Reciprocity Effect between Cognitive Style and Mixed Learning Method on Computer Programming Skill

: Many universities undertake mixed learning to meet the required needs. Mixed learning is a blend of F2F classroom education and online learning education. The strength of mixed learning is that it supports student cognitive styles more than non-mixed learning. The right mix of mixed learning provides more constructive and conducive learning. Meanwhile, the programming language is the primary skill that students must master to create computer application programs. The question is: Is there an effect on student cognitive style and learning methods on mixed material 30% F2F and 70% asynchronous online and on the contrary mixture on student programming skills? Therefore, this study aims to determine the effect of reciprocal interaction between cognitive styles and mixed learning methods on programming skill achievement. This research method is experimental research. The study found that: Although there is no difference in the achievement of student learning skills based on tests on mixed learning methods, further test on student cognitive styles found that there are differences in the achievement of student learning skills in mixed learning methods; students with auditory and visual cognitive style who learn with mixed learning-2 have better programming skill achievement than students with auditory cognitive style who learn with mixed learning-2; students with kinesthetic and visual cognitive styles who learn with mixed learning-2 have superior programming skill achievement compared to students with kinesthetic cognitive styles who learn with mixed learning-1. The research novelty is: There has been no previous research on the reciprocal effect of cognitive styles and mixed learning methods with a mixture of 30% F2F and 70% online and vice versa.


Introduction
The rapid advancement of Information and Communication Technology (ICT) makes it easier to realize multimedia in learning to support student cognitive styles. Besides that, the use of ICT also has a positive impact on the learning process as well as realizing learning efficiency (Aljuboori et al., 2020). However, the rapid development of ICT has increased pressure for universities to include greater use of technology and innovation in the curriculum (Tyler-Wood et al., 2018). Therefore, it is not surprising, if many tertiary institutions adopted mixed learning approaches as a solution (Nazarenko, 2015). Mixed learning is a perfect blend of F2F classroom education and asynchronous online learning education (Pierce, 2017).
Mixed learning research has long been a concern of researchers and lately, it has become an important research topic because it has a combined advantage of learning in the classroom and outside the classroom. Unfortunately, research on mixed learning skills is still limited (Nazarenko, 2015). The benefits of mixed learning are the best approach to learning strategy by taking the strengths of Face-To-Face (F2F) and online learning (Sleator, 2010). One of the strengths of mixed education is that it provides a conducive learning environment for students and supports a variety of student cognitive styles (Pierce, 2017). That makes sense that prior research indicated, mixed learning is superior to non-mixed learning in learning achievement (Van Niekerk and Webb, 2016). Mixed learning can maximize student learning outcomes by applying appropriate technology learning to fit student cognitive style in transferring skills correctly and at the right time (Lieser and Taff, 2013).
The right combination of mixed learning provides social support and constructive learning for students (Opina et al., 2011) and creates conducive interactions in learning activities (Pierce, 2017). However, the question is:  How is student learning skill achievement based on student cognitive style?  How is the student' skill achievement from two different mixed learning and is there a reciprocal influence between student cognitive style and mixed learning methods?  What will the results be if there is a mutual influence between cognitive styles and mixed learning methods?
Therefore, this study aims to determine the effect of reciprocal interaction between cognitive styles and mixed learning methods on learning skill achievement.
The mixed learning method combines education between F2F classroom education and online learning education (Hogarth and Biggam, 2009;Almpanis et al., 2010). F2F class education strength is the high intensity of interaction between students and lecturers in facilitating cooperative learning and clarity of lecture materials (Agosto et al., 2013). The F2F classroom education offers real and meaningful interactions between students and teachers, while pure online education cannot replace it (Tang and Chaw, 2013). The problem is, the F2F classroom education requires higher tuition fees, especially in wellknown tertiary institutions (Norman, 2016). Online learning is a significant part of university education to support conventional F2F learning (Seta et al., 2018). Communication that occurs in online learning is synchronous and asynchronous (Clark and Barbour, 2015).
