ARTÍCULO ORIGINAL  
INNOVA Research Journal, ISSN 2477-9024  
Septiembre-Diciembre 2023). Vol. 8, No.3 pp. 37-50  
(
Data-Driven Learningꢀin an EFL class: aꢀstudy of Ecuadorian learners’  
perceptions  
Aprendizaje basado en datos en la enseñanza de inglés como idioma  
extranjero: percepciones de estudiantes ecuatorianos  
Hover Ismael Conza-Armijos  
Universidad Nacional de Loja, Loja, Ecuador  
Liliana Fernanda Celi-Celi  
Universidad Técnica Particular de Loja, Loja, Ecuador  
Recepción: 17/06/2023 | Aceptación: 29/08/2023 | Publicación: 10/09/2023  
Cómo citar (APA, séptima edición):  
Conza-Armijos, H. I., y Celi-Celi, L. F. (2023). Data-Driven Learning in an EFL class: a study of  
Ecuadorian learners’ perceptions. INNOVA Research Journal, 8(3), 37-50.  
https://doi.org/10.33890/innova.v8.n3.2023.2297  
Abstract  
Previous research has reported an increased interest in the application of Data-Driven Learning  
(DDL), which was derived from corpus linguistics, to aid English as a foreign or second language  
learning. DDL emerged as a learner-centered approach that emphasizes the discovery of linguistic  
patterns by analyzing natural occurrences of language. Many studies have investigated the effect  
of DDL on several linguistic components and have reported positive outcomes. Consequently, its  
application has germane importance in an Ecuadorian context as it has not been widely explored.  
This study seeks to document the experience of 14 EFL students from an Ecuadorian public  
university and examine their perceptions towards the implementation of DDL activities. The  
intervention included two stages: assisted guidance for students to retrieve information on their  
own and corpus-based materials developed by the teacher. Students' perceptions were gathered via  
a questionnaire, a reflection, and a follow-up interview. The findings indicate that students  
Esta obra se comparte bajo la licencia Creative Common Atribución-No Comercial 4.0 International (CC BY-NC 4.0)  
Revista de la Universidad Internacional del Ecuador. URL: https://www.uide.edu.ec/  
37  
Hover Ismael Conza-Armijos y Liliana Fernanda Celi-Celi  
ISSN 2477-9024. Innova Research Journal (Septiembre-Diciembre, 2023). Vol. N8, No. 3, pp. 37-50  
perceived DDL as a practical and innovative approach to support grammar discovery in an EFL  
class. Nevertheless, students encountered some challenges associated with the use of DDL, for  
example, the limited number of searches, an issue that can be resolved by a complete registration  
in the COCA web. This article concludes with pedagogical implications and recommendations for  
future DDL applications and research in Ecuadorian EFL classes.  
Keywords:  
COCA;  
corpus linguistics; data-driven learning (DDL); EFL class; learners'  
perceptions.  
Resumen  
Muchas investigaciones han divulgado la implementación del aprendizaje basado en datos,  
conocido en inglés como Data-Driven Learning (DDL), que surgió de la lingüística de corpus para  
contribuir con el aprendizaje del inglés. El DDL se enfoca en el estudiante y fomenta el  
descubrimiento de patrones lingüísticos a través de un análisis de ejemplos suscitados en un  
contexto natural. Numerosos estudios han investigado los efectos de DDL en ciertos componentes  
lingüísticos y han reportado resultados positivos. Por lo tanto, la implementación de DDL es  
crucial en un contexto ecuatoriano ya que hasta el momento no se ha explorado. Este trabajo  
documenta la experiencia de 14 estudiantes en una universidad pública ecuatoriana y recoge sus  
percepciones sobre la implementación de actividades con el DDL. Esta intervención se realizó en  
dos etapas: la primera, incluyó una capacitación acerca del uso de corpus para obtener ejemplos  
de forma independiente. La segunda etapa estuvo a cargo del docente y consistió en la creación de  
materiales basados en corpus. Con respecto a las opiniones de los estudiantes se utilizó un  
cuestionario, una reflexión y una entrevista. Los resultados mostraron que los estudiantes  
consideran el DDL como un enfoque práctico e innovador para coadyuvar con el aprendizaje de  
la gramática del inglés. Algunas limitaciones giraron en torno al número restringido de búsquedas  
que la herramienta permite, lo cual, se resolvería con un registro previo en la web de COCA. Este  
artículo concluye con recomendaciones pedagógicas para aplicaciones e investigaciones futuras  
de DDL en un contexto similar.  
