The Impact of Integrating an Intelligent Personal Assistant (IPA) on Secondary School Physics Students’ Scientific Inquiry Skills

Intelligent personal assistants (IPAs) carry massive potential in enhancing students’ performance through individualized dynamic scaffolding strategy. Despite IPAs being increasingly recognized among educationists, little is known about their application in the development of students’ scientific inquiry skills, particularly in physics. This study integrated the use of Apple's Siri in physics learning and examined the impact on secondary school student's scientific inquiry skills, and how these interactions affect their learning experiences. This mixed-method study conducted two field quasi-experiments (School A: science boarding school; School B: vocational boarding school) and post-experimental focus group discussions. Each school had two classes that were randomly assigned as experimental and control groups. The quantitative data showed that the experimental groups developed significantly higher scientific inquiry skills in comparison to the control groups (School A: p = 0.050; School B: p < 0.001) with moderate (d = 0.62) to large effect size (d = 1.82). The qualitative data further revealed the perceived ability of the IPA to support students’ individualized learning by providing timely scaffolding. This study offers substantial empirical data to support the effectiveness of the IPA in increasing secondary school students’ achievement in physics and scientific inquiry skills.

The Impact of Integrating an Intelligent Personal Assistant (IPA) on Secondary School Physics Students' Scientific Inquiry Skills Nurfaradilla Mohamad Nasri , Nurfarahin Nasri, Nur Faraliyana Nasri, and Mohamad Asyraf Abd Talib Abstract-Intelligent personal assistants (IPAs) carry massive potential in enhancing students' performance through individualized dynamic scaffolding strategy. Despite IPAs being increasingly recognized among educationists, little is known about their application in the development of students' scientific inquiry skills, particularly in physics. This study integrated the use of Apple's Siri in physics learning and examined the impact on secondary school student's scientific inquiry skills, and how these interactions affect their learning experiences. This mixed-method study conducted two field quasi-experiments (School A: science boarding school; School B: vocational boarding school) and post-experimental focus group discussions. Each school had two classes that were randomly assigned as experimental and control groups. The quantitative data showed that the experimental groups developed significantly higher scientific inquiry skills in comparison to the control groups (School A: p = 0.050; School B: p < 0.001) with moderate (d = 0.62) to large effect size (d = 1.82). The qualitative data further revealed the perceived ability of the IPA to support students' individualized learning by providing timely scaffolding. This study offers substantial empirical data to support the effectiveness of the IPA in increasing secondary school students' achievement in physics and scientific inquiry skills.
Index Terms-Individualized dynamic scaffolding, intelligent personal assistant (IPA), scientific inquiry skills, secondary school physics learning.

