Research in Developmentnl Disabilities, Vol. 12. pp. 229-241.1991 0891-4222191 $3.00 + .OO Printed in the USA. All rights reserved. Copyright 8 1991 Pergamon Press plc

Attentional Aspects of Classroom Behavior and Discrimination Learning

Steve C. Strand

Plymouth Polytechnic

Teachers’ ratings of their mentally handicapped pupils’ classroom behavior were examined as correlates of the child’s performance on a discrimination learning task. Teachers completed the 23-item version of the AttentionlDistraction InhibitionlExcitation Classroom Assessment Scale (ADIECAS) for a total of 77 pupils. Factors identified as attentionldistractibiliry and responsivity to conse- quences correlated signijcantly with number of errors and number of trials to criterion during acquisition of a simultaneous visual discrimination. Attentionldistractibility scores also correlated significantly with number of dimensions “attended to” during discrimination learning.

Two-stage theories of discrimination learning (e.g., Sutherland & Mackintosh, 1971; Zeaman & House, 1963, 1979) emphasize the impor- tance of attentional processes in discrimination learning. For example, Zeaman and House (1963) ascribe the “learning deficit” of mentally handi- capped persons on two-choice visual discrimination tasks to a low proba- bility of attending to certain relevant stimulus dimensions, rather than poor ability to learn which of two observed cues is correct. They hypothesise that mentally handicapped individuals may attend to stimulus dimensions which are irrelevant to task mastery (e.g., position). Other research (e.g., Koegel & Wilhelm, 1973; Lovaas, Koegel, & Schreibman, 1979) has sug- gested that mentally handicapped people often have difficulty acquiring complex discriminations since they may respond “overselectively,” attend-

The author wishes to express his appreciation to the headteachers, staff and pupils of Mill Ford and Downham schools for their generous help in supporting the reported research.

Requests for reprints may be sent to Dr. Steve Strand, Head of Research and Evaluation, Education Department, London Borough of Wandsworth, Education Professional Centre, Franciscan Rd., Tooting, London SW17 8HE, England.

229

230 S. C. Strand

ing to only one or a restricted number of cues of a compound stimulus. These accounts emphasise concepts of direction and breadth of attention, respectively. However, while attention in these terms is a cognitive con- struct referring to information processing, attention also has been assumed to have a behavioral component (Wood, 1988). The aim of the present study was to determine whether teacher ratings of attentional aspects of their mentally handicapped pupils’ classroom behavior relate reliably to “attention” as exhibited in a visual discrimination learning task.

Siegel, Crawford, and Evelsizer (1985) have also addressed this ques- tion. While teachers’ ratings of their mentally handicapped pupils’ inat- tention did correlate significantly with PPVT MA and IQ they report no significant correlations between ratings of inattention and the child’s performance on a visual discrimination learning task. However, the mea- sure of inattention employed by Siegel et al. consisted of a single dichotomously scored item: The child’s teacher was simply required to indicate the presence or absence of “inattention (e.g., short attention span, easily distracted)” (p. 390). It may be argued that this measure entails the loss of considerable information, since attention may be pre- sumed to be a broad dimension upon which individuals vary, rather than a dichotomous variable.

In this regard, Evans (1975) has developed a psychometric instrument termed the Attention/Distraction Inhibition/Excitation Classroom Assessment Scale (ADIECAS) for use with mentally handicapped children. The scale consists of 23 items related to the child’s overt behavior in the classroom, and each item is rated by the teacher on a 7-point scale. Factor analysis extracted two factors, identified as attention/distraction (AD) and inhibition/excitation (IE) (Evans, 1975). Evans’ principal concern was with the inhibition/excitation scale, and subsequent research has indicated a relation between teachers’ ratings of inhibition/excitation and aspects of both pupils’ performance on a “go-no-go” discrimination learning task (Evans & Hogg, 1975), and their classroom interactions with teachers and peers (Beveridge & Evans, 1978). In the only research to follow-up the attention/distraction scale, Evans and Hogg (1984) reported significant cor- relations between attention/distraction and EPVT Verbal Age and Verbal Quotient. No investigation of the relation between AD and discrimination performance has been TepOrted.

