How to Cite:

Kaharuddin, K. (2021). Assessing the effect of using artificial intelligence on the writing
skill of Indonesian learners of English. Linguistics and Culture Review, 5(1), 288-304.
https://doi.org/10.37028/lingcure.v5n1.1555

Assessing the Effect of Using Artificial

Intelligence on the Writing Skill of Indonesian
Learners of English

Kaharuddin
Faculty of Education and Teacher Training, Universitas Islam Negeri (UIN)
Alauddin, Makassar, Indonesia

Abstract---Electronic device with artificial intelligence application

facilitates the process of learning-teaching in Higher School. This
research integrates the use of artificial intelligence in teaching reading
and giving feedback as independent variables to improve students’
writing skill as dependent variable. Students’ attitude on such
teaching is the intervening variable. The research was conducted in
Higher Schools in Makassar City. Research instrument was
questionnaire distributed to 100 students’ of Higher Schools, but only
83 were processed. The data were analyzed by path analysis. The
research findings: 1) teaching reading with artificial intelligence
application has no direct but indirect effect through attitude on
students’ writing. 2) Lecturer’s feedback virtually has direct and
indirect effect through attitude on students’ writing. 3) Reading and
feedback through attitude contributes 34.8% to students’ writing skill
improvement.

Keywords---artificial intelligence, attitude writing, feedback, reading.

Introduction

The outbreak of COVID-19 pandemic is a global calamity disrupting various

aspects of human life. However, we must wisely deal with this problem. In
educational field, it imposes the use of technology in learning-teaching virtually. a
thing many lecturers don’t think to apply optimally. It is undeniable that before
this COVID-19 outbreak came about, the learning-teaching process was still
dominated by traditional face-to-face learning models in the classroom without
utilizing the facilities provided by technology. On the other hand, technological
advances should be responded by utilizing the various vi+rtual facilities provided
fon learning-teaching process. Therefore, the researchers view this virtual
learning-teaching as government necessitates it through policy may posively or
negatively affect the learning teaching in particularly Higher Schools in Makassar

Linguistics and Culture Review © 2021.

Corresponding author: Kaharuddin, K.; Email: [email protected]
Manuscript submitted: 27 June 2021, Manuscript revised: 09 Sept 2021, Accepted for publication: 18 Oct 2021
288
 289

City. In fact lecturers and students are busy to learn to apply the appropriate
software application provided by technology product in learning-teaching
virtually.

Utilizing technology as a learning medium is the only option so as to call online

learning, or virtual learning or e-learning. “Technology is facilitating the lecturer's
pedagogical approach” Tsui & Tavares (2021), and “creating new ways of
teaching” (Zhao et al., 2006). Therefore, teachers using electronic device with
artificial intellihence make possible to pull out the best results from students. In
other words, the interaction between technology and pedagogy can bring changes
to the way the lecturers teaching and the way the students learning. In the
process, artificial intelligence of technology is integrated to help smoothing the
learning process (Kaharuddin, 2020). Many artificial intelligence applications are
useable in learning teaching process. For example, google translation provides
translation facilities for students to translate from English into Indonesian or vice
verse to understand lecturer’s teaching materials. Artificial intelligence shows its
effect in automatic translation and language processing, and even “in a 2015
Google report, a new computer model called D-Wave 2X is capable of performing
complex Artificial Intelligence operations” (Caplar et al., 2017).

This paper does not discuss the extent technology contributes to improving the
quality of education, but limited to the process of English learning teaching in
Higher Schools as Fahimirad & Kotamjani (2018), state “the effective application
of artificial intelligence methods is considered as a means of improving the quality
of teaching and learning.” One of the English skills highlighted in this research is
writing which require mastery of grammar and abundant vocabulary that
students may get from reading and feedback. The fact shows that the students’
ability to write in English is still low, even though this skill which needs good
mastery on grammar and rich of vocabulary has been taught using traditional
model since elementary school (Arafah, 2019). This paper discusses Higher School
students' perceptions of the use of artificial intelligence applications in English
learning teaching process. The use of this application is particularly intended to
spur the improvement of students’ ability in writing skill (Mohan & Kumar, 2018;
Astawa et al., 2017).

Since writing skill needs students to enrich themselves with new vocabularies and
mastery of grammar, reading practices could also facilitate them to get many new
vocabularies.Their efforts to understand the reading passage is a way to improve
grammar, because getting the message in reading passage perfectly needs
understanding on grammar. Therefore, reading comprehension is used as an
independent variable to analyze its effect on students’ writing skill. Reading
lesson in learning teaching is given to students by using of artificial intelligence
application. In addition, every learning requires feedback from tecturers and
technology has basic features to make lecturers and students communicating
virtually. This application is useful for delivering feedback to students and in
verse. With this feedback from lecturers, students can find out their mistakes
and the accuracy of vocabulary they use. Therefore, feedback is the second
independent variable. It is assumed therefore that reading and feedback using
artificial intelligence have effect on students' writing skills (Absatova et al., 2021;
Asri et al., 2021).
290

To accelerate students’ English mastery, positive attitude is necessary by which

learning motivation is boosted. A research by Kaharuddin et al. (2021), stated
that “attitude effected significantly students’ motivation”. In this research,
students' learning attitude is an intervening variable to explain the effect of
reading comprehension and feedback on writing skill where in the learning
process the use of artificial intelligence applications is integrated. Students'
perception on digital-based learning of English during COVID-19 was analyzed in
relation to improving their writing skill in English as dependent variable. Learning
reading and providing feedback as independent variables with artificial
intelligence in learning teaching process are highlighted with their effects on
students' writing skill both directly or indirectly through students' learning
attitudes as an intervening variable (Widana et al., 2020).