I. INTRODUCTION
T ODAY, more important than ever, educators carry immense responsibility to prepare a highly competent workforce that is capable of making informed decisions and actions based on critical analysis of scientific-based information. By driving revolutionary innovation and building competitive advantage in a global economy, a focus on nurturing scientific This work involved human subjects or animals in its research. The author(s) confirm(s) that all human/animal subject research procedures and protocols are exempt from review board approval.
Digital Object Identifier 10.1109/ TLT.2023.3241058 inquiry in addition to attaining academic achievement is central in any education plan or policy. At the same time, educators face constant challenges in catching new waves of technological advancement and riding them to diversify existing pedagogical practices. In this context, there is an urgent need to integrate easily available and user-friendly technology to help educators develop scientific inquiry skills among students. Based on the National Science Education Standards [1], scientific inquiry describes the various methods by which scientists investigate the natural world and develop understandings following the facts gained. Basic principles governing well-grounded scientific inquiry involve hands-on research activities, objectivity and rigorous thinking, and most importantly, it undertakes a systematic investigation approach. Primarily, scientific inquiry entails more than just "doing experiments" as it necessitates a high level of imagination and inventiveness [2]. However, to leverage this concept at a level appropriate for secondary schools, fundamental scientific inquiry skills are outlined as follows: 1) identify a problem; 2) devise a hypothesis; 3) plan procedures; 4) carry out a scientific experiment; 5) gather, organize, and interpret data; 6) employ statistical techniques to support conclusions; 7) report findings [3], [4]. A point to note, scientific inquiry varies significantly in the amount of freedom and autonomy given to the students. For this reason, scientific inquiry can be categorized into the following three main approaches: 1) structured scientific inquiry, where students conduct the investigation process according to the prescribed procedure; 2) guided scientific inquiry, where students investigate collaboratively with teachers; 3) open inquiry, where students take responsibility and ownership to independently design the investigation procedures [5], [6]. The first two approaches place great emphasis on external forms of assistance, whereas the third approach views students as the key figure to direct their own learning, resonating valuable virtue of competent learners with high scientific inquiry skills.
In the world of mobile devices, intelligent personal assistants (IPAs) gain strong recognition due to its potential in supporting students' learning. By combining artificial intelligence in its design, IPAs are able to perform user commands and provide communication or interaction. Voice-controlled IPAs, such as Amazon's Alexa, Samsung's S Voice, and Apple's Siri, can be utilized for self-learning purposes as they leverage natural spoken language and semantic comprehension approaches to obtain information required by the users. In other words, IPAs can be described as having a higher level of engagement and intelligence in comparison to other user assistance systems as IPAs can instantly respond to human input and offer appropriate guidance through a complex and complicated task.
Developed on the premise of automation and convenience, most IPAs operate on smart mobile devices with an advanced voice user interface. This new consumer technology commonly establishes effective learning experiences through an individualized dynamic scaffolding strategy. Generally, scaffolding is described as offering support to students as needed and reducing that support as the students' competence grows [7]. There are two forms of scaffolding, namely static scaffolding and dynamic scaffolding [8]. As derived from Vygotsky's Social Development Theory, interaction is a crucial aspect or component in scaffolding. Considering that there are no interactions between the system and the students, static scaffolding maintains constant across time and the same for all students [9].
In contrast to static scaffolding which often provides fixed, predictive support based on tasks, dynamic scaffolding strategy requires the system to constantly assess the students' understanding and responsively provide contextual assistance to facilitate the learning process [10]. Examining the impact of dynamic scaffolding on middle school students within a computer-based learning environment [8] discovered that scaffolding improved learning outcomes when compared to the nonscaffolding control group. This type of research has led us to conclude that dynamic scaffolding in IPAs may also have a favorable impact on students' learning in physics education. Similarly, Luckin et al. [11] described the potential of IPAs in replicating one-to-one human tutoring as they are able to make evidence-based decisions by engaging the students in an active dialogue while at the same time assessing the student's learning needs before suggesting appropriate content and providing necessary scaffolding.
Many researchers underline the following key characteristics of dynamic scaffolding [12], [13]: 1) it aids students' understanding of how to complete the work and why it should be completed in that manner; 2) it details disciplinary thinking and strategies; 3) it provides expert guidance; 4) it supports the organization of complex tasks and the reduction of cognitive load, allowing students to concentrate on the components of the task that are important to achieve the learning objectives. Furthermore, there exists a general consensus that students require scaffolding tools and external guidance to assist them in organizing and directing their own projects [14], [15], [16] while mastering the essence of scientific inquiry skills [17], [18]. However, providing guidance to individual students may pose critical challenge due to growing classroom size and varying students' learning needs. Therefore, this study suggests that guidance should be made available following students' current demands and should decrease over time.
The application of IPAs focuses on interactive engagement between the individual student and technology which can be explained using the Interactive, Constructive, Active, Passive (ICAP) framework by Chi and Wylie [19]. The ICAP framework postulates effective learning progression from the passive stage, to the active stage, followed by the constructive stage and, finally, the interactive stage requires responsive support from the educators. Within this study context, the IPA is viewed as a responsive educator who provides the necessary scaffolding to support students' learning. More importantly, the ICAP framework delivers an understanding that students who maintain increasing engagement with the learning content, are more likely to learn and master more skills [19]. Given this knowledge, this study argues that providing students with the opportunity to receive IPA guidance during the scientific inquiry process is essential for learning physics more effectively.
IPAs have drawn a lot of interest because of how well they support students' learning. For instance, IPAs have been demonstrated to be successful in improving students' speaking and listening abilities [20]. This effect was noticeable since IPAs enable the students to speak with a system and listen to their recorded voices back. However, some studies did not find a significant difference between the usage of IPAs and the improvement of students' listening skills [21], [22]. These researchers hypothesize that one reason could be the absence of explicit guidance for students to develop listening scaffolding when engaging with IPAs. Therefore, the use of different IPAs with different student cohorts can cause the empirical evidence to be contradictory. The results of these studies demonstrate that the relationship between IPAs and learning is too complex to make broad assumptions about its impact when used in the context of physics learning.
When discussing scientific inquiry skills, researchers typically address the use of inquiry-based learning to assist students in developing these skills. However, there are many different ways to apply inquiry-based learning. Several studies promoted inquiry-based learning in the classroom by using small group activities [23], [24]. Other studies concentrate on encouraging scientific inquiry by utilizing outdoor classroom pedagogies at an animal sanctuary, an ethanol plant, and a forest [25], [26]. The use of IPAs for inquiry-based learning is however limited as previous studies are only found to employ computer simulations to assist students' development of scientific inquiry skills [27], [28].
In summary, the novelty of this study lies in the following three key aspects that have an influential impact on how the IPA supports students' learning through scaffolding: 1) functionality of the underlying technology; 2) open scientific inquiry; 3) individual control system. Grounded by the constructivist learning paradigm [29] and the ICAP framework [19], this study aims to provide a comprehensive understanding of the usefulness of integrating the IPA in the development of scientific inquiry skills in physics learning.
The following two research questions were formulated to guide the investigation of this study: a) RQ1: Does using the IPA help students develop their scientific inquiry skills? b) RQ2: How does using the IPA influence students' physics learning experiences?