In summary, the present research attempts to: (a) establish an empirical link between behavioral and cognitive constructs of “attention” by asking whether teacher ratings of attention/distractibility correlate with the cogni- tive construct of attention, measured from performance on a discrimination learning task; and (b) to further evaluate the validity of the ADIECAS by examining the correlation of all ADIECAS factors with pupils’ discrimina- tion performance.

Attentional Aspects 231

METHOD

Subjects and Apparatus

Eighty severely mentally handicapped children attending schools for children with Severe Learning Difficulties (SLD) participated in the study. Chronological age (CA) ranged from 8.8 to 17.9 (mean 13 years 10 months, SD = 29 months). Mean PPVT Verbal Age (VA) was 4 years 3 months (SD = 25 months), and mean PPVT Verbal Quotient (VQ) was 3 1.2 (SD = 13.1). Discrimination learning tasks were presented using a 32K BBC(B) micro-computer, double disk drives, and color monitor. Located beneath the screen was a two-button response box. Stimuli presented on the screen appeared directly above the response buttons. All data was automat- ically recorded and saved at the end of each session.

Procedure

Factoring the ADIECAS. Nine teachers completed the 23-item version of the ADIECAS (Evans, 1975). Teachers were required to rate each of their pupils on a 7-point scale for each item. Completed questionnaires were obtained for a total of 77 pupils. These ratings were then subjected to factor analysis to replicate the factor structure of the ADIECAS.

Evans (1975, 1982) used principle components analysis with an eigen- value of 1.0 as the cut-off, followed by varimax rotation. Two factors were extracted which were identified as attention/distraction (AD) and inhibi- tion/excitation (IE). However, when this analysis was replicated with the present sample, four factors emerged. The attention/distraction factor was recovered quite clearly. The same nine items Evans identified as loading on his first factor (AD) also grouped together and loaded on the first factor extracted in the present analysis. However, the other items of the ADIECAS proved less stable: inhibition/excitation items split over two fac- tors - one consisting of the IE items relating to speech and a second con- sisting of the other IE items - and a separate factor relating to responsive- ness to the consequences of behaviour was identified. Details of the factor analysis, assessment of the reliability and validity of the extracted factors, and a full discussion and comparison with Evans’ results is contained in Strand (1989) and Strand, Sturmey, and Newton (1989). For the purpose of the present research, the factor loadings obtained by Strand were employed.

Table 1 presents the items with significant loadings on each of the extracted factors. These four factors were identified as attentiomdis- tractibility (AD), inhibition/excitation (JE), responsivity to consequences (RC), and verbal expressiveness (VRB). Factor scores were generated by the regression method for each pupil on each of the four factors. These fac- tor scores were subsequently employed in the correlational analysis.

232 S. C. Strand

TABLE 1. Brief Description of ADIECAS Scale Items

Attention/Distraction (AD) Requires additional stimulation to continue a task: always to never Concentration is: v. poor to v. good Capacity for work is: v. low to v. high Reaction to a problem is: v. slow to v. fast Perseveres for a long time: not at all to v. great extent Listens to instructions well: not at all to all the time Works well towards a set objective: not at all to v. great extent Is v. easy to v. difficult to distract To get attention for a new task is: v. difficult to v, easy

Inhibition/Excitation (IE) In action is: v. restrained to v. impulsive Is v. calm to v. excitable Is v. inhibited to v. uninhibited

Responsivity to Consequences (RC)

Performance on a task tends to be disrupted by failure: not at all to v. great extent Performance on a task improves with success: v. great extent to not at all Responsive to praise: not at all to v. great extent Is upset by being told off: v. great extent to not at all Actively seeks approval: always to never

Verbal Expressiveness (VRB) Speaks: v. indecisively to v. decisively Speaks at length: not at all to v. great extent Speaks: v. slowly to v. quickly

Discrimination Training

Pre-training. A pretraining procedure was employed to familiarize the chil- dren with the apparatus and the association between the position of the stimuli on the screen and the buttons beneath them. A single white square was presented randomly on either the right or left of the screen. A correct response (selection of the button beneath the stimulus) produced a tune from the computer and was rewarded with a small edible and verbal praise. An incorrect response produced no consequences. A noncorrection proce- dure was used so that every response terminated that specific trial. Initially, the experimenter modelled the correct response, and provided extra physi- cal, verbal, and imitative prompts whenever necessary. The criterion for successful completion of the problem was 20 consecutive, unprompted, correct responses.

