Você está na página 1de 51
‘SERVIGO DE ACONSELITANENTO EDUCACIONAL FACULDADE DE PStcoLoGia E DE CIENCIAS DA. Enucacio Usivensmane DE Lispon MFF - 20 (ADAPTADO DE ED CAIRNS & TOMMY CAMMOCK, 1978) © Matching Figures Test - 20 (MFF-20) ¢ uma medida de avaliagio da dimensio impulsividade-reflexdo, Este instrumento foi desenvolvido por Ed Caims e Tommy Cammock em 1978, % Constituido por 20 itens seleccionadas pelos autores com base nas suas correlacdes item- total (preciso) © na capacidade para discriminar entre rapazes impulsivos e reflexivos (validade). Nos seus estudos também foram encontradas boas correlagdes teste-reteste por bipartigao e apds periodos de 2 de 5 semanas. Os resultados mantiveram-se para amostras constituidas por criangas de ambos os sexos, dos 7 aos 11 anos. Esta anilise dos itens foi feita a partir de uma versio mais longa (30 itens), que inclula itens das 3 formas do MFF desenvolvidas por Kagan nos anos 60 (formas F, $ e 2). Para além das preocupagdes psicométricas, os autores limitaram o nimero de itens para a ordem dos 20 tendo em consideracéo os efeitos da fadiga e do cansaco das criangas durante a administrago do teste, A dimensio reflexdo-impulsividade tem sido conceptualizada como uma variavel que descreve © processo cognitive que envolve “a reflexio da precisio de uma hipétese disponivel” (Kagan & Messer, 1975, p. 224). Operacionalmente, 2 variével tem sido definida pela composigao de dues dimensdes: ~ 0 tempo de laténcia da primeira resposta (0 tempo que a crianga leva a dar a sua primeira resposta); ~ © precistio da escola (0 nimero total de erros até a crianga acertar na resposta certa. O MFF-20 avalia estas duas dimensdes para um conjunto de 20 itens, Cada item consiste nna apresentagdo @ crianga de uma figura (figura-alve) e um conjunto de 6 figuras semelhantes, de onde apenas uma ¢ idéntica a figura-alvo. A tarefa da crianga é seleccionar ; de entre essas 6 figuras aquela que é igual ao alvo. Se a crianga acertar a primeira, passa para o item seguinte: Quando nao acerta, é-the dito para tentar outra vez até escolher 2 esposta certa: Sdo medidos, para cada item, o tempo de laténcia e o niimero de erros e, no: final, $30 calculados o tempo de laténcia médio ¢ o niimero de ertos médio (as médias das respostas dadas aos 20 itens). Tradicionalmente os resultados no MFF eram interpretados com base na mediana das respostas da amostra/idade, sem recorrer a normas. Assim, as criangas eram consideradas reflexivas quando tinham um némero médio de erros abaixo da mediana da sua idade e um tempo de laténcia acima dessa mediana (lentas e precisas - slow/accurate S/A). As eriangas eram classificadas de impulsivas quando tinham um mimero médio de erros acima da mediana da sua idade e um tempo de lat€ncia abaixo dessa mediana (répidas ¢ imprecisas - Sastinaccurate FM). As criangas cajos desempenhos recaiam nas outras duas categorias cram referidas por F/A (rpidas ¢ precisas - fastlacoiiftié) ¢ por S/T (Ientas e imprecisas - slow/inaccurate). : Mais tarde, Neil Salking (Universidade de Kansas) desenvolveu normas para criangas dos 5 a0s 12 anos de idade, Verificou-se assim uma tendéncia para a impulsividade evoluir para uma maior reflexio com a idade, por uma diminuiglio do tempo de laténcia com a diminuigéo ¢ depois a estabilizagio do nimero de erros (logo, uma maior eficécia). As diferencas entre sexos no parecem ser importantes, embora haja uma tendéncia para os rapazes fazerem mais erros. Este instrumento tém-se vindo a revelar itil para a disoriminagdo entre criangas impulsivas e reflexivas, O teste também se verificou ser sensivel a efeitos da interveng&o terapéutica, tanto pela comparagao dos resultados antes © apés a intervengéo, como pela capacidade de discriminagao dos efeitos de tipos diferentes de programas terapéuticos (p. ex., apds treinos de resolugéo de problemas e de auto-instrugdes, ambos os grupos se revelaram mais eficazes do que 0 grupo controlo, mas as criangas no primeiro beneficiaram mais do que as do segundo) Medianas do N° de Erros (ME) e do Tempo de Laténcia (ML) por Grupo Etdrio Idade (enos) N ME, ML, > 21 21.96 6.42 6 443 17.52 3.21 7 461 14.02 10.22 8 408 12.25 121 9 563 i> 13.71 10 227 6.68 13.67 nN 297 126 11.75 12 226 1.66 10.68 Médias (M) ¢ Desvios-Padrao (SD) para o N° Total de Erros por Idade e por Sexo Tdade Rapazes Raparigas Total N M sD N M SD N M sD 5 94 23.15 7.53 127 -21.50 6.64 = 221 22.20 7.06 6 245° 1858 7.17 198 «17.17 7.08) 443 17.96 7.16 7 215° 15.00 6.68 246 1425 6.10 461—«14.60 6.38 8 214 13.05 632-194 11.66 5.77 408-—«12.39«&.10 9 252° 9.48 5.74311 8825.63 5639.12 5.69 100 108 7.31 «4.62119. 7.33 5.23227 7324.94 1 M47 839-482-150 7474.45 2977.92 4.65 12 118 8.00 485 108 8.05 4.43226 = 8024.64 Médias (M) ¢ Desvios-Padrao (SD) para o Tempo de Laténcia por Idade e por Sexo Tade Rapazes Raparigas Total N M SDN MSD. NM SD 94 759 556 127 7343.16 221745435 245° 9.70 6.00198 1081-870 443.—«10.19 7.34 215 1251-884 246 11,74 620 461. 12.10— 7.55 214 1298 8.24 194-1417 874-408 13.55 8.50 9 252, 16.69 10.16 311. 16.18 9.355636 AI9.72 10° 108 17.28 11.12 119 17.16 10.06 -227,-—«17.22—«*10.55 Uo 147 13.61 10.08 150 1421837297 13.91 9.25 128 12.68 8.34 408_—12.37 6.02" 226 «12:53:38 en aw TESTE DE EMPARELHAMENTO DE FIGURAS FAMILIARES. MFF - 20. (Cairns, E. & Cammock, T. - Development of a more reliable version of the Matching Familiar Figures Test, Developmental Psychology, 1978, 14, 553-560) Anilise dos resultados ‘Tempos de laténcia Niimero de erros ‘Acima da mediana Abaixo da mediana ‘Acima da mediana Lentos - imprecisos Reffexives, Abaixo da mediana Impulsives Rapides - precisos Resultados obtidos com a amostra irlandesa dades N Eros Laténcias m dp m dp TI anos 7 18.18 313 13.40 Tas Fanos 35 273 11.89 Thi 4a 7 anos 7 35.19 i159 101g 934 'SERVICO DE ACONSELHAMENTO EDUCACIONAL Facuunaps De Pstcotocia & ne Citncias Da Epucacio Usivensipape De Lisnon MEF - 20 FOLHA DE APRESENTACAO DAS FIGURAS : Itens de Treino 1. caneca . 2. régua tens 1. fotha I 2. tesoura 3. dculos 4. cowboy S.casa 6. nave espacial 7. fotha II 8. girafa 9. avio 10. flor 11, navio 12. tenda de indios 13. gato 14, bareo /Jancha 15. TV 16. pato 17. cand 18. vestido 19. urso 20. candeeiro I ‘SERVIGO BE ACONSELAANDINTO EDUCACIONAL, FAcuLpADe be Pstcotoata & De CiENCIAS DA EDUCAGAO ‘Usiversipape DE LisB0A MFF - 20 FOLHA DE INSTRUGOES “Vou-te mostrar o desenho de uma coisa que tu conheces ¢ depois varios desenhos parecidos. O que tu tens de fazer & apontar para o desenho desta folha (apontar para a folha de baixo) que seja exactamente igual a0 desenho que esté nesta folha (apontar para a folha de cima)*. Vamos fazer dois (alguns) para aprenderes/praticares”. Mostram-se os itens de treino e ajuda-se a crianga a encontrar a resposta correcta. “Agora vamos fazer alguns que so um pouco mais dificeis. Verds um desenho na folha de cima ¢ seis na folba de baixo. Procura um na folha de baixo que seja exactamente igual ao que esté na folha de cima e aponta para ele”. Regista-se o tempo de laténcia até & primeira resposta (20 1/2.mn.), 0 niimero total de erros cometidos ¢ a ordem em que se cometem os erros. Se 0 sujeito acertar € gabado (Muito bem”), Se se engana, diz-se-the “Nao, esse nfio é 0 desenho certo. Procura © que 6 exactamente igual a este (apontando para o de cima)”. Continua-se @ anotar as respostas (amas no as laténcias) até que o sujeito cometa um méximo de 6 erros ou até que assinale ‘a resposta correcta, Se no consegue atingi-la, o experimentador mostra qual é. HE necessétio ter um apoio para o caderno que contém o teste, de tal forma que o sujeito possa ver claramente e ao mesmo tempo o estimulo ¢ as alternativas. As duas paginas de cada item deveriam formar um angulo de cerca de 90", * Nota: E recomendavel introduzir as paginas do teste em folhas