The Impact of the ITLC on Children

Differences Among Classroom Types
First Year Results on the BIT and the ILA. First year BIT and ILA scores were analyzed according to the Ganova multivariate statistical analysis program (Brecht & Woodward, 1987). Univariate analysis of variance with prior contrasts was performed on both the BIT and the ILA scores. No significant differences were found on either set of the mean pretest scores. At the beginning of Year 1 both the BIT and the ILA scores among all three classroom types were similar. However, at the end of the year, there were statistically significant differences and linear trends on the scores in favor of the children in the ITLC classrooms. There was a significant difference for the ILA between Types I and II (p<.04) and between Types II and III (p<.03). Differences were more pronounced for the BIT, with significant differences between Types I and III (p<.006) and between Types II and III (p<.00001). This finding suggested that perhaps the prior experience with computer technology of Type II's teacher and children had a positive effect on the ITLC, although we suspect that teaching style and classroom conditions played strong roles. Interestingly, the Type III teacher and children had just as much computer experience as found in Type II, but that factor alone did not contribute to increased literacy behavior in the children. Neither did it contribute to increased computer skills.

Second Year Results on the BIT and the ILA.
Across ITLC sites in the second year, gains in literacy behaviors were seen in 3-, 4-, and 5-year olds. An Analysis of Variance testing for linear trends on the gains scores of the 12 item ILA measure is shown in Figure 8. The classrooms were grouped according to the following four categories: (1) technology plus the ITLC plus close supervision and assistance from the Macomb Projects research team, (2) technology plus the ITLC implemented by teachers, (3) technology alone, and (4) no technology. Although the distance between the classes is not equal, the order corresponds to the strength of the treatment. The highly significant linear trend means that the more powerful the application of technology and the curriculum, the more powerful the results with children. Although the gains are not large, it should be remembered that the ILA is not a sensitive instrument and other benefits from the ITLC have been observed in the qualitative data which covers a longer time period and observation of behaviors children demonstrate in classrooms.

Figure 8. Analysis of Variance Testing for Linear Trends in Year 2

A = ITLC Implemented by Macomb Projects (N = 64)
B = ITLC Implemented by teachers (N = 40)
C = Control Site Technology (N = 26)
D = Control Site No Technology (N - 22)

Second year results on the BIT were similar to first year results. Differences between the Type I and II classes and Type III classes were significant, p<.001. Over a two year period, 3-, 4-, and 5-year-old children in both the two ITLC conditions (N=114) and the non-ITLC implementation sites (N=48) showed gains on the ILA from the pretest to the post-test. However, the ITLC children showed greater increases on two thirds (66.7%) of the 12 items as compared to the non-ITLC children who made gains slightly better than the ITLC children by only three percentage points on one item ("points to pictures while reading") and equal to the ITLC group on "retells story in own words". The greatest gains in the ITLC group were shown on relatively sophisticated emergent literacy skills related to pretend "reading" with vocal inflection (40%, N=42 in the ITLC group, 25%, N=13 in the non-ITLC group), and sequencing stories in appropriate order (24%, in the ITLC group, 8% in the non-ITLC group). Interestingly, one third of 3-year-olds in ITLC sites turned book pages at the appropriate time (33%, N=8) but not one in the non-ITLC group did so. Fewer differences on ITLC items were found between ITLC 3-year-olds and non-ITLC children of the same age. However, differences separating the two groups were greater at four years of age, and even greater at age five, suggesting that the ITLC rather than maturation accounts for children's gains in literacy.

Third Year Results of the BIT and the ILA.
At the end of the third year of the study, scores on both the BIT and ILA were analyzed using the Brecht Woodward Ganova 4 program of multivariate analysis of variance. The dependent variable was the difference between pre and post scores. Because of the theoretical and logical independence of the variables, each of the subtests of the BIT and ILA were considered as independent univariate variables. As a first step, all data were tested for normality. Where evidence of non-normality was determined, that fact was noted but was not compensated for in the analysis. It often occurred when there was clearly no suggestion of significance. The meaning of significant skew and kurtosis is mentioned but not elaborated on. Subsequently, data in the cells of the analysis were tested for equality of variance. Where there were significant differences in variance among groups, Wald's Heteroscedastic ANOVA was used. Analysis of variance was then conducted for each of the variables and subsequently, planned comparisons were used to answer interesting questions suggested by the first analysis.

The independent subtests of the BIT, determined by factor analysis, were (A) Attending, (B) Resisting the computer, (C) Cause and Effect, (D) Expression, (E) Follows directions, (F) Independence, (G) Planning, (H) Calling attention to self in a positive way, (I) Calling attention to self in a negative way, (J) Peer interaction, (K), Cooperating with peers, and (L) Competing with peers.

Thirty-six paired t-tests were first performed on each of the three groups and 12 subtest variables to see if gains had been made. No significant gains were noted in any of the three comparison groups for variables B, I, and L. These three behaviors, resisting the computers, calling attention to self in a negative way, and competition at the computer were seldom observed. Pre-post differences for each child were usually 0 on those variables. Because of the lack of variability, these three items were excluded from further analysis.

