THEEFF ECTI VENESS OF A SHORT-TERMI~l'I'ERPERSONAL PROBLEM-SOLVINGPROGRAM FOR AGGRESSIVE
PRESCHOOLCHI LDREN WI.TH LOW ACCEPTANCE AMONGPEERS
BY
CDonna Michelle Bennett, B.Se.(Ha ns .)
Athesis submitted toth e Schoolof Graduate Studies in partial fulfillmen t of th e
requirementsfor thede g r e e of Mas terof Science
Departmentof Ps yc holo g y Me mo r i a l UniversityofNe wfo und land
Fe b r ua ry1995
St.John's Newf ou n dla nd
Probl em-S olving i i Abstract
The purpose of th e present st udy was to evaluat e the effectiveness of a short-term int e r p e r s o na l problem-solving programfor preschool childr en who are perceived byteac he rs as aggressive and have lowsocial acceptance among their peers. Thirty-fourpreschool ch ild r e n identified asreje c t e d (5 females,11 ma l e s)or average (1 0 females, 8 males ) based on ratings of peer acceptance andteacher ratings of behavior in the preschool settingwere assigned to either lreatment or attention control groups. Resultsindicatedthat childrenin the treatmentgroup demo nstratedsignifican::. imp r o veme nt in problem-solving skills following 13 training sessions and gains were maintainedat a 4 week follow -up.Child r e nin the treatment group also showed a significant decrease in aggressivebe h avio r at fol low-up.Trainingwas not found tobe differentiallyeffective for rejectedand average child ren . There were no significant changes in ratin3s of peer acceptance.The implications ofthe s efi ndi ng s and sugges tions for future res e arch are discussed.
Probl em~ S olving iii Acknowledgements
To my familyand husban d ,Wayne fortheir encouragement andlo ve. To my supervisor,Dr. ChristineArlettwho provided me wi th the necessarydirectionand who never gave up on me.
Tomyco mmi t t e e , Dr.Al Kozmaan d Dr. Mary Courage fortheir helpful comme n t s and advice.ToWayne Rideoutand DavidOsmond for the ir support and fait h in me. To th e pr eschools and childr en who as sistedme withthisstudy.
Problem-Solving iv Table of Contents
PAGE
Abstract .
Acknowledgments ..
Table of Contents
List of Tables .
List of Figures .
Introduct ion .
Reviewof the Literature As s e s s ment of Soc ialStatus
As s e s s me n t ofSoc i al Status withPreschool
Children .
i i iii iv vii ix
Summary 12
SocialStatusand Aggression 12
Stabil ityof Social StatusandAg g r e s s i o n 16
Summary 18
Interpersonal Problem-Solving 18
InterpersonalProblem -solvingSkillsof Pr e s c h oo l Children ..
Summary .
Problem-Sol vingIntervention wi t h Preschool
Children .
Summary .
The Present Study
Me t h od .
20 23
24 35
3.
41
Problem-Solvin g
participants Measures .. Procedu re
pr-etraatrae nc Treatment .
soactreatmcnt;and Follow-up Results ...
Pret:reatmentAnalyses .
Treatment andPo I Low- upAnalyses Assessme nt of Social St at u s
Ass e s s ment ofBe hav i or alAdj u s t me nt .. . Assessment of problem -solvingSkills
Discussion .
Effectivenessof Treatment Effects on Problem-Solvingskills Effects onBehav ioralAdjus tmen t
Ef f ec t s onSocialStatus .
Ot he ref f e c t s . Summar y
PAGE
"
44 50 50 51 53 53 53 57 67 67 71
a' as as ee
a'
"
"
FutureResearch 92
Re f e re nce s
Appe ndixA Consent Le tter Se nt to Parents . 95 107 Appen d ixB Pr eschool Int e rpe r s o n al Pro b l em-Solving
(PIPS) Tes t 110
Probl em- S o l v ing vi PAGE AppendixC What Happens Next ? Ga me (WHNG) 115 Appendix D Tra i ni ng Se s s i o n s : Treatment Gr ou p ., 119 Append ix B Sessions: AttentionCont r o l Group 145
Table
Problem-Solving vii Li s t ofTables
PAGE
Summary of anaLyse s of varianceon age, receptive vocabulary,social status, preschool behavior questionnaire (total), and factor 1(a gg r e s s i o n) at pretreatment 55 Means and standard deviations on
pretreatment measures for treatment
and attention control groups 56
Summary of analysisof variance on social
status overtime 58
Summary of analysisof variance on preschool behavior questionnaire (total)over time ...
Summary of analysis of varianceon factor 1 (aggression) over time . Summary of analysis of varianceon preschool interpersonal problem-solving test over
time .
Summary of analysis of variance on preschool interpersonal problem-solving test (part 1) over time .
59
60
61
62
63 Pr ob lem - Solv ing vii i
PAGE
Summary of ana lysis of vari ance on preschool interpersonal problem-solving te st (p a r t 2) over time
Summary of analysisof variance on wha t happensnext?ga me over time 10 Su mma r y of analysis of variance on what
happens next?game (part 1)ov er time ... 65 11 Summaryof analysisof variance onwhat
ha pp e n s next?game (p a rt 2)overti me .. . 66
Figure
Prob lam-Sol v ing ix Lis t of Figure s
PAGE
Group meansover timeon the preschool behaviorquestionnaire (total) 58 Group means overtime on factor1
(agg r e s s i on) of the preschool behavior
questionnaire 70
Treatmentmeans overtime onfa ctor1
(a gg r e s sio n ) of the preschoolbehavior
questionna ire .
Treatmentmeans over time on th e preschool inte r p e r s onal problem-solvingtest
(t ota l ) .
Tr e a t me n t meansover time on part 1 (p e e r problema) ofth e preschool in t e r p e r s o nal
problem-solving te s t .
Groupmeans overtime on thewhat happens
next z game (t ota l) .
'J'reatment me a n s over time on the what 72
74
76
78
happ ensnext?game (to tal) 80 Groupmean s over time onpart1 (p eer problems) of thewha tha ppe n s nex t?
ga me. . ... . ... .. 82
Problem-solvi ng
Treatment means over timeon part1 (pe e r problems) of the what happens next?
game ...••.. ... . . .. . ..• .
PAGE
83
Proble m·Solving Int roduc:tio n
Inre c ent years, bo ththe importanceof children's pee r relations in the development of so c i a l competence and theeffectiveness ofin t e r vention s aimedat improving poor peer relationshave received considerableattention . Concern about the quali ty of children'speer relations has been motivatedinlarg e part by findings fromstudies suggesting tha t poor peer relationsareassocia tedwith ad j u stme n t prob lemslaterin childhood or adulthood (f or reviews, see Kupersmidt, Coie, s,Dodge, 1990; Parker&
Asher, 1987) . Poor peerrelat ionshave been showntobe relatedto schoc.;dropouts (Ollendick, Weist, Borden, &
Greene, 1992 ), behavioralproblems and psychopatho logy in adolescence (c o r e , Lochman, Terry, & Hyman, 1992;
Kupersmidt & Coie, 1990; Morison & Masten, 1991;
Ollendick et al.,19 9 21,externalizingproblems in middle childhood (Hymel, Rubin, Rowden, & Le Ma r e, 1990 ) , criminal behavior and delinquency (Kup e r s mi d t Patterson,1991;Ol lendicket al., 1992; Roff, 1961; f.ofE
&Sells, 1968),academic difficulties(Bo nn ey, 1971; Li,
1985; Ollendick et al., 1992 ), youngadult psychosis (Roff, 19 6 3) , and psychiatric pr o b l e ms during adulthood (Cowen, Pederson, Babigian , Izzo, &Trost, 1973).
