This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type o f computer printer.
The quality o f this reproduction is dependent upon the quality o f the copy subm itted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction.
In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.
Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand comer and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back o f the book.
Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6” x 9” black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order.
UMI
A Bell & Howell Infomiation Company
300 North Zeeb Road, Ann Arbor MI 48106-1346 USA 313/761-4700 800/521-0600
in Clinically-Referred Children by
Elisabeth Mary Suzanne Sherman B.A., University of Ottaw a, 1991 M.Sc., University of Victoria, 1995
A Dissertation Submitted in Partial Fulfilment of the Requirements for the Degree of
DOCTOR OF PHILOSOPHY in the Department of Psychology
We accept tliis dissertation as conforming to the required standard.
Dr. M.: iperc’îsor (Department of Psychology)
Dr. E. Strauss, Supervisor (Department of Psychology)
Dr. H. Kadlec, Departmental Member (Departm ent of Psychology)
Dr. M. Uhlemann, O utside Member (Departm ent of Education)
r. J. MacDonald, External Examiner (B.C. Rehabilitation - G. F. Strong Centre)
© Elisabeth Mary Suzanne Sherman, 1997 University of Victoria
All rights reserved. This dissertation may not be reproduced in w hole or in part, by photocopying or other means, w ithout the permission of the author.
Supervisors: Drs. Michael Joschko a n d Esther H. Strauss
Ab st r a c t
Case studies and a sm all n u m b er of group studies in the neuropsychological literature on adults and children w ith brain dysfunction su g g est that certain cognitive skills are im p o rtan t d eterm in an ts of social skills. H ow ever, standardised m easures of social skills designed expressly to m easure this construct have not been used previously in th e neuropsychological field. The goal of this study w as to determ ine neuropsychological correlates of social skills in ch ild ren referred for neuropsychological assessm ent, a n d com pare the findings to the cognitive skills identified in social inform ation processing m odels from developm ental and clinical psychology. In younger children (6 to 9 years), only a m easure of conceptual ability w as related to social skills. H ow ever, from a clinical standpoint, differences in conceptual skills betw een socially im paired an d socially skilled children w ere m inim al. In older children (10 to 13 years), sustained atten tio n and verbal reasoning w ere strong, u nique predictors of social skills. H ow ever, only differences in sustained attention betw een socially skilled and unskilled children w ere clinically significant. Results w ere discussed with reg ard s to 1) social inform ation processing m odels, 2) th e effect of age on the cognitive correlates of social skills; and 3) the relationship betw een insight an d social skills.
Examiners:
Dr. M. Joschl^rf^^upefvT^r (Department of Psychology)
Dr. E. Strauss, Supervisor (Department of Psychology)
Dr. H. Kadlec, Departm ental Member (Departm ent of Psychology)
Dr. M. Uhlemann, Outside Member (Department of Education)
_____________________________
TABLE OF CONTENTS
ABSTRACT... ü
TABLE OF CON TENTS...iv
LIST OF TABLES... vüi ACKNOW LEDGMENTS...xi
IN TRO D UCTIO N ... 1
Def in it io n sa n d Mo delso f Exec utiv e Fu n c t io n i n g...1
Me a s u r in g Executive Fu n c t i o n s... 3
Th e Dev elo pm en tof Ex ec u t iv e Fu n c t io n s in Ch il d r e n... 6
Executive Fu n c t io n in g a n d So c ia l Sk il l s...8
COGNITIVE De t e r m in a n t so f So c ia l Skillfrom Ne u r o p sy c h o l o g y...10
Neuropsychological Case Studies and Clinicians' Reports...10
Neuropsychological Empirical Group Studies... 15
Th eo r iesof So c ia l Be h a v io u r a n d So c ia l Co m peten c e... 19
Me a s u r in g So c ia l Sk il l s...24
Su m m a r yoft h e Liter atu reo n C o c N r n v E Correlatesof So c ia l Sk il l s... 28
Go a l sa n d Hy p o t h e s e s... 32
Pa r t ic ip a n t s a n d Pr o c e d u r e... 35
Ma t e r ia l s...38
Social Skills...38
Intellectual Ability and Achievement Level...40
Executive Functions...42
Reasoning and Conceptual A bility...45
Attention and M em ory...46
Visual-Spatial and Verbal Skills...47
Motor Skills...48 Impairment Classifications...50 R E S U L T S ... 52 Ex p e r im e n t-Wise Err o ra n d Sta t ist ic a l Co n s i d e r a t io n s... 52 Descriptive Da t a forth e En t ir e Sa m p l e... 52 Social Skills... 53
Intellectual Ability and Achievement Level for All Children...56
Executive Functions...57
Other Neuropsychological Tests...58
Intercorrelations betxveen Tests...60
So c ia l Skill Levels a n d Ag e...63
Co r r e l a t io n a l An a l y s e s... 63
General Intellectual A bility...69
Executive Functioning...70
Other Neuropsychological Tests...71
Testing for Statistical Differences between Correlations...74
De t e r m in in g Un iq u e Predicto rso f So c ia l Skillsw it h Reg ressio n An a l y s is ..74
Da t a Re d u c t io n a n d Testingof Th e o r e t ic a l As s o c ia t io n sbetw een Me a su r e s76 Factor Derivation...77
Factor Interpretation...79
Ag e, Hig h/ Lo w So c ia l Skills, a n d Ne u r o p s y c h o l o g ic a l Fa c t o r s...81
Id e n t if y in g So c ia l Skill Subtypes: Clu ster An a l y s i s... 8 7 Derivation o f Social Skill Subtypes...8 7 Cluster Solution Replicability...90
Social Skill Subtypes, Age, and General Intellectual Skills...94
Social Skill Subtypes and Neuropsychological Factors...96
Social Skill Subtypes and Behaviour...9 7 Me a s u r in g In s ig h t: Tea c h er-Self Dis c r e p a n c ie sa n d Neu r o psy c h o lo g ic a l Fa c t o r s...100 D I S C U S S I O N ...1 0 2 Su m m a r ya n d Discussio nof Fi n d i n c s...1 02
Sample Characteristics and Incidence o f Social Problems...1 0 2
Executive Skills and Social Skills... 108
Neuropsychological Factors...109
Social Skill Subtypes... I l l Insight, self-axvareness and Social Skills... 112
Ge n e r a l Lim i t a t io n so ft h e St u d y... 115
Design and Sample Characteristics... 115
Measurement...116
CONCLUSION... 120
BIBUOGRAPHY...122
APPENDIX A CASE EXAMPLES... 142
APPENDIX B SELECTED EXAMPLES FROM THE BLISHEN SCALES... 145
APPENDIX C FORMULA FOR THE CORRECTION FOR ATTENUATION...147
APPENDIX D FORM ULA FOR DETERMINING STATISTICAL SIGNIFICANCE BETWEEN CORRELATIONS...148
LIST OF TABLES
Table 1 List o f N europsychological Tests and N europsychological Functions
P u rp o rte d to be M easured by Each Test... 49
Table 2 Classification Table for Norm -Based Test S c o re s... 51
Table 3 M eans and Standard Deviations for the SSRS Scales... 55
Table 4 Raw Scores for the Teacher-Rated SSRS... 56