Palabras claves: aprendizaje basado en datos; COCA; inglés como lengua extranjera; lingüística  
de corpus; percepciones de estudiantes.  
Background to the study  
Corpus linguistics  
Lindquist (2009) defines corpus as "a collection of texts which is stored on some kind of  
digital medium used by linguists to retrieve linguistics items for research" (p 3). For Crawford and  
Csomay (2016), corpus stores a large number of texts (written or spoken) from where the study of  
language in use can be conducted (see also Boulton, 2012; O'Keeffe, McCarthy, and Carter, 2007).  
In this realm, the study and analysis of these data is conceptualized as corpus linguistics (Cheng,  
2
012).  
Over the past few years, the contributions of corpus linguistics have spread to many fields,  
one of which is language teaching because it provides complete insights into how people use the  
language (Conrad, 2010). Huang (2011) asserts that corpus linguistics enables distinguishing  
language forms at various levels of formality. By the same token, this comprehensive analysis  
shows how different speech functions help language users across contexts. This is supported by  
Esta obra se comparte bajo la licencia Creative Commons Atribución-No Comercial 4.0 International (CC BY-NC 4.0)  
Revista de la Universidad Internacional del Ecuador. URL: https://www.uide.edu.ec/  
38  
Data-Driven Learning in an EFL class: a study of Ecuadorian learners’ perceptions  
Biber and Reppen (2002) who claim that individuals do not often use language as prescribed in  
grammar books and that lexical meanings are contingent on contexts and users.  
Currently, there are different types of corpus that are representative of certain aspects of  
language that can be employed to analyze, describe and teach the language. For example, the  
British National Corpus comprises more than 100 million words that represent the universe of  
contemporary British English. The Corpus of Contemporary American English (COCA) (Davis,  
2
008) is the largest of its kind and contains more thanꢀone billionꢀwordsꢀof text.  
COCA  
The source of data for this work is COCA (https://www.english-corpora.org/coca/) largely  
because learners in this context are taught American English, so it is a good way of exposing them  
to more authentic input. COCA is well equipped with analytical tools that enrich the search such  
as frequency, charts for genres and years, collocates, comparisons, KWIC, POS tags, etc. In the  
pedagogical world, these tools allow manipulating the language, which provokes learning and  
retention (Reppen, 2010).  
One of the main tools of COCA is a word concordance called 'Key-Word-In-Context'  
(KWIC). KWIC enables viewing corpus data in which samples of real language are presented in  
lines of text known as concordances. Users can also access an extended context for each occurrence  
of the input search (Pérez-Paredes et al., 2019). This tool helps identify aspects of a particular word  
or phrase such as lexico-grammatical patterns or semantic prosody. (Boontam & Phoocharoensil,  
2
018).  
When retrieving results with KWIC, the word occurrences are displayed vertically with the  
node word in the center of the line and the surrounding context highlighted in different colors. The  
colors signal the part of speech as follows: orange for adverbs, purple for verbs, turquoise for  
nouns, yellow for prepositions, green for adjectives, blue for pronouns, and grey for articles,  
demonstratives, and connectors. Learners may find the structural analysis more manageable  
provided that the interface is visually appealing.  
Data-driven learning in ELT  
Corpus Linguistics has made evident that language exhibits a highly patterned structure. A  
direct application of this language characteristic in teaching is the Data-Driven Learning (DDL)  
approach, which involves examining patterns in a target language and inferring meaning and use  
(Johns, 1991). In this seminal work, Johns captures three central moments towards language  
analysis: identify, classify, and generalize. Identifying refers to the recognition of the target  
structures. It then leads to a systematic classification of the retrieved information, which the teacher  
or the student can do. Finally, generalizing entails a profound analysis to draw logical conclusions  
regarding language rules.  
DDL offers greater insights to be applied in foreign language environments because it  
compiles natural and authentic examples of language in use (Reppen, 2010; Luo, 2016). Within  
DDL, learners play a pivotal role as they become "language detectives" (Johns, 1997, as cited in  
Esta obra se comparte bajo la licencia Creative Commons Atribución-No Comercial 4.0 International (CC BY-NC 4.0)  
Revista de la Universidad Internacional del Ecuador. URL: https://www.uide.edu.ec/  
39  
Hover Ismael Conza-Armijos y Liliana Fernanda Celi-Celi  
ISSN 2477-9024. Innova Research Journal (Septiembre-Diciembre, 2023). Vol. N8, No. 3, pp. 37-50  
Flowerdew, 2012) because they need to recognize and examine the emergent patterns to draw their  
conclusions (Smart, 2014).  