Literature review

Artificial intelligence for learning-teaching

In educational contex during the outbreak of COVID-19 pandemic, especially in

learning teaching of English in Higher Schools, lecturers use the software facility
of artificial intelligence application to carry out the English learning process. This
application helps students improving English language skills and important to
students developing their skills in writing English. Artificial intelligence is "a
simulation of human intelligence on machine programmed to think like a human
and imitate its actions" (Vasiljeva et al., 2021). Hence, artificial intelligence is a
software program specifically made to do a certain job. Students learn grammar,
spelling, word matching and sentences construction by this artificial intelligence
application. It can detect students' mistakes in writing and give them feedback.
Mozgovoy (2011), states that “grammar examination is important in text writing
and language learning”.

Artificial intelligence provides feedback on students’ assignments that they can

make improvements. This can affect student learning activities. They will be
motivated to learn if the mistakes they made in grammar, diction or sentence
construction are corrected and the corrections are returned to students. Mistakes
in writing can also be traced by the application of artificial intelligence. Cotos
(2011), said that students get feedback from artificial intelligence, then reread
and improve their writing and practice to become independent learners (Boden,
1998; Miller, 2019).

Reading on writing improvement

Reading is one of the four language skills whose purpose is to get messages from
reading passage. The message can only be obtained if we understand what is
being read which of course requires mastery of vocabulary. It means that by
reading a lot and understanding the text, students can enrich themselves with a
lot of new vocabulary. Having a lot of vocabulary is important for students
learning languages, because vocabulary is an element of language that contains
meaning and through vocabulary messages can be conveyed both orally and in
writing (Ruslan et al., 2020).
 291

“New vocabulary is mostly learned through reading” (Renandya, 2007). In relation

learning teaching of English in Higher Schools, most English lecturers teach
reading separately from writing (Kroll, 1993). In almost all writing classes
students are not involved in reading as much as expected" (Jolliffe, 2007). Even in
composition classrooms, reading is largely overlooked by both students and
teachers (Hirvela, 2004). However, because vocabulary, in addition to mastery of
grammar, is needed in writing, it can be said that there is a relationship between
reading skill and writing skill. A theory by Horning, & Kraemer (2013), stated that
“reading can or should be used as an effective method to teach/learn writing
effectively in high schools and colleges”. Durukan (2011), stated that among the
four language skills, reading and writing are the first to be learned by students
and that there is a high correlation between reading comprehension and writing
achievement. Even, Esmaeili (2002), affirmed that if reading and writing can be
used together, then there is a positive influence on students' academic success.
Related to the effect of reading on writing skills, this research uses artificial
intelligence of technology to facilitate the communication between lecturers and
students in virtual learning teaching process. This research discussed the effect of
reading on writing directly and indirectly through students’ attitude toward the
learning teaching process using artificial intellegence (Bench-Capon & Dunne,
2007; Graham et al., 2007).

Feedback on writing improvement

A number of questions may arise in students' minds: have the lecturer checked
my assignment?, whether the assignment I have done is correct?, is there
anything I need to improve?, and so on. To answer this question, feedback is the
answer. Feedback is an important part that cannot be separated from the
learning process, but the majority of lecturers do not receive much training on
how to provide feedback, resulting in ineffective feedback which will affect student
learning outcomes (Corwin, 1976). In addition, there are still many
misconceptions about feedback which is understood only as information or
comments given by lecturer on the assignments students have done (Boud &
Molloy, 2013).

In general, feedback is understood as information or comments given by lecturer

regarding the results of student work. Hattie & Timperley (2007), stated that
"feedback is information provided by teacher related to students’ works or
understanding." It is clarified by Henderson et al. (2019), that "feedback is the
process students understanding information or comments about their
assignments, then students using the information or comments to improve their
learning process". Feedback is basically student centered and not the final part of
a learning process. In feedback, lecturer’s comments are only information Boud &
Molloy (2013), that must be followed up by students. Therefore, effective feedback
is to fulfill the principles of being timely, clear, educative, proportional to
assessment rubric, communicate through feedback, and motivating students
(Henderson & Phillips, 2014). Research by Boggs (2019), stated that providing
students with corrective feedback would help them improving their writing skills.
The same study by Chandler (2003), found that error feedback reduces
grammatical and lexical errors in students’ writings over time. Khadawardi
(2021), said that the lecturer implicit written feedback had a significant effect in
292

helping students improve their writing for a short time (Crosthwaite et al., 2017;
Kwakernaak, 1969).