Problem·Solv ing AS will be discussed inthe literaturereviewthat follows, children's peer relations have typicallybeen assessed by determining a child'ssocial status or level of social standing relative to otherch ildrenin a peer group. There is a growing body of evidence to suggest that the types of behaviors children display among peers play an important role in determining their social status. One fi ndingthat has eme rge dconsist ently from studies examining the behavioral correlates of social status is that rejected children, across a wid e age range, are described by peers , teachers, andind e p e nde n t
observers aggressive, disruptive and
uncooperative th a n their more accepted peers. A predominant link between aggression and peer rejection suggests that reducing tHe frequency of these behaviors should be considered an essential feature ofin t e r v e n tio n programs designed to imp r ov e the peer re l a tio n s of childrenidentified tobe at risk.
One approach to reducing the frequency of aggressive behavior is direc:t behavioral intervention where discrete observable behaviors are targeted. An alternative approach is to target in t e r p e r s o na l problem-solving skillsby teachi ng childrenskills such as alternative
Problem-Solvi n g thinking and consequential thinking, as research suggests that poorly adjusted children are deficient in problem- solving skills. For example, poorly adjusted children have been found to generate fewer, more aggressive solutions to interpersonal problems than do their more adjusted peers. Researchsuggests that: through training in interpersonal problem-solving, a child learns to generate a wider range of more appropriate. nonaggressive solutions to interpersonal problems, thus strengthening the re l a t i ons h i p between problem-solving abi lities and behavioraladjustment.
In thepresent study, aggressive preschool children with lowacceptancealiiong peers were selected and trained in interpersonal problem-solving skills. Specifically, the study was designed to examine the following research questions with respect to this population:
1. Is th e r e a significant relationship between social statusand problem-solving skil ls?
2. will there be significant improvement in the pr oble m- s o l v i ng skills as a result of the present problem -solvingtrailling program?
Prob lem -Solving 3. If there are improveme nts in problem-solving skillsasa result ofth e present training program,wh a t are the effe c ts, i fany, onag g r e s sive l:.e h avior?
4. I f there improve ments in pro b lem- s o l v i ng skills, what are the ef fects, i f any, on peer grou p acceptance?
ReviewoftheLite r a t ure
AssessmsntofSocialStatus
Behavioral observation strategies, exemplified by therate-af- interaction appro a c h , definesocial status in termsofth e frequencyof pe e r interactionas opposed to other more qualitati ve as p ect s of the interaction.
Childrenwh o s e total ra te of interaction is conside rably belowaverage forthe groupar e re f err ed toas socia l ly wit h d r awnor sociallyisolated.Thisrate -of-interaction approac h has been criticized because of a la c k of pred i c t i v e and concur r e nt validity (Asher, Ma rk e ll , « Hyme l , 1981 ; Gottman, 1977 ; Gottman, Gonso, «Schuler, 1976; Li, 1985).
Pro blem-Solving As anal te r na t ive , sociometric measures nave been usedextens ivelyas a meansof assessi nga chi ld'ssocial stat us, relat i v eto that of theother childrenin a peer gro up (for reviews, see Bullock, Ir o n s mlt h, &Poteat, 19 8 8 ; Hymel, 198 3 ). Sociometric meas ures pr o vi d e an evaluation of a child's peer relations from the perspectiveofthe peers themse l ves,ratherthanre l y ing on external,adultsourcesof informatioll(Hyme l,19 8 3 ). The two moatcommonl yusedsoc i ome t ric measuresare the peernominationandrating - scaleprocedures.
Thepeernomi na t i o nme a s ur e de ve lopedbyMoreno (1934, as describe dinBullock et a1.; 1988 andHymel;
19!Dl was the most fre quent l y us e d procedure to assess socia lstatus in earl yresearch. Thi sproce durere q u i r e s childre ntono mi na t eapredetermined nu mbe r of classmates accordingto specifiedc:citeriasuchas most pre fe r redor least preferredplaymateorworkma t e. so c i ome t r i c scores of peer acceptance or rejection are derived fr o m the number of positive and nega tive no mi na ti ons re cei v e d in eac h category. Inan attemp t to si mpl ifythe procedure for pres c hoo l children , McCa ndle s s and Mar s hall (1957) suggestedusing photog r aphs of peers. This proc e dur e requires chi l dren to select pic tures of peers from a
Pr obl e m- Solvin g display of all classmates, rather than asking for verbal or written nominations.
Nomination procedures are attractive because of their ease of administration. Although peer nomination acceptance and rejection scores can be considered separately, they can also be combined into social impact and social p r e fexence scores which allows for the classification of individuals intorejected. average, popular, neglected, and controversial social status groups (coie , Dodge, &. coppotelli, 1982; Newcomb &
Bukowski, 1983; Peery, 1979), each of which has been found to be associated with dist inct behavioral characteristics. Of particular in t e r es t is the distinction between two groups of unpopular children, those of rejected and neglected status. Often. however, nomination methods producea skewed distributionof votes and very little or no information on many group members.
Nominations are prone to biases resulting from peer friendship or other possible halo effects (Asher&.Hymel, 1981; Gresham, 1981 ; Schofield&Whitley,1983),
The rating-scale procedure cas become increasingly popularin the literature as a means of assessing social status. Developedby Roistacher (1974), the rating-scale
Prob1em. So1v ing method requires childrento rate classmates accordingto Borne specified Lnte eperecnaI criteria, such as how much they like to play with or work with each one. with older children, this usually takes the "t-xmof as-point, Likert-typescale (Singleton&:Asher, 1977). A simplified version where children have to rate pe e ra on a 3-point scale by assigning photographs of classmates to one of three boxes identified by smiling, neutral, or frowning faces has been described for use with preschool children (As he r , Singleton, Tinsley, &: Hymel, 1979). Ratings received from all classmatesare averaged for each child to obtain a r au Lnq- scale score.
An advantage of the rating-scalemethod is that it al lowsevery childto be rated by each of his or her pee rn thusproviding an index of each child'soverall acceptability in tbu peer group (Asher&:Hymel, ~981i Gresham, 1981; schofield..Whitl e y , 1983).As well, wi t h the use of the rating-scaleprocedure, children'sratings ar e unaffected by group size(Fo s t e r .. Ritchey ,1979),a problem inherent in the peer nomination procedure.
Rating -scalescores appear to be more sensitiveindices of treatment effectiveness than nomi natio ns (ed en &:
Asher, 1977). Afinal ad vantage of the rating-scale
Problem- Solving
procedure is t.nat childrenare not required toIndicate anyoneas particularlydisliked {As he r &Dodge,19861. Theprimarydisadvantageof the procedure ist.h a t rating- scale scores cannot be used to distingui s h bet.ween re jecte d andnegl e c t e ~;; children (Asher &Dodge, 19B6:
Ash er& Hymel, 1981; Dorval &Begin, 1985:Hymel,1983;
Olson& Li f gr e n, 1988),
Assessme ntof social status with pre s ch o ol chi l d r en.
A review of the researchevaluatingthe psychometric propertie s of sociometric procedures with preschool childre n sugge sts th e use of the rating-scale as the assessment tool of choice. The rating-scale has been found to prov ide a more relia bl e in d e x of preschool children'Ssociometricstatus thanthe peer nomination procedur e (Ashe r et al., 1979; Dorva l & Begin, 19 8 5 ;
Olson&Lifgren ,1988).Asher et a1. found a test-retest
correlationof .81forthera t i ng-s c a l e measureover a 4- weekint e rv al , In comparis on,theyfounda correlationof .5 6 for pos i t i ve nominat.ions and .42 for negative nomi nati o n s. Similar ly,Olsonand Lifgre n (1988 ) reported
Problem-Sol ving test-retest coefficients of .52 for positive nominations, .4.8 for negative nominations, and .81 for the rating' scale procedure over a 3 week interval. Hymel (1983) suggested that the greater instability of nomination scores amongpreschool children may be due to the fact that younger children'sfriendships fluctuate more than do those of older children and thatnomi na t i on scores are more sensitive to this fluctuationsince they are based on only a few responses by each child.
