The Effect of Vocal Modeling on Pitch-Matching Accuracy of Elementary Schoolchildren Georgia A. Green Journal of Research in Music Education, Vol. 38, No. 3. (Autumn, 1990), pp. 225 231. Stable URL: hitp://links,jstor.org/siei?sici=0022-4294% 28 199023%2938%3A3%3C225%3ATEOVMO%3E2.0,CO%3B2-K Journal of Research in Music Education is eurrently published by MENC: The National Association for Music Education, ‘Your use of the ISTOR archive indicates your acceptance of JSTOR’s Terms and Conditions of Use, available at hhup:/www.jstor org/about/terms.html. JSTOR’s Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at hup:/ www jstor.org/journals/mene hum Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the sereen or printed page of such transmission, JSTOR is an independent not-for-profit organization dedicated to creating and preserving a digital archive of scholarly journals. For more information regarding JSTOR, please contact support @ jstor.org. hupulwww jstor.org/ ‘Tue Sep 19 17:14:49 2006JRME, VOLUME 38, NUMBER 3, PAGES 225-281 295 The pp of hs study was onsite the efecto dat ole, al mal, on hid eal mating on be pchaching acu of hen rade 1 trngh ‘he sche etd 382s nly on esearch eh iinet cal mda Mal puch ima testing enon were he sme on each ist seasim be mb ferme me med we Re eed sl modeling had on fe nese pch aching eceray. Thee wee mare coe ‘pons tote chide jllsed the font ond te mle model, reece Into pts were mar etn fl [rte fmal nd wale dts ad wor i ‘hap for th Ol mode Fst ad nds deuce hige ercanag ofc! ones oll re mda The het pretenses ws sg OS fintrade ajc Georgia A. Green, Baylor University, Waco, Texas The Effect of Vocal Modeling on Pitch-Matching Accuracy of Elementary Schoolchildren “Traditionally, singing has been considered a major component in the elemen- tary micrantha ma concep can Be ugh In'the primary grades. Because singing i such an integral part of a chi’ early musical education, the fact that many children never lear o find and use thet Singing voices and therefore are never able to sing tunefully should be of Concern to music educators. The most natural way of teaching singing to shildren iby modeling. The teacher sings and expect that the cil will imatate the pitch presented by the model voie- Since the range, quality, and timbre of adult voles, especially the adult male voice, differ from those of children, it aight be expected that these factors would contribute tothe dificly that some children have in matching the pitch presented by the teacher Some generalizations based on experience have been made concerning vocal sodeling: Gould (1968) states that before being asked to reproduce a pater, the student should "hear the pattern sung with a treble voce (or played om the plano ithe teacher isa man) by the teacher or other pupils" Greenburg (1979) fecommends that men sing in'a"lght, tenor voice" with children, and Green burg and MacGregor (1978) state that "mae teachers, whose voices are an octave lower than those ofthe stents, can reinforce their own voices with instruments This article i based on the author's doctoral dissertation, tiled “The Effect of Vocal Modeling on Ptch-Matching Accuracy of Children in Grades One through Six," accepted by Louistana State University, Baton Rouge, in May 1987. For copies ofthis aril, contact Georgia A. Green, School of Music, Baylor University, Waco, TX 76798.226 GREEN so that students can match their voices with another tone of the same pitch." Nye and Nye (1985) suggest that occasionally a child may be confused in tying to match the octavelower voice of the male model. In this case, Nye and Nye recommend that a male teacher explain that he “cannot sing as high as the children” and that he “play the song on an instrument that gives proper pitch, Sing faleto, or have another child sng the song.” Also to be considered the timbre difference between male and female voices. Roe (1983) suggested that some children may be more inclined to attempt to reproduce the timbre of the model voice, rather than the pitch, which would affeet the resulting pitch ‘accuracy of the child ‘Adding to the conclusions produced as a result of these experiences, more controlled empirical research has provided some insight into the variables affecting pitch accuracy. Smith (1963) suggests that the use of abartone singing ‘oice in his study and the resulting octave transposition may have presented a problem in his investigation ofthe effect of group vocal training on the singing ability of nursery schoolchildren. The results of Hermanson’s (1971/1972) study ‘on the effect of timbre on vocal pitch acuity of kindergarten and third-grade children showed significant