As learning teaching process during the COVID-19 pandemic takes place

virtually, feedback can be done through technological mediation where artificial
intelligence application is provided. In this case, feedback can be given in written
form through the use of comments feature in Microsoft Word or Google Docs.
Another innovation is audio feedback to make students feeling invited to
communicate by lecturer and the feedback given can be felt more personal
(Cavanaugh & Song, 2014). Some examples of audio feedback are Vocaroo and
Mote. Vocaroo is a website-based application that can be directly accessed at
(https://vocaroo.com/) then start the recording process or download the
recording file. Mote is an extension from Google that can be integrated with
Google features such as Classroom, Google Docs, Google Slides, and Google
Sheets. In addition, Screencast-O-Matic, software, can be used to record activities
and sounds on a laptop and can be saved in video format (Lunze & Lehmann,
2010; Price & Mechelli, 2005).

Learning attitude on writing improvement

Before the COVID-19 pandemic, students studied offline with traditional methods
and sometimes lecturers used electronic devices to mediate the learning-teaching
proses. However, during this pandemic, they must study virtually by utilizing
technology as a learning medium. Newton (2003), revealed “the online learning
has 3 main areas: develop access to education and training, developing the
quality of learning, and maintaining competitiveness in university. However, it
does not mean that the three objectives cannot be built with an offline learning
process. The social aspect among students must also be built and that can only
be done by having face-to-face meetings. Regarding online learning by utilizing
artificial intelligence application of technology, questions arise regarding student
learning attitudes. In this case, whether students show a positive or negative
attitude towards the online learning process. A research by Male et al. (2020),
stated that "the students tend to be bored learning from home." This means
students prefer to study in class rather than virtual learning. This research is in
line with the research by Farooq & Javid (2012), which stated that "the students
realized the importance of technology even though the use of technology was not
encouraging." Govindasamy (2001), further emphasized that the students have
negative attitudes towards online learning. In contrary, Guven Ozdemir & Sonmez
(2021), found that The nursing students had mildly positive attitudes toward e-
learning. In this research students’ learning attitude with artificial intelligence
application of technology and its effect on students’ writing improvement is also
studied (Duncan et al., 1997; Küçükoglu, 2013).

Methodology

This research was a descriptive quantitative research by highlighting students'

perceptions of the use of artificial intelligence in online English learning. The
focus of the research was the effect of independent variables (X) on dependent
variable (Y2) through intervening variable (Y1). The independent variables were
reading comprehension (X1) and feedback (X2). The dependent variable was
 293

writing (Y2), and the intervening variable was learning attitude (Y 1). The
respondents were 100 Higher School students in Makassar City to whom the
questionnaires were distributed directly. The research variables were measured
using the five scales of Likert (strongly agree, agree, neutral, less agree, and
disagree).

Test of research instrument

The research instrument was tested for its validity and reliability. A good
research instrument must satisfy the criretia of validity and reliability
(Kaharuddin, 2021).

 Validity Test: In this test of vaility, the value of each item and the total
value of a variable were correleted by using Pearson Correlation Product
Moment ( r ) with significant degree of 5% or 0.05. If the value of rcalculation
of each item of questionnaire is bigger than the value r -table , the instrument
is called valid and useable in this research.
 Reliability Test: In this test of reliability, the consistency of students in
answering the questionnaire is tested. The criteria to determine the
reliability of students’ answer is by using Cronbach Alpha. It is said reliable
if the value of Cronbach Coefficient Alpha is above 0,6. Otherwise, it is not
reliable.

Test of classical assumption

 Normality Test: This test is intended to test the distribution of residual

values in the model by using the test of Kolmogorv-Smirnov. A good
regression model is one that has a normal distribution of residual values.
the basis for making the decision is if the value of asymp. sig. (2-tailed)
greater than 0.05 means the residual value is normally distributed.
 Multiclinearity test: This multicollinearity test is intended to determine the
existence of a linear relationship between the independent variables. A good
regression model is free from the assumption of multicollinearity. The basis
for making decision is that if the value of vif is less than 10 or the tolerance
value is bigger than 0.10, it means there is no multicollinearity.
 Heterocedastisity Test: This heterocedasticity test is intended to determine
whether in the regression model there is an inequality of variance from the
residuals of one observation to another (heteroscedasticity). The basis for
making decision is if the value of Sig>0.05, then there is no
heteroscedasticity.
 Linearity Test: This linearity test is intended to determine whether the two
variables have a significant linear relationship or not. the basis for making
decision is the sign value > 0.05 then there is a significant linear
relationship between the independent variable and the dependent variable.

Inferential statistical analysis

This iinferential statistical analysis was conducted by path analysis with software
of SPSS verse 21.
294

Hypothesis testing

In this hypothesis test, the direct and indirect effects of independent variables on
dependent variable were tested. The hypothesis is:

 H0 : variable X1 and X2, have no direct effect on variable Y2

 H1 : variable X1 and X2, have direct effect on variable Y2.

Research Result

Research instrument is questionnaire distributed directly to 100 students of

Higher School in Makassar City Indonesia. 90 questionnaires were retuned and
among them 7 were flawed. Therefore, there are only 83 questionnaires processed
in this research. The data are analyzed by software SPSS verse 21. The outputs
are studied for the validity and reliability of the instrument. The criteria used to
determine the validity of the instrument is that if the value r-calculation > the value of
r-table, each item of the instrument is considered valid. For this purpose, the
significant level 5% or 0.05 is used from which the value of r-table, 0.213 is got.
This value is smaller than the value of r-calculation or r-calculation > r-table, that all items
of questionnaires show r-calculation above 0.30. In means that the questionnaire has
satisfied the criteria of validity. In reliability test, the output of SPSS shows the
value of crombach alpha 0.778 which is above the value of r-table of 0.213. It
means that the questionnaire has satisfied the criteria of reliability.