Concurrent validity of the rating-scale procedure wjth preschoolchildrenhas beendemonstrated in terms of correlationswithmeasures of observed behavior and peer interaction (Olson&<Lifgren, 1988; Rubin&Clark, 1983;
Rubin, Daniels -Beirness, &Ha yvr e n , 1982). Rubin et a1.
(198 2) found consistent positive correlations between peer rejectionas determined by sociometricratings and indicesof agonistic behavior.Behavioral observations of rough-and -tumbleplay and negative peerinterchangeswe r e positively correlatedwith negative sociometricratings.
They also re por t e d that preschool teachers' ratings of aggressive-hostile and hyperactive behaviors on the Preschool Behavior Que s t i o nna i r e (PBQ) correlated positively with negat iveratings. In a simi l arstudy,
Problem~Solving 10 Rubin and Clark (198 3) foundr-at.Lnq-jsc a l.e scores to be modestly correlated with teachers' ratings oE maladjustmenton the PBQ.Olson and Lifgren (19 88) found that rating-scale scores were modestly positively correlated wit h teachers' ratings of positive peer
mt
erecc Lon, and modestly negative ly correlated wit h teachers' ratingsof aggression.Dorvaland Begin(1985 ) reported convergentva l i d i t y of the rating-scaleprocedure with preschool children in terms of high correlations with variousaspects of. group structure. In te r ms of predictivevalidity, Olsonand Lifgren (198 8 ) found that rating-scale scores we r e predictive of children's ability to generate relevant solutions to hypothetical social problemsituationson the Preschool Interpersonal ProblemSolving (PIPS) test at a 1 year follow-up .Ir o n s mi t h and Poteat (1990)found that peer ratings ob t a in e d in preschool significantly predictedteacherra t i ngs of behavior on the PBQ1year late r whenthe children werein kindergarten.
Bt h i c a l conc e rns have been raised about the co n s equ e nce s of administering sociomet ri cprocedures to chi l dren (Bel l -DoI<ln & We ssle r, 1994). Sur prisingly, however, few pu blishedstudies haveadd r e s s e d this issue.
Prohlem.Solving 11 In a study with fifth grade children, Bell-Dolan,Foster and Sikora (1989) found that the administration of positive and negative nomination procedures had no effects on the children'ssocial interactions or on their reports of mood or loneliness.Hayvren and Hyme l (1984) evaluated the potential negative impact of using sociometric procedures with preschool children. They fo und that the administrationof neither the positive nomination, negative nomination, rating-scale sociometric measures had an immediate or a 10n9- term eff e c t on pc-e schooI children's peer interactions.
Behavioral observations of peer interactions obtained immediately followingtesting provided no indication that childrenmade negative ver ba liza tionsto or about their peers.Al t hou g h children made more positive and neutral initiations and responses to positive peer nominees and high-ratedpeers thanto negativepeer nominees and low- rated peers, they did not differ in the frequency of negative initi ationsor responses as a re s ult of testing either immediately after or several weeks followIng completion of testing.
Problem-Solving 12
sociometric measures have been used extensively in the literature to assess socialstatus, as an alternative to behavioral observation strategies. Although peer nomination and rating-scale procedures have both been used toid e n t i f y childrenof low and high status, they appear to measure differentdimensions of social status.
Rating-scales measure a child's overall acceptability in a peer group and tend to have less potential for bias, compared with thenomi n a t i on procedure. With regard to psychometric properties,the literature suggests that the rating-scale procedure, as compared wi t h the peer nomination procedure, is a more rel iable and valid measure to assess the peer relations among preschool age children.
Social Status and Aggression
In a re ce nt meta-analysis, Newcomb, Bukowski, and Pattee (19 9 3) foun d that childrenfrom differentsocial status groups have distinct behavioral and eccd aL.
Pr o blem-SolYing 13 cogn i tive pr ofiles that influe nce thequali ty of thei r social relat i on s. Comp ared wit h other social st at us gro ups , rejec t ed ch i ldr en we re found to be more aggressive, lesSt sociab le and cognitive ly skilled . Newcomb eta!. (19 93 ) conclude dth a t rej e cted children are at risk in their socia l development and thatthe aggre ssion, withdrawal, and social de fi c its of these chil d r en repre s e nt pote ntial ant e c e dents fo r psycho l og i c al disturbance .
There has been a weal t h of studie s exami n i ng th e beha v ior al correlatesof childre n ' s social st a tus(for a review, see Coie , Dodg e, & Kupe rsmidt , 1990). Using differing met hod olog i es, the behavioral corr elate s of pee r acc epta nce and re jection have beenex aetneu across a wide ag e range. Studies inv o lvingpre school child ren le.g., Olson, 1992; Olson & Brodf e ld, 1.991; Rubin &
Cl a r k,1983;Spen ce, 1987),kindergart en childre n(e.g., Rubi n & Dan i e ls - Belrne s s, 1983; Rub i n et a1., 198 2;
Vita r o , Tremblay, Gagnon , &Boiv i n, 1992 ; Wa sik,1987), elemen ta ry child r en (e.g. , Chen, Rubin , & Su n , 19 92 ; Dodg e,1983;Dodge,Coie,Pe t tit, liePri ce, 1990; Oygdon, Cong er,&Ke ane, 1.987; Er hardt&Hins haw, 1.9 94; Hymel et a!. , 199 0; Tay l or, 1989; Vita ro et al., 19 92) an d
prob~em-So~ving 14 adolescents (Ollendick et al., 1992) are availablein the literature. One finding thatconsistently emerges from these studies is that rejected children are perceived by teachers, peers, and independent observers to be more disruptive, aggressive, inattentive, uncooperative,and socially inappropriate than their more sociallyaccepted peers. In addition, aggression appears to be .J. more significant factor in the rejection of boys than of girls.
Consistent withresearch on olderchildren, studies examining the relationship between peer rejectionand aggression in preschoolchildren indicatethat aggressive and disruptivepreschoolers tend to be disliked by their peers (e.g., Hayes, 197 8 ; Ladd " Mars, 19 8 6 ; Mili ch , Landau, Kilby, " Whitte n, 1982; Olson, 1992; Olson &
Brodfeld,1991 ; Rub in&Clark, 1983 ; Rubin et al. , 1982 ; Spence, 1987). Both peer and teacher assessments of externalizing behaviors such aggression, disruptiveness , and impulsivity have been found to be significantlycorrelatedwi t h peer rejection in preschool children (Olso n , 1992; Olson& Brodfeld, 19 91; Spence, 19 87) . Indicesofpro s oc i a l behavior have been found to be significantly related to likability (Denham s,Holt,
Prob lern·Solving 15 1993; Denham, McKinley, couchoud, & Holt, 1990; Ladd,
Price, &Ha r t , 1988). Friendlier, more cooperative and
less aggressive preschoolchildren werefou nd to be liked
Although there is strong support to ind i c a t e a relationship between peer rejection an d aggressive behavior across a wi d e age range, recent research with bothol d e r and youngerchil d r e n suggest.s that there may be considerablehe te r o ge n e it.y among the group of rejected chfidren (nt e r me o , Smoot, &Aumiller, 199]; Cillessen, van IJzendoorn, van Lieshout, &Hartup, 1992; French, 1988; Hodgens&McCloy, 1989). French(1988) examinedthe possibilitythat subtypes exist wi t h i n a population of peer-rejec te d 8-10 year old boy .. in two studies, one using the peer nomination procedure as the method of selection andoneus i ng the rating-scale procedure. In both studies, itwa s foun dtha t approxi matel y sot of the rejected group exhibitedan agg ressi ve behaviorprofile wh i l e the re ma inde r of children in this group wer e rejectedforotherreasons. Similarly, ina study wi th a younger populationof 5-7year old boys, Cillesse net aL, (1992) found that about 50% of the boys id e n t if ied as rejectedwer e of the aggressive, impulsive ,disruptive,
Pro blem-Solving 16 and noncooperative subtype. Cillessen et all also examined the relation between rejection subtype and sociometric stability and found tha t 57% of the boys identified as rejected-aggressive continued to be rejectedby their peers 1 year later as compared wi t h only34%ofthe rejected-non aggressivebOYS, suggesting that peer rejection involvingaggressionis more stable than rejectionthat does not involveaggression. These studies provide support for heterogeneityamong children identifiedas rejectedand suggest thatag g r e s Rio n is a mea ni ngfuldimension on whic h to distinguish them.