differences among responses to osillator, woman’ Yoice, piano, and 8-year-old girl's voice. When singing simultaneously with different prerecorded pitch timbres, the subjects produced the most accurate Pitches when singing with a woman's voice and the least accurate when singing With a piano. Results also showed that subjects did not sing significantly better With another chie’s voice than they did with other timbres. Petzold’ research (1966; 1969) involved similar investigation and showed that responses 10 a female voice were more accurate than were responses to patterns played on the flute ‘The question of timbre as a complicating factor in pitch-matching accuracy was also investigated in an experiment that deat withthe question of which presentation mode was easiest forthe child to reproduce voealy (Clegg, 1966) ‘This study involved 796 first, second-, and third-grade children. Results showed thatthe female voice was the easiest forthe child to match, foliowed by piano, Autoharp, and pitch pipe. The male voice, ecorder, utophone, and song bells ‘were more dificult for subjects to match. “The results ofa recent study involving 5- and 6-year-ld children support the {dea that itch-matching ability in young children may be complicated by octave transfer necessitated by male vocal modeling (Sims, Moore, & Kuhn, 1982) Children were tested on vocal accuracy as a result of male and female modeling, tonal pattern length, and age. Results indicated that there were significantly more correct responses tothe female model than tothe male model. Also, of the approximately 48% of pitches that were sung incorrectly, 64% were one or more half steps fat, 23% were one or more half steps sharp, and 13% were omited. Tn contrast the previous mentioned findings, resus of study involving first grade children showed that male voeal modeling did not necessarily present 4 pitch-maiching problem to children of this age-group (Small & McCachern, 1983), Students were pretested using female and male models, then assigned to ‘one of three treatment groups. One group practiced with a female model, one with a male model, and one group served asthe control group with no practice ‘After several practic periods, all subjects were tested again using both models. Results indicated that, prior to practice, subjects who could match pitch with one model could generally match pitch with the other. Practice with voeal models did not produce significant diferences in accuracy among the three groups. ‘contlicting results of these studies, as well a the added implications of the possible effect of subject age and. the analysis of the incorrect responses, provided the impetus forthe study. The purpose of this sudy was to determine the effect of female, male, and child vocal modeling on the pitch-matchingJRME 227 accuracy of children in Grades 1 through 6 and to examine and analyze the incorrect responses to each of the three models. METHOD Subjects for this study were 282 elementary students enrolled at a major southern university laboratory school during the 1986-1987 school year. This number included the entire population of Grades 1 through 6 with the exception of two students who were not present at the beginning of the study and were eliminated from further participation. Audiotapes of three different ‘model voices—an adult female soprano, an adult male tenor, and a 9-year-old child's unchanged voice—were prepared for use in the study. ‘The descending minor third was chosen as the stimulus interval because ofits inclusion in the Kodaly curriculum as the interval most easily sung by young. children (Daniel, 1978). Pitches constituting the interval were G and E above ‘middle C, because these notes are both found in the comfortable singing range of preschool and first-grade children (Nye & Nye, 1985). Model stimuli were sung on the neutral syllable "la" to decrease changes in vocal timbre (Roe, 1983). ‘The descending minor third was sung and recorded by each of the three model voices at a tempo of 60 beats per minute, with each pitch receiving one beat. Each subject in the study was tested individually for pitch-matching accuracy ‘on three separate occasions, each time responding to a different vocal model. ‘The interval and pitches sung were identical for each of the three tests, the only difference being the model voice. The order of the tests was identical for all subjects: female model, male model, and child model. The three tests were given, at 7-day intervals to correct for the confounding variable of tonal memory. When each subject entered the testing room, she or he was told to stand behind the microphone and sing the notes heard on