Table 1
The output of SPSS for test reliability
Reliability statistics

Cronbach's Alpha
Cronbach's Based on
Alpha Standardized Items N of Items
,778 ,768 16

The classical assumption tests conducted in this research are normality test,
multicollinearity test, linearity test, and heterocedasticity test. For the normality
test, the test is conducted on the residual value. The output of the test using
SPSS version 21 shows the value of Asymp. Sig. (2-tailed) of Kolmogorov Smirnov
0.200 which is bigger that 0.05 (0.200 > 0.05). It means that the residual variable
is normally distributed.

Table 2
The output of SPSS for normality test
One-sample Kolmogorov-Smirnov test

Unstandardized Residual
N 83
Normal Parametersa,b Mean ,0000000
Std.
,35598364
Deviation
Most Extreme Absolute ,080
 295

Differences Positive ,080

Negative -,069
Test Statistic ,080
Asymp. Sig. (2-tailed) ,200c,d
a. Test distribution is Normal.
b. Calculated from data.
c. Lilliefors Significance Correction.
d. This is a lower bound of the true significance.

To identify whether there is linear relationship between independent variables,

multicolinearity test is applied. The independent variables are Reading (X1) and
Feedback (X2). To make conclusion, the tolerance value of each variable is
compared with 0.10 and the value of VIF is compared with 10.00. The output of
SPSS indicated that tolerance value of each variable (X 1: 0.871, X2: 0.595) is
bigger than 0.10 and the value of VIF (X 1: 1,148, X2: 1,681) is smaller than 10.00.
The conclusion is that there is no linear relationship between independent
variables and therefore there is no multicolinearity problem in the regression
equation model.

Table 3
The output of SPSS for multicolinearity test

Coefficientsa
Unstandardized Standardized Collinearity
Coefficients Coefficients Statistics
Std.
Model B Error Beta t Sig. Tolerance VIF
1 (Constant)
,541 2,686 ,009
1,454
RD ,084 ,150 ,055 ,562 ,576 ,871 1,148
FB ,399 ,116 ,405 3,438 ,001 ,595 1,681
ATT ,159 ,086 ,218 1,841 ,069 ,586 1,706
a. Dependent Variable: WR

Heterocedasticity test is intended to know whether there is in regression model

variance insimilarity from residual of one research to another research. A good
model does not show heterocedasticity problem. The basis to make conclusion is:

 Value of Sig> 0.05 = no heterocedasticity

 Value of Sig<0.05 = there is heterocedasticity

The coefficient table column sig. denotes the value of sign. Of X1 0.785 for, X2
0.328 and Y1 0.062. These values of sign. Are bigger than 0.05 which means that
there is no heterocedasticity in the regression model.
296

Table 4
The output of SPSS for heterocedasticity test

Coefficientsa
Unstandardized Standardized Collinearity
Coefficients Coefficients Statistics
Std.
Model B Error Beta t Sig. Tolerance VIF
1 (Constant) ,464 ,342 1,357 ,179
RD -
,095 -,032 -,274 ,785 ,871 1,148
,026
FB ,072 ,073 ,140 ,984 ,328 ,595 1,681
ATT - -
,055 -,271 ,062 ,586 1,706
,103 1,890
a. Dependent Variable: RES2

Linearity test is intended to know whether the independent variable and

dependent variable show significant linear correlation. Criteria to determine the
correlation is if the value of sign. Is bigger than 0.05, there is then significant
linear correlation. Otherwise, there is no linear correlation. The output of SPSS as
donoted by Anova table shows that the value of sign. In deviation from linearity is
0.099 bigger than 0.05. It means that there is significant linear correlation
between independent variables and dependent variable.

Table 5
The output of SPSS for linearity test

ANOVA Table
Sum of Mean
Squares Df Square F Sig.
WR * Between (Combined) 6,577 8 ,822 6,505 ,000
FB Groups Linearity 4,983 1 4,983 39,420 ,000
Deviation
from 1,595 7 ,228 1,802 ,099
Linearity
Within Groups 9,353 74 ,126
Total 15,931 82

As this research uses intervening variable (Y1), the path analysis is conducted
twice and two models are produced: model 1 and model 2. The independent
variables are Reading (X1) and Feedback (X2). The dependent variable is Writing
(Y2) and the intervening variable is Attitude (Y1).