In synchronous online, education is delivered remotely in real-time by the teaching lecturer to students (Alammary, 2019;Anggrawan and Satria, 2020). In contrast, learning material in asynchronous online education is given indirectly to the student (Anggrawan and Satria, 2020). In asynchronous online, students can access material or modules stored on the server computer anytime and anywhere through computers connected via the internet to specific web addresses (Anggrawan and Satria, 2020). Asynchronous online learning constitutes independent learning for students (Anggrawan and Satria, 2020) or collaborative learning (Alammary, 2019) by some students who agree to study. Students and lecturers agree that one of the main weaknesses of online learning is the lack of F2F interaction (Król, 2016). Meanwhile, asynchronous online education strength takes advantage of various multimedia forms: Text, audio, visual still and moving and other forms for learning purposes supporting student cognitive styles (Clark and Barbour, 2015).
Every student has a cognitive style that reflects a way of learning that is preferred and easier for students to understand. Students with high specific cognitive styles have more significant difficulties acquiring knowledge than weaker cognitive styles (Psycharis et al., 2014). There are three types of cognitive styles: Visual, auditory and kinesthetic (Rhouma, 2016). Students with visual cognitive styles prefer and easily understand the lessons presented in writing, pictures, graphs and tables (Rhouma, 2016;Anggrawan et al., 2019). In other words, students with visual cognitive styles in learning rely on their sense of sight. Auditory students prefer the lesson presentation in voice or lecture form (Anggrawan et al., 2019). So, in other words, students with auditory cognitive style rely on the sense of hearing. Kinesthetic students prefer to learn in interactive information media and special situations (Anggrawan et al., 2019). In essence, learning should support student cognitive styles to improve performance and success learning (Eudoxie, 2011). So it makes sense and inevitable, if later, more and more universities have created systems and organized innovative mixed education to accommodate student cognitive styles and the learning needs of 21stcentury students (Lieser and Taff, 2013).
Substantially, mixed learning provides better effectiveness than education that relies only on the conventional F2F education method (Van Niekerk and Webb, 2016). Besides, students respond positively and easily adapt to mixed learning (Anggrawan et al., 2019); actually, mixed learning brings excellent opportunities for students to master the subject matter and achieve success in education (Lieser and Taff, 2013). Moreover, mixed learning patterns provide a more conducive learning environment and increase student learning achievement (Bazelais and Doleck, 2018). In fact, despite online learning education or conventional F2F classroom education has disadvantages, but mixed learning can overcome it as long as mixed learning is mixed with the right mix (Opina et al., 2011). According to Kanuka and Rourke (2013), some experts argue that the portion of the online learning mix in mixed learning is 25 to 50%; meanwhile, other experts determine the amount of the online learning mix in mixed learning is between 30 and 70% (Kanuka and Rourke, 2013). In short, there is no certainty about the portion of the online learning mix in mixed learning. Thus, the mixed learning main obstacle lies in the accuracy of the mixture; of course, the right mix in one subject does not mean it is suitable for another course (Anggrawan et al., 2019). Likewise, although certain courses have produced evidence of satisfactory learning success with mixed learning, this does not mean that the same conditions apply to other courses. So, the mixture accuracy and the suitability of mixed learning in each subject must be scrutinized scientifically. Thus, in essence, scientific research on mixed learning is necessary to determine how well the student learning success due to the mutual influence between cognitive styles and mixed learning methods. In other words, if there is a mutual influence between cognitive styles and mixed learning methods, it is necessary to know what the results are?
In short, in connection with this research, the formulation of the research problem explicitly is: (a) What learning methods are better than the two kinds of mixed teaching methods; is the mixture of 70% learning in class and 30% online learning outside the classroom or 30% learning in class and 70% online learning outside the classroom? (b) Is there a reciprocal influence between student cognitive styles and mixed learning methods applied in teaching? (c) If there is a reciprocal influence, how will it affect the achievement of student programming skills?
