After these years of development in the computer-based assessment (CBA) field, I have heard a lot of expectations from many institutions, … thanks to my customers and also thanks to FLIP+ association that gathers a lot of institutions involved in education assessment. All actors are searching for a better workflow, from the item conception to the test administration and the retrieval of results. All actors were faced with building specifications and had to choose between building their own assessment platform or using an existing product (open source or not). Once a solution is chosen, after a while, new needs appear and that is the moment of truth : how will the platform handle a new development? The first unpleasant surprise could be technical : the evolution was not or poorly planned and the modification will concern the platform code that could destabilize the whole assessment solution. Even if the platform has a modular architecture, the surprise could be the cost of the plugin, the extension, or the module ? The last surprise is the development skill : because the solution is an integrated platform, new developers must learn the technical constraints… if documentation is available… If not, the institutions are completely dependent on the original developer team. That could be a problem.
Many assessment institutions are confronted with bringing their assessment tools to the classroom, as OECD did with “the PISA for School project”. It is a logical evolution of large assessment programs : sharing the assessment method with teachers to better prepare pupils to the test. How item banks and assessment platforms are ready for this new mission, should institutions have to recreate from scratch their process ?
Adding new functionality to your workflow concerns all the steps of a CBA process. Let us focus on the item building. Why ? Because it is the most complex part : item building is facing didacticians and teachers’ creativity. That is a real challenge…Or a nightmare for developers. IMS Global proposed to classify the main interactions, encoding the element that composes items (QTI). Their solution to keep open the interaction paradigm was to add to the standard the possibility to develop portable custom interaction (PCI).
Developing PCIs is a part of my professional activity, during years, I developed PCIs in TAO platform to add interactive animations, spreadsheets, blocks-based programming language (SNAP), chatroom simulation… and soon GeoGebra and 3D simulations.
Before the pandemic, I had the opportunity to work with NFER (National foundation for education and research from the UK) and Vretta (Toronto) on the basic interaction : choice, making pairs, sequencer… The idea was to help the item builder to create items with a lot more options than TAO item editor offers. The NFER team did a great job defining their expectations. With Vretta, we had to manage the technical context and that cost us a lot of time and effort !
During the first lockdown, I decided to revisit this project, and explore a quite different solution to answer the NFER initial need. I have decided to develop an independent item editor dedicated to “choice interaction” (named WonderChoice), which will offer the possibility to export item ZIP packages as a QTI or a PCI format. The export process will be modular and could be enriched. As I defined my own technical context, it was easier to add functionalities and reinvent the item editor process.
My first direction was to use a graphic editor, the idea was to recreate a design environment (drawing software / photo editor) with the possibility to add, drag and drop images and text. Parameter panel, context menu, and toolbars help the item builder to format and organize the content as she/he wants. I have integrated the usual features : multichoice, single choice, feedback, hint, clickable image or text, shuffling, choice limiter, elimination mode, chronometer and count down. Concerning the score, the user can define : the right answer, the choice to eliminate, a penalty if chosen. A preview is possible and shows exactly the item in a test context.
My second direction was to create the most efficient workflow. With the “rich editor”, the user can paste formatted content (from Microsoft Word for example) and defines a list to automatically inject all choices to the future item. Once the user is satisfied, he can add it to the item package and start to create another item. But he keeps the control on his work, it is still possible to see and edit an item from the package to change a detail and check all the options and metadata.
The quickest way to produce items is the use of the #code in a text editor (tags for each component). Copy and paste the text to the rich editor and press the parser button. It automatically generates the item package (it can handle 150 items and more in one click). The user can of course edit items to adjust their presentation.
I’m not going any further in the description of WonderChoice, a documentation is available on the website for more details.
You can try this application here: https://wondertest.org/wonderChoice.
WonderChoice is a project under development, I’m looking for funding and testers of course but also, I’m looking for simple collaboration about the project about ideas and needs.
In conclusion, I wanted to insist on the fact that external item generators could be a fantastic solution to add completely new items to an assessment platform. Respecting the PCI format will facilitate the work, but we can also imagine many ways to import this kind of item to any platform that will have a modular approach in their item editor system. So, keep in mind this important point If you are writing some specifications for your future assessment platform.
Toulouse France, 01/26/22
Jean-Philippe Rivière, www.wiquid.fr
jpriviere@wiquid.fr