Janet Read (et al)'s Endurability, Engagement and Expectations: Measuring Children's Fun provides an interesting and empirically grounded perspective on the nature of fun...
The paper abstract introduces:
...a toolkit for measuring fun with children aged between 5 and 10. The relationship of fun to usability is discussed and three dimensions of fun; Endurability, Engagement, and Expectations are described. A set of tools that can be used in empirical studies to measure these dimensions of fun is described in detail.
First, a caveat boldly stated in the paper's introduction: adult usability does not equate to children's fun- the underlying motivations differ. Janet's approach revolves around methods and measurements in three dimensions (emphasis added):
As an experience that the user has of a system, fun can be considered to have three dimensions. These dimensions have been previously identified as Expectations, Engagement, and Endurability (Read & MacFarlane, 2000). These dimensions are derived from expectancy theory, user expression of frustration / satisfaction and the Pollyanna principle.
Expectations would appear to measure the difference between the reported and expected fun experiences of children. Engagement is grounded in observed reactions. Endurability has two components: Remembrance, the likelihood of recall an enjoyable activity; and Returnance - the desire to repeat an enjoyable activity. The paper also describes a number of tools and approaches for measuring points in this space, all of them simple, but a few seem subtle ( e.g. Fun-Sorter).
There is always a danger reading papers outside of one's field, especially when they involve clincal work. Nonetheless, the paper's discussion stimulates thinking:
1.) towards an empirical approach to evaluating "fun"
2.) measuring expectations as part of the evaluation process
3.) counter-weighting shorter-term engagement reactions versus longer-term endurability impacts.
Fresh from our recent discussion of the Theory of Fun (ToF), might we consider how high level patterns of practice suggested in game design centric descriptions can mesh with more direct measures of individual experiences? "Patterns of practice" may serve well as a source of guiding principles for shaping reading of a lower level system of measures, yet these measures, whatever they turn out to be, may provide the sharper edges to our understanding.
Their approach was experimental and therefore note really grounded, but I suppose you could do a grounded study using either experience sampling or thinking aloud.
Posted by: Ola Fosheim Grøstad | Feb 06, 2005 at 10:10
Nathan, I think this post is misleading. The paper at hand appears interested in developing tools for evaluating fun, not "...an interesting and empirically grounded perspective on the nature of fun..."
I can personally attest to the difficulty of reading papers outside one's field (I have no proper field), but it seems to me that the intent of this paper is fairly clear, even if one were only to read the abstract; the focus is clearly on tools. "A discussion of the efficacy of the tools presents the findings that there is a sufficient subset that can be used to give a reliable measuer of fun. Dependent on the nature of the event being evaluated and on the number of activities being compared; the subset will be differently constructed." If one reads just a bit further, it's hard to avoid This paper presents a 'toolkit' for measuring fun with young children.
The authors cite a model of fun they (sans Casey) presented in 2000; "...derived from expectancy theory, user expression of frustration / satisfaction and the Pollyanna principle." Maybe that's the paper to take a look at. :)
Or have I misunderstood?
Posted by: Isaac | Feb 07, 2005 at 02:36
Ola>
Their approach was experimental and therefore note really grounded,
Isaac>
Nathan, I think this post is misleading. The paper at hand appears interested in developing tools for evaluating fun, not "...an interesting and empirically grounded perspective on the nature of fun..."
...
This paper presents a 'toolkit' for measuring fun with young children.
Point taken - this paper wasn't proposing a model of fun. I wonder though, per the Q of my last paragraph, whether empirical measurements, ala some 'toolkit' could be used to develop/validate game-design centric models/notions of, well, fun?
Posted by: Nathan Combs | Feb 07, 2005 at 07:23
I think measurement of fun in a multi-user context is going to be difficult as having our "identity" confirmed by others most likely is a very important source of adult "fun". Since no metric can cover such issues in a reliable way we have to rule out the whole "identity" aspect...
In order to understand a measurement we also need a baseline to compare against. If we are designing a game we should try to avoid comparing apples and oranges, so how do we obtain the baseline? Maybe the more interesting application would be, not in game design, but in game balancing. Then we do have an opportunity to get a decent baseline (initial design) to compare against and it would be interesting to measure dimensions of "fun" and "unfun" (frustration, confusion, helplessness). I suspect that "unfun" would be easier to measure, though.
Posted by: Ola Fosheim Grøstad | Feb 07, 2005 at 08:19
I suspect that "unfun" would be easier to measure, though.
Ola,
On a related note: "Rude Software Causes Emotional Trauma" at http://www.theregister.co.uk/2005/02/07/rude_software/ is an interesting bit of info...
Posted by: Andres Ferraro | Feb 07, 2005 at 12:37
Andres, that was a very fun news item!
So, may that's what Microsoft's paper clip is good for: healing our pain by always being so darn inclusive... ;)
Posted by: Ola Fosheim Grøstad | Feb 07, 2005 at 13:03
For sure, Nathan. The paper is a good read and certainly makes me interested in following up. It's likely that any comprehensive theory of "fun" will be quite a long way off, but tools and functional models are important steps. A one researcher confided in me just this afternoon, "...this is why I went into the cognitive sciences; It's one of the few fields left where you are free to theorize and it's almost impossible for anyone to refute with any semblance of actual proof."
Posted by: Isaac | Feb 07, 2005 at 18:25
It's likely that any comprehensive theory of "fun" will be quite a long way off,
I wonder if there is an easier first step.
In linguistics is the distinction between descriptive adequacy vs. explanatory adequacy - (Chomsky) as two different levels of accountability for evaluating grammars as models of language. The former need only account for the data, the former needs to also imply something about the underlying properties/mechanisms of mind.
By analogy... "a descriptively adequate model of fun" would be a model to account for fun where it occurs, whereas an "explanatorily adquate model of fun" would provide a deeper understanding into the cognitive properties of fun etc...
I'd settle for the former...
Posted by: Nathan Combs | Feb 07, 2005 at 19:17
The danger is that you wind up with a test akin to IQ tests, which despite their name, probably don't really measure Intelligence. They measure something akin to it, and they might be useful, but there's far more to Intelligence than one's IQ and there could be a lot more to "fun" than these tools measure. I guess what I'm arguing is they may be descriptively adequate of something, but whatever that something is, it might not be "fun".
Bruce
Posted by: Bruce Woodcock | Feb 08, 2005 at 05:52
Yeah, that's what I meant to indicate by "important steps." I think most of us (designers) can settle for an approximate model — we have little choice for the foreseeable future. But I also favor Churchland's notion that we shouldn't reason from models contradicted by neuroscience. With the current resolution of this knowledge being so low, many useful "descriptively adquate" models can develop without threat.
This is not to say that convenience is big on my list of reasons to remain satisfied personally. :)
Posted by: Isaac | Feb 08, 2005 at 06:12
Hi all,
Here is a workshop that might be of interest to several in this blog:
Workshop on "Adaptive Approaches for Optimizing Player Satisfaction in Computer and
Physical Games" in conjunction with SAB'06: From Animals to Animats
9,(http://www.sab06.org), October 1st, Rome, Italy.
For further details, please visit the workshop's web-page at:
http://www.mip.sdu.dk/~georgios/gamesWorkshop
Posted by: Georgios Yannakakis | Mar 01, 2006 at 05:08