Table 5 M eans an d Standard D eviations for Intellectual Ability and A chievem ent Tests for the Entire Sam ple... 57
Table 6 M eans an d Standard Deviations for Executive Functioning T ests...58
Table 7 M eans an d Standard Deviations for O ther N europsychological Tests..57
Table 8 T eacher SSRS Subtest Intercorrelations...60
Table 9 P arent SSRS Subtest Intercorrelations...61
Table 10 Self SSRS Subtest Intercorrelations...61
Table 11 Intercorrelations Between N europsychological T e s ts ... 62
Table 12 Test-Retest Reliability Coefficients for the T eacher SSRS... 65
Table 13 Test-Retest Reliability Coefficients for the P aren t SSRS... 66
Table 14 Test-R etest Reliability Coefficients for the Self-Rated SSRS...67
Table 16 Intercorrelations Betw een T eacher-R ated Social Skills and T ests of
Executive F u n ctio n in g ...71
Table 17 Intercorrelations B etw een Teacher-R ated Social Skills and O th er N europsychological T ests... 73
Table 18 Sum m ary of Sim ultaneous R egression Analysis fo r N europsychological M easures Predicting Teacher-R ated Social Skills fo r O lder C h ild ren ....76
Table 19 Factor L oadings an d R otated Factor M a trix ...78
Table 20 M eans Differences Betw een G ro u p s o n the A tten tio n Factor for Y ounger an d O ld er C h ild ren W ith Low (SS-) a n d H igh (SS+) Social Skills S cores... 82
Table 21 H igh an d Low Scorers on the Teacher SSRS (Ages 6 to 9): N europsychological P erfo rm an ce...84
Table 22 H igh a n d Low Scorers o n the Teacher SSRS (Ages 10 to 13): N europsychological P e rfo rm a n ce ... 86
Table 23 M ean SSRS Subscale Scores for the Five S ub ty p es...88
Table 24 SSRS Subscale M eans for th e K -m eans Cluster R eplication...91
Table 25 SSRS Subscale Scores for W ard 's M ethod C luster R e p lica tio n ... 93
Table 26 SSRS Subscale M eans for the C om plete Linkage M ethod C luster R eplication... 94
Table 27 A ges for the Five Social Skill S u b ty p e s... 95
the Social Skill S u b ty p es... 98
ACKNOW LEDGMENTS
I w ish to thank Esther Strauss an d M ichael Joschko for their support, encouragem ent, and example th ro u g h o u t g rad u ate school. I w ould also like to thank Dr. Helena Kadlec for m ethodological an d statistical advice.
De f i n i t i o n s a n d M o d e l s o f Ex e c u t i v e Fu n c t i o n i n c
N europsychology as a field has a rich tradition o f attem pting to characterise higher cognitive functions such as those th a t underlie judgem ent, problem solving, goal attainm ent, and the m odulation of behaviour. In particular, early neuropsychologists and neurologists w ere struck by th e severe behavioural deficits of certain patients with frontal lobe d am age w ho w ere unable to carry o u t daily activities and sustain social interactions despite seem ingly norm al intelligence, m em ory an d sensory-perceptual functions (e.g., Feuchtw anger, 1923, and Welt, 1888, as cited by Benton, 1991b; Ackerly, 1937; Brickner, 1934; 1936; H ebb & Penfield, 1940; Hebb, 1945). In the 1920s, G oldstein introduced the notion of "abstract attitude" (1936a; 1936b; 1944), a cognitive skill that encom passed the neuropsychological functions of abstract reasoning, initiative, foresight, self- aw areness, behavioural flexibility, and com plex problem analysis (Benton, 1991b). The loss of "abstract attitude" in certain brain-dam aged individuals w as tho u g h t to explain their behavioural im pairm ents in the absence of deficits in other cognitive abilities. M ore recently, the term "executive functions" has replaced earlier term s to describe those cognitive operations th a t m odulate and control m ore prim ary cognitive skills such as language, memory, perceptual skills, and m otor behaviour.
functioning into com ponent processes have also em erged in the last decade. A review of the better know n m odels follows.
Welsh, Pennington, and Groisser (1991) defined executive functions as consisting of planning, the ability to shift set, im pulse control, and the organised search an d retrieval of items from m em ory (W eyandt & W illis, 1994). U n d er the executive function rubric, Lezak (1995) included volition (the ability to form an intention and form ulate a goal), planning (the ability to identify and organise the steps needed to reach a goal and identify alternative solutions), purposive action (the ability to initiate, maintain, switch an d stop behaviour sequences in an orderly, organised m anner), and effective perform ance (the ability to m onitor, self- correct, and regulate performance). A ccording to her m odel, deficits in executive functions are m ost evident in planned, intentional activities an d m ost ev id en t in situations dem an d in g appropriate an d socially responsible conduct. Lezak noted that th e ability to initiate behaviour and the capacity for self-awareness are preconditions to volitional behaviour. In particular, Lezak posited th at self- aw areness could be broken dow n into the aw areness of o n e's physical self w ithin the im m ediate environm ent, aw areness of self as distinct from the environm ent and from other individuals, and aw areness of the self as p a rt of social interactions.
M ateer and W illiams (1991) view ed executive functions in term s of three general classifications: (1) self-regulatory abilities, (2) the allocation of attentional
Williams posited th at executive functions include the direction of attention (orientation to a specific targ et an d resistance to interference), the recognition of patterns of priority, the form ulation of intentions, the d ev elo p m en t of plans for attainment, the execution of plans, and the recognition of goal attainm ent. According to this m odel, d ism p tio n s in executive functioriing processes cause stim ulus-bound behaviour, perseveration, lack of initiation, difficulty sustaining attention and m otivation o v er tim e, failure to recognise an d use feedback, failure to regulate behaviour in d ependently, and lack of aw areness o f deficits (Mateer & Williams, 1991).
M e a s u r i n g Ex e c u t i v e Fu n c t i o n s
In clinical neuropsychological investigations, the m easurem ent of executive functioning skills poses a n in h eren t challenge. N europsychological testing is a highly structured situation w here the examiner form ulates goals, determ ines projected outcomes, and pro v id es the subject w ith specific instructions so th at the subject performs a task w ith in defined param eters and given m aterials, while the examiner m onitors perform ance. According to Lezak (1995), the problem w ith assessing executive functioning in the context of a neuropsychological assessm ent is to "transfer goal-setting, structuring, and decision m aking from the clinician to
paradigm s have been developed th at allow the m easurem ent of executive functions within the structured context of neuropsychological testing.
