. .2 2. .2 2 . . . . . 2 ...? .2 q o , . ..w . . . _ . . I 9 . .2 . . 2.22:9 . :... 22...... .o .... . . . . 2v. . . . I 2.... . . 2 . 2 O .. _ . ‘20. . .. . .. 2 . . I . I 2 . . _ . . 2I .. o. .. ....I... ”.... gtfuw.21...__ 2..... :... ......2... I. 2 .......:u... . . .22... ..H: 2... ....... .. 2 . 2 2 2 . . . ... nn2. .2 . 2. .... . . ,. _. . .. . 2 . .2. . 2 ... . . . . .... 5 .b0 ...)L 1...... 2.1. PO 9 - .\a 2.. o u‘ . . _. cl ‘0 ._ I. ... .2 o o 0 I .2 2 . J 2 .I H .. 292 ._ . a '. 2 2 .. o n 2 . 2 I 0 - . 2 2 o . 22.2 52 s. _. 2 ... ... 2 237.05.232.32 . f 2 o. . . , 2 9.. .. ... 2 .2 . I .2. 2 . 222 . I O2 . . 2 2.2.. :qu9 2n. . . ...2..22...2.. . 9.02 I 22.2 2 I . .2. .2 .I. . 2 . .2 2.2 2 2 . a 2 .. 0.2 I 2 .I 2 2 .2. ... 222 o. '4'. 00.22....2.... O. ... 22k.“ '2. .c .. .o. 2‘... .Q t . .2 I .2 .. .. Io. . .22. O o 2 I2 0‘ . I . I 2 I 2.. .II 2 22.. o ... 2 “I... . 2 .22. 2 . 2 I . . . . . . l. . . . . . . 2 . I I o . . 2. 21." 2.. ... . v; 2:. ..a... 5.2. 2.... 22.23272... .26. u. . .. .22. . .. . 22 v 2 ._ 2 . 2O. . 2 ... .. .o ... .... 2 2 .12 . 2 . .. ... 2 _ . . .. .. . ... . r : .2....2... .. . ....t..- ...,2 2 . .2. . ... . . . . I 2 . 2 I . 2 . . . . ... . 2 2.. 2 2 2 . . .2 . 2 2.2. ... .. 22. 2. t: 0 2.2.. 5.. ‘0 . .. 2. 2.6.... .. .. 2 .. 2.2. 2 . 2 .. 2. .. Q . . 2 . 2 2 2 .I ._ . o 2 0.2 2 .... 2 ..ls...«.r-' ' 2.“ .‘9..... ‘2 2.....2.N...Ii v. 2!. . . .n 2 I2 I O .. . J 2 '2' I2 ..Os. . O2 2 . I . 9.. . . I . 2 .0 . . .... .I. .. 2 .2 2 Q . .2 o . 2 2| 2 . 2 2 2 I o 2 I. . 2 I . 2 216232....»m \’ “Rx-.2 .29 2. .2 ..I. ....v... ...I... .... 2 22.10. I... I, . v2 ...... .I I. . I I 22;. . 2 2 _ ... lo 2. ...2 2. 2.1 .. ... .I 2 2. 2. 2. . .2! . 2.2 . 2.. .2.. . .. . I2... . . ... .22.. .2.2 .. . I . ‘ I . 2 2 . . . . 2 .. . . . _ . 2 2 2..... I . .2 .. 2 .2. . 2 . x. .2 O A o . 2 2 ..I 2‘ .2aI u . 2. 2 . 2 .:2.......2.....2:.2 -..; .. . .... ...2.. . .. .. .. .. .. . .. .2. . . . . -. ..., . 2 . . . . . .. . _ u. . .5 . 4...... hurry... .... 2 ... . 2 2 2 o 2 . 2 I 2 I o . . . 22.0 v. .0 ’.o.~. . u 2, 2. 2 OJ 2 c 22‘ n . 2 .. Q . 9 q 2 . .2 2. . I 2 . 2 . o 2 .202 T I o 2 I .2. I . . 2 2 . .2 O. 2 C . . I . 2 2 2 2” . ... ._ u 3.. O 90.4r0 I.“ ...? . r 2 . n. 2 .2 I Q . 2 u ov 2 Q 2 2 o 0 2 . I C I l . 2 .9 I . . a 2 0 . O 2. .0 q”- “nf \ I! o 4. cl s. ..Q: ... 2 2 o 2 o 2.. .... l2 2 . . 2 .2 us 2r . .1 v. 2 2 O .. 90 2 2.226 I. I 2 ‘. . d I .9. I. 2 . i 20. 2 . O 22 .4 . v If. .2. 22 . .... 2 .2 . O‘Ro- v.o..”2 ‘ L 4” $.fiot‘an1 s. ..I. . . o .. 2 2 . .2 2 .. . v v 2 2 2 2. r . 2 ..I I .I .2 2. ... 2 2 v 2 o 2 . 2 a . 2 I a . .2 .. _. 2 2 .. ‘32:. 2%. 0.2.. 0 2 2. 2. 2 O .. .2. . ... o I... ... g. 2 . 2 2 2l22 2- I. . . . .2 o . . .. I. .... . . 2 I. 2 p'. ‘ . 4 o. s i. ... ..... .. . : . 2 .. .. .... . . . .. .. .. n; . .. ...2. 2 ...: .. ......_. 1:..2. ... 2.25:... 2.2.... ... ..h2c ' 2 2 .. .. 2.. .2‘. 2 2 2 I. I“. . .2 V . 2 2 .. .2'. o. .2 . 2. . ¢ I 2 . o O .0. .22 . .. . ‘. III . J . , v Q . . 2 0 ..I I . .2., 2 A 2.. . o . o 2 . Us. a!“ 2 ‘2'2 . 2 2. o _ 0 2.... 2.. 2 ... . . . .. I . 2 .022 2 C I . 2.2 .2 . o. I 2 2 I 2.2 . . I. .. o :2. ..2 2. *3 2 . . 2. 2 . 2 . .2 C. . . o . .. . . . .I .. . 2 2 . . 2 . . 2 2 . 2 2 . . .2 . .. 2 ... 0.8-: _ . ‘2. I . s. o . . I I .2 ' O. I Q . 2 . , 22.22 I I .20. .. . I... I 2'. .. a I .‘32 o .222. .2. ... I o .\ . 2 . 2 .... 2 . 3.2 \.. . c 2 2 .21 o 02 . ..II 2“ .2 3 ... 2 2 I. . I.“ . 02 2I O I! c . ' '4: I I. I... ‘2 m. 4.. I W2 2 2 I o 2 I0 2 I I. . o oo . 2. 2 I . .W2. 2. .I if.) ....v I‘Ihg‘u ’0’». LA I, ’0.” . I . . .v .0 o .. .... C. II. 22 , $.22 .. O ... . 2 .. o n. I: 2 , 2. . 2 2 . O . . 2 .‘f . . 2 2 2 . . . ... .. .. . .O.4 ...I..r .. . .o . o . . 2 o. "I. ...-A . . 0.92 ..." a. . I. o 2. _ . 2.2 . 2.. II I . o . I I. .2 o I '2 g. 0. to. . 2.. v .2 ‘I 0| .0 u 2 2|- . .. ‘ . o a...- .0 24 . .2 2. . . 2. 2 I v r I I l . I 9‘. O . o. ....I . '2 2 2 ”Ju‘lt..JoV2. 22”. ... '2 Pic .3) § 1": 4H2 . 2 . a I 2 I y 2 . 2 .2.. “a... s d —.oI...OI.-.~ I _ 2 I u 2 . o 2 . I 31¢. 2. v 2 '22 I... 2. 2. o 9 . 2v 2 2 I .I 2... 2 .— O 2 2... o 2< § . 2 . 2 n 2 o .0 Q I o 2 2 o I .. I II 2 O m 2 2 o . . ..v: 2 ”#021 .o .2 I '02. .... 2 22 T... r. a ... ‘1“.«1. 2 . thng.‘ 2... I2. I. 2. .2 2 9 v o I 2 . 2... 2 2.02 2 . I .22 . 2 I . I 2 I I... 2 I II. c 2 2 I o 2 2 ‘00 .... D. 2 . .I. 222. r’? 0 ‘2 III- ‘I on. urol 22. 2.2 a... I‘.s. 2 . 2 2 . . 2 2. 22 ‘ I .2 2 2. 2 . 22 . II. . 2. .0. 2 I II I . 22 . . o o. . I. II 2 I . . .. c...!l52..go .3701 .0. VI ..A .2.. I . 2 .0 2 2. 22 .L 2 . 2 ...... o 22 ... . 2 I. 0 I . 2. 2 . .. ... .I ...}. . ... . I2. . I .2 2 02:722. . .0 ‘2020202 . .2 I .2. . » ...... n. 2 . . 22 ..22 . . . I! . ..2.... 2. . .02. . 2 . l h 2 . .. a. . 2 2 .2 . . ... . 2 I .. 2 .2.... .2 2 2.. .I. 2 I 2.2 . 2 2 r. a... u . _. . . .... .. . . . 2. . . ....4w2vufolau2v... osufi.w.2n.~§1n..22 "M2204. . c. 2 I 2.: o 2 ..I 2 . 2 2. 2 9 I 2 22a 2 I 2 2 . I. 2 I . 2I 2 202 2 I2 . 2 o 22 2 I 2 . 2 a Q I . I 2 2 4 . 22...I .12 IA. . . 22:.alk20k" I 2 . . o 2 . o 2 .2 .. <2 I 2 2' . I 02- I. 2. . V 2 2 . . 2-.z' . 2 .0 2‘ . . 2 2 . ..2. ..2H . .2 .v 0 22. 2 . 1.2 .q- 2 f . I..I. ..2 . I... 2 2n ... 2 o I 2 2 . . .. 2 . 2 o 2 r. . .. 2 . a . . o I. III . I 2 2.2 . I a 2 2 2 .4 12.53.02.353... “.2 o . .1904. ’ o ‘ o o . ..o q 2 2 2 o 2. 2 I' . .c ‘ o o. 2 4 I 2 24 o o 2 o. I . .22 2 O o 2 o 2 . A l 2 I . . I 2. I 2O 2 2‘. .22 II o . I .2. I. 2 2 . .42 o: 2’0. 2 ‘EVMP’OQQ‘ 2’. . 1 "O I I ..2 \II . . I: I 2 .2.. . 2. 2 . . I 2| . 2 2 I . 22. 2.. . O 20 o- . .. . . . . 0.. 2 . I . . .2....2 . 2 2. 2 2 ...I. I . . . ... . ... . .0. o v. 2 22 ... . .. . . . .. .. . 2 .. 2.. I . ... 2 I2 2 .. ..I 2 .. 2 2.1.3. «0... 2 9.62:...“ 2 I53 738.6. no 2 . Jv I2. 4! a2 . . 2. n . 2..... 2 2. .2 .2 22.. I . 02. 2.... . I. II .2.. .. :22"? . 2 . I .2. 22 2 .. . 2 . .. . I. . . ... ... I. . . 2 2 n . . 2‘24‘2 ...JIPHIIO. 22.32.?»FQ“ 2”?!“«13 .2... I. .. a . .2 . . .. 2. 2 -2 2 — 2 2 2 2. I . . 2 . .0. I I. I . . 2.2 v 2 .I . o 5... . .2 . _ ... 0.2 2 2 2 . 23.2 at. a . . 4-, ‘2‘ 2 I. . . . . . 2 . 0 2 2 2 .. . 2 2 .. 2 I 2 I .2 .... . . 2 o ... . . .2). I I I2 I . 22 II I- . I . . . I 2. .. 2. H ...-5.): C. . .n. . I. 22 . ... 2 :22; 1‘. . 2 92 I. 2 . . ... 2..! 3. I ...-I t . If. 2 .2 . 22 2 2 .I .20 . . ..III . .2 . . v . 0. 9I. O. 2202.“. . 2.. 2. .. .. 2 2 2.. 2 2.” LI . 2.1."... W...» I 1.33.03 .....Doiargnuon2' . . . . . . . I .I ... I . . 2 0" 2 2 4 ,— . _ . 2 , . ... .- . .. . . . _. .. 2.. .. .... .2.. :.. .2. . .. r .. .. :3. .2... 2.2.2.5 52...... It 0 . ‘- . 4. . 2 2 II . .Iu . I .52 I . 3‘; ..J? 20 . . . 22.20-105 v C ...I 2, .. . I ... .22 . . .o .2 2.2 ..I .30 .22 22.22 . . . . 2 . 2 .. 2o . .2..' 2 .. I . . 29 or a. ....I2 ... . b 2 2.... _. . .... .2... r... . . 2 . u. . 2 I I 2 .{‘..22 (I. .I , I .2 .- .2222 . I ......O .' o a .5 2.22 .. .2. . . ..2....: 2.. .422... 2221 I. I . .2IOI.¢. I2... .2. .. t I .2 2 I h I .2.. 1 2 . a .. ‘ . , .o"..‘ [32.0 . I.2..2uc v.22. . o2. . .I. 3 V I .2 . . 2 2 .. II. 2. 5.52.222. 1.1.5.. 2. .2 2 .2 o a. o 2 .9. 22.712 2 .1. ol' ... .2 2... . . r\. 722‘ n. . 2 2 I I.22 2 . .2923} .....u. .. . .2I.....I2. ...Ousr ......v... ...... 202.222 0 .. ' ‘2...‘ ...I‘. I. .2 2 2 . O 2 Id 2..... 2 72‘ n. 2. 0‘0 I 2 02.2.2HW...1’0422§‘0.A 3a,... I t .2? .2.. 2 2 I f. . t . 2..”.- JOoI I 2 c .2 2A2. ’.IJDT ‘2... .n' \C I . .2I 2' . . . 40...! IIK’.H2'22. 0' p 0. 2 . II. I 2 2 v . . . . 2 . . . 2 . 2. . . . . 2 . . . . 2, o .06. ”a?! 2.22...) f"? .v'fi”: ‘ C p . 2 . . 2 . 2 2 . .. . . . . 2 . . , . . .. . .2 .222’2321..II’ 2 .hfi... . 2 _ 2. . . . . 2 . . . . . .. H .. 2 , . . . . . . . .. .32 522...: .22. o . 2 . . .... ... . . . 2 .. . .. ... 2. . 2 . .2 . . . . . . . . . . . .. . 2 . 2 I....I.....2I2O .. . . . . . 2 . .. I .. . 2 . 2 . .... . . . . . . . . . .. . . 2 . I . ...3? . 22 2292...! 2. .2 2 2 . . . 2 2 . . . 2.. . . . . fl . 2 I . . I a .2 . . . .22. .2 .2 2 ......H 0 i2. . . . .. . .3. I s I. I. 2 I 2 I . o 21,. it.’ I I. _ . 2 . .2 5.9.22.1. . 12:122.! ‘ c. l. . 2. 2 .02 I2 . 2 2 . 2 2.2.9.7... ....I.-.2.2.221~L2H2IW2 . 2 . 2 . I. 2 2 . 2 .10.. . . . . . .. . . . . . . ,. .23-.........3»......2..... m V 2 I . 2. I . 2 . 2 .. v .2. 2 2 .2.. . -..-2.14.2922! ....2 _ 0.2.. 20.“. .32V‘ N . 2 . 2 .2 I 2 I . I 2 . a . 2.4 .2. 2 . . . . . . . 033.0 . 'w.r2r2II. “’21 2 324'. 22.2 o I 9 ‘2 . I I? 2 2 O . . . . . . . . 2 2 . . . .....di 22.2'2‘0I2')122 ‘1 2 .‘JsI’v2n . _ . _ . .2 . -2 .. ... - .. . . . . . _ . . . 2.22.29.52.22. .. .. .2._22...2.........I . O . . . . . , . 2 2 _ . .. . 2. . d .. 2 2. . . . .. .. . . 2. . . . 2. . .. _ . 22.. .12..t.bnh.uhp. IILIIRLVMIRI I . 2 . . . 2 .2. . .. . . 2 2 2 I . . . . . . . . . . . .t. . II 2 .. . . 2 2 2r 2 . I: . . . . 2 . . . . . . . . . . . . . _ :iirhutrrl; 2.29.222 ” .2 . 2 2 2 .2- H... . . . . . 2 : . . . . . . i . . . I . ». . $ . 2 n.9.’vu2.Jw.thIII:.-... 20.— I...” u.u.22I2M.2‘wwI2u.’I2‘!.h26w2u.2 ' . In . I '2 I ‘O ‘n I. :2 0" .. . 1 Nu. . 2 2.22.2...qu IIYQh‘l-ov‘ . 2‘ 2.4.2.1“... ’lbmunh .35 .... 2..}..2..22r.2....... .2 . I. .21 figs-0.2.0 I .0 022‘ 0220.... v“! 7.I2I 2 2'1.” . . . _ 2 . .. 3.....J2....2.}........2.r...II.... .2.. . .. . . . . 2. .. ......‘O . . . 2. . . . . . . . 2. . . . 226.2"v25'hl2‘1fufi22 20.“...(«“ I 2 . 2.2.0.222 ‘Iv 2 (If 0.22... l... 22 JuniuIl . 2 .I . 202.22 22.. . . . 9D .2 . . . .2 . ... o .22 2 ”.222... u 222. ...tvfl. I... 2.}I,I..MI.725 - .02 . . .2 II. 2" 4.0.2 ‘11.. 2"...! .002. . 2 .... It 2 v4.2. I.a’ I 2 c 2 ‘10 I 2 i O .1.- ' u I I. . ; v ' I OI I . ‘ 222 2 f . .0 .'. . I. C I.“ 2.. 2'. o 'I C . . I I O ‘ ‘2 .‘ :. r C I " 5 :6 “:9. : .mufl‘!