Model 1

This model is constructed by X 1 and X2 as independent variables and Y 1 as

dependent variable. The output of SPSS is shown in table below.
 297

Table 6
The output of SPSS for model 1

Coefficientsa
Standardiz
Unstandardized ed
Coefficients Coefficients
Std.
Model B Error Beta T Sig.
1 (Constant
-,507 ,699 -,725 ,471
)
RD ,321 ,191 ,152 1,684 ,096
FB ,785 ,122 ,579 6,428 ,000
a. Dependent Variable: ATT

The table 6 shows the value of standardized coefficient of the independent

variables as path coefficient of X1 and X2. The path coefficient matrix is:

𝜌𝑦1𝑥1 0.152
( )=( )
𝜌𝑦1𝑥2 0.579

The value of determinant coefficient or the value of R square is denoted in the

model summary below:

Table 7
Output of SPSS for determinant coefficient (R2) of model 1

Model Summary
Adjusted R Std. Error of
Model R R Square Square the Estimate
1 ,643a ,414 ,399 ,46982
a. Predictors: (Constant), FB, RD
b. Dependent variable: Attitude

The R square value is 0.414. To calculate manually the R square value, the path
coefficient matrix X1 and X2 is changed into row matrix and multiplied them with
the Y1 column matrix. From the R square, the path coefficient of other variables
outside the model is calculated, pY1ɛ.

ρY1ɛ = √1 − 0.414 = 0,586

 Path coefficient of ρY1X1

For the path coefficient of pY1X1, namely from the path X1 to Y1, the column
sig. in table 6 is 0.096 bigger than 0.05. However, the value of t -calculation
1.684 is bigger than the t-table 1.663. It is decided therefore that Ho is
rejected and Hi received which means the path coefficient of X 1 to Y1 is
statistically significant.
 Path coefficient of ρY1X2
298

For this path coefficient of pY1X2, the column sig. is 0.000 smaller than 0.05
and the value of t-calculation 6,428 is bigger than the t-table 1.663. It means
therefore that Ho is rejected and Hi is accepted and so the path coefficient
X2 to Y1 is statistically significant.

Model 2

In model 2, the students’ Writing skill (Y2) is dependent variable. Reading (X1),
Feedback (X2), and Attitude (Y1) are independent variables. Below is the output of
SPSS for this model 2.

Table 8
The output of SPSS for model 2

Coefficientsa
Standardiz
Unstandardized ed
Coefficients Coefficients
Std.
Model B Error Beta T Sig.
1 (Constan
1,454 ,541 2,686 ,009
t)
RD ,084 ,150 ,055 ,562 ,576
FB ,399 ,116 ,405 3,438 ,001
ATT ,159 ,086 ,218 1,841 ,049
a. Dependent Variable: WR

The value of standardized coefficients as path coefficient for X1 is 0.055, X2 is

0.406 and Y1 is 0.218. The matrix path coefficient is:

𝜌𝑦2𝑥1 0.055
(𝜌𝑦2𝑥2 ) = (0.406)
𝜌𝑦2𝑦1 0.218

Table 9
Determinant coefficient (R2) of model 2

Model Summary
R Adjusted R Std. Error of
Model R Square Square the Estimate
1 ,590a ,348 ,323 ,36268
a. Predictors: (Constant), ATT, RD, FB4
b. Dependent Variable: WR

The value of R Square is 0.348 calculated manually by altering the path

coefficient matrix X1, X2 and Y1 to path matrix and multiplied by the column
matrix Y2. The path coefficient of other variables outside the model ρY 2Ɛ is
calculated:
 299

ρY2Ɛ = √1 − 0.348 = 0,652

 Path Coefficient ρY2X1

For the path coefficient from the path X1 to Y2, the column sig. in table 8 is
0.576 bigger than 0.05 and the value of t -calculation 0,562 is smaller than the t-
table 1.663. Therefore, Ho is accepted and Hi rejected means the path
coefficient of X1 to Y2 is statistically not significant.
 Path Coefficient ρY2X2
For the path coefficient from the path X2 to Y2, the column sig. in table 8 is
0.001 smaller than 0.05 and the value of t -calculation 3,438 is bigger than the t-
table 1.663. Therefore, Hi is accepted and Ho rejected means the path
coefficient of X2 to Y2 is statistically significant.
 Path Coefficient ρY2Y1
For the path coefficient from the path Y1 to Y2, the column sig. in table 8 is
0.049 bigger than 0.05 and the value of t-calculation 1,841 is bigger than the t-
table 1.663. Therefore, Ho is rejected and Hi accepted means the path
coefficient of X2 to Y2 is statistically significant.

Hypothesis testing

The effect of Reading Comprehension (X1) on Students’ Learning Attitude (Y1) is

tested. The analysis of path coefficient of ρY2X1 indicated that the significant
value of X1 is 0.576 is bigger than 0.05, and the value of t-calculation (0,562) is
smaller than the value of t-table (1.663) or t-calculation<t-table. It proves Reading
Comprehension has not effect on Students’ Learning Attitude. Hence:
Hypothesis 1: “Reading has significant effect on Attitude” is rejected.

With regard to the effect of Reading Comprehension (X1) on students’ Writing

Improvement (Y2), the analysis denoted that the significant value of X 1 is 0.576
bigger than 0.05 and the value of t-calculation (0,562) is smaller than the value of t-
table (1.663) or t-calculation > t-table. The conclusion is Reading Comprehension (X1)
has no direct effect on students’ Writing Improvement (Y 2). Hence, the hypothesis:
Hypothesis 2: “Reading has direct effect on Writing” is rejected.

