A preliminary study on supportive feedback strategies in language education

George S. YpsilandisA preliminary study on supportivefeedback strategies in languageeducationIntroductionThe term feedback is probably the most known

among a number of other terms used in theliterature (operand conditioning, reinforcementor evidence) to describe “a reaction, aresponse that is usually triggered and receivedby the learner and provided by the teacher”, atleast in most cases of educational environments(Ypsilandis, 2002:171). As early as 1959MacDonald described feedback as a “controlmechanism for signalling him (the learner) thathis responding is bringing him toward or takinghim away from his goal”(p.78). This concept offeedback is generally accepted by mostresearches as an error correction andhypothesis testing mechanism “so that the erroris eliminated from further production” or toplace it in Long’s (1998:16) words “…explicitnegative feedback, i.e. error ‘correction,’etc”. Although at early stages feedback wasprovided in the form of yes or no (or othermethods of reward and punishment), a number ofdifferent corrective feedback types have beensuggested by different scholars in thebibliography, such as, recasts, expansion(Brown and Bellugi, 1964), repetition,backchannel cues, metalinguistic, requests forclarifications, elicitation or negotiation (toname a few). These are among those discussed indetail in Psaltou (2004) and Ypsilandis (2006).The impact to learning for some of thesefeedback types has been tested by experimentalor empirical research since the beginning ofthe programmed instruction behaviouristSkinnerian theories, initially with not soconvincing results but lately with moreconcrete ones. This is reviewed in short below,in section 2.1. Similarly, a number ofadjectives have also been added to the nounfeedback, such as immediate (Skinner, 1954),delayed, corrective, formal or informal, frequent (Gagne,1977), direct and indirect (Hendrickson (1978) andincidental (Hounsel, 1999). All the above feedback

types require student active participation inthe form of performance or response to astimuli or a task. It may thus be understoodthat, this corrective mechanism operates at aproductive stage of the learning process. Although the aim of feedback provision has

always been to support learning, from thebehaviourist era, the term supportive feedback hasinitially been advocated in Ypsilandis (2002)and initially tested in Ypsilandis (2006) todescribe a different concept. Supportivefeedback, does not aim to reward or correct andrepair the performance of the learner butrather to assist the learning process at a muchearlier stage. It aims to offer support for thecomprehension of learning material (input) andfurther aid retention, and as such, it has adifferent type of operation than its correctivenamesake. Supportive feedback does not requireany output from the part of the student. Thestudent expresses his need to receiveclarifications upon initial input and theteacher, the software, a co-learner or alistener supports him by providing explanations(supportive feedback). By that token, theprovision of supportive feedback to studentinquiries is not a new concept all together.Rather, it is a typical teacher reaction tostudents’ inquiries which is provided in theform of clarifications to new input. Thisconcept of feedback is beginning to receivesome attention in the psychology of learning(Menezes de Oliveira e Paiva, 2003) and it isdiscussed in more detail in section 2 of thispaper. It is at this point that the mainquestions for this preliminary study are asked:a) what are the best supportive feedbackprovision strategies that would assist shortand long term learning? b) Are all thesupportive feedback strategies equallybeneficial for all learners irrespective oftheir cognitive or learning style? c) Do allfeedback provision strategies equally engagethe learner in the learning process? In mostempirical research studies of correctivefeedback, student personality factors, such asthe cognitive or learning style of thelearners, are not taken into consideration. Itis my view that these personal characteristics

may be worth tested as it is probable thatinteresting liaisons may be uncovered of onefeedback type over another. The present study, which contains several

pilots, was undertaken to investigate successto the learner’s short and long-term memory(dependent variable), a number of differentteaching strategies (independent variabletested) and the role of the learners’ cognitiveand learning style (intervening variabletested). The intervening variable here is takeninto consideration to offer a better, morerefined understanding about the relationshipbetween the two main variables. This paperinitially offers a short presentation of thenotion of feedback and discusses findings ofexperimental or empirical research mainly oncorrective and less on supportive feedback. Inthe Method section the tools and procedurescreated and followed to test our hypothesis arepresented and the analysis of results. At thelast (Discussion) section, findings are seen inrelation to existing knowledge on the subject.The research questions aim to address the

supportive feedback issue both quantitativelyand qualitatively, i.e. measure theeffectiveness of the specific strategies bothin terms of volume (number of words retained),and number of students engaging in the learningprocesses in relation to the learners’ shortand long term memory. By that token thisresearch may add to the discussion on thetopic, by providing new data resulting from theexperiments conducted, and supply usefulinsights to the language teacher or thedesigner of language teaching software.Findings of this study and in particularidentified patterns of successful supportivefeedback strategies would indirectly be able toprovide information about human brain functionsrelated to language learning and as such, addto the discussion on human brain operations andfurther lead to a more refined research on thetopic.

FeedbackIn a short definition provided by McDonald

(1959:75) of the notion of feedback, as a “linkbetween an information source and an

information user” one could identify the twomajor points related to the behaviouristunderstanding of education, which seem to bereappearing in most definitions provided later,i.e. the ‘information source’ which is theteacher and the ‘information user’ which is thelearner. The idea behind this is that learnerscan only learn from teachers, an assumptionthat was questioned much later. Incommunication theory this link develops as areaction from a receiver (not necessarily ateacher) which helps the sender of a message“to check on the efficiency of hiscommunication” (Crystal 1985:119) or “theresponse to efforts by the learner tocommunicate” (Ellis, 1985:296). Although thereceiver is then understood not to be theteacher the notion of feedback still works as acorrective and testing hypothesis device.Feedback, in the related literature, is oftendiscussed through a set of binary divisions.