‘ ”W'P‘vnu"""" O O 1 ’3 O ‘. I I .0 W .51: .5 0's Q I ,. .I . . .II I $92.2 .l 02.12.222rflltlxln III 20. .. . .7022l2’ 10 V oo' [2. ”.0“... I . I. 006- O ‘2 . . O .222 .... 2..“....!2I 2. 2 .... t 2 22'2‘0’ v '9': I.'2.. 22..-: O . 5“ p.22- 2'23?! I ”cum. ‘HIWu ' | O I C 4 .. 'e - I .. 2; I I 0.2! I 2 I 2 I _ .I .. 2. O ‘ o _2 . - . . . . . 2 . 2 2 . I 2 - I . . . _ . 2 2 . . 2 2 \ 2 o .I _. 2 . . . . "I . 2 2 I . .u 2 . 2 2 2 2 2 I . _o . . 2 .. . 2 .u 2 .I I . . 2 . 2 2 I . I . o 0.1 . 1| I I .m. . I . 2 . 2 2 2 o 20 2 I 2. . 2- 2 I . .2 u I 2 _ t 20 I2 . w 2 _ 2 I I I 2 . o I .5 2 .22 I. 2 O I 2 I I 2 ' _ . O I . 2 2 2 ’_ 2 2 2 2 . 2. 2 I a 2. a. . . . _ 22 . 2 I 2 . .2 I 0‘ I2 2 2 2 2 . O. . 2 a 2 I .2 . . .0 . I . 2+ .. 2 . _ .2 2 O . . . 2 2 2 .2 . I . 2 2 2 v 2 2 2 2 2 . O .Q . . . 2 22 o . I O . . . . I 2 2 2. . 2 2 2 . 2 I _ ‘ Q I n. 22 . I I 2 I I 2 2 c n u '1 I 1. 9 2 _I.. . I . O . . . 2 . 2 . . . . I .2 _' o . 1. I . ”22 I . I 2 . I 2 . . 2 2 2 2 2 Q. . v2 . 2 u 02 2 . ...I . _ O 630‘" I O ’ I I I .. . 2. 2 2 . . .I.2 . .. I. 2 . .. 2 . . 2.2!. ... . ... . ‘2. 2 2. VII O. I. 2.. .\.2. .....833A2 2 $22.2. l 2.5 .‘4 .72IP1JI 3&1... 2..... 29.22.3A42.222I.I ‘51. ... 2.. I’M 2 .fi. I I 2 . I . I .. o . v. 2 2.. . .2 . . 2 . av «0—9 3 II. ...-.22.) .. 2.0.2.2“. wflewfi *454'hraanw3 .229l5. b‘b.;. . .1. a .Ialr.%.. 2 . 2%. ..w.» J 2 .Qoxgg I 0 ”21¢- 20 2 In 2 ..IP. 1 OJ? .222 2. .2... .....22 .‘u. 2.1. .2.. I. .3 I 2‘»: v“; L2 3 . "In... ...s.‘vfi.r.~t.~n2 ...... 442,0. .3 22¢. .13?”— .2~... .W...u .. v. 2.. 2.4... 2.322. .. ..h‘ .2.? term?“ a . 2“? 22 ....gc. ‘ f #3322. . 2. t2... .0 .2222... 2 I 44 .2 V’I..Oo2_2.q.J..H.o..g . .‘3..., 25......”2'2 Au'nlw 9.. ..2_.2 ...h‘ . .4 .....a .00 . 2” Aka-‘3.” 2w ”.2..r.Am..bJ2.a.\ 22omhi..U 82.2.“.37KFL...“ 9. .I fir. 22% kfl.‘ § “v2’s..£. «Cm 9—er m£~ 2’ ... . . 'Bcfi.i ABSTRACT ACQUISITION OF SYNTAX VIA TELEVISION BY TWO AND ONE-HALF TO-THREE AND ONE-HALF YEAR OLDS BY Diane Roscile Moholy It is generally agreed that the quality of the child's linguistic environment is the most important external factor affecting the rate of language development. It has not been determined what constitutes a quality environment. There is little doubt that television viewing has a poten— tial for enhancing the quality of a child's linguistic environment. A young child acquires the basic rudiments of syntax by either repeating verbatim what is uttered by a model (modeling), or by hearing and repeating expansion made on the child's utterance by a model (expansion). A child tends to reduce adult utterances, resulting in the process called reduction. Whereas an adult tends to expand a child's utter- ance, resulting in the process of expansion. It works much like a telegraphic message:- a parent or model expands on the child's utterance by adding functors to the child's content words of nouns, verbs and adjectives. Diane Roscile Moholy Since media sources account for a very large percentage of some children's total language exposure, the positive effects are well-worth establishing. Sesame Street, a pro- duction of Children's Television Workshop of New York has adopted effective verbal communication as one of its primary objectives for young viewers. Hence, this program was uti- lized in this study. Expansions seem ideally suited for assisting the ac- quisition of grammar, while modeling appears to limit the child's language exposure. It has been argued that what is important for the child is not a particular kind of exposure but simply ample exposure to well-formed speech. This ex- periment was designed to isolate these two processes of language development and to compare their effects. Three hypotheses were tested: 1. Children viewing the television segments containing expanded material will show a greater language growth than will children receiving neither treat- ment. 2. Children viewing the television segments containing modeled material will show a greater language ' growth than will children receiving neither treat- ment. 3. Of the two treatments, children viewing the tele- vision segments containing expanded material will demonstrate greater language growth than will children viewing the modeled material. The subjects were twenty-seven two and one-half to three and one-half year old children enrolled at the Diane Roscile Moholy Married Student Activities Unit day-care center in East Lansing, Michigan. Three randomly assigned groups, each containing nine subjects, were formed: two experimental and one control. One group (expansion) was exposed to 30 minutes per day of televised expanded material; another group (modeling) was exposed to 30 minutes per day of televised modeled material. A third group (control) re- ceived no treatment. These respective situations existed for two weeks, without exposure on the weekends. All twenty-seven subjects were pretested and post- tested on three dependent measures of language development: mean length of utterance, ability to count, and raw score on the Peabody Picture Vocabulary Test. All three indices have been used as a measure of language development in previous research, with the mean length of utterance being the most widely acclaimed measure of language growth. In evaluating the validity of the dependent measures it was assumed (1) that a developmental characteristic must either stay at the same level or increase; (2) that a valid measure of a developmental characteristic will yield scores conforming to that pattern; and (3) that converse fluctua- tion in the scores indicates the presence of unaccountable superficial influences. Means on the pretest and posttest scores were computed. The results indicated a trend Diane Roscile Moholy insomuch that all the means, excluding two, increased from the pretest to the posttest. For statistical analysis, gain scores were computed for each group between the pretest and the posttest scores. A t-test was computed between experimental and control groups on these gain scores. A t-test was also computed between the two experimental groups to determine which treatment had the greater effect. A non-parametric test was used to determine the number of subjects who demonr strated an increase in one direction by chance alone. Two major findings emerge from these analyses. First, there was a low correlation among indices between the ex- perimental and control groups. Four possible explanations for this finding are discussed: (1) low internal reliabil- ity among the assistants administering the tests; (2) chil- dren's language skills and motor skills alternate in developing, causing plateaus or dormant periods of learning in one skill while the other is developing; (3) control group scores were initially higher, indicating the possible presence of a level of exposure to language in their non- treatment environment, and (4) limited amount of actual treatment time. The full explanation is probably some combi- nation of all of the above possible explanations. The second major finding is that expansion was the more effective treatment for increasing a child's mean length of Diane Roscile Moholy utterance count. While this was the only dependent measure that was significantly proven, the other two measures indi- cated expansion was slightly superior. On the basis of these results, none of the research hypotheses are confirmed. Four alternative hypotheses are confirmed, giving positive enforcement towards the three research hypotheses. These results tend to support the statement that the quality of a child's early linguistic environment is the most important external factor affecting the rate of lan- guage development. Providing the child with an enriched linguistic environment of ample well-formed sentences paired with meanings the child can understand will most adequately facilitate a young child's acquisition of syntax. ACQUISITION OF SYNTAX VIA TELEVISION BY TWO AND ONE-HALF TO THREE AND ONE-HALF YEAR OLDS BY Diane Roscile Moholy A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Television and Radio College of Communication Arts 1974 Accepted by the faculty of the Department of Television and Radio, College of Communication Arts, Michigan State University, in partial fulfillment of the requirements for the Master of Arts degree. Directo of Thesis ii ACKNOWLEDGEMENTS With sincere appreciation I wish to thank my thesis advisor Dr. John Abel for his continual guidance and assistances. I am also indebted to Andy Gilpin for his knowledge and support concerning the linguistic develop- mental aspects of this study. I would like to thank the nine undergraduate assis- tants who gave their time and comments to this research project: Pam Drelles, Audrey Garon, Michelle Holda, Debbie Medel, Saunga Miller, Ruby Peete, Jeanette Scroggins, Susan Swirsley and especially, Pam Mitchell. I am also most grateful to Randy VanDalsen for his competent assistance in recording and editing the video tape segments. A thank you is also extended to the faculty and staff at the Married Student Activities Unit day-care center for their cooperation and use of their facilities. To Mom and Daddy, I wish to express my gratitude for their omnipresent personal support and inspiration. iii TABLE OF CONTENTS CHAPTER 1. THE PROBLEM AND RELATED RESEARCH. . . . . The Need . . . . . . . . . . . . . . . The Problem. . . . . . . . . . . . . . The Hypotheses . . . . . . . . . Review of Related Research . . . . Acquisition of syntax by Young children Learning from television and Sesame Street 0 O O O O O O O O O O O O 2 Q PROCEDURE 0 O O O I O O O O O O O O O O 0 Subjects . . . . . . . . . . . . . Treatment. . . . . . . . . . . . . Selection of treatment material Pretesting procedure. . . . . . Schedule of treatment . . . . . Posttesting procedure . . . . . Measures of Language Development . Mean length of utterance. . . . Ability to count. . . . . . . . Peabody Picture Vbcabulary Test Statistical Analyses .