In addition, DDL favors a guided inductive approach to language learning where the  
elements of explanation and corpus use are suited to students' needs (Johansson, 2009; O'Keeffe,  
2
020). In this respect, the teachers' roles entail guiding students whilst students perform the  
analysis (Boontam & Phoocharoensil, 2018). There are two possible methods to exploit DDL; for  
one, teachers can prepare the materials in advance or direct students to conduct the searches  
independently. Consequently, the learners' role is twofold: they act as language learners and as  
language researchers simultaneously (Cheng, 2010).  
Guan (2013) and Smirnova (2017) assert that DDL fosters problem-solving skills as  
learners explore and discover language structures, unlike teacher-centered methodologies.  
Szudarski (2018) highlights the role of corpus in increasing learning autonomy and understanding  
real-life linguistic communication. This is because corpus does not interpret data; it is rather the  
duty of the researcher (Szudarski, 2018; Gavioli, 2005). The outcomes derived from this process  
result in a richer and more meaningful language learning experience.  
Previous Studies  
A great deal of research has investigated the application of DDL in EFL settings and  
students' attitudes towards the use of corpus-based learning methodology. For example, in 2014,  
Geluso and Yamaguchi conducted a study on DDL use to discover formulaic language for a 15-  
week period. This study took place in a private university in Japan and involved 29 students at the  
A2-B2 CEFR level. For data collection, the authors employed a questionnaire, follow-up  
interviews, and student reflection logs. The findings revealed that students had positive attitudes  
regarding DDL use despite some minor difficulties such as unfamiliar vocabulary.  
A study by Boontam and Phoocharoensil (2018) examined the effectiveness of paper-based  
DDL method for the learning of English prepositions with 35 A2 students at a private school. The  
participants were taught the prepositions using DDL for about 6 weeks. The researchers gathered  
data through a pre and post-test and complemented the process with a questionnaire and an  
interview. The results depicted those students found the learning experience with DDL fun,  
interesting, and challenging.  
A small-scale study by Yoon and Hirvela (2004) reported positive findings in the  
implementation of a corpus approach in an American university. These 22 international students  
viewed the corpus approach as valuable to support the L2 writing process. In addition, participants  
pointed out that DDL boosted their confidence to develop this productive skill.  
A recent longitudinal study by Lin and Lee (2019) analyzed the pedagogical use of DDL  
for EFL grammar classes in Taiwan. Three instructional formats were used to compare students'  
grammar performance: a traditional deductive approach (TDA) only, a combination of 40% DDL  
with 60 % of TDA, and a blend made of 60% DDL plus 40% of TDA. It was found that all the  
treatment arrangements helped students to learn grammar. However, in terms of attitudes, students  
tended to favor the DDL approach.  
Esta obra se comparte bajo la licencia Creative Commons Atribución-No Comercial 4.0 International (CC BY-NC 4.0)  
Revista de la Universidad Internacional del Ecuador. URL: https://www.uide.edu.ec/  
40  
Data-Driven Learning in an EFL class: a study of Ecuadorian learners’ perceptions  
Overall, these previous studies show that DDL may produce a meaningful learning  
experience in EFL contexts. To this date and to the best of our knowledge, no studies have been  
undertaken in Ecuadorian institutions to implement DDL and collect students' or teachers'  
perceptions. For this reason, we deem necessary to conduct a study to capture students' perceptions  
towards DDL and, by the same token, to shed light on pedagogical applications.  
Research Questions  
This study attempts to answer two research questions:  
What are Ecuadorian EFL university learners' perceptions of adopting a DDL approach for  
grammar learning?  
What approach to learning grammar do students prefer?  
Methodology  
Participants  
The participants in this study were taking general English lessons as a requirement prior to  
obtaining their formal qualification in an Ecuadorian public university. Students would normally  
attend face-to-face sessions but given the health emergency caused by the Covid-19 pandemic,  
they received virtual lessons via Zoom. The deliberate instruction took place twice a week for 60  
minutes per session, where grammar teaching was paired with language skills development.  