the tape. Each subject was given ‘only one chance to respond to the model voice; however, all subjects were familiar with echo singing and the sol-mi interval because of their Kodaly-based ‘music instruction. Headphones were not used for playing the stimulus voice so as not to interfere with the subjec’s ability to hear his or her own voice. At the conclusion of the three tests, all audiotapes containing. subjects’ responses to each of the three models were heard and evaluated. A Korg Auto Chromatic Tuner, Model No. AT-12, was used to evaluate the accuracy of each taped response. Each individual pitch of the interval was evaluated, resulting in six pitches per subject, two for each model. Deviation (sharp or flat) from the ‘model pitch stimulus was measured in increments of 100 cents. A pitch response ‘was considered accurate if the deviation from the model pitch was less than 100 cents. The two pitches for each model were then averaged into one deviation, score, resulting in three data points per subject, one for each model. Each data, point represented either a correct, flat, or sharp response to each model for each subject. Three trained evaluators independently analyzed 40% of the total responses with the Korg tuner. Reliability (agreements divided by agreements plus disagreements) was 93. RESULTS, The number of subjects singing sharp, flat, and correct responses to each of the three models was counted (see Table 1). The results were analyzed using the cchissquare and Cochran Q statistics. All possible combinations of model, child's sex, and grade level were compared. Data demonstrated that the three vocal models (adult female, adult male, and child) had significant effect on whether subject's responses were correct. Of the three models, the highest number of correct responses was to the child model,228 GREEN ‘Table 1 Frequencies of Caret and Incorrect Responses to Each Model by Grade and Sex (N = 282), Boys Gis Model Grade 0 0+ T uw Bo 6 03 2 7 2 5 8 3 8 6 6 2 1 Female 4 1 BR 2 2 no 2 5 2 Bo 0 1 6 7 om 4 5 3 1 woo wooo 2 noo O 9 7 0 3 Ib 180 6 a 2 Male 4 6 8 4 3 2 5 8 5 od nr) 6 noo 8 Hoo 4 1 wo W383 4m 5 2 ro 6 4 9 3 3 5 wT o 7 B coed 4 1 wo 4 rou 3 5 3 8 8 bow 2 6 2 3 6 Loa 6 Nae ~ = fat; O= correc + = sharp. followed by the female mode, with the lowest number of correct responses being fiven to the male model (Q'= 21.88, = 2, p-=/05, N = 282). Further analysis Indicated that there were more correct than incorrect responses to the female model, with the incorrect responses being more often fa than sharp (x2, 282) 152.84, p <.08). The same findings were indicated for the male model, with ‘more correct an incorrect responses and more fla than sharp responses [x2 282) = 100.84, p <_05]. However, wile there were more correct than incorrect responses to the child model, the incorret responses were more often sharp than flat (2, 282) = 147.9, p <8) ‘When comparisons were made of the subject's responses based on the sex of the subjec, data indicated that for the female model, both boys and girls sang more correct than incorrect responses and more lat than sharp responses [2 150, boys) = 7081, p = 05; y°@, 143, gir) = 81.67, p <5) For the male model, boys sing more fat respontes than correct o sharp responses (x12, 139) 5.3 p © 05) while girs sang more correct than fat responses and more fat than sharp responses for the same model [2, 148) = 6387, p <-05]. For the ‘hild mode, boys sang more corret than incorrect responses and more sharp than fat responses [y42, 199) = 86.74, p < 05). Dat indkated the same Findings for Fesponses tothe child model sung by girs [x¥2, 143) = 987. p= 0) “The responses of the subjects in each grade level were compared for each sodel separately (see Table 2) For fis, second, and sixthegrade subject, the highest percentage of correct responses was othe child model It should alo be noted that only for firs grade subjects were the total responses othe mae and female models less than 508 correc, except Tor sixth-grade responses (othe tale model, which were also les than 80% correct. For third, fourth and ith rade subjects, the highest percentage of correct responses was tothe female ‘model Av each grade lve, the highest percentage of fat responses was tothe tale model, and the highest percentage of sharp responses was tothe child modelJRME 229 Table 2 Percentage of Corect and Incorrect Response in Back Grade to Each Model Grade Model T 2 3 a 3 é 0 Ho G2 090 767 909 are Female = 535315510. + B47 BSS 8S 0 303 G13. 618 600-600 a2 Male - 5728S 4B 300367 + 35. 