de pléstico para se manterem limpas. The following guide was developed for recording and scoring private speech on the MEF. ; PRIVATE SPEECH SCORING STANDARDS FOR MEF During administration of the MEF, record all statements made by the child. These are subse quently scored in hioth of the following ways: 1) the total number of inner-directed, insecure, and ether comments, and 2) the number of cards on which one or more relevani and irrelevant com: ments were made, MEF —Scoring Speech for Inner-Directed, Insecure, | and Other Comments 1, Inner-Directed: Anything that appears self-guiding ' Examples: "I can tell by the eyes." The other arm is sticking out. “They all look the same. “This is curled.” “Too big.” “I's the striped one.” 2. Insecure: Anything that indicates child is unsure of 1) his performance, 2) whether the ex: aminer likes him, or 8) whether he likes himsell Examples: (1) Open references to performance or fectings about self or examiner “W's hard.” "This is too hard, “They all look alike.” ‘How are they different “Tean't do it (2) Manifestations of acute indecision “Ws this—no wairie's, no, (8) Asking rather than telling “is this 10" (4) Excuses “1 didn’t see the legs." (8) Blatant distractions “My stomach hurss." (after 2 series of wrong answers) "Can we stop now?" 3. Other: All other verbalizations (social, word play, outer-directed comments, laughing, sing: ing, statements of choice) Examples: “It's this one." °s this.” "The cat’s eyes are mad.” “These are hard.” "Do you have seme more games for me?” MFF — Scoring Speech for Relevance : Score cach card for Relevant or Irrelevant speech only once. Each card may have a total of 10 scores but only one of each type. Relevant: Anything directed to task Examples: “Let's ee." “This one.” “This is hard They all look the same.” 1 figured it out from the pickets “I should have known it was this. 270 Measurement of Private Speech | : Irrelevent Examples: All other verbalizations Um" ‘Um hum” Laughing, singing Ts that a rooster or a giraffe’ ‘My stomach hurts “A cowboy!” ‘What time will we be done?’ ‘Do I get a prite like Rayrnond?” Final scores are derived as follows: MEF-Inner = Total number of inner comments MFF-Other = Total number of other comments MFF-Insecure = Total number of insecure comments MFF-Relevant MFF-lrrelevant Total number of cards on which one or more relevant comments occur Total number of cards on which one or more irrelevant comments occur ‘Measurement of Private Speech 27 Development of a More Reliable Version of the Matching Familiar Figures Test ED CAIRNS axp TOMMY CAMMOCK. New University of User Coleraine, Northera treba Because of the inadequate relinbility of Mutching Familine Figures Test (MFFT) cerror scores, a longer, more reliable version was developed. As a result of ilem analysis of 30 MEF T-type items, 20 were selected on the basis of ilem-total error correlations and ability to discriminate reflective from impulsive I -year-oll boys, ‘Two subsequent studies with subjects of similar ages suggested that split-hall correlations for the new 20-item MFI over? weeks were 9 forlatency and .89.or rors, while test-retest correlations over S weeks were valculated as RSand .77for Fvency and errors, eespectively. These results plus those of further study with.7- sand 9-yenr-old boys and girls led to the suggestion that the new MFF (the MFF20) is, suitable for use with childven in the age range 711 years Although many test procedureshave heen cited as adequate measures of the reflec- tion-impulsivity dimension, the Matching Familiar Figures Test (MFFT) is generally accepted as the "primary index.”” Kagan, developed three forms of the MFFT in the 1960s. Egeland and Weinberg (1976) point out that “none of the sources has discussed in detail the psychometric de- velopment of the tests; in fact, different pri- nary references are often cited for the same form of the test’* (p. 484). Form F of the MFFT is the most popular version, and Form S, which has generally been used as a positest measure in pretest and posttest in- tervention designs, is no longer distributed Form K, the younger children's version, also has 12 test items but an array of four instead of six alternatives. The reported test-retest and internat con sistency reliabilities of the MFFT tests (Ault, Mitchell, & Hartmann, 1976; Egeland ‘We are grateful to J. Kagan and T, Zelniker for prom viding the tes lems. ‘Copies of the MFF20 may be obtained. at cost from the first author Requests for reprints should be sent to Fd Cairns, Department of Psychology, New University of Ulster Colersine, County Londonderry. Northern ireland. 3 2 [Conran ny he Ame & Weinberg, 1976) despite their statistical significance have tended to be low to moxler- ate by most psychometricstandards. Ault et al, suggest increasing the test length and es- timate that “a test of 80-96 items would be necessary 10 increase the reliability to .90 average of .52°"(p. .230). OF course, consideration of such factors as fatigue, boredom, etc., necessitates limiting the test length to not more than 24 items when testing children. Consequently. any ‘ultempt to enhance the reliability of MFFT error scores should concentrate on improv- ing the discriminative power of the items, thus eliminating the need for a protracted version of the lest. The most efficient way to select items that yield a good discriminative index is by item analysis based on the corre fation of each item with the total test score (Nunnally. 1967). Nunnally argues. that those items that correlate highest with the total test score should be selecied because they are probably less ambiguous, cannot be extreme in difficulty ineither direction, have more variance relating to the common factor among the items, and will tend to make the final test highly reliable: The aim of this study was to improve the reliability of MFFT error scores by increas- Pasblege Avssnn, I OB AYERS ONSEN 585 dl 556 ED CAIRNS AND TOMMY CAMMOCK. ‘Table 1: Hem Analysts of the 30-Iegm Matching Faniitlar Figures Test (Errors Only) =f Mea errors if Retectives”—Impulaves ToialY Selection Nees Form TP Hoae F 7 ry Ye ry. z ' 9 Yes Actoplane 3 1 ra Ye Salser 3 1 u ne Dew F 1 ° Yes Spaceship z 2 Pa Yer Sh F 2 2 Yer as 5 2 et Ne Tree e 2 it Ne Lat 5 2 = Yer instr s 2 ry Yer Dor 5 3 cy No cat e 3 oe Yer Dress z 3 2 ver Telephone é 3 a Ro bret tt s ‘ 41 No Graph 3 ‘ BY No Comey F 4 o Yes Specibost z 4 a ves inate F ‘ a ver Bel s 5 2 No Flower 5 5 a Ye Lame t F 5 co va res Ht F 3 Ff No Duck HHS! a Ye eat Fo6 3 Yer Baby 5 6 a No Wigwam 2 6 a5 Yer Late 5 5 2 Ye Seis e é 9 yea if the umber of items and selecting, through item-total error performance, only those items providing an efficient dis- native index Experiment 1 Method Subjects. Subjects were 98 boys with an age range of H.