6. PRESUPUESTO TENTATIVO DEL PROYECTO
6.1 COSTOS DIRECTOS
As was mentioned previously, the data collected to reflect achievement at 11-years was the level awarded to the children in their end o f Year 6 Standard Attainment Tests (SATs). Children are given an overall score relating to their English, Mathematics and Science test performance. A level o f 4 is awarded to children who have reached the target scores set for their age group. A level 3 score indicates that children have failed to meet that target, while level 5 indicates above average functioning. Table 3 shows the means o f the levels achieved by the BabyTalk and control group children in each o f these areas. The use o f 2-tailed significance criterion is used in this chapter only due to the exploratory nature o f these investigations.
Table 2.3 : Comparative mean SAT level achievement for BabyTalk (experimental) or control status children at 11 years o f age.
Experimental Mean (s.d.) n = 15 Control Mean (s.d.) n = 17 t-value Sig. English 3.93(1.3) 3.88(1.2) .104 .918 Mathematics 4.00(1.3) 3.89(1.2) .262 .795 Science 4.20 (3.94) 3.94(1.2) .604 .550
Table 2.3 reveals that there are no statistically significant differences at the p<.05 level between the means for the experimental and control groups on any o f the SAT tests. It is the case, however, that in all instances the experimental group means are higher. It is also most interesting that for Mathematics and Science scores the means o f the Experimental group, 4.0 ( ^ . = 1.3) and 4.2 (s.d. = 1.2) respectively either meet or surpass the government-set achievement target for their age-group. The control group means fall short of the attainment targets, with scores for mathematics and science o f 3.89 (s.d. = 1.3) and 3.94 (s.d. = 1.2) respectively.
Thus a statistically significant difference in SAT achievement scores is not detectable at 11 years, however it could be argued that there is meaningful difference in the context o f the assessment measure used.
Following the remarkable differences at the 7-year follow-up, these results are rather disappointing. Admittedly, the SAT is a rather clumsy achievement measure which is arguable not a very sensitive test at this age group, and given the small sample size, more impressive results might have been anticipated with more children and more sensitive achievement measures. None-the-less these results rather questioned whether the 11-year sample was truly representative o f the 7-year sample. Consequently, those children (n = 45) who were seen at 11-years and who were also seen at 7-years (n = 32) were selected for analysis of their achievement scores at 7- years to see if the raised IQ scores for the experimental group in this sub-sample remained present. The results o f this analysis are displayed in Table 2.4.
Table 2.4 : Comparative mean WISC IQ scores for BabyTalk (experimental) or control status children at 7 years o f age (sub-group who were also follow-up at 11 years). Experimental Mean (s.d.) n = 15 Control Mean (s.d.) n = 17 t-value Sig. Verbal IQ 98.4 (20.3) 101.1 (15.0) .434 .667 Performance IQ 102.3 (20.6) 100.2(16.1) .313 .757 Full Scale IQ 100.1 (20.7) 100.8 (17.3) .090 .929
The results in table 2.4 clearly indicate that differences between the control and experimental groups did not remain. There is no difference between IQ scores for the control and experimental group at 7 years for those children included in the 11 year follow up. In fact, there is a non-significant difference in favour o f the controls!
It seems that the sub-group o f the BabyTalk sample who have been traced and followed up at 11 years are not representative of the larger group at 7 years in terms o f higher IQ functioning. Indeed it would seem that the current sample is made up o f all the lowest achieving BabyTalk (experimental) children and highest achieving
control children. It is no longer surprising that an achievement benefit for the BabyTalk children was not evident for this sample at 11 years. In this context, the slightly higher (but o f course insignificant) SAT scores o f the experimental group are surprising, and perhaps indicate an effect on later achievement that is independent o f IQ.
A concern remains that the anticipated benefits o f the intervention will not be identifiable in this unrepresentative sub-sample. Further analysis o f the seven-year data for this sub-group revealed that there is also, unlike the group as a whole, no difference in the language skills between the control and experimental group. Reported below (table 2.5) are the language skills test results at 7 years for the sub group o f children followed-up at that age who were additionally followed up at 11 years in the context of the current study. The assessments used are widely used reliable and valid measures developed by Dr John Rust and published by the psychological corporation. 'Word' (Weschler Objective Language Dimension, Rust, 1996) assesses language skills and has sub-scales BR (basic reading), SP (spelling) and RC (reading comprehension) and an overall average score WC (or word composite). The 'Wold' (Weschler Objective Language Dimension, Rust, 1992) has sub-scales o f LC (language comprehension) OE (oral expression) and WE (written expression).
Table 2.5 : Comparative Word and Wold Language skills scores for BabyTalk (experimental) or control status children at 7 years o f age (sub-group who were also follow-up at 11 years). Experimental Mean (s.d.) n = 15 Control Mean (s.d.) n = 17 t-value Sig. (2-tailed) WORD.BR 96.5 (17.3) 95.7(17.3) .145 .886 WORD.SP 81.3 (29.0) 83.7 (32.1) .224 .824 WORD.RC 89.5 (12.5) 91.5(17.7) .361 .727 WORD.WC 92.1 (17.8) 93.1 (20.2) .145 .886 WOLD.LC 98.9(15.1) 102.5(17.1) .638 .528 WOLD.OE 105.0(11.4) 99.9 (14.5) 1.10 .281 WOLD.WE 100.9(14.4) 99.9(16.9) .166 .869
Only in the Wold Oral Expression test did the Baby Talk experimental group (m = 105.0 , s A = 11.4) out-perform the control group (m = 99.9, s A = 14.5) by a noticeable margin, and this failed to reach statistical significance t (30) = 1.10, p = .281 (2-tailed).
It is rather disconcerting that despite the apparent success o f the Baby Talk intervention group followed up at 7 years, the sub-group o f BabyTalk children in the current study appear to be those who have least benefited from the intervention. It is rather disappointing in terms o f our prospects o f identifying improvements in other areas o f functioning, such as empathy, and social understanding. However, if significant differences in these capabilities are detected, we can anticipate that the magnitude o f the effect must therefore be strong and meaningful, and also operating independently o f the enhanced language or IQ functioning o f the children.