Los
efectos del método de enseńanza de idiomas basado en tareas en la habilidad
oral: Una síntesis de investigación sistemática y meta análisis.
Alberto Medina Fernández. [1]
Recibido: 12-08-2021 / Revisado: 22-08-2021 /Aceptado=
: 09-09-2021/
Publicado: 05-10-2021
=
Abstract <=
/span> DOI: =
https://doi.org/10.33262/cienciadigital.v5i4.1801
Introduction. For more than three decades Task-based language
teaching strategies have constituted an object of research in second langua=
ge
acquisition. However, a TBLT approach has been fully embraced in some parts=
of
the world while not in others. Despite the fact that there have been previo=
us
systematic reviews and meta-analyses investigating TBLT implementation, the=
re
is still the need for a synthesis of the effects of tasks on oral productio=
n.
Objective. The study aims to estimate the effects of TBLT
implementation on the speaking skill development holistically and through t=
he
components of complexity, accuracy and fluency. Methodology. This pa=
per
provides a systematic research synthesis of different TBLT interventions on
spoken production. In total, 61 published studies were retrieved, coded, and
analyzed following established coding procedures for systematic reviews.
Additionally, 11 studies were also meta-analyzed so as to research the over=
all
effects of the application of a TBLT speaking strategy as treatment in
different contexts. Results. The findings unveiled a medium to large
effect of (d=3D0.78) for TBLT interventions on oral production. The synthes=
is
across qualitative and quantitative data also revealed a positive impact of
Task Repetition and Pre-task planning for the enhancement of linguistic
complexity, accuracy and fluency dimensions of the speaking skill. Concl=
usion.
The study concludes with implications for the implementation of a TBLT
strategy in L2 teaching-learning process. It is evidenced that a TBLT strat=
egy
offers improvement possibilities for the development of the speaking skills=
in
intermediate level students.
Keywords: TBLT, Pedagogical Task, Speaking Skill, Spoken
Interaction
Resumen
Introducción.
Durante más de tres décadas, las estrategias de enseńanza de lenguas basada=
s en
tareas han constituido un objeto de investigación en la adquisición de segu=
ndas
lenguas. Sin embargo, un enfoque TBLT se ha adoptado plenamente en algunas
partes del mundo, pero no en otras. A pesar de que ha habido revisiones
sistemáticas y metanálisis
previos que investigan la implementación de TBLT, todavía existe la necesid=
ad
de una síntesis de los efectos de las tareas en la producción oral. Objetivo. El estudio tiene como
objetivo estimar los efectos de la implementación de TBLT en el desarrollo =
de
la habilidad de hablar de manera integral y a través de los componentes de
complejidad, precisión y fluidez. =
Metodología.
Este artículo proporciona una síntesis de investigación sistemática de
diferentes intervenciones TBLT sobre la producción hablada. En total, se
recuperaron, codificaron y analizaron 61 estudios publicados siguiendo los
procedimientos de codificación establecidos para las revisiones sistemática=
s.
Además, también se realizaron metanálisis de 11
estudios para investigar los efectos generales de la aplicación de una
estrategia de habla TBLT como tratamiento en diferentes contextos. Resultados. Los hallazgos revelar=
on un
efecto de medio a grande de (d =3D 0,78) para las intervenciones de TBLT en=
la
producción oral. La síntesis de datos cualitativos y cuantitativos también
reveló un impacto positivo de la repetición de tareas y la planificación pr=
evia
a la tarea para mejorar la complejidad lingüística, la precisión y las
dimensiones de fluidez de la habilidad de hablar. Conclusión. El estudio concluye con implicaciones para la
implementación de una estrategia TBLT en el proceso de enseńanza-aprendizaj=
e de
lenguas extranjeras. Se evidencia que una estrategia TBLT ofrece posibilida=
des
de mejora para el desarrollo de la habilidad oral en estudiantes de nivel
intermedio.
Palabras clave: TBLT,
Tarea pedagógica, Habilidad oral, Interacción oral.
Introduction
Task-based
Language Teaching rationale
TBLT methodology has been applied in different par=
ts
of the world as an alternative to other traditional methods for the develop=
ment
of communicative competence. As Ellis & Shintani
(2014) put it, TBLT aims at engaging students in meaning-focused communicat=
ion through
the performance of tasks. It is based on providing opportunities for learne=
rs
to focus not only on language but also on the learning process itself (Nuna=
n,
2004). While learners do the tasks they interact=
in
the target language to reach a pre-established outcome. Generally, the kind=
of
tasks used in TBLT tends to link classroom language learning with language =
use
outside the classroom (Nunan, 2004). Ever since the advent of the communica=
tive
approach to second or foreign language teaching, language educators and
policy-makers have attempted to devise new ways of teaching oral production=
in
language courses in more productive and meaningful ways. One of these propo=
sals
was Task-based language teaching which has been extensively researched
throughout the last 30 years as an approach that attempted to increase
students’ involvement in more meaningful, real-world and as part of
form-function relationship in language acquisition (Van den Braden, 2006). =
Thus,
there persists the need to synthesize the current available research and
corroborate the advantages and limitations of TBLT practices and potentiali=
ties
for Second or Foreign Language Oral Production enhancement. The present stu=
dy
represents a synthesis of the body of work that takes into account the impa=
ct
of TBLT interventions in the acquisition and development of speaking skills=
in
Second Language and Foreign language learners.
Tasks as the
driver of the teaching-learning process
In TBLT, real-world tasks become pedagogical tasks
that resemble communication in the real world. However, to define tasks, on=
e of
the most important aspects to take into account is that, as Richards &
Rodgers (2001) stated, tasks do not necessarily involve the production of
language since they further the processes of meaning negotiation, modificat=
ion,
and experimentation. Therefore, tasks can be oriented towards comprehensible
input and to the production of language while emphasizing meaning exchange.=
In
essence, pedagogical tasks are classroom pieces of work that engage the
learners in comprehension, manipulation, and interaction; or language
production while mobilizing their grammatical knowledge to express meaning
(Nunan, 2004). While carrying out tasks, learners are free to use different
grammatical structures and linguistic resources previously acquired, to ach=
ieve
an outcome.