II. RESEARCH METHODOLOGY
A quasi-experimental research design was used to answer the research questions, with two field quasi-experiments (RQ1) supplemented by post-experimental focus group discussions (RQ2). According to Fraenkel et al. [30], quasi-experimental design suits favorably for educational studies due to the impracticality of random group assignment in this study field. Since this study involved two different schools, pre-experimental tests were conducted to review similarities across all classes based on several predetermined characteristics, namely gender, age, initial scientific inquiry skills, and prior experience with an IPA. This way allowed the researchers to randomly assign one class as the experimental group experiencing the IPA integration program and the other class as the control group undergoing the traditional teaching method. Ethical review and approval were waived for this study, as this study involves no more than minimal risk to subjects.

A. Research Setting
This study was conducted at two upper secondary schools in one of Malaysia's northern states. It took place during the first semester of 2020. The experimental and control groups were selected from two schools (School A: science boarding school; School B: vocational boarding school). The major difference between a science boarding school and a vocational boarding school is the curriculum. Students from the science boarding school complete both the Cambridge Secondary Curriculum and the Malaysia National Secondary School Curriculum concurrently as part of a dual certification program. While students from the vocational boarding school primarily undertake the Malaysia National Secondary School Curriculum. The decision to implement the program in this setting offered two major benefits: 1) the experimental and the control groups in each school were relatively comparable. Both groups were taught by the same physics teacher and used similar learning materials, making it ideal to conduct a field quasi-experimental design. The teachers at both schools had at least 10 years of teaching experience in the subject; and 2) the different schooling types offered an opportunity for this study to make a critical contextual comparison, broaden the scope of the investigation and strengthen the overall results. Moreover, both schools have tight policies against students using smartphones or other gadgets. As a result, both schools were part of an educational scheme that provided students with iPad tablets for use in the library only.

B. Sampling Approach
A convenience sampling approach was employed to select two secondary schools that were willing to participate and contribute  I  STUDENT CHARACTERISTICS   TABLE II  PRE-EXPERIMENTAL TEST RESULTS to this study. Each school had two classes that were randomly assigned to either experimental or control groups. All groups consisted of active students studying physics following the national secondary school physics curriculum. The pre-experimental test results revealed similar characteristics in terms of gender, age, initial scientific inquiry skills, and prior experience with an IPA. Table I outlines the student characteristics, namely gender, age, and prior experience with an IPA. The distribution of males and females was presented. The average age of each group was reported. Prior IPA experience was determined by the average number of years the students had used an IPA, specifically Siri. Table II describes the mean scores for students' initial scientific inquiry skills based on three different tasks, namely Task 1: scientific knowledge, Task 2: scientific content, and Task 3: science processes. Further details on this measurement are provided in the subsequent section. However, it is important to note that the mean and gain scores provided for pre-and posttests after this point are scaled to 0-100 to aid in better interpretation for the reader.
1) School A: The experimental group in School A consisted of n = 8 males and n = 14 females, with an average age of 16.8 years. While control group in this school had n = 12 males and n = 11 females, with an average age of 17.2 years. The experimental group recorded 3.2 years of experience with an IPA, whereas the control group recorded 3.1 years of experience with an IPA. The level of statistical significance was set as pvalue less than 0.05. The analysis of variance (ANOVA) tests showed no significant differences between groups in terms of gender (p = 0.670), age (p = 0.790) and prior experience with an IPA (p = 0.960). The mean scores and standard deviation showed small differences between the experimental and control groups, indicating that both groups possessed a similar level of scientific inquiry skills. The experimental group was later involved in the post-experimental focus group discussion.
2) School B: The experimental group in School B consisted of n = 11 males and n = 11 females, with an average age of 17.2 years. Meanwhile, the control group had n = 11 males and n = 12 females, with an average age of 16.9 years. The experimental group recorded 2.8 years of experience with an IPA, whereas the control group recorded 2.4 years of experience with an IPA. The level of statistical significance was set as p-value less than 0.05. The ANOVA tests showed no significant differences between groups in terms of gender (p = 0.167), age (p = 0.691), and prior experience with an IPA (p = 0.932). The mean scores and standard deviation indicated small differences between the experimental and control groups as both groups possessed a similar level of initial scientific inquiry skills. This experimental group was also later involved in the post-experimental focus group discussion.

C. Design Phase
The IPA integration program was codeveloped by the researchers and the two respective physics teachers. A physics project assignment was designed following the electromagnetism module derived from the national curriculum and the National Science Teacher Association's [4] principles of scientific inquiry tasks. The project was built to incorporate and stimulate scientific inquiry skills such as identifying a problem effectively, formulating a hypothesis, designing an appropriate experimental procedure, conducting a scientific experiment in a systematic manner, gathering, organizing and analyzing data, applying statistical methods to support conclusions, and reporting findings. The individual students were specifically instructed to plan, perform, collect, analyze, and write a report about their findings. They were required to complete the project assignment over the course of four weeks.
The experimental groups used Apple's Siri to help them with the assignment. The use of Siri was mainly due to two main reasons, namely: 1) the iPad tablets were readily available at both schools to be used by the experimental groups; and 2) Siri offered one of the most developed, yet friendly and easy-to-use software that employed a variety of advanced as well as sophisticated machine learning technologies to understand a command before providing appropriate feedback using natural language. The students in the experimental groups were required to use Siri in the school library. The students requested internet access codes from the librarian before using the IPA. More specifically, in addition to using Siri as a Google Search engine, the students frequently utilized it to define words, translate them into their mother tongues, conduct mathematical calculations or metric conversions, and display a model of the maglev train. The students were also required to record their questions and Siri's responses in their journal writing reports. The sessions ended after 3 h, and the students returned the iPads to their respective librarians.
On the other hand, the control groups received paper-based learning materials related to the assignment and verbal guidance from their physics teachers. The printed reading materials could only be used in the libraries and the students were not permitted to take the materials outside of the libraries. Additionally, the students did not have access to the internet. The students were also required to record their library sessions and consultation sessions with their teachers in their journal writing reports. Table III shows the details regarding the physics project assignment.

D. Implementation Phase
Before the IPA integration program, the researchers organized a meeting session with the physics teachers to explain the study and to ensure that all groups were taught following the designated plan. All groups were given a 30-minute introduction on electromagnetism and a short brief regarding the project assignment that they would be working on for the next four weeks. More specifically, the students researched and applied several physics concepts related to electromagnetism, such as magnetic force, magnetic field, direction or magnitude of force, and magnetic levitation to complete the project assignment. The groups began their project assignment in week 1, and they were allocated 45 min per week. At the end of week 4, all groups were required to submit their completed project assignments and answer the post-experimental test. Both experimental groups were later invited for post-experimental focus group discussions.
Figs. 1 and 2 indicate a completed model from one of the students. She chose a car over a train because the former was lighter. This improved its stability and capacity to withstand the horizontal and vertical motion. Based on Fig. 1, she successfully demonstrated the concept of magnetic levitation. The model was able to levitate due to the repulsive forces generated by the magnets aligned on the track and the one attached below it. In this case, the magnetic force eliminated friction between the model and the track. Furthermore, the rulers were positioned vertically at that height to avoid the model from being forced off the track. Based on Fig. 2, the magnets were arranged close to each other at the beginning of the track to generate sufficient magnetic force for the model to levitate and the subsequent magnets were spaced apart at a similar distance to maintain a balanced, continuous movement. Since the model was not powered by any source, the model only began to move after being pushed and remained in movement obeying Newton's First Law of Motion.

E. Data Gathering and Analysis Approach 1) Quantitative Data:
Pre-experimental test that was applied before the introduction session was used to measure the initial scores of students' scientific inquiry skills before the IPA integration program. Post-experimental test was applied after the students submitted their project assignments to measure the students' scientific inquiry skills after they completed the program. Both pre-and post-experimental tests contained the same questions which belonged to three different tasks, namely Task 1 assessed students' scientific knowledge, Task 2 evaluated students' application of scientific content, and Task 3 measured students' execution of science processes. Each task had six openended questions that were scored on a 5-point Likert scale. For example, questions in Task 1 covered topics related to students' understanding of electromagnetism, electrostatics, electromagnetic fields, and magnetism (e.g., what is meant by a magnetic field?). While questions in Task 2 focused on the application of evidence-based knowledge related to electromagnetism across a broad range of daily activities (e.g., a single bar magnet is placed horizontally on a table. Please draw the magnetic field lines that most closely represent the actual magnetic field configuration). Finally, questions in Task 3 encompassed various situation-based procedures related to the basic and integrated scientific skills (e.g., you have one magnet with the north and south poles labeled. How can you use this magnet to identify the north and south poles of other magnets?). The minimum score for each task was 6 and the maximum score was 30. Mean scores and standard deviations for both experimental and control groups were calculated. The students were given 30 min to answer all questions in both pre-and posttests. The duration was considered to be sufficient for the students to respond to all of the questions based on the results from a pilot study. Following the scientific inquiry rating framework proposed by the NSTA [4], three raters comprising one research member and the respective physics teachers were appointed to independently examine the pre-and post-experimental tests results. This framework allowed the raters to evaluate and give a score to each scientific inquiry task. The final rating scores were derived from the average score from each individual rater appraisal. The gain scores arose from the difference between results from postand pre-experimental tests. Light's Kappa for multiple raters was used to check the interrater agreement on the scores for all groups. The strength of agreement was interpreted based on Cohen's [31] recommendation where a Kappa value of 0.61-0.80 indicated good or substantial agreement and a Kappa value of >0.80 showed very good or almost absolute agreement. Table IV illustrates the agreement scores for Quasi-experiment 1: School A and Quasi-experiment 2: School B.
One-way analysis of covariance (ANCOVA) was employed to analyze quantitative data obtained from both quasi-experiments. The four parameters or characteristics which were averaged were gender, age, initial scientific inquiry skills, and prior experience with an IPA. In this analysis, the post-experimental test score served as a dependent variable, the treatment group as an independent variable, and the pre-experimental test score was used as a covariate. Cohen's d [32] was used to measure and interpret the effect size, in which d = 0.2 was considered a small effect size, whereas d = 0.5 represented a medium effect size and d = 0.8 indicated a large effect size. The effect size was calculated by dividing the mean difference of posttest scores with pooled standard deviation. Finally, all statistical tests were conducted at a 95% confidence interval using SPSS Version 28.
Several statistical tests were carried out to assess the assumptions required to proceed with ANCOVA. Shapiro-Wilk test was used to check the normality (p = 0.894), Levene's test was performed to determine the homogeneity of variance (p = 0.772), and interaction analysis was conducted to check the homogeneity of regression slopes (p = 0.613). The normality, homogeneity of variance, and homogeneity of regression slope assumptions were met in this analysis. Homogeneity of variance describes the variance of the dependent variable which should be equal across all groups. While homogeneity of regression slopes refers to the absence of interaction between the covariate and the independent variable.
2) Qualitative Data: Qualitative analysis was conducted to investigate the students' experiences of using the IPA to complete the physics assignment project. Post-experimental focus group discussions involving experimental groups from School A and School B were employed to explore the students' perceived learning experiences with the use of the IPA. This method was useful to discover in detail how the IPA affected the students' development of scientific inquiry skills while understanding and applying physics concepts. Generally, the focus group discussions were conducted in two systematic steps. First, the facilitator who was also one of the researchers began outlining the aim and purpose of the discussion. Second, each experimental group was divided into five workgroups and was instructed to critically discuss their views on how they used Siri to complete their project assignment and the benefits or setbacks that they had experienced while using and engaging with Siri. The discussions lasted for 45 min on average and were audio-recorded and manually transcribed. A thematic analysis approach suggested by Ryan and Bernard [33] was used to analyze the transcripts. In this approach, keywords or statements indicating learning aspects associated with the IPA were carefully identified. A rigorous search was done to locate instances of relevant words or phrases and was later categorized into emerging themes. Once all themes emerged, a respondent validation strategy was carried out for theme verification and accuracy. A codebook was also maintained throughout the process to ensure a comprehensive and systematic data analysis.
Moreover, to further examine the experiences and student-IPA interactions, a content analysis of the students' journal writing reports was also undertaken. The analysis of these reports served two purposes, namely: 1) to analyze the impact of student-IPA interactions on the development of scientific inquiry skills; and 2) to supplement the findings from focus group discussions. The derived codes from this analysis were later incorporated into the findings of the thematic analysis for establishing consistency and congruency.
It is important to note that there had been an attempt to compare the written journals of the experimental and control groups. However, there were several reasons why this comparison was discontinued. First, the researchers found insufficiently relevant responses in the journals of the control groups, suggesting that these students might have spent time building models instead of writing journals. Second, given that the main goal of this study was to examine the experiences and student-IPA interactions, it may not be very helpful to review the journals of the control groups. Therefore, it is pertinent to analyze journals from the experimental groups to look into the impact of student-IPA interactions on the development of scientific inquiry skills.

A. Quantitative Results
This section presents quantitative results to answer the first research question; RQ1: Does using the IPA help students develop their scientific inquiry skills? 1) Quasi-Experiment 1: School A: The results of ANCOVA showed a significant relationship between the IPA and scientific inquiry skills: F (2, 42) = 4.527, r 2 adjusted = 0.526, p = 0.050, confidence intervals for mean difference = −0.603 and 5.733, and n = 45. The effect size was moderate, measuring at d = 0.618. Moreover, a comparison between gain scores showed that the experimental group acquired significantly more skills than the control group (p = 0.040). Overall, the IPA integration implied a higher influence and positive effect on the development of students' scientific inquiry skills when compared to the traditional teaching method. Table V shows the means, standard deviations, and gain scores of the pre-and post-experimental test results. Gain scores were derived in this table by subtracting pretest results from posttest results.
2) Quasi-Experiment 2: School B: Assumptions for AN-COVA were tested using the following statistical tests; Shapiro-Wilk test of normality (p = 0.795), Levene's test of homogeneity of variance (p = 0.782) and homogeneity of regression slopes (p = 0.613). The test results indicated that all assumptions were met for ANCOVA.
The results of ANCOVA showed a significant relationship between the use of the IPA and scientific inquiry skills: F (2, 42) = 30.575, r 2 adjusted = 0.53, p <0.001, confidence intervals for mean difference = 3.9273 and 8.2331, and n = 45. A large effect size was measured at d = 1.82. Furthermore, a comparison of the gain scores showed that the experimental group acquired significantly more skills than the control group (p <0.001). Therefore, the IPA integration had a positive effect on developing students' scientific inquiry skills in comparison to the traditional teaching method. Table VI shows the means, standard deviations, and gain scores of the pre-and post-experimental test results.

B. Qualitative Findings
This section presents qualitative findings to answer the second research question; RQ2: How does using the IPA influence students' physics learning experiences? The analysis revealed three emerging themes, namely meaningful engagement, scaffold-alignment, and individualization. Table VII presents the main themes, their relative extracted statements, and the number of appearances of related statements.

1) Theme 1: Meaningful Engagement:
The majority of students reported meaningful engagement when using the IPA as they perceived a sense of familiarity in their interaction with the technology. The students highly recognized the ability of the IPA to replicate human communication and use words describing emotion. The students found the responses that were given by the IPA as positive, spontaneous, and supportive. They specifically highlighted some encouraging words that they had received such as "Good question!," "Clever guess!," and "Nice try!" According to the students, they found these words to be extremely helpful in giving them confidence to successfully complete the project assignment without relying much on their teachers. As a result of the IPA integration, the students were able to study more effectively and engage in a low-stress, nonthreatening physics learning environment. One student reported, "The teacher frequently calls out name during the class, that makes me feel very anxious. Siri helps me learn in a more calming and comfortable manner." Therefore, learning and engaging with the IPA was regarded as pleasant, especially for students who struggled with performance anxiety and social anxiety in a traditional classroom setting.
2) Theme 2: Scaffold-Alignment: A common statement made by the students was the IPA's ability to listen and comprehend their questions leading to desired responses or outcomes. The students commonly asked Siri factual and mathematical problem-solving questions. One student, for instance, inquired as to what a maglev train was. Siri then responded with a detailed explanation of a maglev train that uses polarised magnets and minimal friction. Although the responses made by the IPA were at times not within the assignment context, the students were receptive and found the responses as interesting. When a student asked Siri, "What is the impact force of the maglev train?," Siri first described the electromagnetic forces that were acting on the train and then went into greater detail about how these forces influenced the train's overall safety. Siri's responses increased the students' awareness of the need to balance the effects of these forces to ensure the safety of the train. They also noted that these responses were thought-provoking and further challenged their understanding of the concept. This way taught the students to be more aware of catching useful hints, connecting relevant information, reasoning from facts and making a smart guess to plan for future steps in the project assignment. Accordingly, the students expressed great satisfaction with their learning experiences with the use of the IPA. For instance, one student noted, "I liked Siri because it assists me to progress and advance through the next step very easily." Moreover, some students acknowledged the IPA's distinctive way of providing assistance only when needed. They affirmed their learning preferences to first create their own idea or solution and sought further feedback or comment in later stages. The students were able to gain control of their learning process and be more focused on developing their scientific knowledge and skills. One student asserted, "I have my own time to think that encourage me to create my own understanding first and only ask for help when I need it." 3) Theme 3: Individualization: Several students described individualized learning experiences when using the IPA for completing the assignment. Based on the analysis of the students' journal writing reports, each student in the experimental groups had an average of 15 student-IPA interactions during the 3-h period per day. The student in the control groups recorded an average of two teacher consultations daily. The IPA was beneficial in promoting students' scientific discovery following their own learning pace because the technology attained the ability to respond to the students' individual learning needs. One student stated, "It was interesting to hear how Siri helps my other friends because I thought we all would receive similar responses." Furthermore, the IPA provided immediate direct responses that were not always possible through traditional teaching methods as the teachers were more focused on assisting low-performing students who required more assistance and guidance. One student reported, "I feel that my queries are always entertained by Siri. So, there is no time gap between questioning and receiving answers." It was observable that Siri played a major role in attending and responding immediately to the student's questions upon receiving them.

IV. DISCUSSION
The results and findings from this study indicate an overall positive impact of integrating the IPA on physics students' scientific inquiry skills. The method of advancing physics exploration with the support of the IPA has a significant moderate to large effect on developing scientific inquiry skills when compared to the traditional teaching method. In addition to this, the interaction and application of the IPA during the project assignment results in a distinctive transformation of the student's learning experiences. For instance, the students were more engaged in their learning process to an extent that they took charge and directed their learning by themselves. These results, in fact, correspond to previous studies that reported similar effects with the use of IPAs in their teaching and learning (e.g., [34], [35], [36], and [37]). These studies indicate that using IPAs increases student engagement because they feel more in control of their learning and therefore more responsible for making it meaningful.
Interestingly, the effect size difference between science and vocational boarding school poses a substantial focus of discussion in this study. The former recorded a moderate effect size (d = 0.618), while the latter recorded a large effect size (d = 1.82). Results from students in science boarding schools and students in vocational schools differ for one or more plausible reasons. Science boarding school students are expected to take up a dual-curriculum program, so they devote more time to mastering the theoretical side of their physics subject. With only one national curriculum to follow, students in vocational boarding schools are exposed to more hands-on, practical learning opportunities. This may show us that, while students from vocational schools perform well independently and require less supervision to improve their scientific inquiry skills, students from science boarding schools may need explicit training or guidance. This finding could paint an early image of how subject specialization affects students' acceptance and ability to use an IPA in guiding their own learning, which warrants further studies.
Despite the quantitative data being derived from a relatively small sample that does not permit generalization, the qualitative data reveals an insightful perspective regarding the positive and favorable impact of integrating the IPA on students' scientific inquiry skills. IPAs facilitate meaningful learning engagement by providing human-like communication and replicating a close resemblance of intimate human tutor interaction [36], [37]. The IPA, in this instance, uses natural human language software to offer repeated positive reinforcement in the form of verbal praise. This finding reaffirms Skinner's reinforcement theory within a technologically assisted learning environment whereby positive reinforcement that is offered by the nonhuman tutor positively increases students' learning behavior and motivation [38]. Consequently, this functional aspect promotes self-assurance for the students to learn productively in a safe, low-stress environment. In this context, the learning environment that is fostered by the integration of the IPA in formal teaching and learning arguably feeds into existing cultural learning expectations of these students who prefer learning in a harmonious and nonconfronting condition. Reflecting the importance of aligning students' cultural expectations to achieve desired learning outcomes, the finding may suggest the benefit of IPAs in creating a learning environment that reduces learning-related anxiety.
Another important finding to be discussed is the characteristic approach used by the IPA in guiding students to explore and pursue their own knowledge discovery and skills acquisition. This approach reflects the principles of individualized dynamic scaffolding that emphasizes providing support on demand, since the instructor is required to monitor students' learning progress and provide scaffold only when needed [39]. Guided by the self-determination theory by Deci and Ryan [40], the learner's authority plays a critical role in advancing students to self-direct their learning strategies to achieve their respective learning goals. This finding serves as good evidence of the IPA's role in developing learner agency by allowing the students to learn at their individual rate and providing assistance when necessary. Although the effects of receiving individualized learning support are repeatedly confirmed by many other researchers in the field of personalized learning [40], [41], [42], [43], [44], [45], this study advances a new prospect of understanding the IPA through the lens of self-directed learning.
Within the context of physics education, this study demonstrates that Siri is capable of both providing accurate physics formulas or equations and performing basic calculations. The teaching and learning of classical physics, such as electromagnetism, mechanics, and thermodynamics, largely depends on basic mathematical operations. As a result, these classical branches of physics would benefit greatly from this particular feature of Siri. However, Siri's inability to perform complex high-level calculations is another characteristic-or perhaps limitation-that deserves further attention. Such an additional characteristic is crucial for assisting students especially undergraduates in handling complex mathematical calculations. The ability to compute high-level mathematical operations such as integration and differentiation is useful for application in modern or advanced physics such as mechanics, quantum, particle, or matter physics.
Most importantly, this study delivers two major contributions to educational technology research areas. First and foremost, it contributes to the field of computer tutoring by offering empirical evidence on the impact of integrating IPAs on students' development of scientific inquiry skills. To the best of researchers' knowledge, there is a lack of empirical evidence investigating the role of IPAs in improving scientific inquiry skills, especially within physics education. Second, this study expands the knowledge of technology-mediated learning by suggesting that IPAs integration has the potential to positively impact students' physics learning experiences. IPAs could also be highly effective in today's learning contexts that operate on large student-teacher ratios. This can be accomplished through strong engagement between students and IPAs, which can subsequently contribute to an increase in students' learning motivation [37]. This study also adds to the body of knowledge on scaffolding by demonstrating that scaffolding when employing IPAs only has an effect when it is dynamic and enables meaningful student engagement or interaction. In terms of its practical consequences, this study exemplified how educators may construct and integrate IPAs into an existing learning environment. Our study has shown that Apple's Siri can be a potent tool for helping physics students understand fundamental physics principles, which in turn promotes students to build their scientific inquiry skills.

