Comparing the Effects of Two Types of Output-Pushing Feedback on Adult EFL Students' Oral Accuracy

ELIA 16, 2016, pp. 49-78 49 Hanne Roo hoo DOI: h p://dx.doi.o g/10.12795/elia.2016.i16.03 Es udios de lingüís ica inglesa aplicada COMPARING THE EFFECTS OF TWO TYPES OF OUTPUT- PUSHING FEEDBACK ON ADULT EFL STUDENTS’ ORAL ACCURACY UNA COMPARACIÓN DE LOS EFECTOS DE DOS TIPOS DE RETROALIMENTACIÓN CORRECTIVA EN LA PRODUCCIÓN ORAL DE LOS ALUMNOS ADULTOS DE INGLÉS. Hanne Roo hoo Uni e sidad Pública de Na a a, Spain [email p o ec ed] DOI: h p://dx.doi.o g/10.12795/elia.2016.i16.03 Despi e g owing e idence ha p o iding immedia e o al co ec i e eedback can help second language s uden s speak mo e accu a ely, he way in which eedback should be gi en emains con o e sial. While a g ea numbe o s udies suppo he use o ecas s, au ho s such as Lys e and Sai o (2010) claim ha ou pu -pushing eedback could be mo e bene icial o class oom lea ne s. Howe e , ew s udies ha e a emp ed o ease apa he e ec s o sepa a e ou pu -pushing eedback- ypes o p omp s, which is why he p esen small-scale s udy compa es wo ELIA 16, 2016, pp. 49-78 50 DOI: h p://dx.doi.o g/10.12795/elia.2016.i16.03 Compa ing he e ec s o wo ypes o ou pu -pushing eedback... kinds o p omp s; me alinguis ic eedback and elici a ion. Two g oups o in e media e uni e si y s uden s (n=31) ecei ed ei he elici a ion o me alinguis ic eedback while ca ying ou communica i e speaking ac i i ies. A s a is ical analysis o s uden s’ gain sco es om p e- o pos - es shows ha me alinguis ic eedback had g ea e e ec s on s uden s’ accu acy, al hough hese we e no signi ican . Me alinguis ic eedback was also much mo e success ul a helping s uden s co ec hei e o s du ing he class oom in e en ion. Key wo ds: o al co ec i e eedback, EFL, speaking, me alinguis ic eedback, elici a ion. Aunque nume osos abajos han demos ado que la e oalimen ación o al inmedia a puede ayuda a los alumnos de una segunda lengua a habla más co ec amen e, no hay acue do sob e qué ipo de co ección es más adecuado. Si bien un g an núme o de es udios apoyan el uso de e o mulaciones, au o es como Lys e y Sai o (2010) a i man que, en el con ex o escola , puede esul a más bene icioso p opo ciona a los es udian es los denominados “p omp s”, que cons i uyen un ipo de e oalimen ación que obliga a los alumnos a p oduci “ou pu ”. A día de hoy, apenas se conocen los e ec os de di e en es ipos de “p omp s”, po ello, es e es udio a pequeña escala compa a dos ipos: la e oalimen ación me alingüís ica y la elici ación. Dos g upos de alumnos uni e si a ios de ni el in e medio (n=31) ecibie on o bien elici ación o bien e oalimen ación me alingüís ica mien as e ec ua on ac i idades o ales comunica i as. Los esul ados mues an que los alumnos que ecibie on e oalimen ación me alingüís ica mejo a on más en e el es p e io y pos e io , aunque es a mejo a no ue signi ica i a. Asimismo, es os alumnos ue on más capaces de co egi sus e o es du an e la in e ención en el aula. Palab as cla e: e oalimen ación o al, enseñanza del inglés como lengua ex anje a, des eza o al, e oalimen ación me alingüís ica, elici ación. ELIA 16, 2016, pp. 49-78 51 Hanne Roo hoo DOI: h p://dx.doi.o g/10.12795/elia.2016.i16.03 1. In oduc ion The ques ion o how o p o ide eedback o second language lea ne s on hei spoken pe o mance is an issue o immedia e impo ance o class oom eache s ha has gi en ise o conside able con o e sy among esea che s. A g owing body o esea ch ca ied ou wi hin he in e ac ion app oach o second language acquisi ion (SLA) suppo s he use o immedia e o al co ec i e eedback (CF) du ing communica i e speaking asks, as se e al me a-analyses ha e ound posi i e and du able e ec s o CF on second language acquisi ion (e.g., Li, 2010; Mackey & Goo, 2007; and Russell & Spada, 2006). The in e ac ion app oach o SLA (Gass & Mackey, 2015) is based on a hypo he ical link be ween in e ac ion and language de elopmen , since in e ac ion p o ides lea ne s wi h modi ied inpu and eedback on hei ou pu . Despi e hese posi i e indings on he use ulness o in e ac ion and eedback, esea che s con inue o deba e he bes me hod o p o iding CF, ei he implici ly in he o m o ecas s, o by using mo e explici echniques. Mos s udies showing posi i e e ec s o ecas s ha e been ca ied ou in a labo a o y se ing, whe e CF is p o ided