composition of the groups for the video recordings was not predefined. Of the remaining group of 42, we tested to determine whether there were significant effects by comparing the results from one session in the period prior to the first, second, and third series of measurements. If there were differences, then these were consistent with those of the analysis of the data obtained from the research task (Van der Aalsvoort & Langeslag, 2001; Van der Aalsvoort & Lidz, 2002). The findings revealed that the results with the standardized and nonstandardized tasks were comparable.
We investigated a second explanation by analyzing the verbal protocols that were made of the research tasks because we suspected that the behavior of the teachers and group leaders might curb the initiatives offered by the children. The language use in the final session in which the research task was administered was compared with a session in which the research task was administered without the presence of either the teacher or group leader. This comparison showed that there were obvious differences. In all sessions in which children played the research task without their teacher or group leader, they were able to play the game correctly, and they spontaneously experimented with their own variations of the game (Kats, 2001; Van der Aalsvoort & Harink, 2001). This variable, then, could have affected the results of this study.
Comments on the Research Method
We intended to create a design that would show that an increase in the teachers' quality of interaction would have a positive effect on the level of competence of their students with intellectual impairment. The use of a pretest–intervention– posttest design seemed appropriate in order to avoid intervention effects that might be attributable to maturation, testing, or events outside of the intervention (history). Furthermore, two of the three measures of competence were known from the literature to be instruments with evidence of good stability (Riksen-Walraven, Smeekens, & Stapert, 1999). For the Scales of Social Support and Response to Social Support, the correlations between the subscales were calculated in order to evaluate whether they were suitable for the assessment of the behavior of the teachers and students because these instruments were originally used primarily for assessing the quality of social interaction between parents and their children (Erickson et al., 1985). The alphas were found to be satisfactory when the scores for the subscales Confidence in Help and Enthusiasm were removed from the original scale of Response to Social Support (Erickson et al., 1985).
In intervention studies in special education, researchers should indicate as explicitly as possible what the intervention involves and the extent to which the intended effects are included in the regular activities of the students (Gersten et al., 2005). Both of these tasks were described in detail in this study. However, by describing social awareness, social support, and response to social support as indicators of competence in a small group, we investigated the background of the effects of video interaction counseling at the cost of effect size (Simmerman & Swanson, 2001).
Recommendations for Future Research
The Discussion section reveals that there are improvements possible concerning future investigations. The suggestion that we offer is implementing video interaction counseling as a main vehicle to guide teachers' efforts to make inclusive education come true. A relatively short period of time elicits major discoveries about how social interactions can be improved by taking the teacher's behavior as the starting point to allow unfolding active involvement by intellectually impaired preschoolers, revealing their adaptability in mainstream classrooms. Especially small groups composed of students with and without special needs will show how active involvement evolves among the students and in relation to their teacher. Implementing the same design as described in the study will reveal how inclusive education works.
We are grateful for grants from Stichting Kinderpostzegels Nederland, Stichting Nationale kollekte voor geestelijk gehandicapten and Stichting Fontys, Teacher Education School, Tilburg, The Netherlands.
G. M. van der Aalsvoort, PhD (firstname.lastname@example.org), and G. Gossé, MA, Leiden University, Centre for the Study of Early Learning Problems and Impairments, Wassenaarseweg 52, 2300 RB Leiden, The Netherlands