£-12 2 years atiending secondary schoo! ina woek- ing class area. Boys were chosen because there af€ ‘moze data on boys" conceplual tempo and hecshse in several studies researchers have fudné sigiicant sex Aiferences in responding to the MFFT (e.g. Lewis, Rausch, Goldtere & Dodd, 1968), Based on Moray Howse Verbal 1Q scores. obsained 8 months porto this experiment, subjects were selected to represent the verbal 1Q range #0-130, i Apparatus. The apparais consistedof32 MFFTitems, 2practice and 30testitems (Form: Fand’S ofthe MFFT tests and 6 cther items supplied by T. Zeliker). ot testitems, which had been thoroughly shulied pir to ing of each subject, were presented according io tirections (Kagan, 1965). Latency (hist re sponse and number ofchoiees lo correct response were ‘noted for each item, Results ‘The criterion for selection was highest itentota error corzeation with he added provisions tha the iter cleuty discriminate between reflectves" and impul sives" error scores and also thal each target positon be adequately represented. The ltem-total enor correla tiuns were ealeulatedusing Pearson's product-moment technique. Subjects were classed at reflective oF im pulsive using the double median split entedon. The ‘median errarand ateney scores were errors and 11-7 See, respectively. This resulted in the cassifcaion of ives. 3vimpulsves,and 13fastaccurateand 10 accurate subjects. The mean error scores of ives end ipusives on each item were computed fand t tests conducted to determine which Siem dls: Criminated between the eror scores of refectives and Iimpulsives 98 for ise Taxed or (em-tatal¢ ing from sponse tim those item: posttions {oval ervor adequate 1 maximizing The ait bility coe in Experi tion bet apart as To calcul it was order pot items int. cation of was also presentat ture inve order pos trated in ‘Table 2:1 em Tat Seine Gates Cow Howe Spaceship Grate Aeroplane Blower Gr Speedbost Lino Bent ot the 0 Tata tion of and 10 res of puted odie es an AMORE RELIABLE MEFT 387 impytsives. Coefficient aps forthe Mbitem MEFT was Ok for Tatency and A Tee errs, both pe Ttasoin evoor scores ry, five items fi eins reflective fromm Htenntoal conor correlations were disap From 18 to 60 fegmparable cortl spunse fines treed from 70 10 90) those iteray in whieh the target occu pions tice, Positions 29 8) reales feqvate Fepwesentaion of al figet positions, (O05 sndmiring response uncestaity. lems were selected from ategovies ever the the vecatons were (ene al cme em ie heiween efectve and Experiment 2 Phe aim of this study was to obtain relia. yy coefficients for the MFF20 developed in Experiment | using the corrected correl- tion between splitchalves given 2 weeks apart as recommended by Nunnally (1967). To calculate reliability using this technique, it was necessary to appoint each item an ‘order position on the test anxl to divide the tems into two comparable halves, The allo- tation of order positions to items on the test ‘o necessary to provide a standard presentation order for the reference of f- ture investigators using the measure. The order positions of items on the test are illus tried in Table 2. bil ‘Order Position on the MFF20 Form Order pasion Tiuget ies Scie Ghaseee Cour fiowse Spin tee Girte toner Ship ‘Tole 3: Meant and Standard Destations of Scores (btained on the Delayed Spiit-Halt Admintstration of the MFF20 Errors Rr Group Set. The allocation of items to order positions was random except for the provision that items with similar target positions were re- quired to oceupy equivalent order positions onthe first 10 compared to the latter 10 order positions on the test, This provision ensured that items were assigned a fixed-order posi- tion and organized into two comparable halves (items 1-10 were Set A, and 11-20 were Set B). Method ‘Subjectn, Subjects were 30 boys with at age range of Tr boll Dyent attending a secondary school ina work, fag close area, Based on Moray House Verbal 1Q Scores, nbtained f months prior (0 this stndy, the sub- jects were selected To represent the verbal 10 ange Ks llas- this sample was equivalent to but ikependent tthe subjects tested in Experiment 1 Procedure, Subjects were aesigned rnslomly to recive Bev of Set Afest followed By he altersate vet days, ern eachoccasion to practice tems were admin tered frst the standned procedure was followed (Ka fan, 1968), and latency 10st response and numberof Ehoices to correct response were noted foreach item. Results The delayed splithalf (alternative set) product-moment correlations for errors and fatency were estimated at .80 and .83, re- spectively (for both, df = 28, p <.01). As these correlations were derived from cor- relating scores on two subsets, each only Cone-haif the length of the complete test, the reliability of the full-length test was obtained by applying the Spearman-Brown formula. 558, ED CAIRNS AND TOMMY CAMMOCK Application of this formula indicated that the 2G-item MFFT should have reliabilities of .89 (errors) and .91 (jatency) Discussion : Coefficient alpha for the 30-item MFFT in Experiment ! was .98 forllatency and .81 for errors (both ps <.01). Thus reducing the test length from 300 20 items had no appre- ciable effect on its reliability; in fact, error score reliability improved slightly. testifying lo the efficiency of the item analysis In order to assess if Set A and Set Bitems were of comparable difficulty, the mean total error and latency scores of the 30 sub- jects on each set were compared. On Set A items, there was a mean error score of 12.23 (SD = 5.90) and a mean latency score of 11.06 (SD = 6.76), compared with a mean error score of 11.40 (SD = 6.27) and a mean latency score of 11.25 sec (SD = 6.94) on Set B (see Table 3). No significant differences between the means for total error, 1(28) = .53, ns, and mean latency, (28) = .12, ns, were te- vealed, indicating that the sets are of equiva- lent difficulty. These results therefore sug- gest that the MEF20 may be divided into two comparable halves, sufficiently equivalent for pretest and posttest research purposes, pending the development of an alternative 20-item form. Experiment 3 According to Nunnally (1967) the cor- rected, delayed split-half correlation is ‘more appropriate measure ofa test's reliabil- ity than the correlation yielded by the test- retest procedure, However, MFFT reliabik ity has typically been determined using the latter design, and consequently the reliabik ity coefficients derived in Experiment? have ‘no comparison in the conceptual tempo Ii crature. A third study was therefore con- ducted to investigate the test-retest reliabil ity of the MFF20. This also provided an ‘opportunity to examine the lateney-error re- lationship, Method Subjects, Subjects were 37 boys, with un age T3119 years, attending a sevondary schoo! in working cls aren, ated on Moray Honse Verbal 1Q scores, obizined & months prior (@ this experiment, subjects were selected fo represent the verbal IQ range 85-125. Tht sample was equivalent to but independent of children tested in Experiments 1 and 2, Procedure, Subjects were tested individually, Fellow Ingtwo pracice tems, tke 20st items were presented inthe Bxed.order postions allocated in Experiment 2 (See Table 2) according to standard procedure (Kagan, 965) Five weeks later the sameexpecimenter teviited the Schoo! andretestedthe subjects with the same 20items, Results ‘The mean total error and mean latency scores of the 37 boys on the first administra- tion of the test were 18.84 errors (SD = 9.13) and 13.40 see (SD = 7.44), respectively. At retest the mean total error and mean latency scores were 16.22 errors (SD = 9.15) and 12.2 sec, (SD = 7.20), respectively. ‘The product-moment stability coefficients were calculated at r(35) = 85 and r(35) = .77, both ps <.01 for latency and errors, respectively. For latency and error scores obtained on the first administration of the test, (35) = —.57, P< 01; and on the second administration, 735) = -57.p < 01. Discussion ‘The stability coefiicient for errors on the MPF20 (.77) compares favorably with the best error stability coefficient (.36, p <.05) reported for the MFFT (Form F) when ad- ministered to a similarage group (Egeland & Weinberg, 1976). Furthermore, the results indicated that the latency-error correlation (r= =.67) isin accordance with the range of correlations (° = =.50 to ~.60) that Kagan (Kagan & Messer, 1975) hitherto has con- sidered adequate. Kagan has always maintained that MEFT response time is “indepenuent’ of or "or- thogonal” to verbal skills (see Kagan, 1965; Kagan & Kogan, 1970), and the evidence generally has supported this suggestion. In the present study Moray House Verbal 1Q scores. obtained some 8 months earlier yielded correlations of ~.04 and =.06 (for