du ing one-on-one in e ac ion be ween a na i e speake and a lea ne (e.g., Ishida, 2004; Leeman, 2003; Loewen & Nabei, 2007; Mackey & Philp, 1998; McDonough & Mackey, 2006). In class oom se ings, obse a ional s udies almos in a iably conclude ha eache s p edominan ly use ecas s o deal wi h hei s uden s’ spoken language e o s (e.g., Sheen, 2004). Howe e , e idence om class oom s udies indica es ha s uden s may bene i mo e om being pushed o sel -co ec by means o p omp s o ou pu -pushing CF (Lys e & Sai o, 2010). While se e al s udies ha e compa ed he impac o ecas s o ha o a g oup o p omp s (e.g., Amma & Spada, 2006; Dilans, 2010; Lys e & Izquie do, 2009; Lys e & Sai o, 2010; Nassaji, 2007), e y ew s udies ha e a emp ed o compa e sepa a e p omp s o each o he , e en hough p omp s comp ise a ange o CF- ypes which may well a ec language de elopmen di e en ly. To ill his gap in he li e a u e, he p esen s udy aims o compa e he e ec s o wo ypes o p omp s o ou pu -pushing CF, elici a ion and me alinguis ic eedback, on EFL s uden s’ abili y o use he English pas simple ense in o al p oduc ion. ELIA 16, 2016, pp. 49-78 52 DOI: h p://dx.doi.o g/10.12795/elia.2016.i16.03 Compa ing he e ec s o wo ypes o ou pu -pushing eedback... 2. Li e a u e Re iew 2.1. Resea ch on Di e en Feedback- ypes When ying o accoun o he e ec s o a ious o al CF- ypes, esea che s ha e classi ied CF-mo es based on ei he hei deg ee o explici ness, o on whe he hey con ain he co ec o m. Acco dingly, ecas s a e gene ally classi ied as implici (e en hough hey can be mo e explici i hey a e educed o s essed, see o ins ance Sheen, 2006) and inpu -p o iding, whe eas p omp s a e all ou pu -pushing echniques a ying in explici ness. The mos explici ype o p omp is me alinguis ic eedback, while elici a ion can be classi ied as mo e implici . Table 1 p o ides examples o implici and explici inpu - and ou pu -p o iding CF in esponse o an example o a lea ne e o in he use o he pas ense. I needs o be no ed ha deg ee o explici ness is a con inuum and ha elici a ion is classi ied as explici by some au ho s (Lys e , Sai o & Sa o, 2013). S uden e o : Yes e day I go o he cinema. Inpu -p o iding ( e o mula ions) Ou pu -pushing (p omp s) Implici eedback Recas : Oh, you wen o he cinema. Elici a ion: Las weekend, I… Explici eedback Explici co ec ion: No go, wen . Me alinguis ic eedback: You need pas ense. Table 1: Classi ica ion o CF- ypes based on Lys e e al., 2013 Inpu -p o iding CF, pa icula ly in he o m o ecas s, is hypo hesized o os e language acquisi ion because i p o ides a b ie oppo uni y o he s uden o ocus on o m while emaining ocused on he meaning o he message (Long, 1991). Since ecas s immedia ely ollow he lea ne ’s e oneous u e ance, and exp ess he lea ne ’s o iginal meaning, esea che s (e.g., Goo & Mackey, 2013) ha e sugges ed ha his ype o eedback allows lea ne s o compa e he co ec o m wi h he e o hey ha e jus made. ELIA 16, 2016, pp. 49-78 53 Hanne Roo hoo DOI: h p://dx.doi.o g/10.12795/elia.2016.i16.03 On he o he hand, ins uc o -p o ided p omp s can be expec ed o p omo e language acquisi ion based on Swain’s (1995, 2005) Ou pu Hypo hesis, which posi s ha pushing lea ne s o exp ess hemsel es mo e p ecisely and mo e accu a ely may lead o in e language de elopmen . Acco ding o de Bo (1996), ac i ely looking o he igh o m, a he han ecei ing i om he eache o in e locu o , s eng hens o m-meaning connec ions in lea ne s’ minds, and hus helps hese lea ne s p ocedu alize hei decla a i e knowledge. Many esea che s epo ing subs an ial suppo o he e ec i eness o ecas s do so by compa ing he g oup ecei ing his ype o CF o a con ol g oup ecei ing no CF, and by conduc ing hei s udies in a labo a o y — a he han in a class oom— se ing (e.g., Han, 2002; Iwashi a, 2001; Leeman, 2003; Mackey & Philp, 1998). As men ioned in he in oduc ion, Lys e and Sai o’s (2010) me a-analysis o 15 class oom s udies ound la ge e ec s o p omp s han ecas s. Howe e , he e ec s o CF a e no independen o ac o s such as he lea ne s’ p o iciency o he na u e o he a ge s uc u e. Fo ins ance, Yang and Lys e (2010) obse ed ha ecas s we e equally e ec i e o he acquisi ion o he egula and i egula pas ense, bu p omp s had g ea e e ec s on s uden s’ accu a e p oduc ion o he egula pas . Rega ding p o iciency, Amma and Spada (2006) ound ha ecas s bene i ed