Because they allow the subject to independently select a chosen path from a set of possible alternatives according to a self-initiated plan, m aze tracing tasks have been used frequently in research on executive functions (e.g., G rodzinsky & Diam ond, 1992; Porteus, 1959; Segalowitz, Unsal, & Dywan, 1992; W eyandt & Willis, 1994) as m easures of planning and foresight (Lezak, 1995). Other w ell- know n tests that are pu rp o rted to m easure planning are the tow er tests (i.e.. T ow er of London, Tower of Hanoi, and Tow er of Toronto). Briefly, these tests involve problem solving paradigm s th at allow a subject to select responses in o rd er to attain a pre-set goal, according to given rules, such as having to m ove coloured disks from their initial position o n a peg to a different peg, one d isk at a time, in the least num ber of m oves possible (Lezak, 1995). An effectively planned strategy is one that uses the least num ber of moves and does not include any rule violations. In addition to m easurem ent of visual-spatial skills, complex design copy task are also thought to d ep en d on planning ability (e.g., Rey-Osterrieth Com plex figure test; G rodzinsky an d D iam ond, 1992; Lezak, 1995). The cognitive process of interest targeted by design copying tasks is the approach the subject takes in executing the design, rather than accuracy itself. How ever, because of the com plexity of the design, poor planning may be accom panied by inaccurate
unorganised, unplanned, piece-m eal approach to problem solving.
Response inhibition (i.e., control of im pulsivity and perseverative tendencies) has traditionally been assessed w ith Go-No-Go tasks a n d card-sorting tasks requiring set shifting in response to examiner feedback. The W isconsin C ard Sorting Test (WCST; H eaton, C helune, Talley, Kay, & Curtiss, 1993) is currently the m ost w idely used m easure of executive functioning (Stuss & Benson, 1984). Poor perform ance on certain m easures of the WCST such as the perseverative errors score reflects difficulty inhibiting a previously learned response set. W ord generation tasks have also been used to m easure initiation and strategic m em ory search. The subject is presented wdth either a letter or category, and th en instructed to provide as m any w ords beginning w ith the particular letter or exem plars from a category as possible w ith in a given tim e limit. Lezak (1995) reports that perform ance on w ord generation tasks involves the use of strategy to guide the search of m em ory stores, such as the use of conceptual clustering along phonological or semantic categories, in addition to initiation and self-m onitoring skills.
Like m o st cognitive skills, the developm ent of executive functions in ch ild ren ap p ears to follow a n ag e-d ep en d en t pattern (Becker, Isaac & H ynd, 1987; McKay, H alperin, Schwartz, & Sharma, 1994; Passler, Isaac, & H ynd, 1985; Stuss, 1992; Welsh, Pennington, & Groisser, 1991; W eyandt 6 Willis, 1994). The developm ent of executive functions in ch ild ren has been investigated prim arily by assessing children on various paradigm s know n to detect executive functioning deficits in brain injured adults. The age a t w hich children reach a d u lt perform ance o n an y given executive functioning test is therefore thought to reflect the age a t w hich the particular executive function becom es fully developed. A lthough this ap p ro ach has been criticised because of th e underlying assum ption th at ch ild ren 's cognitive skills are in a sense "partially form ed" versions of adult skills instead of distinct cognitive skills th a t are ap propriate, complete, and well suited to each developm ental stage, these types of investigations have contributed to the u n d erstan d in g o f neurocognitive developm ent in children. The follow ing is a brief review of this research em phasising the developm ental progression of perform ance on executive functioning tests.
A dult-level perform ance on executive functioning measures reflecting set shifting, resisting perseverative tendencies, and self-monitoring appear to be reached by age 10 to 12 (C helune & Baer, 1986; C helune & Thompson, 1987; Levin e t al., 1991; Passler a t al., 1985) or older (W eyandt & Willis, 1994) on such tasks as
the WCST and Go-No-Go tasks (Drewe, 1975). O ther research suggests that response inhibition develops m ost rapidly betw een 7 and 9 years of age, reaching ad u lt levels at age 11 (McKay et al., 1994). Planning and strategy use increase substantially from 12 years to later adolescence (Levin et al., 1991; Welsh, Pennington & Groisser, 1991) as reflected by perform ance on the T ow er of L ondon (Shallice, 1982) and T w enty Q uestions procedure (Denny & D enny, 1973). Tem poral ordering, a m em ory organisation skill, attains a peak by age 12 (Becker, Isaac, & H y n d , 1987). O ther m em ory organization skills such as clustering responses into sem antic categories to facilitate retrieval on w ord list learning tasks increase from age 7 to adolescence (Levin et al., 1991) on such tasks as the California Verbal L earning T est - C h ild ren 's Version (Delis, Kramer, K aplan, & Ober, 1986). Perform ance on w ord fluency tasks, a skill that depends on the use of rapid m em ory retrieval strategies, attains a plateau between the fifth and seventh grade (i.e., betw een approxim ately ages 11 and 13; Halperin, Healy, Zeitchik, Ludm an, & W einstein, 1989).
Based on the research of Welsh, Groissier, and Pennington (1988) and other researchers, Stuss (1992, p. 17) proposed a developm ental progression in executive functions which follow s three stages: (1) sim ple planning and organised visual search by age 6; (2) set m aintenance, hypothesis testing, and im pulse control by age 10; and (3) complex planning, m otor sequencing, and verbal fluency d u rin g adolescence. The developm ent of self-awareness, insight and self-reflectiveness.
presum ably m atu rin g later th an other m ore basic executive functions, is not yet well charted, b u t presum ably take a more protracted course.
Ex e c u t i v e Fu n c t i o n i n g a n d So c i a l Sk i l l s
Because executive functions m odulate, organise, and integrate other cognitive processes a n d behavioural responses, the im plications of deficits in this area are far reaching. In particular, Pennington and W elsh (1985) reported that social behaviour requires the integration of m ultiple sources of information, the ability to make inferences regarding others' m ental states, the inhibition of inappropriate re s p o n d in g and the rapid shifting of attentional resources. Because of these requirem ents, Pennington and Welsh posited th at social skills would be expected to be highly dependent on executive functioning skills. O ther researchers have reached sim ilar conclusions. For example, in his exam ination of factors im portant after h ead injury, Ylvisaker et al. (1987) listed a num ber of cognitive deficits th at have adverse effects on social skill. These included Limited awareness of social and com m unicative events, inadequate retrieval of the rules of social interaction, reduced ability to take alternative perspectives, organisational problem s leading to problems introducing m ain tain in g and term inating conversational topics, and disinhibition and self-m onitoring deficits in verbal and non-verbal behaviour causing the brain injured individual to repeat information.
uninterested or unm otivated. All these difficulties, to some extent, appear to reflect deficits in executive functioning.
A lthough sim ilar deficits occur w ith injury to other p a rts of the brain, executive deficits are com m only reported after frontal lobe d am ag e in children (Ackerly, 1964; Dennis, 1991; Dennis & Bames, 1990; Eshnger, G rattan, Damasio, & Damasio, 1990; G rattan & Eslinger, 1991; Mateer & Williams, 1991). Ylvisaker summarised the social deficits th at are commonly rep o rted after frontal lobe injury in children, highlighting th e association between executive functioning deficits and problems in the social realm . These include "disinhibition, socially inappropriate behaviour and language, lack of initiation, difficulty reading social cues and interpreting social situations, difficulty organising com plex responses, rigid social behaviour, and others" (p. 373).