both, uf = 35, us) for latency and error scores, respectively. Even if these correla- tions are corrected for attenuation due to low score reliability, the correlations rise to ten err A MORE RELIABLE MFFT 559 ~.04 (latency 1) and ~.07 (errors 1Q), both nonsignificant jeriment 4 This series of studies has outlined the con struction of a more reliable version of the MEFT, So far, however, it has only been demonstrated that the new test, the MFF20, is suitable for use with boys in the 1+ 10 1-year age range. Because the items in the MEP20 were selected in order to be of op- timum difficulty for the 11- to 12-year-old: we undertook an examination of the reliabi ity of the MFF20 with a younger age sample ‘and one that included girls as well as boys in order to examine sex differences in perfor- mance on the MFF20. Method Sete Subjects inched 63 bays(a » 31) and iri ay slihs mean ape of 7.6 yearsand 32 boys (n = 29) Jad pinks (a 2 2) with 9 mean age of 9.7 years, all of ‘whom attended the same elementary school Procedure. The chikien were tested individually. aod the MEE2O was administered foltowing the usual 0 practice items ana! according (0 the standard proce Tres tkapan, 196%), Two weeks later, the Verbal Henning subtest of tie Primary Mental Abilities test {CTinwstone feThurstone, 1963) wasaudministered tothe eileen in sro. Results For the 9-yeur-old group the mean latency and mean total error scores were 11.14 sec (SD = 4.84) and 27.73 errors (SD) = 11.89). Coefficients alpha for the two principal scores were .7R for errors and .94 latency. with latency and error scores yielding r(50) 67.p <1. Correlations between ver- bal intelligence test scores and latency anc error scores were ~.07 and ~.06, respec- tively (if = $0. ns) For the T-year-old group the correspond- ing results were as follows: mean latency = 10.14 sec (SD = 9.54); mean total errors = 35,19 (SP = 11.59) with coefficient alpha Calculated as .69 for errors and .92 for k teney. The correlation between kntency and errors was —.£2(61),p < 01. and these two measures yielded (61) = «I and r(61) = 2°21, both nonsignificantly, with intelli gence. Inorder to assess the influence of age and sex on MPF20 performance, two-way anal- yyses of variance were performed on the la- tency and error scares. ‘The analysis of the latency scores revealed no significant main effects or interactions. The analysis of error scores, however, revealed a significant main effect die toage, F(1, 114) = 12.19.p < OL, ‘but no main effect due to sex nor any Sex x ‘Age interaction Diseusston ‘The results of the present study suggest that at the 9-year-old level the MFE20 dis- plays the same qualities of reliability (for both latency and error scores), high negative latency-error correlation, and a lack of cor- relation with verbal intelligence as it did in the earlier studies with Jl-year-olds. The results from the 7-year-old group are equally satisfactory except that the reliability of the trror scores is somewhat lower than might be desired. However, error score reliability i still higher than that usually reported for the MFFT (Form F), especially for this age group (Aull et al., 1976). Further, the analysis of variance suggests that sex differences are not important for MEF20 performance, while age differences are important but only for error scares. tn fact, if one compares the MFF20 latency scores for the 1 -year-olds in Experiment 3 (13.40 sec) and the 9-year-olds (11.14 sec) and T-yenr-olds (10.14 sec) in the present Study, it would appear that even between ages 7 and 11 years there is little increase jn Tatency. Similarly, the respective error scores are 35.19, 27.73, and 18.14, which suggest a decrease in errors over the age Tange 7-11 years, These results are identical in trend to those obtained by Cairns (in press) for MFFT latency scores and to those obtained for error scores except that with the MFPT it appears that errors only de- creased between ages 5 and 7 but not there- In conclusion the MFF20 would appear to be a test that can be safely recommended for tise within the age range 9-11 years for both boys and girls. Caution must be exerted When the testis used for those younger than 9 years, and it is not recommended for use ‘with chiklen uncler 7 years. 560 ED CAIRNS AND TOMMY CAMMOCK, REFERENCES ichell, C,, & Hartmann, D. P. Some 1 problems in relleclion-impulivty ‘esearch. Child Development, 1976, 47, 227-23), Cairns, E. Age and conceptual empo. Journal af Ge netic Poychology, in press Egeland, B.,& Weinberg, R.A. The Matching Famiae Figures Test: A look at its psychometric credibly Child Development, (916,47, 483-491 Kagan, J. Iempalsive and reflective children: Sig ‘afieance of conceptual tempo. In J. D. Krumboliz (Ed), Leoming and the educational process Chicago: Rand-MeNaly, 1965 Kagan, J. & Kogan, N. Individval variation in cogt tive processes. InP. H. Mussen (Ed.),Carmichael's manual of child psyehology (318 ed). New York Witey, 1970 Kagan J, & Messer, 8. B.A reply 10 "Somme iisgiv- ings about the Matching Familiar Figures Test as a measure of rellection impulsivity.” Developmental Prvcholugy, 1995, 1, 244-268, Kagan. J, Pearson, L., & Welsh, L, Conceptual imput slvity ad inlucive reasoning. Child Development, 196, 37, 583-594, Lewis, M., Rausch, M., Goldberg, 8., & Dodd, C. Error response and time and 1Q: Sex differences in cognitive style of preschool cildeen, Perceptual and Motor Skis, 1988, 26, 363-568, Nunnally. J.C. Psychomewrie theory, New York ‘Thurstone.L, L., & Thurstone, TG. Primary mental abilies. Chicago: Science Research “Astociates, 1961 Yando, R,M.,& Kagan, J. The effects of teacher tempo (nthe child, Child Developmen, 1964, 29, 27-34, (Received November 18, 1977) Manuscripts Accepted (Cominved from pope 473) Children’s Moral Judgments as a Function of Intention, Damage, and un Actor's Physical Harm. Jerry Suls (Department of Psychology, State University of New York at Albany, 1400 Washington Avenue, Albany, New York 12222) and Robert J. Kalle. Development of Selective Listening and Hemispheric Asymmetry. Gina Geffen (Psychology Discipline, School of Social Sciences, Flinders Ur ersity of South Australia, Bedford Pari, South Australia, Australia $042) and,tocelyn Wade. Effect of Language on Preference for Responses to a Moral Dilemma. Joseph J. Moran (Department of Behavioral and Humanistic Studies, State University of New York College at Buffalo, Buffalo, New York 14222) and Andrew J. Jona. Androgyay Across the L fe Span. Janet Shibley Hyde (Department of Psychology. Bowling Green State University, Bowling Green, Ohio 43403) and Diane E Phillis. Perceived Determinants of Hi Runyan (1203 Tolman ghs and Lows in Life Saiisfaction. William MeKinley |. Institute of Human Development, University of California, Berkeley, California 94720), Care:Giving and Infant Behavior in Day Care and in Homes. Judith L. Rubenstein (Division of Child Psychiatry, Tufts New England Medical Center, I71 Harrison Avenue, Boston, Massachusetts 02111) and Carollee Howes. Children’s Orientation of a Listener to the Context of Their Narratives, Carole L: Menig-Peterson (Memorial University of Newfoundland, St. John's Newfound. land, Canada) and Allyssa McCabe. Development of Activity Level in Chibdren Revisited: fects of Mother Presence, Donald K. Routh (Department of Psychology. University of lowa, Lowa City, Towa 52240), Marsha D. Walton, and Efrat Padan- Belkin, Neonatal Precursots of infant Behavior. Raymond K.. Yang (Department of Child and Family Development, University of Georgia, Athens, Georgia 30602) und Howard A. Moss. Running Head: MPF Noz=s MFF Norms Abstract ‘This report presents a set of normative tables for the Matching Familiar Figures test for ages five through twelve for sales and fenales. Descriptive