Stabi1ity of social statusandaggress ion!
Socialstatus, pa rt icularlyrejected socialstatus, appears to be a relatively stable phe no me no n . Rejected status hasbeenfoundto be stable across time (Bukowski
&Newcomb, 1984;Coie&Dodge, 1983; Howes, 1990;Hymel
et aL., 19 90 ; Rubin & Daniela-Beirness, 1983; Taylor, 1989; Vita ro, Gagnon, & Tremblay 1990; Wasik, 19t17), ac rosssocial situations (Luftig, 19 87 ), and acrossne w social groups of unfamili a r peers (Coie & Kupersmidt, 19 83 ; Dodge et al., 19 90 ) .
Problem-solving 17 'rhe stability of rejected status has been demonstratedwithpreschool age children (Denham et al., 1990; Olson, 1992; OlsonIi<Brodfeld, 19 91 ) . Olson and Brodfe ld (1 991) found that peer rejection in preschool boyswa s moderatelystableover a 6 month period.Olson (19 92) found that preschool bo y s id e nt if i e d as rejected and aggressive at the beginn ing of the year tended to remain so at the end of the year. Den ha m at al. found preschoo le rs'sociometricratings,particula rlynegative ones, to be stable over both a 1 and a 9 month time period.
Aggressive behavior has also been found to be a relativelyst a ble phenomenonacross a wide age range.
Lad dan d Mars (1 966) found preschoolers' perceptionsof peer aggression and co op e r a t i v e play to be the most stableof the be ha v i o r s assess ed. Bot hpeerandtea cher measuresof ag g res siv e- d i s r u p ti v e behavio rwere foundto be highlystableoverthe cours eof the preschool year (Olson, 1992; Olson Ii< Brodfeld, 19 91). cummings, Iannotti, and Zahn- Waxler (1989) found aggressive behavior to be stable from toddlerto late preschool age and to be more st a ble for boysthangirls. Rusher,Ware, an d Cross (19 94) reported stability of disrup tive
Pro blem- So lving J.8 behavioral characteristics for2year old children over a period of 4we e k s .
The link between peer rejection and aggressive behavior has beer- well established and both peer rejectionand aggressive behaviorhave beenfound to be relativelystablein childrenas young as the preschool age.Giventhat researchers havereceneIy recognizedthe he t e r og e ne o u s natureof rejected children and the fact that aggression may only be characteristic of some chdLdz-en, an increased focus on aggressive-disruptive behaviors as targetsfor change is likely.As~1Uggested by Bullock etaL, (l988), there is ane e d to not only identify an d target for interventionthose children who are rejec tedby their peers, but alsoto determine the specificbehaviors that are associ -rt ed with difficulties in their peerrelat i on s .
InterpersonalProblem-S olving
Re s ear c h has suggest edthat int erpersonalcognitive problem-sol ving (repS) skills are one subsetof socially
Problem- Solving 19 competent behavior which is positively correlated with, and predictiveof, emotional and behavioral adj uat.ment;. Interpersonal problem-solvingskillshave been found to be associated with social competence in a wide variety of populationsand across a wide age-span (f o r a revie w,see Tisdelle & St. Lawrence, 1986). In a meta-analysis, Denham and Almeida (1987) found that Ieps measures significantly different iated between adjusted and nonadjusted children.
spivack, Shu re,and their colleagues have described several Ieps skillsas important 1J"'. therel a t i o ns hi p to behavioral adjustment, independent of intelligence (Spivack, Platt,&Shure, 1976;Spivack&Shure, 1974).
Theseskillsincludeproblem sensitivity,or the ability to percei ve problem sicuations and to focus on the aspects of interpersonal confrontation that create problems; alternative think ing, or the ability to conc eptualize alternative solutions to typical age- relevant in t e r personal problems;consequentia lthinking , the abi lity to conceptuali ze the po t e nti al consequencesof an inte r persona lact;causal th i n k i n g , or the abilityto relateoneeve nt to anotheroverti mewi t h regardto the "Why" that mi g hthave pre cip i t atedtheactI
Probl8ll'l -So lving 20 and means-end thinking, or the ability to articulate to a problem solution. The significance of the relationship of each skill with behavioral adjustment appearsto differ as a function of age and developmental level of a child.
Inte rpersona l problem- s olv i n gskills of pre school children.
Initialwo r k examining the relationship betweenIeps skills and behavioral adjustment in preschool children was carried out by Shure, Spivack, and colleagues in several earlystudies. Shure and Spivack(1970, as cited in Spivack & Shure, 1974) identified a relationship between alternative thinking and both behavioral adjustment as rated byteachers and socioeconomic level.
They found thatlower-classchildren as a group and less we ll- a d j u s te d children within both the lower and middle class groups offeredfewer solutions andana r r owe r range of solutions to problems on the PIPS. Lower class child ren al s o produced an increas ed proportion of forceful solut ions.
Problem-Solving 21 Shure, Spivack, and Jaeger (1 97 1 ) examined the r-e Lat Lorra h i.p between behavioral adjustment and alternative, consequential, and causal thinking in low socioeconomic preschoolers. Children we i-e classified as aberrant or adjusted based on seven items describing inability to delay, emotionality, and aggression factors.
The aberrant group consisted of children who acted out and children who were considered inhibited or withdrawn.
Results indicatedthat alternativethinking was the only measure related to behavioral adjustment. Children classified by teachers as less well adjusted offered significantly fewer relevant solutions to peer and authority problems on the PIPS and gave ana r r owe r range of types of solutions than did better adjustedchildren.
Results also indicated that those with lowe r problem- solvingscores gave ahig h e r ratioof forcefulsolutions (e.g ., grab it) to peer problems. These results were foundindependent of receptive vocabularyas measured by the Peabody Picture VocabularyTest (PPVT).There were no overal lgender differences and no relationship was found between behavioral adjustment andeitherconsequential or causal thinking.
Problem- Solving 22 consistent wit h theea r l i e r studies, Shure, Newman , and Silve r (1973 , as cited in sp i v a ck & Shure, 1974) found a re lat ion ship betwee n alternative thinking and be h a v i oral adjustment as as se s s e d by the Hahnemann PreschoolBehavior Rating (HPSE) Scale . However, they al s o fou n d are l a t i on s h i p betweenconseque ntia lthinking andbehavioral adjustment. Adjus ted childr e n provideda greater number of consequ e n ce stopr o b l e ms thanchil dre n rated as either impulsive or inh ib i te d. Results we r e found indepen d ent of language ab i.l i t y , general inte llig e n c e, andWil li ngn e s s to talk. There we r e no gender diffe r e n ces.
The rela t i on s hi p demon s tratedby Shure, spivack and colleagues betwe en ICPS skills and the behavioral ad justment ofpre s cho ol children wa sno t substa n t".iatedby Rickel andBu r gio (1982).Inthefi r s t phase of a project carried out atWa y ne State Unive rs ity, Ricke l and Burgio attempted to replicate thewo r k of Shure et a!. (1971) with low income preschoo l children. Children we re classifiedas imp ulsi v e ,in hibited, oradjusted using the HPSB rat i ng sca l e. In orde r torepl icate Shureet al .·s proced ure ,in hi b itedand impu ls ivech i ld r enwe recomb ined to fo rm one aberrant group. No significantdiffere nr.:es
Prob lem- solving 23 were found between aberrant and adjustedgroups on either the PIPS or What Happens Next Game (WHNG) measures of problem-so lving ability. However, children rated by teachersas adjustedhad signif icantlyhigher achievement scores on the CaldwellPreschoolIn v e n t o r y than children rated as aberran t. Observation by ind e pen d e n t raters using the Sharp Behavior Identification Checklis t indicatedthat aberrant chil d r e nen g a g e din more verbal and physical aggre.s.sionth a n did adjusted children. In additi on, aberrantmales exhib ited sign i fi c a n t l y higher le v e l s of aggressionth a n aberrant females.Gouze(1 98 7 ) also failed to find a relationship betweenthe number of alternativesoluti onsgenerated by preschoolboysonthe PIPS and adapt ivebehavior in the classro om.Differences in these stud iesmay have been due to the select ionof chi ldr e nwit h more extreme behavioraladjustment problems inthe Shure et al. study.