Framework for =
TBLT
and task types
At a syllabus level, tasks must be linked together
through ‘task chaining’ (Nunan, 2004). This means that they are organized
according to units of work, and at a unit level, in terms of functions, gra=
mmar
and lexis that are expressed through them.
In a task-based lesson, the general agreed on phas=
es
that have been proposed are pre-task (can be seen as a planning phase), main
task (performing the task in groups), and post Task or Language Focus phase=
(a
focus on reporting language, and extra language work needed).
As stated earlier, the types of tasks in TBLT are
those that resemble what people do with language in the real world. Pedagog=
ical
tasks have been classified according to different criteria and aims. Thus,
tasks can have a rehearsal rationale or an activation rationale. They can b=
e sorted
in terms of the operations required, such as listing, ordering, comparing,
problem-solving, sharing information, or creation. Besides, tasks could inc=
lude
an information gap, an opinion gap, or a reasoning gap as defined by Prabhu,
N.S. (1987). These represent the most frequent tasks for the development of
speaking skills, which will be analyzed in this work.
Focus on form =
in
tasks
A much-debated topic on TBLT research and applicat=
ion
has been whether communicative tasks employed in a task-based approach negl=
ect
a focus on linguistic forms. While an explicit focus on form in tasks is no=
t a feature
of strong versions of TBLT, early versions of task cycles included a focus =
on
grammatical forms in the pre-task phase. More recently, a combination of fo=
cus
on form and expression of meaning has been incorporated in the task cycles =
utilizing
recasts and prompts that offered the learners a more effective control of f=
orms
while expressing themselves.
Methodology
The =
current
study followed a bibliographical design characteristic of systematic reviews
and meta-analysis which includes the search and selection of studies accord=
ing
to a set of criteria. The particular method chosen for this study was mixed
research synthesis due to the nature of the research questions aimed at
exploring the most common effects and practices, as well as the prevalence =
of
pretest-experimental-posttest designs, generally used in TBLT research. Thu=
s,
it required a mixed-method synthetic approach for the integration of the
findings of several published qualitative, quantitative and mixed primary l=
evel
articles (Heyvaert, M. et al., 2011).
Search keywords
At the beginning of this research synthesis, keywo=
rds
related to TBLT and the SPEAKING SKILL were established as:
(TASK-BASED LANGUAGE TEACHING / PEDAGOGICAL TASKS/
COMMUNICATIVE TASKS / SPEAKING SKILLS /SPOKEN INTERACTION/ FLUENCY/
PRONUNCIATION/ L2 ORAL PRODUCTION)
To
conduct a thorough search, the following combinations were used to retrieve
reports about the intended research questions:
Combinations
used:
[“TASK-BASED
LANGUAGE TEACHING”] + [“COMMUNICATIVE TASKS”] + [“PEDAGOGICAL TASKS”] + [“S=
PEAKING
SKILLS”] + [“SPOKEN INTERACTION”] + [“FLUENCY”] + [“PRONUNCIATION”] + [“L2 =
ORAL
PRODUCTION”]
Data sources:
Firstly, an initial search was conducted using the
search keywords through combinations on the following databases and journal=
s:
DATABASES:
1-&n=
bsp;
ERIC
2-&n=
bsp;
SCIENCEDIRECT
3-&n=
bsp;
MICROSOFT
ACADEMIC
Additionally, searches w=
ere
carried out in GOOGLE SCHOLAR, and the following journals:
JOURNALS
•
Applied Linguistics •
Advances in Language and Literary
Studies • ELT Journal • English Language Teaching • International
Journal of Applied Linguistics • Journal of English as an International
Language •
Language Teaching Research • Modern Language Learning •Sage Journal •TESOL Quarterly=
• The
Language Learning Journal.
It was searched on articles’ abstracts, titles, and
keywords. The literature search was
conducted on referred journals that focus on SLA research and were only
peer-reviewed journals.
It resulted in 172 articles.
Inclusion and exclusion criteria<= o:p>
This section gives an account of the procedure and
steps taken to search for the published literature pertinent to this resear=
ch
synthesis. It also describes the inclusion criteria used to filter out
documents that did not pertain to the object of this systematic review or d=
id
not provide relevant data for the research questions proposed. As a second
step, the retrieved documents were sifted through a set of pre-established
criteria for their inclusion or exclusion. As the last step, studies were
evaluated according to the research questions posed previously. The documen=
ts
selected had to satisfy the eligibility criteria as follows:
1)&n=
bsp;
the
study was mainly about developing speaking and L2 oral production within TB=
LT
frameworks.
2)&n=
bsp;
all
studies were dated from a 15-year-span, from 2006 to 2021, due to an increa=
se
of interest in task-based language teaching research publications.
3)&n=
bsp;
all
studies had an intervention on the speaking skills as a whole, or one or mo=
re
of its aspects such as fluency, interaction, comprehension, accuracy, input=
or
output constructs.
Inclusion and exclusion process
As
previously stated, the queries led to a total pool of 172 articles that mat=
ched
the selection criteria and resulted from the keyword combinations. This num=
ber
was then reduced by the following elimination procedure:
Phase
1: elimination based on duplication of titles from various sources which
reduced the number from 172 to 141.
Phase
2: after examining titles and abstracts, a reduction from 141 to 122.
Phase
3: elimination based on the reports after reading the text, from 122 to 61.=
At
this phase, reports were excluded on the basis of the directions, quality, =
and
interventions.
The
flow of the elimination procedure for screening the articles is represented=
in
Figure 1.
Figure
1
PRISMA flowchart
for article inclusion.
Identification of studies via databas=
es
and registers Screening PHASE
1: Databases (n =3D89) Registers and journals (n =3D83=
) Duplicate Articles removed (n =3D 31) Articles removed for other reasons (n =3D0) PHASE 2: PHASE 3: (n =3D34=
) Identification
Included
<=
/span> =
Source: Page, M.J. et =
al.