The positive / negative division. Behaviourist feedbackoperated in the forms of reward and punishmentwhich was later developed in parallel aspositive and negative feedback. Positivefeedback would then be understood as a reactionthat would encourage someone to repeat orexpand on an action. The problem arises when anaction is not correct, accepted or expected.By contrast then, negative feedback would beany information that would aim to errorcorrection. According to (Chaudron, 1988:150)this is delivered in different layers: a) thatof informing “the learner of the fact oferror”, b) the next which is “to elicit arevised response,” and finally c) to assurethat the error is corrected and does notreappear in the future. In all the above layersof corrective negative feedback it becomesapparent that: a) it requires an initial outputfrom the part of the learner (productivestage), b) it is provided predominantly by theteacher (or an external source) and c) it aimsat noticing, and then at short and long termrepetition or correction. It may be concludedat this stage that the target of positive ornegative feedback is purely pedagogical as thefinal objective is to support the learningprocess. This division is discussed in more

detail in relation to theoretical stances andempirical research findings in section 2.1 andattempts to answer the question as to what kindof feedback should one offer.

The internal / external division. It becomes clear frommost references of feedback that this istypically provided by an external source, i.e.the teacher, a co-speaker, a listener or amachine/computer (external feedback). Internalfeedback may by contrast then be thought of asan internal mechanism of the individual whichself-monitors his outcome. Evidence for theexistence of this mechanism could be the factthat many people correct themselvesautomatically when they speak without anyexternal notice or when they write (rephrase oredit). External corrective feedback, as it isunderstood by most researchers, leaves littlespace for the internal mechanisms of thelearner to monitor and reformulate his initialhypotheses and outcomes, although theseinternal procedures exist in all individualsand seem to be triggered automatically insidethe learners’ brains (read results of Holleyand King, 1971 in 2.1). The fact that theseinternal feedback (to add another term to thealready existed long list) potentials of thelearner were not taken into consideration isnot coincidental as, the essence of thepedagogy of language learning for years wasbased on an understanding of the teacher (andlater the native speaker) as the only andunquestionable source of information in thelanguage classroom. Today, learners areaccepted to be able to learn by themselves fromstructured or authentic materials by decodingand further by comparison with previousknowledge on the subject, although lowerlanguage level learners are claimed to lacksufficient knowledge for self-correction(Brown, 1991 and the other citations reportedin Ferris, 2006). The ability of self-correction is not a question that can beanswered in a binary (yes/no) approach as thereare other variables that may be intervening,some of which could be previous linguisticexperiences, the cognitive and the learningstyle or the age of the learner. Inferencingand success of decoding may be reasonably

expected to increase as the language level andlanguage learning experiences grow and thusbecome more effective at later stages oflanguage instruction or acquisition. The skilland the process of inferencing, which revealsthe full linguistic potential of the learnerand which is “automatically activated in thehuman brain”, is a strong indication of thepotential of internal feedback (Tsopanoglou,Ypsilandis and Mouti, 2014:47). The internal /external division refers to the question ofwho should provide feedback and brings forwardthe next division of direct and indirectfeedback which builds on the potential of thelearner for self-correction.

The direct / indirect division. This division isrelated to the approach of feedback provision.The terms suggested by Hendrickson (1978) havebeen given different content by differentscholars. Direct feedback may be understood asa straight and immediate correction mechanismin the hands of the teacher in various explicitmodes, while its opposite (indirect) operatesas an indicator of a flaw without anyspecification on the error (for more on thisdivision, see Ferris 2005). Indirect feedbackapplies to the internal capacities of thelearner for error correction (internalfeedback). The learner at this instance useshis decoding skills and prior knowledge on thesubject, his knowledge of the target language,or his general knowledge of languages. Self-correction could be activated eitherautomatically or it could be called toattention by an external source (guided).Indirect feedback may thus be assumed as amechanism guiding the learner to activate hisinternal feedback capacities for self-correction. Finally, Ferris (2006) and thecitations within suggested the quality of errortype to be a criterion or a determinant fordirect or indirect feedback, while errorgravity (weighting) and cognitive and learningstyle of the individual could be more criteriaadded to this list.

The explicit / implicit division. This division can beseen in relation to the two umbrella areasoperating in language pedagogy, i.e. ForeignLanguage Learning (FLL) and Second Language

Acquisition (SLA). By definition these twoareas see feedback in two different modes. FLLfocuses on the explicit feedback (typicallyprovided by the teacher) which is expected toenhance the learning processes. SLA on theother hand proclaims implicit feedback toassist acquisition, which could be offered byany external source, possibly a co-speaker. Theposition scholars have undertaken on thisdivision are in some cases connected to theirinitial stance in language pedagogy. Long(1996) takes a middle stance and argues forenvironmental input in SLA and suggests theterms positive and negative evidence (recasts,repetitions, clarification requests, facialexpressions, etc.) which on the surface lookparallels to the terms positive and negativefeedback. Long makes this distinction as in SLAthere is not a typical teacher / learnerclassroom situation and as a result no formalteaching and learning. In his positive evidencehowever, there are no rewards but authentic ormodified (simplified or elaborated) acceptablelanguage which works by providing the speakerwith models of acceptable utterances andprobably acts as a reinforcement device.Negative evidence operates by offering director indirect information about what isunacceptable. In Long (1991) and later in Longand Robinshon (1998,) a model of reactivenegative evidence follows the explicit /implicit division with overt error correction(explicit) and communication breakdown orrecasts (simple or complex) for implicit. Thisis part of Long’s (1991) and Long andRobinson’s (1998) theory of focus on forminstruction (read also Pool, 2005). Thedivision here is related with the feedback typeto be chosen.