“. . . . . . . a. . . . . . . . . 3. ANALYSIS OF RESULTS . . . . . . . . . . . Evaluation of Measures of Language Develop- ment. 0 O O O O O O O O O O O O O 0 Test of Hypotheses . . . . . . . . . . 4. SUMMARY AND DISCUSSION. . . . . . . . . . APPENDICES A. Research Policy . . . . . . . . . . . . . B. Rules for Calculating Mean Length of Utter- ance O I O C O O O O O O O O O O O O 0 iv Page mama-NH I-‘ 11 16 16 23 24 25 26 29 30 32 34 35 38 38 40 46 50 53 TABLE OF CONTENTS--Continued APPENDICES Page C. Rules of Administration . . . . . . . . . . . 55 D. Scoring Procedure for Peabody Picture Vocabu- lary TeSto O O O O O O O O O O I O O O O O 57 E. Form A of the Peabody Picture Vocabulary Test 58 F. Statement of Instructional Goals for Numbers Goals on Sesame Street . . . . . . . . . . 60 G. Guidelines for Letters to Parents from Researchers. . . . . . . . . . . . . . . . 62 H. Letter to Parents . . . . . . . . . . . . . . 63 I. Permission Slip . . . . . . . . . . . . . . . 64 BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . 65 LIST OF TABLES Demographic Breakdown of Groups. . . . . . . . Mean Scores and Mean Gain Scores for Measures of Language Development. . . . . . . . . . . . t-Test of Gain Scores Between Experimental and Control Groups . . . . . . . . . . . . . . . . Sign Test of Direction of Change in Individual Scores on Three Language Measures for Combined Expansion and Modeling Group Scores. . . . . . Sign Test of Direction of Change in Individual Scores on Three Language Measures for Control Group Scores . . . . . . . . . . . . . . . . . vi Page 23 39 43 45 45 CHAPTER 1 THE PROBLEM AND RELATED RESEARCH The Need Language is the basis of communication with others. It is a highly complex form of behavior entailing systems of sounds, grammar and meaning. The process by which a child changes from an inarticulate organism to an articulate organism continues to defie understanding. The phenomenon warrants in depth study and has only received limited and fairly unsophisticated research. The role of language in our society vehemently neces- sitates further study of this complex process of acquisition of syntax. Language and thinking are so closely related that experimenters must ordinarily use the child's language to study his thoughts. The ability to successfully communi- cate with others is a prerequisite for social acceptance. Television is continually taking over the role of educating, or some may call it "babysitting" today's chil- dren. This extremely influential medium has the vast potential of becoming a powerful educational tool for children. Establishing television's role within a young child's environment will certainly aid the multifaceted process of child develOpment. Studies combining the effects of linguistics and tele- vision are long overdue. Adequate television programming for young children has just begun to be developed. This success story has stirred researchers to seek more knowl- edge about two very omnipresent and yet clandestine phenomena: a child's acquisition of syntax and the role television can play in facilitating this process. This study incorporates both phenomena in an attempt to contribute to our limited understanding of a child's acquisition of language in a television assisted situation. The Problem As previous studies indicate, a child acquires lan- guage by either modeling a model's sentence (Cazden, 1964) or by expansions made on the child's utterance by a model (Slobin, 1965) or by an exposure to an environment utiliz- ing both expansion and modeling methods of language acquisi- tion. Expansions, which are constructed much the same way telegraphic messages are, seem ideally suited for assisting the acquisition of grammar. Modeling, which is contingent upon an original utterance in order for direct repetition to occur, appears to somewhat limit advancement of language acquisition. It can be argued that what is important for the child is not a particular kind of interaction but simply ample exposure to well-formed speech (Brown & Bellugi, 1964). This experiment was designed to separate the modeling and expansion methods of acquiring language skills and at the same time expose the children to well-formed speech. The latter design element of the experiment was achieved not through personal interaction as with all the other studies cited, but rather through an electronic medium-- television. Children can learn language fundamentals by viewing television (Ball & Bogatz, 1970). If this is the case, both well-formed sentences and an exposure to either one of the language acquisition processes should establish an ideal learning environment. "All the available evidence . . . supports the general prediction that the quality of a child's early linguistic environment is the most important external factor affecting the rate of language development" (Carroll, 1961, p. 340). The experiment reported here was designed to test the effects of a particular process of language acquisition, either modeling or expansion, in a consistent, ideal lin- guistic environment, viewing television. The Hypotheses While all the studies providing background data deal with a personal two-way communication interaction process, their general conclusions can be applied to a non-personal one-way communication process. Belief that exposure alone is sufficient receives some support from the occasional stories of children who remained silent for a long period of time and then started to speak in mature patterns (Ervin, 1964, p. 163). A pure example of this could be that of a child who heard speech only over television or radio. Exposure to well-formed sentence structure over television is only one process of creating an ideal linguistic learn- ing environment for a child. An argument that more than exposure is necessary could be made on the evidence that children from the lowest social class groups, who are known to be retarded in language development, watch as much as or more television than do culturally more advantaged children (e.g., Wortis, 1963). This qualification of well-formed sentences is matched with simultaneously paired meanings the child can under- stand (Brown & Bellugi, 1964). There was no discrimination of television viewing for children in Wortis' study. Presently children's programming is developing both in qual- ity and quantity which helps to create an ideal linguistic environment by restricting children's viewing to meanings they can understand. This study employed what is considered one of the finest children's programs available today--Sesame Street. A one-way communication process of viewing television, where the child is not an active participant, is designed to measure only one source of verbal stimulation available to a child during this process of language acquisition. An increased exposure to language would assist the child by providing more models for imitation and more instances of the concepts to be learned, while expansions would addi- tionally provide more feedback. On the question of the relative assistance to the child provided by models and expansions, prior research predicts that modeling would be more helpful. This prediction is embodied in the three research hypotheses which this re- search tests: 1. Children viewing the television segments contain- ing expanded material will show a greater language growth than will children receiving neither treat- ment. 2. Children viewing the television segments contain- ing modeled material will show a greater language growth than will children receiving neither treat- ment. 3. Of the two treatments, children viewing the tele- vision segments containing expanded material will demonstrate greater language growth than will children viewing the modeled material. Review of Related Research Acquisition of syntax by young children In order for a child to comprehend and use a particular language, it is necessary that it be expanded or modeled (Odem, Liebert, & Hill, 1968). Reduction can be defined in terms where a child shortens an adult's utterance much the same way an adult shortens a message when sending a telegram: high-information or content words are retained and the func- tion words (functors) are omitted. When a child imitates an adult utterance he reduces its length but maintains the word order. When an adult adds to a child's utterance, such responses are called "expansions". Brown and Bellugi (1964) suggested that they might provide optimal data for the acquisition of grammar. Child: John dinner Adult: John is having dinner. Child: Throw Daddy Adult: Throw the ball to Daddy. Just as the child preserves the word order in reducing the adult utterance, the adult preserves the word order in expanding the child's utterance. A parent frequently at- tempts to verify his interpretation of the child's utterance when he responds by expanding. Brown and Bellugi precisely define "expansions" as a process in which the mother or the model does not exactly reproduce the model sentence of the child, but instead adds something to it or expands it. What is added is a functor, e.g., the inflection for third-person on the verb, the very form the child omits when he imitates a model. According to Brown and Bellugi, expansion is a kind of com- munication check between the mother (model) and the child: "Is this what you mean?" Since the expansion must come