Brit J Educational Tech - 2023 - Xia - The Mediating Effects of Needs Satisfaction On The Relationships Between Prior

14678535, 0, Downloaded from https://bera-journals.onlinelibrary.wiley.com/doi/10.1111/bjet.13305 by Orta Dogu Teknik Universitesi, Wiley Online Library on [05/02/2023].
See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Received: 5 July 2022 Accepted: 17 January 2023
DOI: 10.1111/bjet.13305
ORIGINAL ARTICLE
The mediating effects of needs satisfaction

on the relationships between prior knowledge
and self-regulated learning through artificial
intelligence chatbot
Qi Xia1 | Thomas K. F. Chiu1,2 | Ching Sing Chai1 |

Kui Xie3,4
1
Department of Curriculum and Instruction,
Faculty of Education, The Chinese University Abstract
of Hong Kong, Hong Kong, Hong Kong SAR The anthropomorphic characteristics of artificial intel-
2
Centre for Learning Sciences and ligence (AI) can provide a positive environment for
Technologies, The Chinese University of
Hong Kong, Hong Kong, Hong Kong SAR self-regulated learning (SRL). The factors affecting
3
Department of Educational Studies, Faculty adolescents' SRL through AI technologies remain
of Education, The Ohio State University, unclear. Limited AI and disciplinary knowledge may
Columbus, Ohio, USA affect the students' motivations, as explained by
4
Research Laboratory for Digital Learning, The self-determination theory (SDT). In this study, we
Ohio State University, Columbus, Ohio, USA
examine the mediating effects of needs satisfaction in
Correspondence
SDT on the relationship between students' previous
Thomas K. F. Chiu, Department of Curriculum technical (AI) and disciplinary (English) knowledge
and Instruction, Faculty of Education, The and SRL, using an AI conversational chatbot. Data
Chinese University of Hong Kong, Hong were collected from 323 9th Grade students through a
Kong, Hong Kong SAR.
Email: tchiu@cuhk.edu.hk questionnaire and a test. The students completed an
AI basic unit and then learned English with a conver-
Funding information sational chatbot for 5 days. Confidence intervals were
Quality Education Fund calculated to investigate the mediating effects. We
found that students' previous knowledge of English
but not their AI knowledge directly affected their SRL
with the chatbot, and that satisfying the need for
autonomy and competence mediated the relation-
ships between both knowledge (AI and English) and
SRL, but relatedness did not. The self-directed nature
of SRL requires heavy cognitive learning and satis-
fying the need for autonomy and competence may
more effectively engage young children in this type of
This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which
permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifi-
cations or adaptations are made.
© 2023 The Authors. British Journal of Educational Technology published by John Wiley & Sons Ltd on behalf of British Educational
Research Association.
Br J Educ Technol. 2023;00:1–20. wileyonlinelibrary.com/journal/bjet 1

14678535, 0, Downloaded from https://bera-journals.onlinelibrary.wiley.com/doi/10.1111/bjet.13305 by Orta Dogu Teknik Universitesi, Wiley Online Library on [05/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2 XIA et al.
learning. The findings also revealed that current chat-

bot technologies may not benefit students with rela-
tively lower levels of English proficiency. We suggest
that teachers can use conversational chatbots for
knowledge consolidation purposes, but not in SRL
explorations.
KEYWORDS
artificial intelligence, K-12 education, prior knowledge,
self-determination theory, self-regulated learning
Practitioner notes
What is already known about this topic
• Artificial intelligence (AI) technologies can potentially support students' self-regulated
learning (SRL) of disciplinary knowledge through chatbots.
• Needs satisfaction in Self-determination theory (SDT) can explain the directive
process required for SRL.
• Technical and disciplinary knowledge would affect SRL with technologies.
What this paper adds
• This study examines the mediating effects of needs satisfaction in SDT on the
relationship between students' previous AI (technical) and English (disciplinary)
knowledge and SRL, using an AI conversational chatbot.
• Students' previous knowledge of English but not their AI knowledge directly affected
their SRL with the chatbot.
• Autonomy and competence were mediators, but relatedness was not.
Implications for practice and/or policy
• Teachers should use chatbots for knowledge consolidation rather than exploration.
• Teachers should support students' competence and autonomy, as these were
found to be the factors that directly predicted SRL.
• School leaders and teacher educators should include the mediating effects of
needs satisfaction in professional development programmes for digital education.
INTRODUCTION
Artificial intelligence (AI) technologies can potentially support students' self-learning of disci-
plinary knowledge through chatbots that engage in conversation, for example, by correcting
pronunciation, spelling and grammatical errors. Some technologies have strong anthropo-
morphic characteristics and can interact with students (Blut et al., 2021; Pelau et al., 2021),
thus providing interactive environments for self-learning. Students must take the initiative
and develop strategies to learn through these technologies and engage in a proactive learn-
ing process, referred to as self-regulated learning (SRL). SRL is ‘not a mental ability or
an academic performance skill, but it is the self-directive process by which learners trans-
form their mental abilities into academic skills’ (Zimmerman, 2002). This directive process
NEEDS SATISFACTION FOR SELF-REGULATED LEARNING WITH AI 3
is strongly associated with students' motivations to learn through digital technologies.

Self-determination theory (SDT) suggests that motivation stems from three universal and
psychological needs that all individuals are driven by: autonomy, competence and relat-
edness (Ryan & Deci, 2000). If these needs are satisfied, they can motivate self-initiated
behaviour and provide energy for personal growth, joyful effort and the acceptance of chal-
lenges. This energy can foster the directive process. The satisfaction of these three needs
can explain how students regulate learning through AI technologies.
Digital competence and prior disciplinary knowledge are associated with students'
SRL capacity and performance in digital environments (Cleary & Kitsantas, 2017; Dong
et al., 2020; Greene et al., 2010; Hirt et al., 2021; Sun et al., 2018; Taub et al., 2014; Taub
& Azevedo, 2019). This implies that prior digital knowledge can affect their capacity to learn
through AI technologies; this is supported by modelling studies that have examined learner
technology acceptance, including the technology acceptance model (TAM) (Chiu, 2017;
Davis, 1989) and the knowledge, attitudes and practice model (KAP) (Kaliyaperumal, 2004).
These models suggest that computer self-efficacy is associated with attitudes towards
learning and the behavioural intention to learn with technology. Thus, student AI knowl-
edge can be viewed as a type of computer self-efficacy that affects their attitudes and
behavioural intentions. Students with a good understanding of AI will understand the affor-
dances and limitations of an AI educational application. Thus, they will feel more capable
of using the application (competence), have more options to interact with it (autonomy) and
be less disappointed when they fail to get satisfactory responses (relatedness) from the
application.
However, AI in education is generally domain-specific (eg, pronunciation error correc-
tion or chatbots for English learning). Student prior disciplinary knowledge can predict
their SRL (Taub et al., 2014) that is supported and explained by four phases in Winne and
Hadwin (2008)'s Model of SRL. Students with different levels of prior knowledge process
information in each of the phases differently. Students with strong prior knowledge are more
likely to use more effective approach to complete their tasks for SRL. Moreover, students
with good disciplinary knowledge may better comprehend the less specific or agreeable
feedback given from an AI education application (competence) and thus feel less frustrated
by the feedback (relatedness). They may also have more disciplinary ideas/knowledge to
communicate with the application (autonomy). In summary, AI and disciplinary knowledge
are likely to be related to the satisfaction of perceived needs, which would promote SRL
through AI conversational chatbots.
Most education studies of AI have focused on system or application development (Song
& Wang, 2020; Zawacki-Richter et al., 2019). SDT is also used as the basis for development,
for example of AI-assisted decision-making design (De Vreede et al., 2021), conversational
agent design (Yang & Aurisicchio, 2021) and AI algorithm development (Ferreira Chame
et al., 2019). Most studies of learner technology acceptance have focused on techni-
cal knowledge (ie, computer self-efficacy or digital literacy; see Alfadda & Mahdi, 2021;
Cheng, 2019; Chow et al., 2022; Dorfsman & Horenczyk, 2022; Khlaisang et al., 2021),
rather than subject domain knowledge. To our knowledge, no SDT-based study has inves-
tigated how students' prior knowledge affects SRL with AI applications. This study fills this
research gap from the two perspectives of needs satisfaction and disciplinary knowledge.
We investigate the mediating effects of needs satisfaction in SDT on the relationships
between student prior knowledge—technical and disciplinary—and SRL using a conversa-
tional chatbot. Conversational chatbots have the ability to converse with learners in natural
language, process and comprehend their language, and respond based on the rules and
data. They are often developed using machine learning, and Natural Language Processing
technologies.
4 XIA et al.
LITERATURE REVIEW
In this section, we propose using SDT as the theoretical framework to explain how the satis-
faction of the three needs can mediate the relationships between SRL with AI applications.
We present our hypotheses and research model.
SDT and SRL
SDT is a motivation theory that aims to explain how basic needs satisfaction affects human
motivation and well-being within a social context (Ryan & Deci, 2017, 2020). It suggests
that everyone has the three fundamental psychological needs of autonomy, competence
and relatedness. These needs satisfaction will determine an individual's motivation to act.
Autonomy refers to individuals' abilities to control their own choices and strive to progress in
whatever direction they find most fulfilling; competence refers to the belief that one knows
what one is doing, can complete a task and possesses the necessary knowledge and skills
to succeed; and relatedness refers to the sense of belonging or connectedness to significant
others, groups, or objects. The more the needs are satisfied, the greater the self-determined
autonomous motivation (Chiu, 2021, 2022). SDT can therefore provide a theoretical frame-
work for examining students' SRL abilities (León et al., 2015; Luo et al., 2021; Mouratidis
et al., 2013).
This theory connects these needs to the motivational processes that encourage and
foster student SRL (Grolnick & Raftery-Helmer, 2015). When students' needs are satisfied,
they become highly autonomously motivated and thus effectively self-regulate their learning
(Liu et al., 2014; Sierens et al., 2009). Students who feel they are autonomous demon-
strate more enthusiastic, focused and purposeful learning, and exemplify self-regulation
(Deci & Ryan, 2004). The satisfaction of the need for autonomy has positive consequences
for students' abilities to regulate independent learning (Holzer, Lüftenegger, et al., 2021;
León et al., 2015; Ng et al., 2012; Radel et al., 2014). Studies of English language learn-
ing in various digital environments have demonstrated that supporting student autonomy
can encourage SRL, for example, through reading-annotation systems (Chen et al., 2020;
Chen & Chen, 2014; Chen & Huang, 2014), vocabulary learning apps (Chen et al., 2019),
mobile-assisted grammar learning systems (Wang et al., 2021) and online collaborative
learning (Lin et al., 2017). Thus, we propose that the satisfaction of the need for autonomy
can determine SRL.
Perceived competence can be regarded as a motivational belief in terms of SRL (Grolnick
& Raftery-Helmer, 2015). Competence can predict students' motivations to engage in SRL
(Holzer, Lüftenegger, et al., 2021; Mouratidis et al., 2013; Ryan & Deci, 2000), and they are
more likely to learn independently in an environment that supports their need for competence
(Grolnick & Raftery-Helmer, 2015). For example, Pelikan et al. (2021) found that students
who perceived themselves as highly competent used SRL strategies more frequently in
distance learning than those who felt they were less competent, due to the significant differ-
ences in their intrinsic motivation. Luo et al. (2021) found that perceived competence was
positively related to perceived usefulness and enjoyment of online SRL. Wan et al. (2012)
found that learners with higher levels of digital competence were able to leverage a virtual
learning environment and used SRL strategies to facilitate their online learning. We therefore
propose that satisfying the need for competence is related to SRL.
Students must internalise the regulation of learning in SRL. Relatedness, together
with autonomy and competence, is an important part of this process (Deci & Ryan, 2004).
Students who feel that they have a positive connection with their teachers are more likely to
develop identifiable and integrated regulations for learning tasks than those who feel discon-
nected from their teachers (Niemiec & Ryan, 2009). Those with a greater sense of relat-
edness and who feel more confident have been found to demonstrate more positive affect
and to perform better in schools than others (Furrer & Skinner, 2003; Holzer, Lüftenegger,
et al., 2021). By internalising extrinsic motivation, they can reorientate their own motiva-
tion and become more autonomous. For example, studies have shown that in a 3D virtual
learning environment, students' perceptions of relatedness are associated with their intrinsic
motivation (Huang et al., 2013; Ryan et al., 2006). Games can engage students in learning
by satisfying their need for relatedness (Deniz et al., 2014). Student engagement in SRL is
therefore likely in a learning environment that supports relatedness (Furrer & Skinner, 2003;
Grolnick & Raftery-Helmer, 2015). We propose that the satisfaction of the need for related-
ness is associated with SRL.
How AI and disciplinary knowledge affect needs satisfaction and SRL
AI knowledge is typically technical, and as the TAM (Davis, 1989) and the KAP model
(Kaliyaperumal, 2004) suggest, can be associated with perceived competence and related-
ness. In the TAM, perceived computer efficacy is strongly associated with perceived ease of
use and attitudes towards using technology. Hong (2022) found that individuals with stronger
AI self-efficacy were more likely to adopt AI technologies as they found them useful and
easy to use. Al Shamsi et al. (2022) found that perceived usefulness was strongly positively
correlated with the behavioural intention to use AI voice assistants. Malik et al. (2021) found
student perceptions of chatbots (eg, perceived convenience, ease of use) positively affected
their intentions to use the technology in their learning. Gado et al. (2021) focused on the
KAP model and suggested that students were able to learn about AI when they felt in control
(autonomy) and capable of using AI (competence). Those with knowledge of how AI works
are more likely to understand the affordances and limitations of technology in supporting
their learning. Thus, they will feel confident to effectively interact with AI in their learning.
They may also be tolerant of receiving negative feedback and responses through AI. More-
over, AI applications that analyse students’ activities in real time give students feedback and
hints that benefit their SRL (van der Graaf et al., 2022). Students need to receive person-
alised feedback on whether they used effective learning practices, which benefit the SRL
(Markauskaite et al., 2022). AI knowledge is seen as a computer efficacy that affects student
motivation and intention to use AI for SRL. Accordingly, AI knowledge can therefore be asso-
ciated with needs satisfaction and SRL.
AI technologies are typically applied to specific learning domains. For example, an adap-
tive learning system for mathematics requires a knowledge structure for generating recom-
mendations and chatbots for English learning require data about language conversations
(Adamopoulou & Moussiades, 2020). Therefore, disciplinary knowledge can influence SRL
that involves AI and also the satisfaction of needs. Prior knowledge of a discipline can contrib-
ute to student motivation and their approaches to learning (Winters et al., 2008). The exten-
sive knowledge of a subject can improve self-reflection and monitoring (Greene et al., 2008,
2010). The extent of students' prior knowledge can determine the SRL strategies they use in
digital learning environments (Yang et al., 2018). This is supported by explained by Winne
and Hadwin's Model of SRL (2008). The model acknowledges that SRL occurs across 4
phases—(i) definition of task, (ii) goals & plans, (iii) studying tactics and (iv) adaptations.
It differs from other SRL models (eg, Pintrich, 2004; Zimmerman, 2002) by examining how
information processes occur within each phase. It is expected that students with stronger
prior disciplinary (English in this study) knowledge will progress through each stage differently
from those with weaker prior knowledge. In the first phase, there is no difference between
6 XIA et al.
students with strong and weak prior knowledge in defining what the task they were asked to
do (Taub et al., 2014). However, students with strong prior knowledge who is more likely to
recall how they approached a similar problem in the past will be more aware of the environ-
mental factors such as available resources. For Phase 2, students with weak prior knowl-
edge who are not familiar with the discipline are more likely to have difficulties to plan their
learning by setting sub-goals for the tasks while those with stronger prior knowledge have no
issues (Greene et al., 2008, 2010). In Phase 3, all students could deploy the planned strate-
gies; however strong prior knowledge students could choose more effective approaches (eg,
use more sophisticated and concise words) to deliver the strategies than weaker one (eg,
use more simple and redundant words) (Greene et al., 2010; Taub et al., 2014). In the final
phase, students with strong prior knowledge are more likely to reflect on their learning and
consolidate what they have learnt, whereas students with weak prior knowledge may fail to
make a meaningful reflection.
In the context of this study, a strong knowledge of English will lead to a better understand-
ing of the responses of a language learning chatbot. The more language proficiency of these
students will also lead to a better communication with the chatbot. However, students with
lower levels of English language proficiency may find it difficult to approach a chatbot. A chat-
bot may also fail to respond appropriately to students (Adamopoulou & Moussiades, 2020).
More proficient students may attempt to communicate with the chatbot in another way
(Jeon, 2022) while less proficient students may give up. Students with good English profi-
ciency will thus be more likely to feel satisfied with their levels of autonomy, competence and
relatedness when learning with conversational chatbots. We therefore propose that discipli-
nary knowledge is associated with need satisfaction and SRL.
THIS STUDY
The aim of this study was to investigate the mediating role of students' needs satisfaction
on the relationships between their English (disciplinary) and AI (technical) knowledge and
self-regulated learning through AI. Our proposed research model is given in Figure 1, and
we offer the following hypotheses.
.40*** Perceived
AI Autonomy
.07 .17*
.32***
.31***
Perceived .10 Self-regulated
relatedness learning
.30***
.19**
.20*** .25***
English Perceived
.29*** competence
F I G U R E 1 Research model. *p < 0.05, **p < 0.01, ***p < 0.001.