The corrective / supportive division. As pointed outearlier the external corrective feedback (CF)mechanism aims to the noticing of a flaw, theimmediate correction by repetition of anacceptable / expected version and to a morepermanent long term correction of unacceptablelanguage. Supportive feedback (SF) differssignificantly by CF in a number of ways: a) SFaims to assist the learning process withouthaving to correct or require an outcome from

the learner, b) it operates at a much earlierstage of the learning process, that ofcomprehension, and can be designed or besupplied with initial input, c) it does notreward or punish, d) it has a pedagogical aim,to assist the learning process by both quantity(increase the numbers of learners who learnsuccessfully or increase the amount of materiallearned from initial input) and quality (make abetter use of acquired knowledge in practiceand keep acquired knowledge in the long termmemory). As argued above SF could also be aimedto be guided and be supplied internally, andfurther, it could be provided implicitly(aiming at inferencing) or explicitly (withmorhosyntactic information). Finally, it couldbe presented with a direct strategy approach orwith an engaging developmental method (seeDesign3.1).It seems necessary at this point to look at

some of the findings of empirical orexperimental research on the topic, whichconcentrate on the corrective feedbackimplementation hypothesis, as research onsupportive feedback is comparatively less andfocusses on what researchers in the past calledclarifications.

Corrective feedback and researchDespite claims in the literature (Cathcart

and Olson, 1976, Kaufman 1991) of studentsdesiring to receive corrective feedback,research results are not conclusive as theyhave produced mixed views of the effectcorrection has on learners’ errors. The oldestview is rather discouraging. Dulay, Burt andKrashen (1982:35) find student correction to be“immensely frustrating” and further presentsome empirical studies to support their claim,which are offered below in the next paragraph.This claim is also supported by Edge (1989) whostates that overt correction may discouragelearning steps by prohibiting students fromexperimenting with the language and work outnew and better ways of saying things. Finally,Chomsky (1975) in his nativist theory alsoadvocates against negative evidence and claimsthat any changes resulting from negativeevidence would be on language behaviour and not

on the learner’s grammars, which are the resultof positive linguistic evidence. In a study by Cohen and Robbins (1976), error

correction of written university student workswas found not to have influenced errorproduction. The authors attributed this findingto lack of quality and systematicity of thecorrections, supported also by Allwright’s(1975) and Fanselow’s (1977) observations. In amore systematic study by Hendrickson (1977) theresult was however the same. The author did notfind any significant correlations in students’written proficiency between the systematiccorrective feedback (all errors or a systematicselective correction) and their correct use.The same conclusion is reached by Plann (1977)in speech grammatical and morphological errors(comparison between oral corrections withrepetition and formal teaching withexplanations on the blackboard accompanied bywritten exercises and drills). In the abovestudies: a) there was not any distinction toerror, i.e. between mistakes (lapsus calami)and errors (lapsus liguae) (the latin termswere originally suggested by Pertounias, E. ina discussion on errors), that is, betweendeviant sentences as a result of processinglimitations or speed (mistakes) and lack ofknowledge (errors), b) errors were ofphonological or syntactical origin (languageform), c) errors were corrected by an externalsource (the teacher) and little attention waspaid to the internal corrective mechanisms ofthe learner, and d) Dulay, Burt and Krashen’s,Edge’s and Chomsky’s (to a certain extend)comments were related to the psychologicalstate (pressure) on learner while the otherstudies concentrated on the learning outcome.The strongest opposition against correctivefeedback on grammar was expressed by Truscott(1996) as reported in Ferris, 2006), whoexpresses the view that the time spent oncorrections in writing classes is indeedharmful, as it reduces the time and energy thatcould be allocated on more important concernsrelated to writing. In a study focussing oncontent correction by Holley and King (1971)without any correction on form, it was foundthat, in more than half of the cases of the

sample used for the study, participantscorrected themselves and improved theirperformance by being given uninterrupted timeto rephrase their responses (is this aninstance of successful internal feedback?).These studies and in particular the latter byHolley and Kind (1971) advocate againstcorrective feedback or to no feedback provisionat all while the last study leaves space forinternal feedback to be further tested. Rebuttals to the above claims in more recent

studies on corrective feedback (negativeevidence) theoretically relate to: a) thenoticing hypothesis (external in this case) whichis claimed to function as an attention gettingdevice of the mismatch between learner outcomeand input (Gass, 1991:136). This hypothesismaintains that noticing is required from thepart of the learner for effective learning,which is triggered by corrective feedback, andleads to a modification of existing L2 (Gassand Varonis, 1994:299), while lack of thisexternal signalling would make “nothing in thetarget language…” to be available “…for intakeinto a language learner’s existing systemunless it is consciously noticed.” Ellis (1991)claims that this procedure includes threesteps, noticing, comparing, and integrating, b)the testing model hypothesis of acquisition by whichthe learner is believed to formulate hypothesesabout the TL which he further tests vis a visthe target norm. This seems to be an internallytriggered possibly automatic, action whichcomes in support of Mc Donald’s (1959:78)initial claim by which he declared the learnerto be an information-processing, goal directed,organism who wishes to achieve certaindesirable states and uses feedback to achievehis personal goals. The testing modelhypothesis denotes the learners’ internal willand active participation in the learningprocess while the former (noticing) implies anexternal more conscious interference. Ohta(2001) argues that if the correct form isprovided, with corrective feedback, learnersare lead to comparison and hypothesis testing,while if the correct form is not providedlearners are directed to use their own internalresources. Advocates of the above theories

believe that corrective feedback has a crucialrole to play in second language acquisition(Bley-Vroman, 1986, 1989) while frequentfeedback can be of “considerable value inincreasing the efficiency of learning” (Gagne,1977:298). In support of negative evidence forthe benefit of learning is also White(1991:133) who comes to a conclusion by anexperimental study that “form-focused classroominstruction, including negative evidence, ismore effective in helping L2 learners to arriveat the appropriate properties of English thanpositive input alone.”Several corrective feedback types have been

tested either empirically or experimentallywith positive outcomes: a) students improvetheir accuracy in L2 by receiving form-focussedinstruction and corrective feedback (Lightbownand Spada, 1990) or positive / negativeevidence (White 1991), b) explicit and implicitfeedback types were found to be beneficial andboth leading to the learning of abstractlinguistic generalizations (Caroll and Swain1993) with the latter (providing the learnerwith implicit metalinguistic information),scoring higher than the explicit, c) recastswere found to play a positive role in L2learning (Brock, Crookes, Day and Long, 1986;Long, Inagaki and Ortega, 1998; Lyster, 2001),d) immediate provision of feedback assists tothe abandonment of the wrong form and theacquisition of the correct version (Nicholas,Lightbown and Spada, 2001), e) intensiverecasts were found to present greater increaseat learners of higher developmental levels,than learners who participated in interactionwithout intensive recasts (Mackey and Philip1998), f) fine-tuned feedback is moresuccessful than teachers corrective feedback(Han, 2001), and g) recasts were found to bethe most widely used form of feedback of thegroup teachers investigated in a study byLyster and Rada (1997).