Training program. If subjects acquired the pretraining task, they subse- quently received a second discrimination task consisting of a large, pale blue diamond (8 cm high) and a small, yellow circle (4 cm high). Thus, the

Attentional Aspects 233

stimuli differed from each other on three dimensions; size, color, and form. For half the subjects the large blue diamond was designated S+, while for the other half the small yellow circle was designated S+. The left/right position of S+ was varied randomly, with the constraint that it could not appear in the same position on more than three consecutive trials.

At the start of each session the child was given the following directions: “Now (child’s name) we are going to play a game, this game is called play the tune. We will see some pictures on the television, one will be a winning picture which will play the tune and the other will be a losing picture. Your job is to find out which is the winning picture and play the tune.” A correct response produced a tune from the computer and was rewarded with verbal praise and a small edible. An incorrect response produced no conse- quences. If the child failed to respond within 6 s, or attempted to elicit a direction from the experimenter before responding, one nondirectional prompt such as, “Which picture will play the tune? Where is the winning picture?” was used. If the child still failed to respond, an error was scored. The screen was blank during each 6-s intertrial interval. The learning crite- rion was 10 consecutive correct responses, and training was continued until the subject attained the learning criterion or to a maximum of 200 trials, whichever occurred first.

Dimensional Testing

As soon as the child attained the learning criterion for the above discrim- ination, a dimensional testing phase was initiated. There were three types of probe (test) trials: a form test, 6 cm green diamond versus 6 cm green circle; a color test, 6 cm pale blue “II” versus 6 cm yellow “II”; and a size test, 4 cm green “IT” versus 8 cm green “IT’. The position of S+ was coun- terbalanced across probe trials so that position perseveration would result in chance level performance. The order of presentation of probes was also counterbalanced across subjects. Each of the three probes was repeated six times, giving a total of 18 probe trials. Probe trials were presented in alter- nation with 18 stimulus complex trials which served to maintain the sub- jects’ responding and discrimination of the correct training stimuli. No responses during probe trials were rewarded to prevent training of positive components. Subjects were told the trainer had some work of his own to complete, and after reinforcing the fist stimulus complex trial the trainer retired to the far end of the room to complete his “work.”

If a cue was correctly selected on at least 5 of 6 probe trials for that cue, the cue was assumed to have been functional in learning. Since the discrim- inative stimuli differed on three dimensions (size, form, and color), breadth of learning was assessed by the total number of cues the child responded to during dimensional testing (dimensional score).

234 S. C. Strand

RESULTS

Discrimination Performance

Fifteen subjects failed to acquire either the pretraining or training dis- criminations. In the majority of these cases pupils exhibited prolonged position perseveration or simple position alternation patterns. For the 65 children who acquired the training task, median trials to criterion equalled 16.5 (range 10-150) and median errors equalled 3 (range O-60). Following dimensional testing, data for a further four subjects was elimi- nated since the subjects failed to maintain at least 13/18 (72%) correct responding on the stimulus compound trials. Of the remaining 61 subjects, 8 responded to all three cues, 17 responded to two of three cues, and 36 responded to at most a single cue. On average, pupils responded to a mean I.5 cues.

Correlations between Variables

Table 2 presents the correlations between the individual difference mea- sures and discrimination performance. 1 All correlations are Pearson prod- uct moments except those involving dimensional score, where Spearmans rho was employed. Attention/distractibility (AD) correlated significantly with trials to criterion (r = .56), errors (r = .56), and number of dimensions (rs = .42) (all p < .OOl). These significant correlations indicate that children rated as less distractible acquired the discrimination more rapidly, with fewer errors, and responded to more of the component dimensions. Attention/distraction also correlated significantly with VA and VQ (r = .43, and r = .50, respectively, p < .OOl). CA was not significantly correlated with attention/distraction (r = .17, p < .lO). Responsivity to consequences (RC) did not correlate significantly with either CA, VA, or VA (all p < .lO). However, responsivity to consequences did correlate significantly with both trials to criterion (r = .23, p < .025) and number of errors (r = .21, p < .04). VA correlated significantly with both trials to criterion (r = .55) and errors (r = .53), and produced a low but significant correlation with dimensional score (rs = .23, p < .05).

No other variables correlated significantly with discrimination perfor- mance. Inhibition/excitation correlated significantly only with CA (r = .25, p < .05) (older children were more likely to have higher inhibition ratings than younger children), and verbal expressiveness correlated significantly only with VQ (r = .28, p < .005).

‘Trials to criterion and errors were significantly positively skewed, and log linear transformations were applied to these variables prior to the correlation analysis (cf., House & Zeaman, 1960).

Attentional Aspects 235

TABLE 2. Correlations Between Individual Difference Measures and

Discrimination Learning

Variable Trials FST0r.S Dimension@

Chronological Age .078 .092 .192 Verbal Age -.553*** -.529*** .225* Verbal Quotient -.449*** -.437*** .115 Attention/ -.561*** -.556*** .424**‘* Distraction Inhibition/ -.112 -.124 .021 Excitation Responsivity -.229** -.206* .OOl to Consequences Verbal -.119 -.118 -.056 Expressiveness Multiple Correlation .634*** .650*** -

*p < .OS; **p < .025; ***p < .OOl. aAIl correlations are Pearson product moments, except those with dimensional score, which are Spearmans rho.

Multiple regression analyses and partial correlation coefficients were also computed. The results indicated that only attention/distraction and verbal age were independently related to learning. With VA held constant, the correlation between responsivity to consequences and discrimination errors was no longer significant (r = .17, p z .05). However, the correla- tion between attention/distractibility and discrimination errors remained highly significant even after controlling for differences in VA (r = .43, p < .OOl). Thus, attention/distraction could account for some unique part of the variance in discrimination errors that could not be accounted for by VA. Similar results were obtained when trials to criterion was the depen- dent variable.

In summary, VA and attention/distractibility ratings explained substan- tially independent parts of the variance in discrimination performance. Consequently, the best predictor of discrimination performance was a com- bination of attention/distraction scores and verbal age which together accounted for 42% of the variance in discrimination errors (multiple r = .650, F(2, 74) = 27.02, p < .OOl), and 40% of the variance in trials to criterion (multiple r = .634, F(2,74) = 24.74, p < .OOl).

Ordinal Classification

Nineteen children had no dimensional score, since they failed either to acquire or maintain the discrimination. If these children had more extreme

236 S. C. Strand

scores on the PPVT and ADIECAS factors, the resulting correlations with dimensional score may suffer from attenuation in range. To test this hypothesis, the sample was divided into three groups. Group 1 consisted of the 19 pupils who failed to acquire or maintain the discrimination, who were referred to as “weak discriminators.” The 61 pupils who successfully completed dimensional testing were split into two groups. Twenty-five pupils performed at a superior level, responding to multiple (2 or 3) cues during dimensional testing. These children were referred to as “multiple cue discriminators.” They were distinguished from the remaining 36 chil- dren, who by default were referred to as “single cue discriminators.” Thus, three groups were identified: weak discriminators (n = 19), single cue dis- criminators (n = 36), and multiple cue discriminators (n = 25).

The relevant data are presented in Table 3. Univariate analyses of vari- ance were computed for all variables. Subsequent Newman-Keuls multi- ple comparisons of means revealed weak discriminators (Group 1) had significantly lower VA and VQ (p < .Ol), were rated as more distractible

TABLE 3. Mean and SD of Scores on the Individual Difference Variables for Weak Discriminators

(Group I), Single Cue Discriminators (Group Z), and Multiple Cue Discriminators (Group 3)

Variable Group 1 (II = 19)

Group 2 (n = 36)

Group 3 (?I = 25)

F Value

Significant Differences

CA _ 174.1 158.6 169.8 (32.3) (27.6) (28.0) 2.21

VA 31.3 52.5 (12.9) (20.9) (E)

12.65 Gl vs. G2** Gl vs. G3** G2 vs. G3*

vQ 18.5 33.3 38.1 18.15 Gl vs. G2** (7.5) (11.4) (12.4) Gl vs. G3**

ADa

IEa

RCa

VRBa

-.75 -.ll .I2 16.04 Gl vs. G2** (.83) (.86) (I@) Gl vs. G3**

G2 vs. G3** -.25 .lO

(1.0) (1.1) (E) 0.72

-.43 .18 .Ol 2.31 Cl vs. G2* (.88) (1.1) (.79) Gl vs. G3* -.27

(1.3) (1:; .14 0.89

(.63)

aFor the whole sample factor scores had a mean of zero and standard deviation of 1. CA = chronological age; VA = PPVT verbal age; VQ = PPVT verbat quotient; AD = attention/distraction; IE = inhibition/excitation; RC = responsivity to consequences; VRB = verbal expressiveness.