from the mother, it is important to find out how she comes to expand on a child's utterance. Consider the utterance "Eve lunch." So far as gramr mar is concerned this utterance could be appropriately expanded in any of a number of ways: "Eve is having lunch;" "Eve had lunch;" "Eve will have lunch;" "Eve's lunch;" etc. On the occasion when Eve produced the utterance, however, one expansion seemed more appro- priate than any other. It was then the noon hour, Eve was sitting at the table with a plate of food before her, and her spoon and fingers were busy. In the cir- cumstances "Eve lunch" had to mean "Eve is having lunch." A little later when the plate had been“ stacked in the sink and Eve was getting down from her chair the utterance "Eve lunch" would have suggested the expansion "Eve has had her lunch." Most expansions are not only responsive to the child's words but also to the circumstances attending their utterance. (p. 142) Brown and Bellugi were not implying that a child learned grammar by storing the expanded versions of his telegraphic utterances, since he could not in this way learn more than the finite set of sentences he had at some time attempted to produce. Rather, they recognized that expan- sions were only raw data for the child's mental processes and that syntactic knowledge was a system of rules somehow derived from those data. They contended that the data pro- vided by expansions might be particularly relevant and seemed to be delivered with ideal timing. "Well-formed sentences which are simultaneously paired with meanings the child can understand" affected his language development (Brown & Bellugi, 1964). If this pairing is important, it seems likely that conversations in which the child is a participant would be more likely to provide it. Based on these assumptions, relevancy and timing of expan- sions are important. Whereas Brown and Bellugi's study demonstrated that expansions were not necessary for learning either grammar or a construction of reality, Slobin (1965) states that a. child can only make so many decisions per sentence and by the process of expansion, this ability can be slightly stretched. According to Slobin, a child has three choices when speaking: a. "simply repeating his original utterance without picking up anything the adult has added b. "say something even shorter than what he originally said c. “add something to his original utterance, something picked up from expansion". In his experiments with two children, Slobin found that they utilized (c) most of the time. Also from observ- ing these children, Slobin concluded that as a child grows older, he imitates less and that the adult expands less. This gives reference to the possibility of a critical age for expansion. According to Slobin, this critical age would be around 36 months. Another test on this critical age group concerning acquisition of grammar found that modeling alone was of considerably greater value than expanding alone (Cazden, 1965). Her study hypothesized that "children whose language is expanded or who merely hear more language spoken will show a greater language growth than do children who receive neither treatment" (expansion or modeling). Also hypothe- sized was that "of the two treatments, expanding will be more effective." Cazden and two trained tutors worked daily for twelve weeks with twelve lower-class Negro children: four in the modeling group, four in the expansion group and four in the control group. The children's age varied between 28 and 38 months, but all within this critical age range. The expansion group received 40 minutes per day of in- tensive and deliberate expansions; the modeling group re- ceived 30 minutes per day of exposure to an equal number of lO well-formed sentences that were deliberately not expansions. The control group received no treatment. On the analysis of variance with covariance control, the adjusted means were higher for the modeling group. This data indicates that while both expansion and modeling forms of dialogue facilitate language development, modeling alone is of considerably greater value than expanding alone. The order of grammatical improvement scores for the three groups is (1) modeling, (2) expansion, and (3) control. The only explanation I can offer is that as the con- centration of expansions goes up--in this case far above that naturally occurring in parent-child conver- sation--the richness of the verbal stimulation goes down. Expansions are by definition contingent on the child's speech in content as well as in timing. To the extent that they are pure expansions just filling in the child's telegraphic utterance to make a complete sentence, they will have less variety of vocabulary and grammatical patterns than the adult's non-expanding speech normally contains. The suggestions that the richness-improverishment dimension may be critical thus gains some support. (Cazden, 1965, p. 91) The process of acquisition begins when the child is born and under normal maturation conditions is a dynamic continuous process. The language stimulation available to a child during this process can and clearly does vary in quantity and quality. It seems intuitively obvious that differences in quantity and quality should affect a child's language development. 11 Learning from television and Sesame Street One source of language stimulation for today's child is derived from electronic media. Within this array of electronic sources of language exposure, television viewing ranks the highest for young children. A study by Friedlander (1971) partly dealt with the role television plays in the percentage of total language exposure for two families. The study tabulated the sources of all utterances a child heard for the Smith and Jones families. While the study was not designed to test the direct effect varying amounts of television exposure could have on a 12 month infant, the families showed considerable difference in this category. Percentages of Systematically Sampled Utterances in the Homes of Two 12-Month Infants, Showing the Source of All Utterances and the Sources of Infant-directed Utterances Source of Utterance Smith Family Jones Family All Utterances Mother 22% 11% Father 18% 4% Baby 35% 15% Other (radio, TV, Guests) 25% 70% Infant-directed Utterances Mother 67% 73% Father 30% 24% Guests 3% 3% Source: Friedlander, 1971. 12 The upper section of the table shows that the families were radically different in terms of the infants' exposure to television (the principal element of the "Other" cate- gory). This study has only fragmentary information on which to estimate the language learning impact of television exposure, which occupied so much of the Jones' non-personal linguistic time. Mrs. Jones reported that at the age of 22 months, the little girl was learning to count without parental tuition, solely by watching Sesame Street. The mother's report was verified by a visit to the home by observers who concluded that it was apparent the child had learned at least this much speech by watching and listening to television. The presence of the mass media is pervasive in many American homes. Media sources account for very large percentages of some children's total language ex? posure. . . . It would hardly seem realistic to try to explain language acquisition without reference to these major sources of influence. Yet surprisingly, one hardly ever find mass media factors discussed in the contemporary professional literature on language development. (Friedlander, 1971, p. 270) Since Friedlander's statement, there have been a number of studies attempting to explain language development in reference to the children's television program entitled Sesame Street. Children's Television Workshop (CTW) of New York, has done perhaps "more than any other producer of instructional programs to use research creatively and to guide creativity 13 with research and in the course of so doing has produced the all-time hit series of children's television, Sesame Street" (Schramm, 1972, p. 105). The fact that educational television is an effective medium for teaching certain skills to very young children has been demonstrated in two research studies by Ball and Bogatz, 1970 and 1971, in conjunction with Educational Testing Service. Sesame Street verifies that television can be an effective medium for teaching 3-to-5-year-old children important simple facts and skills, such as recog- nizing and labeling letters and numerals, and more complex higher cognitive skills, such as classifying and sorting by a variety of criteria (Ball and Bogatz, 1970). The potential of educational television as a teaching medium is suggested by three primary findings, that . . . children who watched the most learned the most . . . . . . the skills that received the most time and attention on the program itself were, with rare ex- ceptions, the skills that were learned the best . . . . . . the program did not require formal adult super- vision in order for children to learn in the areas the program covers. . . . (Ball and Bogatz, 1970, pp. 3 & 4) The sample numbered 943 children covering areas of disadvantaged children from the inner city, advantaged children from suburban areas, children from rural areas and disadvantaged Spanish-speaking children. With this diverse 14 sample, Ball and Bogatz found that 3-year-old children gained the most; 5-year-old children gained the least. This suggests that 3-year-old children are able to learn many skills that have traditionally been introduced at later ages. More time on Sesame Street was devoted to letter- related skills than to any other single