H1 English (disciplinary) knowledge will contribute positively to the satisfaction of each of
the three needs.
H2 AI (technical) knowledge will contribute positively to the satisfaction of each of the three
needs.
H3 The satisfaction of each of the three needs will contribute positively to self-regulated
learning through AI.
H4 Needs satisfaction will mediate the relationship between the two types of knowledge
and self-regulated learning with AI.
METHOD
Participants and research procedure

The participants were 323 9th Grade students from three middle schools in Hong Kong
SAR. They were 12–14 years old (51% male, 49% female) with average academic perfor-
mance standards and socio-economic status. The number of students came from each
of the schools ranged from 105 to 110. They were all learning English as a second
language. They had completed an AI basic unit before this study. The unit covered the six
subtopics: (i) What is AI?, (ii) Does AI always work as expected?, (iii) Input-Process-Out-
put framework and AI systems, (iv) Big data, (v) Basic learning models in AI, and (vi)
model training and testing. According to the apriori sample size calculator for structural
equation models, the recommended minimum sample size is 256 when the power level
equals 0.8 and the numbers of latent and observed variables are 6 and 16 respectively
(Soper, 2020). The sample size in this study was therefore appropriate for testing the
proposed model.
We first obtained ethical approval from the corresponding author's institution and received
consent from all of the participants. The students completed a 20-minute test consisting of
two parts, English language and AI knowledge, before engaging in learning with a conversa-
tional chatbot in their schools. Over 5 days, they learned English with the chatbot, and then
completed a 20-minute questionnaire.
To choose the chatbot for this study, three English teachers from the schools explored
and picked their own preferred commercial chatbot mobile Apps to support their student
learning. The corresponding author compiled their preference into a list of five Apps.
The teachers tried and rated the five Apps. The highest rated chatbot App was used in
this study. The App include a chatbot with different sub-themes—hello, restaurant, hotel,
tickets, conversations, shopping, meeting and rental cars—that were used for student
learning in study. To make conversations with the chatbot, the students could choose the
suggested responses, text (typing, spelling) or voice (speaking) input their own ideas to
answer the questions promoted by the chatbot. The chatbot provide the students with
different options to set how they learn—listen to the audio of the questions and suggested
responses, and read Chinese translation of the questions and responses. The students
can set the learning environments they want. To begin the chatbot, the mother language
was set to Chinese, and the learning language is English as all the student participants
were Chinese. The students had choices to learn with the chatbot in any ways such as
typing any input, learn more than one time. They did not have to finish all the subthemes.
Figure 2 shows the screenshots of the chatbot. The AI conversational chatbot decides
if student responses are appropriate for the promoted questions. It guides the students
to answer by giving a gentle question and more suggested answers. The chatbot uses
XIA et al.
Chinese translation of the question

The students can choose to read
The students can listen the audio of
the suggested answers, typing and
questions and suggested answers.

answer the questions by choosing
and suggested answers.

The students have choices to
speaking
F I G U R E 2 Screenshots of the chatbot.

8
AI speech to offer students with learning opportunities for speaking and listening, and AI
read to offer students with Chinese translations for more understanding.
Questionnaire and test
In addition to demographic data, the questionnaire assessed four measures: the satisfaction
of each of the three perceived needs and SRL with the chatbot. Each measure consisted
of four 5-point Likert-scale items. The test consisted of an English language part and an AI
knowledge part.
To assess the satisfaction of students' needs for autonomy and competence, as facili-
tated by the chatbot, we adapted items from Hew and Kadir (2016), which were validated by
Chiu (2021, 2022) in studies that had similar demographic samples as this study. The four
items for perceived autonomy satisfaction, with an original reliability of α = 0.82, were ‘I feel
like I can make a lot of input in deciding how I use the chatbot in learning’; ‘I feel a sense of
freedom when using the chatbot’; ‘I have many opportunities with the chatbot to decide for
myself how to learn’; and ‘I have a say regarding what input I want to learn with chatbot’. The
items for perceived competence, with α = 0.71, were ‘I think I am pretty good at learning with
the chatbot’; ‘I have been able to learn interesting new knowledge with the chatbot’; ‘I feel a
sense of accomplishment from learning with the chatbot’; and ‘I am pretty skilful at learning
with the chatbot’. To measure the satisfaction of students' need for perceived relatedness,
we adapted four items from Furrer and Skinner (2003), with α = 0.86. The items were ‘When
I learn with the chatbot, I feel supported (changed to comfortable/important/valued)’.
The items measuring self-regulated learning were adapted from Sha and colleagues (2012),
with α > 0.70. The items were ‘When learning English with the chatbot, I will normally set
learning goals for myself so that I can decide how and what I want to learn’; ‘When learning
English with the chatbot, I will normally try to identify the knowledge that I do not understand
well’; ‘When learning English with the chatbot, I will normally ask myself questions to help me
focus on what to study’; and ‘When I am not sure about any English language, I will go back
and try to figure it out on my own using the chatbot.’
The two parts of the test had ten questions each. The first part tested students' English
language proficiency, such as ‘Please write down a question for the following scenario:
saying hello to a waiter, booking a table in a restaurant, buying a ticket in a cinema, asking
ways to go to a museum, and looking for a bag of candy in a supermarket’. The second part
tested students' basic AI knowledge, for example, ‘Artificial Intelligence is about’, ‘Which
one is an AI application?’, ‘Which of the following applications misuses AI technologies?’,
‘Bias in AI can be caused by’, ‘Which of the following is an example of “Data Bias” in collect-
ing assessment scores in a class?’, ‘Which of the following is the correct description of
“data authenticity”?’, ‘Which of the following will enhance model performance?’, ‘Which of
the followings will improve AI model?’ The three schools were participating schools of the
corresponding authors' projects on two AI education projects—AI curriculum planning for
schools (Chiu & Chai, 2020), and the development of AI and education in schools. The tests
of basic AI knowledge were adopted and modified from an AI curriculum development project
that involved 63 the middle schools and developed a local curriculum content and assess-
ment (Chiu et al., 2022). Moreover, these questions were validated by three content experts
in AI (two engineering chair professors and one post-doctoral fellow) and two education
professors, two English teachers and two technology teachers. All the tests are marked by
the teachers, and their maximum scores was 10. Accordingly, they are appropriate for this
modelling study.
10 XIA et al.
Data analysis
Mplus 8.3 and SPSS 26.0 were used to analyse the data. First, Cronbach's alpha (α) was
calculated for every subscale in the instrument to determine its internal consistency. Second,
the construct validity was examined by a confirmatory factor analysis. Last, path models
were constructed and tested to examine the relationship between the variables. The effect of
mediating was tested using a bootstrapping approach.
RESULTS
Descriptive statistics and reliabilities
Table 1 gives the descriptive statistics and reliabilities for each factor. The skewness ranged
from −0.21 to 0.17 and the kurtosis from −0.21 to −0.08, indicating that the data were reason-
ably normally distributed and could be used in the analyses. The Cronbach's α coefficients
of SRL, autonomy, competence and relatedness ranged from 0.98 to 0.99, indicating an
acceptable level of internal consistency.
Measurement model
We conducted a confirmatory factor analysis to examine the construct validity. The research
model had a hierarchical structure, and we found that the second-order model did not
reach an acceptable level of data fit (χ 2 = 369.794, df = 98, CFI = 0.972, TLI = 0.966,
RMSEA = 0.093), so we deleted one item from SRL according to the model modifica-
T A B L E 1 Descriptive statistics and reliabilities (n = 323).
M SD Skewness Kurtosis α
1 SRL 3.37 0.87 −0.21 −0.17 0.99
2 Autonomy 3.48 0.92 −0.16 −0.55 0.99
3 Competence 3.50 0.87 −0.16 −0.40 0.98
4 Relatedness 3.25 0.71 −0.08 −0.38 0.98
5 English knowledge 6.95 2.04 −0.15 −0.71 –
6 AI knowledge 6.64 1.93 0.17 −0.73 –
T A B L E 2 Factor loadings and correlation matrix for all variables.
Item CR Factor loadings 1 2 3 4 5 6

1 SRL 0.979 0.969 (0.949–0.996) 0.970
2 Autonomy 0.985 0.970 (0.930–0.990) 0.473*** 0.971
3 Competence 0.981 0.964 (0.940–0.985) 0.494*** 0.463*** 0.964
4 Relatedness 0.975 0.953 (0.928–0.971) 0.375*** 0.410*** 0.442*** 0.953
5 English knowledge – – 0.485*** 0.528*** 0.469*** 0.359*** –
6 AI knowledge – – 0.443*** 0.566*** 0.475*** 0.419*** 0.565*** –
Note: The bold values on the diagonal are the square root of AVE; *p < 0.05, **p < 0.01 and ***p < 0.001 (2-tailed).
tion indices. The modified model was examined and found to fit the data moderately well
(χ 2 = 218.189, df = 84, CFI = 0.984, TLI = 0.981, RMSEA = 0.070) (Hu & Bentler, 1999).
As shown in Table 2, all of the indicators had standardised factor loadings above 0.60
and were significantly loaded (p < 0.001) on each latent variable. The composite relia-
bility (CR) of the variables were all higher than 0.60, indicating that all of the subscales
reached a high level of internal consistency. The average variance extracted (AVE) was.
50, indicating that all of the subscales achieved a high level of convergent validity (Fornell
& Larcker, 1981).
AI and disciplinary knowledge were treated as manifest variables. As depicted in Table 2,
AI knowledge was positively correlated with SRL, r (321) = 0.443, p < 0.001, and the satis-
faction of the needs for autonomy, r (321) = 0.566, p < 0.001, competence, r (321) = 0.475,
p < 0.001, and relatedness, r (321) = 0.419, p < 0.001. English knowledge was also correlated
with the satisfaction of the needs for autonomy, r (321) = 0.528, p < 0.001, competence, r
(321) = 0.469, p < 0.001, and relatedness, r (321) = 0.359, p < 0.001. The three needs were
positively associated with SRL.
Mediating effect analysis
To further examine the relationship between the variables, a path model was constructed
and tested. We considered the nested nature of the data and used the maximum like-
lihood robust (MLR) estimator. The SEM suggested that the model shown in Figure 1
fitted the data well (χ 2 = 298.591, df = 109, CFI = 0.979, TLI = 0.973, RMSEA = 0.073).
The standardised direct, indirect and total effects among the variables are presented in
Table 3.
The relationship between English knowledge and needs satisfaction (H1)
Student English knowledge positively and significantly predicted the satisfaction of each of
the three needs of autonomy (β = 0.30, p < 0.001), competence (β = 0.29, p < 0.001) and
relatedness (β = 0.19, p < 0.01).
T A B L E 3 Standardised total, total, indirect and direct effects among variables.
Predictor Mediating/criterion variable Direct effect Indirect effect Total effect

AI Perceived autonomy 0.40 (p < 0.001) _ 0.40
Perceived relatedness 0.32 (p < 0.001) _ 0.32
Perceived competence 0.31 (p < 0.001) _ 0.31
Self-regulated learning 0.07 (p = 0.295) 0.18 0.25
English Perceived autonomy 0.30 (p < 0.001) _ 0.30
Perceived relatedness 0.19 (p = 0.004) _ 0.19
Perceived competence 0.29 (p < 0.001) _ 0.29
Self-regulated learning 0.20 (p < 0.001) 0.14 0.34
Perceived autonomy Self-regulated learning 0.17 (p = 0.015) _ 0.17
Perceived relatedness Self-regulated learning 0.10 (p = 0.068) _ 0.10
Perceived competence Self-regulated learning 0.25 (p < 0.001) _ 0.25
12 XIA et al.
The relationship between AI knowledge and needs satisfaction (H2)
Student AI knowledge positively and significantly predicted the satisfaction of each of the
three needs of autonomy (β = 0.40, p < 0.001), competence (β = 0.31, p < 0.001) and related-
ness (β = 0.32, p < 0.001).
The relationship between needs satisfaction and SRL (H3)
The results indicated that autonomy significantly predicted SRL (β = 0.17, p < 0.01), as did
competence (β = 0.25, p < 0.001), but relatedness did not significantly predict SRL (β = 0.10,
p > 0.05).
The mediating role of needs satisfaction (H4)
We used a bootstrapping approach to examine the mediating effect of needs satisfaction.

This approach does not require distribution assumptions; thereby, it avoids the issue of coef-
ficient product testing violating these assumptions. We calculated 95% confidence intervals
(CIs) for the effects of AI and English knowledge on SRL through bootstrapping with 1000
random samples, to further examine the mediating effects of autonomy, competence and
relatedness. The coefficient of the indirect effect of AI knowledge on SRL was 0.143 (95%
CI = [0.097, 0.293]) and the direct effect was not significant. In this indirect mediation, the
coefficient of the indirect effect of English knowledge on SRL was 0.176 (95% CI = [0.087,
0.221]), as the direct effect was significant and pointed in the same direction as the indi-
rect effect, indicating complementary mediation. The 95% CIs did not contain 0, and thus
both AI and English knowledge were significant indirect predictors of SRL through needs
in SDT. The total effects of AI knowledge and English knowledge on SRL were 0.143 and
0.176 respectively. The non-significant specific indirect effects via relatedness indicated that
the media tion of needs in SDT mainly resulted from autonomy and competence. Thus, our
results indicated that the three forms of needs satisfaction partially mediated the relation-
ships between AI knowledge and SRL and between English knowledge and SRL.
DISCUSSIONS AND CONCLUSIONS
This study provides a useful examination of the mediating role of the satisfaction of students'
needs in the relationship between their knowledge of English (disciplinary) and AI (technical)
and their SRL through AI in the K-12 context. This study makes three empirical contributions
and two theoretical contributions to the literature. We also offer practical suggestions for
researchers, practitioners and educational technology developers.
Empirical implications
First, we found a significant direct effect between English language skills and SRL (H1), which
is aligned with the studies of Greene et al. (2008, 2010), and Taub and colleagues (2014)
The students with good English profanely engaged more in a self-directive process to apply
their abilities to English learning with the chatbot. Prior disciplinary knowledge could predict
the SRL in digital learning environments. This is also explained by the 4 phases in Winne and
Hadwin (2008)'s model of SRL. Students with good English proficiency are more likely to use
an effective approach to complete the tasks in each of the phrases.
Second, we found an insignificant direct effect between AI knowledge and SRL (H2).
These findings suggest that having good AI knowledge is less likely to motivate student
engagement in the self-directive process, possibly because the process requires students to
use subject content knowledge to interact with the chatbot (Zimmerman, 2002). Chatbots do
not understand complex expressions (Bibauw et al., 2019; Jeon, 2022; Yin & Satar, 2020),
and good vocabularies are required to communicate with the chatbot. In addition good AI
knowledge alone does not facilitate the self-directive process, because the use of the chat-
bot may change students' learning behaviour but not their learning cognition and perception
(Kumar, 2021). In addition, the chatbot actively asks the students questions and responds to
their questions in a learner-friendly environment. This does not require sophisticated techni-
cal knowledge, as the students encounter very few technical problems. Our results together
imply that English knowledge is a prerequisite for self-regulated English learning using a
chatbot. The students are required to pay attention to language conversation rather than to
how the chatbot processes their input. Our findings differ from those of studies that identified
a relationship between computer self-efficiency and the motivation to learn through technol-
ogy (such as the original TAM; see Alfadda & Mahdi, 2021; Cheng, 2019; Chow et al., 2022;
Dorfsman & Horenczyk, 2022; Khlaisang et al., 2021). Most studies have used generic tech-
nologies (eg, mobile device, LMS), and AI educational technologies are often focused on a
specific discipline. Our study identifies an underexplored empirical relationship between prior
knowledge and SRL through AI technology.
Third, we found that satisfaction of the needs for autonomy and competence were
predictors (H3) of SRL, and mediators (H4) in the relationships between prior knowledge
(AI and English) and SRL. Our findings indicate that the two types of knowledge indirectly
predict SRL via the need for autonomy and competence. However, satisfaction of the need
for relatedness was not a mediator in the relationships between prior knowledge (AI and
English) and SRL. These findings are in line with the studies by Alrabai (2021) and Rutten
et al. (2012), who suggested that satisfying the needs for autonomy and competence are
important mediators of learning motivation or cognitive outcomes, but the effect of related-
ness is insignificant. Thus, a chatbot that supports the needs for autonomy and competence
can encourage self-direction. The theory of mind in AI (Langley et al., 2022) suggests that the
process is primarily cognitive rather than emotional, as social information can be processed
regardless of whether it comes from AI or another person (Lee, 2004). An appreciation of the
subject (ie, English) rather than the chatbot itself can encourage students to invest cogni-
tive effort in the process. This supports the findings of Kumar (2021), who suggests that
chatbots can improve student learning performance but not affective-motivational learning
outcomes such as perceptions of learning and motivation. However, whether the interaction
between students and chatbots stems from an innate response to social information or from
the emotional attachment as learning partners remains unclear. Further studies are required
to explore how the anthropomorphic characteristics of AI affect human cognitive processes.
Finally, our results indicate that supporting the needs for autonomy and competence
rather than relatedness is more important when encouraging student SRL through chat-
bots. Student experiences of SRL may therefore be more cognitive in nature. Their cogni-
tive processes can thus be explained by the desire to satisfy the needs of autonomy and
competence (Chiu, 2022; Sierens et al., 2009). Our findings support those of Holzer and
colleagues (2021), and suggest that relatedness in general is not positively related to SRL,
as the assumptions in SDT are context-dependent and student psychological needs vary
between education sectors and learning environments (Ryan & Deci, 2000). In addition,
SRL is a complex issue that involves cognitive, metacognitive, behavioural and motivational
14 XIA et al.
aspects of learning, so additional research is needed to determine the predictors of SRL with
AI applications.
Theoretical contributions
This study contributes to learner technology acceptance research by identifying the type
of prior knowledge and needs satisfaction as important factors. We distinguish student
knowledge into disciplinary (ie, English) and technical (ie, AI) knowledge. Most research on
technology acceptance (eg, TAM; Alfadda & Mahdi, 2021; Cheng, 2019; Chow et al., 2022;
Dorfsman & Horenczyk, 2022; Khlaisang et al., 2021) suggests that technical knowledge
such as computer self-efficacy and digital competence influences the motivation to learn
through technology. We suggest that disciplinary knowledge is more influential than techni-
cal knowledge when learning with domain-based technologies. This finding may encourage
researchers to re-examine their theories and models by identifying the relationships between
discipline and the type of technology (domain vs. non-domain).
Second, the research into psychological needs in AI for education remains limited
(Al Shamsi et al., 2022; Gado et al., 2021; Seufert et al., 2021). This study contributes to
SDT-based research by examining AI educational technology and presenting evidence of
how needs satisfaction can explain the development of SRL in English. The digital learning
environments investigated in most SDT research are non-disciplinary, such as LMS and
games (Bedenlier et al., 2020; Chiu, 2021; Chiu et al., 2022; Xie & Ke, 2011). We take a
different perspective, emphasising how needs satisfaction can inform AI disciplinary educa-
tional technology.
Practical suggestions
This study provides insights for teachers into how prior knowledge can affect student SRL
using conversational chatbots for English language learning. Our findings have three prac-
tical implications for K-12 teachers and school leaders. First, teachers should be aware that
current AI chatbot technology may not be sophisticated enough for SRL in school language
learning, although personalised learning is a common approach in such technologies
(Bibauw et al., 2019; Jeon, 2022; Vázquez-Cano et al., 2021). Not all students are capable
of regulating their learning with chatbots. Current chatbot technology may be appropriate for
more advanced students who have good language proficiency. We suggest that teachers
can use chatbots for knowledge consolidation rather than exploration (Fryer et al., 2019;
Vázquez-Cano et al., 2021), and thus in the final stages of learning. This can be more effec-
tive in increasing students' English language knowledge.
Second, teachers should support students' competence and autonomy, as these were
found to be the factors that directly predicted SRL. They are essential in shaping student
motivation and thus SRL in an AI-interactive environment. Teachers should support student
competence by giving step-by-step guidelines or sample phrases, sentences and questions
(Hartnett, 2015). This can encourage successful learning experiences with chatbots. The
students can learn from example conversations and develop their own input and ideas in
learning through interacting with chatbots.
Third, K-12 school leaders can offer professional development training in AI technology to
teachers, as this technology is still new and challenging (Chiu et al., 2022). Our results reveal
how students' prior knowledge can determine the effectiveness of SRL with conversational
chatbots and also highlight their affordances and limitations. School leaders and teacher
educators should include discussion of these characteristics in professional development

programmes for digital education.
Finally, we provide insights for researchers and developers of AI conversational chatbots.
We reveal that current chatbot technologies are not sophisticated enough for young students
with limited English language proficiency. These students often have no idea about what to
ask or offer meaningless input. Thus, the developers should use more data from weaker
students or identify their levels of expertise (academic skill or cognitive capacities) for further
conversations (Chiu & Mok, 2017).
LIMITATIONS AND FUTURE DIRECTIONS

This study has four main limitations. First, the proposed model is examined using structural
equation modelling. Studies using other research methods should be conducted to validate
our findings. We suggest using intervention-based or experimental studies to test the effec-
tiveness of chatbots in terms of language learning. Second, the participants are non-native
English speakers in Hong Kong middle schools. Their native Chinese is more a graphical
language (Chen & Zhou, 2010). It would be helpful for future research to use a sample of
non-native English speakers familiar with a non-graphical language such as Thai. Third, our
research relies on questionnaire surveys as a quantitative approach, but this does not allow
in-depth analysis of the relationships between needs satisfaction and prior knowledge in
terms of SRL with chatbots. Further studies should take a more qualitative approach to gain
a more thorough and deeper understanding of this issue.
Our empirical implications also suggest future research directions. These include (i) estab-
lishing the empirical relationship between prior knowledge and SRL using AI technology; (ii)
exploring how the anthropomorphic characteristics of AI (conversation in chatbots) affect
human cognitive processes; and (iii) determining the predictors of SRL with AI applications.
ACKNOWLEDGEMENTS
The authors thank the school teachers and students for participating in this study.
F U N D I N G I N F O R M AT I O N
This study is funded by Quality Education Fund (project code: 6906035)
C O N F L I C T O F I N T E R E S T S TAT E M E N T
There is no conflict of interests between the author and participants.
D ATA AVA I L A B I L I T Y S TAT E M E N T

The data that support the findings of this study are available on request from the correspond-
ing author. The data are not publicly available due to privacy or ethical restrictions.
E T H I C S S TAT E M E N T
This study got ethical clearance from the author's university.
16 XIA et al.
ORCID
Qi Xia https://orcid.org/0000-0003-0538-7665
Thomas K. F. Chiu https://orcid.org/0000-0003-2887-5477
Ching Sing Chai https://orcid.org/0000-0002-6298-4813
Kui Xie https://orcid.org/0000-0002-7173-4859
REFERENCES
Adamopoulou, E., & Moussiades, L. (2020). Chatbots: History, technology, and applications. Machine Learning with
Applications, 2, 100006. https://doi.org/10.1016/j.mlwa.2020.100006
Al Shamsi, J. H., Al-Emran, M., & Shaalan, K. (2022). Understanding key drivers affecting students' use of artifi-
cial intelligence-based voice assistants. Education and Information Technologies, 21, 8071–8091. https://doi.
org/10.1007/s10639-022-10947-3
Alfadda, H. A., & Mahdi, H. S. (2021). Measuring students' use of zoom application in language course based on
the technology acceptance model (TAM). Journal of Psycholinguistic Research, 50(4), 883–900. https://doi.
org/10.1007/s10936-020-09752-1
Alrabai, F. (2021). The influence of autonomy-supportive teaching on EFL students' classroom autonomy: An exper-
imental intervention. Frontiers in Psychology, 12, 728657. https://doi.org/10.3389/fpsyg.2021.728657
Bedenlier, S., Bond, M., Buntins, K., Zawacki-Richter, O., & Kerres, M. (2020). Facilitating student engagement
through educational technology in higher education: A systematic review in the field of arts and humanities.
Australasian Journal of Educational Technology, 36(4), 126–150. https://doi.org/10.14742/ajet.5477
Bibauw, S., François, T., & Desmet, P. (2019). Discussing with a computer to practice a foreign language: Research
synthesis and conceptual framework of dialogue-based CALL. Computer Assisted Language Learning, 32(8),
827–877. https://doi.org/10.1080/09588221.2018.1535508
Blut, M., Wang, C., Wünderlich, N. V., & Brock, C. (2021). Understanding anthropomorphism in service provision: A
meta-analysis of physical robots, chatbots, and other AI. Journal of the Academy of Marketing Science, 49(4),
632–658. https://doi.org/10.1007/s11747-020-00762-y
Chen, C.-M., & Chen, F.-Y. (2014). Enhancing digital reading performance with a collaborative reading annotation
system. Computers & Education, 77, 67–81. https://doi.org/10.1016/j.compedu.2014.04.010
Chen, C.-M., Chen, L.-C., & Yang, S.-M. (2019). An English vocabulary learning app with self-regulated learning
mechanism to improve learning performance and motivation. Computer Assisted Language Learning, 32(3),
237–260. https://doi.org/10.1080/09588221.2018.1485708
Chen, C.-M., & Huang, S.-H. (2014). Web-based reading annotation system with an attention-based self-regulated
learning mechanism for promoting reading performance. British Journal of Educational Technology, 45(5),
959–980. https://doi.org/10.1111/bjet.12119
Chen, C.-M., Li, M.-C., & Chen, T.-C. (2020). A web-based collaborative reading annotation system with gamification
mechanisms to improve reading performance. Computers & Education, 144, 103697. https://doi.org/10.1016/j.
compedu.2019.103697
Chen, S., & Zhou, J. (2010). Creative writing strategies of young children: Evidence from a study of Chinese emer-
gent writing. Thinking Skills and Creativity, 5(3), 138–149. https://doi.org/10.1016/j.tsc.2010.09.002
Cheng, G. (2019). Exploring factors influencing the acceptance of visual programming environment among
boys and girls in primary schools. Computers in Human Behavior, 92, 361–372. https://doi.org/10.1016/j.
chb.2018.11.043
Chiu, T. K. F. (2017). Introducing electronic textbooks as daily-use technology in schools: A top-down adoption
process. British Journal of Educational Technology, 48(2), 524–537. https://doi.org/10.1111/bjet.12432
Chiu, T. K. F. (2021). Digital support for student engagement in blended learning based on self-determination theo-
ryk. Computers in Human Behavior, 124, 106909. https://doi.org/10.1016/j.chb.2021.106909
Chiu, T. K. F. (2022). Applying the self-determination theory (SDT) to explain student engagement in online learning
during the COVID-19 pandemic. Journal of Research on Technology in Education, 54(Suppl. 1), S14–S30.
https://doi.org/10.1080/15391523.2021.1891998
Chiu, T. K. F., & Chai, C. S. (2020). Sustainable curriculum planning for artificial intelligence education: A self-deter-
mination theory perspective. Sustainability, 12(14), 5568. https://doi.org/10.3390/su12145568
Chiu, T. K. F., Meng, H., Chai, C. S., King, I., Wong, S., & Yeung, Y. (2022). Creation and evaluation of a pre-tertiary
artificial intelligence (AI) curriculum. IEEE Transactions on Education, 65(1), 30–39. https://doi.org/10.1109/
TE.2021.3085878
Chiu, T. K. F., & Mok, I. A. C. (2017). Learner expertise and mathematics different order thinking skills in multimedia
learning. Computers & Education, 107, 147–164. https://doi.org/10.1016/j.compedu.2017.01.008
Chow, M. C. M., Hung, M. S. Y., Chu, J. W. K., & Lam, S. K. K. (2022). Factors affecting nursing students' intention
to use a 3D game to learn field triage skills: A structural equation modelling analysis. BMC Nursing, 21(1), 46.
https://doi.org/10.1186/s12912-022-00826-0
Cleary, T. J., & Kitsantas, A. (2017). Motivation and self-regulated learning influences on middle school mathematics
achievement. School Psychology Review, 46(1), 88–107. https://doi.org/10.17105/SPR46-1.88-107
Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology.
MIS Quarterly, 13(3), 319–340. https://doi.org/10.2307/249008
De Vreede, T., Raghavan, M., & De Vreede, G.-J. (2021). Design foundations for AI assisted decision making: A self
determination theory approach. https://doi.org/10.24251/hicss.2021.019
Deci, E. L., & Ryan, R. M. (2004). Handbook of self-determination research. University Rochester Press.
Deniz, E., Victor, L., Dirk, I., Xun, G., & Raymond, M. (2014). An investigation of the interrelationships between moti-
vation, engagement, and complex problem solving in game-based learning. Journal of Educational Technology
& Society, 17(1), 42–53. http://www.jstor.org/stable/jeductechsoci.17.1.42
Dong, A., Jong, M. S.-Y., & King, R. B. (2020). How does prior knowledge influence learning engagement? The medi-
ating roles of cognitive load and help-seeking. Frontiers in Psychology, 11, 591203. https://doi.org/10.3389/
fpsyg.2020.591203
Dorfsman, M., & Horenczyk, G. (2022). The coping of academic staff with an extreme situation: The transition from
conventional teaching to online teaching. Education and Information Technologies, 27(1), 267–289. https://doi.
org/10.1007/s10639-021-10675-0
Ferreira Chame, H., Pinto Mota, F., & da Costa Botelho, S. S. (2019). A dynamic computational model of motivation
based on self-determination theory and CANN. Information Sciences, 476, 319–336. https://doi.org/10.1016/j.
ins.2018.09.055
Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and meas-
urement error. Journal of Marketing Research, 18(1), 39–50. https://doi.org/10.1177/002224378101800104
Fryer, L. K., Nakao, K., & Thompson, A. (2019). Chatbot learning partners: Connecting learning experiences, inter-
est and competence. Computers in Human Behavior, 93, 279–289. https://doi.org/10.1016/j.chb.2018.12.023
Furrer, C., & Skinner, E. (2003). Sense of relatedness as a factor in children's academic engagement and perfor-
mance. Journal of Educational Psychology, 95(1), 148–162. https://doi.org/10.1037/0022-0663.95.1.148
Gado, S., Kempen, R., Lingelbach, K., & Bipp, T. (2021). Artificial intelligence in psychology: How can we enable
psychology students to accept and use artificial intelligence? Psychology Learning and Teaching, 21, 37–56.
https://doi.org/10.1177/14757257211037149
Greene, J. A., Costa, L.-J., Robertson, J., Pan, Y., & Deekens, V. M. (2010). Exploring relations among college
students’ prior knowledge, implicit theories of intelligence, and self-regulated learning in a hypermedia environ-
ment. Computers & Education, 55(3), 1027–1043. https://doi.org/10.1016/j.compedu.2010.04.013
Greene, J. A., Moos, D. C., Azevedo, R., & Winters, F. I. (2008). Exploring differences between gifted and grade-
level students’ use of self-regulatory learning processes with hypermedia. Computers & Education, 50(3),
1069–1083. https://doi.org/10.1016/j.compedu.2006.10.004
Grolnick, W. S., & Raftery-Helmer, J. N. (2015). Contexts supporting self-regulated learning at school transitions. In
Self-regulated learning interventions with at-risk youth: Enhancing adaptability, performance, and well-being
(pp. 251–276). American Psychological Association. https://doi.org/10.1037/14641-012
Hartnett, M. K. (2015). Influences that undermine learners' perceptions of autonomy, competence and relatedness
in an online context. Australasian Journal of Educational Technology, 31(1), 86–99. https://doi.org/10.14742/
ajet.1526
Hew, T.-S., & Kadir, S. L. S. A. (2016). Understanding cloud-based VLE from the SDT and CET perspectives:
Development and validation of a measurement instrument. Computers & Education, 101, 132–149. https://doi.
org/10.1016/j.compedu.2016.06.004
Hirt, C. N., Karlen, Y., Merki, K. M., & Suter, F. (2021). What makes high achievers different from low achievers?
Self-regulated learners in the context of a high-stakes academic long-term task. Learning and Individual Differ-
ences, 92, 102085. https://doi.org/10.1016/j.lindif.2021.102085
Holzer, J., Luftenegger, M., Kaser, U., Korlat, S., Pelikan, E., Schultze-Krumbholz, A., Spiel, C., Wachs, S., &
Schober, B. (2021). Students’ basic needs and well-being during the COVID-19 pandemic: A two-country study
of basic psychological need satisfaction, intrinsic learning motivation, positive emotion and the moderating
role of self-regulated learning. International Journal of Psychology, 56(6), 843–852. https://doi.org/10.1002/
ijop.12763
Holzer, J., Lüftenegger, M., Korlat, S., Pelikan, E., Salmela-Aro, K., Spiel, C., & Schober, B. (2021). Higher educa-
tion in times of COVID-19: University students' basic need satisfaction, self-regulated learning, and well-being.
AERA Open, 7, 23328584211003164. https://doi.org/10.1177/23328584211003164
Hong, J.-W. (2022). I was born to love AI: The influence of social status on AI self-efficacy and intentions to use AI.
International Journal of Communication, 16, 172–191. https://ijoc.org/index.php/ijoc/article/view/17728/3632
Hu, L. T., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional
criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6(1), 1–55.
https://doi.org/10.1080/10705519909540118
18 XIA et al.
Huang, Y.-C., Backman, S. J., Chang, L.-L., Backman, K. F., & McGuire, F. A. (2013). Experiencing student learning
and tourism training in a 3D virtual world: An exploratory study. Journal of Hospitality, Leisure, Sport & Tourism
Education, 13, 190–201. https://doi.org/10.1016/j.jhlste.2013.09.007
Jeon, J. (2022). Exploring AI chatbot affordances in the EFL classroom: Young learners' experiences and perspec-
tives. Computer Assisted Language Learning, 1–26. https://doi.org/10.1080/09588221.2021.2021241
Kaliyaperumal, K. (2004). Guideline for conducting a knowledge, attitude and practice (KAP) study. AECS Illumi-
nation, 4, 7–9.
Khlaisang, J., Teo, T., & Huang, F. (2021). Acceptance of a flipped smart application for learning: A study among
Thai university students. Interactive Learning Environments, 29(5), 772–789. https://doi.org/10.1080/104948
20.2019.1612447
Kumar, J. A. (2021). Educational chatbots for project-based learning: Investigating learning outcomes for a team-
based design course. International Journal of Educational Technology in Higher Education, 18(1), 65. https://
doi.org/10.1186/s41239-021-00302-w
Langley, C., Cirstea, B. I., Cuzzolin, F., & Sahakian, B. J. (2022). Theory of mind and preference learning at the
interface of cognitive science, neuroscience, and AI: A review. Frontiers in Artificial Intelligence, 5, 778852.
https://doi.org/10.3389/frai.2022.778852
Lee, K. M. (2004). Why presence occurs: Evolutionary psychology, media equation, and presence. Presence: Tele-
operators and Virtual Environments, 13(4), 494–505. https://doi.org/10.1162/1054746041944830
León, J., Núñez, J. L., & Liew, J. (2015). Self-determination and STEM education: Effects of autonomy, motiva-
tion, and self-regulated learning on high school math achievement. Learning and Individual Differences, 43,
156–163. https://doi.org/10.1016/j.lindif.2015.08.017
Lin, C.-H., Zhang, Y., & Zheng, B. (2017). The roles of learning strategies and motivation in online language learn-
ing: A structural equation modeling analysis. Computers & Education, 113, 75–85. https://doi.org/10.1016/j.
compedu.2017.05.014
Liu, W. C., Wang, C. K. J., Kee, Y. H., Koh, C., Lim, B. S. C., & Chua, L. (2014). College students’ motivation and
learning strategies profiles and academic achievement: A self-determination theory approach. Educational
Psychology, 34(3), 338–353. https://doi.org/10.1080/01443410.2013.785067
Luo, Y. M., Lin, J. P., & Yang, Y. (2021). Students’ motivation and continued intention with online self-regulated learn-
ing: A self-determination theory perspective. Zeitschrift für Erziehungswissenschaft., 24, 1379–1399. https://
doi.org/10.1007/s11618-021-01042-3
Malik, R., Sharma, A., Trivedi, S., & Mishra, R. (2021). Adoption of chatbots for learning among university students:
Role of perceived convenience and enhanced performance. International Journal of Emerging Technologies in
Learning, 16(18), 200–212. https://doi.org/10.3991/ijet.v16i18.24315
Markauskaite, L., Marrone, R., Poquet, O., Knight, S., Martinez-Maldonado, R., Howard, S., Tondeur, J., De Laat,
M., Shum, S. B., Gašević, D., & Siemens, G. (2022). Rethinking the entwinement between artificial intelligence
and human learning: What capabilities do learners need for a world with AI? Computers and Education: Artifi-
cial Intelligence, 3, 100056. https://doi.org/10.1016/j.caeai.2022.100056
Mouratidis, A., Vansteenkiste, M., Michou, A., & Lens, W. (2013). Perceived structure and achievement goals as
predictors of students' self-regulated learning and affect and the mediating role of competence need satisfac-
tion. Learning and Individual Differences, 23, 179–186. https://doi.org/10.1016/j.lindif.2012.09.001
Ng, J. Y. Y., Ntoumanis, N., Thøgersen-Ntoumani, C., Deci, E. L., Ryan, R. M., Duda, J. L., & Williams, G. C. (2012).
Self-determination theory applied to health contexts: A meta-analysis. Perspectives on Psychological Science,
7(4), 325–340. https://doi.org/10.1177/1745691612447309
Niemiec, C. P., & Ryan, R. M. (2009). Autonomy, competence, and relatedness in the classroom: Applying self-de-
termination theory to educational practice. Theory and Research in Education, 7(2), 133–144.
Pelau, C., Dabija, D.-C., & Ene, I. (2021). What makes an AI device human-like? The role of interaction quality, empa-
thy and perceived psychological anthropomorphic characteristics in the acceptance of artificial intelligence in
the service industry. Computers in Human Behavior, 122, 106855. https://doi.org/10.1016/j.chb.2021.106855
Pelikan, E. R., Lüftenegger, M., Holzer, J., Korlat, S., Spiel, C., & Schober, B. (2021). Learning during COVID-19:
The role of self-regulated learning, motivation, and procrastination for perceived competence. Zeitschrift für
Erziehungswissenschaft, 24(2), 393–418. https://doi.org/10.1007/s11618-021-01002-x
Pintrich, P. R. (2004). A conceptual framework for assessing motivation and self-regulated learning in college
students. Educational Psychology Review, 16(4), 385–407. https://doi.org/10.1007/s10648-004-0006-x
Radel, R., Pelletier, L., Baxter, D., Fournier, M., & Sarrazin, P. (2014). The paradoxical effect of controlling context
on intrinsic motivation in another activity. Learning and Instruction, 29, 95–102. https://doi.org/10.1016/j.
learninstruc.2013.09.004
Rutten, C., Boen, F., & Seghers, J. (2012). How school social and physical environments relate to autonomous
motivation in physical education: The mediating role of need satisfaction. Journal of Teaching in Physical
Education, 31(3), 216–230. https://doi.org/10.1123/jtpe.31.3.216
Ryan, R. M., & Deci, E. L. (2000). Intrinsic and extrinsic motivations: Classic definitions and new directions. Contem-
porary Educational Psychology, 25(1), 54–67. https://doi.org/10.1006/ceps.1999.1020
Ryan, R. M., & Deci, E. L. (2017). Self-determination theory: Basic psychological needs in motivation, development,
and wellness. The Guilford Press. https://doi.org/10.1521/978.14625/28806
Ryan, R. M., & Deci, E. L. (2020). Intrinsic and extrinsic motivation from a self-determination theory perspec-
tive: Definitions, theory, practices, and future directions. Contemporary Educational Psychology, 61, 101860.
https://doi.org/10.1016/j.cedpsych.2020.101860
Ryan, R. M., Rigby, C. S., & Przybylski, A. (2006). The motivational pull of video games: A self-determination theory
approach. Motivation and Emotion, 30(4), 344–360. https://doi.org/10.1007/s11031-006-9051-8
Seufert, S., Guggemos, J., & Sailer, M. (2021). Technology-related knowledge, skills, and attitudes of pre- and
in-service teachers: The current situation and emerging trends. Computers in Human Behavior, 115, 106552.
https://doi.org/10.1016/j.chb.2020.106552
Sha, L., Looi, C.-K., Chen, W., Seow, P., & Wong, L.-H. (2012). Recognizing and measuring self-regulated learning
in a mobile learning environment. Computers in Human Behavior, 28(2), 718–728. https://doi.org/10.1016/j.
chb.2011.11.019
Sierens, E., Vansteenkiste, M., Goossens, L., Soenens, B., & Dochy, F. (2009). The synergistic relationship of
perceived autonomy support and structure in the prediction of self-regulated learning. British Journal of Educa-
tional Psychology, 79(1), 57–68. https://doi.org/10.1348/000709908X304398
Song, P., & Wang, X. (2020). A bibliometric analysis of worldwide educational artificial intelligence research devel-
opment in recent twenty years. Asia Pacific Education Review, 21(3), 473–486. https://doi.org/10.1007/
s12564-020-09640-2
Soper, D. S. (2020). A-priori Sample Size Calculator for Structural Equation Models [Software]. https://www.daniel-
soper.com/statcalc
Sun, Z., Xie, K., & Anderman, L. H. (2018). The role of self-regulated learning in students’ success in flipped
undergraduate math courses. The Internet and Higher Education, 36, 41–53. https://doi.org/10.1016/j.
iheduc.2017.09.003
Taub, M., & Azevedo, R. (2019). How does prior knowledge influence eye fixations and sequences of cognitive
and metacognitive SRL processes during learning with an intelligent tutoring system? International Journal of
Artificial Intelligence in Education, 29(1), 1–28. https://doi.org/10.1007/s40593-018-0165-4
Taub, M., Azevedo, R., Bouchet, F., & Khosravifar, B. (2014). Can the use of cognitive and metacognitive self-reg-
ulated learning strategies be predicted by learners' levels of prior knowledge in hypermedia-learning environ-
ments? Computers in Human Behavior, 39, 356–367. https://doi.org/10.1016/j.chb.2014.07.018
van der Graaf, J., Lim, L., Fan, Y., Kilgour, J., Moore, J., Gašević, D., Bannert, M., & Molenaar, I. (2022). The
dynamics between self-regulated learning and learning outcomes: An exploratory approach and implications.
Metacognition and Learning, 1–27, 745–771. https://doi.org/10.1007/s11409-022-09308-9
Vázquez-Cano, E., Mengual-Andrés, S., & López-Meneses, E. (2021). Chatbot to improve learning punctuation in
Spanish and to enhance open and flexible learning environments. International Journal of Educational Tech-
nology in Higher Education, 18(1), 33. https://doi.org/10.1186/s41239-021-00269-8
Wan, Z., Compeau, D., & Haggerty, N. (2012). The effects of self-regulated learning processes on e-learning
outcomes in organizational settings. Journal of Management Information Systems, 29(1), 307–340. https://doi.
org/10.2753/MIS0742-1222290109
Wang, X., Chen, J., & Zhang, T. (2021). Facilitating English grammar learning by a personalized mobile-assisted
system with a self-regulated learning mechanism. Frontiers in Psychology, 12, 624430. https://doi.org/10.3389/
fpsyg.2021.624430
Winne, P. H., & Hadwin, A. F. (2008). The weave of motivation and self-regulated learning. In D. H. Schunk & B. J.
Zimmerman (Eds.), Motivation and self-regulated learning: Theory, research and applications (pp. 298–314).
Lawrence Erlbaun.
Winters, F. I., Greene, J. A., & Costich, C. M. (2008). Self-regulation of learning within computer-based learning
environments: A critical analysis. Educational Psychology Review, 20(4), 429–444. https://doi.org/10.1007/
s10648-008-9080-9
Xie, K., & Ke, F. (2011). The role of students' motivation in peer-moderated asynchronous online discussions. British
Journal of Educational Technology, 42(6), 916–930. https://doi.org/10.1111/j.1467-8535.2010.01140.x
Yang, T.-C., Chen, M. C., & Chen, S. Y. (2018). The influences of self-regulated learning support and prior knowl-
edge on improving learning performance. Computers & Education, 126, 37–52. https://doi.org/10.1016/j.
compedu.2018.06.025
Yang, X., & Aurisicchio, M. (2021). Designing conversational agents: A self-determination theory approach. Paper
presented at Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems, Yokohama,
Japan. https://doi.org/10.1145/3411764.3445445
Yin, Q., & Satar, M. (2020). English as a foreign language learner interaction with chatbots: Negotiation for meaning.
International Online Journal of Education and Teaching, 7(2), 390–410. http://iojet.org/index.php/IOJET/article/
view/707
20 XIA et al.
Zawacki-Richter, O., Marín, V. I., Bond, M., & Gouverneur, F. (2019). Systematic review of research on artificial
intelligence applications in higher education—Where are the educators? International Journal of Educational
Technology in Higher Education, 16(1), 39. https://doi.org/10.1186/s41239-019-0171-0
Zimmerman, B. J. (2002). Becoming a self-regulated learner: An overview. Theory into Practice, 41(2), 64–70.
https://doi.org/10.1207/s15430421tip4102_2
How to cite this article: Xia, Q., Chiu, T. K. F., Chai, C. S., & Xie, K. (2023). The
mediating effects of needs satisfaction on the relationships between prior knowledge
and self-regulated learning through artificial intelligence chatbot. British Journal of
Educational Technology, 00, 1–20. https://doi.org/10.1111/bjet.13305

Brit J Educational Tech - 2023 - Xia - The Mediating Effects of Needs Satisfaction On The Relationships Between Prior

Uploaded by

Copyright:

Available Formats

Brit J Educational Tech - 2023 - Xia - The Mediating Effects of Needs Satisfaction On The Relationships Between Prior

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Brit J Educational Tech - 2023 - Xia - The Mediating Effects of Needs Satisfaction On The Relationships Between Prior

Uploaded by

Copyright:

Available Formats

14678535, 0, Downloaded from https://bera-journals.onlinelibrary.wiley.com/doi/10.1111/bjet.13305 by Orta Dogu Teknik Universitesi, Wiley Online Library on [05/02/2023].

The mediating effects of needs satisfaction

Qi Xia1 | Thomas K. F. Chiu1,2 | Ching Sing Chai1 |

Br J Educ Technol. 2023;00:1–20. wileyonlinelibrary.com/journal/bjet 1

learning. The findings also revealed that current chat-

is strongly associated with students' motivations to learn through digital technologies.

SDT and SRL

How AI and disciplinary knowledge affect needs satisfaction and SRL

F I G U R E 1 Research model. *p < 0.05, **p < 0.01, ***p < 0.001.

Participants and research procedure

Chinese translation of the question

questions and suggested answers.

and suggested answers.

F I G U R E 2 Screenshots of the chatbot.

Questionnaire and test

Descriptive statistics and reliabilities

T A B L E 1 Descriptive statistics and reliabilities (n = 323).

T A B L E 2 Factor loadings and correlation matrix for all variables.

Item CR Factor loadings 1 2 3 4 5 6

Mediating effect analysis

The relationship between English knowledge and needs satisfaction (H1)

T A B L E 3 Standardised total, total, indirect and direct effects among variables.

Predictor Mediating/criterion variable Direct effect Indirect effect Total effect

The relationship between AI knowledge and needs satisfaction (H2)

The relationship between needs satisfaction and SRL (H3)

The mediating role of needs satisfaction (H4)

We used a bootstrapping approach to examine the mediating effect of needs satisfaction.

DISCUSSIONS AND CONCLUSIONS

educators should include discussion of these characteristics in professional development

LIMITATIONS AND FUTURE DIRECTIONS

D ATA AVA I L A B I L I T Y S TAT E M E N T

You might also like

F I G U R E 1 Research model. p < 0.05, p < 0.01, p < 0.001.