This small-scale research took place in an elementary English class, Level 4 from the EFL  
program. The class consisted of 23 students, of whom 14 students volunteered to participate (11  
female and 3 male). The participants were majoring in teaching and education, and their ages  
ranged from 20 to 24 years old. All participants shared the same nationality and L1 (Spanish).  
They had received English for 6 years as part of Ecuador's national curriculum for secondary  
education and nearly 300 hours of instruction in the English language program.  
DDL Treatment and materials  
At the beginning of the semester, students took a diagnostic test that was an adaptation  
from the Key A2 by Cambridge Assessment English (2020). The test originally consisted of  
reading, writing, listening, and speaking, but the speaking part was removed due to practicality  
issues. This test served the purpose of gauging students' language proficiency.  
The treatment was implemented in weeks 5, 6, and 7 of the 16-week semester, and each  
treatment lasted for two hours. The topic unit was food, and the grammar revolved around  
countable and uncountable nouns and quantifiers. See figure below:  
Esta obra se comparte bajo la licencia Creative Commons Atribución-No Comercial 4.0 International (CC BY-NC 4.0)  
Revista de la Universidad Internacional del Ecuador. URL: https://www.uide.edu.ec/  
41  
Hover Ismael Conza-Armijos y Liliana Fernanda Celi-Celi  
ISSN 2477-9024. Innova Research Journal (Septiembre-Diciembre, 2023). Vol. N8, No. 3, pp. 37-50  
Figure 1  
The screenshot of COCA interface  
The first synchronous session included the presentation of food semantic sets and activities  
to promote repeated encounters with the vocabulary. For the asynchronous work, students searched  
for two examples for the quantifiers a, an, some, and any followed by a food item.  
The teacher started the second synchronous session by having students share one of their  
examples on Google Jamboard. The rationale behind this activity was to guide students' attention  
to singular and plural nouns before presenting the grammar topic of countable and uncountable  
nouns. After some guided practice, students were prompted to work in groups of four in Zoom  
Breakout Rooms. It is worth mentioning that students chose their peers they wanted to work with  
to create a quiz with 5 multiple-choice questions on Quizziz. For the asynchronous work, students  
had to take the quizzes created by their peers to allow further practice and consolidate their grasp  
of the content. Also, to ensure accountability, students had to keep a record of the examples they  
retrieved on a digital foldable template. The teacher provided this template and included an  
organized layout and detailed instructions of the tasks at hand.  
Esta obra se comparte bajo la licencia Creative Commons Atribución-No Comercial 4.0 International (CC BY-NC 4.0)  
Revista de la Universidad Internacional del Ecuador. URL: https://www.uide.edu.ec/  
42  
Data-Driven Learning in an EFL class: a study of Ecuadorian learners’ perceptions  
Figure 2  
Students' examples of their COCA search for quantifiers  
Figure 3  
Students' examples of their COCA search for Some and Any  
The third synchronous session entailed a lexical review as a lead-in to food containers.  
Then, to facilitate the content of How much...? and the quantifiers a lot, a little, not much, and  
none, the teacher provided students with a handout. It contained contextualized examples taken  
from COCA along with two key questions to trigger inferences about its use. Once again, students  
worked collaboratively in Zoom Breakout Rooms to thoroughly analyze and identify patterns.  
After responding to the questions, the participants had to produce two examples following the  
Esta obra se comparte bajo la licencia Creative Commons Atribución-No Comercial 4.0 International (CC BY-NC 4.0)  
Revista de la Universidad Internacional del Ecuador. URL: https://www.uide.edu.ec/  
43  
Hover Ismael Conza-Armijos y Liliana Fernanda Celi-Celi  
ISSN 2477-9024. Innova Research Journal (Septiembre-Diciembre, 2023). Vol. N8, No. 3, pp. 37-50  
identified patterns. Finally, to check comprehension, the teacher called on some students to share  
their examples on the Zoom whiteboard and gave corrective feedback when required. As part of  
the asynchronous activities, learners focused on completing the fourth flap of their digital foldable  
writing two examples per category.  