Attentional Aspects 237

(p < .Ol), and were less responsive to consequences (p < .OS), than single cue discriminators (Group 2). The results also indicated the single cue discriminators (Group 2) had significantly lower VA and were rated as more distractible than multiple cue discriminators (Group 3) (p < .Ol in both cases).

DISCUSSION

Evans and Hogg (1984) reported significant correlations between atten- tion/distraction scores and EPVT verbal age and VQ (r = .54 and r = .49, respectively), and concluded the attention/distraction scale may therefore reflect attentional abilities related to general cognitive ability. The present study replicates these correlations with verbal age and VQ. Additionally, the present results also reveal that attention/distraction scores correlated significantly with three measures of discrimination learning; trials to crite- rion, number of errors, and dimensional score. Importantly, the correlation between attention/distraction and discrimination performance remained significant when the effect of verbal age was controlled. Attention/distrac- tion score accounted for a significant part of the variance in discrimination performance that could not be accounted for by general cognitive ability.

The significant correlation between the attention/distraction factor of the ADIECAS and dimensional score is of particular interest. Many reviews have considered attention to be a multibehavioral process (e.g., Alabiso, 1972; Mercer & Snell, 1977; Stankov, 1983), and distinctions have been drawn between “cognitive” as well as “behavioral” components of atten- tion (Wood, 1988). In attention theory (Zeaman & House, 1963, 1979), emphasis is placed upon a cognitive or information processing component (the ability to scan or sample the available stimuli, and to subsequently select relevant stimuli). Dimensional score, a measure of the number of stimulus dimensions a child responded to following acquisition of a dis- crimination with multiple relevant cues, appears to accord with this con- ception of “selective attention” (Zeaman & House, 1963). However, behav- ioral aspects of attention may also be important in determining discrimina- tion performance, since in order to process relevant stimuli, the child must also maintain attending behavior over extended spans of time. This capaci- ty, variously described as sustained attention, distractibility, or attention span, may be one element assessed in the behavioral items composing the AD scale. The convergence between measures of “sustained” and “selec- tive” attention observed in the present study lends strong support to the validity of the attention/distraction factor. The obtained correlations may indicate the attention/distraction scale measures more than a molar (class- room behavior) aspect of attention, and also measures a more cognitive construct of attention.

S. C. Strand

These results contrast with the only previously reported research (Siegel et al., 1985). These authors found no significant correlations between teachers’ ratings of their mentally handicapped pupils’ inattention and the number of errors and number of dimensions attended to during acquisition of a visual discrimination (r = .15 and r = .18, respectively). However, the two studies differ in several respects. For example, the pupils in the present study had lower VA than the subjects in the Siegel et al. study (mean VA 4.3 vs. 6.8), and Siegel et al. employed a match-to-sample task rather than a two-choice discrimination task. However, the most striking difference between the two studies lies in the measure of attention employed. The dif- ferent results may plausibly be ascribed to the ADIECAS providing a more differentiated measure of attentional aspects of children’s classroom behav- ior than the single dichotomous item employed in the Siegel et al. study.

Evans and Hogg (1984) also reported that, in contrast to the attention/distraction factor, inhibition/excitation scores did not correlate significantly with EPVT verbal age and VQ (r = .04 and r = .17, respec- tively). This result was also replicated in the present research. Evans and Hogg (1984) conclude the inhibition/excitation scale may be better consid- ered as assessing behaviors relating to temperament rather than ability. In support of this conclusion, Beveridge and Evans (1978) have reported that children rated as excitable made significantly more initiations to teachers and peers than children rated as inhibited. However, some earlier research (Evans & Hogg, 1975) did report a relationship between inhibition/excita- tion and indices of inhibition derived from children’s performance on a dis- crimination learning task. Children rated as excitable rather than inhibited made more responses to S- during discrimination acquisition, and more responses to stimuli adjacent to S- during generalization testing. However, while Evans and his associates employed a successive discrimination train- ing procedure, the present study employed a simultaneous discrimination training procedure. In terms of conditioning-extinction theory (Hull, 1950; Spence, 1956) simultaneous discrimination learning requires both excita- tion of responding to S+ and inhibition of responding to S- simultaneously; hence, the two processes are confounded. In contrast, the successive tech- nique presents S+ and S- as discrete trials and separates these component processes; subjects are required to respond only during S+ trials and with- hold responses during S- trials. The use of a simultaneous rather than suc- cessive training procedure in the present study may therefore be a factor accounting for the absence of significant correlations between IE and dis- crimination performance.