high-p o iciency lea ne s and low-p o iciency lea ne s equally, bu p omp s had g ea e e ec s on he acquisi ion o English possessi e p onouns by low-p o iciency lea ne s. Ra he han opposing ecas s o p omp s, some au ho s ha e compa ed ecas s o me alinguis ic eedback o explici co ec ion in o de o add ess he ques ion whe he implici o explici CF is mo e e ec i e. I has been ound ha ecas s, due o hei implici na u e, a e no always in e p e ed as co ec i e by lea ne s (Lys e , 1998), especially in he case o mo phosyn ac ic e o s (Kim & Han, 2007; Mackey, Gass & McDonough, 2000). Ellis, Loewen and E lam (2006) compa ed he e ec s o ecas s o me alinguis ic eedback on he acquisi ion o he pas simple ense and ound la ge e ec s o he mo e explici eedback. Ano he s udy by Sheen (2007) showed g ea e bene i s o explici eedback han o ecas s on he acquisi ion o English a icles. ELIA 16, 2016, pp. 49-78 54 DOI: h p://dx.doi.o g/10.12795/elia.2016.i16.03 Compa ing he e ec s o wo ypes o ou pu -pushing eedback... The e a e hus some indica ions ha p omp s a e use ul o helping lea ne s acqui e ce ain g amma ical s uc u es in a class oom con ex . A he same ime, he e is e idence ha mo e explici CF has a g ea e impac on lea ne s’ accu acy han implici co ec ion echniques. Al hough p omp s encompass a wide ange o echniques, bo h implici and e y explici , ha dly any esea ch has compa ed p omp s wi h each o he . While some s udies ha e ocused on me alinguis ic eedback (e.g., Ellis, 2007; Ellis e al., 2006; Loewen & Nabei, 2007), o he bes o ou knowledge no s udies ha e deal wi h elici a ion as a sepa a e ea men condi ion. Elici a ion is none heless used by eache s, and appea s o be e y success ul a pushing lea ne s o e o mula e hei ung amma ical u e ances (Lys e & Ran a, 1997; Sheen, 2004). To con ibu e o he sca ce body o li e a u e on his issue, he p esen s udy compa es he e ec s o wo ypes o p omp s, me alinguis ic eedback and elici a ion, on s uden s’ o al p oduc ion o he English simple pas ense. Se e al s udies ha e shown ha he pas ense can be in luenced posi i ely by CF (Dough y & Va ela, 1998; Ellis e al., 2006; Han, 2002; McDonough, 2007; Takashima & Ellis, 1999; Yang & Lys e , 2010). Mo eo e , he pas ense lends i sel well o he use o p omp s, since s uden s a he in e media e o highe p o iciency le el should al eady ha e some knowledge o he s uc u e and, he e o e, may be able o e ie e he co ec o m when pushed o ou pu by hei in e locu o . 2.2. The Ta ge S uc u e: he Regula and I egula Pas Simple Tense E en hough English ex books such as New English File (Ox o d), To al English (Longman) o Face o Face (Camb idge) in oduce he pas simple ense a he elemen a y o A1 le el, i con inues o be a p oblema ic a ea o lea ne s as hey p og ess owa ds mo e ad anced le els o English, especially in o al p oduc ion. Mo pheme s udies such as Dulay and Bu (1974) o La sen- F eeman (1975) show ha he pas ense is acqui ed la e in compa ison wi h o he mo phemes such as p og essi e –ing o a icles. ELIA 16, 2016, pp. 49-78 55 Hanne Roo hoo DOI: h p://dx.doi.o g/10.12795/elia.2016.i16.03 To explain why ce ain mo phemes appea o pose g ea e di icul y o lea ne s han o he s, Goldschneide and DeKeyse (2001) no e ha pe cep ual salience is one o he ac o s which may in luence he o de o acquisi ion o hese mo phemes. In he case o he egula pas ense, he egula –ed ending is pe cep ually non-salien , because i is usually ealized as an al eola s op (ei he oiceless [ ] o oiced [d]) and as such does no cons i u e a syllable. Apa om i s lack o saliency, he pas ense ending is also o en edundan , as bo h con ex ual in o ma ion and ad e bial ph ases such as “yes e day” o “las week” can be used o in e p e a sen ence as pas (VanPa en, 1996). This means we can expec lea ne s o make e o s when using he pas ense. As s a ed by Ellis e al. (2006: p.251), “ he ypical e o made by lea ne s is he use o he simple o p esen o m o he e b in place o V–ed: *Yes e day I isi my sis e .” Mo eo e , in case o i egula e bs, lea ne s may o e gene alize he –ed ending o i egula pas ense o ms. 2.3. Resea ch Ques ions Two main esea ch ques ions unde lie he p esen s udy. The i s ques ion pe ains o he immedia e impac o hese ypes o o al CF, which was s udied by looking a he a es o up ake and epai ollowing he eache ’s CF. Up ake e e s o he lea ne ’s esponse o a eache CF mo e and indica es ha he lea ne has no iced he CF. Repai e e s o he s uden ’s success ul co ec ion o his o he ini ial e o , ollowing he eache ’s CF. The second ques ion is ela ed o he e ec s o he wo ypes o CF on lea ne s’ subsequen language p oduc ion on an immedia e pos - es . 