Figure 4  
Example of DDL materials introducing How much  
Figure 5  
Example of DDL materials introducing quantifiers  
The last synchronous session complemented the grammar topic of quantifiers, and it  
followed the same approach as the previous session. Students had to work in Zoom Breakout  
Esta obra se comparte bajo la licencia Creative Commons Atribución-No Comercial 4.0 International (CC BY-NC 4.0)  
Revista de la Universidad Internacional del Ecuador. URL: https://www.uide.edu.ec/  
44  
Data-Driven Learning in an EFL class: a study of Ecuadorian learners’ perceptions  
Rooms for a deep analysis of examples, but on this occasion, with an emphasis on How many...?  
and the quantifiers a lot, a few, many, and any. To wrap up the unit, the teacher reviewed the  
grammar content and utilized a formative assessment on Quizziz.  
Data Collection  
To gather data about students' attitudes towards DDL use, we relied on three instruments:  
a questionnaire, a reflection that was included in a digital foldable, and a follow-up interview at  
the end of the treatment.  
The construction of the questionnaire required a sound review of the literature which led  
to the selection of 15 statements adapted from Yoon and Hirvela (2004), Geluso and Yamaguchi  
(
2014), and Boontam and Phoocharoensil (2018). The questionnaire comprised four categories:  
tool, grammar, experience with DDL, and future application. The participants rated the statements  
using a four-point Likert scale (4: strongly agree, 3: agree, 2: disagree, 1: strongly disagree).  
At the end of the intervention, participants had to complete a reflection section mainly to  
evaluate DDL. The reflection prompts were written in students' L1 so they could express their  
ideas clearly irrespective of their English proficiency level (A2). This instrument served two  
purposes: to underpin the creation of the interview questions, and to bridge the information from  
the other instruments.  
Finally, the researchers conducted an individual semi-structured interview with 5  
volunteers (4 females and 1 male) to have a more comprehensive view of their experience. The  
questions were devised to elaborate on the information provided by participants in the  
questionnaire and the reflections. The researchers interviewed participants in Spanish to aim for a  
more comfortable interaction and avoid any misunderstandings due to participants' English  
proficiency level. The recordings were transcribed and then translated into English by the  
researchers.  
Results and discussion  
Table 1  
Perceptions about the corpus tool  
Statement  
Strongly  
agree  
Agree  
Disagree Strongly Mean Mode SD  
disagree  
n
%
n
%
n
%
n
%
1
2
.
.
The searching tool  
KWIC was easy to  
use.  
I had difficulty in  
retrieving the  
examples with the  
target forms.  
3
21 10 71  
1
7
3,14  
2,86  
3
3
0.51  
0.64  
2
14 57  
8
4
29  
Esta obra se comparte bajo la licencia Creative Commons Atribución-No Comercial 4.0 International (CC BY-NC 4.0)  
Revista de la Universidad Internacional del Ecuador. URL: https://www.uide.edu.ec/  
45  
Hover Ismael Conza-Armijos y Liliana Fernanda Celi-Celi  
ISSN 2477-9024. Innova Research Journal (Septiembre-Diciembre, 2023). Vol. N8, No. 3, pp. 37-50  
The items for the first category are related to the evaluation of COCA, particularly the use  
of KWIC. The results reflect an overall positive trend regarding the use of KWIC. Further data in  
the journals showed that the tool is described as "easy to use and intuitive", "user-friendly and  
manageable", and "free of charge".  
However, a great majority mentioned that they had difficulty in retrieving examples, which  
can be attributed to the little or no previous experience with COCA. These perceptions are  
supported by the interviewees who stated that they had not used COCA before besides the initial  
guidance. It was found that, for some students, the interface was complex to navigate at first, they  
lacked knowledge on how to navigate the tool, and they had to follow multiple steps to conduct  
the word search. Moreover, they also reported that attempting more than 10 searches constituted a  
hindrance to retrieving fast information.  
Table 2  
Perceptions about grammar learning  
Statement  
Strongly Agree Disagree Strongly Mean Mode SD  
agree disagree  
n
%
n
%
n
%
n
%
3
4
.
.
The real texts in the  
corpus are too difficult to  
understand.  
The use of COCA was  
good for understanding  
the differences between  
countable and  
uncountable nouns  
I like to discover the  
language patterns by  
myself.  
Learning through DDL  
makes me understand the  
lesson better.  
5
7
36  
3
21  
3
21  
3
21  
2,71  
3,50  
4
1.16  
0.5  
50  
36  
7
8
50  
57  
4,3 *  
5
6
7
.
.
.
5
3
1
7
3,29  
3,21  
3
3
0.59  
0.45  
21 11 79  
I completely understand  
how to use how much  
and how many by  
learning through DDL  
method.  