information, as well as percentile ranks for raw total error and mean latency scores are presented in tabular forn. From a total data bank of 8172 entries, 2846 were used to chart the developmental progression fron impulsivity towards ears of age, after which perfornance is increasing reflectivity through. ten characterized by a decrease in latency coupled with a stabilization in errors, 3 The utility of the norms and their implications hence an increase in efficiency. for further cognitive tenpo research is discussed. : . Introdvetion | dn increasing mnber of studies in developmental psychology has focused on a dimension of cognitive developnent referred to as "reflection-impulsivity” or "conceptual tempo" (Kagan, Rosnen, Day, Albert, and Phillips, 1964). Within this group of studies, inforsation has accumilated dealing with the dimension and its relationship to a host of other.yariabies,.such as sex, problea solving strategies, color-fora preference, and self-control. H Reflection-impulsivity has been conceptualized as an individusl variable describing the cognitive processes involved in "reflecting on the accuracy of available hypotheses" (Kagan & Messer, 1975, p. 224), Operationally, the variable has been defined as a composite of two dinensions: latency to first response, and accuracy of choice or total errors: These two dimensions are combined in the use of the Matching Familiar Figures (NFF) test, (Kagan et al., 1964), regarded as the prinary (and cften the only used) index of | geflection-inpulsivity. Using the standard fornat of the test, each child is presented with a picture of a faniliar object (the standard), and a set of highly similar variants, where only one of the set of variants is exactly the sane 2s the standard. The task is to select fron the set of variants the one that is exactly the same as the standard. If the response is correct, the ‘child goes on to the next iten. If incorrect, the child is told to choose again, usually with a maximm of five errors per iten. The dependent measures exployed are mean tine to first response across 211 itess (mean latency), and total nuber of errors across all items. Traditionally, those subjects whose : =e scores fall below the sample median for errors and abeve the sample nedian for latency (slow/accurate oF S/A) ave referred to si Feflective) while those subjects whose scores fall above the sample median for errors, and below the sample median for latency (fast/inaccurate or F/I) are referred to as Sapulsive> Subjects whose scores fall in the other tvo cells have usually been referred t0+ FF Norns and g the two 4. slow/inaccurate (or S/1). stov/inaccurate (or S/1)-. “one of the primary problems, however, which results from the median split classification procedure is that individual performance is merely typed, and not scaled. That is, since all subjects are placed in one of four classification cells (S/As F/T, F/A, or S/I), no within cell discriminations in perfornance . between subjects are made, This traditional method of using median splits on poth the tine and error dimensions and the specification of four groups =. generated by sample characteristics has been the most comon classification strategy. Based on such a classification strategy, the generalizability of MEF test scores; as well as associated research findings are at best questionable, regardless of the internal validity of the various forns of the instrument. Salkind (1975) has shom how the use of sample-genezated norms for classifi- cation severely limits réneralization of research findings. Even when means and standard deviations acress different samples are comparable, (which is infrequent in the liter: e), medians are often quite different (i.e., different classification for individuals in different sazples with sinilar seores). These discrepancies are striking in light of the frequent ‘comparisons readily drawn between experinental populations based on their similarity in perfornance on “the MFF. Often the only criterion that is used to establish intersample. sinflarity is the consistency of the negative corzelation between errors and Jatency, rather than any systematic comparison of nore descriptive parameters such as means and standard deviations. 2) the use of the pera eere traditional classification strategies such as the median spl De re ere eee Accordingly, the following points are stresse wate and presents sone potential for (based on sample medians) is inade nisclassification of subjects, and 2) the total lack.of normative data for the Matching Familiar Figures test constitutes a very serious_zestriction on the MFF Norms z Procedure qhe initial step in the development of the data pool was to contact 4% BAny investigators as possible who had used the MEF in 2 research setting: From arches of the published (e.g. professional journals such as Child Dev Lopsent, eee.) and unpublished Literature (e.g. convention prograns, ERICy ete.) 2 total of 350 potential sources were identified. A letter was sent to potential participants with instructions to return the stamped self-addressed return postcard indicating their decision as to whether to take part in the norm development. Of the 580 solicitations made, 179 (51%) responded, and of that 179, 97 respondents (54%) indicated they would contribute MFF data 0 the data bank. Eaeh respondent was then sent instructions and data sheers to assist in the preparation and submission of the data to the céntral pool. Since a variety of forms of the MFF other than the standard have been used, it was necessary to collect information from participants such as nenber of ~ Geens, number of alternatives, ere.3 In addition, procedures, such as munber of errors alloved per iten, differed from researcher to researcher, 2 aid scoring practices (e-g- mean latency to first response vs. total latency to fizst response), and therefore explicit format instructions were given to each parti- eipant. Although the final pool resulted in relatively few MEF contributions which were highly deviant fron the standard form (approxinately 58), in retro spect, there were enough discrepancies to warrant these precautions. Sone of the information requested from participants (see Table 1) required elaboration beyond a mmerical response. For example, number 10, “Description of Sample" required a description of the sample and any outstanding characteristics the participant believed were inportant, Although this left the decision as to - what constitutes other than normal" to the researcher, it was the only practica way to minimize the mixing of samples of children differing on sone clearly ' relevant, although possibly unclearly defined variable such 2s “pyperactive” oF EEE eCeeCeeeeeegeeeeeeeeeeeees cece eee ESS eee MFF Noras insegrity (i.e, external validity) of the construct. of conceptual terpo 25 seasured by the MPF, Although the construct validity of the MFF has recently heen questioned (Block, Block, & Harrington, 1974), for the purposes of this report, and in Light of other evidence providing such validity (lesser, 1976), this issue will not be discussed. * me purpose of this noraing project was to establish bank of MFF data, and to use this infornation for the developrent of normative tables. Noms are to be developed for two dependent variables, total exrors and pean Tatency to first response. y wprain danaged", The same situation was encountered for social class, where an arbitrary decision was required on the participant's behalf. Regardless of the metric originally used to record age, it was transforned fon, Each contributor was into years (26 months) for the final nora prepazs requested to record mean latency to first response, ss well as total errors for each subject. Although it would have been very helpful to