subject, and it was within the area of letter and number skills that the chil- dren's gains were the most prevalent. It was found that in addition to what was directly and deliberately taught on the program, some transfer of learning occurred (Ball and Bogatz, 1970). In a continuing evaluation of the educational potential of Sesame Street there was "empirical evidence that in view- ing Sesame Street, children were not passive and their learning included more than the accumulation of important basic skills through simple continuous associations (rote learning)" (Ball and Bogatz, 1971, p. 9). In support that Sesame Street can contribute to creat- ing an ideal linguistic environment, and is not intended to replace all other sources of verbal stimuli, "it (Sesame Street) was meant as an ingredient for the educational diet of the millions of 3-to-5-year-old children who do not have the opportunity of going to preschool" or receiving an ade- quate amount of verbal stimuli from other sources (Ball and 15 Bogatz, 1971, p. 9). Again, it should appear intuitively obvious that differences in quantity and quality should affect a child's language development. In defining television's potential for teaching young children, modeling of effective verbal communication has been a guiding principle behind many writing and production methods by the writers for Sesame Street (Lesser, 1972). Expansion of a concept allows the message to be com- municated to the child with an embellishment of language stimuli. Inherent to this process is the possible pitfall of including irrelevant and extraneous material, either visual or auditory, that confuses the child and causes him to loose interest. Sesame Street has been designed to screen out irrelevancies and reduce extraneous material when expanding on a concept (Lesser, 1972). CHAPTER 2 PROCEDURE. Briefly: Twenty-seven children, ranging in age from 31 to 41 months, were randomly assigned to one of three treatment groups. Nine assistants pretested the children and when the treatments concluded, posttested the children. One group (expansion) viewed segments of Sesame Street that were of an expansion format. Another group (modeled) viewed segments of Sesame Street that were of a modeling format. A third group (control) received no special treat- ment. Two 15 minute daily treatments extended over an 18 day period, with no treatment on weekends. These procedures will be discussed in detail under four headings: subjects, treatments, measure of language develop- ment, and statistical analysis. Subjects The subjects were twenty-seven children enrolled in a day-care center at Michigan State University, East Lansing, Michigan. The children were mixed racially, with Caucasian outnumbering any other (21 children). There were 14 boys 16 17 and 13 girls participating in this study. At the time treatment began they ranged in age from 31 to 41 months. The process of selecting the children was limited to two approaches: using existing day-care centers, or by soliciting children through Church organizations. The latter procedure was abandoned because it would have been extremely time consuming and had no guarantee of producing an adequate number of children for this study. Since it was clear from previous research (Brown and Bellugi, 1964) that the age range needed was approximately 30 to 40 months, the first task was to locate a day—care center that enrolled children under three years of age. The most logical place to look was on the campus of Michigan State University, since there are three day-care centers located on the campus. After a lengthy discussion with a representative of the Institute of Family and Child Studies, which is the department responsible for the three day-care centers, it was decided that only one of the centers would be appro- priate in terms of children, space and schedules for the study. That center was the Married Student Activities Unit Day-Care Center (MSAU). Upon proposal approval by the Department of Television and Radio, an "Application for Research" form was submitted for approval by the Institute of Family and Child Study and 18 by the Preschool Committee. Approval of the research by these two committees forwarded the application to the Assistant Director of Early Childhood Laboratories who in conjunction with the unit coordinators made final judgment as to the starting date. Research policy by the Institute requires that all researchers give only general results, not individual data. Compliance with research policy was upheld throughout this study (see Appendix A). The Married Student Activities Unit Day-Care Center is located within a married student housing complex, south- west of the main campus. The day-care center has a capacity of 110 children and was full to its capacity Spring term, 1974. The center has five units: one infant, one toddler, and three pre-school. Since this study only concerned the preschoolers, no references will be made to the other units. For a child to qualify for enrollment in any one of the three preschool units, he or she must be between the age of 2% and 5 years at the time of desired enrollment, be toilet- trained, and have a complete physical check-up. One of the child's parents must be enrolled in the University and carrying a minimum of six credits. There is no discrimina- tion of sex or race for admittance to the center. There is a breakdown within the units, with Unit 1 and Unit 3 being full-day attendance and Unit 2 being half-day attendance. In order to provide continuity for the children, 19 the day-care center requires the preschool child to be enrolled and attend a minimum of two-half days a week in the same class. The Married Student Activities Unit Day-Care Center is open from 7:30 a.m. until 5:30 p.m., Monday through Friday from the first day of registration to the last day of exams each term. The following information about the staff relates to the preschool units only. The preschool units maintain a staff ratio of one assistant for every five children (1 to 5). There are four A.M. head teachers, four P.M. head teachers, four assistant teachers, and students from the University balance out the l to 5 ratio. The center has one nurse, one cook and one cook assistant, one secretary and one unit coordinator. On a typical day for all three units, the children arrive at the center from 7:30 to 9:00 a.m. The block of time is designated as quiet activities and is tended to by assistants. Between 9:00 and 9:15, the children go into circle time which is lead by the head teacher. This could be construed as the instructional block of time. Activities in this block vary from making peanut butter from fresh peanuts to having a photographer come in and simplistically demonstrate his equipment. Toilet time follows this, although the children are allowed to go to the rest room at 20 any time. Between 9:30 and 9:50 the children have their morning snack and clean up afterwards. Weather permitting, the next thirty minutes are designated for outdoor play. The center has a fairly large backyard with two swing-sets, a slide, tricycles, a sandbox and large cement cylinders for the children to climb on. From 10:30 to 11:15 the children go inside and are involved in a free play period. Various activities are set up, three or four types, and the children are allowed to move freely from one activity to another. The free play period usually involves one activity for the motor skills, books and records, art and a science project. From 11:15 to 11:30 the children are divided into small groups and are instructed in developmental activities. These range from language, small motor skills, large motor skills, and math skills. This is followed by another 15 minute toilet routine. At this point, the k day morning children are usually picked up by a parent. The 8 day afternoon children arrive at 1:00 p.m. Eleven forty-five is lunch time, lasting between 20 and 30 minutes. Afterwards the children clean up and get ready for their nap. Nap time runs from 12:15 to 2:30, with another snack following. Between 3:00 and 3:30, the P.M. staff comes on the floor and the A.M. staff leaves. The rest of the afternoon, 21 until 5:30 or until the child is picked up by a parent, is filled with flexible activities. The children play out- side, weather permitting, or have free play inside. In selecting subjects from among the children at these three units no attempt was made to obtain demographic information on the families or make any observations of the child's home environment. Rather, it was assumed that these children were from fairly homogeneous family-liferstyles by the requirements of the center for enrollment. It was determined that, at a minimum, there should be nine subjects in each experimental and control group. This would require a minimum of 27 subjects for this study. The three units combined had a total of 40* children who were between 31 and 41 months old. Letters of introduction and explanation of the study were sent to all the parents whose children were in the designated age range.