5
36  
9
64  
3,36  
3
0.48  
Note* Responses occurred equally.  
The statements for the second category revolved around grammar learning aided by COCA.  
It is observed that the responses for the first item about real texts being difficult are evenly  
distributed among the four options. Yet, more respondents are inclined to the agreement end of the  
continuum. Despite these results, some participants ascertained that COCA retrieves natural and  
complex sentences, contextualized and high-quality examples.  
Esta obra se comparte bajo la licencia Creative Commons Atribución-No Comercial 4.0 International (CC BY-NC 4.0)  
Revista de la Universidad Internacional del Ecuador. URL: https://www.uide.edu.ec/  
46  
Data-Driven Learning in an EFL class: a study of Ecuadorian learners’ perceptions  
Regarding the preferred grammar learning approach (See item 5), the majority of students  
pointed out that they like to discover language patterns, which favors the inductive approach. To  
this end, the interviewees claimed that COCA offers well-structured examples, which eases  
comprehension. Overall, most participants were satisfied with DDL implementation because it  
enhanced the lesson delivery, which increased understanding of the target grammar.  
Table 3  
Experience using DDL  
Statement  
Strongly Agree Disagree Strongly Mean Mode SD  
agree disagree  
n
%
n
%
n
%
n
%
8
9
1
1
.
.
I understand the purpose  
of using the corpus in  
this course.  
The guidance from the  
teacher was efficient to  
explore COCA.  
6
9
2
43  
7
50  
1
7
3,36  
3,64  
2,79  
3
4
3
0.61  
0.48  
0.77  
64  
14  
5
8
36  
57  
0. I think that DDL method  
is more challenging than  
other traditional methods  
1. I was encouraged to  
direct my own learning,  
actively think, and  
express my reflections  
during the DDL sessions  
2. When I learn grammar, I  
prefer a learner-centred  
method like DDL.  
3
4
21  
29  
1
7
4
3
29 10 71  
3,29  
2,93  
3
3
0.45  
0.70  
1
21  
7
50  
This third category sought information pertaining to participants' experience with DDL.  
One of the most salient details from this section was that participants were aware of the rationale  
behind the use of corpus. To put it simply, students highlighted that COCA has a plethora of  
information paired with authentic examples taken from different contexts. Interestingly, students  
identified that COCA allows for increasing vocabulary as an additional benefit.  
Likewise, the majority of participants acknowledged the guidance from the teacher as  
efficient to explore COCA. It is worth mentioning that students only received one entire  
introductory session to navigate COCA.  
The remaining three items (10, 12, and 13) illustrate that students perceive DDL as more  
challenging than traditional methods and helpful for fostering autonomous learning. That is the  
case of one of the interviewees who mentioned that she was encouraged to think critically when  
choosing examples. Another interviewee believed that DDL stimulated active participation. This  
is also supported by one participant who wrote in his journal: "DDL promotes independent  
research".  
Esta obra se comparte bajo la licencia Creative Commons Atribución-No Comercial 4.0 International (CC BY-NC 4.0)  
Revista de la Universidad Internacional del Ecuador. URL: https://www.uide.edu.ec/  
47  
Hover Ismael Conza-Armijos y Liliana Fernanda Celi-Celi  
ISSN 2477-9024. Innova Research Journal (Septiembre-Diciembre, 2023). Vol. N8, No. 3, pp. 37-50  
Table 4  
Future application  
Statement  
Strongly Agree Disagree Strongly Mean Mode SD  
agree disagree  
n
%
n
%
n
%
n
%
1
1
1
3. I will use the corpus for  
my English learning in  
the future.  
4. I recommend using DDL  
for the same unit with  
future cohorts.  
5. I want to study other  
English lessons through  
DDL method in future  
units.  
2
5
14 11 79  
1
7
3,07  
3,36  
3
3
0.46  
0.48  
36  
29  
9
9
64  
64  
4
1
7
3,21  
3
0.56  
As for the last category about future applications, the majority of the participants indicated  
that they are likely to employ COCA in the future. In fact, the journal analysis revealed that 12  
respondents would use COCA in the long term, and all the interviewees corroborated this.  
Similarly, most participants agreed with item 14, thus recommending DDL to cover the same unit.  
Finally, when participants were asked about implementing DDL in future units, the overall  
response to this question was very positive.  