examine individual jtea data, such an effort would have been nore than many respondents vere willing to contribute. In many cases individual item responses were no longer retrievable, and more often participants found it beyond their tine constraints to supply such information. : Description of the Sample Of the 97 individuals who indicated an interest in participating in the developnent of noras, 51 (S34) actually contributed data, accounting for a total of 8,172 separate MFF administrations. Table 1 presents a "breakdown" by Table 1 about here Exequency and percent for each of the relevant variables fron the infornation sheet. From an examination of the first three variables (number of test itens, nunber of practice items, and musber of alternatives), it is clear that the coverwhelning majority of researchers used a format, identical to that of the original MFF, having 12 test items (92.3%), two practice itens (98.3%), and six alternatives (96.08). O£ the total population, 97.0% was described as being “pomal", 96.8% as middle class), with a breakdown by sex of 52.6% males and 47.4% females, The range of age groups was from 5 to.85 years with poe to 12-yearzold range the majority of the population falling within the (46 months). : Data Selected for Inclusion in the Norms ‘A decision had to be made as to what specific groups or samples would be ineluded in the final norn tables. Based on the results of Table 1, it was decided to include all data collected using the standard form of the MFF, ineluding both sexes, and results from populations desczibed as normal and middie class. At best, the last tvo descriptors mininize the possibility that there are extrene test scores included in the final nora tables. In addition, ed the the age range fron five to 12 years was selected since 1) this represen largest number of data entries (at least 100 of egch sex), 2) it is the most frequently tested population of children as reported in the literature, and 3) there is some question as to the appropriateness of the MFF for younger caildren (Kogan, 1977)- one other inportant criterion which vas considered in defining the norning population was the nationality of the sample. From the initial requests for ese, Israeli, participation, data were received fron administrations using Js British, French, Canadian, German and Australian subjeccs- Tt was planned a priori that, “phy shibiereteyted.in tha continua gnited Seetes wo! a be “Included, and indeed sone cross cultural comparisons were made which indicated large differences between groups. ‘Descriptive information on these groups is available on request, and at least one formal comparison has been made (Salkind, Kojima, and Zelniker, 1977). The final data pool, including only children Sizst Table 5 about here a response for each of the eight age groups. = Table 6 presents the medians for total errors and mean latency by respond Table 6 about here per age group, Each row represents 2 set of data from an individ For example, six different contributors sent in data based on adninistrations of the MFF to five year olds. One subsample of five year olds consisted of 57 data entries with medians of 21.2 for total errors, and 6.4 for mean latency 7 MEF Norns Hee 10 to first response, and so on. This table docunents the inappropriateness of using the traditional sample-generated median splits 25 classification serategy. Figure 3 illustrates the intersection of the median scores for total ——__—__ Figure 3 about here —_—_—— eqrors and nean latency to first response within each circle for each age group. For exanple, for all five year olds (¥ = 221), the median for total errors is 21.96, and the median for mean latency to first response is 6.42 (fron Table §), Surrounding each of the eizeles in Figure 3 are various nates for a single subgroup from qusbers, each one representing a set cf coord: table 6, showing the distribution of nedian scores when specific age subsamples obtained from different contributors are compared with one another, Only subsanples larger than ten were included in Figure 3. Each of the circles denoting the intersection of medians for the total age yroup is divided into fourequal quadrants, as is done with the traditional Jouble median split classification nethod. According to the code presented in the upper right hand corner of the figure, one can determine where each of the separate subsamples belongs according to the nedians for the total age srove Clearly there is sone diserepancy in the classification of children based on dian scores. For sample median scores, versus those based on population ckample one group of seven year olds (N #148) would all be classified as slow/inaccurate eight year olds according to the total age group nedians. To generate a fourfold table based on the characteristics of this sazple, and then to generalize to other seven year olds is at best very chancy and not productive. Jables 7 through 10 show percentile ranks by raw scores for total errors, and mean latency for males and females. For all tables, the left hand column —_——_—_—_—_—_—— Tables 7, 8, 9, 10 about here —_—— Lists the percentile rank, while the age group is listed horizontally at the MEF Noras 1 top of the table. Percentiles range from zero to 99, with the uppermost linit of any one percentile entered in the body of the table. For example in gable 7, the median or S0th percentile for five year old sales is 22, while for gix year old males the nedian is 17. It is not untit the S4th percentile that the raw score for six year old nales is accompanied by change in the percentile rank, To deveraine percentile ranks the user should 1) compute total errors, or nean latency to fizst response for alt 12 test itess for one individual, 2) locate the child's score under the appropriate 5°% and age group, and Jefe the corresponding percentile rank. For exanple, a ten year old ferale has a total error score of 19, and a pean latency to first response of 13.2. Using Table 8, one should locate the score for total errors of 19, and read to the left the percentile rank of 97. Likewise, one should proceed go Table 10, locate the score of 15,2 and read to the left the percentile score of 49. percentile ranks represent the relative position of an individual's perforsance within his or her age group. Users of these tables should be aware that it is questionable to assume that percentile ranks are at the interva: jevel of measurenent and can be used in lieu of raw scores. within the donain of psychonetrics, the usefulness of percentile ranks has been debated and reader caving a further interest in this area should consult the following reference, (angoff, 1971). : The neans-and standard deviations reported in Tables 2 and 3 form the central elenent of these norms, and it is through their use that such a norning attenpt assumes’ a degree of validity, Altheugh the percentile ranks are only useful for within age-group cocparisons, the descriptive information provided for each age group allows for single-subject testing and evaluation within 2 specific age group. This elizinates the need for obtaining sarple nedians fer classification purpeses- MFF Noras : 12 An Alternative Scoring Procedure Salkind and Wright (1977) have recently presented a model conceptualizing reflection-impulsivity as two orthogonal dinensions; impulsivity and efficiency, heir model employs these parameters in an attenpt to resolve sone of the | with the } scoring problems traditionally associate: Fe The proposed model defines the specd/accuracy donain in tems of txo BeBe peered ereerer deco ee eee eer ulsivity and efficiency, The axes of the original seazte-plot constructs, =—_ —- are rotated so that impulsivity is defined as a dimension of individual pe differences ranging from