Visiting the Epistemic Games Group

I’m currently spending some time at University of Wisconsin-Madison with the Epistemic Games group there.  Some very brief (i.e. they probably won’t make much sense without knowing something about the epistemic games group) notes…

I’d read about, and seen demoed, the Epistemic Network Analysis (ENA) for exploring students commitments to particular: skills, knowledge, identity, values and epistemology (SKIVE).  Until today I hadn’t actually seen the games though.

Today I had the chance to view two games – the Landscience game, and Nephrotex – both engineering games, the former targeted at high school aged students the latter undergraduates. The former involves a planning exercise working with multiple stakeholders to change zoning in a town, the latter involves developing some new medical technology and balancing 5 key metrics for assessing the quality of your output.

Epistemic Game Structure

Both games run on a similar model, with users progressing through ‘rooms’ as they are unlocked (via ’email’) by a mentor.  These rooms then form the basis both for structuring the learning environment (and support), and for breaking down data into analytic segments (stanzas) for research purposes.

Both games are collaborative – with users placed into teams at the start of the game, and later ‘jigsawed’ into other groups.  This is a nice approach to swap people up and make sure everyone is exposed to information from an area they haven’t closely looked at (in the landscience case a stakeholder group, in the nephrotex case a material), and also forces the students to engage with new information (and rely their own) throughout the game.

Both games involve mentors supporting students in their game play (by scaffolding their collaboration), and some ‘canned phrases’ are available for that purpose (and some auto-coding is being explored here).  Mentors also check the outputs of some of the exercises against a rubric to ensure the appropriate considerations are being made, and information retrieved.  In particular, students are required to engage in producing some product (a land map or a dialyzer), balancing requirements for it, and discussing the alternatives.  The notes, emails, products, and chat data produced all form part of the data for student outcomes and research purposes.

In both games one way to ensure students are learning and picking up on the sometimes implicit content of the discourse is through reflective activities, and the use of ‘canned phrases’ from the mentor to prompt students to consider particular aspects of their work…there’s interesting scope here  for autocoding on those written and ‘chat’ based reflections, in addition to autocoding other products produced.


Both rooms start with some introductory material, including some stuff on self-efficacy and prior knowledge (although the interest of the games is largely around conceptual change rather than content assessment).  They also both include a stage to define some personal characteristics – an avatar (Landscience) or a personnel page (Nephrotex) – possibly affording some chance to explore the ‘identity’ facet of SKIVE – how do participants in the game perceive their own identity in relation to the task (of being an engineer).


At the analysis stage, data are split into message level segments – chat utterances, emails, notes, etc.  I think there’s some segmenting into smaller portions (by punctuation), but it might also be interesting to explore for example, the use of subheadings to split up notes (particularly where these might have epistemic significance – “advantages”, “disadvantages”…).

In any case, these segments are then coded for the presence of keywords related to Skills, Knowledge (of the domain), Identity, Values, and Epistemology, with careful removal of some keyterms (e.g. “know that” is controlled for when looking for “now that”).  In each instance, the interest is not just did you get the node (a single SKIVE code on an utterance) but did you make the connection – e.g. an epistemology of some knowledge facet.

Multiple Document Processing Scope?

One of the things I was struck by when looking at the games was the scope for exploring the type of multiple document processing task I’ve blogged about a few times recently.  At a few rooms in the games students are asked to summarise documents in a notebook (and search sources on the web).

Rather than giving fact retrieval documents, it might be interesting to give a set of documents with some conflicting information, and varying document qualities (perhaps methods learnt about earlier in the game could be referred to, differing source qualities, varying corroboration of information across varying article qualities, etc.).

Print pagePDF pageEmail page

This Post Has 1 Comment

Leave A Reply

You must be logged in to post a comment.

%d bloggers like this: