Research report

December 2004
Mary Ulicsak, Futurelab

The full version of this report is available to download in pdf format - see box below. On this page you'll find the paper's executive summary.

Executive summary

Astroversity is a game for students aged 12 to 14 played in groups of three. It is designed to give the students an opportunity to develop collaborative and scientific enquiry skills. It does not teach these skills, but the task is structured so that an individual cannot complete it in one attempt nor is it easy to solve by memory alone. The most effective solution requires three students to work together, systematically recording data and coming to a shared solution.

The prototype development process

The project was developed iteratively. The International Centre for Digital Content (ICDC) and Futurelab developed a puzzle that was trialled by students on paper. Once it was agreed that the task was of suitable difficulty and could be made engaging ICDC created a two-dimensional version in Director. In an iterative process this was tested and revised with observations and feedback from the students used to refine the text and format. These findings were used to create a three-dimensional version, which again was iteratively revised. Finally, this was tested within a classroom context over a period of three weeks.

The details of the various stages can be found in Table 1.

In addition to these formal trials Futurelab sent comments on the Astroversity software to ICDC after each release.

Non-technical prototype description

The final prototype is a game for three 13-15 year-old students playing on a computer network. In the scenario they are students at the Astroversity, an orbiting space academy. They are currently undergoing search and rescue training, ie learning how to control a probe, infer information from sensors, and plot routes for a rescue vehicle. During this exercise an alien vessel crashes into the Astroversity releasing three toxic substances into the atmosphere. Bloppo causes brain swelling and eventual explosion. Moob increases the heart rate leading to a cardiac arrest. In contrast, inhaling gunk causes liquids to collect in the lungs leading to death by drowning. The robot headmaster instructs them to form groups of three and rescue their peers on various levels of the academy.

The most efficient way to rescue the casualties is to work together to identify a strategy for recording data, record this accurately, and in the later levels divide the sensors among their individual probes to cope with multiple toxins. Communication skills are enhanced as they must explain their findings, route suggestions and theories about toxicity levels and the relationships between the toxins to each other. They are aided by their chosen robotic tutors who offer suggestions and question them on strategy, use of data, and their performance as a group.

The groups can repeat the rescues in order to effectively save the casualty.

Key innovations of Astroversity project

The following aspects of the experience are novel in games for learning:

• the use of multiple methods of representation, requiring students to switch between a virtual online world and a paper-based representation which they create as a consequence of exploring this world
• the requirement that team members do not simply work together by fulfilling different roles in the same task, but have to contribute information to a single activity with a collective outcome
• the explicit encouragement of students to self-assess and reflect on the skills being developed within the task.

General findings from Astroversity research

The following are the key findings from the studies:

• students need a structured environment when practicing group skills or performing a task requiring scientific enquiry as these are not instinctive skills
• self-assessment is not performed reliably when not mediated by an external person
• well designed games for learning can be motivating and engaging, shown by more students turning up for the final session with the three-dimensional than the first despite the fact it was end of term.

Key learning findings and recommendations

For researchers

• a task can be structured to encourage reflection but direct written prompting is often ineffective
• non-textual interfaces are important
• the findings are highly dependent on the method of assessment; clear questions need to be set, are we looking at improved technique, engagement, motivation, or use of context?
• if Astroversity was to be a useful research tool then data logging is required to observe changes in strategy, performance and responses to reflection prompts.

For teachers

• Astroversity needs to be supported by a teacher or expert student, written online instructions about the concept and strategies appear to be ignored
• fitting in with broad curriculum aims does not limit the enjoyment of the activity within the classroom
• learning gains may be associated with listening and talking skills more than intended scientific enquiry
• the lack of recording of data means that progress cannot be reviewed and the students will have to be asked to record progress in terms of damage to casualties.

For ICDC

• as found in commercial games providing 'training' before the mission is beneficial
• Astroversity needs to be more portable; despite recommendations for machine specification (which is higher than some schools currently have), the software had difficulty in collecting machine addresses and setting up a network
• although the issue of gas diffusion has been glossed over the system still needs to improve the underlying physics - it is currently possible to make probe go through floors with sufficient velocity
• team selection process requires improvement - currently it is hard to identify what team you belong to once selected
• identifying current sensor selections, for example, what sensor has been chosen during and after the exploration and what it impacts
• the original proposal was to develop Astroversity into a Playstation game, further studies are needed to show that the concept of using physical and virtual maps works outside a classroom environment before this is commenced.

Policy makers

• the importance of developing a cross-curricula tool - Astroversity has a science focus but as observed by an English teacher, it can also be seen as a listening and speaking activity
• provision of a non-text based science activity benefits all students but in particular those with poor reading abilities
• games are not necessarily negative, although little improvement in strategy was noted the engagement and motivation of the students to attend lessons improved.

Industry

• studies show that schools are willing to have non-traditional tasks within a syllabus setting
• games can be overtly educational and still enjoyable within a formal setting
• the possibility of mixing various representations does not detract from enjoyment - but no information exists as to how this would transfer to an informal setting.

Futurelab recommendation for the next stage of Astroversity

As it stands Astroversity requires more puzzles before it becomes a viable commercial game. The puzzle implemented is engaging but limited. It is therefore suggested that minor changes be made to the current version, they are:

• addition of teacher guidelines
• amending narrative so that the robotic tutors are external and therefore cannot perform the task
• amending graph to show likelihood of survival rather than damage.

This is then made widely available so that anyone can download the instructions, paper map grids, and software and play it on their school networks. Information about Astroversity as a research project would also be released and if wished trainee teachers and other researchers could use it to investigate how to support collaboration, scientific enquiry etc in the classroom.

Finally, it would be interesting to do a comparison of the two-dimensional and three-dimensional versions of Astroversity. It appears that task focus occurs earlier in the two-dimensional groups though both are poor at systematically recording data. Groups of students could alternate so each do a different level in a different format. One research focus would be on the type of talk, apparent engagement and strategies in each environment leading to a discussion of the overall favourite format and why. Another research issue is the impact of the shared environment, ie seeing each other in landscape? It appears motivating but the studies indicate it does not lead to an increase of recording data or speed at task although the difference in the current format, you cannot see the others in the two-dimensional version, means further work is needed for a comparison to be made.

top of page