** A permission slip was enclosed requiring a parent's signature in order for the child to be tested and to participate in this study. Because the original design of this study included weekend treatments, which entailed the parent making arrange- ments for the child to get to the center, there were only * Two children were eliminated because they had non- English speaking parents. ** . Copy of letter and permission slip are included in Appendices G, H, and I. 22 16 affirmative responses by the parents. Departure from this original design had to be accepted in terms of rede- signing the.study not to include weekend treatments. Telephone calls were made to each of the original 40 par- ents, which resulted in exactly 30 affirmative responses. Ideally, the study was designed to include either all male subjects or all female subjects. This was more in deference to the generally held belief in differential verbal development than to any substantial amount of recent confirming evidence. McCarthy (1954) states that girls develop language skills faster than boys. But Templin (1957) found smaller sex differences in linguistic develop- ment, with boys ahead at the age of three, though not sig- nificantly so. Berko (1958) found no sex difference in the ability of 4 to 7 year old children to handle English morphology. There are too few cases with significant data to permit any generalization about sex difference and lan- guage development abilities. This study involved 14 boys and 13 girls. The subjects were randomly assigned to be pretested by one of the nine assistants and then randomly assigned to one of the three treatment groups: expansion, modeling, or control. As the study progressed, three children, one from each group, were dropped due to lack of cooperation on the part of the children. Thus, each cell or group contained 23 nine subjects, for a total of 27 subjects participating in the study. Once the assignments had been made, the treatments were begun. The parents of the children were not informed as to which group their child had been assigned. Although the parents were free to contact the investigator if they had any questions once the study had begun, none did. Table 1. Demographic Breakdown of Groups Mean Age Number of Number of Group (months) Females Males Expansion 39 5 4 Modeling 40 4 5 Control 38 4 5 Treatment Nine female undergraduates at Michigan State Univer- sity were given Psychology independent credit for their role as assistants in this project. These students had varied majors, so instruction and assigned readings in language development were necessary. Winter term, 1974 was spent instructing these students on how to obtain mean length of utterance counts, the difference between modeling and ex- pansion methods of acquiring language skills, and how to 24 administer and score the Peabody Picture Vocabulary Test. They were also requested to View Sesame Street periodically during the term. The assistants were also instructed in the process of video tape editing. After this instruction they were re- quired 1K) make two-15 minute tapes: one containing expan- sion material and one containing modeling material. Selection of treatment material. The assigned readings and practice in class demon- strated the difference between segments on Sesame Street that were of expansion or modeling format. Expansion seg- ments were defined in terms of taking one original concept and expanding on it within that short segment. An example of this would be the segment involving a little girl and ping-pong balls. She took the concept of "what would happen if" she dr0pped the ping-pong balls on her cat. Each time a result of this action was reported, the preceding video and audio of the story were repeated verbatim. Modeling segments were defined in terms of repeating one word, letter or number more than once. The most promi- nent example of the repetition was the segment concerning numbers: "1,2,3,4,5,6,7,8,9,10." The audio was constant with changing video or picture. Another example would be the repetition of a letter: "2 - z - Z- Z." Again, the video changed while the audio repeated. The assistants 25 were asked to make judgments on each segment and classify it as either expansion or modeling. If a segment did not fit into one category or the other, it was eliminated. Fourteen segments of Sesame Street were video taped directly off the local Public Broadcasting Station, Channel 23, WKAR-TV, East Lansing, Michigan. The programs were then edited into X inch monochrome video tape and categor- ized into one 15 minute block of modeling material and one 15 minute block of expansion material. The assistants were usually able to obtain this amount of material from one-one hour program of Sesame Street. Each of the nine assistants was required to turn in one 15 minute tape containing expansion material and one 15 minute tape containing model- ing material. Pretesting procedure. For pretesting, the assistants were randomly assigned three subjects, with three assistants testing four subjects. Pretesting consisted of recording 100 original utterances by the subject, ascertaining the subject's ability to count, and the subject's Peabody Picture Vocabulary Test raw score. From the 100 original utterances, the subject's mean length of utterance count was calculated. Due to the limited space at the center, the subjects were pretested throughout the center: in the observation booth, in the conference room and outside on the playground. 26 The Peabody Picture Vocabulary Test was generally adminis- tered in the quiet conference room. The other places in the center were more conducive for initiating the subject's speech, and thus were utilized for obtaining the subject's 100 original utterances and ability to count. The treatment situation was set up in the conference room. This room is equiped with two tables, a soft drink machine, a table cluttered with instructional materials, a rest room and a sink, and water fountain. Since there was no television set at the day-care center one had to be supplied. A Sony 3500 video tape recording machine was connected to the television set in order to establish a playback system for the pre-recorded and edited treatment tapes. This system remained in the conference room until the conclusion of the study. Schedule of treatment. So as not to disrupt the continuity of the normal daily activities in the units, a schedule for treatments was established. This schedule ran Monday through Friday for two consecutive weeks. Monday through Friday Morning Schedule 8:35 to 8:50 treatment 8:55 to- 9:10 treatment 10:30 to 10:45 treatment 10:50 to 11:05 treatment 27 Two make-up sessions were scheduled on the tenth and eleventh days of the experiment. These sessions were in the after- noon. Afternoon Schedule for Make:gp 2:30 to 2:45 treatment 2:50 to 3:05 treatment In the letter sent home to the parents, emphasis was placed on having the subjects at the center no later than 8:30 a.m. Sporadically, a few subjects arrived late, which required a time delay in the treatment schedule. But in general, treatments began promptly at 8:35 a.m. Whichever group, 1 or 2 had the full nine subjects at the center first, was given the treatment first. They were also given the treatment first on the second round of morning treatments. Thus, if all nine subjects in group 2 (modeling) arrived at the center prior to 8:30 a.m., they viewed the treatment from 8:35 to 8:50 and then again from 10:30 to 10:45. The study had been designed to alternate the first exposure sessions between the two groups, but the sporadic arrival of the subjects did not allow this systematic scheduling. There were nine original treatments of expansion and nine original treatments of modeling on video tape. These tapes were numbered 1 through 9 and shown in that order. The same treatment tape was not shown twice in one day. Repetition of the tapes with this time lapse between expo- sures was not considered redundant for the subjects. 28 At this age, repetition and continuous associations are necessary for the rote learning of basic knowledge (Ball & Bogatz, 1972; Cass, 1970). Within many of the one hour programs on Sesame Street repetition of certain segments often occurs. Each head teacher explained to all the children in each unit that some children whose parents had given special permission, would go with Ms. Diane and watch television. The subjects in group 1 (expansion) were only told that they were in group 1. In a like manner, the subjects in group 2 (modeling) were only told they were in group 2. The first two days of getting the subjects to remember and recognize which group they were in was time consuming and confusing for them. On the third day, the subjects were able to recognize the other subjects in their group and what group number they were. This facilitated the process of getting the subjects from the units and into the conference room for treatment. At the onset of the study, there were 10 subjects in each cell or group. Within the first week it became appar— ent that two subjects, one from group 1 (expansion) and one from group 2 (modeling) were not going to cooperate accords ing to the design of the study. In order to maintain a low level of anxiety for all the subjects involved in the ex- periment, these two were dropped from the study. Another 29 subject, in the control group, was ill at the time of pre— testing and subsequently also had to be dropped from the study. The eliminations resulted in a total of 27 subjects participating in the study; 9 in each group. As encouragement and incentive for the subjects to sit quietly and watch their respective segments, each subject was given a piece of candy after each treatment.