The indirect effect of Reading Comprehension (X1) through Attitude (Y1) on

students’ Writing Improvement (Y2) is tested. The indirect effect of X 1 through Y1
on Y2 is the multiplication between β values of X1 on Y1 with β value of Y1 on Y2.
As shown in table 6, the β value of X1 on Y1 is 0,152 and β value of Y1 on Y2 is
0,218. Thus, 0,152 x 0,218 = 0,033. The total effect of X1 on Y2 is direct effect
plus indirect effect: 0,055+0,033= 0,088. It means that the indirect effect is bigger
than the direct effect of Reading Comprehension on the students’ Writing
Improvement. Hence, the hypothesis:
Hypothesis 3: “Reading has indirect effect through Attitude on Writing” is accepted.

The analysis of the effect of Feedback (X2) on Students’ Attitude in learning

writing with artificial intelligence (Y1), the significant value of X2 is 0.000 smaller
than 0.05 and the value of t-calculation is 6,428 is bigger than the value of t-table
(1.663) or t-calculation > t-table. It satisfied the criteria of concluding that Feedback
has effect on Attitude. Hence, the hypothesis:
Hypothesis 4: “Feedback has effect on Attitude” is accepted.
300

Regarding the direct effect of Feedback (X 2) on the Students’ Writing Improvement

(Y2), the analysis shows the significant value of X 2 is 0,001 smaller than 0.05 and
the value of t-calculation (3,438) is bigger than the value of t-table (1.663) or t-calculation <
t-table. It also fulfills the criteria of saying that Feedback has direct effect on
students’ Writing Improvement. Hence, the hypothesis:
Hypothesis 5: “Feedback has direct significant effect on Writing” is accepted.

The indirect effect of X2 through Y1 on Y2 is calculated by multiplying the β value

of X2 on Y1 with the β value of Y1 on Y2. The result is 0,579 x 0,218 = 0,126. The
total effect of X2 on Y2 is the direct effect plus the indirect effect, 0.405+0,126=
0,531. The result is that the value of indirect effect (0,531) is bigger than the
value of the direct effect (0.405). It means that X2 has indirect effect through Y1 on
Y2 or Feedback has indirect effect through Attitude on Students’ Writing
Improvement. Therefore, the hypothesis:
Hypothesis 6: “Feedback has indirect effect through Attitude on Writing” is
accepted.

Effect of Attitude (Y1) on Writing (Y2) is tested. The criteria is if the significance
value of Y1 is bigger than 0.05 or T t-calculation > t-table, Y1 is considered to have
effect on Y2. The result of the analysis shows that the significant value of Y 1 is
0.049 smaller than 0.05 and the value of t-calculation (1,841) is bigger than the value
of t-table (1.663) or t-calculation < t-table. It means that Attitude has effect on Writing.
Hence, the hypothesis:
Hypothesis 7: “Attitude has effect on Writing” is accepted.

Discussion

During COVID-19, the process of learning-teaching of English is forced to be

conducted virtually by using electronic devices such as computer, laptop or
mobile phone. Selecting the appropriate Microsoft application is then important to
run well the learning-teaching process. This paper studies the use of artificial
intelligence in teaching reading and giving feedback for students’ work to improve
their writing skill. Students’ attitude to the use of artificial intelligence is also
studied to explain the effect of reading and feedback on writing skill. The benefit
of using this artificial intelligence in learning-teaching is that it can do the
automated tasks assigned to students. However, it is not imaginable that they can
do the more complicated tasks of higher learning (Soto et al., 2012). Therefore,
technological progress will be a common solution Bengio et al. (2013), for the
problems of learning teaching in Higher Schools.

In teaching reading virtually, the reading passage can be uploaded on google

classroom in form of PowerPoint. Pence (2019), said that meetings are held online
as presentations on a slide show presentation program like PowerPoint. Students
can also be given assignment through the google classroom and send it back to
lecturers after working out the assignment. Here is lecturer’s feedback applied for
informing students about mistaken they have done both in diction and grammar.
Students’ understanding on the reading passage sent virtually by google
classroom is a progress to boost writing competence, because their understanding
must supported by good grammar and vocabulary, especially those used in the
reading passage. Of course the students’ progress in learning virtually must
 301

supported by their positive attitude toward the use of artificial intelligence in

learning-teaching process.

This research proves teaching reading online using the artificial intelligence with
students’ positive attitude has significant effect on their writing skill
improvement. It proves by the total effect, 0,088, of Reading on Writing through
attitude which is bigger than the direct effect 0,055. Without the students’
positive attitude, the teaching of reading virtually does not contribute significant
effect on students’ writing skill. Therefore, the improvement of students’ writing
skill must be boosted by their positive attitude on the learning teaching of writing
using artifical intelligence. Furthermore, feedback has direct effect on writing. In
this case, feedback contributes significant effect on students’ writing skill, even
with the absence of students’ positive attitude. This research proves that virtual
teaching of reading and feedback using artificial intelligence with the intervening
of positive attitude contributes 34.8% on students’ writing skill improvemnts. The
rest, 65,2% are contributed by variables outside this model. Therefore, improving
students’ writing skill requires good mastery on grammar and rich of vocabulary
which can be developed by learning teaching of reading and feedback from
lecturer and of course supported by students positive attitude. Electronic device
with application of artificial intelligence could facilitate the process.