fast-inaccurate to slow-accurate performance. E¥ficiency is defined as a dimension conceptually and computationaliy orthogunal te impulsivity. along which individual d nees range from slow inaccurate to fast acew Impulsivity and efficiency scores (I and E respectivel") are generated from raw latency and error scores by the following formul; dl di ei TLE \ Fete 7 Fay » where I, + inpulsivity for the ith individual; E, + efficiency for the ith individual; z_j a standard score for the ith individual's total errors and 24, * a standard score for the ith individual's mean latency Large positive I scores ars indicative of inpulsivity, and large negative 1 scores indicate reflectivity, Highly positive E scores indicate inefficiency and highly negative E scores indicate efficiency. The two resulting dimensions are represented graphically in Figure 4, Hypothetical data points are also Figure 4 about here ‘shown, each representing an individual's location by standardized error and latency scores. For each data point perpendiculars dropped to the rotated axes indicate the I and E scores for that individual, Further discussion of the 13 + significance of this nodel for the construct of cognitive tempo is contained dn Salkind and Wright qs77). * tn sum, the use of I and E scores as dependent variables in conjunction with the infornation from the norsing effort, is recommended when individual children or groups of children are adninistered the Matching Faniliar Figures test. In place of a typology based on classification, the use of I and — scores not only yields continuous dimensions (raw latency and exrors do that), but also separates a dinension of effectiveness of infornation processing nce (speed (speed-cun-accuracy) from one of stylistic and strategic preft versus accuracy). It is an interesting sidelight of the norms that the developnent of MFF perforsance with age indicates that a change in stratesy (I-scors) from impulsive to reflective appears to precede a change in efficiency not become (E-score) fron reflective to efficient. On this test at least one c efficient until one has adopted th: nore effective strategy. 7 To the extent that subsequent studies employ these norms (with or without Land E scores), it is hoped that confidence in and comparability of MTF results will be enhanced. Validation and generalization of the construct of conceptual teapo will perhaps be facilitated as well. MEF Korzs - REFERENCE NOTES : salkind, N. Errors and latency on the MFFT: A reassessment of classification strategies. Paper presented at the Biennial Meeting of the Society for Research in Child Development, Denver, April 1978. , 7. Cognitive tempo in Japenese, Salkind, Nw, Kojima, H., and Zelnike american, and Israeli children, Unpublished manuscript, University of Kansas, 1977. A 2 -Psycholo. MFF Norns 1s REFERENCES Ancoff, W. H, Seales, Norms, and Equivalent Scores, In Robert L. Thorndike (ed.), Educational Measursnent, (2nd edition), 1971. Block, J., Block, J. H., and Harrington, D. M. Some misgivings about the Matching Faniliar Figures test as a neasure of reflection-inpulsivi Develonment2l Psycholory, 1974, 10, 611-652. Kagan, J., Rosman, B. Le, Day, D., Albert, J., and Phillips, 4. Ingoraation processing in the child: Significance of analytic and reflective attitudes. cal Monographs, 1964, 78 (1, Whole No. $78). Kagan, J., and Kogan, N. Individual variation in cognitive processes. In d Psvcholory. Vol. 1. P. Mussen (ed.), Carmichael's Manual of © New York: Wiley, 1970. Kagan, J., and Messer, S. B. A reply to "Sone misgivings about the Matching Faniliar Figures test a: a measure of reflection-impulsivi: Developmental Psychology, 1975, 11, 244-248. Kogan, N. Cognitive styles in in:‘ancy and early childhood. Hillsdale, N. Js: Lawrence Erlbaun Associates, 1976. Messer, §. 8. Reflection-inpulsivity: A review. Psychological Sulistin, 1976, 83, 1036-1052. ‘ : Saliind, N., and Wright, J. C. The developnent of reflection-inpulsivity and cognitive efficiency: An integrated nodel. Hunan Develoonent, 4977, 20, 377-387. HF Norns 16 FOOTNOTES lonis project was made possible through the support of NIMH Grant MH28487. Zee author would like to thank the following people for assistance in data preparation, Elns Cox, Lisa Lowry, Jo NeGreth, and chris Nelson; for typing Prepareial and final dragts, Pat NcGregor and Ann Trusty; and for substantive coments and guidance, the Developnental Group at the University of Kansas. seetospletion cf this project would have been impossible without the fooperation of those researchers, too numerous to nane, who contributed data ght for support and assistance. fo the nors. Special thanks go to John C. Ne: ‘J. Salkind, Bailey Hall, “Copies of this document are available frea Nei, University of Kansas, Lawrence, Kansas 66045. 3yq general these different forns of the NFF, as well as di: ferent copies of, Thevsane Zora have Little quality control during the reprocuccicn procsss. Eince there was no “formal” publicatica of the FF, users should de ayare that inperzections in the reproduction process may alter the degree of re tiariy between the variants contributing to sone possible confusion en the part of the user. TABLE 1 BREAXDONN OF VARIABLES FOR TOTAL POPULATION (118172) YALU. Es 4 Test Iteas 6 as 1.8 10 as 5.1 i 70 2 qt 7542 92.3 # practice Items 1 139 17 2 8033 98.3 f Alternatives 4 206 25 on 7842 96.0 a 124 15 Sample Nornal* 7929 97.0 bnornal 243 3.0 Social Class Upper 125. Hiddle* 7913 Lower ist ‘Sex Male 4302 Fenale 3870 Age in Yeats (26 nos.) 3 6 a 4 18 3 s* 802 3.8 or 2101 13 § 7 913 12 at 1702 20.8 oe 1332 16.3 10* 749 9.2 qt 478 5.3 12° aig 5.1 13 60 7 14 60 7 15 70 3 16-25 24 3.4 26-38 $6 7 36-45 19 2 46-55 6 a 56-65 s a 66-75 21 3 76-85 18 2 ‘Groups selected to comprise the norming population. 18 MEF Norns oy sor vG"h 69's ong aro ore 907k as zoe wk wk a6 ora o9"bt 9G'Lt ort it ‘WLOL zz oz ze 9s eur 19F shy Wz = ssMOuil try sorb z's 59'S L's oro got r9'9 ja “IVLOL UOd SNOLLVIATG GUVANVLS ONY SHV So's tye sete ze 99 st'yt aneat osiz - bet syiweid (9pGZ=N) XaS UNV HOV AT van gut ost ot ur vor one act aut zl sa'y way zw'y PLS wg 29°9 uve tS°L as 00'8 1 6ce ik are so°st ou'st es'st SUEZ let sami ait Let sot zst viz siz sz 16 w et zu ats or | 19 MEF Noras ore rs'ur oz or" aerate sor vee 89°21 at a sz6 est or ace az | ost 90°01 19st sei u . SS‘Or wut ere 90°Ot one oir ct ezLt BOT or wL6 Wor £95 se'G Biror ur onor Gyo sz 6 os's ss"st gor ree arr vor berg sorat? wz g sok olzt lov org Pour one yore Is'zr siz “ vere 6tar tbr one 8rot GT o0'9 0L°6 She 9 sty sre wz ore pre zi 6s's ose +6 s G xX Wi we xX W ws W "VIOL Swed (ovezen) xas ONY dov an JONUUVT NVM Od SNOLAVIATG qIVONVELS ONY SVAN Sivas 20 Tee oUt ist azz x95 sor 19F soy lez =I ere os*~ wid (9rez=N) xas QNY SUOWWa JO SNOILVIANHOD sor ost ou us rot ove act at TINE, sami ait tyr got wt viz SIZ spe. +6 zl at i or av TABLE 8 (CONT.) PERCENTILE RANK BY RAW SCORE FOR TOTAL ERRORS (FEMALES) Tae AGE (YRS) f = & nm Seo tect 40 tH aura 49 16 47 nn 43 20 10 6 42 is 12 39 19 s “37 14 os” 38 18 5 iM 13 u cot vv ; ' 8 28 12 s 4 27 “10 23 ou 5 21 16 , 4 2 3 19 10 : 3 17 1s 13 16 nu 13 a 10 12 4 2 2 wn o w nu 6 10 ue Oe NU RUA TH. uae ee iw MEF Norns TABLE § MEDIANS FOR ERRORS (ME) AND LATENCY (ML) BY AGE AGE (YRS) x ae se 5 221 21,96 6.42 6 443 17.82 8.21 7 461 14,02 10,22 8 408 12.25 4n.2t 9 563 8.13 “33.71 10 227 6.68 13.67 u 297 7.26 117s 12 Re 7.661" 10.68 MFF Norns TABLE 6 HEDIANS FOR ERRORS (ME) AND LATENCY (ML) BY AGE SUBGROUPS NOUSROUNUH BnwooD > eomusthret sue Bubb aoa 23 MFF Norns TABLE © (CONT.J MEDIANS FOR ERRORS (ME) LATENCY (ML) BY AGE SUBGROUPS al YRS) NEM Meannaoon ue 4 aonunnnoocan Baddsadaeds Raunsvoarins 10 1 ee Z 24 o a TABLE 6 (CONT:) MEDIANS FOR ERRORS ote) AND. LATENCY (ML) BY AGE SuacRours AGE_(YRS) x Me ML 32 29 34 14.8 19 5.8 15.4 3 11.0 16.4 51 6.1 9.0 12 13.0 13 . 