Shure, Spivack and their colleagues have demonstrated a relationship between problem-solving skills and adj us tment inpreschool children. Theyfound
that alternative thinkin g,
Prob l em- So lving 24 the ability to conceptuali ze alternative solutions to interpersonal problems, is the problem-solving skill most strongly related to behavioral adjustmentin preschool ch ild r e n . Children ratedbyteachersas poor l yadjustedtended to generate fe we r, aggressive so l u t i o n s to interpersonal problems than theirmore adjusted peers.
The relat i o n s h i p between behavioral adjustment and consequential thinkingwas foundto be less signifi.cant. In contrast, Rickel and Burgio (1 9 8 2 ) failed to find support fora relationshipbetweenproblem-solvingskills and adjustment in preschool children.The re s u l t s of a meta-analysisbyDenhaman d Almeida(1987)however,found strong support for therelationship be twee n Ieps sk ills and behav i oral adjustment in children.
Problem-solvi ng int e rventi o nwithprescho ol
Given the results of studies indica t ing a signific an t relat i on s hipbe tw e e npr o b l em-s o lvi ng skills andbehavio r a l adj u s t me n t , Shure and Spivackpredicted that it should be pona i.bLe to enhance the behavioral
Problem~Solving 25
adjustment of young children by imp r o v i ng those problem- solvingskills relatedto adjustment (Sp i v a c k&Shure, 1989). In an early pilot study, Shure , Spivack, and Gordon (1972) examined the effect of problem-solving training on behavioral adjustmentwith lowsocioeconomic preschoolers. Training consisted of fifty, 20 minute sessions. In i t i a l sessionsinvolvedtraining of specific linguisticconcepts associated with problem-solving while later sessions involved training in problem-solving skills. In addi tion to the training group, attention control and no treatment groups were also included.
Prior to training there were no significant differences between the three groups on either the PIPS or the PPVT. All children who received training gave a significantlyhighernumber of relevant eofuttons to both peer and authority problems, a greater numberof solution categories, an d fewer irrelevantre s p o ns e s than did the two control groups. Within the training group, those children who had thl'! lowest PIPSscores prior to training re c e i ve d thegr e a t e s t benefit from the program.Although not st'atlstically significant, only the delay of gratificatio n factor on the Devereux Child Behavior Rati ng Scale showed a posit ive change as a result of
Problem-S olving 26 training. There was no change on the emotionality dominance-aggression dimensions.At a 7 week follow-up, trained children continued to show improvement on delay of gratificationand were also rated as less aggressive.
Trained children who improved most in problem-solving also improved more behaviorally, particularly those who were most poorly adjusted initially.
Shure and Spivack (Shure, 1993; Shure & spivack, 1979, 1980,19B:;!) examined the impact of problem-solving training on the behavioral adjustment of 219 low income preschool and kindergarten children over a 2 year period.
Children received training during the preschool year,the kindergarten year, both years, or neither year. Format of the training program was determined by a script developed by spivack and Shure (1974) to be used by preschool teachers for instructing children in problem-sol ving skills. The program included dialogues, games, and activities for the teacher to use with the children in a series of forty-six, 20 minute sessions. Early sessions focused on developing the language skills which Spivack and Shure (1974) believed to be prerequisites for effective problem-solving and la t e r sessions were designedto teach alternative, consequential,and causal
Problem- Solv i ng 27 thinkingskills through a series of real-lifeproblems.
Training included dialoging which involved teachers following thr oug h with problem-solving training when actualproblems arose inthe classroom.
Prior to intervention, the r e we r e no significant differences between training and control groups on measures of intelligence, problem-solv ing skills, or behavioral adjustment ai'- assessed by the Hahnemann Preschool Behavior Rating scale. P.esults indicatedtha t children tr ai ne d in either year showed significant improvement in bo t h the ability to conceptualize alternative solutions to problems and the ability to conceptualizeconsequences of problems.Trained children also showed a decrease in the number of coercive or forceful solutionsgiven and this was most significant for those children classified as LmpuLe Lve . Children trained in the preschool ye a r alsoshowed an improved tendencyto conceptualizecause-and -effectwhen presented withan interpersonal event. Children wh o receivedno training showed some impr o vement over time on both alternativeand consequentia l thinking.
Inboth years,those childrenwho showed improvement in behaviora ladjust mentwere also those who improved ir
prob l em-Solving 28 alterna tive and conse q uent i al thi nk i ng sk i ll s. This rela tionship was st rong e s t foralte r native thi nk i ng and therelationshipwithconsequent i althi nk ingwas stronge r for the kindergartenye a r.At 6 monthand 1 yearfollo w- up s , trainedchildren remainedahead of controls on both problem-solvingandbehavioralmeasures.Controlchildren showedsome imp r ove me n t on ;;>IPS sco r es at the 6 mo nt h follow- up butdi d not catch up totr ai ned child r e n.
Attempts to re p lica t e Shureand Spivack' s find ings with preschool children have met wit h conflicti ng re sul t s. In the second phase of the project at Way ne StateUniversity, Sharp (1981)assessed the impactofth e spivackand shure(1974) training program wi t h low in c o me preschoolers.However, unlikeShure and Spivack' swork, theprogramtrai nerswere unfamiliar withthe chi ld ren, classroomteachers were unawareof group assignment and of the contextofthe training program, trainingdidnot invo lve dialoging,andan attentioncontr o l grouprather than a no treatment contro l group was included . In ad d i t i o n to the re g ula r training program, Sharp also includeda modifie d training group wh i c h followed the spiv"lckand Shure(1974) script withtheexcl usio nof the first 12 sessions an pre r e q u i s i t e la ngu a g e skills.
Prob l em-Solving 29 Analyses at posttest revealed that the aberrant children who received training increased significantly more in PIPS scores compared with both childrenin the control group and the adjustedchildren in eachof the three conditions. The difference in gains between the complete and modified training groups not significant. No flignificant training effects we r e found for consequential thinking. Improvements in alternative thinking, however, did not mediate improvement in b'ahavicz-a I adjustment. In fact, both adjusted and impulsive groups showed an increase in aggression and dominance from pretest to posttest.
In phase three of the project at Wayne State University, Rickel, Eshelman, and Loigman (1983) conducted a 6 month follow-up of the children who participated inthe Sharp (1981) study. Although aberrant children who were trainedgained significantly in their abilityto generate alter nativesolutions from pretestto posttest, there was no significantchange from post test to follow-up. In fact, all groups chewed significant improvement from pretest to follow-up on the PIPS and WHNG measures and in achievement. No relationship was found betwe e n problem-solving measures and behavioral
Probl~-Solving 30 adjustmentat follow-up.However,there was a decr e asein aggr e s sion fr om posttesttofo l l ow- up inboththe control gro u psand the tr eatedadjustedgroup.
Feis and Simon s (l98S) replicated the trai n i n g pr ogr a mof Spivack and shure (1974) with low incom e pre s c hoo l child r enov e r a 3ye a r pe ri od .Ch ildrenwere randomly as sig ne dtoeither trai ningor con t rol groups. Th e r e were no si gnif icantdiffere n c es be t weenthegro ups at pretest inthe numberof solution s generat e d orth e numb e r of solution ca t e g orieson thePIPS . In all 3years of the st udy, trained ch ild ren showed a significant improveme nt in the numbe r of sol ut ions an d number of solutioncategorieson thePIPSfrompret estto posttest as compa r ed withthe co n tro l group. Inthe thirdyear, trained childr en also showed significantl y tewe r behav i oral pr oble ms the Preschool Beha vio r Ques tionnaire at pos ttest co mpa r ed with the control grou p . Specifically , trai ni n g wa s found to have a sig n ifi c a nt moder ating ef Eect; on children's anxi ous - fearf ul andhype r active- di stra ct i bl e behavio rsbut not on agg r e s siv ebehavior.