Programming language courses are essential for students in mastering programming skills (Yindi, 2016;Anggrawan, 2018). The programming skill is the ability of students to create computer applications with programming languages. The visual programming language is event-based programming (Yindi, 2016;Anggrawan, 2018). In other words, program code is made based on specific events or functions so that the sequence of program execution is also an event. In contrast to structured or procedural programming, it will execute program code from the beginning to the end of the program sequentially (Anggrawan, 2018). The visual programming language that is most popular today is Visual Basic.Net or VB.Net (Anggrawan, 2018). VB.Net is an object-oriented programming language (Yindi, 2016;Othman et al., 2018). VB. Net has the advantage of being of the visual programming language and high interest from students who learn it (Zhang et al., 2011). VB.NET is useful for applying network interfaces and operate records databases (Othman et al., 2018).

Related Work
This subsection provides a brief literature review of some of the most recent scientific articles relating to cognitive styles, programming and mixed learning: between cognitive style and gender on mixed learning in Algorithm and programming lesson. This study has a limitation that only investigates mixed learning by mixing 40% F2F material and 60% online material. This means that this earlier study was not a reciprocity effect study; that is, it did not investigate mixed learning with the opposite mixture of mixing 40% F2F material and 60% online material. This earlier study found differences in learning outcomes between students who had different learning styles. Male gender students were more successful than students with the female gender, using the experimental research method  Alammary (2019) conducted an assessment of the comparison of programming learning experiences between conventional treatments and mixed care methods. This study concluded that mixed learning is more effective in constructing traditional education to improve student learning experiences. This study also confirms that there is an increasing trend in the application of mixed learning programming lessons. This previous researcher also warned that there was still little research related to programming education and mixed learning methods Literature review of the relative work as mentioned above: (a). Did not examine the comparison of the learning achievement of two mixed learning with the opposite percentage of teaching material mix between F2F and online learning materials; (b). did not examine mixed learning with a blend of 70% learning in class and 30% learning online outside of the classroom and 30% learning in class and 70% learning online outside the classroom; (c). did not research experimental methods on student learning achievement in two mixed learning associated with student cognitive styles and learning methods.
In essence, the authors in this study conducted research that no one has examined, namely the effect of back and forth between mixed learning methods and cognitive styles on computer programming education. Besides, the authors conducted this mixed learning research with mixed teaching materials divided into 70% online and 30% F2F mix and vice versa, which so far, no one has researched.

Research Methodology
This study is experimental research. Two different classes get a mixed learning treatment of VB.Net computer programming lesson materials with a mixture of different portions between F2F learning and asynchronous online learning. The combination tested was 70% versus 30% between classroom learning and asynchronous online learning in the first class and vice versa, 30% mixture versus 70% between classroom learning and asynchronous online learning in the second class.

Learning Treatment
Two classes received learning VB.Net Programming courses. Two treatment classes resulted from the random selection from 5 classes in the Computer Science study program at Bumigora University. The number of students in each treatment class consists of 50 students in the first semester of the academic year of 2019/2020.The first mixed learning class (mixed learning-1) got treatment by combining around 30% F2F learning and about 70% asynchronous online learning. Meanwhile, the second mixed learning class (mixed learning-2) got treatment by combining around 70% F2F learning and about 30% asynchronous online learning.
Students acquire VB.Net programming skills through F2F mixed learning materials in class and online learning materials. Students can learn online lessons independently in teaching material modules (or asynchronous online forms) prepared on a computer server. Besides, students can access them anywhere and anytime via the internet and study according to student needs and speed.
This research is an experimental study with two factors. The first factor is mixed learning with two levels and the second factor is the cognitive style with three levels. Thus, this research is an experimental study with a 23 factorial design.
Mixed Learning-1 (ML1) and Mixed Learning-2 (ML2) are two learning class groups prepared to realize this experimental research. Table 1 shows the model construction methodology of a 23 factorial design. So, the reciprocal effect examined in this study is the combined effect of two independent variables (two factors) on mixed learning and cognitive styles in influencing the dependent variable on learning achievement. Figure 1 shows the diagram of the 23 factorial design model.