* As the study progressed, the candy played an important role in facilitating the subjects to attend the treatments. The purpose of the control group was to serve as a base-line against which to measure change in language development beyond that due to maturation and other uncon— trolled variables. Once the group assignments had been made, and these subjects were pretested, there was no further contact made with them until posttesting. They participated in the daily routine at the center. Posttestingpprocedure. For posttesting, which followed the conclusion of two weeks of treatment, the nine assistants were again randomly assigned three subjects. The subjects were posttested on the same three indices on which they were pretested. The procedure for the posttesting was identical to that of the pretesting. * M & M candies were found to provide the most "incentive" for these subjects. 30 A few of the subjects in the control group and many of the older children at the day—care center wanted to watch the television. As a compromise for these children, after posttesting was completed, a television was taken into each unit so all the children could watch Sesame Street. Measures of Language Development Trying to study the acquisition of grammar by more direct experimental techniques is complicated by working with children at a young age when this process takes place. Probably the most difficult element of this research was establishing valid and reliable dependent variables, measures of a child's language development which together constitute an operational definition of that development. Theoretically, there is a vast difference between what a child can do and what he actually does. Likewise, there are variations in styles of television viewing by children. Some children can view television for hours with their eyes rarely leaving the set. we were so struck by this viewing style when we first began doing research on appeal that we coined the term 'zombie viewer"to refer to the child that sat seemingly hypnotized, in front of the set. Other children constantly keep a check on all outside activities in the room while they view. We found these styles to be no guarantee of.how much the child was absorbing from the program. '(Reeves, 1970, p. 11) The only means of ascertaining what a child knows or can do is by the child spontaneously responding and then rating this response on some predetermined precise measuring scale. 31 Acquisition of grammar refers to the increasingly com- plex structures which the child "has succeeded in mastering and internalizing, whether or not he utilizes them in practice, without interference from the many other factors which play a role in actual behavior" (Chomsky, 1964, p. 36). There is general agreement that linguistic competence in the sense of these structures is not the same thing as linguistic performance in the sense of overt verbal behavior (Cazden, 1965). Obviously one can find out about competence only by studying performance (Chomsky, 1964), but drawing conclusions about the child's development of linguistic competence from an analysis of samples of spontaneous speech should be done with caution and discrimination. Studying the acquisition of grammar by more direct ex- perimental techniques, by definition is dependent upon the response on the part of the subject. Soliciting responses is further complicated when working with young children, when this process occurs. An experimental situation had to be designed in which a 31-month-old child was able to spontaneously respond. Such tests or indices were used in this research. The de- sign relied heavily on the analysis of samples of spontaneous speech by the subjects. The tests also had to relate to the segments being utilized as the treatment. Children's Television Workshop 32 producers designed Sesame Street primarily as a tool for teaching language concepts and auditory and visual dis- crimination (Polsky, 1974). CTW's programming objectives conveniently correspond to the objectives of the treatment segments for this study. The three dependent variables are measures of certain aspects of spontaneous speech. They are: mean length of utterance, ability to count, and the raw score on the Pea- body Picture Vocabulary Test (PPVT). Mean length of utterance. Mean length of utterance has been the single most wide- ly used measure of language development (McCarthy, 1954). The validity of this test is based on the widespread finding that it increases with age (Cazden, 1965) and language development. The mean length of utterance for each subject was com- puted in morphemes, not words. This process requires prior decision about what is an utterance and what is a morpheme. The boundary of an utterance is based on "the usual criteria of either a prolonged pause or a shift of speakers" (Brown & Frase, 1964, p. 52). The process of determining what is a morpheme is more difficult and requires many arbitrary decisions. A morpheme is not a word; conversely each word is not a morpheme. A word may include more than one morpheme 33 (e.g., nowhere and running), or a word may be combined into a single morpheme, as a proper name (e.g., Santa Claus and Easter Bunny). Morphemes are further classified as free morphemes (e.g., blue and gap) which can appear alone, and bgund morphemes (e.g., ing and 1y) which never appear alone. Determining what is a morpheme is difficult because morphemes are defined as the "smallest units of structure which embody grammatical or lexical meanings" (Carroll, 1959, p. 24). Since meaning is relative to the individual, a set of rules for totaling the number of morphemes and com- puting the mean was established. Appendix I lists the rules for ascertaining the subjects mean length of utterance count. The assistants had the option of recording the subjects speech with a tape recorder or writing it down as the child spoke. The assistants were allowed to converse with sub- jects and/or record verbal interaction between the subjects and another child or teacher. The assistants, on a written transcript of each sub: ject's utterances, computed the mean length of utterance for the pretest and the posttest using the same set of predeter- mined rules. Each pretest and posttest computation was then recalculated by the director of the study in order to raise the level of consistency in the process of computing the mean length of utterance. 34 Abilityyto count. One of the specific goals of Sesame Street is categor- ized as "numbers goals" (Ball & Bogatz, 1971). There is a breakdown within this category that includes: 1. "recitation--the child can recite the numbers from 1 to 20. 2. "enumeration--the child can define a set or subset of up to 10 objects from a larger set. example: Here are some pennies? How many are there? Appendix F lists the other objectives under the numbers goal category. Since this cognitive skill is one of the main objec- tives of Sesame Street it was included in the battery of dependent variables to be tested. This skill involves rote learning and a high level of repetition. Most of the segments on Sesame Street concerning num- erals were classified under the modeling treatment. This was a result of the high repetition of low verbage content of these segments. There were some cases where the concept of numbers was incorporated into a musical format. These cases resulted in those segments being classified under the expansion treatment, due to high verbage and low repetition. The subject's ability to count was established by three processes: 1. the subject was asked to "recite" or count as far as he was able 35 2. the subject was asked to count the number of ob- jects in front of him (these varied: pennies, blocks, leaves, pinecones, etc.) 3. the subject was asked once again to count as far as he was able. If the subject's ability varied during the three tests, all three steps were repeated until the subject achieved the same count on all three trials. This constituted the accurate level of the subject's ability to count. This process was done with all three groups for the pretest and the posttest. Peabody_Picture Vocabulary Test (PPVT) "The Peabody Picture Vocabulary Test is designed to provide an estimate of a subject's verbal intelligence through measuring his hearing vocabulary" (Dunn, 1959, p. 25). Because of the young age and accompanying inabilities of the subjects, an intelligence test had to be appropriated to their abilities. Since the PPVT does not require the subjects to be able to read, the scale is applicable to any "English Speaking resident of the United States between 2 years 6 months and 18 years who is able to hear words, see the drawings, and has the facility to indicate "yes“ and "no" in a manner which communicates" (Dunn, 1959, p. 25). A number of other advantages of the PPVT include: (1) the test has high interest value and therefore is a good rapport establisher (2) extensive Specialized preparation is not needed for its administration 36 (3) it is quickly given in 10 to 15 minutes (4) scoring is completely objective and quickly accom- plished in one or two minutes (5) it is completely untimed and thus is a power rather than a speed test (6) the test covers a wide age range. The process for administering and scoring the Peabody Picture Vocabulary Test is described in Appendices C and D. The reliability of the PPVT was established by calculat- ing Pearson product-moment correlations on the raw scores of the standardization subjects for Forms A and B at each age level. Validity data for the PPVT, or the extent to which it measures what it purports to measure, are of two types: rational and statistical. Content and construct are estab- lished validities under rational; congruent, concurrent and predictive are established validities under statistical. The Peabody Picture Vocabulary Test was included in the battery of tests administered to the subjects of "The Second Year of Sesame Street: A Continuing Evaluation" (Ball and Bogatz, 1971), in order to assess the level of vocabulary. The PPVT is a standardized test and was also used for the first year evaluation of Sesame Street (Ball and Bogatz, l970).