Conclusion

The global problem of COVID-19 outbreak is actually enforcing educators to effort

using electronic devices with artificial intelligence application in learning teaching
process. Teaching reading virtually with artificial intelligence application boosted
by students’ positive attitude is proved contributing significant effect on students’
writing skill improvement. Teachers’ feedback of students’ works has also direct
and indirect effect through attitude on students’ writing skill improvement. This
research proves that the virtual teaching of reading and feedback boosted by
positive attitude contributes 34.8% to students’ writing skill improvements.

References

Absatova, M., Aituarova, A., Ospanbekova, M., Azhgaliyev, M., Uanbayev, E., &
Doshybekov, A. (2021). Experimental research of motivational and reflexive
components of students’ reading culture. Linguistics and Culture Review, 5(S1),
670-683.
Arafah, B. (2019). The Idol: A Model for Incorporating Literature in Elt. KnE Social
Sciences, 43-59.
Asri, D. N., Cahyono, B. E. H., & Trisnani, R. P. (2021). Early reading learning for
special needs students: challenges on inclusive primary school during COVID-
19 pandemic. Linguistics and Culture Review, 5(S1), 1062-1074.
Astawa, I. N., Handayani, N. D., Mantra, I. B. N., & Wardana, I. K. (2017). Writing
English language test items as a learning device: a principle of habit formation
rules. International Journal of Social Sciences and Humanities, 1(3), 135-144.
Bench-Capon, T. J., & Dunne, P. E. (2007). Argumentation in artificial
intelligence. Artificial intelligence, 171(10-15), 619-641.
https://doi.org/10.1016/j.artint.2007.05.001
302

Bengio, Y., Courville, A., & Vincent, P. (2013). Representation learning: A review
and new perspectives. IEEE transactions on pattern analysis and machine
intelligence, 35(8), 1798-1828.
Boden, M. A. (1998). Creativity and artificial intelligence. Artificial
intelligence, 103(1-2), 347-356. https://doi.org/10.1016/S0004-
3702(98)00055-1
Boggs, J. A. (2019). Effects of teacher-scaffolded and self-scaffolded corrective
feedback compared to direct corrective feedback on grammatical accuracy in
English L2 writing. Journal of Second Language Writing, 46, 100671.
Boud, D., & Molloy, E. (2013). Rethinking models of feedback for learning: the
challenge of design. Assessment & Evaluation in higher education, 38(6), 698-
712.
Caplar, N., Tacchella, S., & Birrer, S. (2017). Quantitative evaluation of gender
bias in astronomical publications from citation counts. Nature Astronomy, 1(6),
1-5.
Cavanaugh, A. J., & Song, L. (2014). Audio feedback versus written feedback:
Instructors' and students' perspectives. Journal of Online Learning and
Teaching, 10(1), 122.
Chandler, J. (2003). The efficacy of various kinds of error feedback for
improvement in the accuracy and fluency of L2 student writing. Journal of
second language writing, 12(3), 267-296.
Corwin, R. G. (1976). "Dan C. Lortie": SCHOOL TEACHER (Book Review). Social
Forces, 54(4), 948.
Cotos, E. (2011). Potential of automated writing evaluation feedback. Calico
Journal, 28(2), 420-459.
Crosthwaite, P., Cheung, L., & Jiang, F. K. (2017). Writing with attitude: Stance
expression in learner and professional dentistry research reports. English for
specific purposes, 46, 107-123. https://doi.org/10.1016/j.esp.2017.02.001
Duncan, L. G., Seymour, P. H., & Hill, S. (1997). How important are rhyme and
analogy in beginning reading?. Cognition, 63(2), 171-208.
https://doi.org/10.1016/S0010-0277(97)00001-2
Durukan, E. (2011). Effects of cooperative integrated reading and composition
(CIRC) technique on reading-writing skills. Educational Research and
Reviews, 6(1), 102-109.
Esmaeili, H. (2002). Integrated reading and writing tasks and ESL students'
reading and writing performance in an English language test. Canadian
Modern Language Review, 58(4), 599-620.
Fahimirad, M., & Kotamjani, S. S. (2018). A review on application of artificial
intelligence in teaching and learning in educational contexts. International
Journal of Learning and Development, 8(4), 106-118.
Farooq, M. U., & Javid, C. Z. (2012). Attitude of students towards E-learning: A
study of English language learners at Taif University English Language
Centre. NUML Journal of Critical Inquiry, 10(2), 17-28.
Govindasamy, T. (2001). Successful implementation of e-learning: Pedagogical
considerations. The internet and higher education, 4(3-4), 287-299.
Graham, S., Berninger, V., & Fan, W. (2007). The structural relationship between
writing attitude and writing achievement in first and third grade
students. Contemporary educational psychology, 32(3), 516-536.
https://doi.org/10.1016/j.cedpsych.2007.01.002
 303