Ridley andVaughn (Ridl e y &;Va u g hn,1982; Vau ghns, Ri d l ey , 1983; Vaughn, Ridley, Ii<Bulloc k , 19 8 4) have
Prob l em- So lvi ng 31 evaluated the effectiveness of problem-solvingtraining wi th middle class preschool children. Ridley and Vaughn (1 9 82 ) evaluated a program that included training in problem-solvingskills and the utilization of an empathic communication mode. Training consisted of forty, 15-20 minute sessions and classroom teachers were unaware of the content of the training program. The components of the program were also used to solve problems that occurredduring training (e.g . a child wantingto sit on another child's mat) .
Relative to the control group, Ridley and Vaughn (1982) found that the training group showed a significant increase in the number of solutions generated topeer problems on the PIPS atpos t test and at a 3 month follow- up. Wh i l e there were no significant differences between the groups in the number of solutions given to mother- relatedproblems an the PIPS at pasttesting, th e r e were signif icant differences at foll ow- u p , suggesting generalization of the program effects from pr a b l em - solving with peers to thatwi t h an adult.Nosignificant differences were found bet wee n the groups on the relevancy ratio at either po s t test or fol:'-",,-up. The Behavioral Preschool Interpersona l Problem Solving
Problem -Solving 32 (BPIPS) test, a modified versi on of th e PIPS, administered to assess the children 'sabilit yto generat e alternatives tointerpersonal problemswith apeerin simulatedreal-li fesituations.Asi gni fi c ant differenc e wa s found becween the two groupson both the number of solutionsgiventopeer problemsand the relevancy rat i o at bothpostte s ti n g and follow-up.
va u g hn and Ridley (1983) evaluated th e effects of training on the be havior of preschoolch i ld re n inthe classroom setting. The training program involved 50 sessionsandwasba s e don the sameco n c e pts as the Ridl ey and Vaughn (1982 ) program. Results indicated that the training group, co mp a r e d with an atte nt i on co nt ro l gr oup showed a significant in c r e a s e inth e frequencyof bot h positiveve r b alinte r a c t i o n with peers from pre tes t to posttest . The increase in positiveverbal inte r a c tio ns reflectedan increaseinsuchposit iveverbalizationsas pr aise, positive regardof another, and engagingot he r s in activity whiletheincrease in nonverbal interact ions reflectedan inc r e a s e in behav iors such as cooperat i ve play, sharing, and helping . Va ughn and Ridley (1983) suggestedth at skillstaught in trainingqenerelteed to
Problem ~ Solv ing 33 th e child 'Bin t e ra ct i o ns withotherchi l d r en. Therewere no significant. eerece eon interactions with adults.
Ina final st udy,vaughn,Ridley,and Bul lock (1 9 84) eva luated the effects of traini ng ontheproblem-solving ski ll s ofaggres s ivepreschoolchildren .Of165 children screened using the Hahne-mannPreschool BehaviorRating Scale,24 chil d r en(1 9ma l e s , 5 females) were iden t i f ied as aggress iveand ra n d oml y assignedto eithera training gro u p that ut i l i zedth e Ridley an d Vaughn(19 8 2) program oran at c ent Ion contr o l gr o u p. Resul ts ind ica t e dthatthe tra i ni ng group shew ed a sign i fic an t inc r e ase in the nu mb e r of relevant sol u t i ons to peer problems and relevancyratioas measuredby the BPIPS at both posttest an d a 3 month follow-up . unfortunately, Va u g hn et al.
(1984 )didnot evaluate the effects of tr aining on the ag g r e s s i v e behavior of th e children in th e preschool se t ti ng .
De n h a m and Almeida (1 9 87) examined the effectsof tr ai ni ng on ICPSski l ls in a meta-analysisandfound that trained ch ild r e n exhibi tedsignificantlyhigher scoreson measures asses singICPS skills at post testcompared with co ntrol children. A link was foundbetweengains inICPS ski lls and improvemen ts in adjustment, and this
Probl em- Solv ing 34 relationship appeared to be stronger for younger children. They found tha t programs that included dialoguing cbt.a Lned higher posttest scores on ICPS measures. Denham and Almeida (19B7) recommended that ICPS training studies need to include groups of children that are at risk or deviant in their behavior such as aggressive children and that sociometricmeasures be used to evaluate the effects of training.
In recent meta-analyses, both Schneider (1992 ) and Beelmann, pfingsten, and Losel (1994) reportedmoderate effectiveness of training. Although not statistically significant,Schneider (1992) found stronger treatment effects for younger children. Seelmann et a L, (1994) reported a trend in the literature towards complex, multimodal training programs but indicated that monomodal programs appear to be more effective with preschool-age children while multimodal programs appear more effective with older children. They recommended the need for future social competence programs to fit the specific social deficits of the children and bring about. not only narrowly defined, short- termmodification but also more comprehensive, long- t e rm follow-upeffects. Schneider (19 9 2 ) also recommended follow-up of a longer duration.
Prob lem -Solving 35
The effectiveness of programs designed to te ach prob l em-solving skills to preschool children and to enhance their behavioral adjus tment have met with con f li c t ing re su lt s . Shu re and Spivack have fou nd that pres c ho o lchildrencanbe taught togen e rat ealternative solu t i ons andconseque nc e s to interpers o nalproblemsand thatgains , parti cu l arlyin alterna tive thinking, were as s o ciat e d wit h improvemen t in behavioral ad j us tm en t. Sharp (1981) found thatpr esc ho o l children canbetaug ht togenerate alternative solutio ns to problemsituations but faile d to find support for a relationshi p betwe en gains in proble m-solving skills and improvement in behavioral adjustment. Altho ug h Ridley and Vaughn have demonstratedimp r o v eme n t s inth e problem -solvingskills ofpresch o ol children asa re sultof tr a i n i ng , theyhav e prov ide dli t t leinformationon the effectsof trai n i ngon thebehav i or ofthe childre n.
Followingme t a··a nalyses, bothschnei der (1 992) and Seelmannet; al. {199 4 l reportedmoderate ef.fectivenessof trai n i ngprogra ms. Denh a m and Alme i da (1 9 8 7 ) foundthat child r en ....ho rece ived training showedsignificantgains
Problem-So lving 36 in ICPS skills. They also provided support for the relationship between gains in reps skills and improvements in behavLoz-, and indicated that the relationship appeared stronger for younger children.
Although not statistically significant, Schneider (1992) also found stronger treatment effel"ts for younger children.
ThePresentStudy
The present study was designedto address the needs of preschool children identified as having low acceptance among their peers and perceived by teachers as agg ressive. This stems from research suggesting that children with poor peer relations are more likely to develop later adjustment problems and therefore, should be considered a group of childrenwhoare at risk and consequentlyin need of intervention. In addition to the rating-scale which appears to be the most reliable and valid measure of assessing peer status in preschool children, a measure of aggressive be hav i o r wa s also includedin thepresent study.This enableda distinction between those childrenwith low peerac c ep t a n ce who were
Prob1em- So1 v ing 37 aggressive and those childrenwith lowpeer acceptance who were not aggressive.
Giventhat earlyide n t ifi c a t i on andint e rvent i o n is essentialfor the prevent ionof potentialproblems, the re l a t i v e l y high stability of poor peer relations, the fact tha t it becomes increasinglydifficult to modify the social reputation of children as they getol d e r and that behavioral repertoires may be more amenable to interventionat an earlier point in development (Ro gos c h
&Newcomb,1989),preschoolchildren were selected as the
target population. Providing intervention forpreschool children identified to be at risk in their peer relations may reduce th e potential for adjustment difficulties later in life.
