Roger Gilabert Guerrero and Judit Serra

Technology has been increasingly used for educational purposes. Nowadays, schools are equipped with diverse technological devices, such as computers, tablets, digital panels, educative robots… Teachers find themselves struggling to incorporate them in their lessons and seek to take advantage of the alleged technological perks: motivating students, personalisation and learning effectiveness. 

Serious games (games with education purposes rather than just entertaining), have recently began to have an adaptive component, which is the ability of a system to adjust instruction based on learner abilities or preferences to improve the efficiency and efficacy of learning (Oxman & Wong, 2014). They are promising, but we do not have enough evidence yet to answer the question.

Why is the issue of sequencing important and how can technology do something about it?  

Teachers are in charge of building up a syllabus that determines what is to be taught (selection) and the order in which this is done (sequencing). They usually do so by applying increasing difficulty based on curriculum guidelines, teacher training, textbooks, intuition and experience. Thus, students go through all the same learning content in the same order, regardless of their previous acquired knowledge and their personal learning characteristics. As opposed to that, digital adaptivity calculates what every player should be facing next thanks to a complex algorithm that takes into account the learning processes, gameplay performance, as well as game mechanics. 

We found out there was almost no research that compared the way teachers sequence learning and how an algorithm do so. Hence, we ask ourselves if games, with their related algorithm and their potential for personalization, pose a learning advantage over decision-making by teachers. In this case, our research was done in the field of reading acquisition in a foreign language (English). Adaptivity algorithms in games clearly affect sequencing, but little research has been conducted on the issue. 

Luckily, we could put the issue of sequencing to test that with the iRead software, specifically using the app “Navigo: Pyramid of the Lost Words” and the teacher tools. While the students played the different game mechanics, the teacher tool gave teachers the autonomy to make their own choices regarding features and games in case they wished to assign them to each individual student or to the whole class. 

Who participated in your research and how did you organize it?

Among the six Spanish schools that were using the iRead system during the school year 2019-2020, there were two schools that participated in this study. There were 107 students involved, although only 67 of them completed all the tests. We need to remember this happened during the pandemic and conditions kept changing.  

Students were asked to use iRead one hour per week during most of the school year. They were tested at the beginning and at the end of the school year on different skills related to reading in their mother tongue and English. The iRead system also recorded several internal measurements, including the total number of games played, games played per feature, the number of different features, and the number of errors among many  others.

In every school, there were two classes participating in the study. We asked the teacher to determine the learning sequence through the teacher tools in one of the groups and let the system completely choose accordingly with the adaptivity rules in the other group (so there was not any teacher intervention whatsoever).  

The adaptive principles in the iRead system involved content difficulty and competence showed in the games already played. Therefore, features known to be easier for learners would appear earlier in the game. As each feature was mastered, the system unlocked the next more difficult features. Besides, players could move on after playing three games correctly on a feature, while others could keep playing on the same feature until they mastered it. 

What did you find out? Did children learn differently under each sequence? 

With our research, we wanted to know if there was a differential impact in reading skills acquisition when following the algorithmic and the teacher-led learning sequence. After making sure the two schools and the two groups within each school were comparable, we ran different statistical tests and concluded that the type of sequence did not affect gains in reading skills differentially. We expected that learners in the adaptive condition would learn more than in the non-adaptive condition and this did not happen. This was the same results as in other previous studies such as Vanbecelaere et al. in 2019.  

We believe the algorithm might not have been sophisticated enough to make a large difference from the teacher led sequence. Also, participants used iRead for a shorter period than we expected, as we had to stop because of the world pandemic. Also related to the length, playing one hour a week for four months, even if it entailed playing hundreds of games by each student, may simply not have been sufficient to have more effects.

We also asked ourselves if different game features could affect the learning of reading skills. One of our findings was that playing a narrower variety of features resulted in more gains in decoding skills than playing a wider variety of features. 

Furthermore, players in the teacher-led sequence made significantly more errors. This may be attributable to the fact that in the teacher-led sequence the features learners had to deal with different dimensions (e.g., morphology, morphosyntax, syntax) that were coded as more difficult features. On the contrary, players in the algorithmic sequence played a wider variety of features, although most of them were related to the same category, which was grapheme-phoneme correspondence. To sum up, both groups showed a similar amount of gains even if they followed two different paths, one with lots of features but fewer games on each feature (algorithmic feature) and another with fewer features but more games on each feature (teacher led sequence). One positive finding, however, was that lower level learners progressed more than more advanced learners, which brings hope for teachers.

What are some of the things you have learned with this research project?

This research project helped us to expand our knowledge on different features. First of all, interestingly enough, the type of sequence followed in the games did not seem to have an impact (although all learners displayed significant progress in reading skills from the start to the end of the iRead project). They took different routes to learning depending on the sequence followed: learners in the algorithmic sequence playing fewer games on a larger number of features and learners in the teacher-led sequence playing more games on a narrower set of features. Nevertheless, overall, we believe the issue of sequencing remains unresolved and future algorithms should include more learning characteristics.

From these findings, we find great potential of the adaptive component, which could impact on the path of language development in general and reading skills in particular. The possibility of choosing an algorithmic or a teacher-led sequence as teachers did through the teacher tool is without a doubt an innovative aspect of serious games design which allowed for the creation of two sequences. Combining both possibilities might be another possibility that could have great benefits: allowing the algorithm to choose the learners path but monitor it and being able to intervene at some point if needed. Although not the focus of this particular study, we are now also ready to look at what the system can tell us about linguistic feature difficulty. We believe that all these has enormous potential for personalization of L2 learning.

***

The iRead project is dedicated to researching personalised learning technologies that support children struggling with reading. Please read more about the importance of personalised reading technologies here: https://iread-project.eu/about/vision/ The project website also features useful tips and strategies for parents and teachers to support children’s learning journey: https://ireadprojecteu.files.wordpress.com/2020/06/navigo-at-home-event-session-2-leaflet-1.pdf  The research above has been written into an academic paper collaboratively by iRead University of Barcelona team members Professor Roger Gilabert Guerrero and Dr Judit Serra and is currently under review.