(2020). The PRISMA 2020 statement: an updated guideline for reporting
systematic reviews.
Developed by: The
author
Coding the data from reports
For the
analysis-synthesis process of the included reports, a coding scheme was
developed to classify the studies according to their substantive and
methodological features. It was used to organize the data relevant to the
research questions proposed and includes the features that appear in Table =
1.
Table
1
Coding
Procedure
|
Variable |
Operationalization |
|
Publication Characteristics: |
Name of author(s), Title of publication |
|
Author, Title, Journal |
Journal name and features |
|
Design Characterist=
ics: |
|
|
Setting |
Foreign or Second Language |
|
Institution Type |
Elementary, High school, University,
Institute =
|
|
Country |
The country where the intervention took place =
|
|
TBLT Program |
The presence or absence of a TBLT syllabus |
|
Duration |
Length of treatment in hours (total class time=
) |
|
Modality =
|
Face-to-face, multimedia, online or multiple |
|
Participant
Characteristics: |
|
|
Participants
L1 =
|
First
language of the participants |
|
Participants
L2 =
|
Target
Language |
|
Proficiency
level =
|
Beginner,
intermediate, advanced, multiple |
|
Methodology: |
|
|
Research Design <= o:p> |
Experimental, quasi-experimental, qualitative<=
span
class=3DGramE>, , mixed-method. |
|
Dependent Variable(s) =
|
TBLT intervention |
|
Outcome Variable(s) |
Speaking skill dimensions |
|
Speaking skill measures |
CAF measures, Oral Production measurements |
|
Pre-test
|
Inclusion of a pre-test |
|
Delayed post-test |
Inclusion of a delayed posttest |
|
Instrument reliability |
Reliability reported for outcome measure instr=
ument =
|
|
Statistical test |
Type of statistical tests |
|
Effect size
|
Effect sizes reported, or sufficient data for
calculation |
|
Findings |
Interpretation of results |
|
Study Limitations |
Limitations found |
Source: Coding categories for research synthesis=
and
meta-analysis. (Norris & Ortega, 2006)
Developed by: The
author
This type of coding allowed for t=
he retrieval
of both qualitative and quantitative information to be synthesized. The kin=
ds
of interventions more frequently used by researchers are described in Table=
2.
Analysis
In general, the most commonly used independent
variables for oral production emerging from the analysis are TBLT Framework=
and
Holistic Strategy, Task repetition, Pre-task planning conditions, Task
sequences, Focus on form, Mobile, and Web-supported TBLT.
Table 2
TBLT types of interventions in the included stud=
ies
|
Intervention
type =
|
No.
of studies |
|
TBLT Holistic Strategy |
21 =
|
|
Task Repetition =
|
10 |
|
Task Planning =
|
8 |
|
Task Types =
|
7 |
|
Task Complexity =
|
4 |
|
ICT-supported TBLT =
|
3 |
|
Task Sequence =
=
|
2 =
|
|
Pre-task Modelling =
|
2 |
|
Task Rehearsal =
|
2
|
|
ACCESS =
=
|
1 |
|
Scaffolding =
|
1 |
|
|
T=3D61 |
Source: Number of stud=
ies
reviewed by intervention
Developed by: The
author
The Quantitative
Meta-analysis
After t=
he
application of the coding scheme to the included articles, 11 studies were
selected for meta-analysis. These studies were chosen based on the presence=
of
an experimental or quasi-experimental pre-posttest design, as well as having
TBLT holistic strategy as a treatment for the experimental group, versus ot=
her
methodological strategies or teaching methods for the control group. Coded
studies that had a descriptive or correlational design were not included. F=
urthermore,
if studies had TBLT types of intervention on both experimental and control
groups, they were also not included in the analysis of effect sizes. For the
effects of treatment and changes between pretest and posttest, Cohen’s d
formula was used with a 95% confidence interval. A correction factor was
employed for samples smaller than 50 participants, Hedges, L.V. (1981):
Theoretical Framework=
Development of=
the
components and dimensions of the skill of speaking in an ESL or EFL
teaching-learning process.
One of the most recurrent concerns in the SLA or FL
teaching-learning process has been to develop speaking skills in the learne=
rs.
As a result, there has been a shift towards approaches and methods that foc=
us
on the process of developing this skill, such as CLT and TBLT. To develop t=
his
skill effectively, its different components must be analyzed since they
constitute the subject matter of most SLA/FL courses. In fact, the componen=
ts
of speaking have been extensively researched by TBLT practitioners and rese=
archers
all over the world to determine the effectiveness of communicative tasks
previously applied. Therefore, it is necessary to determine the theoretical
underpinnings of the speaking skill and their significance in TBLT empirical
studies. To do so, it is paramount to ponder on the dimensions of complexit=
y,
accuracy, and fluency which have been frequently operationalized as the CAF
triad in applied linguistics research (Housen e=
t al.
2012). As well as, the particulari=
ties
of negotiation of meaning, interaction, and output which are aspects that
researchers have systematically paid attention to.
Comprehensible
input
For learners to be able to speak and communicate
effectively in L2 there must be tasks that are geared towards comprehensible
input in the first place. A speaker must understand first what is being sai=
d in
order to engage in verbal exchanges effectively. According to Pica et al.
(1987), all learners in the second language classroom should participate in
interaction and ask questions in order to clarify input. In a sense, language input must precede
output. Language input can be oriented towards content, or it can be orient=
ed
towards linguistic forms.
Complexity
The notion of complexity has been defined by sever=
al
authors and researchers in applied linguistics. In general terms, complexity
has been understood as features of language that are particularly difficult=
for
learners to make use of. For the purpose of this synthesis and for overall =
L2
performance analysis, complexity can best be defined as comprising three
components: propositional complexity, discourse interactional complexity, a=
nd
linguistic complexity (Housen et al. 2012). Whe=
re
propositional complexity refers to the number of idea units that a speaker
processes in a task. Discourse interactional complexity is the number and t=
ypes
of interactional moves that a learner takes. And finally, linguistic
complexity, which has been defined as the complexity in terms of lexical,
morphological, syntactic, and phonological features used by learners (Housen et al. 2012).