The number of classroom learning meetings of the mixed learning-1 and mixed learning-2 takes 7-time meetings (does not include the Exam). F2F mixed learning-1 activities occurred at different times, days, classes and buildings to prevent the threat of spreading external validity. Mixed learning-1 and mixed learning-2 learning activities are as follows: Lecturers provide F2F class lessons, which are structured modules of F2F learning materials that have been classified materials and materials according to F2F learning schedules. Likewise, online learning with structured material has been prepared on a computer server that can be accessed by students (anytime and anywhere) with the asynchronous (or shared) independent learning method via the website.

Data Collecting
The data collected in this study are data on the learning achievement and cognitive style of each student. The instrument in assessing student skill attainment at the end of the lesson is in essay questions. The test instrument used to evaluate student learning achievement has passed the reliability and validity test before being used in the experimental class in this study. Student cognitive style data were collected using a questionnaire conducted in a mixed learning class. The student cognitive style questionnaire instrument uses standard Visual, Auditory, Reading/Writing, Kinesthetic (VARK) instruments that have been tested for reliability and validity.

Research Methods
Learning skill achievement data in this study is ratio data. Due to in this study two classes are treated, then this research method is experimental research, but based on the type and analytical data, this research method is inferential quantitative parametric research, Testing for normality and homogeneity of data and the instrument validity and reliability was carried out using Shapiro-Wilk, Levene, Pearson Correlation and Cronbach's Alpha. A two-way Anova test was conducted to ascertain a reciprocal influence between student cognitive styles and learning methods; differences in learning achievement due to differences in student cognitive styles and differences in learning achievement between mixed learning methods with 70% learning in class and 30% via online compared to mixed learning methods with 30% learning in class and 70% via online. The Tukey post-hoc test is conducted to analyze the reciprocal effects that occur between cognitive style and learning methods.
Threats to internal validity in this study were overcome by means of students with the same background, namely fresh graduates from high school, meaning that students in this study have equal initial cognitive competence in computer programming, thus can overcome the threat of internal validity in the form of death/friction. Control group in the form of classroom lessons as part of mixed learning prevents this research from threatening history internal validity. The instrument used was a standard instrument or tested instrument of validity and reliability to free from instrument internal validity threat. In overcoming external validity threats in this study, other lecturers (not researchers) carried out the teaching process, thus preventing bias or the deliberate or carelessness of researchers in influencing student achievement. Mixed learning in new students was a new method for students; besides that, students were not aware of the research, thereby preventing the threat of external validity of reactive influence and treatment diffusion. In this study, students only got one experimental treatment so that interplay did not occur before and after treatment, thereby avoiding the threat of multiple treatment disorders.

Research Results and Discussion
The survey results using the VARK instrument show that students who have visual cognitive styles are 35 students, auditory cognitive styles are 37 students and kinesthetic cognitive styles are 28 students, as shown in Table 2.
Table 2 also shows that in the mixed learning-1 class, the number of students is 50. As many as 20 students have a visual cognitive style, 17 students have an auditory cognitive style and 13 students have a kinesthetic cognitive style. In the mixed learning-2 class, the total number of students is 50 students. Fifteen students have a visual cognitive style, 20 students have an auditory cognitive style and 15 students have a kinesthetic cognitive style.
A research instrument becomes a useful measuring tool if the instrument measures appropriately (or validity) and can be trusted (or reliable). Therefore, the instrument used in this study must meet the validity and reliability requirements of the test.
The Pearson correlation coefficients of the learning achievement instrument (Question-1, Question-2) using Product Moment were 0.799 and 0.917 (Table 3).
Meanwhile, according to Sugiyono (2004), the minimum requirement to be a correct (valid) instrument using Pearson Moment correlation (or Product Moment correlation) is if the correlation between items with a total score is greater than or equal to 0.3. It means that the instrument to measure learning achievement in this research has high validity.