Guven Ozdemir, N., & Sonmez, M. (2021). The relationship between nursing
students' technology addiction levels and attitudes toward e‐learning during
the COVID‐19 pandemic: A cross‐sectional study. Perspectives in Psychiatric
Care, 57(3), 1442-1448.
Hattie, J., & Timperley, H. (2007). The power of feedback. Review of educational
research, 77(1), 81-112.
Henderson, M., & Phillips, M. (2014, September). Technology enhanced feedback
on assessment. In Australian Computers in Education Conference.
Henderson, M., Phillips, M., Ryan, T., Boud, D., Dawson, P., Molloy, E., &
Mahoney, P. (2019). Conditions that enable effective feedback. Higher
Education Research & Development, 38(7), 1401-1416.
Hirvela, A. (2004). Connecting reading & writing in second language writing
instruction. University of Michigan Press.
Horning, A. S., & Kraemer, E. W. (Eds.). (2013). Reconnecting reading and writing.
Parlor Press LLC.
Jolliffe, W. (2007). Cooperative learning in the classroom: Putting it into practice.
Sage..
Kaharuddin, A. (2020). Principles Behind Semantic Relation between Common
Abbreviations and their Expansions on Instagram. International Journal of
Criminology and Sociology, 9.
Kaharuddin, A. (2021). Analysis Of Syariah Credit Cardfrom Islamic Economic
Perspective: Evidence From Bni Bank In Indonesia. Academy of Strategic
Management Journal, 20(4).
Kaharuddin, K., Ahmad, D., Mardiana, M., & Rusni, R. (2021). Contributions of
Technology, Culture, and Attitude to English Learning Motivation During
Covid-19 Outbreaks. Systematic Reviews in Pharmacy, 12(3).
Khadawardi, H. A. (2021). The Effect of Implicit Corrective Feedback on English
Writing of International Second Language Learners. English Language
Teaching, 14(1), 123-139.
Kroll, J. F. (1993). Accessing conceptual representations for words in a second
language. The bilingual lexicon, 53, 481.
Küçükoğlu, H. (2013). Improving reading skills through effective reading
strategies. Procedia-Social and Behavioral Sciences, 70, 709-714.
https://doi.org/10.1016/j.sbspro.2013.01.113
Kwakernaak, H. (1969). Optimal low-sensitivity linear feedback
systems. Automatica, 5(3), 279-285. https://doi.org/10.1016/0005-
1098(69)90070-3
Lunze, J., & Lehmann, D. (2010). A state-feedback approach to event-based
control. Automatica, 46(1), 211-215.
https://doi.org/10.1016/j.automatica.2009.10.035
Male, H., Murniarti, E., Simatupang, M. S., Siregar, J., Sihotang, H., & Gunawan,
R. (2020). Atittude Of Undergraduate Student’s Towards Online Learning
During Covid-19 Pandemic. PalArch's Journal of Archaeology of
Egypt/Egyptology, 17(4), 1628-1637.
Miller, T. (2019). Explanation in artificial intelligence: Insights from the social
sciences. Artificial intelligence, 267, 1-38.
https://doi.org/10.1016/j.artint.2018.07.007
Mohan, N., & Kumar, S. S. (2018). From the individual to the historical: a
commentary on amitav ghosh as a writer of historical fiction with reference to
304

the glass palace. International Journal of Social Sciences and Humanities, 2(3),
79-85.
Mozgovoy, M. (2011, September). Dependency-based rules for grammar checking
with LanguageTool. In 2011 Federated Conference on Computer Science and
Information Systems (FedCSIS) (pp. 209-212). IEEE.
Newton, R. (2003). Staff attitudes to the development and delivery of
e‐learning. New library world.
Pence, H. E. (2019). Artificial Intelligence in Higher Education: New Wine in Old
Wineskins?. Journal of Educational Technology Systems, 48(1), 5-13.
Price, C. J., & Mechelli, A. (2005). Reading and reading disturbance. Current
opinion in neurobiology, 15(2), 231-238.
https://doi.org/10.1016/j.conb.2005.03.003
Renandya, W. A. (2007). The power of extensive reading. RELC Journal, 38(2),
133-149.
Ruslan, I., Sunubi, A. H., Halidin, A., & Kaharuddin, A. (2020). Paraphrasing
Technique To Develop Skill for English Writing Among Indonesian College
Students of English. Sys Rev Pharm, 11(11), 291-297.
Soto, J. A., Armenta, B. E., Perez, C. R., Zamboanga, B. L., Umaña-Taylor, A. J.,
Lee, R. M., ... & Ham, L. S. (2012). Strength in numbers? Cognitive reappraisal
tendencies and psychological functioning among Latinos in the context of
oppression. Cultural Diversity and Ethnic Minority Psychology, 18(4), 384.
Tsui, A. B., & Tavares, N. J. (2021). The Technology Cart and the Pedagogy Horse
in Online Teaching. English Teaching & Learning, 45(1), 109-118.
Vasiljeva, T., Kreituss, I., & Lulle, I. (2021). Artificial Intelligence: The Attitude of
the Public and Representatives of Various Industries. Journal of Risk and
Financial Management, 14(8), 339.
Widana, I.K., Dewi, G.A.O.C., Suryasa, W. (2020). Ergonomics approach to
improve student concentration on learning process of professional ethics.
Journal of Advanced Research in Dynamical and Control Systems, 12(7), 429-
445.
Zhao, Y., Frank, K. A., & Ellefson, N. C. (2006). Fostering meaningful teaching
and learning with technology: Characteristics of effective professional
development. Meaningful learning using technology: What educators need to
know and do, 161, 179.