24 28 10.4 9 4.9 1.8 60 9.5 10.8 12 1s 11.3 7 1S ad 25 TABLE 7 ‘ NTILE RANK BY RAW SCORE FOR TOTAL ERRORS (MALES PERCE: Rs Q 2 AGE_(YRS) s 6 ft Bog 10 cay 12 99 3938 32 28 24 2 22 98 38 36 30 27 22 16 20 21 87 370 19 18 36 StH 3S) 29 28 21 1s 85 32 24 200~C 18 94 34° 26 23 19 14 16 7 33 n 15 16 $2 29 2s 18 13 4 92 23 24 2 30 3300227 7 3 89 23 22 13 88 26 22 87 32 21 20 16 as at 84 28 20 19 12 83 30 24 4s iL 82 . 12 81 qt 80 29 23 18 79 14 7 : v7 10 78 : 22 19 10 7 28 2 a 72 16 a : 18 6 . 10 6s 26 1s 9 67 21 i . Bae 66 17 6 20 63 28 14 9 62 10 61 19 16 so 24 56 43 35 “4s 9 8 7 6 5 4 3 2 1 oO TABLE 7 (CONT.) PERCENTILE RANK BY RAW SCORE FOR TOTAL ERRORS (MALES) AGE YRS) Sarena zh a | “ 17 21 il 13 7 6 10 20 12 1s 6 Ed ig ll 14 18 era: s lo i 3 12 ? 4 16 9 il 8 6 3 14 10 s Be 2 13 6 9 12 5 4 8 - 1 4 1 10 7 3 3 2 9 6 1 2 1 PERCENTILE PERCENTILE RANK BY RAW SCORE FOR TOTAL ERRORS 30 35 33 32 a 30 29 28 27 26 25 24 23 22 2 26 25 23 2 21 20 19 18 7 20 19 18 7 16 1s 14 13 TABLE 8. AGE (YRS) 20 19 18 16 1s 14 13 12 12 Ww 10 v7 16 13 12 aL 10 12 lu: 10 (FEMALES) 12 Ww 10 TARLE 9 : PERCENTILE RANK BY RAV SCORE FOR MEAN LATENCY (MALES) AGE_(YRS) : & ua 7 30 u a 39 19.7 34.3 43,8 52.3 62.8 98 18.4 31.4 38.8 48.7 45.9 37 16.6 25.7 30.8 39.8 41.6 96 25.0 29.1 Mea 40.2 7 95 18.7 20.2 28.9 36.1 40.1 94 13.8. 19.4 27.2 33.7 36.3 93 13.4 18.8 26.2 RS Mid 92 17.0 24,3 M.7 34. 31 13.3 16.8 24.0 30.7 32. 90 12.8 16.0 23.5 30.3 32 89 15.4 23.3 2.400 33 12.8 15.3 23.0 23.100 31 “87 W160 15.1 22.2 27.8 30. 86 WS 14.6 21.8 27.0 30.0 as W140 14,30 214 26.7 26.5 84 10.6 14.1 19.8 26.4 25.9 83 10.5 13.7 18.8 26.2 15,7 82 10.4 13.7 18.2 25.2 81 10.30 13.1 17.7 25.0 28.1 80 8.9 13.0 17.0 18.0 24.2 2204 73 9.7 11.9 16.8 17.7 23.6 78 9.4 11.7 15.9. 170 234 21.6 7 9.3. 11.6 14.5 °° 16.3 23.30 2115 16 8.8 112 144 154 22B 21d 78 Wt 14,3 22.6 20.7 4 86 10.9 14.1 24.9 (2201 73 8.5 10.8 14.0 14.7 “21.9 20.6 72 8.2 106 135.4 116 21.2 20.5 n “8.0 M.D 20.9 20.3 70 B 20.0 20,0 14.0 13.2 69 7.9 10.5 . +7 19.5 19.0 13.0 68 10.4 12,7 13.6 19.3 13.8 12.9 6 7.6 10.3 125 13,3 190 18.0 13.3 12.6 66 7.5 10.2 33.200 186 0174 ZL 65 7309.9 12.5 13.1 184 17.0 1300 as 64 7.0 9.8 12.0 12.9 17.9 16.9. 12:70 12.1 63 800128 178 161 12.6 62 9.5 11.7 12.6 371 15,9 12.0 61 9.4 1S 12.2 16.8 15.8 12.4 - 60 6.9 93 114 119 16.3 12.5 59 6.7 9.0 11.2 118 15.6 15.5 1200 11.8 58 8.7 lt 7 154 1.6 s7 8.6 11.0 11.5 15.2 15.2 114 $6 65 85 108 14 149 15:1 49 Ld ss (207 13 148 15.0 18 ko Sa 6.4 10S ILL 14.60 eons. 53 6.2 8.3 10.3 11.0 14,2 14.70 15104 52 8.2 10.1 10.8 45 11.2 10.8 31 6.1 9.9 10.7 14.0 14.0 41.0 © 10.7 50 8.1 10.6 13.8 9613.9 10.8 10.6 MEF Norns 30 TABLE 9 (CONT.) 3 Nk BY RAW SCORE FoR MEAN LATENCY (MALES) ret ive wins nano 224 tae St 33 SSdang Sis 6 Ree S208 al 3282 usene ane ane ee Eto er tere al segs SERSCK oes Soe Gas SR oRS See 8] dd 88298 2 52 noneun BeEeNeN © a we Sloe Sidad 4 3% RSSSSs gddeda ¢ 3 22 ao NT2en Cwemnoamn eae one eee g Riddadd easdgeges RStedis Gogg eee Boat Ine awonen SIRS a Men en 0 BSER Seco SI S2E2I2 Nee ge “o See e Eg TLIBNeLCnee nena, en PROG Se OeEEs Cy ; RM whence na S3S982252 etende ies, RN Ree R RSS wo VOCS SGT wuvinian feteettttet 22 4Mxane om now At st ean oi eS “a aa adds Vw add vie Ss vad S88 eure RRaaaaR een eesnnnenenen Ths SUS 28 8antaanuasasnaesauaeannes Nt 4 nen “ne ve 6 sui SoG 22° aNnenwow COS Mune 1 Ree S1s40% © SNS: SSuuGS 4 Rao gItenses PROS SSS SMCS exe was AMT HONG VANS id cy vey Qo nea a Weinie od MEF Norns TABLE 10 ea PERCENTILE RANK BY RAW SCORE FOR MEAN LATENCY (FEMALES) PERCENTILE AGE (YRS). & ae az 99 19.2 32.4 29.0 98 16.0 29.0 28.0 97 14,0 26.8 76.4 96 12.5 25.3 26.0 95 12.3 24.6 25,3 94 11.8 23.4 25.5 93 1S 20.5 9 214 23.8 92 20.4 21.0 22.0 gl id 18.9 20.1 20.0 90 11.0 18,5 19.9 $9 10.9 18.0 19.1 20.6 88 16.5 18.4 20.5 87 10,8 16,0 18.3 20.4 86 10.5 18.7 w7 18.7 88 10.4 15.6 17.4 18.5 84 10,2 15.0 16.8 18.0 83 10.0 14.7 16,7 qt 32 9.7 14.5 16.4 17,0 81 3.5 13.6 16.2 16.9 80 13.30 16.1 16.4 79 9.3 13.2 15.9 16.3 78 o.L 12.7 15.4 26.0 7 8.9 12.6 15.1 * 76 12.3 14.8 15.6 7 Ee 12.2 14,7 15.5 74 + B48 12.1 14.6 15. 73 8.6 12.0 14,3 15.1 72 B.S 11.8 14,2 15.) 1 8.407 13.9 14.8 70 8.1 11.6 13.8 4h 69 11s 13.5 14.0 63 8.0 114 13.3 13.9 67 11,0 12.8 13.8 66 10.9 12.6 : : 13.7 65 10.8 12.4 14.8 169 197 1 64 ng 12.200 14.0 16.8 18.7) 14.73. 63 7.8 10.7 12:1 16.6 18.8 14.6 13.0 62 7.6 10.5 12.0 13.8 16.4 18.3 14.3 61 7s 10.4 11.9 13.6 16.0 14.1 60 10.3 118 13.5 15.9 17.8 12.8 59 7.4 10.0 107 154 158.7 16:7 13.9125 58 7.30 97 13.2 15.4 16.5 13.7 12.4 37 7.20 96 16 130 15:3 - 16.4 13.6 36 71 9S 1d 15.8 13.8 12.0. 55 9.4 11.2 15.4 15.1 0 11.8 34 7.0 89 10 129 15.2 13.0 011.7 33 8.7 10.9 1208 14.9 11.2 582 6.9 84 10.8 12.6 14.40 12.8 11.0 $1 6.8 10.7 11.9 13.7 1207 50 66 8.3 105 i1.8 13.3 1206 32 ‘TABLE 10 (CONT.) oy SEMIS genase en a ec eaciy anuve ey sone gp cl SS88sd daddaa Sad og ¢ Rxken SddSS Wa waned 3 3 Z. s eat Sgeqenve aewnnwoneane o on SESS Shasdddcs Keener sssds @ Gen “AT HEAT CAanNe Nanenoanoaunensnane nae Adi $8 Seegsddda ddiddcdrenecauunacde NOSSBBissantaneamensn maenwacevennnen BAASASSeeraaanaiwian Senn Knee OMS AGE_(YRS) ANK BY RAW SCORE FOR MEAN LA’ Boo G 3 $S8828e sesp tas ge 2 sen ceenen deqenun 2 Sg skkdeed “ess 6 Ss SS G GRad GLII2 Bsgsns a g eB oo $3 32 2 82 yensat 22 25 ene ane on geen a 6 S$ 6s S$ vw Meiwdds oS SOS SIE 33S 32 gee3 PERCENTILE Mee horss 33 FIGURE CAPTIONS FIGURE 1: TOTAL ERRORS FOR MALES AND FEMALES BY AGE FIGURE 2: MEAN LATENCY TO FIRST RESPONSE FOR MALES AND FEMALES BY AGE FIGURE 3: MEDIAN ERROR AND LATENCY SCORE BY AGE SUBGROUP FIGURE 4: INTEGRATED SHOWING ROTATED AXES FOR INPULSIVITY AND EFFICIENCY TOTAL ERRORS FOR MALES AND FEMALES BY AGE 23 20. — Moles v7 —— Females TOTAL ERRORS 3 6 7 & 3 10 1 12 AGE FIGURE 1 MEAN LATENCY TO FIRST RESPONSE FOR MALES AND FEMALES BY AGE MEAN LATENCY 5 z 6 — Femoles — Maies 3 6 7 @ o 10 iy 2 AGE FIGURE 2 MEDIAN ERRORS 5 5 MEDIAN ERROR AND LATENCY : aa SCORES BY AGE SUBGROUP . cove: @ 7oTaL, Poe Avion MEDALS ® SUDGROUP MEDIANS SuION aah ss teere neste si Heri Herie a a6 saa ost MEDIAN LATENCY IN SECONDS : Shee” DIRECTIONS FOR MATCHING FAMILIAR FIGURES Adolescent/adult Set : "I am going to show you a picture of a familiar item and then some pictures that look like it. You will have to point to the picture on this bottom page (point) that is just tice." | E shows practice items and S$ selects the correct item. |"Now we are going to do.some that are a bit harder. You will see a picture on ton and eight pictures on the bottom. Find the one that is just like the one on top and point to it." | ° E will record latency to the first response to the half- second, total number of errors for each item and the order in q like the one on this top page (point). Let's do some for prac q which the errors are made. If S is correct, E will indicate this to him. I€ wrong FE will say, \'No, that is not the xight one. Find the one that is just like this one (point).)' Continue to code responses (not tines) until S makes a maximum of eight errors or gets the item correct. If incorrect E will show the right answer. The test should be set up in a notebook. It is necessary to have a stand to place the book on so that both the stimulus time. The two pages should be practically at right angles ito one another. Note: It is desirable to insert the pages in clear plastic which helps to keep the pages clean. q and the alternatives are clearly visible to the § at the same re eae YO MATCHING FAMILIAR FIGURES Answer Sheet Adolescent/Adult Set Practice 1. boat..., 2. cowboy.. Item Le dogs... e048 2. rose +66 3. soldier...2 4. graph.....7 5. baby... 25.4 6. lamo......8 dress.....1 lion......5 9. glasses...7 lO. plane.....4 Ll. leaf......2 12. bed... 5

Você também pode gostar