Interestingly, Tranel, A nderson, and Benton (1994) include social conduct as a subset of executive functions. H ow ever, Bennett (1989) counters the view that social consciousness is a n in d ep en d en t cognitive process; he states th at the loss of social consciousness in brain injured patients such as those w ith frontal lobe damage "reflects the general difficulties that a frontal patient exhibits in interacting with the environm ent. H e or she is impulsive, perseverative, m onitors, his or her own behavior poorly, a n d is slow to self-correct" (p. 20). In B ennett's view, social deficits are not caused by dam age to a separate "social m odule", b u t are instead
the behavioural m anifestations of a m ore pervasive loss of executive functioning skills. Presum ably, social situations, because of their complexity, an d by virtue of their requirem ents for initiation, flexibility, an d self-m onitoring, u n m ask executive functioning problem s.
Co g n i t i v e D e t e r m i n a n t s o f So c i a l Sk i l l f r o m N e u r o p s y c h o l o g y
N europsychological explorations of the relationship betw een cognitive skills and social difficulties are of two m ain types: case studies a n d clinicians' reports of social im pairm ent following neurological disorders o r dam age to specific areas of the brain, and empirical g ro u p studies. Both types o f approaches have contributed to o u r understanding of the cognitive determ inants of social skills.
Neuropsychological Case Studies and Clinicians' Reports
U ndoubtedly one of the m ost fam ous cases in neuropsychology, Phineas Gage was a y o u n g w orker w h o sustained extensive frontal lobe d am ag e w hen a large iron bar w as accidentally propelled th ro u g h his left frontal lobe, exiting at the posterior right frontal lobe (Benton, 1991; H arlow , 1848; MacMillan, 1986;
Steegm an, 1962). Remarkably, despite the extensive size of the lesion. Gage show ed preservation of intellect and m em ory. However, although previously an honest an d reliable man, impulsivity, profanity, and p o o r judgem ent lead to interpersonal difficulties and disruption in his social and occupational functioning. Thus, despite selective preservation of certain cognitive skills. Gage was incapacitated, from a social standpoint, by his acquired deficits in im pulse control and judgem ent.
The classic reports of patient JP by Brickner (1936), patient DT (Grattan & Eslinger, 1992; Eslinger, G rattan, Damasio, & Damasio, 1990), and patient EVR by Eslinger and Damasio (1985) have also contributed to clarifying the relationship betw een cognitive functions and social behaviour. JP h ad congenital, bilateral frontal lobe atrophy (Ackerly, 1950,1964; Ackerly & Benton, 1947). JP's scores on tests of general intellectual ability w ere w ithin the average range, y et he had im paired performance on tests m easuring cognitive flexibility, planning, and abstraction. From a young age, his everyday behaviour w as characterised by im pulsivity, poor decision-making, and social problems described as the inability to sustain friendships, and superficiality in interpersonal relationship (G rattan & Eslinger, 1991). Grattan and Eshnger (1991) concluded th at JP had deficiencies in decision m aking that were especially sah en t in his everyday social behaviour. They proposed that his social difficulties w ere influenced by his deficits in
cognitive flexibility. Based on their observations of JP, Ackerly and Benton (1947) characterised his dysfunction as a "prim ary social defect".
G rattan and Eslinger's (1992; Eslinger e t al., 1990) patient DT sustained a subarachnoid haem orrhage at age 7 th at caused bilateral frontal lobe damage. She had executive function im pairm ents that included im paired cognitive flexibility, planning, self-monitoring, and organisation skills. A lthough she initially functioned socially w ithin age expectations, as she grew older, she began to exhibit progressive social and emotional dysfunction. The researchers concluded that DT's deficits in executive functions interacted w ith increasing dem and on social skills as she entered adolescence and adulthood, leading to pervasive social difficulties.
In the case of EVR (Eslinger & Dam asio, 1985), an accountant who un d erw en t bilateral orbital frontal lobe resection d u e to a tu m o u r at age 35, Saver and Dam asio (1991) dem onstrated a dissociation betw een know ledge of sociaUy- appropriate behaviour and the ability to deploy these behaviours in everyday life. Despite norm al-range scores on tests of the ability to generate hypothetical behavioural alternatives to given social situations, EVR w as unable to select and deploy appropriate social behaviours in real-life interpersonal situations, suggesting pervasive executive functioning deficits.
Price, Daffner, Stowe, and M esulam (1990) provided tw o case reports of adults w ho sustained early bifrontal lobe dam age, one at birth and the other a t age
4. Both h a d chronic and pervasive behavioural difficulties, abnorm al social conduct an d lack of empathy. Standard neuropsychological assessm ent revealed deficits in the areas of attention, organisation, a n d m ental flexibility despite adequate m em ory, visual-spatial, and language functioning. D evelopm ental psychology p aradigm s m easuring Piagetian cognitive stages (i.e., sensorim otor, preoperational, concrete operational and form al operations stage), perspective taking, an d m oral developm ent (i.e., the H einz dilem m a) w ere adm inistered to these patients. Consistent w ith their im paired executive functioning on neuropsychological tasks, both patients' perform ance on the developm ental psychology paradigm s was deficient. Their perform ance suggested that they had not developed past the concrete operational stage an d had difficulty inferring the perspective of others. In one case, im m ature m o ral developm ent w as found. In the other, inconsistencies between knowledge of m oral behaviour and deploym ent of m oral behaviour in real life w ere evidenced. These results suggest that bilateral frontal lesions in childhood may lead to deficits in m aturation of judgem ent, insight, foresight, abstract reasoning, perspective taking, em pathy and m oral developm ent th at are translated into aberrant social conduct despite the intactness of other cognitive skills such as m emory, language, an d visual-spatial skills.
M arlow e's (1992) patient PL sustained a rig h t prefrontal penetrating injury a t age 3 d u e to a law n dart. H e had executive functioning deficits including m arked self-regulation deficits despite above av erag e scores on conventional
intelligence tests. H e also had m arked difficulty ad ap tin g his behaviour after feedback (i.e., learning from experience), and h e show ed consistent difficulty in social situations.
Biddle an d Eslinger's (1996) case of an 11 y ear old boy w ho sustained right frontal d am age after an arterio-venous m alform ation a t age 7 highlights the role of executive functioning an d language in m ediating social behaviour. Despite norm al elem entary language and verbal cognitive abilities, p atien t JC had severe expressive language problem s th at consisted of difficulty m anaging discourse production, especially the planning, organisation, flexibility and m onitoring of discourse. The researchers argued that such im pairm ents cause "significant 'listener b u rd en ' w ith potentially disruptive effects on social interaction and academic achievem ent". Ylvisaker (1993) noted th a t other language problems in children w ith brain dysfunction such as im paired w o rd retrieval, and rea d in g /a u d ito ry com prehension problem s m ay actually be related to executive functioning deficits w hich preclude effective search of the sem antic netw ork and failure to use organising schemes to benefit com prehension. In addition, Ylvisaker em phasised th at children with traum atic brain injury m ay have conceptual difficulties th at interfere w ith their com prehension of verbal abstraction (e.g., m etaphors and figures of speech), and u n d erstan d in g im plied and indirect m eanings (e.g., irony), problems th at have significant im pact on their ability to interact in interpersonal situations.
