Teachers’ beliefs and practices about oral corrective feedback in university EFL classes

This study examined (in)congruences between beliefs and practices of EFL university teachers on in-class oral corrective feedback (OCF). The participants were 20 English language from a private in Turkey. Data collected via video-recorded non-participant detached observation, task about OCF to determine the teachers, stimulated recall interview. results showed incongruence what said they believed did. However, teachers’ similar regarding whether errors should be corrected, when who correct them. Particularly notable this was finding that those with greatest almost always stood by their decisions, even after watched unsuccessful practices. Bu çal??ma, üniversitede ?ngilizceyi yabanc? dil olarak ö?reten okutmanlar?n s?n?f içi sözlü düzeltici dönüt (SDD) konusundaki inançlar? ve uygulamalar? aras?ndaki uyum(suzlu?)u incelemi?tir. Kat?l?mc?lar, Türkiye'deki bir vak?f üniversitesinde çal??an okutmand?r. Veriler videoya kaydedilmi? ba??ms?z gözlem, SDD ile ilgili inançlar?n? belirlemek için tasarlanm?? etkinlik ça?r???m tekni?ine dayal? görü?me yoluyla toplanm??t?r. Sonuçlar, inand?klar?n? söyledikleri yapt?klar? aras?nda tutars?zl?klar oldu?unu göstermi?tir. Ancak, hatalar?n düzeltilmesi gerekip gerekmedi?i, ne zaman gerekti?i kimin düzeltmesi konusunda ö?retmenlerin benzerlik çal??mada özellikle dikkat çekici olan, en büyük uyumsuzlu?a sahip okutmanlar?n, ba?ar?s?z uygulamalar?n? izledikten sonra bile de?i?tirmek istememeleri verdikleri kararlar?nda ?srarc? olamalar?d?r. Oral (OCF) is one most common teacher moves classroom discourse. Feedback imperative for learners, as it helps them improve understanding of, communicative ability in, (Mackey & Goo, 2007; Nassaji, 2016; Russell Spada, 2006). Besides, contrary arguments against error correction, we can come across “even more numerous much convincing reasons why an integral part teaching practices” (Pawlak, 2014, p. 49). As Li (2018, 4) puts it, “research has unequivocally demonstrated benefits facilitating L2 development”; therefore, not question effective or which type (Nassaji Kartchava, 2020). So, rather than investigating values different types, stipulate need investigate because influence effectiveness (Sheen, see also Lyster et al., 2013). Li's (2017) review seven studies CF revealed only 39% thought important. More point, recent review, Vuono (2019, 99) found “teachers inconsistency consistency” terms OCF. A deeper topic, requires timely investigation. main aim therefore shed light upon interviews. In state-of-the-art article, al. (2013) provide overview research feedback, identifying frequencies, preferences, theoretical issues, then moving discussions effectiveness, linguistic targets, learners’ age, peer aspects. They conclude make own decisions drawing techniques available means ultimately, individual teacher's influences implementation feedback. (2017, 143) defines “attitudes, views, opinions stances learners hold utility learning how implemented classroom.” literature (see Vuono, 2019, past 25 years relevant published System) shows some focus learner (Agudo, 2015; Jernigan Mihai, 2008; Lee, 2013; Loewen 2009; Zhu Wang, 2019), others 2014; Gurzynski-Weiss, 2010; Hernandez Mendez Reyes Cruz, 2012; Rahimi Zhang, 2015), both (Davis, 2003; Pawlak, Schulz, 1996, 2001). There are address relationship (Bao, 2019; Basturkmen 2004; Dong, Junqueira Kim, Kamiya, 2006; Mori, 2002; Olmezer-Ozturk, Roothooft, 2014). line last focus, present aims explore congruence real 40 ago, Hendrickson (1978) asked five overarching questions challenge mechanisms These “Should corrected?” “How corrected” “When “Who correction?” “Which examination used these guide study. our scan literature, examine conducted ESL settings (e.g., 2002), while there exceptions non-ESL such Bao (2019), Dong (2012), Olmezer-Ozturk (2019) Roothooft (2014). investigated eight Chinese second China, two foreign at US master's thesis. Olmezer-Ozturk's Turkish setting working intensive program. Roothooft's Spanish context, 10 public institutions reported (should corrected?), (which provider (who correction?), timing (when corrected?) (how Regarding OCF, Kim's four Kamiya's (2014) providing ineffective practicing differently. When comes (i.e., all (2004) favored meaning-oriented mistakes but often addressed language-related mistakes, On level Kim stated pronunciation grammar-related errors. other hand, Mori's (2002) each Among previous conspicuous (2012) focused ideal Both Dong's self-correction favorable same time opting correction. however, Bao's study, six (out eight) pointed out followed practice during lessons. she would activity did activity. half consistent performed To gap research, topic findings explanatory sequential (QUAN?QUAL) mixed-methods (EFL) larger number covering aspects suggested (1978), includes interview where reflect if mismatch encountered What (a) (b) (c) provider, (d) (e) OCF? reasons, any, classroom-based design (Creswell Plano Clark, 2017) had stages data collection (first quantitative, qualitative) three methods (two quantitative stage). For transparency response shown Table 1. stage one, observation gathered a-three-hour video recordings regular hours. Also stage, identified based analyzed using Spearman's Rank Order Correlation (See Appendix). After analyzing first sources collection, two, highest