Supportive feedback and researchResearch on supportive feedback is

comparatively less. It may be worthwhile tomention two types of studies, the first underthe term clarifications and the second with the

label supportive feedback. Laboratory andclassroom studies on clarifications arereported in detail in DeKeyser (2006) wherefrom some of the studies are reviewed andpresented in short. Implicit and explicit explanations of new L2

material were studied in a laboratory context.One of the earliest studies was N. Ellis (1993)with three different types of clarifications tothree different groups. It was found that, thegroup which was offered implicit explanations,in the form or numerous random examples of thephonological phenomenon, was the fastest tojudge the well-formedness of sentences seenbefore, while it was the slowest to generalizethis knowledge in new sentences. The secondgroup, which received explicit ruleexplanations, revealed knowledge of the rulesbut little ability to apply them in practice.The third group, which received both explicitrule explanations and the same random examplesas the first group, performed well in explicitrule knowledge and grammaticality judgements.It was the explicit rule treatment thatoutperformed the other two groups in showingawareness of how rules apply to examples. In analmost analogous experiment Michas and Berry(1994) and de Graaf (1997), come to the sameconclusions of the advantages of explicit rulepresentation in the pronunciation of Greek andeXperanto words by English natives. The sameconclusion is reached by Alanen (1995) andDeKeyser (1995) in favour of the explicit ruleexplanation treatment group. Classrooms studies of explicit / implicit

feedback treatment, reported in DeKeyser(2006), are very few, namely Scott (1989,1990), VanPatten and Oikkenon (1996) and Levin(1969). In Scott’s (1989, 1990) studies theexplicit rule groups registered a significantadvantage. Similar findings were witnessed inVanPatter and Oikkenon (1996) although scoreswere not significantly different. In Levin’s(1969) study no difference between the twotreatments is found although the explicitmethod was registered superior at all the otherintervening variables of the study (age,proficiency and aptitude levels). For moredetails and more experimental and classroom

studies, read (DeKeyser, 2006). In Ypsilandis(2006), two supportive feedback strategies arepresented and experimentally tested. In hisstudy Ypsilandis argues that electronicallyhyperlinked supportive feedback, provingmorpho-syntactic information with the use ofoff-screen tasks, seems to assist short andlong-term memory. It may be concluded so farthat the explicit rule clarifications assupportive feedback and negative feedback as acorrective mechanism seem to be the mostsuccessive tools for feedback provision.

MethodDesign This preliminary brief study is part of a

large longitudinal research which involvesthree repetitive experiments conducted in 2004,2005, and 2006, all following the same designand procedure with similar tools (texts),investigating the effectiveness of feedbackprovision strategies as a support mechanism forshort and long-term memory. In all experimentsthere is one group of subjects treated with thesame tool aiming to offer supportive feedbackon vocabulary items. Feedback strategies aretested in twos, the control strategy being theone typically used in most language learningeducational sites or software (in the past),providing word definition and morphologicalinformation in the target language (English),and the equivalent term in the subjects’ mothertongue. This is a supportive feedback strategyused in CDs dedicated language learning in thepast, when the tool was designed for a localmarket. Today, most dedicated language learningsites use the same strategy and either presentinformation in the target language or throughthe support of several languages where from thelearner could select the one that better suitshim. The experimental supportive feedbackstrategies tested in various repetitiveexperiments involved: a) a successiveapproximations strategy attempting to engagethe learner in a hybrid discussion (thisstrategy is proclaimed to be beneficial byKeislar and McNeil (1962) in contract to theresponse easily demonstrated), b) a successiveapproximation strategy aiming at inferencing

from the part of the learner, c) a keywordtarget language / mother tongue method ofvocabulary provision accompanied with targetlanguage non-linguistic image associations, d)awareness and explicit rules with memorytechniques provided (employing synonyms,antonyms, morphology), e) Inferencing througheffort of processing techniques, f) ameaningfulness of processing technique withpersonal significance, g) inferencing inmultiple choice format, h) an off screen task,i) definition and morphological analysis plusone of the following vocabulary teachingstrategies (sound association, effort ofprocessing with personal experiences andmeaningfulness of processing). Some of theseresults were presented very briefly at aconference keynote in Greece (Ypsilandis,2006). These supportive provision strategieswere hypertext and linked to a number of wordsequally and evenly distributed in anelectronically presented text. Short and long-term memory is measured by retention of thevocabulary item after one hour and after oneweek. Liaisons between success of strategy andsubjects’ cognitive and learning style,preferred feedback strategy, opinions of bestfeedback strategy for learning were alsoregistered and tested. Cognitive and learningstyles of subjects were decoded through the useof relevant questionnaires translated in thesubjects’ mother tongue (those in Brown, 1994;Richard and Lockhard, 1996; Cohen and Oxford,2001). Strategies from (a) to (h) were used inthe first case, (h) used tested in the secondcase, and (i) was examined in the third case. Asignalling / attention device was also testedin one of the cases to measure whether thatwould have an effect on learning outcomes.