The instrument reliability test to measure learning achievement using Cronbach's Alpha in this research was 0.677 (Table 4). It is indicating that the internal consistency of the instrument was good. The research instrument has reliable internal consistency if the reliability coefficient is equal to or greater than 0.6 (Siregar, 2014).    The requirement for conducting a two-way Anova parametric statistical test is that the data used must have the same or homogeneous variants (Siregar, 2014). Therefore, this study uses the Levene test to determine whether the learning outcome data is homogeneous or not. Likewise, the data is must normal distribution as a condition to perform a two-way Anova parametric statistical test (Siregar, 2014). Therefore, this study uses Shapiro-Wilk to test the normality of the data.
The Levene test significance value on the test result score (0.060) is higher than the alpha value of 0.05 (Table 5); this indicates that the data variance is homogeneous.
The normality test of learning outcomes data with Shapiro-Wilk shows the significant value of Mixed Learning-1 (ML1) is 0.65 and Mixed Learning-2 (ML2) is 0.68 (Table 6). The significance value for the two mixed learning test scores is greater than the alpha value (0.05), so the learning achievement for both mixed learning is normally distributed.
Using a two-way Anova is to determine how the combination of two independent variables affects the dependent variable (Montgomery, 2012).
Based on the two-way Anova test, there is an interaction between cognitive styles and teaching methods (the significant value of VAK*ML1ML2 is 0.00), which is smaller than the alpha value (0.05) shown in Table 7.
Thus, the cognitive style and teaching methods influence each other in programming learning.
The two-way Anova test showed that the significant value of the difference in learning achievement between students who received mixed learning-1 and mixed learning-2 was greater than the alpha value (ML1ML2 significant value 0.108). Thus, the conclusion is that there is no difference in learning achievement between teaching done with mixed learning-1 and mixed learning-2. There are differences in learning achievement between students with different cognitive styles in mixed learning-1 and mixed learning-2. In this case, the significance level of student cognitive style (VAK) on the two-way Anova test (0.002) is smaller than the alpha value (0.05), which means that there are differences in learning achievement between students who have different cognitive style. So, the conclusion is, even though the two teaching methods show no different student learning achievement, this does not mean that there is no difference in learning achievement based on student cognitive styles.
This finding is the novelty found in this study, that although the accomplishment of learning skills for both teaching methods is equally good, it does not mean that the learning method is suitable for all students. However, it turns out that students with specific cognitive styles may not be ideal for that teaching method. The implication is that students can achieve maximum learning success; the way is, learning methods facilitate learning media that support student cognitive styles.
Tukey post-hoc test is a follow-up test when an interaction occurs in the two-way Anova test (Montgomery, 2012). Tukey post-hoc test can tell precisely where the differences between the two independent variables affect the dependent variable. Therefore, this study conducted Tukey post-hoc test.      Based on the results of Tukey post-hoc test (Table 8), in mixed learning-1, the results of the Tukey post-hoc test show: (a). The learning achievement of students who have auditory cognitive styles is better than students who have kinesthetic and visual cognitive styles; (b). The learning achievement of students who have kinesthetic cognitive styles is not different from students who have visual cognitive styles.
Meanwhile, the comparison of learning achievement on the Tukey post-hoc test between students taught with mixed learning-1 and mixed learning-2 shows: (a) Mixed learning-1 students who have auditory cognitive styles have better skill achievement than mixed learning-2 students who have auditory cognitive styles (b) Mixed learning-1 students who have auditory cognitive styles have no different skill achievement compared to mixed learning-2 students who have kinesthetic and visual cognitive styles (c) Mixed learning-1 students who have kinesthetic cognitive styles have worse skill achievement than mixed learning-2 students who have visual and kinesthetic cognitive styles (d) Mixed learning-1 students with kinesthetic cognitive styles have no different skill achievement than mixed learning-2 students with auditory cognitive styles (e) Mixed learning-1 students who have visual cognitive styles have better skill achievement than mixed learning-2 students who have auditory cognitive styles (f) Mixed learning-1 students with visual cognitive styles have no different skill achievement than mixed learning-2 students who have kinesthetic cognitive styles (g) Mixed learning-1 students with visual cognitive styles have worse skill achievement than mixed learning-2 students with visual cognitive styles The results of this study have also revealed which learning styles are superior in the achievement of student computer programming learning in mixed learning-1 and mixed learning-2. This finding is the strength of the contribution of this study when compared to previous related works.