Clinicians w o rking w ith neurological populations have often observed that im pairm ents in executive functioning skills, abstraction, an d reasoning skills ap p ear to be accom panied by social and occupational difficulties (Hart & Jacobs, 1993; Sohlberg & M ateer, 1989; V am ey & Menefee, 1993). In particular, im pulsivity and poor judgem ent are th o u g h t to contribute to interpersonal difficulties in children w ith traum atic brain injury (McGuire & Rothenberg, 1986). Clinical im pressions also suggest that head-injured persons with executive functioning deficits have reduced em pathy (Cicerone, 1989; Lezak, 1993; Restak, 1984).
In sum m ary, case studies and clinical observations suggest that executive functions (i.e., im pulsivity inhibition, cognitive flexibility, planning) as w ell as other cognitive skiUs such as judgem ent, reasoning, conceptual ability and abstraction, are im p o rtan t variables in m ediating social behaviour. O ther difficulties such as executive-related language problem s and attention deficits also appear to be implicated.
Neuropsychological Empirical Group Studies
There exist few controlled, empirical studies of social skills in the neuropsychological literature. Existing studies cover a range of clinical syndrom es such as alcoholism, head injury, Tourette syndrom e, and other conditions. A
unitary definition of social skills has n o t been u sed across studies, a fact th at complicates interpretation of existing findings. In addition, the w ays in w hich social skills are operationalized vary across studies. Specifically, heterogeneous m ethods have been used in previous research, including reports of clinical impressions, sociometric m easures of likability, observational m ethods, and social skill scales from m easures designed to measure general psychopathology. Despite these limitations, the contribution of these studies to o u r u n d erstan d in g of social behaviour com plem ents case studies and clinical reports th at provide a prelim inary delineation of the relationship between cognitive a n d social skills.
Perrott, Taylor, a n d M ontes (1991) found th at perform ance on m easures of auditory attention (WISC-R Digit Span) and inhibition/ set shifting differentiated between head-injured children w ho w ere socially or behaviouraUy disordered and those whose social behaviour w as w ithin the norm al range. In one stu d y on norm al children, conceptual ability an d flexible strategy use (i.e., num ber of categories and perseverative errors on the WCST) were related to social adjustm ent (Social Problem s subscale of the Teacher version of the C hild Behaviour Checklist) in children aged 9 to 11. H ow ever, the variables w ere n o t significantly related for children younger than 9 years or o ld er than 11 years (Riccio, Hall, Morgan, e t al., 1994).
In another study on m ild to m oderate head injury in children, Papero, Prigatano, Snyder and Johnson (1993) found that boys b u t not girls had borderline
to low average social skills o n the Socialization subscale of the parent-rated V ineland A daptive Behaviour Scale. E xploratory item analysis suggested that social problem s in boys stem m ed from n o t follow ing the rules of social convention and peer interaction. The researchers posited that dysfunction in th e executive control of behaviour such as problem s w ith im pulse control and self-m onitoring m ight lead to failures in exhibiting ap p ro p riate social behaviour in children w ith brain injury.
Stokes, Bawden, C am field, Backman, and Dooley (1991) exam ined the question of neuropsychological test predictors of social adjustm ent in children w ith Tourette syndrom e. U sing a w ell-know n peer n o m in a tio n / sociometric m easure, the Peer Evaluation Inventory, they found th at several verbal tests were positively correlated w ith a peer-rated m easure of likability, a construct th o u g h t to reflect social skills. These included the W echsler Intelligence Scale for C hildren Verbal IQ, as w ell as a m easure of fund of general inform ation (Inform ation subtest) an d a verbal practical reasoning test (Com prehension subtest). A m easure of expressive vocabulary (Peabody Picture Vocabulary Test-Revised) w as also related to the sociometric m easure. In addition, a m easure of verbal m em ory and tw o subtests from the Perform ance scale of th e Wechsler test (Picture C om pletion and Picture A rrangem ent) w ere also m oderately correlated w ith the likability m easure. Achievement, visual-m otor integration, atten tio n /im p u lsiv ity , and m otor tests w ere n o t significantly correlated w ith likability m easures. C onsistent
w ith the view th a t non-verbal perceptual abilities may be related to social skiUs, Peterson (1988) fo und a strong correlation betw een social interactive competence and the ability to detect m eaning in facial expressions.
In ad u lts w ith head injury, Levine, Van H orn, and C urtis (1993) exam ined the ability to problem solve effectively and interpret interpersonal conflict situations presented in the form of stories. H ead injured subjects show ed a more concrete and less socially m ature conceptualisation of the interpersonal situations com pared to controls. In a sim ilar vein. Marsh an d Knight (1991) found th at head injured adults' level of social skiU in a real-life interaction w as related to a verbal fluency task, suggesting th at inflexibility in processing contributed to difficulties d u rin g social interactions. In contrast, level of social skill w as not related to m em ory skills. In a n earlier study, the researchers found sim ilar results in that level of skill d u rin g social interaction w as not related to m em ory, intellectual, or speed of inform ation processing in very severely head injured adults (Marsh, Knight, & G odfrey, 1990). N ew ton and Johnson (1985) also found that intellectual ability was n o t related to level of skill in a videotaped social interaction in adults w ith severe head injury.
McKay an d Ramsey (1984) examined the relationship betw een a peer nom ination m easu re in a d u lt alcoholics seen for g ro u p therapy and neuropsychological test perform ance (i.e., Luria N ebraska Neuropsychological Battery). The researchers concluded that sociometric ratings reflecting group
m em bers' ratings of competence d u rin g therapy w ere related to com m unication skills such as verbal comprehension a n d fluency, non-verbal perceptual skills, and general intelligence. Perceptual skills w ere hypothesised to facilitate aw areness of self an d of others' affective state. M cKay and Ramsey (1984) also em phasise that social behaviour requires a variety o f perceptual, motor, and intellectual skills. In particular, they posited th at "right-hem isphere" non-verbal skills such as prosody and appreciation of hum our m ight also play an im portant part in m ediating social skills.
Th e o r i e s o f So c i a l Be h a v i o u r a n d So c i a l Co m p e t e n c e
To date, existing theories aim ed a t explicating social skills in children have been developed by clinical, developm ental, and educational psychologists. C u rren t conceptualisations consist o f social information processing or social- cognitive problem solving models (e.g.. Dodge, 1980, 1986; McFall, 1982; Pettit, 1992; Spivack, Platt & Shure, 1976). A review of three of the m ore influential m odels, nam ely those proposed by Spivack and Shure (1974), Rubin an d K rasnor (1986), an d D odge (1986), highlights the im portance of cognitive variables in m ediating social skills. M ore specifically, review of these m odels dem onstrates
their sim ilarities to definitions of executive functioning such as those proposed by M ateer an d Williams (1991), Lezak (1995), and W elsh et al. (1991).
