Experimental Measurement of Verbal Development: (an Experimental, Behavioural Approach to the Study of Aspects of Conceptual-Cognitive Development in Certain Groups of Atypical Children.)

Abstract

In this study the author attempts to establish the practicality and validity of developing a technique, first described by Luria (1963,a), using psychophysiological methods, for establishing the functional level of conceptual cognitive development of a particular subject and to describe characteristics of homogenous groups. This was done by measuring the associative orienting responses for a subject when a stimulus word (C.S.) had been paired with a U.C.S. and orienting responses (skin resistance responses) toward both sound (R.S.) and meaning (R.M.) were measured. Literature was reviewed in the following areas: 1) Orienting Responses and the Learning Process, describing differences in Q.R. with increasing chronological age and between normal and subnormal children. 2) Verbal Association and Learning, relating verbal association to cognitive development and describing difficulties in the mediation processes which underlie all learning experienced by subnormal children. 3) Literature relating to Inhibition and Learning in relation to both Operant and Respondent conditioning procedures was also reviewed as these techniques were used in the experimental programme. Differences between normal and subnormal children when aetiology of retardation is controlled were described and related to the first two areas of discussion. Responses from 50 subjects ware studied over three trials; six weeks and six months respectively, from date of Trial 1. These subjects comprised five groups; 30 subnormal Down's Syndrome subjects (I.Q. 30 - 59) randomly assigned to a control group (Group 1), two experimental groups (Groups 2 and 3) and two groups of normal subjects (I.Q. 93 - 138) matched with Group 1 on experimental variables C.A. (6.2 - 14.5) Group 4, and M.A. (3.0 - 6.6) Group 5. Data from Trials 1, 2 and 3 were studied and Test-Retest Reliability of experimental variables R.S. and R.M. over 6 weeks and 6 months reported. Standard Errors o f Measurement were computed for both R.S. and R.M. Groups 2 and 3 were trained (in the Experimental Phase prior to Trial 3) to respond (Group 2) to sound and (Group 3) to meaning. Responses on both variables were compared for Trial 3 (Groups 2 and 3) with those of Group 1. This procedure was used to ascertain the degree to which R.S. and R.M. reflected immediate previous learning and thus acted as a source of unreliability or confounding in the results. Validity o F the Experimental Method was assessed by measuring the degree to which the predictor variables (I.Q., M.A., and C.A.) as measured by standardised tests of intelligence were correlated with the experimental variables (R.S. and R.M.) over all subjects and in Control Groups 1, 4 and 5. It was thus hypothesised that responses for both R.S. and R.M. would be reliable (correlated at .05 level of significance) over all Trials and that these correlations would be highest for Group 4 and lower for Groups 1, 2 and 3 i.e. to be highly reliable. Responses to both R.S. and R.M. were hypothesised to show no significant difference between Groups 1 and 2 and 3 over Trial 3 i.e. not to reflect direction of previous training. A strong significant positive correlation was expected between I.Q. and R.M. and a strong significant negative correlation between I.Q. and R.S., over all groups to establish the validity of the experimental method. Results of the study indicated the following: R.S. and R.M. were highly reliable for normal subjects only and R.S. was less reliable than R.M. with subnormal subjects. It would be feasible to increase the reliability of R.M. for subnormal subjects by lengthening the test. The directional conditioning procedure led to decreased response in the direction opposite from that of training and to a decrease in R.S. for both experimental groups. This was discussed in terms of a stimulus and a response inhibition facilitation effect. Effects of previous learning on reliability were also discussed. Both R.S. and R.M. were established as indicators of conceptual cognitive ability (r, I.Q./R.M. = .358, sig. .05 level; r, I.Q./R.S. = -.599, sig. .01 level) but R.S. was seen to be more useful as a predictor with normal subjects due to the continued presence of Klang (responses to sound) responses in older normal subjects. It was concluded that the experimental method, although valid with respect to relationships between predictor and experimental variables was more useful in its present form for use with normal children.