Accuracy
Since the 1980s, researchers have tried to make a
distinction between fluent and accurate language usage to investigate L2 or=
al
proficiency (Housen and Ku=
iken
2009). What is referred to as accurate language production was expressed in
terms of error-free language output measured against a model, often that of=
the
native language speaker. Nonetheless, in a CAF triad model, accuracy must a=
lso
be operationalized in terms of how appropriate and acceptable the language
produced is (Housen et al. 2012).
Fluency
Fluency is one of the most important aspects of
speaking because being a fluent speaker of L2 has been equated to being a
proficient language user. In other words, fluent speakers respond quickly a=
nd
accurately in spoken interactions. Therefore, some command over a combinati=
on
of linguistic forms and communicative functions is needed for the achieveme=
nt of
fluency. The type of effective tasks in TBLT should include some kind of fo=
cus
on the aforementioned aspects. The measurement of fluency in speaking has t=
o do
with the phonological features expressed, such as the rate of speech, amount of pauses, and repetitions of the speaker while
engaging in interactions (Housen et al. 2012).<=
span
style=3D'mso-spacerun:yes'>
Interaction,
output, and negotiation of meaning.
The role of interaction, output, and negotiation of
meaning has become one of the integral parts of the speaking skill. In task=
s or
activities where students converse or exchange meaning towards reaching a
certain outcome, these components are present. Traditionally, the methods u=
nder
the communicative approach have been devised for the treatment of these thr=
ee
aspects. Conversational skills constitute a key part of the performance of =
TBLT
task cycles. Besides, learners need to understand the social context, roles,
and language functions they express. In
the same way, there must be a communicative output that fulfills the commun=
icative
objectives of tasks. As learners get involved in meaning exchange to reach a
common aim, conversational strategies are to be developed in the learners i=
n a
way that fosters the formation and use of these speaking skills. As previou=
sly
stated, the dimensions of complexity, accuracy, and fluency comprise the
linguistic and functional aspects that have been the object of research in =
TBLT
interventions to develop speaking skills in L2 programs. They have been def=
ined
as the constructs that measure the extent to which any pedagogical interven=
tion
can be regarded as effective or less effective in the appraisal of spoken
production. It has been upon that =
basis
that the current synthesis has been motivated. Ultimately, the motivation h=
as
been to take stock of the current research that measures the extent to which
TBLT methodology and its pedagogical tasks influence the aforementioned
constructs of the speaking skill in the learners. Therefore, the following
research questions guide this study:
Research question 1: <=
span
lang=3DEN-US style=3D'font-family:"Times New Roman",serif;color:windowtext'=
>What
are the benefits and limitations of the implementation of pedagogical task
frameworks in the development of spoken production constructs in terms of
(a)complexity, (b) accuracy, (c) fluency, and (d) spoken interaction in the=
se
studies?
Research question 2:
For
the investigation of the above questions, studies were categorized accordin=
g to
the type of TBLT framework, Task intervention and their overall effects on =
oral
production components.
Results
The research
synthesis
This section gives a broad account of the synthesi= s of the publications, participants, and types of interventions used; as well as= their effects on the different speaking constructs and dimensions. The aim is to provide a wider view on the impact of the treatments for complexity, accura= cy, fluency plus conversational skills, and negotiation of meaning dimensions.<= o:p>
Publications. The
totality of studies collected was varied in research design and objectives.
Studies were classified into 3 subsets as follows: studies that had TBLT ta=
sks
manipulations as independent variables and syntactic complexity, accuracy,
lexis and fluency measures (CALF) as dependent variables (n=3D38), qualitat=
ive
method studies (n=3D2); and between-group studies that compared TBLT strate=
gy
implementation in the experimental group versus other teaching methods in t=
he control
group (n=3D21). Out of 61 experimental studies published from 2006 to 2021,=
11
studies met the previously mentioned criteria for inclusion in the
meta-analysis.
Research designs and settings
The majority of studies used a between-groups
quasi-experimental design. Only 20% of studies used a mixed-method design w=
ith
the inclusion of qualitative techniques such as questionnaires and intervie=
ws.
Treatment tasks included roleplays, information-gap, opinion gaps, decision
making, problem-solving, and oral narrative tasks. The age of participants =
in
the studies ranged from 15 to 58 with the exception of one study where
participants were 9-11-year-olds. The most frequent L1s reported were Arabi=
c,
Chinese, Korean, Japanese, and Indonesian because of an increased research =
and
implementation of TBLT in the Asian-South pacific region. The target langua=
ge
contexts consisted of English as a Foreign Language (90%), English as a Sec=
ond
Language (8%), French as a Foreign Language (1%), German as a Foreign Langu=
age
(1%). The level of proficiency of participants was the following in studies
percentage: beginner level (26%), pre-intermediate level (19%), intermediate
level (53 %), advanced level (2%).
Statistical Analysis
The statistical data reported was varied and in some instances inconsistent. In total, (n=3D53)=
studies
reported statistical analyses. The following tests were employed: independe=
nt t-tests
(34%), paired samples t-test (24%), one-way ANOVA (13%), repeated-measures
ANOVA (11%), MANOVA (10%), Pearson Test for correlation (4%), Wilcoxon sign=
ed
ranks test (2%), and Mann–Whitney U-test (2%).
Effectiveness =
of
TBLT interventions on oral production
Outcome measur=
es
used for complexity, accuracy and fluency
The 38 studies that had complexit=
y,
accuracy and, or fluency as dependable variables used similar measures in t=
heir
operationalization. Complexity measures included the ratio of clauses to
AS-units (the Analysis of Speech Units), the nu=
mber
of complex AS-units. Fluency was measured as speech rate (number of
syllables per minutes). Accuracy measures included: (Errors per 100 words),
Percentage of error-free clauses.