Conclusions and pedagogical implications  
This study set out to explore EFL learners perceptions towards the implementation of a  
DDL approach for grammar learning in an Ecuadorian university. Based on the information  
obtained from the questionnaire, journals, and interviews, it is evident that students had positive  
impressions of the corpus-based approach. The participants described COCA as user-friendly and  
easy to navigate, and those findings are consonant with the results of Tribble's study in 2015.  
Besides, participants highlighted the use of corpus as an additional language exposure to a wide  
range of natural and authentic examples. Nonetheless, some students reported difficulties  
navigating COCA due to their little or no previous experience with the tool. A common  
shortcoming was that searches were limited to 10 if users had not registered previously.  
Another emerging point derived from this study shows that students prefer an inductive  
approach because they mentioned that it guides grammar discovery and enhances their  
comprehension. These results echo the findings of Flowerdew (2015) who believes that DDL and  
a discovery-based approach are inexorable given that learners induce language rules from the  
corpus. Flowerdew further asserts that inductive learning is closely linked with discovery learning  
and triggers higher-order thinking skills such as noticing, analyzing, and making inferences.  
The present study provides additional evidence with respect to the potential benefits of  
DDL in language learning that researchers have already reported in numerous studies. Learners  
recognized the usefulness of DDL as it encouraged active thinking about the target grammar.  
Despite the positive feedback, it is important to note that teachers' time can be compromised when  
Esta obra se comparte bajo la licencia Creative Commons Atribución-No Comercial 4.0 International (CC BY-NC 4.0)  
Revista de la Universidad Internacional del Ecuador. URL: https://www.uide.edu.ec/  
48  
Data-Driven Learning in an EFL class: a study of Ecuadorian learners’ perceptions  
developing materials and becoming familiar with the corpus tools. We concur with Römer (2007),  
and we advocate for a DDL-friendly environment for both teachers and students where they can  
receive appropriate guidance to access corpora. Only then will they be able to exploit other  
resources that corpus consultation tools offer. Ecuadorian EFL teachers could implement the DDL  
methodology in an endeavor to promote learner autonomy, particularly useful amidst remote  
instruction.  
Limitations and further research  
This study was based on pioneering research and was adapted to an unprecedented fully  
virtual environment. However, one of the major limitations was the low number of participants;  
therefore, it is difficult to generalize the findings. Although the intervention was conducted  
throughout a unit, it was short in length as it only lasted 3 weeks. Further research is called for to  
explore DDL use with a longitudinal study over a term or a semester. To this extent, a pre and a  
post-test could be valuable tools to gather quantitative data to support and contrast students' claims.  
Provided that this study was solely focused on gathering students' perceptions of DDL use  
for grammar learning, future research could investigate the effect of DDL on other language  
components. For example, given the importance of vocabulary for accurate and appropriate  
language use, studies on collocations could be conducted to increase students' lexical competence.  
Within the same category, researchers may consider comparing the learning of synonyms and near-  
synonyms such as make/do or win/gain/earn.  
Bibliographic References  
Biber, D., & Reppen, R. (2002). What does frequency have to do with grammar teaching? Studies  
in  
Second  
language  
acquisition,  
24(2),  
199208.  
https://doi.org/10.1017/S0272263102002048  
Boontam, P., & Phoocharoensil, S. (2018). Effectiveness of English preposition learning through  
data-driven learning (DDL). 3L: Language, Linguistics, Literature, 24(3), 125141.  
https://doi.org/10.17576/3L-2018-2403-10  
Boulton, A. (2012). What data for data-driven learning? EuroCALL Review, 20(1), 2327.  
https://bit.ly/484oO1c  
Cambridge  
Assessment  
English.  
(2021).  
A2  
Key  
Preparation.  
https://www.cambridgeenglish.org/exams-and-tests/key/preparation/  
Cheng, W. (2010). What can a corpus tell us about language learning. In The Routledge handbook  
of corpus linguistics (pp. 319332). Abingdon, Oxon: Routledge.  
Cheng, W. (2012). Exploring corpus linguistics: Language in action. Banbury: Routledge.  
Conrad, S. (2010). What can a corpus tell us about grammar? In The routledge handbook of corpus  
linguistics (pp. 227240). https://doi.org/10.4324/9780203856949.ch17  
Crawford, W., & Csomay, E. (2016). Doing corpus linguistics. New York: Routledge.  