ProcedureThe procedure was divided in different stages:a) The reconnaissance stage, in which it was

attempted to locate, 1) the CS and the LS ofthe subjects involved, and 2) the list ofvocabulary items (List A) provided to thesubjects in order for them to provide theirequivalent in their mother tongue. Proceedingin this manner allowed us to identify items

that were known to the subjects prior to thisexperiment. In this case the word waseliminated and subsequently not considered inthe study. b) The treatment stage, in which subjects

read a text (taken from a textbook of Englishfrom a known publisher) which appeared on acomputer screen together with the hypertextedsupportive feedback provision strategies. Inthis text the words from list A appearedtogether with supportive explanatory feedbackfor each word. Each time half the words weresupported by traditionally provided feedbackwhile the other half with the experimental.There were no time constraints.c) The short memory stage, in which (after an

hour) subjects were asked to provide theequivalent, in their mother tongue of, as manyas they could remember, vocabulary items of thesame list used at the reconnaissance stage. d) The preference stage which followed

immediately the short memory stage, in whichsubjects were asked to answer introspectivequestions with the use a small questionnaireregistering their learning preferences ofsupportive feedback, and their views of bestfeedback provision strategy for learning.e) The long term memory stage (after two

weeks), in which subjects were asked to providethe equivalent, in their mother tongue of, asmany as they could remember, vocabulary itemsof the same list used at the reconnaissance andshort term memory stage.

Apparatus and MaterialsThere were two types of tools that were used

in this study, which are classified into twocategories according to the technological meansthese were provided. Thus, in no-technology,paper format tools there were, list A ofvocabulary items where subjects wereregistering the equivalents in their mothertongue before and after the treatment period,the questionnaires aiming to locate theparticipants’ cognitive and learning style,offered in Brown (1994), Richard and Lockhard(1996) and Cohen and Oxford (2001), and theintrospective questionnaire registering thesubjects’ preferences in receiving supportive

feedback together with their views of bestsupportive feedback strategy for learning. Ina technologically more advanced tool there wasan html document providing a different text ineach of the three cases, containing the textwith the 48 hyper-texted words with supportivefeedback on selected vocabulary items.

The subjectsThe subjects were all university students of

the foreign languages departments of theAristotle University of Thessaloniki. A totalof 66 subjects were involved in all threeexperiments, 40 men and 26 women. 34 subjectsin the first experiment, 17 subjects in thesecond case, and 15 subjects in the lastexperiment. Their age varied between 19 and 41years.

AnalysisA descriptive analysis is presented in 5

sections offering scores in the form of actualnumbers and percentages. Initially the actualnumber of words retained after an hour andafter one week is presented for all threeexperiments, findings in relation to subjects’individual scores are following, results ofcomparisons between the variables (traditionaland experimental supportive feedbackstrategies) are later offered, scores relatedto the cognitive and learning style of theparticipants and the subjects’ learningpreferences are finally discussed.

1. Scores of the independent variables tested. Thefollowing table presents the results of wordsretained after one hour and after one week. Thefirst column shows the study case (experiment),the second column presents the supportivefeedback strategy tested, the third columngives the 100% of word occurrences, i.e. words(minus known words) multiplied by number ofsubjects participated in each study, the fourthand fifth columns show the words retained afterone hour and after one week together with therelevant percentage from the total.

CASE MODE WORDS ONEHOUR

ONEWEEK

1st

Experiment

Traditional

269(100%)

96(35.6%

)

47(17.4%

)Experimen

tal269

(100%)82

(30.4%)

47(17.4%

)2nd

Experiment

Traditional

270(100%)

198(73.3%

)

180(66.6%

)Experimen

tal270

(100%)216

(80%)216

(80%)3rd

Experiment

Traditional

240(100%)

82(34.1%

)

68(28.3%

)Experimen

tal240

(100%)86

(35.8%)

63(26.2%

)

From the above table it is possible to note: a) There is no case of 100% retainment in all

experiments. Words remembered begin to drop asearly as one hour after the treatment andcontinue to drop significantly after one week[except the case of the second experiment whichis presented in section (e), below]. b) Drop after an hour. In experiments one and

three retainment after one hour is relativelyclose, irrespective of the supportive strategyused. The drop is considerably high in allcases varying from 187 to 154 word occurrencesmissed. This constitutes a drop between 79.6%and 64.2%. May that be indicative of wordretainment for those who learn a languagethrough dedicated language teaching software? c) Drop after one week. In all cases of the first

and third experiment word drop continues afterone week irrespective of the supportivefeedback strategy employed. The drop isrelatively small in all cases and variesbetween 5.8% and 18.2%.d) Final retainment. In all cases of the first

and third experiment, final word retainment isvery low, irrespective of the treatmentstrategy used. It varies from 17.4% to 28.3% ofthe initial total of 100%. e) Experiment two was kept to be analysed

separately as retainment is significantlyhigher than in the other two experiments.

Initial, one hour later, retainment remains at73.3% for the traditional supportive feedbackprovision strategy and at an 80% retainment forthe experimental method. Interestingly, theretainment after a week remains high also forthe traditional supportive feedback strategy(TSFS) at 66.6% with a loss of only 6.7%.Finally, notice that the experimentalsupportive feedback strategy (ESFS) in thisexperiment, which involved an off-screen taskremains the same after one week. f) Word retainment after one hour is always

higher than after one week. g) The TSFS performed equally well to the

ESFS in all three experiments while in thefirst experiment it outperformed theexperimental in the after one hour retainment. h) Statistical correlations pursued in twos

independently in each experiment did not seemto favour one strategy over another. The sameresult was found for the attention signallingdevice which seemed to have no impact on thelearning outcomes of the participants.

2. Findings in relation to subjects’ individual scores.41.1% of the subjects scored the same after onehour and after one week, in both the variablestested, in all experiments, including cases ofsubjects who found less than one word. A 70.5%of the subjects participated in the study,found less than one word with the traditionalsupportive feedback strategy while the score issignificantly lower at 55.8% for theexperimental variable. Next comparison, whichinvestigated the difference between subjectswho found more than three words showed scoresof 26.4% of the subjects for the traditionalTSFS and 47% for the ESFS.