* *The PPVT may have a racial bias. Only two pictures portray blacks--one is a porter and the other is a spear carrier. However, it has a rich history in research studies, and its dubious value in black and white comparisons is not the variable under research in this study. 37 Statistical Analyses This experiment consisted of a pretest-posttest control group design. This design conveniently controls for all of the seven rival hypotheses1 (Campbell and Stanley, 1963). The design for this experiment was as follows: R 01 X 02 (expansion) R 03 x 04 (modeling) R 05 06 (control) This design assumes equivalent groups are achieved by randomized assignments. The most widely used acceptable test for significance for this experimental design "is to compute for each group pretest-posttest gain scores and to compute a 't' between experimental and control groups on these gain scores" (Campbell and Stanley, 1963, p. 23). This procedure was utilized for analyzing the data for this experiment. A non-parametric technique of data anlysis was used to give additional insight into the data and contribute to the interpretation of results. 1Testing, maturation, history, instrumentation, regres- sion, selection and mortality. CHAPTER 3 ANALYSIS OF RESULTS Evaluation of Measures of Language Development In evaluating the measures of language development, it is assumed (1) that a developmental characteristic must either stay at the same level or increase; (2) that a valid measure of a developmental characteristic will yield scores conforming to that pattern; and (3) that converse fluctua- tion in the scores indicate the presence of unaccountable: superficial influences. Table 2, on the following page, gives the mean scores for pretest and posttest on the three language measures for. the three groups. On all but the mean length of utterance for the modeling group and the ability to count for the con- trol group, the mean scores increased from pretest to post- test. The dependent variables appear to have lacked a high level of correlation at the onset of this study. It is not possible to say which of the three measures of spontaneous speech is most valid, but mean length of utterance seems to 38 39 Table 2. Mean Scores and Mean Gain Scores for Measures of Language Development Pretest Posttest: MeanGain Mean length of utterance expansion 4.656 6.899 .7243 modeling 5.215 4.990 .1762 control 4.654 4.889 .5885 Peabody Picture Vocabulary Test expansion 39.88 44.22 8.000 modeling 43.00 47.11 6.777 control 39.88 40.77 4.777 Ability to count expansion 8.00 12.88 3.55 modeling 10.00 13.33 3.44 control 13.77 13.00 1.000 be the most widely accepted measure. This is in view of its widespread use in language acquisition research. It does appear that of these three measures, the raw score on the Peabody Picture Vocabulary Test is the least valid. It is the measure that fluctuated from the mean most frequently. A plausible explanation would come from the influence the assistants had over this measure. Much of the results of the PPVT are contingent upon the verbal rein- forcement and enthusiasm of the administrators of the test. 4O Varying levels of verbal reinforcement and enthusiasm result in varying achievement scores. It was assumed that randomi- zation would account for these extraneous variables, but it apparently did not. Test of Hypotheses In the process of operationalizing "greater language growth" each of the three research hypotheses were sub- categorized into three alternative hypotheses to test the effect of each index of language development. Gain scores were computed for each group between the pretest scores and the posttest scores. A t-test was computed between experi- mental and control groups on these gain scores for each index. The first research hypothesis, which states children viewing the television segments containing expanded material will show a greater language growth than do children receiv- ing neither treatment was broken down into three alternative hypotheses: 1. children viewing the television segments containing expanded material will show a greater increase in their mean length of utterance count than will children receiving neither treatment; 2. children viewing the television segments containing expanded material will show a greater increase in their ability to count than will children receiving neither treatment; 3. children viewing the television segments containing expanded material will show a greater increase in 41 their raw score on the Peabody Picture Vocabulary Test than will children receiving neither treat- ment. Of these three alternative hypotheses, only one was significant at the .10 level--chi1dren viewing the tele- vision segments containing expanded material showed a great- er increase in their ability to count than did the children receiving neither treatment. The other two alternative hypotheses demonstrated an increase in trend, but not at a significant level. The second research hypothesis, which states children viewing the television segments containing modeled material will show a greater language growth than do children receiv- ing neither treatment was broken down into three alternative hypotheses: 1. children viewing the television segments containing modeled material will show a greater increase in their mean length of utterance count than will children receiving neither treatment; 2. children viewing the television segments containing modeled material will show a greater increase in their ability to count than will children receiving neither treatment; 3. children viewing the television segments containing modeled material will show a greater increase in their raw score on the Peabody Picture Vocabulary Test than will children receiving neither treatment. The control group achieved significantly higher, at the .10 level, on the mean length of utterance count than the modeling group did. The converse effect for this dependent variable is possibly due to the presence of unaccountable 42 influences in both groups. The modeling group did gain sig- nificantly at the .10 level over the control group in their ability to count. While there was not a significant dif- ference between the two groups raw score on the PPVT, there was a positive increase for the modeling group, supporting the trend. To determine which of the two experimental treatments had the most significant effect, the third research hypothe- sis was analyzed by a comparison of gain scores on three alternative hypotheses. These alternative hypotheses state: 1. of the two treatments, children viewing the tele- vision segments containing expanded material will show a greater increase in their mean length of utterance count; 2. of the two treatments, children viewing the tele- vision segments containing expanded material will show a greater increase in their ability to count; 3. of the two treatments, children viewing the tele- vision segments containing expanded material will show a greater increase in their raw score on the Peabody Picture Vocabulary Test. Only one of these three alternative hypotheses support— ed the research hypothesis on the significant level. The expansion treatments did have a significant effect over the modeled treatments for increasing a child's mean length of utterance count. The other two calculations demonstrated some positive support for their respective alternative hypothesis, but not at a significant level. The t-test and corresponding level of significance for all alternative hypotheses are found in Table 3. 43 Table 3. t-Test of Gain Scores Between Experimental and Control Groups t-Value Expansion/Control Peabody Picture Vocabulary Test .8558 N.S. ability to count 1.465 p < .10 mean length of utterance .3447 N.S. Modeling/Control Peabody Picture Vocabulary Test .5321 N.S. ability to count 1.700 p < .10 mean length of utterance -1.520 p < .10 Expansion/Modeling Peabody Picture Vocabulary Test .2680 N.S. ability to count .0543 N.S. mean length of utterance 1.689 p < .10 44 Instead of testing the significance of the amount of mean gain, counting the number of subjects who demonstrated an increase would determine the probability of obtaining that many or more changes in one direction by chance alone. Tables 4 and 5 present the Sign Test for correlated samples (Siegel, 1956) for the combined scores of the ex- pansion and modeling groups and the Sign Test for the con- trol group scores. The sample size was reduced in cases of zero difference. For the combined scores for expansion and modeling groups, the probability of this proportion of subjects show- ing growth was <.05 for the measurement of the ability to count, <.Ol for the measurement on the Peabody Picture Vocabulary Test, and no difference from chance on the sub- ject's mean length of utterance count. In contrast the probability of this proportion of subjects showing growth for the control group was due solely to chance for all in- dices of language measurements. The significant data indicates that the treatments, both for modeling and expansion, were the result of reliable dependent measures and not by chance. It gives encourage- ment that a partially controlled linguistic environment can systematically enhance and increase a child's acquisition of language. Two of the three dependent variables resulted in statistical significant data, strengthening the conclusion 45 Table 4. Sign Test of Direction of Change in Individual Scores on Three Language Measures for Combined Expansion and Modeling Group Scores Test Pretest‘