Responsivity to Consequences (RC) also correlated significantly with discrimination performance (see Table 2). The greater the rated responsive- ness to consequences the fewer trials and errors to criterion. Thus, a mea- sure of responsiveness to reinforcing consequences derived from ratings of

Attentional Aspects 239

the child’s behavior in the classroom correlated significantly with perfor- mance on an experimental task that may also be dependent upon such pro- cesses, since rate of acquisition of a discrimination will in part be deter- mined by responsiveness to the reinforcing and nonreinforcing conse- quences contingent upon responses. This finding suggests some validity for the responsivity to consequences factor. However the correlation, while significant, is of a relatively low magnitude and the factor can account for only a small proportion of the variance in discrimination performance.

The present results offer some support to research suggesting that breadth of learning is positively correlated with verbal age (e.g., Bailey, 1981; Schover & Newsom, 1976; Wilhelm & Lovaas, 1976). While the correlation between dimensional score and VA was low (r = .23), the corre- lation may suffer from attenuation of range in VA. For example, the six children with the lowest verbal age were also those who failed to acquire the discrimination and therefore had no dimensional score, restricting the range of VA in the correlation with dimensional score. Since the “weak dis- criminators” also differed significantly from the other groups in attention/distraction and responsivity to consequences scores (see Table 3) a similar argument may apply for these variables. In short, higher correla- tions may exist with dimensional score for an unrestricted range of VA, attention/distraction, and responsivity to consequences. However, in accord with observations made by Bailey (198 1). there is enough variability in the data to indicate that the reIationship between VA and breadth of learning may be a weak one. The present results reveal attention/distraction score, rather than verbal age, was most highly correlated with breadth of learning.

Finally, it has been argued that where individual differences in cognitive or attentional abilities are identified that correlate with discrimination per- formance, these individual differences may be used as a basis for individu- alizing instruction (Conners & Detterman, 1987). Parameters of discrimi- nation training programmes may be altered for individual learners, to capi- talize on strengths and compensate for weaknesses in these abilities. For example, pupils with low ratings on attention/distraction may benefit most from pretraining of on-task behaviors, such as visual fixation of discrimina- tive stimuli (e.g., Maier & Hogg, 1974) since such attending skills are essential prerequisites for successful learning. These pupils may also bene- fit particularly from programmed training procedures in which stimuli are initially made highly discriminable and gradually manipulated to succes- sively approximate the target discrimination (e.g., Sidman & Stoddard, 1967; Strand & Morris, 1986). There is also some evidence (Strand, 1989) that pupils rated by their teachers as more distractible are less likely to ben- efit from the class of “non-criterion-related” fading procedures than chil- dren with higher ratings of attention/distraction. Further research to sys- tematically evaluate possible “aptitude-by-treatment” interactions is clearly

240 S. C. Strand

required. However, a necessary precursor to any attempt to specify such interactions is the identification of dimensions of individual difference that are relevant to discrimination learning. The present research suggests the ADIECAS may be important in this respect.

In conclusion, the present results indicate that individual differences in the behavior of severely mentally handicapped children, as assessed through a teacher-completed classroom rating scale, correlated significantly with measures of ability (VA, VQ); measures of discrimination acquisition (trials and errors to criterion); and breadth of learning (dimensional score). The results suggest these dimensions of individual difference provide rele- vant information concerning both aspects of the child’s classroom behavior and performance in discrimination learning tasks.

REFERENCES

Alabiso, F. (1972). Inhibitory functions of attention in reducing hyperactive behavior. American Journal of Mental Deficiency, 11,259-282.

Bailey, S. L. (1981). Stimulus overselectivity in learning disabled children. Journal of Applied Behavior Analysis, 14,239-248.