Spivack and Shure (1974) proposed th at children's social competence depends on specific cognitive operations consisting of (1) the recognition th at an interpersonal problem n eed s to be solved; (2) the ability to generate alternative solutions to the problem ; (3) m eans-end th in k in g including the ability to sequentially consider the steps required to reach the goal; (4) causal thinking or the ability to foresee the consequences of future actions; and (5) the ability to identify the future behaviour and current m otives of others (Rubin & Rose-Krasnor, 1992). This m odel em phasises th at perspective taking and the ability to appreciate consequences are essential com ponents of effective social functioning. In line w ith the m odel, Shure (1981) found th at although four-year-olds could n o t spontaneously plan the sequential m eans to a goal, they could produce alternative solutions w h e n asked (i.e., tell w hat m ight happen next). C hildren w ho were poor a t providing alternate solutions were less liked by their peers th an children w ho were able to generate solutions w hen prom pted (for a case exam ple of successful alternate solution generation from Shure, 1981, see Appendix A). Shure (1981) also successfully dem onstrated that problem -solving skills were related to social competence by training these skills in children and then assessing social skills post training.
Rubin an d K rasnor (1986; cited in R ubin & Rose-Krasnor, 1992) introduced the idea of autom atic scripts consisting o f overleam ed actions used in social interactions (for instance, greeting and leave-taking behaviour). Script-driven behaviour, how ever, is precluded w hen unexpected changes occur in the environm ent, or w h e n the enactm ent of th e script is p rev en ted in som e way. In this case, flexible strateg y use is needed in o rd e r to function effectively in the social situation. Rubin an d Krasnor posit th a t social functioning requires several cognitive operations; (1) selection of a goal, including form ing a representation of the projected end-state of the problem -solving process; (2) exam ination of the task environm ent; (3) assessm ent and selection of strategies, including the retrieval of previously stored strategies, dependent in p a rt on the autom aticity of retrieval and the size of the strategy repertoire; (4) strategy outcome assessm ent requiring a feedback system th a t enables com parison of the initially form ulated goal to the end result, w ith re-initiation of the problem -solving sequence if required. A ppendix A, Case 2, illustrates th e use of scripts in children's social interactions.
The m odels p u t forth by D odge are some of the m ost influential and em pirically-tested in their field (Pettit, 1992). For exam ple. D odge (1986) first proposed a five-stage model of social inform ation processing th at involves successive steps: (1) encoding (sensation, perception, and attention), (2) representation (integration of cues to m em ory stores, application of decision rules, feedback to encoding system, and interpretation), (3) response search (generation
of potential responses, application of response rules), (4) response decision (representation of consequences, evaluation of outcomes, feedback to response generation, and selection of response), and (6) enactm ent (use of scripts, m onitoring of enactm ent, and self-regulation).
More recently. D odge and colleagues revised their m odel to reflect an em erging view th at social cognition does not necessarily reflect a sequential, linear process but a m ore plastic, on-line process w here com ponents interact in a dynamic fashion (Crick & Dodge, 1994). They proposed a six-com ponent system composed of feedback loops and a cyclical path. Their m odel attem pted to incorporate the view th a t processing occurs in sim ultaneous parallel paths instead of a rigid sequential structure. The six cognitive processes interact in an ongoing fashion w ith a proposed "d ata base" consisting of acquired rules, social schemas, and social know ledge, stored m long-term m em ory. These stored m ental representations are retrieved in order to guide the processing of n ew social cues. The cognitive com ponents are (1) encoding of cues, (2) interpretation of cues (including evaluation of goal attainm ent and self-evaluation, (3) clarification of goals (including arousal regulation), (4) response access or construction, (5) response decision (including self-efficacy evaluation and response selection), and (6) behavioral enactm ent. Crick and D odge posit th a t individuals are continually and sim ultaneously engaged in m ultiple inform ation processing activities, instead of sequentially em ploying each cognitive com ponent in a step-by-step m anner
until a behavioural goal is reached. For example, a person m ig h t engage in interpretation w hile encoding cues, and continue to evaluate the m eaning of another person's behaviour w hile sim ultaneously accessing a behavioural response (p. 77). N evertheless, Crick an d Dodge em phasise th at a sequential m odel is a useful heuristic in understanding the path from cue to response. In o th er w ords, the path from o n e stim ulus to a behavioural response m ay generally follow a time- related linear structure, even tho u g h feedback loops connect com ponents in a non linear manner.
As illustrated by these models, the developm ental literature suggests that elements of the executive functioning skills, identified in a parallel by neuropsychologists, are im portant determ inants of social skills. C om m on to developm ental approaches is the view that social problem solving consists of a sequence of cognitive processes th at includes attention to cues, consideration of alternatives, decision-making, an d enactm ent (Pettit, 1992). Several o f these skills, in particular, the ability to resist impulsive responding, engage in strategic m em ory search a n d retrieval, and dem onstrate flexible thinking while considering alternative courses of action, easily fall u n d er the neuropsychological label of executive functions. Executive function models developed by neuropsychologists, m ost notably those by Stuss (1992), Mateer and Williams (1991), and Sohlberg and Mateer (1989) are rem arkably sim ilar to social inform ation processing m odels such as th at of Crick a n d Dodge (1994). The difference is the type of behavioural
outcome specified (i.e., restricted to the social realm in social inform ation processing m odels, o r applicable to all behaviour in the case of executive functioning models). O ther non-executive cognitive operations are outlined in social inform ation processing m odels, such as the ability to en g ag e in effective problem solving using logical and conceptual abilities. The sim ilarities betw een neuropsychological m odels of executive functioning and social inform ation processing m odels su g g est th at neuropsychological m easures m ay be useful tools in m apping the relationship betw een cognitive skills an d social skills.
M e a s u r i n g So c i a l Sk i l l s
The selection of tests to m easure particular facets of behaviour is alw ays complex. In social skills research, three questions are particularly relev an t (1) w hat is the accepted definition of socially skilled behaviour?; (2) how are social skills operationalized?, a n d (3) w ho is the rater?
G resham (1986) notes th at an adequate conceptual definition of social skills has not yet been derived, perhaps because the field has been m ore concerned w ith treating social skill deficits th an w ith deriving adequate assessm ent tools. D odge noted that there m ight be as m any definitions of social com petence as there are researchers stu d y in g the phenom enon in this field (1985, as cited in Rubin &
Rose-ïCrasnor, 1992). G resham (1986) notes that definitions of social skills have ranged from the abstract a n d vague (e.g., "the ability to interact [...] in specific w ays that are socially acceptable and valued and at the sam e time are personally beneficial, m utually beneficial, o r beneficial prim arily to others"; Com bs & Slaby, 1977, p. 162) to the concrete an d vague (e.g., "th e complex ability to both em it behaviors that are [...] reinforced an d n o t to em it behaviors th a t are punished or extinguished by others"; Libet & Lewinsohn, 1973, p. 304). Despite these disparities, certain identifiable view points on the conceptual natu re of social skills have evolved.