036 oo 83 0 607 731 864 700 700 748 child = 350 96 673 + M3 7s 84888167203 correc; ~ = fat; + = sharp, DISCUSSION Results from this study indicated that adult female, adult male, and child vocal ‘modeling had an effec on the pitch-matching ability of students in Grades 1 through 6. There were more total correct responses tothe child mode, with the least number of correct responses elicited by the male model. Results also indicated that incorrect responses were likely to be Rat for the female and male ‘models and sharp forthe child model. Boys inthe study tended to sing flat more often with the male model than with the female or child model. Subjects in Grades | and 6 sang the highest percentages of incorrect responses, with first grade subjects singing the highest percentage of flat responses for all three Iodels. For the female model only, there was an increase in accuracy from Grades I to 5, with a decrease in accuracy in Grade 6. For the male and child ‘models, there was not consistent increase or decrease in accuracy among the six grade levels Subjects in this study were most eaily able to match the pitch presented by the child model, Although no existing research literature was found that specifically recommends it, the vast majority of textbook recordings and other children’s recordings use the child voice as singing model. A possible explanation for this finding is that a child can most easily match a pitch presented by a voice similar in timbre to his or her own voice. Conversely, a child may have difficulty ‘matching a pitch presented by a voice or instrument with a timbre different from his or her own voice even when the stimulus pitch is presented inthe octave and range of the child's voice. This theory is supported by results of studies involving the eflects of timbre differences on piteh-matching accuracy (Clegg, 1960: Pewold, 1966, 1969; Roe, 1983). Contradiciory results were obtained by Her- manson (1971/1972) kindergarten and third-grade children were not able to sing simultaneously more accurately with a child's voice than with other timbres. Th the study presented in this article, the male model produced the largest number of incorrect responses. This would seem to suggest dhat the octave transfer necessitated by the octave-lower range ofthe adult male voice presents a problem to young children in accurately matching a pitch stimulus. This Supports the research of Sims, Moore, and Kuhn (1982), in which 5- and 6-yea ‘ld children were more easily able to match a pitch presented by a female voice rather than amale voice. Contrasting results were found in a study involving. first-grade children; for these subjects, the male vocal model did not present & pitch-matehing problem (Small & McCachern, 1983).280 GREEN ‘The deviation of incorrect responses in the study described in this article was more often fiat for the female and male models and sharp for the child model. This supports findings of a similar study in which 5- and 6-year-old subjects sang ’gher percentage of flat responses than sharp to both female and male models, (Gims, Moore, & Kuhn, 1982). This may be due in part to timbre differences in the model voices, oF to pitch-discrimination factors. Madsen, Edmonson, and Madsen (1969) found that when younger subjects perceived pitch incorrectly, the incorrect perception of deviation was most often sharp. This presents an interesting relationship in that more subjects in the study described inthis article ‘who sang incorrectly sang flat than sang sharp. In this study, the only significant effect for deviation and sex was found in responses by boys to the male model. Boys sang consistently fat with the male model, suggesting a pitch-matching problem unique for boys with this model, ‘There was nota significant difference between flat and sharp responses by girls to the male model. This indicates that there may be a sociological issue involved in young boys’ matching of pitches presented by a man’s voice. Results indicated that of six grade levels involved inthe study, the highest centages of incorrect responses were sung by subjects in Grades 1 and 6. The fact that first-grade subjects demonstrated the most difficulty in matching pitch forall three models support the theory that pitch-matching ability isa function of maturation (Petzold, 1966; Smith, 1963). The decrease in accuracy among sixth-grade subjects presents intriguing questions as to possible causes. Serafine (1988) also found a decrement in accuracy among sixth-grade subjects ina series of experiments that compared adult knowledge of musical processes to that of children at ages 5, 6, 8, 10, and 11. This “sixth-grade slump,” as referred to by Serafine, was found to occur in those tasks that required analytical skill in the auditory domain, such