1. Do elici a ion and me alinguis ic eedback on lea ne s’ pas ense e o s du ing class oom communica i e ac i i ies di e in e ms o a es o lea ne up ake and epai ? 2. A e he e p e- o pos - es gains in lea ne s’ accu a e o al p oduc ion o pas ense o ms and is he e a s a is ically signi ican di e ence be ween he elici a ion g oup and he me alinguis ic eedback g oup? ELIA 16, 2016, pp. 49-78 56 DOI: h p://dx.doi.o g/10.12795/elia.2016.i16.03 Compa ing he e ec s o wo ypes o ou pu -pushing eedback... 3. Me hodology 3.1. Pa icipan s The pa icipan s we e 31 unde g adua e s uden s, 5 male and 26 emale, om wo in ac classes (de ined by Po e, 2010, p.41 as “classes o which he s uden s ha e been assigned p io o he s udy i sel ”) in he i s yea o a P ima y Teaching Deg ee a a Spanish uni e si y, who had been placed in an in e media e English g oup based on in-house placemen es s. Pa icipan s we e be ween 19 and 23 yea s old, wi h a mean age o 19.8. They epo ed ha ing lea ned English o be ween 8 and 18 yea s (13.7 yea s on a e age). All o hem we e na i e speake s o Spanish. Thei English classes ook place wice a week o one hou and a hal pe lesson and hey used he ex book Ready o PET. The lessons consis ed o a combina ion o skills-wo k and g amma and ocabula y exe cises, bu ela i ely li le ime was de o ed o speaking. They did no ecei e any ins uc ion on he pas simple ense du ing he ime o he s udy, bu he e is no doub hey did ecei e explici ins uc ion on how o o m he pas ense in English in p e ious English classes, since all o hem had ecei ed se e al yea s o g amma -o ien ed ins uc ion in he pas . The pa icipan s had e y limi ed exposu e o English ou side o he class oom and mos o hem had no spen ime in an English-speaking coun y, as hey indica ed on a backg ound ques ionnai e hey comple ed du ing he i s session o he expe imen . Be o e he expe imen , he pa icipan s we e in o med ha hey would ake pa in a s udy which would allow hem o p ac ice hei speaking skills. They all signed a consen o m in which hey ag eed o being audio- eco ded. E en hough 31 s uden s ook pa in he s udy, only 16 s uden s a ended all ou sessions, one o whom was excluded because she had sco ed ze o on he g amma icali y judgmen p e- es . This means he da a ela ed o p e- and pos - es gains a e only based on 5 s uden s o he me alinguis ic eedback g oup and 10 s uden s o he elici a ion g oup. 3.2. Design The s udy ook place du ing he s uden s’ no mal class ime and he same esea che ca ied ou he asks wi h bo h g oups. Fou sessions o abou 1 ELIA 16, 2016, pp. 49-78 57 Hanne Roo hoo DOI: h p://dx.doi.o g/10.12795/elia.2016.i16.03 hou each, sp ead o e wo consecu i e weeks, we e dedica ed o he s udy. A p e- es /pos - es design was used and he classes we e andomly assigned o a ea men condi ion, esul ing in a me alinguis ic eedback g oup and an elici a ion g oup. As in ac classes we e used, he e we e sligh ly mo e s uden s in he elici a ion g oup (n=18) han in he me alinguis ic eedback g oup (n=13). Figu e 1 illus a es he design. Figu e 1: Design o he s udy 3.3. P ocedu e and Ma e ials Table 2 p o ides an o e iew o he di e en ma e ials and asks used du ing he s udy. To es he s uden s’ explici knowledge o he pas ense, along wi h hei abili y o p oduce he s uc u e in speaking, hey we e gi en a g amma icali y judgmen es and an o al s o y- elling es p io o he ea men . As able 2 shows, a simila p ocedu e was ollowed o he pos es . ELIA 16, 2016, pp. 49-78 64 DOI: h p://dx.doi.o g/10.12795/elia.2016.i16.03 Compa ing he e ec s o wo ypes o ou pu -pushing eedback... Co ec pas ense o ms Pas ense e o s CF Up ake Repai Elici a ion (n=18) 116 102 80 77 (96.3%) 38 (47.5%) Me alinguis ic eedback (n=13) 179 62 59 56 (95%) 49 (83.1%) Table 4: Up ake and epai du ing he ea men sessions The wo ypes o CF a e hus equally e ec i e a p omp ing s uden s o modi y hei ini ial u e ance, bu he e seems o be an impo an di e ence in hei abili y o help s uden s co ec hei pas ense e o s. 4.2. Resea ch Ques ion 2: P e- o Pos - es E ec s o he Two Types o P