Gresham (1986) em phasises the distinction betw een social skill as a
behavioural construct an d as a trait. The behavioural construct m odel (also know n as
the m olecular m odel) operationalizes social skills as a category of discrete, situation-specific, observable behaviours. In contrast, trait theorists define social skills as a hypothetical, global personality trait or predisposition th at is cross- situational. A lthough the trait m odel has received lim ited empirical support (McFall, 1982), the behavioural m odel, in its q u e st to reduce the construct of social skill into observable behaviours, has also incurred some problem s. McFall (1982) outlines these difficulties. First, th e question of w hich specific behaviours should be included in the category of social skill behaviours has n o t been adequately resolved. Second, the characteristics of a social skill situation have n o t been clarified (e.g., is it the physical characteristics of the situation or the persons
involved). T hird, the definition of outcom e for defining social efficacy (i.e., success vs. failure, long- vs. short-term outcome) is not clear.
In the developm ental and educational field, social skills have been operationalized in a variety of ways, including sociometric indices, rating scales, and naturalistic observations of social interactions. A w ell-know n technique, the sociometric m easure, involves instructing groups o f children (typically an entire class) to nom inate other children for certain tasks o r roles, o r to rank their classmates o n different dimensions of likability. In th is conceptualisation of social skills, popularity is equated with social skill. Specifically, those children nom inated o n the highest num ber of positive dim ensions, or nom inated most often by other children as being a friend, are defined as socially skilled. A lthough sociometric m easures are considered by some as the gold stan d ard of social m easurem ent (La Greca, 1981; Vosk, Forehand, Parker, & Rickard, 1982), sociometric ratings do n o t provide an u n d erstanding of the specific behaviours that lead to p eer acceptance or rejection (Gresham, 1986; Van Hasselt, Hersen, Whitehill, & Bellack, 1979), an insight that is crucial from a n intervention standpoint. Additionally, it is not clear that p o p u larity is alw ays synonym ous w ith good social skills, or th at social rejection invariably indicates social skill deficit. Evidently, n o t all rejected children are necessarily socially unskilled. O ther factors such as illness, disability, socioeconomic status, or other factors may also im pact acceptance and popularity am ong children. Im portantly, p eer nom ination
measures that provide indices of popularity and rejection have been criticised because of the risk of disclosure of nominations by children to each other (La Greca, 1981).
Another w ay to m easure social skills is by purely behavioural means. According to this definition, social skills are those behaviours that are reinforced in specific interpersonal situations. In this method, trained observers in naturalistic or role-play paradigm s record behaviour in vivo. Deficits or excesses of specific behaviours are deem ed to be indicative of social skill problems. G resham (1986) notes that the advantage of this approach is that the antecedents a n d outcom es of specific behaviours can be identified. The disadvantage, of course, is that there is no guarantee th at the specified behaviours are actually relevant in o ther contexts.
A third m ethod is to use questionnaires using items th at reflect specific behaviours that have dem onstrated validity as indices of social competence, using raters that habitually see children interacting with their peers in several contexts. Observer-rated questionnaires have advantages over sociometric ratings an d naturalistic observation in th at risk of rating disclosure is m inim ised, d ata collection time is m uch reduced, and cross-situational stability is relatively ensured. Clearly, questionnaires designed specifically to m easure social skills provide more specific inform ation on actual social skill behaviour th a n do broad band questionnaires (e.g.. C hild Behavior Checklist, Personality Inventory for
Children) w hich w ere developed as screening m easures for general behavioural problem s an d as aids in diagnosing clinical psychopathology.
Lastly, the ty p e of rater used in a psychometric m ea su re of social skills influences the validity of th at m easure. Teacher-rated indices of social skill have dem onstrated stro n g relationships to other social skill m easures such as sociometric status a n d are considered valid and reliable w ays of m easuring social skills in children (G resham , 1986). Parent-rated m easures o f social skills have received less attention in the social skills field; the validity of these m easures is therefore currently unknow n. Self-report m easures of social skill o r social competence in children are not recom m ended (Gresham , 1986). Self-rated m easures derived to d a te show little relationship to peer acceptance, popularity, teacher ratings, role-play perform ance, or naturalistic observation (G resham , 1986).
S U M M A R Y ' O F THE L IT E R A T U R E O N C O G N I T I V E C O R R E L A T E S OF S O C I A L SKILLS
Developm ental psychology m odels of social inform ation processing em phasise attention, executive functioning, and problem solving (logical reasoning and judgem ent) as im p o rtan t social skill determ inants. In a sim ilar vein, the neuropsychological literatu re suggests that several different cognitive skills m ay be related to social skills. These include (1) executive functioning, particularly self
monitoring, the ability to inhibit impulsivity, and language-based m anifestations of executive function, as w ell as executive skills enabling the p roper d ep lo y m en t of social knowledge; (2) judgem ent, reasoning, and problem solving; an d (3) non verbal perceptual skills. Som e researchers and clinicians have also implicated memory, language, and se n so ry /m o to r skills.
A lthough the contribution of neuropsychology to ou r u n d erstan d in g of social behaviour is im portant, som e im portant lim itations deserve to be noted. First, a unitary definition of social skills has not been used across studies, a fact that complicates interpretation of existing findings. In a related point, the w ays in which social skills are operationalized vary across studies. Specifically, heterogeneous m ethods have been used in previous research, including reports of clinical impressions, sociometric m easures of likability, and social skill scales from m easures designed to m easure general psychopathology. Care in choosing a definition and a w ay of m easuring social skills in neuropsychological investigations is im portant, g iv en th at clinical, educational, and developm ental psychologists are still debating the m erits and Limitations of existing m easures of social skill. Indeed, no study to d ate has used a psychom etrically ro b u st m easure th at was developed specifically for m easuring social skills. In a d d itio n to case studies and clinical reports, few em pirical studies have been co nducted on the cognitive correlates of social skills in the neuropsychological literature. There is
thus a need to verify previous findings using well-validated m easures of social skill in g ro u p s of interest.
As m easured by neuropsychological tests, executive skills ap p ear to develop in a n age-dependent m anner, as do m any other cognitive skills. A lthough the literature is n o t definitive as to w hich executive skills m ay be m ore im portant in m ediating socials skiUs a t different ages, the neuropsychological literature suggests th at there is an age-dependent pattern of developm ent, w ith m ore basic functions such as response inhibition developing early, an d m ore complex functions such as set shifting, planning and strategy use developing later. Nonetheless, it is clear th at the ability to take p art in successful social interactions does n o t begin exclusively once a child has developed a d u lt perform ance on neuropsychological m easures of executive skills. Despite po o r perform ance on certain executive functioning tests relative to older children, the majority of younger children exhibit age-appropriate social behaviour w h en interacting w ith their peers. If social skills are indeed d ep en d en t on specific cognitive skills at all age levels, then different cognitive skills m ay be m ore critical for social interactions at different ages. For example, attention skills m ay be better predictors of social skills th an executive skills in 6 year olds. Conversely, attention skills m ay no longer be sufficient to m aintain social interactions a t age 10, if the social interactions of older children also require higher-level cognitive skills such as logical reasoning. For these reasons, the im pact of age on the cognitive correlates of social skills m ay
be a crucial factor in determ iriing a p p ro p riate m odels. W ith regard to the issue of age-related differences in social skill correlates, previous researchers have noted the im portance of accounting for age in social information processing m odels (Crick & Dodge, 1994). O ther Limitations of existing models are a failure to account for changing social-environmental d em an d s w ith age (Gottman, 1986).