Davis, M. (2008). The corpus of contemporary American English (COCA). https://www.english-  
corpora.org/coca/  
Flowerdew, L. (2012). Applying corpus linguistics in teaching arenas. Corpora and Language  
Education, 190232. https://doi.org/10.1057/9780230355569_7  
Esta obra se comparte bajo la licencia Creative Commons Atribución-No Comercial 4.0 International (CC BY-NC 4.0)  
Revista de la Universidad Internacional del Ecuador. URL: https://www.uide.edu.ec/  
49  
Hover Ismael Conza-Armijos y Liliana Fernanda Celi-Celi  
ISSN 2477-9024. Innova Research Journal (Septiembre-Diciembre, 2023). Vol. N8, No. 3, pp. 37-50  
Flowerdew, L. (2015). Data-driven learning and language learning theories. Amsterdam: John  
Benjamins.  
Gavioli, L. (2005). Exploring corpora for ESP learning. Amsterdam: John Benjamins.  
Geluso, J., & Yamaguchi, A. (2014). Discovering formulaic language through data-driven  
learning:  
Student  
attitudes  
and  
efficacy.  
ReCALL,  
26(2),  
225242.  
https://doi.org/10.1017/S0958344014000044  
Guan, X. (2013). A Study on the application of Data-driven learning in vocabulary teaching and  
learning in China’s EFL Class. Journal of Language Teaching and Research, 4(1), 105–  
1
12. https://doi.org/10.4304/jltr.4.1.105-112  
Huang, L. S. (2011). Corpus-aided language learning. ELT Journal, 65(4), 481484.  
https://doi.org/10.1093/elt/ccr031  
Johansson, S. (2009). Some thoughts on corpora and second language acquisitions. Corpora and  
Language Teaching (pp. 3344). https://doi.org/10.1075/scl.33.05joh  
Johns, T. (1991). Should you be persuaded two samples of data-driven learning. ELR Journal, 4,  
116. https://bit.ly/45vsGGP  
Lin, M. H., & Lee, J. Y. (2019). Pedagogical suitability of data-driven learning in EFL grammar  
classes: A case study of Taiwanese students. Language Teaching Research, 23(5), 541–  
5
61. https://doi.org/10.1177/1362168817740899  
Lindquist, H. (2009). Corpus linguistics and the description of English. Edinburgh: Edinburgh  
University Press.  
Luo, Q. (2016). The effects of data-driven learning activities on EFL learners’ writing  
development. SpringerPlus, 5, 1255. https://doi.org/10.1186/s40064-016-2935-5  
O’Keeffe, A. (2020). Data-driven learning- A call for a broader research gaze. Language  
Teaching, 114. https://doi.org/10.1017/S0261444820000245  
O’Keeffe, A., McCarthy, M., & Carter, R. (2007). From corpus to classroom: Language use and  
language teaching. Cambridge: Cambridge University Press.  
Pérez-Paredes, P., Ordoñana Guillamón, C., Van de Vyver, J., Meurice, A., Aguado Jiménez, P.,  
Conole, G., & Sánchez Hernández, P. (2019). Mobile data-driven language learning:  
Affordances  
and  
learners’  
perception.  
System,  
84,  
145159.  
https://doi.org/10.1016/j.system.2019.06.009  
Reppen, R. (2010). Using corpora in the language classroom. Cambridge: Cambridge University  
Press.  
Römer, U. (2007). Corpora and language teaching. 112131.  
Smart, J. (2014). The role of guided induction in paper-based data-driven learning. ReCALL, 26(2),  
1
84201. https://doi.org/10.1017/S0958344014000081  
Smirnova, E. A. (2017). Using Corpora in EFL Classrooms: The Case Study of IELTS Preparation.  
RELC Journal, 48(3), 302310. https://doi.org/10.1177/0033688216684280  
Szudarski, P. (2018). Corpus linguistics for vocabulary: A guide for research. London: Routledge.  
Tribble, C. (2015). Teaching and language corpora. In Multiple affordances of language corpora  
for Data-driven Learning. Amsterdam: John Benjamins.  
Yoon, H., & Hirvela, A. (2004). ESL student attitudes toward corpus use in L2 writing. Journal of  
Second Language Writing, 13(4), 257283. https://doi.org/10.1016/j.jslw.2004.06.002  
Esta obra se comparte bajo la licencia Creative Commons Atribución-No Comercial 4.0 International (CC BY-NC 4.0)  
Revista de la Universidad Internacional del Ecuador. URL: https://www.uide.edu.ec/  
50