omp s While he analysis o up ake and epai p o ides use ul in o ma ion abou s uden s’ eac ions o CF in he con ex o a communica i e ask, esea che s ha e wa ned agains in e p e ing up ake as a sign o language de elopmen (e.g., Goo & Mackey, 2013). To ind ou i he wo p omp s had an impac on he s uden s’ abili y o use he pas ense accu a ely in o al asks, we need o analyse hei pe o mance on he o al p e- and pos - es . Tables 5 and 6 con ain he esul s o he s o y- elling es s o he elici a ion and he me alinguis ic eedback g oup, espec i ely. The esul s o he pos es a e di ided in o he accu acy sco es o he same wo s o ies ha we e used in he p e es and he sco es on he new s o y, which he s uden s old o he i s ime in he pos es . S uden P e es Pos es Same s o ies Pos es Di e en s o y E1 1/15 7% 3/15 20% 4/8 50% E2 13/22 59% 9/18 50% 6/14 42.9% E3 7/18 38.9% 6/14 42.8% 4/8 50% ELIA 16, 2016, pp. 49-78 65 Hanne Roo hoo DOI: h p://dx.doi.o g/10.12795/elia.2016.i16.03 E4 7/11 63.6% 12/17 70.6% 6/9 66.7% E5 7/14 50% 12/22 54.5% 10/13 76.9% E6 6/23 26.1% 6/17 35.3% 5/8 62.5% E7 7/20 35% 11/20 55% 7/11 63.6% E8 10/18 55.6% 14/20 70% 6/11 54.5% E9 3/15 20% 10/21 47.6% 7/11 63.6% E10 1/16 6.3% 5/16 31.3% 5/10 50% Mean sco e 36.15 (SD 20.9) 47.71% (SD 16.1) 58.07% (SD 10.3) Table 5: Resul s o he s o y elling es s o he elici a ion g oup S uden P e- es Pos - es Same s o ies Pos - es Di e en s o y M1 4/10 40% 11/17 64.7% 5/9 55.6% M2 1/14 7.1% 7/12 58.3% 7/13 53.8% M3 10/16 62.5% 13/17 76.5% 9/11 81.8% M4 7/16 43.8% 11/22 50% 5/12 41.7% M5 1/16 6.3% 11/22 50% 6/11 54.5% Mean sco e 31.94% (SD 24.6) 59.9% (SD 11.1) 57.48% (SD 14.7) Table 6: Resul s o he s o y elling es s o he me alinguis ic eedback g oup An unpai ed - es ( α=0.05, p=0.73) compa ing he p e es sco es o he elici a ion g oup o hose o he me alinguis ic eedback g oup showed no s a is ically signi ican di e ence be ween he wo g oups on he p e es . As seen be o e, he wo g oups we e also ound o be compa able on he g amma icali y judgmen sco es. Two pai ed - es s (α=0.05) compa ing he p e es and pos es sco es (same s o ies) in bo h g oups showed bo h g oups imp o ed signi ican ly a e he ea men (p=0.009 o elici a ion, p=0.03 o me alinguis ic eedback). When compa ing he p e es s o ies wi h he esul s o he new s o y in he pos es , he elici a ion g oup also imp o ed signi ican ly (p=0.01), whe eas he me alinguis ic eedback g oup imp o ed bu he di e ence was no qui e signi ican (p=0.0572). This indica es ha he imp o emen s made in accu a e p oduc ion o he pas ense we e no jus due o p ac ice e ec s o ask epe i ion. ELIA 16, 2016, pp. 49-78 66 DOI: h p://dx.doi.o g/10.12795/elia.2016.i16.03 Compa ing he e ec s o wo ypes o ou pu -pushing eedback... To answe he ques ion whe he hese wo ypes o eedback a ec he acquisi ion o he pas ense di e en ly, gain sco es we e calcula ed by sub ac ing he p e es sco es om he pos es sco es (only aking in o accoun he sco es o same s o ies). These gain sco es a e displayed in able 7. S uden E1, o ins ance, made gains o 13 pe cen om he p e es o he pos es o he same s o ies, whe eas s uden E2’s sco e wen down, which esul s in a nega i e sco e o -9 pe cen . The mean gain sco e o he elici a ion g oup was ound o be 11.56, while he mean gain sco e o he me alinguis ic eedback u ned ou o be somewha highe , i.e. 27.96. An unpai ed - es compa ing he gain sco es o bo h g oups esul ed in a p- alue o 0.0518, which means he di e ence is no qui e s a is ically signi ican , bu he e a e some indica ions ha me alinguis ic eedback migh be mo e e ec i e han elici a ion o he accu a e p oduc ion o he pas ense. S uden Elici a ion Me alinguis ic eedback 1 13% 24.7% 2 -9% 51.2% 3 3.9% 14 % 4 7% 6.2% 5 4.5% 43.7% 6 9.2% - 7 20% - 8 14.4% - 9 27.6% - 10 25% - Mean 11.560% (SD 10.9) 27.960% (SD 19.1) Table 7: Compa ison o gain sco es in he elici a ion and me alinguis ic eedback g oup. When in e p e ing hese esul s, i is impo an o ake in o accoun ha only 5 s uden s could be s udied o he me alinguis ic eedback g oup, compa ed o 10 in he elici a ion g oup. Any di e ences be ween he gains de i ed om hese echniques could hen be a ibu ed o his dispa i y, and no o he na u e o he echnique hemsel es. Mo eo e , majo di e ences ELIA 16, 2016, pp. 49-78 67 Hanne Roo hoo DOI: h p://dx.doi.o g/10.12795/elia.2016.i16.03 exis be ween he imp o emen o indi idual s uden s, and some o hem only imp o ed sligh ly ( o example s uden s E4 and M4). 