as pitch-discrimination ability; and in those tasks that required sensitivity to conventional or cultural musical rules. In a 1973 study involving sixth-grade subjects, Smith found that in-tune singing ability was significandy related to pitch-discrimination ability and tonal memory. Further investigation with sixth-grade subjects is needed to confirm this decrease in accuracy among children of this age-group as well as to determine which variables affect the musical abilities of sixth-grade children, For subjects in Grades 2 through 5, responses to all three models were more often correct than incorrect. The absence of a significant difference among these grade levels in the deviation of the incorrect responses to all three models indicates that, inthis study, maturation did not affect the direction of deviation Of the subjects incorrect response. Results indicated that only in the case of the female model did accuracy gradually increase with grade level, suggesting maturation asa factor affecting pitch-matching accuracy with this model. It might also be noted that all subjects Involved in this study had received regular music instruction with a female music specialist since enrollment in kindergarten, indicating the effect of training on increased accuracy with the female model. For the male and child models, the absence ofa consistent increase or decreas in percentages of accurate responses among the grade levels suggests that, for these two models, maturation may not hhave been a significant factor in pitch-matching accuracy. Further research is needed to investigate other variables that affect pitch-matching ability in ‘children, such as pitch-discrimination ability and maturation. In the area of vocal ‘modeling, timbre studies need to be conducted to investigate the effect of voices Of differing timbres on children's pitch-matching ability. Specific research is needed in the area of male vocal modeling in order to establish guidelines for imale teachers in the elementary music classroom. Studies involving male normal range as opposed to male falsetto and different ranges of male voices are indicated. With regard to vocal modeling, studies involving the effects ofJRME 231 training with a particular model voice are needed to investigate the effectiveness of certain types of model voices in the elementary music classroom. REFERENCES Clegg, B. (1966). A comparative study of primary grade children’s abily to match tones blshed master’s thesis, Brigham Young University, Provo, UT. Daniel, KS. (1978). Kody approach, method book one. Champaign, iL: Mark Foster. Gould, O. (1968). Finding’ and learning to wse the singing vaice (Contract OEB-10-016), ‘Washington, DC: US. Office of Education, Greenburg, M: (1979). Your children need music. Englewood Cliffs, N}: Prentice-Hall Greenburg, M., & MacGregor, B. (1972). Music handbook forth elementary schoal. West Nyack, NY: Parker. Hermanson, LW. (1972). An investigation of the effects of timbre on simultaneous vocal pitch acuity of young children (Doctoral Dissertation, Teachers College, Columbia University, New York, 1971), Diseraion Abstracts Intemational, 32, 3558 ‘Madsen, C. K., Edmonson IIL, F. A, & Madsen, Jr, C. H. (1968). Modulated frequency ‘discrimination relationship to age and musical raining. Journal of the Acoustical Society of America, $6, 1468-1472. ye, RE, & Nye, V. T. (1985). Musi nthe elementary schol (5th ed). Englewood Cliff, NI: Prentice-Hall. Petzold, RG. (1966). Auditory perception of musical sounds by children inthe first si grades (Cooperative Research Project No. 1051). Madison, WI: University of Wisconsin. Pewold, RG. (1969). Auditory perception by children. Journal of Research im Music Education, 17, 82-87. Roe, PF. (1983). Choral music education. Englewood (Cliffs, NJ: Prentice-Hall, Serafine, M,L. (1988). Music as cognition, New York: Columbia Universiy Pres Sims, W. L. Moore, R.S.,& Kuhn, TL. (1982). Effects of female and male vocal stimuli, tonal pattern length, and age on vocal ptch-matching abies of young children from England and the United States [Special Issue] Psychalogy of Music, 104-108, ‘Small AR. & McCachern, FL (1983) The effet of male and female voc modeling on pitch-matching accuracy of frst grade children. Joural of Research in Music Education, 31, 227-283, ‘Smith, R. B. (1963). The effect of group vocal training on the singing ability of nursery ‘school children, Journal of Research in Music Education, 11, 187-141 Smith, R. 8. (1973). Factors related to children’s in‘tune singing. abilities (Doctoral, ‘dissertation, West Virginia University, Morgantown). Disertation Abstracts Inernational, 34, THT February 3, 1989