V. LIMITATIONS AND FUTURE RESEARCH
This study has some limitations. First, a field quasiexperimental research design was conducted to investigate the impact of the IPA on students' scientific inquiry skills. The researchers were aware that pretreatment group variations between the experimental and control groups may potentially influence the study outcomes. Recognizing this risk, the issue was addressed by gathering pre-experimental data and comparing experimental and control groups using ANOVA to check for character similarity. Furthermore, both groups in each school were taught by the same teacher using predetermined plans and strategies. Nevertheless, a future study involving the use of other quantitative analytical methods or tests is suggested to further validate and extend the results of this study.
Second, the sample size used in this study may not be considered as large enough for ANCOVA to be used. Further investigation into subsets of the sample was, therefore, rendered impossible. However, the medium to large effect sizes that were obtained in this study presents empirical evidence on how the use of the IPA can significantly enhance the development of scientific inquiry skills among students. Future studies would be interested to examine or compare the impacts of IPAs integration between high and low achieving students.
Finally, this study conducted the experiment during a short period of time and under favorable conditions, which may result in better results due to the novelty effects and increased focus on details [34]. However, it is worth noting that many students were unfamiliar with the use of IPAs in the context of completing project assignments, which in turn evokes a critical discussion on how far this approach can be employed in the long run. Therefore, conducting longitudinal studies to investigate patterns or long-term changes in students' scientific inquiry skills would add value to the research area.

VI. CONCLUSION
This study provided answers to two research questions. First, it examined whether an IPA could support students in improving their scientific inquiry skills. The quantitative results from two field quasi-experiments showed the use of the IPA over a period of four weeks exerted a positive significant impact on the development of students' scientific inquiry skills. Second, this study examined how students' learning experiences were affected by the integration of the IPA. The qualitative findings revealed a positive and favorable impact of the IPA in providing dynamic scaffolding which offered rather responsive and on-demand support needed by the students. In sum, although this study delivers a positive impact of the use of the IPA in supporting students' scientific inquiry skills, it is imperative to inform that this study does not consider renouncing the roles and positions of teachers. More precisely, this study recommends the use of IPAs within a teacher-guided inquiry learning environment since this way would meet every individual student's learning needs. The results and findings of this study add to the growing body of knowledge about computer tutoring and technology-mediated learning.
Nurfaradilla Mohamad Nasri received the Ph.D. degree in curriculum and pedagogy from the University of Edinburgh, Edinburgh, U.K., in 2017.
She is currently an Associate Professor in curriculum and pedagogy at the Centre of Educational Leadership and Policy, Faculty of Education, Universiti Kebangsaan Malaysia, Bangi, Malaysia. Her main research interests include the development of a culturally responsive curriculum and instruction, teachers' professional development, and self-directed learning.
Nurfarahin Nasri received the M.D. degree in medicine from Universiti Putra Malaysia, Malaysia, in 2019, and the M.Ed. degree in curriculum and pedagogy in 2021 from Universiti Kebangsaan Malaysia, Malaysia, where she is currently working toward the Ph.D. degree in curriculum and pedagogy with Universiti Kebangsaan Malaysia.
Her main research interests are innovative curriculum, responsive pedagogy, and qualitative research method.
Nur Faraliyana Nasri received the B.Sc. degree in physics from the University of Wellington, New Zealand, in 2013, and the M.Ed. degree in curriculum and pedagogy from Universiti Kebangsaan Malaysia, Malaysia, in 2019.
She is currently a physics teacher with Mara Junior Science College, Baling, Malaysia.
Mohamad Asyraf Abd Talib received the B.Ed. degree in teaching English as a second language from Universiti Teknologi Mara, Malaysia, in 2018. He is currently working towards the M.Ed. degree in teaching English as a second language with Universiti Kebangsaan Malaysia, Malaysia.
As a Research Assistant he was involved in a few research works related to STEM and culturally relevant science teaching.