5. Discussion The esul s o his small-scale in es iga ion in o he e ec s o CF on s uden s’ o al accu acy seem o be in line wi h p e ious esea ch showing ha CF p o ided in he con ex o communica i e asks helps s uden s imp o e hei accu a e p oduc ion o g amma ical s uc u es (e.g., Amma & Spada, 2006; Ellis, 2007; McDonough, 2007; Yang & Lys e , 2010). As we ha e seen, bo h he s uden s ecei ing elici a ion and hose ecei ing me alinguis ic eedback signi ican ly imp o ed hei accu a e p oduc ion o he English pas ense du ing s o y elling ac i i ies. Howe e , as discussed in sec ion 2.1, ou pu -pushing CF can ake many o ms, which is why we compa ed he mo e implici elici a ion wi h he mo e explici me alinguis ic eedback. Indeed, we ound ha hese wo ypes o p omp s appea ed o a ec s uden s’ pas ense usage di e en ly, as bo h he immedia e and delayed e ec s o me alinguis ic eedback u ned ou o be g ea e han hose o elici a ion, al hough no s a is ically signi ican di e ence was ound o he gains made om p e- o pos - es . Since bo h CF- echniques push lea ne s o sel -co ec in a simila way, i can be assumed ha he obse ed di e ences a e ela ed o he deg ee o explici ness o each CF- ype. Looking back o he analysis o up ake and epai , we no e ha s uden s a emp ed o espond o almos all o he elici a ion- and me alinguis ic eedback-mo es, which indica es ha hey no iced he eedback. Up ake ollowing ecas s, which akes he o m o a epe i ion o he ecas , has been ound o be no indica o o subsequen acquisi ion (Mackey & Philp, 1998). Howe e , up ake a e ou pu -pushing eedback usually esul s in modi ied ou pu , which can play an impo an ole in language de elopmen acco ding o Swain’s (1995, 2005) ou pu hypo hesis. In suppo o his hypo hesis, McDonough (2005) concluded ha he oppo uni y o modi y one’s speech was he bes p edic o o ELIA 16, 2016, pp. 49-78 68 DOI: h p://dx.doi.o g/10.12795/elia.2016.i16.03 Compa ing he e ec s o wo ypes o ou pu -pushing eedback... acquisi ion in he s udy on he impac o CF on he acquisi ion o ques ion o ms. Gi en ha no di e ences we e obse ed be ween bo h g oups’ pas ense knowledge o o al p oduc ion on he p e- es , i is s iking ha me alinguis ic eedback esul ed in a high a e o epai , whe eas elici a ion only esul ed in he use o he co ec o m abou hal o he ime. To unde s and his inding, we need o ake a close look a speci ic ins ances o lea ne up ake in bo h g oups. In he case o elici a ion, i appea s ha he e we e 22 cases in which s uden s esponded o he CF-mo e by simply epea ing wha hey had said be o e, as in he example below. (4) S: Bu she go o house T: Bu she…? S: Bu she go o house o 11 o’clock The s uden in his example may ha e hough he eache had no hea d he u e ance clea ly, and hus he co ec i e in en o he elici a ion-mo e seems o ha e gone unno iced. Twen y- wo o he 77 elici a ion-mo es leading o up ake simply con ained a epe i ion o he ini ial o m. In se en o he cases, s uden s modi ied hei ini ial u e ance, bu hey seemed o hink he p oblem was ela ed o ocabula y, as in (5): (5) S: And a li le man a i e… T: A li le man…? S: Go He e he s uden is simply using ano he e b, ins ead o he pas ense o “a i e.” Finally, he e we e wo cases in which s uden s we e clea ly a a loss as o he eache ’s in en ion: (6) S: When he p ince climb up he owe … T: And when he p ince…? ELIA 16, 2016, pp. 49-78 69 Hanne Roo hoo DOI: h p://dx.doi.o g/10.12795/elia.2016.i16.03 S: Wha ? In he case o me alinguis ic eedback, he eache is ob iously e e ing o a g amma ical p oblem, as she explici ly s a es ha he s uden s need o use pas ense. This may explain why he e we e only 6 cases o up ake wi hou epai in he me alinguis ic eedback g oup. In h ee o hese cases, s uden s seemed o misunde s and he in en ion o he eache , as in example (7): (7) S: She had one lis o names bu i ’s inco ec . T: In he pas ? S: don’ ? The e was also one case in which a s uden chose a di e en e b, as in example 5 in he elici a ion g oup. The wo o he cases o up ake wi hou epai in he me alinguis ic eedback g oup occu ed a e i egula e bs and hey esul ed in o e gene aliza ion o he egula e b ule, as in (9): (9) S: A e his he… Cinde ella hea he clock T: In he pas S: Hea ed, no sí, hea ed he clock. O e gene aliza ion o he –ed ending o i egula e bs also happened once in he elici a ion-g oup: (10) S: wan … o un he all and he’s die T: he…? S: he alled O e all, i seems ha he low a e o epai a e elici a ion is no due o a wo se command o he pas ense in he elici a ion g oup, bu a he o a lack o unde s anding o he eache ’s in en ion. ELIA 16, 2016, pp. 49-78 70 DOI: h p://dx.doi.o g/10.12795/elia.2016.i16.03 Compa ing he e ec s o wo ypes o ou pu -pushing eedback... The g ea e explici ness o me alinguis ic eedback p obably also explains he inding ha s uden s in he me alinguis ic eedback g oup p oduced a ewe pas ense e o s han hose in he elici a ion g oup (see 4.1), e en hough he e we e no signi ican di e ences be ween he wo g oups on he p e- es . Since he o he s uden s we e awa e ha he eache was p o iding eedback on hei classma es’ pas ense usage, i is likely ha hey s a ed ocusing ha de on ge ing he pas ense o ms igh . E en hough me alinguis ic eedback was mo e success ul a p omp ing s uden s o sel -co ec hei pas ense e o s du ing he s o y elling asks, his does no mean me alinguis ic eedback is mo e e ec i e a p omo ing long- e m imp o emen o lea ning. As men ioned abo e, se e al au ho s ha e poin ed ou ha up ake and epai do no p edic subsequen acquisi ion, especially in he case o ecas s (Goo & Mackey, 2013; Mackey & Philp, 1998). Indeed, no wi hs anding he low a e o epai in he elici a ion g oup, bo h he elici a ion and me alinguis ic eedback g oup made signi ican imp o emen s in hei accu a e o al p oduc ion o pas ense o ms on he pos - es . On he whole, howe e , he s uden s in he me alinguis ic eedback g oup appa en ly imp o ed mo e, al hough he di e ence was no qui e s a is ically signi ican and he esul s a e based on a small sample size. 6. Conclusion The esul s o his small-scale s udy on he e ec s o ou pu -pushing CF on s uden s’ accu a e p oduc ion o he pas simple ense du ing s o y elling asks indica e ha me alinguis ic eedback may be mo e e ec i e han elici a ion. As men ioned in sec ion 5, we a ibu e his inding o he mo e explici na u e o he o me CF ype. One possible eason why explici p omp s could wo k be e han mo e implici ou pu -pushing CF is ha i is easie o s uden s o no ice hei co ec i e in en , since language lea ning canno ake place wi hou a ce ain deg ee o no icing (Schmid , 2001). Indeed, as shown in sec ion 5, he s uden s in his s udy did no always in e p e elici a ion as being di ec ed a pas ense usage. ELIA 16, 2016, pp. 49-78 71 Hanne Roo hoo DOI: h p://dx.doi.o g/10.12795/elia.2016.i16.03 Fu he mo e, i is possible ha me alinguis ic eedback was mo e success ul wi h hese s uden s because o hei ela i ely low le el. As hey we e in e media e le el s uden s who had no had many oppo uni ies o pu ing hei heo e ical knowledge in o p ac ice, he mo e explici CF was p obably mo e in line wi h hei de elopmen al le el. Simila ly, Aljaa eh and Lan ol (1994) posi ha mo e explici o ms o CF could be mo e use ul o low-le el lea ne s, while highe -le el lea ne s may be be e able o egula e hei pe o mance in he o eign language and hus mo e implici CF migh be su icien o hem. Amma and Spada (2006) also ound p omp s, which a e gene ally mo e explici , o be mo e e ec i e o low-p o iciency lea ne s han mo e implici ecas s. Ob iously, u he compa a i e s udies on he e ec s o p omp ype on s uden s a di e en le els o p o iciency a e wa an ed o co obo a e hese claims. Ano he possible eason o he obse ed success o me alinguis ic eedback can be ela ed o skill-acquisi ion heo y (e.g., DeKeyse , 2007), since his heo y posi s ha explici knowledge can be au oma ized and p ocedu alized h ough meaning ul p ac ice. The esul s o he g amma icali y judgmen p e- es show ha he s uden s gene ally had a ai le el o explici knowledge o pas ense o ms, which was ac i a ed h ough he s o y elling ac i i ies and he eache s’ p omp s. E en hough his s udy shows explici p omp s o be mo e e ec i e, Li (2010) and o he esea che s ha e sugges ed ha he e ec s o implici eedback ake longe o mani es hemsel es. In Li’s (2010) me a-analysis o 33 expe imen al CF-s udies, he au ho ound ha ecas s had g ea e e ec s on he acquisi ion o second language g amma on he delayed pos - es , in compa ison wi h mo e explici eedback- ypes. Since his s udy only las ed wo weeks and no delayed pos - es could be pe o med, we we e no able o e i y whe he his is also he case o implici p omp s, such as elici a ion. The e o e, u he esea ch on he long- e m e ec s o elici a ion and o he mo e implici p omp s needs o be ca ied ou . Ano he impo an ca ea conce ns he indi idual di e ences be ween s uden s. As can be seen in he esul s sec ion, no all o he s uden s imp o ed hei o al p oduc ion o he pas ense in he same way. ELIA 16, 2016, pp. 49-78 72 DOI: h p://dx.doi.o g/10.12795/elia.2016.i16.03 Compa ing he e ec s o wo ypes o ou pu -pushing eedback... E en hough he s uden s we e o a compa able le el and no signi ican g oup di e ences we e ound be ween hei explici knowledge o he pas ense and hei accu a e p oduc ion o he a ge s uc u e on he p e- es , one o he elici a ion-g oup s uden s in ac dec eased he accu acy sco e by 9%, whe eas gain sco es in he me alinguis ic eedback g oup anged om 6.2 o 51.2%. Such an impo an a ia ion in s uden s’ pe o mance canno be igno ed. To accoun o hese di e ences, u u e s udies need o ake in o accoun s uden ac o s such as p o iciency, wo king memo y and mo i a ion, amongs o he s. Finally, we acknowledge ha his s udy has se e al limi a ions. Fi s o all, i was a small-scale s udy, in which only a limi ed numbe o he ini ial pa icipan s we e p esen h oughou he en i e expe imen . Second, no con ol g oup in which s uden s did no ecei e eedback was used and he e o e i is possible ha he obse ed p e- o pos - es gains we e me ely due o he ac ha s uden s had he oppo uni y o p ac ice he pas ense du ing he ea men . Thi d, only one pos - es was ca ied ou h ee days a e he second ea men session. As men ioned abo e, a delayed pos - es could ha e gi en us mo e in o ma ion abou he long- e m e ec s o he ea men . Despi e i s limi a ions, we belie e ha his s udy p o ides impo an insigh s in o s uden s’ eac ions o wo unde - esea ched ypes o o al co ec i e eedback. Mo eo e , i p o es ha i is necessa y o s udy di e en ypes o p omp s sepa a ely, since di e ences we e ound be ween explici and implici p omp s. Re e ences Acklam, R., & C ace, A. (2005). To al English. Ha low, UK: Pea son Educa ion. Aljaa eh, A., & Lan ol J.P. (1994). Nega i e eedback as egula ion and second language lea ning in he zone o p oximal de elopmen . The Mode n Language Jou nal, 87, 465-483. h ps://doi.o g/10.1111/j.1540-4781.1994. b02064.x ELIA 16, 2016, pp. 49-78 73 Hanne Roo hoo DOI: h p://dx.doi.o g/10.12795/elia.2016.i16.03 Amma , A., & Spada, N. (2006). One size i s all? Recas s, p omp s and L2 lea ning. S udies in Second Language Acquisi ion, 28 (4), 543–574. h ps://doi.o g/10.1017/s0272263106060268 Ba do i-Ha lig, K. (2000). Tense and Aspec in Second Language Acquisi ion: Fo m, Meaning and Use. Ox o d, UK: Blackwell. de Bo , K. (1996). The psycholinguis ics o he ou pu hypo hesis. Language Lea ning, 46, 529 – 555. h ps://doi.o g/10.1111/j.1467-1770.1996. b01246.x DeKeyse , R. (2007). In oduc ion. Si ua ing he concep o p ac ice. In R.M. DeKeyse (Ed.), P ac ice in a second language. Pe spec i es om applied linguis ics and cogni i e psychology (pp.1-18). Camb idge, UK: Camb idge Uni e si y P ess. h ps://doi.o g/10.1017/ CBO9780511667275.002 Dilans, G. (2010). Co ec i e eedback and L2 ocabula y de elopmen : p omp s and ecas s in he adul ESL class oom. The Canadian Mode n Language Re iew, 66 (6), 787-815. h ps://doi.o g/10.3138/cml .66.6.787 Dough y, C., & Va ela, E. (1998). Communica i e ocus on o m. In C. Dough y & J. Williams (Eds.), Focus on o m in class oom second language acquisi ion (pp. 114–138). New Yo k, NY: Camb idge Uni e si y P ess. Dulay, H., & Bu , M. (1974) Na u al sequences in child second language acquisi ion. Language Lea ning, 24, 37-53. h ps://doi.o g/10.1111/j.1467-1770.1974. b00234.x Ellis, R. (2007). The di e en ial e ec s o co ec i e eedback on wo g amma ical s uc u es. In A. Mackey (ed.) Con e sa ional In e ac ion and Second Language Acquisi ion: A Collec ion o Empi ical S udies. (pp. 339–360). Ox o d, UK: Ox o d Uni e si y P ess. Ellis, R., Loewen, S., & E lam, R. (2006). Implici and explici nega i e eedback and he acquisi ion o L2 g amma . S udies in Second Language Acquisi ion, 28, 339 – 368. h ps://doi.o g/10.1017/S0272263106060141 Gass, S., & Mackey, A. (2015). Inpu , in e ac ion and ou pu in second language acquisi ion. In B. VanPa en & J. Williams (eds.). Theo ies in Second