3. Comparisons between the two variables. A pairedsample t-test was employed with the use of SPSSto determine whether there is a significantdifference between the average values of thesame measurement, made under the two differentconditions in each experiment separately. Thisdid not reveal any statistically significantdifferences between the independent variablestested.

4. Scores related to the cognitive and learning style of theparticipants. Out of the sixty six (66) subjectsparticipated in all studies, a number of

different cognitive styles were listed.Following the extroverted / introverteddivision, it was found that 70.5% of theparticipants were extroverted, 5.8% introvertedand 8.8%registered as middle. The intuitive /concrete division provided the followingscores: 35.2% were found to be intuitive, 23.5%concrete and 41.1% middle. In the brainorientation dominance test: 14.7% were foundleft to be brained, 20.5% moderately leftbrained, 29.4% with no particular dominance,23.5% moderately right brained, and an 8.8%with a clear right brain dominance. As far asthe openness / closure division, it was foundthat 44.1% were found to be open, 20.5% hadboth inclinations and 35.2 were closed. Anumber of learning styles were also listed. A47% was found to be visual, 5.8% visual andhands-on, 5.8% visual and auditory, 23.5% handson, 5.8% auditory and hands on, 2.9% auditoryand hands-on, and finally an 8.8% was found tomaintain all three cognitive styles. Out ofthose subjects who retained more than threewords after one week with the TSFS 57.1% werevisual and 14.2% were hands-on. Out of thosesubjects who retained more than three wordsafter one week with the ESFS, 28.5% were hand-on, 28.5% were visual and 14.2 used all three.The other percentages are insignificant toregister. Statistical correlations between thisvariable and success retainment did not showany statistically significant correlations,perhaps due to the very low number ofparticipants for every individual case.

5. Scores related to subject’s learning preferences.20.5% of the subjects claimed to learn betterwith the TSFS while 79.5% selected the ESFS. Itwas however recorded that only for a 39.5% ofthe subjects their claims were supported by theevidence as the 70.3% was uncovered to learnbetter from the other SFS they selected. Onlythe 29.7% was able to trace the SFS they wereactually scoring better with. 94.1% recognizedthat engaging feedback was best to assistknowledge although there were cases ofparticipants scoring higher with the TSFS.

Conclusion - Discussion

Although the initial hypotheses have not beensupported by the evidence in this preliminarystudy, a number of significant conclusions canbe made and are presented below in the hopethat they are found fruitful by languageteachers and language learning softwaredesigners. The design and the procedure of thestudy did not present any problems for theapplication of the experiments and thecollection of data relating to the hypothesis,although an external and internal reliabilitytest was not pursued at this stage. Despite thelimitations of this study offered below, anumber of conclusions could be made.1. Results from point (a) in the analysis

show that there is a significant drop of wordsretained in all pilots. This shows clearly thatthe human brain is not able to retain all wordsencountered, irrespective of the supportivefeedback strategy, or that it (the brain)selects the words to retain. In either case itbecomes clear, and it might be considered as anunquestionable finding, that not everythingtaught remains in the brains of the students,irrespective of the supportive feedbackstrategy used.2. Results from point (f) in the analysis

clarify that short term is always higher thanlong term retainment. This clearly signifiesthat eliciting an immediate revised responsefrom the learner after corrective feedback isprovided, as it is suggested by Chaudron(1988), does not guarantee long termretainment. It also shows that again the brainmaintains a selective memory in favour ofcertain words. 3. The surprisingly high retention of words

in the second experiment, on the surface, itmay indicate the success of this ESFS. A closerlook however, uncovers that the same result isfound for the TSFS as well. Thus, at this stageit cannot be interpreted to support our initialhypothesis of expected difference resultingfrom the SFS employed. It may relate to thewords selected to be tested and the languagelevel of the subject in this study (difficultyindex of words being closer to the languagelevel of those participants).

4. Although the actual number of wordsretained was not significantly higher in termsof statistic comparisons, the number oflearners retaining more words with the ESFSswas significantly higher in absolute numbersand percentages (almost double). This is aclear indication that the ESFSs used in allexperiments invited more learners in thelearning process while it proved to have abetter distribution in the learning style ofthe subjects (section four of the analysis). 5. Findings related to the cognitive and

learning style showed that the citizens ofGreece which participated in this study areextroverted, split between openness andclosure, split between intuitive and concretewith both, left and right brain dominance. Thedominant learning style was registered to bevisual and hands-on. It should be noted herethat most subjects respond to thisquestionnaires by offering a vision ofthemselves (how they like to see themselves)rather than what they truly are. 6. Findings related to the subjects’ actual

scores preference, for SFS and best perceivedSFS for learning, clearly indicate that theselearners were not aware of the feedbackstrategy they learned better. It should be noted here that the study on

feedback (supportive and corrective) is a keyissue in language education research because itmay provide valuable information that wouldhelp the language teacher better shape histeaching and corrective strategies and further,it may lead to inferences about the processeswhich take place within the human brain. Thisis supported by most learning theoristsdiscussed in the book by Olson and Hergenhahn(2007: 3) who agree in that “the learningprocess cannot be studied directly; instead,its nature can only be inferred from changes inbehaviour.” Olson and Hergenhahn (ibid.) seethis process as an intervening variable whichoccurs between the dependent variable (thebehavioural changes) and the independentvariables (observed stimuli). It is suggestedby the authors that the “independent variablescause a change in the intervening variable

(learning), which in turn causes a change inthe dependent variable (behaviour).”