Beveridge, M. C., & Evans, P. (1978). Classroom interaction: Two studies of severely subnormal children. Research in Education, 19.39-48.

Conners, F. A., & Detterman, D. K. (1987). Information-processing correlates of computer-assist- ed word learning by mentally retarded students. American JOWMZ of Mental Deficiency, 91, 606-612.

Evans, l? (1975). Inhibition and stimulus generalization in the discrimination [earning of ESN(SJ and ESN(M) children. Unpublished Ph.D. thesis, University of Manchester.

Evans, P. (1982). Inhibition and individual differences in inhibitory processes in retarded children. In N. R. Ellis (Ed.), International review of research in mental retardation (Vol. 11, pp. 79-122). New York: Academic Press.

Evans, I?, & Hogg, J. (1975). Individual differences in severely retarded children in acquisition, stimulus generalization, and extinction in go-no-go discrimination learning. Journal of Experimental Child Psychology, 20.377-390.

Evans, P., & Hogg, J. (1984). A classroom rating scale for use with mentally retarded children. British Journal of Clinical Psychology, 23, 187-194.

House, B. J., & Zeaman, D. (1960). Visual discrimination learning and intelligence in defectives of low mental age. American Journal of Mental Deficiency, 65,51-58.

Hull, C. L. (1950). Simple qualitative discrimination learning. Psychological Review, 57, 303-313.

Koegel, R. L., & Wilhelm, H. (1973). Selective responding to the components of multiple visual cues by autistic children. Journal of Experimental Child Psychology, 15,442-453.

Lovaas, 0. I., Koegel, R. L., & Schreibman, L. (1979). Stimulus overselectivity in autism: A review of research. Psychological Bulletin, 86,1236-1254.

Maier, I. & Hogg, J. (1974). Operant conditioning of sustained visual fixation in hyperactive severely retarded children. American Journal of Mental Deficiency, 79,297-304.

Mercer, C. D., & Snell, M. E. (1977). Learning theory research in mental retardation: Implications for teaching. Columbus, OH: Merrill.

Schover, L. R., & Newsom, C. D. (1976). Overselectivity, developmental level, and overtraining in autistic and normal children. Journal of Abnormal Child Psychology, 4,289-298.

Attentional Aspects 241

Sidman, M., & Stoddard, L. T. (1967). The effectiveness of fading in programming a simultane- ous form discrimination for retarded children. Journal of the Experimental Analysis of Behavior, 10,3-1X

Siegel, P S., Crawford, K. A., & Evelsizer, Z. (1985). Attention in discrimination learning in rela- tion to certain teacher-rated behavior anomalies. American Journal of Mental Deficiency, 89, 389-392.

Spence, K. W. (1956). Behavior theory and condidoning. New Haven: Yale University Press. Stankov, L. (1983). Attention and intelligence. JOWMI ofEducational Psychology, 75.471-490. Strand, S. C. (1989). Learning and instructional success in people with a mental handicap.

Unpublished Ph.D. thesis, Plymouth Polytechnic. Strand, S. C., & Morris, R. C. (1986). Programmed training of visual discriminations: A compari-

son of techniques. Applied Research in Mental Retardation, 7, 165-18 1. Strand, S. C., Sturmey, P., & Newton, J. T. (1989). A classroom rating scale for use with mentally

handicapped children: A replication of the ADJECAS. British JOU~M~ of Clinicul Psychology, 29.121-123.

Sutherland, N. S., & Mackintosh, N. J. (1971). Mechanisms of animal discrimination learning. New York: Academic Press.

Wilhelm, H., & Lovaas, 0. I. (1976). Stimulus overselectivity: A common feature in autism and mental retardation. American Journal of Mental Deficiency, 81.26-31.

Wood, R. L. I. (1988). Clinical constraints affecting human conditioning. In G. Davey & C. Cullen (Bds.), Human operant conditioning and behaviour modification (pp. 87-118). Chichester: John Wiley & Sons.

Zeaman, D., & House, B. J. (1963). The role of attention in retardate discrimination learning. In N. R. Ellis (Bd.), Handbook of menfal deficiency (pp. 1.59-233). New York: McGraw-Hill.

Zeaman, D., L House, B. J. (1979). A review of attention theory. In N. R. Ellis (Ed.), Handbook of mental deficiency (pp. 63-120). New York: McGraw-Hill.