difference rates chosen (Teachers 6, 10, 12, 18) gain insight into observe hours speaking classes recorded, total making 60 h recordings. “detached” observations, researchers classroom, avoiding “observer's paradox,” causing students behave certain way presence observer (McKinley Rose, 2019). Each transcribed verbatim episodes identified. episode later coded content analysis 2019) following help reduce bias, immediately completed. Within week interviews selected teachers. recorded transcribed, qualitative analysis. Before any took place, accordance ethical guidelines, informed free withdraw without showing reason could purposes possible publication. All agreed signed consent form given explanation. program university. age varied 24 47 (M = 32.4) experiences ranged 1.6 22 7.8). Some held bachelors’ degrees fields Language Teaching (n 7), Literature 4), Translation Interpretation 2) Linguistics 1). Six graduate Educational Statistics 1) Four CELTA certificates. 16 recording 18.6). general consistency, B1 level, decided in-house placement progress tests materials prepared experienced staff mostly whole-group teacher-centered interaction very few no instances pair group work activities. courses, although skills practiced, observations assuming student teacher. 20) one-hour times. made Based Hendrickson's developed collect beliefs. Participants hypothetical correction excerpts choose Ranta's (1997) classification. belief aimed match through observations. collecting necessary relationships third incongruences invited least within remember it. allow speak easily, occasional code-switching English. During interviews, encouraged report choices type. important point practices; watching, asking talk audio episodes, considered Ranta, 1997; researcher went videos note Then, do same. training manual prepared, included definitions examples. inter-rater reliability (Hallgren, 2012), Cohen's kappa good agreement external reviewer (k 0.887), 0.874). presented practices, organized according key (stage task), differences interview). use English, below have been translated identify perceived rate percentile scale task. compared details responses participant together averages actions provided 2. seen 2, percentages quite 65.25 63.5, respectively. Moreover, average ratios slightly higher still close. participants’ rank order correlation correlate significantly, Spearman test (r 0.73, p < 0.01). way. One category error. task, rated corrected vocabulary, grammar 3 illustrates actual obtained participated correcting (the 2.1%) vocabulary 4.4%) bigger 10.9%). focusing may really portray picture further correlated significantly 0.74, 0.01) 0.53, 0.05) whereas significant 0.46). noted providers setting. themselves self-correction), peers feedback) feedback). state questionnaire observed lessons taught. comparison yielded mixed results. 9.9%) 11.35%) 1.45%) 4 details). show (self-correction r 0.14, 0.12 0.04). decide students’ errors, keep flow communication. preferred classroom. 5 class. asked, opted balanced approach (immediate: 48% vs. delayed 52%) (however, individually side other). practice, imbalanced favor immediate 79.6% 20.4%). analyzed, 0.81, 0.01 0.78, 0.01, respectively). Teachers repertoire types mistakes. implicit recasts elicitations explicit task's samples (57%) (23.5%). chose clarification requests (6.5%), metalinguistic (4%), repetition (5%) (4%) generally resulting low 6). proportion slight changes. Recasts (61.35%) (24.4%) frequent type, rest inconsistent. recast elicitation, recast; 0.85, elicitation). request 0.35), 0.17), 0.03) 0.30) levels correlation. (T6, T10, T12 T18) in-class. selection inquire like re-evaluate answers gave questionnaire. appear 7. According 7, appears might believe negative evidence interlanguage development learners; contextual factors played role Of factors, dynamics discussion “The generated lively discussion, so I want interrupt correction”), characteristics “I waited self-repair.” “He successful student”; “That shy introverted give feedback”; explicitly he little bit behind his friends needs help”), management lost talking off-topic, regroup students. Normally, that”) necessary”). wanted change stance appropriate specific contexts, thus revealing inconsistencies illustrate better, episodes. intentionally elicit reactions context reasons. taken Excerpt 1 (taken T6) Topic: future plans 1, class plans. 113, makes mistake, provides goes unnoticed mistake again. interview, watching video, realized notice chance, “No” continuation small there. 2 T12) summer holiday holiday. 82, starts yes/no question. next line, 84, does understood request, student. Later, [use niece instead nephew] required extra explanation.” (either second) “the corrections rational” (teacher's words). 3, discussing Early word choice (instead excellent, uses excel). continues student's lesson, her chance. another 4, grammatical mistake. continues. this, noticed didn't correction.” “No,” “fluency accuracy grammar.” T10) 5, Speaking session. 43, tries floor whole facilitate everyone stays silent chance class, nobody responded pause there.” preference decision himself. We protocols. Following classification, (namely time, type), course recording, approximately 65%. Similarly, 63% ratio 68%. (2017), 39%, positive (50.5%), (68.5%) (69%) corrected. application (grammar 61.4%, 70.6%, 73.4%) built (2004), well matched Vuono's studies, large extent th