Significant gains were made on variables A, C, F, G, H, and J for all three technology groups, Technology Only, ITLC, and ITLC + staff. Significant pre-post differences were found for two groups for variables E (follows directions) and K (cooperating with peers). An almost significant pre-post difference, (p = .051) was found for variable D (expression). The Technology Only group was the group of children who had technology but no ITLC; Group ITLC plus staff were classrooms during their first year presenting with the help and supervision of research staff. The ITLC group represented the same teachers who were in the ITLC plus staff group during a year of practice and instruction with research staff. These pre-post differences are shown in Table 5.

Table 5. Behavior Interaction Tool Pre-Post Scores and Probability for Independent Variables in ITLC with Staff, ITLC, and Technology Only Groups on the Behavior Interaction Tool

BIT Results.
On the first variable, Attending, there was no significant difference across all three groups. Perhaps the novelty or potency of the computer and software, regardless of how it is presented, affected all groups equally. The curriculum seemed to make no difference.

On the second variable tested, Cause and Effect, the mean gains were as follows:

The ANOVA was significant with p<.05. This would imply that the curriculum plus the research staff was significantly different from the technology only group. Further tests showed that curriculum only was not different from technology only and that technology plus staff was significantly different from curriculum only, p<.05.

For variable D, Expression, there was no significant difference among the scores of the three groups.

Variable E, Following directions, had highly significant skew and kurtosis of the data in all three cells. There was also a significant difference in variances so the Wald statistic was used. There was a significant difference among the scores of the three groups, p<.05. The mean differences were:

The contrast for curriculum only and curriculum plus staff showed no significant differences. The average of those compared to technology only showed a significant difference of p<.025. In other words, both ITLC groups were similar in gains and did significantly better than the technology only group.

The mean differences for Independence, variable F, were significantly skewed and kurtotic for all three groups. Also, the variances were significantly different. Therefore, the Wald Heteroscedastic ANOVA was used. It showed an overall significant difference with p<.001. The mean differences were:

From the overall significance it is apparent that the largest value is significantly different from the smallest. Thus, the presence of the staff was an advantage over the teachers presenting the curriculum by themselves, even though their data represents their second year of experience with the program. Observational data supported this finding. When teachers and program assistants implemented the ITLC in the second year, they were more directive with children and more restrictive in allowing children access to software. A planned contrast comparing technology only with curriculum only showed no significant difference, p=.092.

Planning, variable G, showed a significant skew for all variables and significant differences in variance. Therefore, the Wald Heteroscedastic ANOVA was used again. The overall difference in means was significant, p<.01. The mean differences were:

Use of the curriculum with assistance from the research staff was clearly superior to use of technology without the curriculum. The planned contrast between curriculum only and technology only was also significant, with p=.002. The difference, then, was due to the ITLC, whether or not the staff was involved.

Variable H, Calling attention to self in a positive way, showed mean gains which were not significantly different. They were 1.5,1.4, and 1.6 respectively. We believe that this means that children called attention to themselves more as the year went by, but this tendency was not related to the way the program was presented. Perhaps, as children matured and grew more accustomed to classroom procedures, they had more to say about themselves in general.

Variable J, Peer interaction, had all groups showing normality of the mean differences for variances. The variances, however, were significantly different which again called for use of the Wald Heteroscedastic ANOVA. This test showed an overall significant difference among groups with p=.01. The mean gains were:

Table 6. Pre-Post Scores and Probability for Independent Variables in ITLC with Staff, ITLC, Technology Only, and No Technology Groups on the Informal Literacy Assessment

A sequence of planned contrasts showed significant difference between the curriculum plus staff and both technology only and curriculum only, indicating a greater amount of positive social, peer-to-peer interaction among children when the research staff implemented the ITLC.

In summary, it appears that the presence of the research staff was crucial to obtain positive results. When teachers were left on their own, even with support, the results were significantly superior to technology only groups in the cases of planning and following directions. Thus, the ITLC, when presented with fidelity and with supervision, is effective and produces significantly superior results on most of the BIT variables. However, a year of training teachers does not seem to be sufficient to obtain the high quality results obtained with closer technical support.

ILA Results The ILA was divided into four subtest of four items each, Book Handling, Pictures, Text, and Story. In contrast to the BIT, an additional "no technology" comparison group was possible. As in the analysis of BIT, the first step was to do paired t tests between pre and post results. No significant gains were noted for the Text variable. However, there were significant gains on Book Handling for the ITLC group. In addition, there were significant gains on the Picture variable for that group and the curriculum plus staff group. There were significant gains on the Story variable for all but the no technology group. Thus, some technology was better than none in all instances. See Table 6.

Table 6. Pre-Post Scores and Probability for Independent Variables in ITLC with Staff, ITLC, Technology Only, and No Technology Groups on the Informal Literacy Assessment

The next step was to determine whether there were significant differences in the gains made by the different groups. The Text variable was excluded because there were no significant gains in any of the groups for that variable. The method of analysis was the same as that used for the BIT results. There were no significant differences among all groups for the first two variables, Book Handling and Pictures. There were significant differences among groups for the fourth variable, Story, p=.004. Indeed, the second weakest group, technology only, is significantly different from no technology on this variable, p=.011. The mean gains are as follows:

The data on Story indicate that any kind of technology is helpful in increasing the concepts; however, adding the ITLC increase the effect somewhat.

• To Next Section
• Back to Index