Task types and
Task Variation Interventions on Complexity, Accuracy and Fluency.
For
measures of the CAF triad, the most recurring interventions used in 35 stud=
ies
have been; variations in the procedures for Task Repetition, Task Planning,=
Task
complexity, and Task Types. In regards to Task Types, the ones used to
intervene for measures of complexity, accuracy, and fluency have been
decision-making tasks, information-gap tasks, dialogic and monologic narrat=
ive
tasks, and picture-based description tasks. The studies that used task
repetition as an independent variable have evidenced gains for accuracy,
complexity, and more discretely for fluency. However, in one particular stu=
dy
where “massed repetitions” of the same task occurred 11 times over a
sixth-month period, statistically significant differences between the
experimental and control groups on all the posttest measures for complexity=
and
fluency were obtained (Ahmadian, J., 2011). Other studies using task repeti=
tion
have been based on Bygate and Samuda (2005)’s a=
rguments
about the role of task repetition in allowing L2 learners to pay attention =
to
both meaning and form in task execution, and thus, the consolidation of
language acquiredIn
these cases, task repetition has been found to aid learners to produce more
task-induced linguistic features and gains in complexity measures in the
subsequent repetitions. Furthermore, identical task repetition was found in=
one
study to raise L2 learners’ gains in accuracy by Gasha=
n
& Almohaisen (2014).
Some
other studies have paid attention to task planning conditions, task complex=
ity,
and task structure to measure its impact on the CAF dimensions. In this cas=
e,
it has been found that the most significant gains have appeared in fluency =
and
accuracy dimensions. In broad terms, it has been inferred from the studies’
results that a more structured and planned task sequence yields better resu=
lts
for complexity and fluency, and a little less evident for accuracy gains. B=
ased
on the Trade-off hypothesis (Skehan, 2009), and=
Levelt’s (1989) model of speech production, the studi=
es
used online planning, strategic planning, pre-task planning and planning
conditions (pressured vs. unpressured, individual vs. collaborative) as
independent variables. According to data form the studies, it can be
interpreted that online planning, where students are given ample time to pl=
an
their speech, aided their performance in terms of complexity and accuracy.
Strategic planning (allowing learners an amount of time to plan), and pre-t=
ask
planning strategies improved learner’s fluency and linguistic resources due=
to
opportunities to conceive and perfect messages before performing the actual
tasks.
On the other hand, precise findings on task types =
used
in these studies indicate that the use of picture-based description and
narration tasks increased the speed of speech production. Spot-the-differen=
ce
tasks, had a significant impact on accuracy improvement. Decision-making ta=
sks
were regarded as more conducive to greater complexity while information-gap
tasks had more effects on fluency development.
Accuracy
treatments in TBLT frameworks for Speaking Tasks and generation of LREs
(Language Related Episodes)
One of the aspects of the TBLT approach that has
received more attention from critics is attention to form within a task-bas=
ed
framework or lesson. For the case of oral production and interaction, resea=
rchers
in the studies included in the current synthesis analyzed the occurrence of
LREs during task interventions such as task repetition, task rehearsal,
pre-task planning, and ACCESS. With the inception of ACCESS, students gener=
ated
more grammatical LREs because this intervention is based on the repetition =
of
sets of utterances and expressions (Dao, P. et al 2016). Some other studies
explored the influence of task repetition, task rehearsal, and task complex=
ity
on form-focus and LREs generation. In sum, it can be stated that task
repetition favors the appearance of self-corrections and more use of L2
language during task performance. Interventional studies for accuracy measu=
res
are represented in table 3. In addition, both pre-task instruction and task
complexity seem to stimulate the production of more lexical LREs and accura=
te
speech production in terms of linguistic forms retrieved.
Table 3
Accuracy focus in Intervention
Treatments
|
Study T=
ests |
Sample<=
span
style=3D'mso-spacerun:yes'> |
Interve=
ntion =
=
|
Focus |
Measure=
s =
|
Stat |
|
Dao, P. et al. (2016) |
33 |
ACCESS =
|
LREs <=
/span> |
Number of LREs
|
- =
|
|
Hawkes, M. (2011) |
60 |
Post-task (Task repetition) |
Vocabulary
|
Conv. Analysis |
- |
|
Kim, Y. (2012) |
45 |
Pre-task Modelling |
Questions
|
Oral Tests <=
/span> |
Wilcoxon |
|
Park, S. (2010) |
110 |
Pre-task Planning |
LREs
|
Number of LREs
|
ANOVA |
|
Salimi, A. (2015) =
|
60 |
Task Complexity
|
Accuracy <=
/span> |
CAF measures =
|
ANOVA |
|
Nguyen, T. (2019) |
48 |
Task Rehearsal |
LREs
|
Number of LREs
|
ANOVA |
Source: Reviewed studies
focusing on accuracy measurements
Developed by: The
author
TBLT effective=
ness
on spoken interaction, negotiation of meaning, and linguistic features.
Out of the corpus of studies in this review, 3 stu=
dies
have yielded a better insight into relations between collaboration and
interaction in oral task performances. There is evidence of a positive impa=
ct
of peer-scaffolding strategies in conversations since the three studies
reported an increase in LRE episodes, as well as better accuracy, vocabular=
y, and
incorporation of linguistic terminology. The inclusion of multimedia-suppor=
ted
TBLT provided learners with further scaffolding strategies which had an imp=
act
on the incorporation of more vocabulary, and linguistic features in
conversation.
Effectiveness =
of a
TBLT framework on holistic approach measures of the speaking skill
It was found that in 21 studies there appeared a
holistic view to measuring the speaking skill. These studies used a
quasi-experimental design with a pretest-experiment-posttest approach, exce=
pt for
one study which used a qualitative approach with descriptive data focusing =
on
words rather than statistics by Yuniarisda, H. =
(2014).