Shortcomings. Among the shortcomings of thisresearch are: a) the words selected to betested were not all of the same size or numberof syllables, b) the words were different ineach experiment and it was not possible toarrive at solid conclusions with highervalidity (results being compatible for all thesubjects’ population in this study). Resultswould have been clearer if words remained thesame as a control variable in all three cases,c) the total number of subjects involved inthis preliminary study is too small to arriveto any firm conclusions, or statisticallysignificant results, except for the point (1)made above, d) the language level of thesubjects needs to be controlled and seen inrelation to the difficulty index of the wordsselected in such experiments as these mayproduce faulty results and jeopardize researchreliability and validity, e) external andinternal reliability of the results was nottested, and finally f) there were not enoughitems of the same strategy used in all threeexperiments, mainly in the first example (toomany strategies tested at the same time). Itwould seem advisable to test each strategyseparately and in comparison to sametraditional.

Further research. In relation to conclusion one,it would be interesting to find out on whatcriteria word retainment selections are made inthe human brain, as this finding would provideinformation of brain functions to languageinput in general. In relation to conclusionpoint four, the off-screen task as a SFS mayneed to be further tested with a greater numberof subjects with words of higher indexdifficulty (all of the same level) than thelanguage level of the subjects.

ReferencesAlanen, R. (1995). Input enhancement and

rule presentation in second languageacquisition. In R.W. Schmidt (ed.), Attentionand Awareness in Foreign Language Learning. Honolulu:University of Hawai'i Press. (259-302).

Allwright, R. L. (1975). Problems in thestudy of the language teacher’s treatment oflearner error. In Burt, M. and Dulay, H.(eds.), New Directions in Second Language Learning,Teaching and Bilingual Education. Washington, D.C.:TESOL, 96-109.Bley-Vroman, R. (1986). Hypothesis testing in

second-language acquisition theory. LanguageLearning, 36(3), 353-376, doi:10.1111/j.1467-1770.1986.tb00559.xBley-Vroman, R. (1989). What is the logical

problem of foreign language learning? In. S.Gass & J. Schachter (Eds.), Linguistic perspective onsecond language acquisition. (pp. 41-68). New York:Cambridge University Press.Brock, C., G. Crookes, R. Day, and M. Long

(1986). Differential effects of correctivefeedback in native speaker-nonnative speakerconversation. In R. R. Day (ed.). Talking to learn:conversation in second language acquisition. Rowley, MA:Newbury House.Brown H.D. (1994). Principles of language learning and

teaching. 4th ed, Prentce Hall Regents EnglewoodCliffs.Brown, J. D. (1991). Accuracy and bias in

self-knowledge. In C. R. Snyder & D. F. Forsyth(Eds.), Handbook of social and clinical psychology: Thehealth perspective. (pp. 158-178). New York:Pergamon Press.Brown, R. and U. Bellugi (1964). Three

Processes in the Child’s Acquisition of Syntax.Harvard Education Review. Vol. 34 133-151Carroll, S. and M. Swain (1993). Explicit and

implicit negative feedback: An empirical studyof the learning of linguistic generalizations.Studies in Second Language Acquisition, 15, 357-386.Cathcart, R., and J. Olson, (1976) Teachers’

and students’ preference for correction ofclassroom conversation errors. In Fanselow,J. and Crymes, R. (Eds.) On TESOL ’76.Washington, D.C.: TESOL, 41-53.Chaudron, C. (1988). Second language classrooms:

Research on teaching and learning. Cambridge, UK:Cambridge University Press.Chomsky, N. (1975). Reflections on language. New

York: Pantheon.Cohen, A. D. And Robbins, M. (1976). Toward

assessing inter-language performance: therelationship between selected errors,

learners’ characteristics and learners’explanations. Language Learning, 26, 54-66.Crystal, D. (1985). A Dictionary of Linguistics and

Phonetics. Basil Blackwell.deGraaf, R. (1997). The eXperanto

experiment: effects of explicit instructionon second language acquisition. Studies in SecondLanguage Acquisition, 19, 249-276DeKeyser, R. (1995). Learning second

language grammar rules: an experiment with aminiature linguistic system. Studies in SecondLanguage Acquisition, 17, 379-410.Dekeyser, R. (2006). Implicit and Explicit

Learning. In Doughty, C.J. and M.H. Long(Eds.). The Handbook of Second Language Acquisition.Blackwell PublishingDulay, H., Burt, M., Krashen, S. (1982).

Language Two. Oxford University Press.Ellis, N. (1993). Rules and instances in

foreign language learning: interactions ofexplicit and implicit knowledge. EuropeanJournal of Cognitive Psychology, 5, 289-318.Ellis, R. (1985). Understanding Second Language

Acquisition Oxford University PressEllis, R. (1991). Grammar teaching practice

or consciousness-raising? In R. Ellis (Ed.),Second language acquisition and second languagepedagogy. (pp.232-241). Clevedon, UK:Multilingual Matters.Fanselow, A. (1977). The treatment of error

in oral work. Foreign Language Annals, 10:583-93.Ferris, D. (2006). Does error feedback help

student writers? New evidence on the short-andlong-term effects of written error correction.In Hyland K. and F. Hyland (Eds) Feedback inSecond Language Writing. Cambridge AppliedLinguistics SeriesGagne, R. M. (1977). The Conditions of Learning.

Holt-Saunders International EditionsGass, S. M. (1991). Grammar instruction,

selective attention, and learning. In R.Phillipson, E. Kellerman, L. Selinker, M.Sharwood Smith, and M. Swain (Eds.),Foreign/Second language pedagogy research. (pp.124-141). Clevedon, UK: Multilingual Matters.Gass, S.M. and E.M. Varonis (1994). Input,

interaction, and second language production.Studies in Second Language Acquisition, 16, 283-302.