Crick and Dodge (1994) h av e proposed some w orking hypotheses regarding the issue of age and social skill developm ent. Specifically, they posit that the data base of social know ledge, including the stored representation of an increasingly skilful strategy repertoire, expands and changes as children are exposed to new experiences with age. In addition, children's attentional ability increases, providing greater accuracy (e.g., attention is focused more often on relevant vs. irrelevant cues) and speed of inform ation processing. W ith age, children also develop m ore efficient w ay s of representing, organising, an d interpreting social information.
G o a l s a n d Hy p o t h e s e s
This stu d y had tw o m ain goals: (1) to identify specific tests th at m ight serve as useful m easures for p red ictin g social skill problem s in clinically referred children, an d (2) to a n sw e r theoretical questions o n the relationship betw een neuropsychological functioning an d social skills u sin g stan d ard ized instrum ents designed expressly to m easu re social skills. From a statistical perspective, tw o general approaches w ere therefore ta k e n to data analysis. Both approaches had their strengths an d lim itations. In th e first approach, analyses w ere test-driven (i.e., d e p en d e n t variables consisted o f the neuropsychological test scores), so as to enable a test-based in terp retatio n o f results. Lim itations included the large n u m b er of tests involved, a n d the resu ltin g risk of increased T ype I error. The second ap p ro ach involved form ing linear com binations of test scores th at m easu red sim ilar constructs, and assessing the relationship of these com posite factors to m easures of social skills. T he interpretability of the results from the perspective of particular tests w as th erefo re less direct; how ever. Type I erro r w as m inim ized by red u cin g the large n u m b er of ind iv id u al d e p en d e n t variables to a sm aller n u m b er of factors. In a d d itio n , this ap p ro ach allow ed for testing theoretical assum ptions reg ard in g th e relationship betw een tests p u rp o rted to m easure sim ilar abilities. A lthough perfect correspondence w as n o t expected, it
w as hypothesised th at separate factors for verbal, visual-spatial, memory, executive, an d m otor skills w ould em erge from d a ta reduction techniques.
The role of m aturation on the cognitive correlates of social skills w as of major interest in this stu d y . Thus, a n im portant hypothesis w as that social skill predictors w ould differ according to age, with m ore basic skills predicting social efficacy at younger ages (e.g., such as attention, verbal ability, non-verbal ability, m otor skills), and more complex, high-level abilities predicting social skills in older children (e.g., executive skills, conceptual ability, problem solving). For this reason, the g roup was d iv id ed into tw o age groups: (1) children 6 to 9 years of age, and (2) children 10 to 13 years of age. Age groupings w ere based on the neuropsychological literature on the developm ent of executive functions reviewed above, w hich seemed to suggest rapid developm ent of executive after m iddle childhood.
Previous studies reporting a lack of association betw een cognitive skiUs and psycho-em otional functioning in the learning disabilities field have been criticised for ignoring the heterogeneous nature of clinical sam ples, and treating patient groups as if they consisted exclusively of children w ith sim ilar characteristics (Fuerst & Rourke, 1995), thus ignoring potentially im portant information on subtype-specific associations between cognitive skills an d psycho-emotional functioning. Therefore, a m o re exploratory goal of this stu d y w as to derive social skill subtypes using cluster analysis techniques, and relate these to (1)
neuropsychological functioning, (2) age, a n d (3) o ther behavioural correlates of social behaviour. A related goal w as to carry o u t a replicability check of the cluster solution obtained. Fuerst and Rourke (1995) em phasise th a t m ultivariate subtyping techniques such as cluster analysis will alw ays produce data groupings, even if random data are employed. Further, they note th a t the clustering analysis literature provides limited evidence of the superiority of som e clustering m ethods over others. A replicability check, by deriving different cluster solutions through different subtyping techniques is therefore useful in determ ining the robustness a particular clustering solution. Cluster solution replicability has been used successfully in deriving neuropsychological subtypes in o th er patient groups, such as learning disability and head injury (e.g., Fuerst & Rourke, 1993; Deshpande, Millis, Reeder, Fuerst, & Ricker, 1996).
Given the assum ed developm ent of self-awareness w ith age, another goal was to conduct exploratory analyses involving the relationship betw een insight and cognitive functions as m easured by neuropsychological tests. Insight was operationalized as the absolute value of the discrepancy betw een teacher- and self- rated social skills. Thus, larger discrepancies betw een teacher- and self-rated social skills w ere interpreted as reflecting low er insight. It w as hypothesised that teacher-self discrepancies w ould decrease w ith age, reflecting the developm ent of self-awareness w ith age.
METHOD
Pa r t i c i p a n t s a n d Pr o c e d u r e
In ord er to obtain d ata for this study, files for all neuropsychological assessm ents conducted betw een 1991 and 1996 w ere surveyed a t the Q ueen A lexandra C entre for C hildren's Health, a children's treatm ent an d rehabilitation centre in Victoria, B.C. The list of confidential identification num bers of children having had neuropsychological assessm ents at this facility w as generated, in o rd er to identify children for possible inclusion in the study.
Each of these children had been referred for neuropsychological assessm ent by their physician, school psychologist, social w orker, occupational therapist, or parent. Each assessm ent w as conducted by a Ph.D. level child neuropsychologist, and included a full neuropsychological battery assessing general cognitive ability, achievem ent, executive functions, attention, co n cep tu al/reaso n in g ability, language, m otor skills, visual-spatial skills, and m em ory. A lthough a standardised battery w as used in m ost cases, individual patients needs som etim es necessitated abbreviation or alteration of the standard battery. As p a rt of the neuropsychological assessment, parents, teachers, an d children completed questionnaires on the child's behaviour an d psychosocial functioning. Each
assessm ent also included parent an d child interview , detailed teacher questionnaires an d adm inistration of a one- to tw o-day neuropsychological test battery by a trained psychom etrist, clinical neuropsychology g rad u ate student, or intern. The assessm ent also included generation of a four to six page report with recom m endations for parents, physicians, and teachers, feedback session w ith the neuropsychologist and parents to discuss test findings, an d school conferences to discuss the implications of the findings for the child's academ ic functioning. Other feedback sessions w ere arran g ed d epending on the in d iv id u al needs of each child. N europsychological assessm ent reports w ere sent to referring physicians and to o th er sources w ith parental consent. All neuropsychological an d psychosocial tests w ere scored using standardised norm ative data from test m anuals or other published an d unpublished sources, either by hand or u sing com puterised scoring packages. D erivation of norm -based standardised scores, percentiles, and clinical classifications w as checked a t least once for accuracy p rio r to rep o rt w riting. All paren ts and children were given the o ption of signing a consent form to allow use of their data in research, prior to initiation of the assessment.
From the initial list, approxim ately 130 patient files w ere considered for inclusion in the study. C riteria for inclusion in the stu d y included (1) signed consent for participation in research, (2) age betw een 6 and 17, (3) adm inistration of the standard neuropsychological test battery, and (4) presence of a completed p a re n t or teacher social skill rating form. Each file w a s surveyed for inclusion