In general, most studies reported gains in the means of experimental groups’
posttests, with only one exception. After data analysis of results was
conducted, it is inferred that the components that benefited more after the
intervention were negotiation of meaning and conversation skills. Few studies reported gains in fluency and
pronunciation. On the whole, the aspect that has been highlighted in the
studies is the increase in students’ involvement in oral production activit=
ies.
The main drawbacks have been found to be the fact that in 70% of the studie=
s a
small sample was used, and the inclusion of measurements other variables su=
ch
as self-efficacy and motivation. This fact presented a difficulty at the ti=
me
of generalizing results which did not belong to this synthesis’ research
objectives.
At this stage, the studies which had a TBLT holist=
ic
intervention were chosen for meta-analysis as they provided sufficient data=
for
contrast. Only 11 studies were selected for the meta-analysis out of 21 that
made use of a TBLT lesson framework implementation. Reasons for exclusion of
the 9 remaining studies were mainly due to: qualitative research design,
within-subject design or insufficient data for effect size calculations in =
the
reports.
To
analyze the effects of TBLT interventions on oral production, studies that =
were
selected had to compare an experimental group with a TBLT framework treatme=
nt
versus a control group that received another condition. Studies that ultima=
tely
involved hypothesis testing of TBLT experimentation and reflected via postt=
est
reports.
For
quantitative pretest-experiment-posttest design studies that provided proper
information of statistical tests, such as independent t-test or paired samp=
les
t-tests, on means and standard deviations, effect sizes were computed.
One of the between-group studies in this section,
Kebede’s (2021), was considered to be an outlier because the calculated Coh=
en’s
d effect value was 5.8, which represented a difference of more than 0.5 from
the mean of the rest of studies. This meant that it constituted an outlying
effect, and was thus removed.
Quantitative
Meta-analysis
After the systematic coding analysis was effected, it was evident that a meta-analysis would en=
able
the researcher to generalize the magnitude of effects in these TBLT strategy
interventions. As well as compare the effects across studies, and thus, be =
able
to answer more general questions than the ones presented in the studies wit=
hin
this research domain. Consequently, for this quantitative meta-analysis, the
effect sizes of 11 unique sample studies were combined to determine an aver=
age
effect of a TBLT holistic intervention. Table 3 summarizes the meta-analytic
findings displaying population sizes and mean effects.
Main effects of TBLT holistic interventional studi=
es
Table
3 summarizes the main effects for between-group studies that included a TBLT
holistic measure. The main effect calculated, represented with the black
diamond in figure 2, determines a medium-large effect d=3D0,78.
Table 3
Summary of c=
oded
studies and effect sizes across TBLT Holistic Strategy quasi-experimental
designs. Meta-Analysis Results<=
/i> (95
% Confidence Interval)
|
95%CI
|
95%CI
|
||||||
|
Label |
Cohen’s
d |
Lower |
Upper |
p
=
|
Diamond
Ratio |
Lower |
Upper |
|
Overall |
0.781 |
0.561 |
1.00 |
<.001 |
1.46 |
1.04
|
2.54 |
|
Studies |
Cohen’s
d |
Lower <=
/span> |
Upper |
Total=
|
N1(exp.)
|
N2
(ctrl.) |
|
|
Aliakbari,
M. & |
0.553 |
0.0371 |
1.068 |
=
60 =
|
30 |
30 |
|
|
Alsagheer, A. & Hasan, A.
(2014) <=
/span> |
1.120 |
0.4834 |
1.756 |
44 =
|
23 |
21
|
|
|
Aziz, A. (2018=
) |
0.554 |
0.0764 |
0.554 |
70 |
35 |
35 |
|
|
Chen, B. &=
Zhang, Y. (201=
5) |
1.385 |
0.8629 |
1.906 |
70 |
35 |
35 |
|
|
El-Sadat, E. (2018) |
0.989 =
|
0.4926
|
1.485 |
70 |
35 |
35 |
|
|
Marzban, A.
& Hashemi, M.
(2012) |
0.168 =
|
-0.2707
|
0.607 |
80 |
40 |
40 |
|
|
Nita,
A. et al.
1.179 |
0.4287 |
1.930 |
32 |
16 |
16 |
||
|
Rahmah, H. (2017) |
0.466 |
0.0472 |
0.885 |
90 |
45 |
45 |
Table 3
Summary of c=
oded
studies and effect sizes across TBLT Holistic Strategy quasi-experimental
designs. Meta-Analysis Results<=
/i> (95
% Confidence Interval) (continuation)
|
Studies |
Cohen’s
d |
Lower <=
/span> |
Upper |
Total=
|
N1(exp.)
|
N2
(ctrl.) |
|
Santhosh, P. &=
amp; Meenakshi, K. =
(2017) |
0.892 |
0.4590 |
1.326 |
90 |
45 |
45 |
|
Mekala, S. (2017) |
0.633 |
0.1727 |
1.094 |
76 |
38 |
38 |
|
Wahidin, I. (2=
016) |
1.083 |
0.4891 |
1.676 |
50 |
25 |
25 |
Source: Meta-analysis results of studies with a =
TBLT
holistic intervention
Developed
by: The author
Figure 2
Forest
plot showing effect sizes and confidence intervals for 11 experimental
between-groups studies
Source: Meta-analysis calculation of studies wit=
h a
TBLT holistic intervention
Developed
by: The author
Discussion
This synthesis examined several research designs a=
nd
speaking constructs measures. In this section, a summary and discussion of =
the
findings on this research synthesis and meta-analysis are presented. Some
suggestions for further research directions are also given.
Research Question 1: <=
span
lang=3DEN-US style=3D'font-family:"Times New Roman",serif;color:windowtext'=
>What
are the benefits and limitations of the implementation of pedagogical task
frameworks in the development of spoken production constructs in terms of (=
a)complexity,
(b) accuracy, (c) fluency, and (d) spoken interaction in these studies?