Han, Z-H. (2001). Fine-tuning correctivefeedback. Foreign Language Annals, 34, 582-599.Hendrickson, J. M. (1977). Goof analysis for

ESL teachers. Unpublished paper. Ohio StateUniversityHendrickson, J.M. (1978). Error correction in

foreign language teaching: Recent theory,research, practice. Modern Language Journal, 62,387-398.Holley, F. and King, J. (1971). Imitation

and correction in foreign language learning.Modern Language Journal. (55-58).Hounsell, D. (1999). The Evaluation of

Teaching. In A Handbook for Teaching and Learning inHigher Education Kogan Page. (pp. 161-174).Kaufman, L. M. (1991). Harmful Backwash from

EFL Tests and Attitudes towards Teaching andPracticing for the Test in CommunicativeLanguage Teaching with the Proceedings of the10th International Conference: Assessment andEvaluation, Vol.6, pp.299-308.Keislar, E.R. & McNeill, J.D. (1962) A

Comparison of Two Response Modes in anAutoinstructional Program with Children inthe Primary Grades. Journal of EducationalPsychology, 53, pp.127-131.Krashen,S.D. (1994). Bilingual education and

second language acquisition theory. Inbilingual Education Office (ed.) Schooling andlanguage-minority students: A theoretical framework (2nd

ed., pp. 47-75). Los Angeles: EvaluationDissemination and Assessment Center, CaliforniaState University. Lightbown, P.M. and N. Spada, N. (1990).

Focus-on-form and corrective feedback inCommunicative Language Teaching: Effects onsecond language acquisition. Studies in SecondLanguage Acquisition, 12, 429-448.Long, M. (1996). The role of linguistic

environment in second language acquisition. InW.C. Richie and B.K. Bahtia (Eds.), Handbook ofsecond language acquisition (pp.413-468). New York:Academic Press.Long, M. (1998). Focus on form: Theory,

research, and practice. In Doughty, C. and J.Williams (ed.) Focus on Form in Classroom SecondLanguage Acquisition Cambridge University PressLong, M. H., S. Inagaki, L. Ortega (1998). The

Role of Implicit Negative Feedback in SLA:

Models and Recasts in Japanese and Spanish. TheModern Language Journal. Volume 82, Issue 3,pp.357–371.Long, M. and P. Robinson (1998). Focus on

form: theory research and practice. In C.Doughty and J. Williams (Eds.), Focus on form inclassroom second language acquisition (pp. 15-41). NewYork: Cambridge University PressLong, M. (2001). Focus on form: A design

feature in language teaching methodology. In K.de Bot, R. Ginsberg, and C. Kramsch (Eds),Foreign language research in cross-cultural perspective (pp.39-52). Amsterdam: John Benjamins. Lyster, R. (2001). Negotiation of form,

recasts and explicit correction in relation toerror types and learner repair in immersionclassrooms. Language Learning, 51, 265-301Lyster, R. and L. Ranta (1997). Corrective

feedback and learner uptake. Studies in SecondLanguage Acquisition, 19, 37-66.Mc Donald, F. J. (1959). Educational Psychology

Wadsworth Publishing CompanyMackey, A. and J. Philip (1998).

Conversational interaction and secondlanguage development: Recasts, responses, andred herrings? Modern Language Journal, 82, 338-356.Menezes de Oliveira e Paiva V. L. (2003).

Feedback in the virtual environment PsychNologyJournal Vol. 1(3) pp.257-282Michas, I.C. and D.C. Berry (1994). Implicit

and explicit processes in the second-languagelearning task. European Journal of CognitivePsychology, 6, 357-381Nicholas, H., Lightbown, P. M., & Spada, N.

(2001). Recasts as feedback to languagelearners. Language Learning, 51, 719-758.Ohta, A.S. (2001). Second language acquisition

processes in the classroom: Learning Japanese. Mahwah,NJ: Lawrence ErlbaumOlson, M.H. and B.R. Hergenhahn (2007). An

introduction to theories of learning Pearson Prentice Hall (eighth edition)Plan, S. (1977). Acquiring a second language

in an immersion class-room. In Brown, H.D.,Yorio, C.A. and Crymes, R. (eds.) Teaching andLearning English as a Second Language: Trends in Researchand Practice. Washington, D.C.: TESOL, 213-25.

Poole, A. (2005). Focus on form instruction:Foundations, applications and criticism. Thereading matrix: An international onlinejournal. Vol. 5(1)http://www.readingmatrix.com/journal.html Psaltou-Joycey, A. (2004). Variable effects of

different types of corrective feedback on L2learning. In Sougari, A.M. and Joycey, E.Challenges in Teacher Education University StudioPressRichards C.J and C. Lockhart (1996). Reflective

Teaching in Second Language Classrooms, CambridgeUniversity PressScott, V.M. (1989). An empirical study of

explicit and implicit teaching strategies inFrench. Modern Language Journal, 73, 14-22Scott, V.M. (1990). Explicit and implicit

grammar teaching strategies: new empiricaldata. French Review, 63, 779-789.Skinner, B. F. (1954). The science of learning and

the art of teaching. Harvard Educational Review,24:86-97. (Reprinted in Lumsdaine & Glaser1960).Truscott, J. (1996). The case against grammar

correction in L2 writing classes. LanguageLearning, 46, 327-369.Tsopanoglou, A., Ypsilandis, G.S., Mouti, A.

(2014). Piloting a polychotomous partial-creditscoring procedure in a multiple choice test. InLanguage Learning in Higher Education   Issue number:4-1. Editors in Chief: Mansfield, Gillian /Little, David. De GruyterYpsilandis, G.S. (2002) Feedback in Distance

Education. C.A.L.L. Journal Vol. 15 No2, Swets &Zeitlinger Publishers. (pp. 167-181).Ypsilandis, G.S. (2006). On Feedback provision

strategies in CALL software. Keynote at theMotivation in Learning Language for Specific and AcademicPurposes conference. Proceedings of the 5thInternational Conference on Language Learningfor Specific and Academic Purposes,  publishedby the University of Macedonia Press, 2006VanPatten, B. and S. Oikkenon (1996).

Explanation versus structured input inprocessing instruction. Studies in Second LanguageAcquisition, 18, 495-510White, L. (1991). Adverb placement in second

language acquisition: some effects of positiveand negative evidence in the classroom. Second

Language Research Vol. 7 n.2: 133-161, SAGEJournals