The present intervention focused on t.r a i ni n g the selected population in interpersonal problem-solving skills. Since the workby Shureand Spivack demonstrating a relationsh ip between impr o v emen t in behavioral ad j us t me nt and gains in problem-solvi ng skillo, interventions have met with conflicting re s ul t s . The present st ud y was a further at t empt to define the relationship between ICPS skills and behavioral adjustment.Training sessions we r e based on the script
Pr obl em- Solving 38 developedbySpivackand Shure (1974). acveve r , in order to reduce the significantlength ofthe Spivack and Shure (19 74) program, children received only 13 tr a i ni ng sessions. In contrast to the Spivack and Shure (1974) tr a i ni n g programwhichinv o l ved the presentationof both peer and adul t problems during training, only peer problemswere utilized inthis study, as peer relations were the prim?ry focus of attention.As well, inord e r to fo s t er generalization, relevant, as opposed to hypothetical , social si tua tions were targeted during training. Th i s was accomplished by requesting the directors from each ofthe participating preschools to describe those situations which they found most problematic amongthe childrenin their preschool. From thisa set of co mmo n problems was selected.
In contrast to work by Shureand Spivack who tended to define adjustment broadly, oftencombininginhibited and imp uls i ve chi l d r e n into an aberrant group, the present study utilized specific criteriafor inclusion into the adjustment groups.As well th e present study included an attentioncontrolgroup to determinewhether therewa s anyimpr ov e ment in problem-so lving sk i ll s as a result of normalmaturat ionalprocesses. The absenceof
Problem-Solving 39 an attention control group in the work by Shure and Spivack, particularly given the length of their training program, makes i t difficult to rule out alternative explanationsof change. Finally, preschool teachers who completed the behavioral ratings in the present study were blind to the treatment status of the children.
In order to conduct the interventionin ways that did not stigmatize children with low peer acceptance, a group of adjusted children was also selected to participate in the study. The inclusion of a normative standard provided an obtainable target range for improvement. The present program also served as a preventive measure for children with higher peer acceptance by further building and reinforcing healthy behavior and thus reducing the risk of future difficulties (Rickel &Burgio, 1982).
'rne following hypotheses were made in the present study:
1. I twas hypothesized that prior to training the rejected group would have significantlylower scores on both the PIPS and WHNGme a s ur e s of problem-solving abilitythan would the average group.
Problem- Solvi n g 40 2. It wa s hypot hesized that the treatment group, relative to the attent ion control group, wo uld show significant improvements in pr oblem- s ol v i n g skills as refle c t e d by anin c r ease in the number of solutionsgiven to peer problemson the PIPSand numberof coneequencea giventope e r prob lemson theWHNG.Itwa s predictedthat improvement would be significantly gre ater for the re j e c ted tr e a t me n t gr oup tha n the average treatment group. No increase in solutionsto mother problems was expected. Nosignificantimprovements we r e expectedfor the attentioncon t rolgroup.
3. It wa s hyp ot he si zed that the tr e atment gr oup , relative to the attention control gro up would show a significa nt improvement in ove rall behavior and a signi fic antdecre aseinaggr essive be havi o r as reflected by the i rra t ings on the Pr e s ch oolBehav i o r Questionnaire.
I t was ex pe c t e d that the improve men t would be significantlygreater for the rejectedtreatment gr oup as oppos e d to the averag e tr eatment group . No si g nificant changes were expectedfor the attention control group.
4. Finally, it washypot hesi zed that thetreatment group, as compa r e d wi t h the attentio n contro l gr oup, wouldshow a significant impro vement inpee r acceptance
Prob~ em- So~vin9' 41 as reflected by their scores on the rating scale. It was expected that the improvement would be significantly greater for the rej ectedtreatment group compared with the average treatmentgroup.No significant changes in peer acceptancewere expe c t ed for the attention control group.
Method
Pa rti cipantE!.
Participants were selected fromfi ve preschoolsin an urban area in eastern Newfoundland (population 130,000). Allof the participants were Caucasian and they represented a wide ra nge of socioecon omic backgrounds.
Letters describing the study and asking for parental consent for thechild to participate were distributed to all parents of children between the ages of 4 and 5 wi thin each of the presc hools (see Ap pendi x A). Of 71 consent forms distributed,permission to participatewas receivedfor all bu t two children.
Following aaae srament; , children were assigned to a group designated as rej ected based on the following criteria :
Problem- Solvi ng 42 1. The childrenwere below the average peer-rating score obtained for their respective echcol.
2. The children's receptive vocabularywa s wi t hi n average li mi t s (standard score of 65 or above on the PPVT·R).
]. The children obtained a total scale score above the ninetieth percentile (a score of17 or above) on r he Preschool BehaviorQuestionnaire.
4.The children obtained a score above the ninetieth percentile (a score of seven or above) on the ace c Ire- Aggressive subscale of the Preschool Behavior Questionnaire.
Children were assignedto a group designated as average based on the follow ingcriteria:
1. The children we r e above the averagepeer-rating score obtainedfor their respectivepreschool
2. The children'sreceptive vocabulary was within average li mi t s (standard score of 85 or above on the PPVT-R).
a. The children obtained a total scalescore below the sixty-fifthpercentile (a score of nine or below)on the Pre s c hool Behaviorcueactcnnatxe.
Problem-Solving 43 4. The children obtained scores below the sixty- fifth percentile (a score of three or below) on each of the three aubaca Lea of the Preschool Behavior Questionnaire.
Based on the above criteria, 34 children were selected to participatein this study. Sixteen or 23%met the inclusion criteria of the rejectedgroup (5 females and 11 males )and 18or 25itmetthein c l u sion criteriaof the average group (1 0 females and 8 males) . All participants were betweenthe ages ...f 49 and 59 months wi th a mean aqe of 53.44 months. The me a n age of the rejected group was 54.06months while the mean age of the average group was 52.89 months. The mean r-eceptiive vocabulary score of the rejected group was 10 2. 25 and 107.11 for the average group.
participants meeting the inclusion criteria for the rejected and average groups were then randomly assigned to either tr e atme n t or attentioncontrol conditions so that there were eight rejected and nine average children in each condition.For thepu r p o s e s of administeri ng the program,fo u r treatmentand four attentioncontrolgroups werethenfo r med . Thesegroups consistedof betweenthree to five children depending onthe number of rejected and
Prob lem~Solv ing 44 average children that were identified within each preschool. There were three gr oups of five, fourgrou p s offo u r , andon egroup of three children.
Pa r t ici p a n t s were evaluated on the fo llow ing
1. The Peabody P~cture Voca bu la ry Tes t·Re v i sed (PPVT- Rl was administered to ensure that for both no r ma t i v e and comparisonpurposes all chil d r e n wereof average abi l i tyon a measureof receptive vo cabulary.
2. Sociometric acceptance assessed by administering the rating-scale procedure described by Asher et al. (197 9 ) . Each child wa s in d i vi d u all y presented wit h colour photo graphs of each of his/her classmates. The childre nwereasked to name each child and then assigneach picture to one of threeboxes on which were drawn either a happyfac e (c hildre n yo u lik e) , a neut ra l face (chi ldrenyo u like somet ime s ) , ora sa d face (c h ildr en you don' t like).I fa childwasunable to name a child ina picture they were not asked to rate that particularchild. positive ratingswe r e accorded a score of three, neutra l ratings a score of two , and
Problem-Solving 45 negative ratingsdsc o r e of one . Each child' s to t a l score was divided by the numberof children in each preschool whora t edthatch ildinor d e r to obta i n an average peer ra t i ng since th e number of children participatingin each preschool differed. This scale has be e n shown to demonstrate bothacceptable reliability (Asher et al., 1979; Boivin &.Begin, 198 6; Do rva l &Begin, 1985) and valid i t y (Rub in , Daniels-Beirness &Hayvren,19 8 2 ) with preschoolchildren.