From
the analysis, it can be stated that task repetition and task planning offer
more gains for the enhancement of accuracy and complexity in oral productio=
n. Different
task planning and task sequence treatments have a positive impact on fluenc=
y.
One aspect that favors conversational skills and negotiation of meaning is
scaffolding. It has been evidenced that peer-scaffolding and
multimedia-supported TBLT strategies aid learners in gaining interactional
skills as well as fluency in completing oral tasks. Due to the aforemention=
ed
reasons, it is significant to increase the number of studies that have
scaffolding as treatment in order to corroborate its effects on conversatio=
nal
skills within a TBLT framework. Th=
is
suggests that peer-scaffolding constitutes an effective strategy during a t=
ask
cycle since it enhances interaction and negotiation of meaning. Also, the
synthesis evidenced that repetition and rehearsal of the same task over a
period of time creates a favorable development of the three CAF dimensions.=
In
particular, the repetition of tasks within a time space favors the developm=
ent
of accuracy in spoken production as evidenced by Gashan & Almohaisen’s (2014) which used
repetition of an information-gap task in a two-week time frame. In the case=
of
studies that implemented accuracy measures, it should be said that the
investigations that experimented with more complex and demanding tasks evid=
enced
the occurrence of more interaction and attention to language forms. On the
other hand, pre-task modelling and task rehearsal have been consolidated as=
key
elements for accuracy gains in TBLT strategies since they allow learners to
focus on their particular linguistic needs.
Research Question 2: <=
span
lang=3DEN-US style=3D'font-family:"Times New Roman",serif;color:windowtext'=
>Overall,
how effective is a TBLT strategy for the development of the speaking skill =
for
L2 learners in classroom environments?
In
experimental and quasi-experimental studies, there is evidence for gains af=
ter
treatment. Specifically, while the effects seem to vary across the studies
included in the meta-analysis, 5 of these studies had a medium to large eff=
ect.
All in all, it has been evidenced that more classr=
oom research
is needed to assess the effects of TBLT in normal classroom conditions. Due=
to
the fact that some studies were carried out in laboratory conditions and wi=
th a
small sample, the impact of TBLT strategies across classroom longitudinal
studies is needed. However, in most studies stud=
ent
engagement was increased, as well as the development of accuracy and fluenc=
y.
Small gains have been evidenced for pronunciation. For further research,
observations must be taken into account since in most studies it was unclear
whether teachers had received preparation for TBLT implementation. The resu=
lts
from the meta-analysis of TBLT treatment effects also show slight variation
with 4 studies having a medium size effect and the rest of the studies
contributing large effects. A significantly large effect was computed in Ch=
en,
B. & Zhang, Y. (2015) study where d=3D 1.38 which makes it larger than =
the
rest. In this study the means form pretest to posttest had an increase of
21,71. Nonetheless, it can be considered a huge effect for only a 3-month
intervention. Adding to this, it must be stated that the large effects
appraised in the 4 studies had an intervention that lasted more than 2 mont=
hs
and more than 60 hours of instruction. Factors such as the emphasis on
cooperative learning, learner engagement, motivation and group work seem to
favor a larger effect in contrast to traditional teacher center approaches =
to
the treatment of the speaking skill as documented in these studies.
Conclusions
·&nb=
sp;
This study applied a =
synthetic
approach to examine research on task-based L2
·&nb=
sp;
learner oral producti=
on.
The results of the synthesis and the meta-analysis are varied due to the nu=
mber
of independent variables of study reports that focused on CAF measures.
Nonetheless, after the analysis-synthesis process was concluded, it has bec=
ome
clear that TBLT offer significant advantages for oral production enhancemen=
t.
Especially, the implementation of task repetitions and pre-task planning se=
em
to favor CAF constructs when applying a TBLT framework. This implies that t=
asks
must be repeated for the improvement of conversational skills in the learne=
rs.
·&nb=
sp;
It becomes evident th=
at
TBLT for the development of speaking skills is beneficial for intermediate
students mostly. The meta-analysis part of this systematic review reveals a
medium to large effect in between-groups designs. However, most studies
included only a quantitative approach to measuring the impact of a TBLT
treatment versus other conventional methods. Although it can be said that a
TBLT framework is effective in developing speaking skills in the classroom,
more randomized experiments are needed to avoid sampling bias.
·&nb=
sp;
Due to the wide array=
of
research designs employed for the measurement of CAF dimensions, more studi=
es
are needed for the meta-analysis of effects of task interventions on fluenc=
y,
accuracy and linguistic complexity. Notwithstanding the=
se
kind of limitations, this synthesis of studies supports the
implementation of task repetition and pre-task planning for the development=
of
these dimensions in speaking activities within the L2 lesson. According to
these results, a TBLT framework favors the enhancement of oral production in
intermediate learners of different cultures and ages.
References
Ahmadian,
J., (2011). The effects of ‘massed’ task repetitions on complexity, accurac=
y,
and fluency: does it transfer to a new task? The Language Learning Journal 39(3): 269-280. http://dx.do=
i.org/10.1080/09571736.2010.545239
Aliakbari,
M., Jamalvandi, B. (2010)- The Impact of 'Role Play' on Fostering =
EFL Learners'Speaking Ability: A Task-Based Approach. =
Pan-Pacific Association of Applied
Linguistics 14(1), 15-29.
Alsagheer,
A., Hasan, A. (2014). The effect of using task-based learning in teaching
English on the oral performance of the secondary school students. International Interdisciplinary Journa=
l of
Education 3(2). https://doi.org/10.12816/0003003
Aziz,
A. (2018)-The effect of Task-based language teaching (TBLT) strategy on
students’ speaking ability in demonstrative speech. Journal of Linguistics and
Language Teaching 4(2). http://dx.doi.org/10.29300/ling.v4i2.1665
Bygate,
M., Samuda, V. (2005). Integrative
planning through the use of task repetition. In R. Ellis, (Eds.), Planning and task performance in second
language. (pp. 37-74). Amsterdam: John Benjamins Publishing.