Conclusion
The two-way Anova test concluded that: (a) There was no difference in the programming skills achieved between students who learn with mixed learning with a mixture of around 30% F2F learning and about 70% asynchronous online learning and students who learn with mixed learning with a blend of approximately 70% F2F learning and about 30% asynchronous online learning (b) There are differences in programming skills acquired between students who have different cognitive styles both in mixed learning with a mixture of about 30% F2F learning and about 70% asynchronous online learning and also in mixed learning with a combination of about 70% F2F learning and about 30% asynchronous online learning The results with Tukey post-hoc test concluded that: (a) Students with auditory and visual cognitive styles who learn with mixed learning with a mixture of around 30% F2F learning and about 70% asynchronous online learning have better programming skills achievement than students with auditory cognitive styles who study with mixed learning with a blend of approximately 70% F2F learning and about 30% asynchronous online learning (b) Students with kinesthetic and visual cognitive styles who learn with mixed learning with a blend of approximately 70% F2F learning and about 30% asynchronous online learning have programming skills that are superior to students with kinesthetic cognitive styles who learn with mixed learning with a mixture of around 30% F2F learning and about 70% asynchronous online learning This means although the student programming skill achievement of the two mixed learning methods when assessed based on the learning method is equally good, it happens that student programming skill achievement of two mixed learning methods differs when evaluated based on the student cognitive styles.
Besides, this study also found that: In mixed learning with a mixture of around 30% F2F learning and about 70% asynchronous online learning, students who have an auditory cognitive style have superior programming skill achievement than students who have a kinesthetic and visual cognitive style. Meanwhile, students who have kinesthetic and visual cognitive styles do not differ in their programming skill achievement in mixed learning with a mixture of around 30% F2F learning and about 70% asynchronous online learning. In mixed learning with a blend of around 70% F2F learning and about 30% asynchronous online learning, students with an auditory cognitive style have worse programming skill achievement than students with a visual and kinesthetic cognitive style. Meanwhile, students who have kinesthetic and visual cognitive styles do not differ in their programming skill achievement in mixed learning with a blend of around 70% F2F learning and about 30% asynchronous online learning.
This research novelty is to study the reciprocal effects of student cognitive styles and hybrid learning with a mixture of 30% F2F subject matter combined with 70% asynchronous online subject matter and vice versa that no one had researched before.
Other new things obtained from this research are: (a) The comparative test to determine which learning method is superior for the achievement of learning skills is not sufficient only with a comparative test based on the learning method but also based on the student cognitive style (b) This study finding can be the beginning of a breakthrough in teaching with certain learning methods based on groups of students with the same cognitive style to achieve better skills. Or in other words, maybe the division of teaching classes no longer contains various cognitive styles with specific teaching methods for better learning achievement This research result implies that each learning method must significantly facilitate learning media that support student cognitive styles. Another implication is the need to test various learning methods to determine which learning method is most suitable for each cognitive style for superior learning outcomes.
Further research will investigate what percentage of cognitive styles affect learning outcomes and compare student achievement between students with strong and weak cognitive styles by grouping or not grouping classes based on student cognitive styles. Thus, further research complements or refines existing research to make it more realistic about how much influence each one has cognitive style towards learning outcomes in mixed learning methods, including strong and weak cognitive styles.
Besides, this study has limitations in the learning factors studied, which only involve cognitive styles and mixed learning method with a combination of mixing 30% versus 70% between F2F subject matter and online subject matter and mixing vice versa, with a mixing variety of 70% versus 30% between F2F subject matter and online subject matter. So, further research needs to involve other internal student factors such as student interest in learning subjects and learning motivation. Likewise, it is necessary to do further research by applying external factors of mixed learning methods with various other variations (other than 30 and 70% or other than 70 and 30%) between F2F and online learning materials and also with other learning methods such as flipped learning and collaborative learning.