3. The PreschoolBehavior Ques tionnaire (PBQ) wa s developed by Behar and Stringfield (1 974 a) and represents a modificationof the Children'sBehavior Questionn aire (Rutter, 19 67),developedfor usewith elementary school - aged children.The PBO is a 30item rating·scale using a 3· po int scalingsystem, for use by preschoolteachers to ratechildren,ages 3-6, inthe context of a peer group . It yie lds a to t al score re fl e ct i ng overall le ve l of adj ustment and three subscale scores lab e l l e d (a) Hostile-Aggressive, (b) Anxious-Fearful, and (e) Hyperactive-Distractible obtained by adding rawscores fo r selected items. For a description of test administration and scoring, see Beharand Stringfield
(1 9 74 b ) .
Problem- Solving 46 Behar and Stringfield (1974a) and Behar (1977) reported interrater and test· retest reliabilities of the PBQ based on ratings providedbypreschool teachers and teacher aides. A mean interrater reliability coefficient of .8 4 was obtained for the total scale, with means of .81, .71 , and .67 reported for the Hostile·Aggressive, Anxious-Fearful. and Hyperactive-Distractible subscales, respectively. A mean ceec-eeceec reliability coefficient of .87 was reported for the total score with means of
.93, .60, and .94 for the three subscales after a 3·4
month interval. Behar andStringfield (1974a)and Behar (1977) have also reported data on the criterion-re lated validity of the measure. The total score and each of the three subscales hav e been shown to discriminate significantly between groups of normal and deviant preschool children. The mean total score for the normal population was 8.007 comparedwith a mean of 21.324 for the deviant population. Hege, Meginbir. Khan, and Weatherall (1985) found evidence for the construct validityof the PBQ as well as strong support for the convergent and discriminant validity of the Hostile- Aggressiveand Anxious-Fearfulsubscales.Rubin and Clark
Prob l e m-S olv i n g 47 (1983) also provided support for the construct validity of the PBQ.
4. Cognitive Measures of Problem-solving.
(a) The Preschool Interpersonal Problem-Solving Test (PIPS) measures a preschool child's ability to think of alternative solutions to two life-related types of problems: fa) ways for one child to obtain a toy that another child is playing with, and (b) ways for a child to avert his/her mother's anger as a result of damage to property.Shure and Spivack (1974a) established a test- retest reliability of .73. Validity is claimed for the PIPS on the basis that the measure consistently discriminates between groups of children who differ in level of behavioral adjustment ex.hibited in the classroom.The script used for the PIPS in this study is presented in Appendix B.
In the peer-type problem, the subject was shown three pictures, two of an age-relevant child (presented on 12.5 cm x 20.5 cm cards) and one of a toy (presented on 7.5 cm x 12.5 cm card). Characters presented were of the same sex as the child being tested.After a series of memory cues to identify the characters, the child was asked what one child could do to get to play with the toy
Pr oblem-Solv i ng 48 that the other child has. Memory cues were given to ensure that the child understood the story being presented.The examiner judged the point at which memory cues wereno longer needed.
In an effort to elicic as many different solutions as possible from each child, the experimenter repeated the same story plot, but substituted pictures of new characters and a new toy. The childwas presented with a minimum of seven similar peer-toy situations, but if seven different, relevant solutions were given, the experimenter continued with additional situations until the child no longer offered new ideas. In order to encourage a different solution, any response not offering a new relevant solution was probed. A maximum of three probes were made for each story.
In the adu'ltet.ype problem,the subject was presented with three pictures, one of a mother (presented on 12.5 cm x 20.5cm card), one of a child(presented on 12.5 cm x 20.5 cm card), and one of an object that was broken (presented on 7.5cm x 12.5 em card), andwas askedwha t the childin the story coulddo to avoid his/her mother' a anger. Characte rspresentedwer e of the same sex as the childbeing tested. New characters and a new object were
Prob lem- Solving 49 presented until a minimum of five stories were completed.
If five different, relevant responses were given, the experimenter continued with additional objects until the child no longer offered newideas. Three probes were made for each problem. Pictures of broken objects were not presented to avoid visual stimuli that might suggest possible solutions (e.g., pu t it back together).
The number of solutions givenby a child to part 1 (p e e r problems) and part 2 (mother problems) are determined and then combined into a total PIPS score.
Scores for the peer and mother problemshave been found to be significantly correlated (Shure, Spivack&Jaeger, 1971). For a description of test administration and scoring, see Shure and spivack (1974a).
(b) The What Happens Next? Game (WHNG) measuresa child'sab i li t y to think of the consequences to peer and adult problems. Each story ends withi'"~child grabbing a toy away from another child or having done something without asking permission fr om an adult.The child is asked "What might happen next in the story?" The script used for the WHNG in this study is present.ed in Appendix C.
Prob l em·So1v ing 50 In the peer problems, the eubject; wa s presented with three pictures, two of an age-relevantchild (presented on 12 . 5 cmx 20 . 5 cm cards) and one of a toy (presented on 7.5 cm x 12.5 cm card). In the adult problems, the subject; was presented with three pictures, one of a mother (presentedon 12.5 cmx20.5 em card), oner,Za child (presentedon 12.5 emx20.5 em card), and one of an item belonging to the mother (presented on 7.5 cm x 12.5 cm ca rd ). There are a mi n i mu m of five basic peer stories and five basic stories involving an adult. If five different, relevant consequences we r e given, the experimentercontinued untilthe child no longer offered new idea s . The same probing procedureis usedas wi t h the PIPS.A child'sscores on part 1 (peer problems) and part 2 (adult pr ob lems ) are then combined into a total WHNG score. For a description of test admi n i s t ra t i o n ar-d scoring, seeShu r e and Spivack (1974b ).
PretreatmentasseSS1I\".n t .
All pr e s c ho o l s we re visited sev eral times before pretesting topr o mot e a degree of famili a rity with the
Problem- Solv ing 51 children. All children in each preschool for whom parental consent toparticipatehad been ob tai n e d were administ ered the rating-scale me a s u r e of social status and thePPVT-R . As well, the PBQ was completed for all ch ild r enbythe preschoolteachermostfamilia r witheach child. Using the information ob t ai ned from these measures , the rejectedand average groupswere selected acco r d i ng to the previouslystated criteria.Theselecte d children were then administered the PI PS and WHNG measures of problem-solving ability. All tests were administered ind i vi d ua':.l y toea chchild by the examiner in separate sessions.
Treatmen t.
Participants in the treatment group received thir t e en , 20 minute training sessions {see AppendixD). All sessions were based on the script developed by Spivack and Shure (1 974). The in i t ial three sessions focusedon developingspecific language concepts to be utilized dur i ng the remaining trainingsessions.Sessions four to eight invol v e d trai ningin alternativethinking skills and we r e designed to encour age the children to think ofdi ff e r e ntwa y s to solve real-life in terpe rso nal
Prob lem-Sol v ing 52 problems. Sessions 9~13 involved training in consequential thinking skills and were designed to encourage the children toth i nk of the consequences of interpersonal acts. During each training ses s ion , a differentproblemwas presented and apos t e r displaying each problem was us e d to aid the participants in understanding the si tuationand to maintain interest.For each tr a i ning session, all ideas sugges ted by the children were recordedon a sheet ofbr i s o l boardfor the childrento see.
Participants in the attention controlgr oup took part in various activitiesfor 13 sessions (s e e Appendix E). These ac tiv i t i e s did not involve problem-solving skills butwe r e designedto stimulate mutualad u l t- c h ild interaction similar to the training sessions. All sessionswe r e 20 minutes in lengthin order to equate the amoun t of attentionthe groupreceivedwith that of the treatment group.
Eachofthe treatmentandatt.e ntLon controlsess ions were conductedon a sep a ra teday. For each respective sessi on, the groups were removed from the regular preschool cl assroomand taken to a nearbyroom to be as free of distractio nsas possible.For trea tme ntsessions,