Chen,
B., Zhang, Y. (2015). An Experimental Study on Task-Based Interaction in Im=
proving
College Students’ English Speaking Ability. 2nd International Conference on Educat=
ion
Reform and Modern Management (ERMM 2015). www.atlantis-press.com
Dao, P., Iwashita, N.,
Ellis, R., Shintani N. (2014). Exploring
Language Pedagogy through Second Language =
Acquisition Research. Routledge Introductions to Applied
Linguistics.
El-Sadat,
E. (2018). Improving speaking skills through task-based language teaching
strategies. Egyptian Journals
105(4):37-46. https://doi.org/10.21608/mae=
d.2019.140740
Gashan,
A., Almohaisen, F. (2014)-The Effect of Task
Repetition on Fluency and Accuracy of EFL Saudi Female Learner’s Oral
Performance. Advances in Language and Literary Studies 5(4). =
http://dx.do=
i.org/10.7575/aiac.alls.v.5n.3p.36
Hedges, L.V. (1981). Distribution Theory for
Glass Estimator of Effect Size and Related Estimators. Journal of Educational Statistics 6(2): 101-128.
<=
/span>
Heyvaert,
M., Maes, B., Onghena, P.
(2011). Mixed metho=
ds
research synthesis: definition, framework, and potential. Springer=
Science+Business Media B.V. 2011. https://doi.org/=
10.1007/s11135-011-9538-6
Housen,
A., Kuiken, F., Vedder, I.
(2012). Dimensions of L2 performanc=
e and
proficiency. John Benjamins
Publishing Company.
Housen,
A., Kuiken F. (2009). Complexity, Accuracy and
Fluency in Second Language Acquisition. Applied
Linguistics 30(4): 461–473. https://doi.org/10.1093/applin/amp048
Kebede,
M. S., (2021). The Effect of Task=
-Based
Language Teaching on Developing
Students’ Oral English Achievement: With Reference to Damot Preparatory School Grade Eleven Students. Journal of Literature,
Languages and Linguistics. Vol 75. https://doi.org/10.7176/JLL=
L/75-03
Marzban,
A., Hashemi, M. (2012). The impact of opinion-gap tasks on the speaking of
Iranian intermediate EFL learners. =
Procedia-Social
and Behavioral Sciences 70: 943-948. https:=
//doi.org/10.1016/j.sbspro.2013.01.142
Nita A., Rozimela, Y.,=
Ratmanida. The Influence of
Task-based Language Teaching on Speaking Skill of EFL Students with Intrins=
ic Motivation.
International Journal of Literature and Arts 7(6): 179-184. https://=
doi.org/10.11648/j.ijla.20190706.18
Norris, J. M., Ortega, L. (Eds.). (2006).
Synthesizing research on language
learning and teaching. John Benjamins Publishing Company.
Nunan, D. (2004). Task-based
Language Teaching. Cambridge University Press.
Page,
M.J., McKenzie, J.E., Bossuyt, P.M., Boutron, I., Hoffmann, T.C., Mulrow, C.D. The PRISMA =
2020
statement: an updated guideline for reporting systematic reviews. BMJ 2021;=
372:
n71. https://doi.=
org/10.1136/bmj.n71
Pica, T., Young,=
R.,
Doughty, C. (1987). The impact of Interaction on Comprehension. Tesol Quaterly =
21(4):
737-758. http://www.j=
stor.org/stable/3586992
Prabhu, N.S. (1987). Second Language Pedagogy. Oxford University Press.
Rahmah,
H. (2017). The Effectiveness of
Task-Based Language Teaching in
Developing
Speaking Skills at SMKN 2 MALANG. <=
span
style=3D'color:black;mso-color-alt:windowtext;background:white'>PARADIGMA: =
Jurnal Filsafat, Sains, Teknologi, dan
Richards,
J., Rodgers, T. (2001). Approaches =
and
Methods in Language Teaching. Cambridge University Press.
Santhosh, P., Meenakshi, K. (2017). Enhancing Or=
al
Communication through Task Based Language Teaching among Polytechnic Studen=
ts. Indian Journal of Science and Technolo=
gy
10(11): 1-4=
.
https:=
//doi.org/10.17485/ijst/2017/v10i11/106481
Shantha,
S., Mekala, S. (2017). The role of oral communicative tasks (OCT) in developing the spoken
proficiency of engineering students. Advances
in Language and Literary Studies 8(2). https://doi.=
org/10.7575/aiac.alls.v.8n.2p.161
Skehan,
P. (2009). Modelling Second Language Performance: Integrating Complexity,
Accuracy, Fluency, and Lexis. Appli=
ed
Linguistics 30(4): 510–532. https://doi.org/10.1093/applin/amp047
Van
den Braden, K. (2006). Task-based Language Teaching, From Theory to Practic=
e. Cambridge University <=
span
class=3DSpellE>Press.
PARA
CITAR EL ARTÍCULO INDEXADO.
Medina
Fernández, A. (2021). The effects of Task-Based Language Teaching on the
Speaking Skill: A systematic research synthesis and meta-analysis . =
Ciencia Digital, 5(4), 72-93. https://doi.org/=
10.33262/cienciadigital.v5i4.1801
El artículo =
que
se publica es de exclusiva responsabilidad de los autores y no necesariamen=
te
reflejan el pensamiento de la Revi=
sta
Ciencia Digital.
El artículo qu=
eda
en propiedad de la revista y, por tanto, su publicación parcial y/o total en
otro medio tiene que ser autorizado por el director de la Revista Ciencia Digital.
=
[1]=
Pontific=
ia
Universidad Católica del Ecuador - Sede Ambato: Ambato, Tungurahua, EC,Maestría en Pedagogía d=
el
Inglés como Lengua Extranjera, Ambato, Ecuador, email: alberto.medina.f@pucesa.edu.ec,
https://orcid.org/0000-0002-4291-720X
www.cienciadigital.org
=
Vol. 5, N°4, p. 72-93, octubre-diciembre, 20